Wildfowl Special Issue No. 4

“Ecology and Conservation of Waterfowl in the Northern Hemisphere”

Proceedings of the 6th North American Duck Symposium and Workshop Memphis, Tennessee, 27–31 January 2013

Editors Eileen Rees, Rick Kaminski & Lisa Webb

Editorial Board Brian Davis, Mike Eichholz, Gary Hepp, Rick Kaminski, Dave Koons, Tom Nudds, Jim Sedinger & Lisa Webb

Published by Wildfowl & Wetlands Trust Slimbridge, Gloucestershire GL2 7BT, UK Registered Charity No. 1030884 Wildfowl

Editor Dr Eileen C. Rees Wildfowl & Wetlands Trust Slimbridge, Gloucestershire GL2 7BT, UK.

Associate Editor Prof Anthony D. Fox Department of Bioscience, Aarhus University Kalø, Grenåvej 14, DK-8410 Rønde, Denmark.

Editorial Board Prof Jeff Black Prof Andy J. Green Department of Wildlife Wetland Ecology Department Humboldt State University Doñana Biological Station-CSIC Arcata, California 95521, USA. C/ Américo Vespucio s/n 41092 Sevilla, Spain. Assoc Prof Bruce D. Dugger Department of Fisheries & Wildlife Dr Matthieu Guillemain Oregon State University Office National de la Chasse et de la Faune 104 Nash Hall, Corvallis, OR 97331, USA. Sauvage La Tour du Valat Le Sambuc, 13200 Arles, France. Cover photographs: King Eider, by Danny Ellinger/Minden Pictures/FLPA; Bewick’s Swan and White-fronted Goose, by Paul Marshall/WWT. The cover images were selected to represent a duck, goose and swan species of the northern hemisphere, consistent with the title of this special issue of Wildfowl. Cover design by Paul Marshall Published by the Wildfowl & Wetlands Trust, Slimbridge, Gloucestershire GL2 7BT, UK Wildfowl is available by subscription from the above address. For further information call +44 (0)1453 891900 (extension 257), or e-mail [email protected] ISBN 0 900806 65 6 Print ISSN 0954-6324 On-line ISSN 2052-6458 Printed on FSC compliant paper by Berforts Information Press Ltd. Recommended citation: Rees, E.C., Kaminski, R.M. & Webb, E.B. (eds.), 2014. Ecology and Conservation of Waterfowl in the Northern Hemisphere. Wildfowl (Special Issue No. 4). Contents

Dedication 1 Dr. Guy A. Baldassarre Rick Kaminski

Foreword 2 Bruce Batt

Preface 5 R.M. Kaminski & J.B. Davis

PLENARY PAPERS Habitat use and selection An introduction to habitat use and selection by waterfowl in the northern hemisphere 9 R.M. Kaminski & J. Elmberg

Habitat and resource use by waterfowl in the northern hemisphere in autumn and 17 winter J.B. Davis, M. Guillemain, R.M. Kaminski, C. Arzel, J.M. Eadie & E.C. Rees

Spring migration of waterfowl in the northern hemisphere: a conservation 70 perspective J.D. Stafford, A.K. Janke, M.J. Anteau, A.T. Pearse, A.D. Fox, J. Elmberg, J.N. Straub, M.W. Eichholz & C. Arzel

Nest site selection by Holarctic waterfowl: a multi-level review 86 M. Eichholz & J. Elmberg

Waterfowl habitat use and selection during the remigial moult period in the northern 131 hemisphere A.D. Fox, P.L. Flint, W.L. Hohman & J-P.L. Savard

Demography, cross-seasonality and integrated population management Drivers of waterfowl population dynamics: from teal to swans 169 D.N. Koons, G. Gunnarsson, J.A. Schmutz & J.J. Rotella Conspecific brood parasitism in waterfowl and cues parasites use 192 H. Pöysä, J.M. Eadie & B.E. Lyon

The effects of harvest on waterfowl populations 220 E.G. Cooch, M. Guillemain, G.S. Boomer, J-D. Lebreton & J.D. Nichols

Cross-seasonal effects and the dynamics of waterfowl populations 277 J.S. Sedinger & R.T. Alisauskas

Managing harvest and habitat as integrated components 305 E.E. Osnas, M.C. Runge, B.J. Mattsson, J. Austin, G.S. Boomer, R.G. Clark, P. Devers, J.M. Eadie, E.V. Lonsdorf & B.G. Tavernia

SPECIAL SESSION PAPERS AND EVALUATION OF THE SYMPOSIUM Ecology and Conservation of North American Waterfowl Implementing the 2012 North American Waterfowl Management Plan: people 329 conserving waterfowl and wetlands D.D. Humburg & M.G. Anderson

Wetland issues affecting waterfowl conservation in North America 343 H.M. Hagy, S.C. Yaich, J.W. Simpson, E. Carrera, D.A. Haukos, W.C. Johnson, C.R. Loesch, F.A. Reid, S.E. Stephens, R.W. Tiner, B.A. Werner & G.S. Yarris

Opportunities and challenges to waterfowl habitat conservation on private land 368 W.L. Hohman, E. Lindstrom, B.S. Rashford & J.H. Devries

Estimating habitat carrying capacity for migrating and wintering waterfowl: 407 considerations, pitfalls and improvements C.K. Williams, B.D. Dugger, M.G. Brasher, J.M. Coluccy, D.M. Cramer, J.M. Eadie, M.J. Gray, H.M. Hagy, M. Livolsi, S.R. McWilliams, M. Petrie, G.J. Soulliere, J.M. Tirpak & E.B. Webb

Annual variation in food densities and factors affecting wetland use by waterfowl in 436 the Mississippi Alluvial Valley H.M. Hagy, J.N. Straub, M.L. Schummer & R.M. Kaminski Atmospheric teleconnections and Eurasian snow cover as predictors of a weather 451 severity index in relation to Mallard Anas platyrhynchos autumn–winter migration M.L. Schummer, J. Cohen, R.M. Kaminski, M.E. Brown & C.L. Wax

Waterfowl populations of conservation concern: learning from diverse challenges, 470 models and conservation strategies J. Austin, S. Slattery & R.G. Clark

Waterfowl in Cuba: current status and distribution 498 P.B. Rodríquez, F.J. Vilella & B.S. Oria

Assessment of the 6th North American Duck Symposium (2013): “Ecology and 512 Conservation of North American Waterfowl” L.P. LaBorde, Jr., R.M. Kaminski & J.B. Davis Sponsors of NADS 6/ECNAW and this Publication

Designed by Justyn Foth 1

Dedication

Unfortunately and sadly, there comes a time in life when it ends despite individual desires to continue onward and to contribute professionally and personally. Dr. Guy A. Baldassarre surrendered life on 20 August 2012 at the age of 59 years after a courageous battle against leukaemia (Moorman 2013, The Auk 130: 194–195). The 6th North American Duck Symposium and Workshop (NADS 6) and this special issue of Wildfowl are dedicated to Guy for his decades of contributions to waterfowl science and conservation. Guy inspired countless people internationally through his teaching, research, writing, mentoring, service, humour and friendships. He had an extraordinary ability to amass and synthesise technical information on waterfowl and wetlands into comprehensive texts of international acclaim; most notably the two editions of Waterfowl Ecology and Management (1994, 2006; with Eric G. Bolen) and Ducks, Geese, and Swans of North America (2014). These treatises will serve as an information resource on waterfowl, and thus our profession, indefinitely. Indeed, Guy so wanted to attend NADS 6 to learn, visit with colleagues and students, and autograph copies of his books, but that trip and symposium were not part of his destiny. We thank Guy for his dedication to family, whom he considered first and foremost in his life, the wildlife profession, humanity, waterfowl, and all the birds he loved so dearly. Guy’s contributions are timeless; his legacy lives on among all who knew this gentleman!

Rick Kaminski

Photograph: Guy A. Baldassarre 1953–2012, by Eileen Baldassarre.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 1–2 2

Foreword

“What a great idea.” That was my immediate reaction when Rick Kaminski told me about collaboration between the Wildfowl and Wetlands Trust (WWT) in the United Kingdom and the team that conducted the 6th North American Duck Symposium and Workshop (NADS 6), Ecology and Conservation of North American Waterfowl (ECNAW). The event occurred in Memphis, Tennessee in January 2013. A key part of that great idea was to broaden the content and title of this publication to include waterfowl from across the Northern Hemisphere. Indeed, since 1997, the NADS have stimulated intercontinental relationships and efforts among colleagues to understand, conserve and sustain waterfowl and their habitats throughout the northern part of the globe. The NADS 6/ECNAW symposium was built on a rich history of progress in waterfowl and wetlands research and conservation, as is nicely explained in the Preface to follow. Briefly, NADS 6/ECNAW combined efforts of three groups of specialists that had previously established separate conferences in North America: the NADS, the North American Arctic Goose Conference, and the International Sea Duck Conference. Each separate gathering had assembled many of the world’s specialists to review advancements in waterfowl and wetlands science; this progress has helped steer ongoing management, conservation and future scientific efforts for their focal waterfowl taxa. These conferences each had an excellent reputation as a “must attend” event for academic, government, and non-government scientists and students. Thus, NADS 6/ECNAW was assured to be an unprecedented assemblage of the world’s waterfowl specialists and, from the level of knowledge exchanged, this proved to be the case for all those attending the meeting. This special issue of Wildfowl is the first published proceedings from any NADS. The specialists in the field will appreciate the effort required to conduct a symposium of this magnitude and then follow with a peer-reviewed proceedings. This outcome was accomplished through an extraordinary level of engagement by all who were involved and they are acknowledged in the following Preface. All those folks also recognise that Rick Kaminski’s leadership was the driving force of the symposium and publication of this volume – a vision he has pursued for over half a decade. Dedication of the symposium to the memory and contributions of Guy Baldassarre was an inescapable decision in view of all participants’ recognition of Guy’s huge lifetime commitments to waterfowl science, conservation and education. His untimely death in 2012 was an overwhelmingly sad event. Guy had made personal contact and was friends with a very large portion of those in attendance at NADS 6/ECNAW, as a result of his superb career. He wrote, with Eric Bolen, the 1994 original and the 2006 second edition of Waterfowl Ecology and Management – the waterfowl profession’s first comprehensive text book. Many conference participants have a tag-eared edition of both treatises on their reference shelf. I cherish the personalised copy of the second edition. Not one to seek praise or credit, Guy would nevertheless surely have been proud of the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 2–4 3 conference and this proceedings – for a moment at least, before he sought conversation with whoever was willing to talk about waterfowl and the latest findings from new studies, especially if the research resulted from students’ efforts. Indeed, Guy had planned to attend the symposium and sign copies of his monumental 2014 revision of “Ducks, Geese, and Swans of North America,” as he did for his and Eric’s book at previous NADS. Francis Kortright and Frank Bellrose, the first and second authors of “Ducks, Geese, and Swans of North America,” respectively, would be thankful to Guy for bringing new life to this classic waterfowl treatise. The WWT brings its own record of excellence in waterfowl research and conservation that extends historically back to the 1940s. They originally focused on the status of waterfowl in Great Britain, on introducing the general public to the beauty of waterfowl, and on emphasising the need for conservation to assure the future of these remarkable species. Since then they have established a singular position of leadership in a multitude of waterfowl and wetlands conservation matters, extending from their local education centers and across the world through engagement in many venues (including the United Nations) and the development of international research and conservation projects, a record that continues to mount to this day. For most North American waterfowl biologists, arrival of the annual Wildfowl Trust Report (which evolved into Wildfowl) was a highlight that provided a window for many subjects that extend from propagation of endangered species to detailed analyses of courtship behaviour and to population inventories of waterfowl throughout the world. The WWT’s record of accomplishment is unequalled. I urge the world’s waterfowl enthusiasts to make the WWT Centres in England, Scotland, Wales and Northern Ireland high priority destinations if they ever find themselves in the United Kingdom. The WWT actually engaged in North American waterfowl conservation from their earliest years of population monitoring, because a portion of Light-bellied Brent Geese, known as Atlantic Brant in North America, fly across the Atlantic Ocean from breeding areas in eastern Arctic Canada to winter in coastal Ireland. My predisposition is to label these geese as “North American” because that is where they breed. I suspect WWT thinks of them as European geese that migrate to Canada to breed before they return “home” to Ireland. Both perceptions are correct and augur well for the future of these birds. On a broader basis, it is critical that people throughout the flyways of waterfowl assume the responsibility of caring for the birds and the places in which they live whether they are breeding, migrating or wintering. Also, Sir Peter Scott, the founder of WWT, conducted the first ever field investigation of the then endangered Ross’ Goose in the central arctic of Canada in 1949. The conservation of the entire world’s waterfowl clearly has been their mission from the onset. Collaborations have since been much more active and ongoing, so this publication should be seen as an important example of the richness of today’s pattern of international cooperation and accomplishment in waterfowl science and conservation. The life histories of most northern hemispheric waterfowl encompass annual passages that extend between northern and southern latitudes and breeding and wintering habitats. The

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 2–4 4 ecology of waterfowl throughout their annual cycle and the occurrence of cross-seasonal, carry-over effects on their survival and recruitment are now a core paradigm guiding waterfowl science, conservation and management. However, few researchers are able to work closely with the birds year round amongst latitudes. Bringing students, managers and scientists together to share and review their findings at gatherings like NADS is one of the most significant benefits to the birds and to the scientific basis for their future. This partnership between WWT and NADS in the production of these proceedings is a natural and welcome step in the continued quest for intelligent and effective application of science to the future understanding of waterfowl ecology and management in the northern hemisphere.

Bruce Batt Chief Biologist (retired), Ducks Unlimited, Inc.

Photograph: Lesser Snow Goose (blue morph) on Wrangel Island, Russia, by Sergey Gorshkov.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 2–4 5

Wildfowl Special Issue No. 4: Preface

This special issue of Wildfowl is the proceedings from the 6th North American Duck Symposium and Workshop (NADS). Archives of NADS 1–5 are filed at http://www.northamericanducksymposium.org/index.cfm?page=home. Because this special issue is the first NADS proceedings to be published in a scientific journal, a brief history of NADS seemed warranted, given that relatively little has been documented on the development of NADS to date. Additionally, this preface reports the goals and themes of NADS 6, the contents of the symposium, and acknowledges people who aided or sponsored NADS 6 and this special issue. Being inspired by the North American Arctic Goose Conference in the 1990s, the conceptual founders of NADS – Alan Afton of the Louisiana Cooperative Fish and Wildlife Research Unit and Robert Helm of the Louisiana Department of Wildlife and Fisheries – envisioned a periodic forum (c. every three years) for waterfowl biologists, managers, researchers, conservationists and especially students to present and discuss current research and management related to ducks worldwide but with an emphasis on North America. They and particularly Dave Ankney, Michael Johnson and Bob McLandress pioneered NADS. Collectively, they believed that research questions and management issues related to sustaining duck populations and maintaining wildfowling traditions and ecological study of the birds in North America were of paramount importance. Moreover, they believed that waterfowl ecologists were among leaders in avian research and sought to ensure that this legacy was perpetuated, in part by creating a venue for discussion of topics significant to waterfowl and their habitats. Therefore, NADS was established in 1997 as an independent meeting free of agency and organisational politics and charged the scientific and local organisational committee of each subsequent symposium with authority to ensure that no group could use the forum for promoting personal agendas without consideration of either alternative viewpoints or scientific-based resolutions to management or other issues. In 2009, NADS, Inc. was created to formalise the organisation, obtain non-profit status, and ensure that NADS continued in perpetuity. The primary mission of NADS, Inc. is to advance the science, management and outreach that guide waterfowl conservation in North America and beyond by convening symposia at regular intervals to present and discuss information related to ducks and other waterfowl. The stated vision of NADS, Inc. is that well-trained and informed educators, researchers and managers will promulgate effective strategies to sustain duck habitats and populations for scientific study, wildfowling, observation, ecosystem services and conservation, and other ecological and societal benefits. Six NADS have convened since inception: NADS 1 at Baton Rouge, Louisiana (1997); 2 at Saskatoon, Saskatchewan (2000); 3 at Sacramento, California (2003); 4 at Bismarck, North Dakota, (2006); 5 at Toronto, Ontario, (2009); and 6 at Memphis, Tennessee (2013). NADS 7 is planned for February 2016 in Annapolis, Maryland, the first NADS to assemble in the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 5–8 6

Atlantic Flyway. When possible, symposium locations were alternated between the United States and Canada and among Flyways in North America. The NADS 6 organisational committee and the board of directors of NADS, Inc. agreed in 2009 at NADS 5 that NADS 6 would include all taxa of Anatidae (i.e. ducks, geese and swans). Thus, the scientific committee dubbed NADS 6 as “Ecology and Conservation of North American Waterfowl (ECNAW)”. Consistent with this taxonomic expansion to all waterfowl, the committee invited the North American Arctic Goose Conference and the International Sea Duck Conference as joint partners of NADS 6/ECNAW (http://www.northamericanducksymposium.org/). In 2010, a 20-member scientific planning committee was formed with representation from universities, agencies and organisations across North America and Europe with expertise in ecology and conservation of ducks, geese and swans (http://www.northamericanducksymposium.org/index.cfm?page=committees). Richard M. Kaminski and J. Brian Davis (both of Mississippi State University) served as co-chairs for local planning, fund raising and logistics of NADS 6/ECNAW, hosting the event at the Peabody Hotel in Memphis, Tennessee, from 27–31 January 2013. A total of 450 conferees attended NADS 6/ECNAW, and a majority of those attending responded to a survey to evaluate the symposium (see Laborde et al. 2014 in this volume). The grand theme for NADS 6/ECNAW was “Science and Conservation: Sustaining Waterfowl Forever.” Although this theme may seem grandiose, it provided an ageless, challenging and inspiring goal for waterfowl scientists and stewards presently and into the future. The scientific committee’s first challenge was to identify broad, prominent topics for plenary sessions consistent with the mission of NADS. By vote, the committee selected the following topics in sequential order occurring Monday–Thursday (28–31 January 2013) of the symposium: 1) habitat use and selection; 2) annual-cycle and biological carry-over effects; 3) life-history strategies and fitness; and 4) population and community ecology and dynamics. These topics were chosen because the committee believed use of habitats and intrinsic resources by waterfowl, relative to myriad exogenous influences, shape annual-cycle and life- history adaptations of individuals and ultimately influence biological outcomes for individuals, populations and communities. Leaders of the plenary sessions were also selected by vote; they and invited presenters represented colleagues with worldwide reputations in the topical areas of the plenaries. In addition to the morning plenary sessions, afternoons of the symposium were filled with concurrent contributed and special sessions wherein professionals and students made oral presentations. There were numerous presentations that spanned theoretical and applied ecology and conservation of waterfowl and wetlands in North America and Europe. Abstracts from presentations are archived at http://www.northamericanducksymposium.org/ index.cfm?page=agenda. The final afternoon of the symposium included a novel “Syntheses and Futures” session wherein senior and junior colleagues presented thoughts on major issues revealed during the symposium and visions for future research and conservation to sustain and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 5–8 7 advance science and conservation of northern hemispheric waterfowl. Many of these progressive notions are presented in the articles in this special issue. Moreover, there were two evening poster sessions and a special session for student mentees to meet and interact with professionals. A special session of the North American Waterfowl Management Plan 2012 (NAWMP) also convened before the official opening of the symposium to discuss implementation and advancement of the revised plan (http://nawmprevision.org/). A synthesis paper of NAWMP 2012 is included in this volume (see Humburg & Anderson 2014). From the wide array of presentations, the scientific committee deliberated and concluded that this volume would be composed of manuscripts from plenary and special sessions of the symposium. As mentioned, the committee selected broad topics for plenary sessions because of their fundamental importance to individual survival, reproductive performance and fitness, as well as how individual biological outcomes impact collectively on vital rates, population dynamics and community ecology of waterfowl. Indeed, this sequential acquisition of knowledge from individual to population and community levels is essential for holistic understanding of waterfowl ecology and guiding effective management and conservation for abundant and rare species, populations, and communities of waterfowl worldwide. Additionally, the committee did not envision the proceedings to be a mere compilation of selected “souvenir” papers from NADS 6/ECNAW but a contemporary compendium of knowledge related to waterfowl and their habitats in the northern hemisphere. Not since 1987 (in Winnipeg, Manitoba, Canada), have scientists, students and managers of waterfowl assembled for an international symposium focused on all waterfowl. The Winnipeg symposium was followed by a seminal publication entitled “Ecology and Management of Breeding Waterfowl” (Batt et al. 1992, University of Minnesota Press), which synthesised knowledge of breeding waterfowl ecology and management from the 20th century. Since this conference and publication, ecologists have greatly advanced understanding of waterfowl ecology and conservation throughout the birds’ annual cycle and range in the northern hemisphere (e.g. Baldassarre & Bolen 2006 in “Waterfowl Ecology and Management”, Krieger Publishing Company; Rees et al. 2009, Wildfowl Special Issue 3; and Baldassarre 2014 in “Ducks, Geese and Swans of North America”, The John Hopkins University Press). Thus, NADS 6/ECNAW served as a cornerstone for this special issue with its own primary goals: 1) synthesising classical and contemporary information related to waterfowl ecology and conservation throughout the northern hemisphere and the birds’ annual cycle and range; 2) comparing this knowledge across taxa of Anatidae with diverse habitat use, annual ecologies and cross-seasonal carry- over effects, life-history traits and fitness strategies, and population and community ecologies, for the purpose of archiving current knowledge from species to communities and enabling cross-taxa generalisations; and 3) using science-based information from the first two initiatives to adapt and advance local, regional, and intra- and intercontinental management and conservation of waterfowl (e.g. NAWMP 2012). Fulfilment of these goals, coupled with presentation here of papers from special sessions at NADS 6/ECNAW, are intended to make significant contributions toward understanding and sustaining northern hemispheric

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 5–8 8 waterfowl and other migratory birds during and after the 21st century. Moreover, because ecologists studying waterfowl have contributed greatly to understanding the ecology of migratory avifauna since the 20th century, this special issue should be useful to ornithologists in general and provide a 21st century standard for guiding and advancing science and conservation for waterfowl and other migratory birds and their habitats in the northern hemisphere. Indeed, we hope our mission has been accomplished. Clearly, our mission would not have been successful without the dedicated efforts of all leaders of the plenary and special sessions and especially the authors contributing manuscripts to this special issue of Wildfowl, the 20-member scientific planning committee for NADS 6/ECNAW, the co-editors of this volume (Eileen Rees, Rick Kaminski and Lisa Webb), the editorial committee (Brian Davis, Mike Eichholz, Gary Hepp, Rick Kaminski, Dave Koons, Tom Nudds, Jim Sedinger and Lisa Webb), the many external peer reviewers of manuscripts, Bruce Batt for penning the Foreword, Jeanne Jones for drawing the logo for NADS 6/ECNAW, the local logistics committee for the symposium (Amy Alford, Bruce Batt, Karen Brasher, J. Brian Davis, Jim Feaga, Justyn Foth, Charlsie Halford, Steve Jones of the Mississippi State University Extension Service, Rick and Loretta Kaminski, Molly Kneece, Jennifer Kross, Joe Lancaster, Joe Marty, Kira Newcomb, Tom Peterson, Jessica Myers, Jessie Schmidt, Clay Shipes, Jake Straub, Lisa Webb and Matt Weegman), Laurie Grace and Justyn Foth for designing and producing wood plaques for student presentation award recipients, session moderators, student poster and oral presentation award judges, student organisers of the mentor/mentee session (Elizabeth St. James, Justyn Foth, Jessi Tapp, David Messmer, Matt and Mitch Weegman), the student mentors, vendors, and the Peabody Hotel staff – especially Shannon Williams and Betsy Wilson. We also sincerely thank Eileen Rees (Editor- in-Chief of Wildfowl) and the Wildfowl & Wetlands Trust for accepting our request to publish the proceedings in Wildfowl, and her editorial staff for reviewing and editing the manuscripts herein. Moreover, we thank the NADS, Inc. Board of Directors who assisted with and supported NADS 6/ECNAW (Brian Davis, John Eadie, Michael Johnson, Rick Kaminski, Tom Nudds, Scott Petrie, Ron Reynolds, Mike Szymanski and Dan Yparraguirre). Rick Kaminski extends sincere thanks to George Hopper, Dean and Director of the College of Forest Resources and the Forest & Wildlife Research Center, Mississippi State University, for providing fiscal and moral support before, during, and after the symposium and for allowing him time to focus work on completion of this special issue. Lastly but not least, we are deeply indebted to the sponsors and conferees who defrayed costs of NADS 6/ECNAW and publication of this special issue (see sponsors’ logo page). This international, premiere symposium and publication are due largely to your support. If we have omitted anyone deserving acknowledgment, we accept full responsibility for the non- intentional oversight and extend our sincere thanks to you now.

Rick Kaminski & Brian Davis (Local co-organisers of NADS 6/ECNAW)

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 5–8 9

An introduction to habitat use and selection by waterfowl in the northern hemisphere

RICHARD M. KAMINSKI1* & JOHAN ELMBERG2

1Department of Wildlife, Fisheries and Aquaculture, Mississippi State University, Mississippi State, Mississippi 39759, USA. 2Division of Natural Sciences, Kristianstad University, SE-291 88, Kristianstad, Sweden. *Correspondence author. E-mail: [email protected]

Abstract This introductory article aims to provide a theoretical framework to the topics of habitat use and selection by waterfowl (i.e. family Anatidae) in the northern hemisphere during the four stages of their annual cycle: autumn migration and winter, spring migration and pre-breeding, nesting and brood rearing, and post- breeding and moulting. Papers addressing each of these seasonal sectors of the annual cycle, which follow this introduction, were presented at the 6th North American Duck Symposium, “Ecology and Conservation of North American Waterfowl” in Memphis, Tennessee in January 2013. Here, we consider the theory and selected empirical evidence relevant to waterfowl habitat and resource use and selection that may affect individual survival and fitness of waterfowl in Nearctic and Palearctic ecozones. Additionally, where possible, a comparative taxonomic approach is attempted in the following papers to identify and generalise patterns in habitat and resource use and selection across waterfowl taxa that may influence biological outcomes for individuals, populations and species through space and time. Each of the subsequent papers use accumulated science-based information to recommend future opportunities and strategies for research and for habitat and population conservation. Collectively, our goals in synthesising information on waterfowl are to help sustain harvestable populations of waterfowl and to protect rare species amid worldwide changes in climate, landscape, economics, socio-politics and growth of human populations.

Key words: Anatidae, annual cycle, biological outcome, conservation, fitness, habitat, habitat use, habitat and resource selection, management, survival, waterfowl.

Habitat is a principal unifying entity exist in the absence of habitat and and concept in wildlife ecology and associated resources (Block & Brennan conservation, because wildlife would not 1993). The English word “habitat” is

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 10 Habitat use and selection by waterfowl derived from the Latin word “habito,” simplest and most common measure of meaning “to live or inhabit.” Herein, we habitat use by individuals of a species is define habitat as environmental space their occurrence within habitats. Habitat occupied by living individuals of a species associations (or correlations) are for any amount of time during their life demonstrated quantitatively when animals’ cycle and which contains resources for presence or abundance varies with measured individual survival and reproduction (Block habitat features (Wiens 1976, 1985). & Brennan 1993; Jones 2001). All waterfowl However, habitat associations should not be species (i.e. ducks, geese and swans: construed as habitat or resource selection, Anatidae) of the northern hemisphere use because selection is influenced by availability three-dimensional habitat space (i.e. aerial, including accessibility and procurement terrestrial and aquatic habitats) and the (hereafter, availability) of habitats and resources therein for physiological associated resources (Block & Brennan maintenance during their annual cycle and 1993; Garshelis 2000; Jones 2001). Although ultimately to gain fitness (i.e. individual unlikely to happen in nature, Fretwell and survival and reproductive success resulting Lucas (1970), Fretwell (1972) and Wiens in genetic representation in subsequent (1976, 1977, 1985) hypothesised that “true” generations; Mayr 1970). Generally, use of habitat selection occurs when animals use only two-dimensional space (terrestrial and habitats disproportionately to their aquatic) by waterfowl has been quantified availability and resource quality without using modern analyses and technologies constraints from extrinsic factors such as (Belant et al. 2012; cf. O’Neal et al. 2010). predation, competition, territoriality, density Nonetheless, much literature exists which dependence, anthropogenic effects or a provides evidence that habitats and combination of these and other factors. associated resources impose critical Likely, habitat selection has evolved through influences on individuals and evolution of increased immediate or short-term benefits adaptive strategies (Hildén 1965; Lack 1968; to individuals using specific habitats and Stearns 1976, 1992; Southwood 1977, 1988; resources, leading to long-term rewards in Kaminski & Weller 1992; Block & Brennan survival, reproduction, and fitness accrued 1993; Jones 2001). through natural selection and possibly other Block and Brennan (1993) and Jones agents of evolution (e.g. gene flow; Lack (2001) provided excellent reviews of habitat 1968; Wilson & Bossert 1971; Jones 2001). use and selection and related concepts, Demonstration of causal linkages between primarily based on the ornithological cross-seasonal and habitat-resource use and literature. To resolve and reduce future fitness is difficult, especially for migratory ambiguity in terminology, they clarified and waterfowl because of disconnects in daily to defined terms such as habitat, niche, habitat seasonal use of resources to meet selection (or preference) and habitat physiological and behavioural needs during suitability, and we have adopted their the annual cycle. Therefore, researchers terminology here for consistency. The generally infer habitat and resource selection

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 Habitat use and selection by waterfowl 11 when these are used disproportionately in philopatry to natal or other areas in their relation to their estimated availability annual range, and juvenile birds, without (Johnson 1980; Mulhern et al. 1985; prior migratory or dispersal experience, Alldredge & Ratti 1986; Jones 2001). For are likely influenced by co-occurring example, Mulhern et al. (1985) proposed and conspecifics or closely related taxa with tested three theoretical models of habitat similar niches (Brennan & Block 1993; use by Mallard Anas platyrhynhos and Elmberg et al. 1997; Thomson et al. 2003). Blue-winged Teal Anas discors in relation to Despite possible social effects on habitat habitat availability, based on structural and resource use, migratory waterfowl characteristics of used and available seemingly make an initial “first-order” wetlands: 1) no selection when use of selection of geographic regions, such as habitats “mapped” availability, 2) “plastic” those important to the birds during breeding selection when use was statistically different and non-breeding seasons (Johnson 1980; from available wetlands but use was Baldassarre & Bolen 2006). Within first- temporally dynamic, and 3) “stenotopic” (i.e. order occupied regions, waterfowl make specialistic or static) selection when use was “second-order” selections of wetland statistically different from available wetlands systems (Cowardin et al. 1979) and possibly but was consistent through time despite associated landscapes (e.g. nesting habitats or varying availability. Therefore, here and in arable fields for foraging). Next, they make subsequent papers in this section on “third-order” selections of local, site- seasonal habitat use, the term “habitat specific wetlands or other locations and selection” is used when referring to results finally “fourth-order” selections of of studies that have demonstrated microhabitats where individuals specifically disproportionate habitat or resource use, or may roost, nest, forage or engage in other relationships between one or both of these activities to acquire food or other resources and measures of biological outcomes (i.e. including mates (Wiens 1973; Johnson 1980; metrics of individual condition or Kaminski & Weller 1992, Baldassarre & performance (fitness correlates); Chalfoun Bolen 2006). Considering non-migratory & Martin 2007; Ayers et al. 2013). species that do not move seasonally among Habitat use and selection by migratory geographic regions (e.g. subtropical species; birds, such as most waterfowl, may be Mottled Duck Anas fulvigula), their local- envisioned as a multi-stage, hierarchical regional, annual home range and inclusive process from macro- to micro-spatial scales habitats constitute their “first-order” throughout the birds’ annual cycle and selected habitats, followed hierarchically by geographic range (Johnson 1980). However, second- to fourth-orders of selected habitats spatial distributions of waterfowl and other within their home range as described above gregarious animals may not be completely for migratory species (sensu Jones 2001). influenced by individuals or habitat and Additionally, the reversal of this process resources. For example, habitat use by male from micro- to macro-habitats also can be waterfowl is often influenced by female envisioned when birds depart micro-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 12 Habitat use and selection by waterfowl habitats to disperse or migrate to different level and individuals vary (e.g. Block & areas and habitats therein. Brennan 1993; Goss-Custard et al. 1995; Ideally, habitat use and selection should Stillman 2008). be investigated at all relevant spatio- Given that habitat quality or suitability temporal scales to identify the scale(s) at (sensu Fretwell & Lucas 1970) influences which possible limiting factors may have habitat and resource use, fitness and greatest impacts on survival, reproduction species’ life-history strategies, ecologists have and fitness. The outcome of habitat use and hypothesised and measured relationships selection analyses can be influenced by the between indices of habitat suitability (i.e. spatio-temporal scale of the investigation biological outcomes) and covariates predicted and associated environmental variation to influence these outcomes. Fretwell and (Wiens 1985). Moreover, habitat use and Lucas (1969) and Fretwell (1972) equated selection should be studied among suitability among habitats to the capacity of individuals of different ages, sexes and habitats to promote fitness of individuals social status to yield accurate inferences and relative to variation in density of coexisting effectively guide conservation at population individuals. Theoretically, in a state of and species levels, always recognising that “ideal-free distribution” of animals, fitness some habitat use by individuals may be prospects should decrease with increasing related to presence and abundance of intra- and interspecific density of and related or unrelated species (e.g. Götmark competition among co-existing organisms 1989; Jones 2001). Additionally, Buskirk and and ancillary negative interactions (Fretwell & Millspaugh (2006) stated that an informative Lucas 1969; Fretwell 1972). Alternative metric of habitat use may be risk to fitness models have been proposed and tested (e.g. (e.g. the probability of individual mortality; ideal despotic model, Fretwell 1972; ideal pre- Lima & Dill 1990). Nonetheless, Ayers et al. emptive model, Pulliam & Danielson 1991), (2013) and Lancaster (2013) recommended but support for these and others has been that researchers should measure individually inconsistent (Kaminski & Prince 1981; based biological outcomes resulting from Rosenzweig 1985; Kaminski & Gluesing habitat and resource use, because the 1987; Pöysä 2001). Thus, researchers have population-level approach common in cautioned that positive associations between wildlife-resource studies limits ability to indices of habitat suitability and population make inferences about the importance of density should not be inferred without habitats and resources by failing to link these supporting data on biological outcomes (Van with fitness metrics or masking effects Horne 1983; Kaminski & Gluesing 1987; through sampling error or model averaging. Kaminski & Weller 1992; Elmberg et al. 2006; Indeed, population-level habitat use and Ayers et al. 2013). selection processes merely reflect the sum Habitat use and selection vary seasonally, of individual responses; thus, individual- especially for migratory species. Anderson based approaches should be emphasised and Batt (1983) stated: “Obviously, a because natural selection operates at this comprehensive understanding of ecology

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 Habitat use and selection by waterfowl 13 and evolutionary relationships of any species recognizing that (1) habitat selection refers requires an appreciation of selective forces to a process (by individuals) and perhaps that act upon individuals of it during all less so a pattern, (2) that there are many seasons and throughout its range.” During extrinsic factors that influence habitat most of the 20th century, research selection, and (3) that a complete test of conducted to understand ecology and habitat habitat selection involves an assessment of use of waterfowl has focused on the whether or not the documented habitat breeding grounds and season (Batt et al. 1992; preferences are adaptive. A second concern Kaminski & Weller 1992; Baldassarre & was that ornithologists do not consistently Bolen 2006). However, since the 1970s, use habitat-related terminology. That lack ecologists have recognised and investigated of consistency can be remedied by the importance of biological events providing operational definitions to limit experienced and habitats and resources used misunderstanding. A third concern was that by waterfowl throughout the birds’ annual methodologies commonly employed to cycle and range (e.g. Weller 1975; Fredrickson document habitat selection do not account & Drobney 1979; Heitmeyer & Fredrickson for the hierarchical nature of habitat 1981; Kaminski & Gluesing 1987; Tamisier selection and do not generate accurate & Dehorter 1999; Baldassarre & Bolen representations of habitat availability. 2006). Nevertheless, ecology and habitat use Comparisons of used habitat with available of vernal and autumnal staging waterfowl habitat are more appropriate than remain understudied compared to other comparisons of used and unused habitats. seasonal sectors of the annual cycle (Arzel et Definitions of habitat availability ought to al. 2006). Nonetheless, increased knowledge be informed by the natural and life-history of waterfowl ecology throughout their characteristics of the species.” annual cycle and range has been paramount Given this background, the papers in shaping waterfowl habitat conservation following in this section of the journal, plans in North America and Europe which were presented during the 6th North (Canadian Wildlife Service & U.S. Fish and American Duck Symposium, “Ecology Wildlife Service 1986; U.S. Department and Conservation of North American of Interior, Environment Canada, & Waterfowl,” in Memphis, Tennessee in Environment and Natural Resources, Mexico January 2013, will synthesise knowledge on 2012; Kadlec & Smith 1992; Baldassarre & habitat and resource use and selection Bolen 2006; Elmberg et al. 2006). across ducks, geese and swans, when As researchers continue studies of possible, throughout the birds’ annual cycle waterfowl habitat use and selection, they and range, focusing primarily on published would be wise to heed concerns and advice studies conducted in Nearctic and Palearctic of Jones (2001). To paraphrase Jones (2001), ecozones. The synthesis is organised “… ornithologists tend not to consistently cross-seasonally to understand carry-over evaluate the behavioural and fitness contexts biological effects related to individual fitness of their findings. That can be ameliorated by and population dynamics in the following

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 14 Habitat use and selection by waterfowl sequence: 1) autumn migration and a grant from the Swedish Environmental wintering habitats, 2) spring migration and Protection Agency. This manuscript was pre-breeding habitats, 3) nesting habitats, approved for publication by the Director of and 4) post-breeding and moulting habitats. the FWRC as a peer-reviewed journal article Additionally, authors will identify new WF-388. questions deserving future research and References suggest how advances in technology and analytical approaches may enhance Alldredge, J.R. & Ratti, J.T. 1986. Comparisons of understanding of waterfowl habitat use and some statistical techniques of analysis of resources selection. Journal of Wildlife selection. Finally, the synthesis of empirical Management 50: 157–165. information will be used to recommend Arzel, C., Elmberg, J. & Guillemain, M. 2006. future scientific investigations and strategies Ecology of spring-migrating Anatidae: a for habitat and population conservation, to review. Journal of Ornithology 147: 167–184. help sustain harvestable populations of Ayers, C.R.,. Belant, J.L, Millspaugh, J.J. & waterfowl and to protect rare species in the Bodinof, C.M. 2013. Directness of resource northern hemisphere amid global changes in use metrics affects predictions of bear body climate, landscape, economics, socio-politics fat gain. Polar Biology 36: 169–176. and human population growth. As we have Anderson, M.G. & Batt, B.D.J. 1983. Workshop increased our knowledge of waterfowl of on the ecology of wintering waterfowl. the northern hemisphere throughout their Wildlife Society Bulletin 11: 22–24. annual cycle and range during the 20th and Baldassarre, G.A. 2014. Ducks, Geese and Swans of North America. John Hopkins University early 21st centuries (e.g. Batt et al. 1992; Press, Baltimore, Maryland, USA. Baldassarre 2014), the future is rich in Baldassarre, G.A. & Bolen, E.G. 2006. Waterfowl opportunities for collaboration of scientists Ecology and Management, 2nd edition. Krieger and managers to advance our understanding Publishing Company, Malabar, Florida, USA. and conservation of Anatidae and their Belant, J.L., Millspaugh, J.J., Martin, J.A. & habitats worldwide. Gitzen, R.A. 2012. Multi-dimensional space use: the final frontier. Frontiers in Ecology and Acknowledgements the Environment 10: 11–12. This introduction benefitted greatly from Block, W.M. & Brennan, L.A. 1993. The habitat reviews by Chris Ayers, Jerry Belant, J. Brian concept in ornithology: theory and Davis, Tony Fox, Andy Green, Matt applications. Current Ornithology 11: 35–91. Guillemain, Joe Lancaster, Aaron Pearse and Buskirk, S.W. & Millspaugh, J.J. 2006. Metrics for studies of resource selection. Journal of Josh Stafford. Kaminski was supported by Wildlife Management 30: 358–366. the James C. Kennedy Endowed Chair in Canadian Wildlife Service & U.S. Fish and Waterfowl and Wetlands Conservation and Wildlife Service. 1986. North American the Forest and Wildlife Research Center waterfowl management plan. Canadian (FWRC) and Department of Wildlife, Wildlife Service and U.S. Fish and Wildlife Fisheries and Aquaculture (WF), Mississippi Service Report, Ottawa, Ontario, Canada and State University. Elmberg was supported by Washington D.C., USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 Habitat use and selection by waterfowl 15

Chalfoun, A.D. & Martin, T.E. 2007. 1995. Deriving population parameters from Assessments of habitat preferences and individual variations in foraging behaviour. I. quality depend on spatial scale and metrics of Empirical game theory distribution model of fitness. Journal of Applied Ecology 44: 983–992. oystercatchers Haematopus ostralegus feeding Cody, M.L. (ed.) 1985. Habitat Selection in Birds. on mussels Mytilus edulis. Journal of Animal Academic Press, Orlando, Florida, USA. Ecology 64: 265–276. Cowardin, L.M., Carter, V., Golet, G.C. & Laroe, Götmark, F. 1989. Costs and benefits to eiders E.T. 1979. Classification of wetlands and nesting in gull colonies: a field experiment. deepwater habitats of the United States. Ornis Scandinavica 20: 283–288. Biological Service Program Report No. Heitmeyer, M.E., & Fredrickson, L.H. 1981. Do FWS/OBS-79/31. United States Fish and wetland conditions in the Mississippi Delta Wildlife Service, Washington D.C., USA. hardwoods influence mallard recruitment? Elmberg, J., Pöysä, H., Sjöberg, K. & Nummi, P. Transactions of the North American Wildlife and 1997. Interspecific interactions and co- Natural Resources Conference 46: 44–57. existence in dabbling ducks: observations Hildén, O. 1965. Habitat selection in birds. and an experiment. Oecologia 111: 129–136. Annales Zoological Fennica 2: 53–75. Elmberg, J., Nummi, P., Pöysä, H., Sjöberg, K., Johnson, D.H. 1980. The comparison of usage Gunnarsson, G., Clausen, P., Guillemain, M., and availability measurements for evaluating D. Rodrigues, D. & Väänänen, V.-M. 2006. resource preference. Ecology 61: 65–71. The scientific basis for new and sustainable Jones, J. 2001. Habitat selection studies in avian management of migratory European ducks. ecology: A critical review. The Auk 118: Wildlife Biology 12: 121–127. 557–562. Fretwell, S.D. 1972. Populations in a seasonal Kadlec, J.A. & Smith, L.M. 1992. Habitat environment. Princeton University Press, management for breeding areas. In B.D.J. Princeton, New Jersey, USA. Batt, A.D. Afton, M.G. Anderson, C.D. Fretwell, S.D. & Lucas, H.L., Jr. 1970. On Ankney, D.H. Johnson, J.A. Kadlec & G.L. territorial behavior and other factors Krapu (eds.), Ecology and Management of influencing habitat distribution in birds. I. Breeding Waterfowl, pp. 590–610. University of Theoretical development. Acta Biotheoretica Minnesota Press, Minneapolis, USA. 19: 16–36. Kaminski, R.M. & Gluesing, E.A. 1987. Density- Fredrickson, L.H. & Drobney, R.D. 1979. Habitat and habitat-related recruitment in mallards. utilization by postbreeding waterfowl. In T. Journal of Wildlife Management 51: 141–148. A. Bookhout (ed.), Waterfowl and Wetlands – an Kaminski, R.M. & Prince, H.H. 1981. Dabbling Integrated Review, pp. 119–131. LaCrosse duck and aquatic macroinvertebrate Printing Co., LaCrosse, Wisconsin, USA. responses to manipulated wetland habitat. Garshelis, D.L. 2000. Delusions in habitat Journal of Wildlife Management 45: 1–15. evaluation: measuring use, selection and Kaminski, R.M. & Weller, M.W. 1992. Breeding importance. In L. Boitani & T.K. Fuller habitats of Nearctic waterfowl. In B.D.J. (eds.), Research Techniques in Animal Ecology: Batt, A.D. Afton, M.G. Anderson, C.D. Controversies and Consequences, pp. 111–153. Ankney, D.H. Johnson, J.A. Kadlec & G.L. Columbia University Press, New York, USA. Krapu (eds.), Ecology and Management of Goss-Custard, J.D., Caldow, R.W.G., Clarke, R.T., Breeding Waterfowl, pp. 568–589. University of le V. dit Durell, S.E.A. & Sutherland, W.J. Minnesota Press, Minneapolis, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 16 Habitat use and selection by waterfowl

Lack, D. 1968. Ecological Adaptations for Breeding Stearns, S.C. 1992. The Evolution of Life Histories. in Birds. Chapman & Hall, London, UK. Oxford University Press, Oxford, UK. Lancaster, J.D. 2013. Survival, habitat use, and Stillman, R.A. 2008. MORPH – An individual- spatiotemporal use of wildlife management based model to predict the effect of areas by female mallards in Mississippi’s environmental change on foraging animal Alluvial Valley. M.Sc. thesis, Mississippi State populations. Ecological Modelling 216: 265–276. University, Mississippi, USA. Tamisier, A. & Dehorter, O. 1999. Camargue, Lima, S.L. & Dill, L.M. 1990. Behavioral canards et foulques: fonctionnement et decisions made under the risk of predation: a devenir d’un prestigieux quartier d’hiver. review and prospectus. Canadian Journal of Centre Ornithologique du Gard, Nîmes, Zoology 68: 619–640. France. Mayr, E. 1970. Populations, Species and Evolution. Thomson, R.L., Forsman, J.T. & Mönkkönen, M. Harvard University Press, Cambridge, 2003. Positive interactions between migrant Massachusetts, USA. and resident birds: Testing the heterospecific Mulhern, J.H., Nudds, T.D., & Neal, B.R. 1985. attraction hypothesis. Oecologia 134: 431– 438. Wetland selection by mallards and blue- U.S. Department of Interior (DOI), winged teal. Wilson Bulletin 97:473–485. Environment Canada & Environment O’Neal, B.J., Stafford, J.D. & Larkin, R.P. 2010. Natural Resources, Mexico. 2012. North Waterfowl on weather radar: applying American Waterfowl Management Plan: ground-truth to classify and quantify bird People conserving waterfowl and wetlands. movements. Journal of Field Ornithology 81: U.S. DOI, Washington D.C., USA. 7–82. Van Horne, B. 1983. Density as misleading Pöysä, H. 2001. Dynamics of habitat distribution indicator of habitat quality. Journal of Wildlife in breeding mallards: assessing the Management 47: 893–901. applicability of current habitat selection Weller, M.W. 1975. Migratory waterfowl: a models. Oikos 94: 365–373. hemispheric perspective. Publicaciones Biologicas Pulliam, H.R. & Danielson, B.J. 1991. Sources, Instituto de Investigaciones Cientificas 1: 89–130. sinks, and habitat selection: a landscape Wiens, J.A. 1973. Patterns and process in perspective on population dynamics. grassland bird communities. Ecological American Naturalist 137 (Supplement): 50–66. Monographs 43: 237–270. Rettie, W.J. & Messier, F. 2000. Hierarchical Wiens, J.A. 1976. Population responses to patchy habitat selection by woodland caribou: its environments. Annual Review of Ecology and relationship to limiting factors. Ecography 23: Systematics 7: 81–120. 466–478. Wiens, J.A. 1977. On competition and variable Rosenzweig, M.L. 1985. Some theoretical aspects environments. American Scientist 65: 590–597. of habitat selection. In M.L. Cody (ed.), Wiens, J.A. 1985. Habitat selection in variable Habitat Selection in Birds, pp. 517–540. environments: shrub-steppe birds. In M.L. Academic Press, Orlando, Florida, USA. Cody (ed.), Habitat Selection in Birds, pp. 227– Southwood, T.R.E. 1977. Habitat, the templet for 251. Academic Press, Orlando, Florida, USA. ecological strategies? Journal of Animal Ecology Wilson, E.O. & Bossert, W.H. 1971. A Primer 46: 337–365. of Population Biology. Sinauer Associates, Stearns, S.C. 1976. Life history tactics: a review of Inc. Publishers, Sunderland, Massachusetts, the ideas. Quarterly Review of Biology 5: 3–47. USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 9–16 17

Habitat and resource use by waterfowl in the northern hemisphere in autumn and winter

J. BRIAN DAVIS1*, MATTHIEU GUILLEMAIN2, RICHARD M. KAMINSKI1, CELINE ARZEL3, JOHN M. EADIE4 & EILEEN C. REES5

1Department of Wildlife, Fisheries and Aquaculture, Mississippi State University, Mississippi State, Mississippi 39762, USA. 2Office National de la Chasse et de la Faune Sauvage, CNERA Avifaune Migratrice Le Sambuca, 13200 Arles, France. 3Section of Ecology, University of Turku, 20014 Turku, Finland. 4Department of Wildlife, Fish and Conservation Biology, University of California, One Shields Avenue, Davis, California 95616, USA. 5Wildfowl & Wetlands Trust, Martin Mere, Burscough, near Ormskirk, Lancashire L40 0TA, UK. *Correspondence author. E-mail: [email protected] Abstract A particular aim of avian ecologists, especially those studying waterfowl Anatidae, in the 20th and early 21st centuries has been to elucidate how organisms use habitats and intrinsic resources to survive, reproduce and ultimately affect fitness. For much of the 20th century, research was mainly on studying species during the breeding season; however, by the 1970s, the focus had changed to understanding migratory waterfowl throughout their annual cycle and range in Europe and North America. Autumn and winter are considered the non-breeding seasons, but habitat and resource use through these seasons is crucial for completing spring migration and subsequent breeding. Here we review the literature on autumnal and winter habitat use by Nearctic and Palearctic waterfowl to determine characteristics of important landscapes and habitats for the birds during autumn migration and in winter. Selection of habitats and resources is discussed (when literature permits) in relation to Johnson’s (1980) model of hierarchical habitat selection. Habitat use by selected species or groups of waterfowl is also reviewed, and important areas for future research into habitat ecology are identified. We suggest that the greatest lack of understanding of waterfowl habitat selection is an ongoing inability to determine what habitats and intrinsic resources, at multiple scales, are truly available to birds, an essential metric in quantifying “selection” accurately. Other significant challenges that impede gaining knowledge of waterfowl ecology in the northern hemisphere are also described. Nonetheless, continued technological improvements and engagement of diverse interdisciplinary professional expertise will further refine understanding of waterfowl ecology and conservation at continental scales.

Key words: autumn, habitat use, migration, selection, waterfowl, winter.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 18 Autumn–winter habitat use by waterfowl

Understanding how wildlife and especially species are largely tied to freshwater birds use habitats and resources to survive systems but several use agricultural, estuarine and reproduce (i.e. promote fitness; sensu and marine environments (Bellrose 1980; Kaminski & Elmberg 2014) has long been Baldassarre 2014). Some waterfowl habitats the subject of ecological research (Darwin are relatively stable and seasonally predictable 1859; Lack 1944; Morrison et al. 1992). relative to hydrology (e.g. estuarine and Studies of waterfowl habitat use and lacustrine wetlands; Cowardin et al. 1979), selection are well represented within the whereas other habitats provide food and substantial avian literature (Block & other resources temporarily but are Brennan 1993; Kaminski & Elmberg 2014). characteristically dynamic, such as harvested David Lack’s (1966) early reference to agricultural lands, riverine and palustrine habitat selection remains valid today, and wetlands (Tourenq et al. 2001; Fredrickson visionaries such as Lack and also Fretwell 2005; Baldassarre & Bolen 2006; Mitsch & (1972) further hypothesised that non- Gosselink 2007; O’Neal et al. 2010). breeding habitats and resources may be Here, classic and contemporary literature important limiting factors for birds of the that revealed habitat and associated resource northern hemisphere, especially migratory use by Holarctic waterfowl during autumn species such as waterfowl. Conditions at and winter is reviewed, with emphasis on the non-breeding habitats (e.g. winter wetlands) latter season of the annual cycle. The review correlate with waterfowl recruitment does not provide an exhaustive summary of (Heitmeyer & Fredrickson 1981; Nichols et habitat and resource use by each species or al. 1983; Kaminski & Gluesing 1987; group of waterfowl, but gives an overview Raveling & Heitmeyer 1989; Guillemain et focusing on habitat use by non-breeding al. 2008). However, understanding habitat waterfowl from macro- to finer spatial scales, use and selection by seasonally mobile when available information permitted such waterfowl remains challenging, because coverage (sensu Johnson 1980; Kaminski & technology, logistics, economics and other Elmberg 2014). Space limitations required us constraints impede monitoring and to review a selected group of waterfowl assessment of resource availability, species and tribes, but planning is underway to exploitation and biological outcomes for address non-breeding seasonal ecology of individuals and populations, from local to lesser known taxa (e.g. Cairini sp., Dendrocygnini flyway scales and cross-seasonally (Elmberg sp. and Anas fulvigula) and better known or et al. 2014; Kaminski & Elmberg 2014; more widely distributed Nearctic species in a Sedinger & Alisauskas 2014). future publication (e.g. A. americana, crecca, The number of waterfowl species and clypeata, strepera, rubripes and Branta canadensis). different populations, and their abundance We begin with a conceptual overview of and geographic distribution in the Holarctic, autumn migration applicable to Nearctic and makes waterfowl dominant fauna of aquatic Palearctic waterfowl, followed by a review of and terrestrial systems in the northern selected eco-regions important to non- hemisphere (Raveling 2004). Many waterfowl breeding waterfowl in the Holarctic and the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 19 aforementioned review of selected species or Autumn migration groups of ducks, geese and swans. Finally, currently perceived challenges in studying Avian migration involves complex habitat selection by non-breeding waterfowl physiological, behavioural, genetic and are conveyed to stimulate further research and ecological influences at individual and flock conservation of these birds and their habitats levels, which can influence population in the northern hemisphere and worldwide. dynamics and demography (Dingle & Drake 2007). Numerous publications focus on Hierarchical habitat use and avian migration (e.g. Dingle 1996; Dingle & selection Drake 2007; Newton 2007; Stafford et al. Kaminski & Elmberg’s (2014) conceptual 2014), but a disproportionate number review of hierarchical habitat selection (sensu address passerines, while relatively few Johnson 1980), indicated that habitat use consider waterbirds. This reality is surprising and selection by migratory birds, such as given the well-known migratory nature of most waterfowl, can be envisioned as a most Holarctic waterfowl (Arzel et al. 2006). multi-stage, spatio-temporal process from Migration involves large-scale movements macro- to micro-scales throughout the birds’ from breeding to non-breeding grounds and annual cycle and range. Migratory waterfowl vernal returns to breeding grounds seemingly make 1st order selection of (Salewski & Bruderer 2007; Zink 2011). geographic regions, such as those important Autumnal migration may be considered to and used by the birds during breeding endogenously and exogenously influenced and non-breeding seasons (Johnson 1980; seasonal movements of birds between Baldassarre & Bolen 2006). Within 1st order breeding and non-breeding areas (Alerstam occupied regions, waterfowl make 2nd order & Lindström 1990; Dingle 1996; Salewski & selections of wetland systems (Cowardin et al. Bruderer 2007). A perplexing aspect of 1979) and possibly associated landscapes for autumn migration in waterfowl is that some species adapted to terrestrial habitats timing of departure in birds is especially (e.g. arable lands). Next, waterfowl make complicated (O’Neal et al. 2010; Krementz 3rd order selections of local, site-specific et al. 2012). Long-migrant passerines wetlands or other locations in their seasonal typically exhibit a time-minimisation home range, and finally 4th order selections strategy (Dänhardt & Lindström 2001; of microhabitats where individuals may O’Neal et al. 2010), and although geese and roost, forage or engage in other activities to swans refuel at staging sites for shorter acquire food or other resources, including periods in autumn than in spring (Madsen mates (Wiens 1973; Johnson 1980; Kaminski 1980; Luigujõe et al. 1996; Beekman et al. & Weller 1992; Baldassarre & Bolen 2006). 2002), some ducks, such as larger-bodied A reversal of this process from micro- to species like Mallard Anas platyrhynchos, may macro-habitats also can be envisioned, as remain at mid-migration stopovers for birds depart micro-habitats to disperse or weeks or longer despite harsh weather migrate to different regions. conditions that seemingly would stimulate

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 20 Autumn–winter habitat use by waterfowl migration (Bellrose & Crompton 1970; newly flooded but ephemeral habitats in O’Neal et al. 2010; Schummer et al. 2010; autumn (Bellrose 1980), whereas Mallard Krementz et al. 2012; Dalby 2013). may have protracted migrations (Bellrose Moreover, autumn migration and winter 1980; Krementz et al. 2012). habitat use are further complicated by Movements, site fidelity and turnover habitat availability and quality and human- rates of waterfowl during autumn-winter are related disturbance (e.g. Väänänen 2001; likely to reveal patterns of habitat suitability Roshier et al. 2006; Legagneux et al. 2009; and trade-offs made by waterfowl during O’Neal et al. 2010; St. James et al. 2013). these periods of the annual cycle (Rodway Life histories of waterfowl vary 2007). Winter site fidelity is known to be considerably among species and confound strong in geese and swans (Owen 1980) simple explanations of migration patterns. but of lesser importance in ducks, which For instance, although body size influences exhibit greater spatio-temporal plasticity in migration and habitat use (Raveling 2004), habitat use (Mulhern et al. 1985; Robertson American Black Duck Anas rubripes (1,100 g; & Cooke 1999). Moreover, interspecific Zammuto 1986; Baldassarre 2014) overlaps comparisons of winter philopatry are in time and space with American Green- confounded by vast differences in the size of winged Teal A. crecca carolinensis (318 g; regions investigated (Robertson & Cooke Zammuto 1986), the smallest dabbling duck 1999). In Europe, studies of individually- species, during migration and winter marked Eurasian Teal highlighted significant (Bellrose 1980, Baldassarre & Bolen 2006). wintering site fidelity among and within Conversely, Blue-winged Teal A. discors winters (Guillemain et al. 2009; Guillemain et (363 g; Zammuto 1986), although ~12% al. 2010a), suggesting that birds were able to heavier than Green-winged Teal, winter evaluate site quality and adapt their use of at more southerly latitudes (≤ 30°N; traditional wintering areas, perhaps resulting Thompson & Baldassarre 1990). Clearly, in increased individual fitness. Of course, waterfowl migration patterns do not strictly such traditions may be jeopardised if abrupt follow ecological generalisations such as habitat changes occur. Indeed, the ecology Bergmann’s Rule (Bergmann 1847). of waterfowl migration in the northern Many Palearctic waterfowl converge from hemisphere remains a frontier for future Fenno-Scandian and Russian breeding scientific investigation (Arzel et al. 2006). grounds toward the Baltic Sea, where they use various habitats as staging sites before Selected important Holarctic gradually moving south during winter. Some regions for non-breeding birds such as Eurasian Teal A. crecca crecca waterfowl move by successive small flights in early Eastern United States autumn, whilst Mallard lag behind and move later in less numerous but longer flights The eastern U.S. historically has been an (Dalby 2013). Others, such as Northern important region for migrating and Pintail A. acuta, may be nomadic and seek wintering waterfowl, particularly lacustrine

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 21 and estuarine coastal wetlands and deep- ~30% of all dabbling ducks in the Atlantic water habitats (Cowardin et al. 1979; Bellrose Flyway including Green-winged Teal, 1980). The region of the Atlantic Coast Northern Shoveler Anas clypeata, Mallard, Joint Venture (ACJV) encompasses 17 states American Wigeon A. americana and in the Atlantic Flyway and is the most Northern Pintail (Gordon et al. 1989). In densely human-populated area in the Florida, the St. John’s and Indian Rivers conterminous U.S., wherein about 35% of basins provide important waterfowl habitat, the population resides (ACJV 2009). supporting nearly 400,000 ducks during Landscape diversity in this region winter (ACJV 2005). Freshwater lakes, such includes ~22% agricultural land and 25% as Lake Okeechobee, also provide important wetlands, which together support ~37 wintering habitats for many waterfowl, native species of waterfowl (ACJV 2009). including Lesser Scaup Aythta affinis, Considering 2nd order habitat selection Ring-necked Duck A. collaris, American within this region, estuarine systems of Wigeon, and Blue-winged Teal (Johnson & coastal Maine are important to wintering Montalbano 1989). American Black Duck, Common Eider Somateria mollissima and scoters Melanitta sp. Mississippi Alluvial Valley that use sheltered ice-free areas for foraging Largely forested prior to settlement by and loafing (ACJV 2005), while fringes of Europeans in the 19th century, flood control saltmarshes and mudflats are important to for agriculture and human inhabitation Mallard and other dabbling ducks (Jorde et influenced a nearly 80% loss of lowland al. 1984). Barrier beaches, back-barrier forests in the Mississippi Alluvial Valley coastal lagoons and salt marshes of Long (MAV) by the late 20th century, with only Island and New Jersey provide additional highly fragmented tracts remaining today important winter habitats for American (MacDonald et al. 1979; Klimas et al. Black Duck and Brent Geese Branta bernicla 2009). The MAV contains flooded (ACJV 2005; Plattner et al. 2010). Farther croplands, wetlands, deep water habitats and south exists the Chesapeake Bay, the largest aquaculture ponds that are important to estuary in the conterminous U.S. with a migrating and wintering ducks and geese watershed that drains 165,760 km2, along (Cowardin et al. 1979; Christopher et al. 1988; with North Carolina Sounds, natural and Reinecke et al. 1989; Stafford et al. 2006; artificial lakes and reservoirs, flooded Kross et al. 2008; Feaga 2013). Swans (e.g. bottomland hardwoods, Carolina bays and Trumpeter Swans Cygnus buccinator) are rarely estuarine and salt marshes that provide sighted in winter in the MAV (R.M. habitat for a diversity of ducks, geese and Kaminski, pers. obs.; MAV Christmas Bird swans (Hindman & Stotts 1989). Counts unpubl. data). Additionally, South Carolina and Georgia Within the flooded agricultural landscape provide habitat for wintering dabbling, (including the aquaculture ponds), migrating diving and sea ducks (Gordon et al. 1989; and wintering waterfowl use 2nd order ACJV 2005). South Carolina alone winters lacustrine (e.g. oxbow and watershed lakes,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 22 Autumn–winter habitat use by waterfowl reservoirs), palustrine (e.g. forested and estimated that rice represented > 41% of moist-soil wetlands) and riverine systems in total food intake by Mallard. However, the MAV (e.g. Mississippi River and because rice, soybean, and other seed tributaries; Cowardin et al. 1979; Mitsch & crops are planted and harvested earlier Gosselink 2007). Considering 3rd order nowadays in the MAV than during the 20th habitat use of agricultural lands and wetlands century, deterioration of waste seed occurs within 2nd order systems, Reinecke et al. because of germination, decomposition and (1992) reported that over half of the Mallard consumption by non-waterfowl species after observed during aerial surveys across harvest but before major wintering flocks most of the MAV used flooded rice and arrive in the MAV (Stafford et al. 2006; soybean fields during winters 1987–1990. Foster et al. 2010; Petrie et al. 2014). Subsequently, during the early 2000s, Pearse Reduction in waste rice from harvest et al. (2012) reported that greatest densities of through late autumn–early winter in the Mallard in the Mississippi portion of the MAV is estimated at 71–99% (Manley et al.; MAV during winter were observed in habitat Stafford et al. 2006). Despite reduced complexes composed of 50% flooded availability of waste rice in harvested fields cropland, 20% hardwood or scrub-shrub in the region, flooded rice fields however wetlands, 20% moist-soil and other have structural characteristics similar to emergent wetlands and 10% permanent natural wetlands (Elphick 2000; Huner water bodies (e.g. rivers, lakes, ponds). et al. 2002; Marty 2013). The mid-winter Greatest densities of other dabbling duck population goal for the Lower Mississippi species were also associated with a similar Valley Joint Venture of the North American habitat composition (Pearse et al. 2012). Waterfowl Management Plant (LMVJV) is Waterfowl associations with flooded > 7.8 million dabbling ducks, and winter- cropland might be expected given that the flooded rice fields provide ~11% of all food MAV is now largely an agricultural energy available to dabbling ducks in landscape. Despite losses of natural flooded habitats in the LMVJV (Petrie wetlands in the MAV and continentally et al. 2014). Approximately 20% of the (Mitsch & Gosselink 2007), migrating and 748,668 ha of ricelands is winter-flooded in wintering waterfowl have adapted to the LMVJV (Petrie et al. 2014). If the flooded agricultural lands and make LMVJV rice fields were able to produce a significant use of them in the MAV to meet second harvested crop intra-seasonally as in nutritional and other physiological needs Louisiana and Texas (i.e. ratoon crop, Marty (Delnicki & Reinecke 1986; Reinecke et al. 2013), the amount of food available to 1989; O’Neal et al. 2010). Indeed, ricelands dabbling ducks from the flooded fields in in the MAV are critical for meeting seasonal the LMVJV would increase 12-fold (Petrie et requirements of waterfowl using this region al. 2014). Development of rice varieties and (Stafford et al. 2006). In the late 1970s and other crops with ability to ratoon at latitudes early 1980s, Delnicki & Reinecke (1986), within the MAV would increase substantially studying food use and body weight, the abundance of waste grain following

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 23 harvest and benefit migrating and wintering bottomlands in the Interior Flatwoods and waterfowl (Wiseman et al. 2010; Petrie et al. MAV in Mississippi during winter. Mallard 2014; Marty 2013). used microhabitats that contained less Despite dominant coverage of agricultural woody understory cover, whereas Wood land in the MAV, Mallard and other waterfowl Duck were associated with microhabitats of use 3rd and 4th order wetland sites in the increased understory vegetation (Kaminski MAV (Reinecke et al. 1989). Reinecke et al. et al. 1993). Within moist-soil wetlands in the (1992) reported that Mallard used forested MAV, dabbling ducks of several species wetlands (3–11%) and moist-soil wetlands foraged in experimental plots with water (3–29%) within and among winters. depths ranging from 3–16 cm (Hagy & Additionally, Davis & Afton (2010), working Kaminski 2012). Such a range of depths in the Louisiana portion of the MAV, may facilitate forage acquisition by a reported that radio-marked female Mallard diversity of species using a common habitat, selected forested wetlands and suggested that at least until food depletion occurs (Greer et continued restoration and establishment of al. 2009; Hagy et al. 2014). these habitats should benefit females. In addition to flooded croplands and However, they did not report any natural wetlands in the MAV, aquaculture relationships between Mallard winter survival ponds for production of Channel Catfish or other correlates of fitness that might Ictalurus punctatus and bait fish have become implicate benefits resulting from female use important staging and wintering habitats used of forested wetlands. Subsequently, Lancaster by dabbling and diving ducks since their (2013), working in the Mississippi portion of construction in the 1970s (Christopher et al. the MAV, investigated habitat-related survival 1988; Reinecke et al. 1989; Wooten & Werner of radio-marked female Mallard. Greatest 2004). Species of waterfowl commonly using rates of winter survival (≥ 75%) were catfish ponds include Lesser Scaup Aythya exhibited by females that used habitat affinis, Ruddy Duck Oxyura jamaicensis and complexes composed mostly of forested and Northern Shoveler, along with lesser emergent wetlands (86% combined) and 12% abundances of Mallard, Gadwall A. strepera, cropland, which was notable considering and introduced resident Giant Canada Geese that most of the MAV landscape now is Branta canadensis maxima (Christopher et al. cropland (Lancaster 2013; Kaminski & 1988; Dubovsky & Kaminski 1992; Vest et al. Davis 2014). Thus, although Mallard may be 2006, Feaga 2013). Dubovsky & Kaminski considered habitat generalists, they also use (1992) estimated that 150,000 ducks used certain habitats disproportionately, affording catfish ponds in Mississippi, with an average increased fitness prospects consistent with of 100,000 individuals using ponds weekly in the concept of habitat suitability (sensu the mid-1980s. Wooten & Werner (2004) Fretwell 1972; Kaminski & Elmberg 2014). collected Lesser Scaup from Arkansas baitfish Considering 4th order microhabitats, ponds and reported scaup primarily ingested Mallard and Wood Duck Aix sponsa Chironomidae larvae, but ~25% of collected differentially used flooded hardwood birds contained fish biomass or bones.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 24 Autumn–winter habitat use by waterfowl

Because of competition from foreign have been among the greatest wetland losses markets, infrastructural and other costs, in the coastal prairies; ~100,000 ha of non- catfish aquaculture has declined in the MAV farmed freshwater wetlands have been lost (U.S. Department Agriculture 2010). There in the coastal plains of Texas since the mid- were 64,000 ha of ponds in Mississippi, 1940s (Moulton et al. 1997). Conversion of Louisiana and Arkansas in 2001, but only rice agriculture to cotton and soybean 25,000 ha remained in operation in those production has further reduced important states by 2012 (Lehnen & Krementz 2013). habitats for waterfowl (Anderson & Ballard Feaga (2013) reported that migrating and 2006). Gulf coastal wetlands are critical to wintering waterfowl and other waterbirds several guilds of wintering waterfowl occurred in densities on catfish production (Weller 1964; Chabreck et al. 1989; Hobaugh impoundments (~130 birds/ha) similar to et al. 1989; Marty 2013), and an estimated idled impoundments (~120 birds/ha). 19% of all waterfowl wintering in the U.S. However, different bird communities existed use marshes in the Louisiana Gulf Coast in production versus idled production (Michot 1996; Bolduc & Afton 2004). The ponds, the latter now managed to provide Texas Mid-Coast once wintered 78% of emergent vegetation, mudflats and shallow the Northern Pintail in the Central wetland areas < 30 cm during summer–winter Flyway (Ballard et al. 2004). Contemporary wetland birds (Feaga 2013; Kaminski & estimates of midwinter population goals for Davis 2014). Diving and dabbling ducks the Gulf Coast JV region include > 5.6 and American Coot Fulica americana were million dabbling ducks (Petrie et al. 2014). primary users of production aquaculture Considering 2nd and 3rd order habitat impoundments (Dubovsky & Kaminski 1992; selection, freshwater and intermediate Feaga 2013), whereas idled impoundments marshes along the Gulf of Mexico are were used by over 40 species of ducks, perhaps the most important wetland habitats shorebirds, waders and other waterbirds for waterfowl in the region (Chabreck et al. (Feaga 2013; Kaminski & Davis 2014). 1989; Batzer & Baldwin 2012). Brackish marshes are the most extensive habitat and Louisiana-Texas Gulf Coast considered historical habitats for wintering The coastal tallgrass prairies of Louisiana Snow Geese Anser caerulescens (Chabreck et al. and Texas once covered over 1 million ha 1989; Batzer & Baldwin 2012), but salt marsh (Chabreck et al. 1989; Hobaugh et al. 1989). habitats are generally regarded as less They have slight topography, relatively favourable to waterfowl in Gulf coastal impervious soils and thus seasonal wetlands systems (Williams III & Chabreck 1986; (Smeins et al. 1991; Petrie et al. 2014). Winter Batzer & Baldwin 2012). In addition to these, rains and tropical storms in summer– lakes (e.g. Grand, White), bays (e.g. Atchafalaya, autumn periodically inundate basins and Terrebonne) and off-shore habitats have been provide habitat for numerous migrating and important historically for scaup and other wintering waterfowl (Petrie et al. 2014). diving and sea ducks in the Gulf region Fresh and intermediate brackish marshes (Harmon 1962; Afton & Anderson 2001).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 25

Scaup wintering off-shore in Louisiana have United States Great Plains comprised 50–86% of the total wintering population and were much more abundant The Playa Lakes Region (PLR) contains off-shore than in in-shore habitats in January 60,000–100,000 playa lakes or shallow (Kinney 2004). Kinney (2004) flew transect wetlands that generally occur at the bottom surveys and determined that only about 15% of large watersheds and are formed by wind of scaup were detected in some years by and water dissolution processes (Smith 2003; traditional Midwinter Waterfowl Surveys. One Venne et al. 2008). Playa wetlands range in hypothesis for scaup wintering farther off- size from < 1 ha to > 300 ha, extend from shore is that Surf Clams Mulinia lateralis were Wyoming and Nebraska to Texas and New historically a preferred food for the species Mexico, and are habitat to a wide diversity (Harmon 1962; Kinney 2004) and recent of life forms including waterfowl (Playa increases in hypoxic areas in the near-shore Lakes Joint Venture 2014). Historic native waters of the Gulf may be causing scaup to grassland has largely been replaced with venture farther off-shore for food. arable crops, and subsequent erosion of Along the Texas Gulf Coast, the Laguna topsoil has contributed to sedimentation of Madre is a large shallow lagoon that contains ~90% of all playas in the Southern High ~80% of the seagrass communities along the Plains (SHP; Venne et al. 2008). Moreover, Texas coast (Ballard et al. 2010). The dominant ~80,000 playas throughout the Great Plains species is Shoal Grass Halodule wrightii and states are currently incapable of recharging ~80% of the continental Redhead Aythya the Ogallala aquifer (Playa Lakes Joint americana population winters in the region, Venture 2014). Historically, one-third of the primarily because of seagrasses (Division: Central Flyway Northern Pintail population Angiospermae) and associated habitats (~300,000 birds) used playa lakes in the SHP, (Weller 1964; Mitchell et al. 1994; Michot et al. but this population has declined 47% since 2006; Ballard et al. 2010). Several studies have 1977 (Bellrose 1980; Luo et al. 1997; Haukos documented the importance of proximate 2004; Moon et al. 2007). Concomitantly, inland freshwater ponds to Redhead and body condition of pintail in the PLR has other ducks including Lesser Scaup (Adair et declined considerably since the mid-1980s al. 1996; Michot et al. 2006; Ballard et al. 2010). (Moon et al. 2007). The proximity of coastal ponds to seagrass The SHP is a southern extension of the foraging areas on the Gulf Coast is important, PLR and is a critical region to waterfowl, as Redhead were never observed using ponds once containing 25,000–30,000 wetlands > 5.7 km from the shoreline or > 8.1 km from (Smith 2003; Baldassarre & Bolen 2006; the nearest foraging area (Ballard et al. 2010). Venne et al. 2008). Obenberger (1982) studied Thus, proximity of freshwater ponds to several species of dabbling ducks from seagrass beds in the Laguna Madre is an autumn–late winter 1980–1982 and reported example of a critical synergistic habitat that ducks generally had a bimodal migration. association, particularly in drier winters Migration phenology of Northern Pintail (Ballard et al. 2010). and Green-winged Teal peaked in November,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 26 Autumn–winter habitat use by waterfowl and autumn abundances were at least double 1986; Heitmeyer et al. 1989; Fleskes et al. their greatest numbers during vernal peaks. 2005; Miller et al. 2010). The state has lost Nearly 30 years later, Baar et al. (2008) ~95% of its historic wetlands (Central Valley conducted similar research in the SHP and Joint Venture 2006) but continues to support observed that duck use of playas was millions of non-breeding waterfowl. Within much more intermittent, protracted or less California, the Central Valley provides intensive compared to previous decades. Baar critical wetland and agricultural habitat for et al. (2008) offered two possible explanations migrating and wintering waterfowl and was for these patterns. First, abundance of playa the focus of one of the original Joint wetlands, irrigation ponds and tailwater Ventures of the North American Waterfowl reservoirs were greatly reduced, and playas Management Plan (NAWMP 1986). The have become more rainfall dependent (Smith Central Valley encompasses ~4.1 million ha, 2003; Baar et al. 2008). Second, playas have stretching 724 km north to south and 64 km been subjected to significant sedimentation, east to west. The valley is dominated by two with negative impacts to hydrologic patterns riverine systems – the Sacramento River and and function (Smith 2003). Moon & Haukos the San Joaquin River, which meet at the (2006) attributed declining body condition of Delta then flow into the Pacific Ocean past Northern Pintail to harassment and stress, the Suisun Marsh, one of the largest resulting from increased movements by contiguous brackish marshes in the western hunters pursuing waterfowl and Ring-necked United States. Pheasant Phasianus colchicus (Baar et al. 2008). The hydrology of the valley determines Generally, evidence suggests that the main habitat types and influences important waterfowl foods, such as waste seasonal and inter-annual patterns of agricultural or natural seeds, are becoming waterfowl use (Fleskes 2012). However, depleted in early winter in the SHP hydrology has been altered drastically from (Baldassarre & Bolen 1984; Bolen et al. 1989; agriculture and urban growth and caused Smith & Sheeley 1993; Moon & Haukos considerable changes in distribution of 2006). As a consequence, exploitation of waterfowl habitats. Before the 1849 Gold these environments by dabbling and other Rush, the valley contained > 1.6 million ha ducks may be more limited during late winter of wetland habitat (Central Valley Joint and spring (Baar et al. 2008) compared with Venture 2006). Most of these wetlands were prior decades (Obenberger 1982). Dedicated seasonal, inundated by riverine flooding in conservation programmes have been the valley, bordered by expansive riparian championed and are needed in the SHP and grassland habitats, which may have (Haukos & Smith 2003; Smith 2003). supported 20–40 million waterfowl during migrations and winter. Central Valley of California Seasonal and permanent wetlands in the California always has been one of the most Central Valley are distributed in four sub- important regions for wintering waterfowl in regions: the southern San Joaquin Valley North America (Gilmer et al. 1982; Miller (including Tulare Basin, which held the now

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 27 dry Tulare Lake, once the largest freshwater Most existing wetland habitat in the valley lake west of Mississippi; Fleskes 2012), the is managed and comprises seasonal, semi- northern Sacramento Valley, the Delta and permanent and permanent wetlands. the Suisun Marsh. Historically, many Seasonal wetlands are flooded in autumn for waterfowl wintering in California would waterfowl and other waterbirds and drawn migrate first to Tulare Lake, a vast shallow down in late winter. Many wetlands are complex of seasonal and permanent managed as waterfowl hunting clubs or state marshes. As winter progressed birds moved and federal wildlife areas or refuges. north, through the San Joaquin Valley, Seasonal wetlands provide critical foraging Delta and Suisun Marsh into the habitat for non-breeding waterfowl. These Sacramento Valley. Prior to land conversion, wetlands are managed annually using several ~40% of waterfowl habitat occurred in the methods (e.g. disking, irrigation and water San Joaquin Valley (including Tulare Basin), management) to promote moist-soil plants while the remaining 60% occurred in the such as Watergrass Echinochloa crusgalli, Sacramento Valley, Delta and Suisun Marsh smartweed Polygonum sp. and Swamp (Fleskes et al. 2005). By approximately 1900, Timothy Crypsis schoenoides (Heitmeyer et al. the Tulare lakebeds were effectively drained 1989). Semi-permanent wetlands are by diversion of water for agriculture, and flooded from autumn to early July, while the lakebeds now remain dry in all but permanent wetlands are flooded throughout extremely wet years. Wetlands in the San the year (Central Valley Joint Venture 2006). Joaquin and Sacramento Valleys were also Semi-permanent and permanent wetlands converted to agricultural land, leading to produce less food, but provide important cotton, orchard, vegetable and rice roosting and brood habitat for locally production in the Sacramento Valley. In the breeding ducks, mostly Mallard and Delta, islands were leveed to grow corn, Gadwall. barley and other grain crops, some of which The most significant change to waterfowl have value to ducks and geese. habitats in the Central Valley over recent Brackish marsh wetlands in the Suisun decades has been the development of rice Marsh historically were significant to agriculture, particularly in the Sacramento wintering waterfowl, but populations of Valley. Planted rice acreage has increased dabbling ducks and geese there have declined. from nearly 41,000 ha (1930s) to almost The Suisun Marsh currently provides 243,000 ha, and now averages > 202,000 ha wintering habitat for > 60,000 waterfowl, of (Petrie et al. 2014). Waste grain remaining in which dabbling ducks are the most numerous fields after harvest provides a valuable food (55,000), followed by diving ducks, geese, sea source for wintering waterfowl (Eadie et al. ducks, and swans (Ackerman et al. 2014). 2008). Along with the increase of planted Following decades of considerable landscape rice, there has been a significant change in changes, the Central Valley is left with merely management of residual rice straw after 162,000 ha of wetlands nested within a harvest. Before the 1990s, fire was the largely agricultural matrix. primary method for rice straw disposal.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 28 Autumn–winter habitat use by waterfowl

However, with air quality concerns, the Rice Sacramento Valley. Cinnamon Teal Anas Straw Burning Reduction Act of 1991 cyanoptera were an exception and did not shift mandated that burning of straw be reduced northward. In contrast to dabbling ducks, and currently less than 10% of all harvested the percentage of diving ducks using the rice fields are currently burned. As an San Joaquin and Tulare Basins increased alternative, rice growers turned to post- concurrently with a decrease in diving ducks harvest flooding, accompanied by disking, using the Suisun Marsh and Delta. Use of the rolling or chopping of straw. The result was Suisun Delta and San Joaquin Valley declined that flooded rice fields provided valuable for geese, with concomitantly large increases foraging habitat to a diversity of dabbling in the Sacramento Valley. Thus, the Central ducks and geese. At the peak, > 141,000 ha Valley has experienced substantial shifts in of harvested rice fields were flooded in the distributions of all waterfowl, reflecting autumn, nearly 70% of the planted rice significant changes at the 2nd order level of acreage (Central Valley Joint Venture 2006; habitat selection. Petrie et al. 2014). Most of these distributional shifts of Waterfowl wintering in the Central Valley waterfowl in the Central Valley have been have responded strongly to these changes at driven by the large-scale changes in habitat both 2nd and 3rd orders of habitat selection. availability and 3rd (and possibly 4th) order Timing and distribution of 2nd order levels of habitat selection. Currently, dabbling selection by waterfowl have been altered ducks in the Central Valley rely on three major considerably with the draining of Tulare Lake habitat types: 1) flooded harvested rice fields, and increase of rice agriculture in the 2) managed seasonal wetlands, and 3) flooded northern reaches of the valley. Fleskes et al. and unflooded harvested corn fields (Central (2005) reported that the total area of Valley Joint Venture 2006). Geese in the valley croplands intentionally flooded in winter also use unflooded rice fields and uplands. increased by 157% in the Sacramento Valley Petrie et al. (2014) estimated that winter- and 58% in the Delta, but declined by 23% in flooded rice fields provided 44% of all food the San Joaquin Valley between 1973 and energy available to dabbling ducks in flooded 2000, leaving only 3% of the total winter- habitats in the Central Valley, while flooded flooded agricultural land in the latter region. and unflooded rice fields provided 49% of all In response, birds have shifted winter food energy available to dark geese but 73% distributions northward. Fleskes et al. (2005) of all food energy for white geese. These conducted extensive surveys and radio- results were corroborated by Fleskes et al. telemetry in 1998–2000 and compared results (2005); they reported the importance of to data from 1973–1982 (Heitmeyer et al. agricultural habitat (relative to managed 1989; Miller et al. 1993; Miller et al. 1995). The wetlands) for Northern Pintail, Mallard recent research indicated that the percentage and Greater White-fronted Geese Anser of dabbling ducks using the Tulare basin and albifrons was greater than 20–30 years ago, the San Joaquin Valley declined, especially in presumably as birds increased their use of late winter, while use increased in the flooded rice fields.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 29

In addition to the above patterns, the producing moist-soil plants. The decline importance of managed wetlands has of Northern Pintail has resulted in increased in the Suisun Marsh. Most management of seasonal wetlands toward waterfowl that winter in Suisun Marsh are more densely vegetated marshes favoured by dabbling ducks, which primarily use Mallard. This technique has reduced amount managed wetland habitats provided by duck of sparse and short vegetation which is likely hunting clubs and state wildlife areas more representative of seasonal flooded (Ackerman et al. 2014). Coates et al. (2012) wetlands sought historically by Northern radio-marked and relocated 330 female Pintail. The greatest recent change in the Northern Pintail in the Suisun Marsh to Central Valley has been the considerable estimate resource selection during non- increase in rice acreage, especially in the breeding months and found strong evidence Sacramento Valley. This change has led to a for selection of managed wetlands. northern shift from the San Joaquin Valley Ackerman et al. (2014) reanalysed Northern by most species (2nd order habitat selection) Pintail telemetry data to examine habitat and a substantial increase in use of flooded selection. They compared spatial patterns of and unflooded rice fields as foraging habitat habitat use by ducks to availability of (3rd order). Indeed, rice landscapes have habitats at two spatial scales and found that become so important to wintering waterfowl Northern Pintail strongly selected managed that decline or loss of this agriculture would wetland habitats at both small and large seem catastrophic to Northern Pintail and scales. Further, Northern Pintail avoided likely other wetland-dependent birds (Petrie tidal marshes, bays, sloughs and some other et al. 2014). Nearly half of all duck-use-days habitats (Ackerman et al. 2014). These in the U.S. portion of the Pacific Flyway results have important implications for occur in the Central Valley, and loss of rice Northern Pintail given current efforts to would have continental impacts on Northern restore large portions of the Suisun Marsh Pintail and other waterfowl using ricelands to tidal wetlands. The consequences for (Petrie et al. 2014). However, the future of dabbling ducks using the marsh have not yet flooded rice as winter habitat for waterfowl been thoroughly assessed, and loss of is in question with recent record droughts, managed wetlands in the Suisun Marsh water requirements for in-stream flows to remain a concern for waterfowl managers meet needs of several species of federally (Ackerman et al. 2014). endangered fish, and ever-growing urban Patterns of habitat selection by waterfowl demands. Petrie et al. (2014) estimated in the Central Valley represent large-scale that > 75,000 ha of additional managed shifts in the area and type of habitats moist-soil wetlands would be required to available; as a consequence, significant replace the waterfowl food value provided by changes in 2nd and 3rd order habitat existing ricelands in the Central Valley. While selection have occurred by many species of rice agriculture is unlikely to disappear from ducks and geese. Most remaining wetlands the valley, the total acreage and the way are intensively managed to produce seed- it is managed post-harvest are uncertain.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 30 Autumn–winter habitat use by waterfowl

Understanding the shifting mosaic of Pacific Flyway population of brant use available winter habitats and bird responses Humboldt Bay as a migratory stopover from will be an ongoing research need to guide late February through to mid-April. conservation initiatives. Inter-mountain West and Great Salt Pacific Coast Lake San Francisco Bay is the largest estuary The Inter-mountain West region comprises along the west coast of the continental U.S. two regions of special importance to and historically important migration and non-breeding waterfowl: Southern Oregon wintering grounds for sea and other diving Northeastern California (SONEC), ducks (Conomos et al. 1985; Hothem et al. including the Klamath Basin, and the Great 1998). More than 85% of the tidal wetlands Salt Lake. The SONEC region covers of the Bay have been lost to agriculture approximately 10% of the Great Basin, and development in the 20th Century although waterfowl habitat comprises a (Nichols et al. 1986; Hothem et al. 1998). much lower percentage (Petrie et al. 2013). Anthropogenic changes and impacts have Historically, peak waterfowl abundance affected numerous waterfowl and other occurred during autumn and spring birds, including Canvasback Aythya valisineria migration. Migrating waterfowl in autumn whose overwintering numbers dropped by likely would have experienced dry 50% during the 1970s–1990s (Hothem et al. conditions and were probably restricted 1998). Despite habitat modifications, San to a few large complexes of permanent or Francisco Bay may harbour nearly 50% of semi-permanent wetlands (Petrie et al. the total population of several diving duck 2013). Few birds remained over winter species during winter (Accurso 1992; Brand because of the below-freezing winter et al. 2014). Given the history of mining in temperatures. Today, nearly all autumn and California, the position of the San Francisco winter waterfowl habitat in SONEC occurs Bay makes it susceptible to accumulating on public land. Two refuges are of particular contaminants such as mercury, cadmium significance: Lower Klamath National and selenium (Heinz et al. 1989; Hothem et Wildlife Refuge (Lower Klamath) and al. 1998). the Tule Lake National Wildlife Refuge Farther up the northern California coast, (Tule Lake). Although these refuges account the coastal lowlands are important for only a fraction of the region, they migration and wintering areas for > 20 support a significant portion of the species of waterfowl, with populations waterfowl that use SONEC in autumn ranging from 25,000–100,000 birds per day and winter (Kadlec & Smith 1989; Fleskes from autumn through spring (Pacific Coast & Yee 2007). In fact, the Klamath Basin Joint Venture 2004). Humboldt Bay is is recognised as a region of continental particularly important for brant because of significance to North American waterfowl its extensive Common Eelgrass Zostera populations (NAWMP Plan Committee marina beds. An estimated > 40% of the 2004).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 31

Management of waterfowl habitats on arriving from northwestern and mid- Lower Klamath and Tule Lake refuges continent Canada and Alaska, and some depends on water supplies. Increasing from the Prairie Pothole Region. Banding demands for water within the Klamath data indicate that many ducks that migrate Basin by farmers, native communities and through the GSL spend the winter in the endangered fish have hindered refuges from Central Valley of California and west coast obtaining sufficient water for waterfowl. A of Mexico (Petrie et al. 2013). Use of GSL recent analysis using bioenergetics models by waterfowl is lowest in mid-winter but (TRUEMET) indicated that food resources increases during spring. Dynamic ebbs and at Tule Lake were adequate to meet energy flows of water and fluctuating lake salinities needs of diving ducks and swans, but were are significant in maintaining this productive insufficient for dabbling ducks and geese. wetland system (Petrie et al. 2013). Food for dabblers was exhausted in early The Inter-mountain West Joint Venture autumn, well before traditional peak estimated 17.4 million waterfowl-use-days migration in November (Petrie et al. 2013). of the GSL during winter of which dabbling Thus, dabbling duck numbers at Tule Lake ducks accounted for 74% (Northern Pintail have declined significantly since the 1970s. = 39% of dabbling duck use-days; Green- The SONEC region is also critical during winged Teal = 23%; Mallard = 21% and spring migration, especially for Northern Northern Shoveler = 11%), while diving Pintail. Over 70% of habitat use by radio- ducks comprised 19% of total waterfowl- marked Northern Pintail in SONEC use-days during winter, with Common (outside of the Lower Klamath) occurred on Goldeneye Bucephala clangula representing privately-owned habitats, primarily flood- 91% of all diving duck use (Petrie et al. irrigated agriculture (Fleskes et al. 2013). 2013). Bioenergetics analyses of food The Great Salt Lake (GSL) is one of the supplies in the GSL needed to support largest wetland complexes in western migratory waterfowl suggested that seed U.S. and is recognised internationally for resources required by dabbling ducks were its importance to migratory waterfowl depleted during autumn migration by late (NAWMP Plan Committee 2004). As many October (Petrie et al. 2013). Yet, there may as 3–5 million waterfowl migrate through have been > 1 million dabbling ducks alone the GSL annually (Petrie et al. 2013). The in the GSL in October and November. GSL is surrounded by >190,000 ha of These results suggest that dabbling ducks wetlands maintained by fresh water from are obtaining unknown but critical energy rivers that flow into the basin. The supplies from perhaps aquatic invertebrates, surrounding marshes are extensive and submerged aquatic vegetation, tubers, or a provide rich diversity of invertebrate and combination of these (Petrie et al. 2013). plant food resources (Petrie et al. 2013). Petrie et al. (2013) concluded that improved Waterfowl use of the GSL is greatest during understanding and estimation of the late summer – early autumn and also in spatiotemporal variability of wetland spring. Peaks occur in September, with birds resources and waterfowl resource selection

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 32 Autumn–winter habitat use by waterfowl in the GSL system were needed to refine abundance and influenced the distribution assumptions about the foraging guilds. and timing of migration of swans and some goose populations. Further south along the Europe flyway, wintering waterfowl, especially ducks As in North America, substantial changes (e.g. Eurasian Wigeon Anas penelope), have in land use and management have occurred largely switched from using marine habitats in Europe since the early 20th century, to freshwater wetlands during daylight hours where landscapes at staging and wintering as the latter have increasingly been managed areas for waterfowl are now a matrix of as nature reserves since the 1950s (e.g. Owen agricultural land and other habitats greatly & Williams 1976; Guillemain et al. 2002). transformed by humans (e.g. industrial and Reserves nowadays not only provide safety residential zones) which envelop small from hunting and other human disturbance, protected areas of remaining wetlands but habitats are managed specifically for (Thomas 1976; Owen et al. 1986; Tamisier waterfowl. Yet despite active habitat & Grillas 1994; Guglielmo et al. 2002). management, there is an increasing Autumn-migrating Western Palearctic awareness that alien species (e.g. Red Swamp waterfowl largely concentrate in a flyway Crayfish, Procambarus clarkia and Water corridor along the Baltic and North Sea Primrose Ludwigia sp. and Swamp Stonecrop coasts (e.g. Scott & Rose 1996; Söderquist et Crassula helmsii) are a threat to protected U.S. al. 2013; Calenge et al. 2010). Here, the habitats and European wetlands (e.g. global concerns of sea level rise and other Dandelot et al. 2005; Meineri et al. 2014). loss of habitat associated with climate Along the Mediterranean coasts, primary change are serious concerns for waterbirds wintering habitats of waterfowl are brackish in coastal wetland habitats (e.g. Clausen & lakes, lagoons and temporary wetlands. Clausen 2014), which are further threatened Wetlands of the Mediterranean region by eutrophication (e.g. declines in seagrass have been reduced by 80–90% by urban beds, Clausen et al. 2012) and the population growth and conversion to encroachment of vegetation that is less agriculture (Toral & Figuerola 2010). nutritious for waterfowl (e.g. Common Fortunately, some of these are now rice Cord-grass Spartina anglica; Percival et al. fields which, as in North America, provide 1998). In contrast, climate warming and valuable resources to wintering waterfowl increased fertilisation of grasslands in (e.g. Tamisier & Grillas 1994) and help northwest Europe may have enhanced compensate lost wetland habitats (e.g. terrestrial habitats for geese, where several Tourenq et al. 2001; Rendón et al. 2008). In populations are flourishing, and some are the Camargue, southern France, portions of short-stopping or becoming partly non- remaining natural wetlands are protected migratory (e.g. Greylag Geese Anser anser, and most are on private estates, wherein Voslamber et al. 2010; Barnacle Goose temporary and seasonal wetlands are Branta leucopsis, Ganter et al. 1999). Hunting flooded beyond natural hydroperiods to restrictions also have likely enhanced the attract waterfowl for hunting and observing.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 33

This practice is detrimental to wetland interior wetlands, cattle ponds, irrigation and biodiversity in general, but it has greatly flood-control reservoirs, playa lakes, seasonal promoted hydrophyte beds on which wetlands, riparian and flooded forest waterfowl forage (Tamisier & Grillas 1994). wetlands, rivers and irrigation canals, plus Such management is mostly beneficial to flooded and dry agricultural lands including herbivorous species (e.g. Gadwall) but the grain and legume crops within their other dabblers also benefit from seeds geographic ranges from the Gulf Coast to spread as bait in these properties (Brochet et southern Canada (Jorde et al. 1984; Chabreck al. 2012). Hunting management practices et al. 1989; Miller et al. 2000; Link et al. 2011). could likely be responsible for considerable In the Atlantic Flyway, Mallard use coastal improvement of wintering body condition and inland freshwater emergent marshes of Common Teal (up to 12%) and other and managed wetlands developed from dabbling ducks in past decades (Guillemain 18th century rice fields (Gordon et al. 1989, et al. 2010b). 1998). Perhaps most intriguing is the winter residency of some Mallard along the sandbar Habitat resources of selected flats of the Missouri River in North northern hemispheric waterfowl Dakota, where these birds tolerate frequent Dabbling ducks inhospitable winter conditions while largely subsisting on Rainbow Smelt Osmerus mordax Mallard (Olsen & Cox, Jr. 2003; Olsen et al. 2011). Mallard challenge clear distinctions of The MAV is considered the ancestral autumn migration and subsequent winter wintering grounds of North American habitat distributions because of great Mallard (Nichols et al. 1983; Reinecke et al. seasonal and annual variation in settling 1989; Heitmeyer 2006). Nichols et al. (1983) by individuals or sub-populations within examined winter distributions of Mallard flyways. The breadth of habitats occupied by and found support for the flexible homing Mallard in North America is particularly hypothesis, given that Mallard wintered fascinating. In the Sacramento Valley farther south in United States during wetter of California, Mallard use agriculturally and colder winters (also see Green & dominated and largely treeless environments, Krementz 2008). Mallard typically migrate where patches of seasonally flooded and in autumn from latitudes of central Missouri emergent wetlands and flooded rice fields after cumulative days of temperatures of mostly occur, notwithstanding the Butte Sink ≤ 0°C, snow cover and ice conditions (i.e. wherein riparian wetlands consisting of weather severity index (WSI) of ≥ 8; willow Salix sp., California Sycamore Platanus Schummer et al. 2010). A quadratic and recemosa, Buttonbush Cephalanthus occidentalis cumulative WSI model explained ≥ 40% of and other woody and herbaceous species the variation in changes in relative exist (Gilmer et al. 1982; Heitmeyer et al. 1989; abundance of Mallard and other dabbling Eadie et al. 2008; Elphick et al. 2010). In ducks in Missouri during autumns–winters Central U.S., Mallard use Gulf coastal and 1995–2005 (Schummer et al. 2010, 2014).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 34 Autumn–winter habitat use by waterfowl

Recent capture-recapture results suggest to foraging, weather, disturbance or a similar patterns in Europe (Dalby 2013). combination of these and other factors Interestingly, satellite-marked Mallard in the related to survival during winter. For example Mississippi Flyway (Krementz et al. 2012) and relative to 3rd and 4th order selection, revealed patterns of incremental migrations Mallard used irrigation canals in Nebraska similar to those described by Bellrose agricultural landscapes over nearby natural (1980). riverine wetlands during harsh winters Mulhern et al. (1985) investigated use and because canals were climatically more suitable selection of wetlands by Mallard broods in than other habitats (Jorde et al. 1984). Saskatchewan and found that broods used Additionally, Mallard may exercise trade-offs structurally different wetlands, but use was in by selecting habitats of perhaps lesser proportion to availability of wetland types foraging quality but prone to fewer and thus not selective. How this apparent disturbances which contribute to greater plastic habitat use by brooding ducks may survival. Krementz et al. (2012) postulated ramify into habitat use subsequently during that Mallard may forego wintering in the autumn and winter unearths interesting Grand Prairie region of Arkansas to avoid questions: 1) What drives individuals to seek this area because of intense hunting pressure. and use diverse habitats? 2) What are survival and fitness outcomes related to these Northern Pintail decisions? 3) What non-breeding habitat Similar to their reliance on rice in California’s complexes are associated with greatest Sacramento Valley, ~52% of all locations survival rates of individuals? 4) Where do (n = 7,022) of radio-marked Northern these birds breed, and what are their Pintail females were in rice habitats, which reproductive outcomes? For example, do included active (18% use) and fallow rice more competitive or fit Mallards occupy the fields (34% use) along the coast of Texas MAV, the supposed region of greatest habitat (Anderson & Ballard 2006). Many radio- quality for the species (Nichols et al. 1983), marked female pintail that were located whereas other Mallard distribute to other > 64 km from the Texas rice prairies flew to regions? Alternatively, perhaps the regions rice field habitats at some point during occupied have little influence on fitness winter, which demonstrated the importance prospects, so long as adequate food, of flooded ricelands to pintail in this region freshwater and potential mates are available. (Anderson & Ballard 2006). In Louisiana, As previously mentioned, evidence exists that Cox, Jr. and Afton (1997) found extensive habitat complexes used by the greatest use of sanctuaries by radio-tagged Northern densities of Mallard and those individuals Pintail during hunting seasons, but less so with greatest winter survival rates in the MAV before and after legal waterfowl seasons. differ in habitat composition (Pearse Female pintail used flooded rice and fallow et al. 2012; Lancaster 2013; Kaminski & Davis fields nocturnally where combined these 2014). Drivers of differential habitat use habitats accounted for 68–93% of nocturnal are not always clear but are likely related use by the birds (Cox, Jr. & Afton 1997).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 35

In California, Fleskes et al. (2007) Baldassarre 2014). Despite being a forested attributed greater survival of Northern wetland specialist, wherein Wood Duck Pintail to increased area of flooded rice forage on red oak Quercus sp. acorns and habitats. Other landscape factors important aquatic invertebrates (Heitmeyer et al. 2005; to pintail survival, such as the size and Foth et al. in press), Wood Duck also use management of sanctuaries, types of flooded croplands where they forage on feeding habitats (e.g. rice, wetlands) and the waste agricultural seeds (Delnicki & juxtaposition of these, may also have Reinecke 1986; Bellrose & Holm 1994; been important (Fleskes et al. 2007). Barras et al. 1996; Kaminski et al. 2003). Nonetheless, contemporary (1998–2000) Much of the non-breeding information survival estimates (87–93%) of adult female about Wood Duck is derived from eastern Northern Pintail in the Suisun Marsh and populations and birds using the MAV and Sacramento and San Joaquin Valleys were southern Atlantic Flyway (Arner & Hepp greater than in any other region of North 1989; Reinecke et al. 1989; Peterson 2014), America (Fleskes et al. 2007). Clearly, but much remains to be learned about non- sanctuaries adjacent to rice and other breeding Wood Duck use and selection of agricultural habitats are critical to survival unique habitats in regions such as the and habitat use by Northern Pintail Central Valley of California and even xeric throughout their wintering range (Cox, Jr. & environments in Nevada that lack traditional Afton 1997; Fleskes et al. 2007). expansive bottomland hardwood forests (Baldassarre 2014). Wood Duck The North American Wood Duck is the Diving Ducks only Aix species in the Nearctic (Birds of Ducks that are among the more ecologically North and Middle America Check list; pelagic have historically used estuarine or http://checklist.aou.org/). Wood Duck freshwater systems, usually along coastlines, are also unique among North American shorelines of lakes and major rivers waterfowl, because they are the only (Bellrose 1980). The significance to diving species with migratory and non-migratory ducks Aythya sp. of myriad bays of North populations (Baldassarre 2014). Wood Duck America, including Chesapeake and San have been widely studied in North America Francisco Bays, has been recognised for since their near extirpation in the early 20th centuries (Audubon 1840; Haramis 1991a,b; century (Bellrose & Holm 1994). Migration Perry et al. 2007). Unfortunately, these routes of Wood Duck are not well defined, systems are often plagued by anthropogenic given the substantial overlap in breeding and effects of shoreline development, boat winter ranges (Baldassarre 2014). Given traffic, increased sediments and nutrients their broad occupancy of geographic areas, and other factors (Perry et al. 2007; Lovvorn Wood Duck use diverse freshwater et al. 2013). Knowledge of niche overlap and wetlands, although they avoid brackish and “carrying capacity” of habitats by these marine systems (Bellrose & Holm 1994; ducks is necessary to understand relations

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 36 Autumn–winter habitat use by waterfowl between birds and potential invertebrate or especially problematic to Lesser and Greater other prey (Lovvorn et al. 2013). Scaup as they comprised as much as 43–47% Diving ducks wintering in Chesapeake Bay of all waterfowl in the Bay. Richman and from 1950–1995 comprised 23% of Atlantic Lovvorn (2004) collected Lesser Scaup in Flyway and 9% of North American winters 1998–2000 and found that 98% of populations of these ducks (Perry & Deller clams consumed by scaup were Asian Clams. 1995; Perry et al. 2007). Some species Asian Clams apparently provide scaup with a wintering in Chesapeake Bay have been more profitable food source, because they mostly adversely affected than others. For example, are distributed in the top 5 cm of sediments Redhead and Canvasback that feed on where scaup intake rates are greatest submerged aquatic vegetation (SAV), seeds (Richman & Lovvorn 2004). Additionally, and tubers have been impacted more than Lesser and Greater Scaup and Surf Scoter species that forage in slightly deeper water on Melanitta perspicillata wintering in San invertebrates, particularly Lesser Scaup Francisco Bay had decreased body mass and (Perry et al. 2007). Increased nutrients and fat and increased foraging effort, causing sedimentation have lessened SAV in shallower them to disperse from upon food limitation. reaches of Chesapeake Bay (Perry et al. 2007). There also was substantial niche overlap and Moreover, recently expanding hypoxic zones opportunistic use of dominant prey species may be negatively impacting sessile prey of by these ducks (Lovvorn et al. 2013). diving ducks (Perry et al. 2007) and have been Lovvorn et al. (2013) concluded that scaup linked to decreased body mass and survival in and scoter did not exploit a substantial Canvasback (Haramis et al. 1986). fraction of food above local profitability Pollutants and invasive species are thresholds before abandoning the habitat, thought to be especially problematic for and encouraged future research to better diving ducks such as scaup and Canvasback understand thresholds of energetic (Lovvorn et al. 2013). In San Francisco Bay, profitability for diving ducks. Hothem et al. (1998) found that mercury Despite vast size and dynamics of San and selenium levels in late winter had Francisco Bay, adjacent habitats in the region accumulated in scaup and Canvasback to provide vital resources for some species levels that impair reproduction in game-farm using the Bay. Specifically, estuarine intertidal Mallard (Heinz et al. 1989). Invasive species, and subtidal mudflats and salt ponds provide such as Asian Clam Potamocorbula amurensis, additional food and water for diving ducks which has displaced the former bivalve (Dias 2009; Brand et al. 2014). Brand et al. prey community (e.g. Macoma balthica), are (2014) found that diked salt ponds, salt pans considered a second primary concern for and managed seasonal wetlands in South San diving ducks in the Bay (Richman & Francisco Bay collectively provided enough Lovvorn 2004; Lovvorn et al. 2013). Asian food energy to sustain 79% of the energy Clams may harbour greater levels of and nutrients required by diving ducks when selenium than other bivalve species birds were at maximum numbers, and (Richman & Lovvorn 2004), which could be basically 100% of the nutrients when

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 37 average bird abundances prevailed. Managed and scaup typically use open-water areas ponds serve as important roosting and only (Korschgen 1989; Roy et al. 2013). foraging habitats in this region. Ponds that Elsewhere herein, Stafford et al. (2014) intake, circulate or discharge water directly to provided a detailed account of scaup habitat or from the Bay or adjacent sloughs use during late winter and spring migration. supported > 95% of the diving duck Diverse coastal and interior wetlands of abundance (Brand et al. 2014). However, south-central Louisiana are critical to diving greater bird and invertebrate abundances and ducks such as Redhead and Canvasback prey energy density occurred in meso-haline (Hohman & Rave 1990; Hohman et al. (i.e. 5–30 ppt) rather than low-hypersaline 1990). Canvasback in the Mississippi River (i.e. 31–80 ppt) circulation ponds (Brand et al. Delta and at Catahoula Lake in Louisiana, 2014). Ruddy Duck Oxyura jamaicensis both important wintering areas to these exercise dietary flexibility in these same species (Hohman et al. 1990), consumed wetland complexes, feeding on amphipods about 97% plant matter at each site, with Amphipoda sp. or polychaetes Polydora sp. below-ground plant biomass composing depending on prey occurrence or abundance 94% aggregate dry mass (Hohman & Rave among different wetland types (Takekawa et 1990). Mudflats with tubers or water that al. 2009; Brand et al. 2014). Thus, similar to permitted Canvasback to tip-up and feed identifying important habitat complexes for were important components of used Mallard or other dabbling ducks (Pearse et al. habitats (Hohman & Rave 1990). Similar to 2012; Lancaster 2013), maintaining diverse plant-eating Canvasback, the importance of foraging wetlands in ecosystems like San Shoal Grass Halodule wrightii to Redhead Francisco Bay is imperative for supporting and several avian guilds has long been waterfowl and other wetland dependent mentioned (Cornelius 1977; Michot et al. birds using this system (Brand et al. 2014). 2008). Redhead wintering in the Chandeleur A primary difference between historical Sound of Louisiana and Laguna Madre, and contemporary habitat use for some Texas consumed as much as 74% dry diving ducks, such as Ring-necked Duck in mass of shoalgrass (Michot et al. 2008). the U.S., has been a shift away from Conserving Halodule beds arguably is the traditional winter habitats to open-water most critical conservation priority within the lakes because of a proliferation of invasive winter range of Redheads, particularly given plants such as Hydrilla Hydrilla verticillata and that most of the North American other species that form dense floating mats population of the species winters along (Johnson & Montalbano 1984; Roy et al. coastal habitats of Texas and Louisiana 2013). Some of the greatest wintering (Michot et al. 2008). concentrations of Ring-necked Duck may Sea ducks occur in managed impoundments of coastal and inland Louisiana (Roy et al. 2013). Ring- North America necked Duck use small marshes adjacent to There are 15 species of North American sea open water, whereas Canvasback, Redhead ducks (Tribe: Mergini) and arguably they are

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 38 Autumn–winter habitat use by waterfowl the least understood taxa of waterfowl and other invertebrates. These foods are (Bellrose 1980; Goudie et al. 1994; Silverman depauperate in energy density, so sea ducks et al. 2013). Evidence suggests that 10 must forage voraciously to maintain positive of these species are in decline, including energy balances (Systad et al. 2000). eight of 12 species that winter off the Surf Scoter Melanitta perspicillata and Atlantic coast of North America, a primary White-winged Scoter M. deglandi in the wintering area for this tribe (Sea Duck Joint Pacific Flyway use soft-bottom habitats and Venture 2004; Zipkin et al. 2010). Eleven forage on bivalves (Bourne 1984; Richman species of sea ducks commonly winter & Lovvorn 2003; Lewis et al. 2008). Scoters in Pacific coastal regions, nine of which encounter considerable variation in clam commonly occur in the Puget Sound of densities and potentially face an exhaustible Washington state (Faulkner 2013). Sea duck food supply (Lewis et al. 2008). However, declines are occurring concomitantly with Lewis et al. (2008) found that scoters in uncertainty about their habitat preferences Baynes Sound (British Columbia) did not (Zipkin et al. 2010). Shoreline development switch winter prey or move extensively to and associated pollution and climate change foraging sites, suggesting clam density was are potential negative influences on sea relatively high there (Kirk et al. 2007). ducks in North America (Zipkin et al. 2010). Sea ducks in the eastern U.S. have been Recent proposals for wind turbines along monitored by the Atlantic Flyway Sea Duck the Atlantic coast and threats from offshore Survey (AFSDS) in at least nine bays and energy development will also challenge sea sounds off of the Atlantic coast to quantify ducks, so further understanding of habitat winter distributions and population indices selection by these ducks is imperative (Migratory Bird Data Center 2009; Zipkin et (Zipkin et al. 2010). al. 2010). Zipkin et al. (2010) modelled Spatial distribution of sea ducks is effects of bottom depths, monthly averages generally determined by winter weather of sea surface temperature, and ocean floor conditions and habitat diversity (Zipkin et al. topography for five species of wintering sea 2010). At greater spatial winter ranges, food ducks. The North Atlantic Oscillation availability, local environmental conditions, (NAO; i.e. fluctuation in sea surface pressure habitat suitability, ocean depths and water across the northern Atlantic Ocean between temperatures influence sea ducks’ use of areas of high (Azores High) and low habitats (Lewis et al. 2008; Zipkin et al. (Icelandic Low) pressure: Ottersen et al. 2010; Dickson 2012). Northern seas are 2001; Stenseth et al. 2002; Hurrell et al. hostile during winter, with below freezing 2003; Zipkin et al. 2010) was the only temperatures, wind, ice and limited daylight environmental covariate that had a because the sun is below the horizon for significant influence on all five species; its two months (Systad et al. 2000). Sea effect was negative for the three scoter ducks, however, remain in these rigorous species and positive for Common Eider and environments during winter and forage Long-tailed Duck Clangula hyemalis (Zipkin et on molluscs, echinoderms, crustaceans al. 2010). These results suggest that climatic

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 39 conditions along the Atlantic coast during declining abundances of macroinvertebrates migration and winter may have direct or during winter. They concluded that indirect influences on sea duck distributions, exploitative competition was likely not perhaps as prey are re-distributed (Zipkin et occurring and interference competition al. 2010). Scoters predominated inshore appeared below thresholds that would cause during cold, snowy winters and Common birds to spatially segregate. Overall, winter Eider and Long-tailed Duck were more forage did not appear to limit habitat use of abundant inshore during wet, mild winters these species in Lake Ontario during winter (Zipkin et al. 2010). Sea surface temperature (Schummer et al. 2008). (SST) negatively affected Long-tailed Duck Chesapeake Bay is considered one of the and White-winged Scoter abundance but most important areas for several species of positively affected Common Eider, although scoters and Long-tailed Duck (Sea Duck there was some interaction of effects Joint Venture 2004; Ross et al. 2009), but between NAO and SST on birds’ habitat little is known about the birds’ use of the distribution. Overall, sea ducks may respond system. Surf Scoter M. perspicillata is thought to a combination of local habitat conditions to forage preferentially in subtidal, sandy and broader-scale weather patterns (Zipkin soft sediment habitats > 6 m deep (Ross et et al. 2010). Collectively, scoters used flatter al. 2009), but will also use hard-substrates bottom sites, which seemed consistent with (Lewis et al. 2007; Perry et al. 2007). Long- knowledge that Black Scoter Melanitta tailed Duck in the upper Chesapeake Bay americana, Surf Scoter and White-winged primarily consume bivalves (Perry et al. Scoter preferred sandier basins along the 2007), likely procuring food from soft- Atlantic shoreline (Stott & Olson 1973; sediment areas (Žydelis & Ruskete 2005; Zipkin et al. 2010). In contrast, Common Ross et al. 2009). Ross et al. (2009) suggested Eider used rugged substrates, but Long- that limited availability of hard substrate tailed Duck have not yet been linked to bottom in Chesapeake Bay might dictate bottom substrates (Perry et al. 2007; Zipkin habitat use patterns among these sea ducks et al. 2010). in the upper Chesapeake compared to other Other important habitats for non-breeding regions. Further concerns are linked to sea ducks in central and eastern North declining water quality since the 1960s in the America include the Great Lakes and lower region of Chesapeake Bay (Ross et al. Chesapeake Bay (Schummer et al. 2008). 2009). Excessive sedimentation and nutrient Mixed species of Bufflehead Bucephala albeola, loading have caused eutrophication and Common Goldeneye and Long-tailed Duck oxygen depletion, negatively affecting use inshore areas of Lake Ontario and forage portions of the Bay’s substrate, and are on energy-dense Amphipoda and larvae of linked to dramatic declines in seagrass beds Chironomidae, both abundant in the shallow- (Chesapeake Bay Program 2007; Ross et al. water zone near shore (Schummer et al. 2008). 2009). These consequences are problematic Despite concentrated mixed flocks of ducks, because seagrasses supply important Schummer et al. (2008) did not detect substrates for bivalves compared to bare

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 40 Autumn–winter habitat use by waterfowl ground under the Bay (Peterson 1982; populations (Skov et al. 2011). Moreover, Peterson et al. 1984; Ross et al. 2009). shipping and offshore wind farms may permanently displace sea ducks from Europe favoured feeding grounds (Petersen et al. Recent count data indicate that most 2006; Skov et al. 2011). Among sea duck European sea duck populations, with the species, the Long-tailed Duck is particularly exception of Common Goldeneye, are sensitive to wind farms (Petersen et al. now in decline (Hearn & Skov 2011; Skov 2006), and plans for offshore wind farm et al. 2011). Common Goldeneye winter construction exist in all Baltic countries. extensively in freshwater habitats along Traffic along the major shipping routes coastlines, whereas other sea ducks tend to (which cross or pass close to Long-tailed have an offshore distribution. The Baltic Sea Ducks wintering sites) is also predicted to is the key wintering area for most European increase (Skov et al. 2011). Oil illegally sea ducks, and it is a region of major discharged from ships continues to kill tens concern. Recent surveys indicate that Long- of thousands of birds each year, despite tailed Ducks, Velvet Scoter and Steller’s enforcement of international regulations Eider have declined by 65%, 55% and 66%, (Larsson & Tydén 2005; Skov et al. 2011; respectively, with declines in Common Brusendorff et al. 2013), and other Eider (51%), Common Scoter (47%), Red- hazardous chemicals are suspected of breasted Merganser Mergus serrator (42%) having a negative impact on Baltic wildlife and Greater Scaup (26%) also recorded (including sea ducks) when birds ingest (Skov et al. 2011). Declines have similarly bivalves or organisms that filter polluted sea been reported in other European countries, water (e.g. Pilarczyk et al. 2012; cf. Skov et al. notably in Britain and the Netherlands, 2011). Additionally, sea duck food resources which are also important wintering grounds in the Baltic Sea have changed substantially for European sea duck populations. in recent decades concomitantly with Generally, wintering sea ducks aggregate in nutrient loading. Increase of nutrient loads shallow coastal waters or over offshore after 1950 might explain rising bivalve banks where they can dive for food on the biomass in shallow waters, which in turn sea floor. In winter, > 90% of sea ducks use may have stimulated sea duck population areas amounting to < 5 % of the Baltic Sea growth. But decreases in nutrient loads (Bellebaum et al. 2012), where they forage (nitrogen and phosphorous) have occurred primarily on Blue Mussel Mytilus edulis. in some coastal regions since the 1990s, Ecosystem changes that have a negative whereas nutrient levels remain high in effect on habitat and food resources during other parts of the Baltic Sea. Declines in the non-breeding season (e.g. extraction of nutrient loads along the coastline and sand and gravel, dredging of shipping subsequent effects on sea duck food quality channels or coastal development), are need further investigation. Nevertheless, potentially the most important explanation Skov et al. (2011) stressed the importance of for the decline in arctic-breeding sea duck eutrophication in spatio-temporal variability

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 41 in food supply for and abundance of Geese waterbirds in the Baltic Sea, with control of Geese and agriculture eutrophication being a plausible reason for the decrease of several benthic species in As for ducks, habitat modifications influence Danish waters. distribution, movement and resource Phytoplankton composition also has exploitation in geese. Geese are generally changed in the Baltic, perhaps through the more adept at exploiting farm crops than increase in water temperatures in recent most duck species (Owen 1980), so their decades or overfishing leading to a decrease autumn and winter habitat use is largely in food quality for filter-feeding bivalve driven by and has changed markedly in mussels. In addition, in warmer waters response to variations in farming practices, mussels metabolise their own reserves both in North America and Europe, during during winter instead of hibernating, which the 20th and 21st centuries. For example, could decrease the quality of mussels for Pacific Flyway Greater White-fronted Geese bivalve feeders (Waldeck & Larson 2013). commonly stage in the SONEC, and then Lastly, overexploitation by commercial migrate and winter in the Sacramento and mussel fisheries (e.g. in the Wadden Sea) may San Joaquin Valleys (Ely 1992; Ackerman cause food shortages for bivalve feeding et al. 2006; Ely & Raveling 2011). species such as Common Eiders (Skov et al. Approximately 80% of foraging flocks of 2011). White-fronted Geese used harvested barley, Concomitant with warming temperatures wheat or oat fields from early September to of the Baltic Sea, ice coverage has decreased mid-October in SONEC, 1979–1982, then and permitted access to new wintering areas switched to potato fields by mid- for waterfowl. Common Goldeneye and October–late November of those years some Aythya species are shifting northward (Frederick et al. 1992; Ely & Raveling 2011). in their wintering distribution in the Baltic When White-fronted Geese migrated to the Sea (Skov et al. 2011; Lehikoinen et al. 2013). Sacramento Valley in autumn and winter, The limited degree of northward shift in the they primarily used complexes of rice field distribution of seaduck feeding offshore habitats (Ely & Raveling 2011). After White- suggests reduced food availability in the fronted Geese departed the Sacramento northern Baltic area, which is now partly ice- Valley for the San Joaquin Valley, green free in winter. Nevertheless, populations of forage, waste corn and other grain and some species including Common Eider have vegetable crops were available to the geese, relocated to the southwest Baltic Sea from but birds disproportionately used corn previous wintering quarters in northwest relative to its availability (Ely & Raveling Denmark. Lastly, European sea duck 2011). The future of Greater White-fronted populations also may be directly or Geese in the San Joaquin Valley is uncertain indirectly affected by commercial fishing because corn acreage declined there by 20%, and the use of gillnets for fishing (Žydelis et largely because of urbanisation (Ackerman et al. 2009). al. 2006; Ely & Raveling 2011). Changes in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 42 Autumn–winter habitat use by waterfowl agricultural practices and crops produced are uplands typical of Arkansas and playa eco- commodity-market driven and largely regions (Alisauskas & Hobson 1993). beyond the control of wildlife biologists, General plasticity of North American thus challenging to conservation planning white geese in exploiting agricultural and (Ely & Raveling 2011; Skalos 2012; Petrie et marsh habitats (Bateman et al. 1988; al. 2014). Alisauskas 1998; Jefferies et al. 2004) creates Another striking example of dynamic complex challenges in arresting the growth of habitat use by geese within agricultural overabundant populations in the 21st century landscapes comes from the North American (Batt 1997; Jefferies et al. 2004; Abraham et al. Snow Geese and Ross’ Geese Chen rossii 2005). However, dwindling rice acreage in (Ankney 1996; Abraham et al. 2005). White Texas may influence white goose population goose use of waste grain is well documented levels. For example, rice acreage was in the literature (Alisauskas et al. 1988; ~203,152 ha and white geese numbered > 1.2 Ankney 1996; Alisauskas 1998; Abraham et million in 1979; whereas ~378,000 geese al. 2005). Recent research has sought to were counted and only > 54,000 ha of rice identify winter origins of white geese existed in Texas in 2013 (K. Hartke, Texas migrating through Nebraska’s Rainwater Parks and Wildlife, unpubl. data). The Basin, a region of continental significance to contemporary estimate of rice acreage is the autumn and spring migrating waterfowl and lowest ever for Texas since records originated Sandhill Crane Grus Canadensis (Krapu et al. ca. 1948 (K. Hartke, Texas Parks and Wildlife, 1984; Alisauskas & Ankney 1992; Alisauskas unpubl. data). 2002; Stafford et al. 2014). Henaux et al. Similarly, contemporary estimates of (2012) used stable isotope analysis and found geese wintering in the Western Palearctic are flexibility in diets and regional landscape use 4.8 million, up from 3.3 million in 1993 (Fox by Snow Geese. They determined origins of et al. 2010). Most species exhibit signs of wintering Snow Geese harvested in the exponential increase, whereas others (e.g. Rainwater Basin as follows: Louisiana (53% the Greenland White-fronted Goose Anser and 9% in 2007 and 2008, respectively), albifrons flavirostris, Red-breasted Goose Texas Gulf Coast (38% and 89%, Branta ruficollis and Dark-bellied Brent respectively), Arkansas (9% and 2%, Goose Branta bernicla) have declined in respectively). However, no birds from the recent years (Fox et al. 2010). Although Playa Lakes region were detected. Beyond reduced hunting pressure on geese in some annual variability in their winter origins, regions probably played an important role, differences in diet also helped to characterise increases in most species of European geese their winter habitat use. Snow Geese relied have likely resulted from exploitation of on rice and wheat fields (C3 plants isotopic grains and root and grass crops, similar to signature) as well as corn and grain sorghum patterns in North America (Abraham et al.

(C4 plants). Geese collected from Texas and 2005; Fox et al. 2010). Since the 1950s, wild Louisiana were generally characterised by geese wintering in the western Palearctic using estuarine and marsh habitats versus have partially or completely switched from

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 43 feeding on natural vegetation to managed Algae Ulva sp. beds also serve as important pastures and agricultural croplands (Madsen food in coastal areas in the Atlantic Flyway 1998; Jensen et al. 2008; Hake et al. (Lewis et al. 2013). Brant exhibit different 2010). Agricultural producers in Europe foraging strategies in Atlantic coastal states have been concerned with losses of wheat of New York, New Jersey and Virginia, and oilseed as goose and swan populations where brant select eelgrass, cordgrass have increased (Dirksen & Beekman Spartina sp. or exploit grasses and clover in 1991; Rees et al. 1997). Several measures upland habitats (Smith et al. 1985). Smith et have attempted to deter geese from al. (1985) attributed diet switching by brant crops, including providing supplemental from eelgrass to other foods because of feed in accommodation fields to influence eelgrass declines. However, brant foraged movements of and use by geese, scaring of on cultivated grass and clovers in New birds, fencing habitats and adjusting farming York, despite an increasing trend in strategies, such as growing barley varieties availability of SAV in the state. They that mature and are harvested before attributed differential feeding strategies varieties used previously (Hake et al. 2010). among regions to the birds’ winter philopatry and social organisation. Black Brant and estuarine-marine systems Brant have been negatively affected by Besides agricultural lands, estuarine and loss of eelgrass habitats in the North marine wetland systems are critical to many American Atlantic Flyway and Europe waterfowl, including Black Brant in North (Vickery et al. 1995; Ganter et al. 1997; Ward America (named Brent Goose in Europe). et al. 2005; Shaughnessy et al. 2012). Brant in Important autumn staging areas for brant those regions use eelgrass where available, include shallow marine waters along but birds also exploit salt marsh habitat. shorelines, within lagoons or behind barrier Moreover, European birds have moved beaches (Shaughnessy et al. 2012; Lewis et al. inland to use golf courses and pastures with 2013). Some of the important habitats for cattle (Vickery et al. 1995; Ganter et al. 1997; the nine-month non-breeding period of Ward et al. 2005; Shaughnessy et al. 2012). brant include the Northeast Pacific United Lovvorn & Baldwin (1996) recognised the States, the lagoons along the west coast value of habitat complexes for wintering of Baja California, areas of Mexico and brant in Western Europe that include Atlantic coastal habitats (Smith et al. 1985; intertidal flats, bays and other permanent Lewis et al. 2013, Martínez Cedillo et al. wetlands that provide sea grasses, as well as 2013). Pacific Black Brant solely use natural nearby farmlands containing waste grains habitats during winter and avoid agricultural and natural seeds. This complex of suitable lands (Ward et al. 2005; Lewis et al. 2013). As habitats allow brant to move and forage mentioned, the unifying food resource for among them and thereby enhance their Holarctic brant is eelgrass (Moore et al. survival (Lovvorn & Baldwin 1996). 2004; Moore & Black 2006; Shaughnessy et However, synergistic effects of climate al. 2012; Lewis et al. 2013). Macrogreen change, possible negative effects on sea level

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 44 Autumn–winter habitat use by waterfowl rise and declining eelgrass communities are 200,000 birds) between 1955–1989, leading emerging concerns for waterfowl ecologists to regulated hunting of the species in some conserving brant (Shaughnessy et al. 2012). states (Serie & Bartonek 1991). In contrast, although the Northwest European Bewick’s Swans Swan population similarly rose from Migratory swans ~16,000 birds in the mid-1980s to a peak of Of the five swan species and subspecies in ~29,000 individuals in the mid-1990s, its the northern hemisphere, the Tundra Swan numbers are now in decline (Rees & (a.k.a. Whistling Swan) C. c. columbianus and Beekman 2010), with several poor breeding Trumpeter Swan of North America and the seasons in recent years probably a major Bewick’s Swan C. c. bewickii (conspecific with contributing factor. the Tundra Swan) and Whooper Swan C. The Eastern Population of Tundra Swans, cygnus in Eurasia are all migratory, whereas which breeds across northern Canada and the Mute Swan C. olor is relatively sedentary north of the Brooks Range in Alaska, in its native Europe and in North America migrates to the U.S. eastern seaboard where it has colonised (e.g. Petrie & Francis (allocating about half their time between 2003). Trumpeter Swans were widespread in boreal forest and northern prairie-Great North America prior to 1900 (Rogers & Lakes habitats during autumn migration; Hammer 1998; Engelhardt et al. 2000), but Weaver 2013), whereas the Western hunting caused their numbers to drop nearly Population, which breeds in coastal regions to extinction by the early 20th century, and of Alaska south of the Brooks Range, use of established migration routes waned migrates to western North America to winter (Gale et al. 1987; Mitchell & Eichholz 2010). mainly on the Pacific coast from Vancouver Legal protection from persecution (since Island to central California, and the inland the 1918 Migratory Bird Treaty) and valleys of California (Bellrose 1980; Ely et al. more recent conservation measures (e.g. 2014). The Northwest European Bewick’s habitat protection and reintroduction Swan population also migrates along a well- programmes) saw Trumpeter Swan numbers defined corridor, from breeding grounds in recover to ~16,000 birds by 1990, and > the Russian arctic along the arctic coast and 34,000 free-ranging swans were estimated in across Karelia to autumn staging sites on the 2005 (Moser 2006; Mitchell & Eichholz Baltic (particular Estonian wetlands) and 2010). Whooper Swan numbers have also wintering grounds in northwest Europe. increased in Europe in recent decades Whooper Swans are thought to migrate (Wetlands International 2014), and the on a broader front (Garðarsson 1991; Tundra Swan – the most numerous and Matthiasson 1991), but like the arctic-nesting widely distributed of North American swans they show strong fidelity to staging swans – is likewise increasing. Indeed, and wintering sites (Bellrose 1980; Black & agricultural foraging opportunities are Rees 1984; Rees 1987). thought to have contributed to a near Historically, migratory swans fed on SAV doubling of Tundra Swan numbers (to > during autumn and winter, often reflecting

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 45 regional and seasonal variation in availability 2010). The drivers of swan exploitation of and dietary requirements. For Tundra Swans, arable lands remain unclear; however, historic this included Arrowhead Sagittaria sp., and novel food availability, nutrition and Sago Pondweed Potamogeton pectinatus and foraging efficiency in croplands may be Wild Celery Vallisneria americana (Bellrose influences (Rees 2006). Several studies have 1980), with Bewick’s Swans also favouring described seasonal variation in the swans use pondweeds (Potamogeton pectinatus and P. of farmland, with birds generally moving perfoliatus) along with hornworts Ceratophyllum from harvest waste (e.g. cereal stubbles, sp., watermilfoil Myriophyllum sp., stoneworts potatoes and sugar beet) to growing cereals Chara sp. and other emergent vegetation (e.g. winter wheat) and then to pasture as the (Rees 2006). However, wetland drainage winter progresses, which has been attributed and intensification of farming (including to a combination of food availability and increased use of fertiliser on grasslands and changes in dietary requirements (e.g. Laubek more extensive planting of arable crops) has 1995; Rees et al. 1997). Weaver (2013), resulted in a large-scale movement of swans studying habitat use by 63 satellite-tagged from wetland habitats to agricultural land. In Tundra Swans, found seasonal differences in Europe, Whooper Swans were recorded habitat selection. Tundra Swans selected feeding on cereals and potatoes as early as the open water over wetlands in autumn, but 19th century, but changes in agriculture saw agriculture was used substantially less during an increase in their use of arable habitats autumn migration (despite representing during the second half of the 20th century 45% and 80% of Tundra Swan habitats in (Kear 1963; Laubek et al. 1999). More the Great Lakes and Northern Prairies, recently, Tundra Swans were first observed in respectively, at this time) than in winter, when grain fields in the mid 1960s (Nagel 1965; swans selected agriculture lands, and wetlands Tate & Tate 1966; Munro 1981), and Bewick’s were used less than their availability. Weaver Swans have been utilising arable habitats (2013) concluded that if adequate aquatic since the early 1970s (review in Rees 2006). habitats were available, swans may not have Trumpeter Swans typically use freshwater made forays to agricultural fields, although marshes, ponds, lakes, rivers and brackish agricultural seeds provided alternative estuaries with abundant pondweed (Gale et al. foods of similar energy value (Kaminski et al. 1987; LaMontagne et al. 2003; Mitchell & 2003), and recommended that wetland Eichholz 2010), but also forage on arable conservationists interested in managing non- land in winter and early spring (Babineau breeding Tundra Swans should conserve and 2004; Mitchell & Eichholz 2010), where they restore wetlands within agricultural avoid soybean and prefer winter wheat and landscapes < 8 km of known roosts and aim corn (Varner 2008). In the mid-west U.S., to protect open water habitats, especially swans use reclaimed surface mine wetlands those containing SAVs. Detailed studies of close to agricultural fields, which rarely freeze Bewick’s Swan feeding ecology have also and are relatively undisturbed compared to illustrated the importance of aquatic habitats reservoirs (Varner 2008; Mitchell & Eichholz for swans arriving in autumn, with swans

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 46 Autumn–winter habitat use by waterfowl feeding on below-ground Fennel Pondweed provisions from humans (Birkhead & tubers at the Lauwersmeer, Netherlands, Perrins 1986; Sears 1989; Gayet et al. 2011). preferably in shallow waters and sandy They also use farmland, for instance moving sediments rather than in areas of deeper to agricultural fields and improved water, likely reflecting increased effort and grasslands during winter (Birkhead & Perrins energy costs (e.g. up-ending as opposed to 1986), but tend to be more widely dispersed head-dipping for food) required to feed in than the migratory species (Rees et al. 1997). deeper waters or where the tubers are in clay In parts of the United Kingdom, where (Nolet et al. 2001). Further analysis of the three swan species (Bewick’s, Whoopers and timing of the swans’ switch from feeding on Mutes) coincide in winter, segregation across pondweed tubers to feeding on sugar beet in habitats has been recorded, with Whooper fields around the Lauwersmeer found that and Mute Swans predominately using most swans switched habitats when the net permanent inland waters and improved energy gain from staying on tubers fell below pasture, whereas Bewick’s Swans were that from feeding on beet alone. However, mostly on arable land (Rees et al. 1997), the swans would attain a substantially indicating a range of habitats are important increased energy and total nutrient gain by for foraging by these swan species. feeding on both beet and tubers, and there On studying effects of patch size and was evidence from van Eerden (1997) isolation on Mute Swan habitat use in that mixed exploitation of tubers and France, Gayet et al. (2011) found that the beet does occur in the Lauwersmeer area. swans’ winter distribution and occurrence Overall, swans seemingly switch to the beet on fishponds was influenced by pond fields long after they would first benefit from structure more than surrounding landscape doing so due to energy gain alone (Nolet et al. and other features. Specifically, fishponds 2002). drained and cultivated for grain the previous year provided crop residues Mute Swans utilised by the swans the following winter. Mute Swan movements tend to be relatively Understanding habitat selection of Mute localised (< 50 km radius; Birkhead & Swans is important because they are Perrins 1986), although some long-distance perceived as having negative influences flights have been recorded (e.g. those at more on other waterfowl, through territorial northerly latitudes heading south in cold behaviour or intensive grazing on aquatic winters). They frequent a wide range of macrophytes, sometimes within their lowland wetland habitats throughout the European range but particularly where they year, including freshwater lakes, estuarine have been introduced to North America wetlands, commercial fishponds, sea lochs (Conover & Kania 1994; Petrie & Francis and shallow coastal waters, where they feed 2003; Gayet et al. 2011), with a Mute primarily on SAV, and are also commonly Swan control programme instigated in found on rivers and canals in urban areas Maryland in 2005 (Hindman & Tjaden where they rely on bread and other 2014; Hindman et al. 2014).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 47

Future challenges and needs breeding waterfowl are typically dynamic and unpredictable. Indeed, many migratory Planning and implementing conservation birds (e.g. Svalbard Barnacle Geese) strategies for waterfowl and their habitats are seemingly cannot assess local resource challenging because some species are conditions from afar and must “sample” declining or remain below long-term averages habitats upon settling in them, though (e.g. Scaup, Northern Pintail American others (e.g. Svalbard Pink-footed Geese Wigeon), whereas others have become Anser brachyrhyncus) appear to use conditions superabundant (e.g. Snow Goose) despite at one site as an indicator of conditions that some using similar resources (e.g. agricultural they might encounter at the next (Tombre et fields used by Northern Pintail and Snow al. 2008). Habitat and other environmental Geese) in autumn and during migration. dynamics may result in patchily distributed Wiens (1989) discussed habitat quality in food and other resources within and across terms of “fitness potential”, whereby habitat seasons, inter-annual site-specific changes in quality may be assessed through demographic, potential foraging areas (e.g. ploughed versus physiological and behavioural approaches. flooded field; 4th order selection), natural Nonetheless, Norris & Marra (2007) alluded inter-annual droughts or flooding, weather to the difficulty in understanding habitat that may dictate where birds winter selection in migrating species, particularly in and exploit resources, disturbance from identifying spatio-temporal connectivity of hunting and other human-related factors, individuals or populations among stages of physiological and behavioural dynamics and the annual cycle. Indeed, there is strong other scenarios (Fig. 1). During winter, some research and conservation interest in species like Northern Pintail, Mallard, teal determining the extent of migratory and diving ducks move inter-regionally, connectivity among birds occupying specific likely in search of suitable habitats (sensu wintering and breeding areas (Norris & Marra Fretwell 1972; Cox, Jr. & Afton 1996; 2007; Guillemain et al. 2014; Kaminski & Heitmeyer 2006; Caizergues et al. 2011; Elmberg 2014). Here we consider some Gourlay Larour et al. 2013). Interpreting challenges hindering understanding of habitat true migration from movements to and fro use and selection by waterfowl during the (i.e. foraging flights) can be challenging non-breeding season and suggest future needs (Dingle & Drake 2007) and documenting for research. We recognise there are other habitat selection across broad landscapes ecological, economic, bio-political and human in brief intervals may be even more dimensional considerations, but believe that equivocal. addressing the following five issues will Arguably, one of the greatest current advance science and stewardship of challenges waterfowl habitat researchers face waterfowl and their habitats in the Holarctic relative to identifying true selection involves and worldwide. an inability to determine true habitat and (1) Habitat and resource availability. resource availability at scales influencing Resources available for migrating and non- biological outcomes for the birds (Kaminski

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 48 Autumn–winter habitat use by waterfowl

Figure 1. A synthesis of primary and secondary factors that influence survival and potential fitness of Holarctic waterfowl.

& Elmberg 2014). For example, non- recent analysis of habitat use by mid- breeding waterfowl that exploit agricultural continent Mallard (Beatty et al. 2013) is environments (thousands of hectares of statistically robust, but may be ecologically agricultural land in one region alone) may tenuous because they could not estimate full suddenly move from dry to shallowly availability of agricultural lands possibly flooded fields during autumn– accessible by Mallard. Despite broad spatial winter (Reinecke et al. 1989). Mallard and temporal scaled information obtainable commonly feed in dry fields in southern from satellite-tracked birds (Krementz et Canada and the northern U.S. prairies, al. 2012; Beatty et al. 2013), sample sizes but not in the MAV where they utilise of marked birds are often small because puddled fields. This typical scenario is of funding limitations (Lindberg & further complicated during winters of below Walker 2007). This limitation constrains average temperatures; then, Mallard use dry determining selection of habitats, because a agricultural fields in winter as wetlands small cohort of individuals is assumed to freeze and foods become inaccessible. These represent the greater population. Moreover, and other scenarios create great resource when making inferences of resource variability across regions, temporal variability selection beyond one or two variables, within regions, and basically constrain sample sizes must be increased significantly researchers’ efforts to categorise and (Lindberg & Walker 2007). Given the estimate available resources. We concur that challenges in capturing environmental

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 49 variability across vast landscapes, we suggest Comparative studies of conspecifics across long-term studies (i.e. ≥ 5 years) should be geographic regions would be interesting invoked to reflect patterns of waterfowl and valuable; thus, studying non-breeding resource selection amid environmental resource use and in regions with the greatest stochasticity. abundance of individuals of a species is a Habitat conservation for non-breeding suggested approach. The genetic variability waterfowl is justified on the assumption that among individuals in these regions should certain important habitats and intrinsic food reveal patterns of resource exploitation resources are limited and thereby ramify important to subsequent breeding success. individual and population implications The greater challenge and future research (NAWMP 2012). However, to our knowledge endeavour is to discover if population (and as emphasised by Stafford et al. 2014), cohorts of a species that occupy ecologically true resource limitation has not been disparate landscapes during non-breeding demonstrated empirically by relating food or seasons contribute differently to population other resource abundance to biological recruitment for the species. Conversely, outcomes for waterfowl. Indeed, further analysis of bands recovered over a large understanding these scenarios is required geographical area have demonstrated that for assessing whether true resource some population boundaries in western limitation exists and is affecting individuals Europe were largely artificial (Guillemain et and populations (Neu et al. 1974; Johnson al. 2005), so that habitat selection studies 2007; Stafford et al. 2014). should be conducted at much greater (2) Populations important to study. geographic scales. Individuals of some species (e.g. Mallard), (3) Functional use of habitats. are widespread in North America during Understanding the range of benefits that autumn and winter (Bellrose 1980). Some birds derive from different habitats is also a Mallard winter along sandbars and adjacent critical need. Time-budget studies have agricultural lands along the Missouri River in been conducted at sites across the Holarctic North Dakota (Olsen & Cox, Jr. 2003), for decades, but new technology such while others predominately occupy the as unmanned aircraft (drones) or GPS southern U.S. (Nichols et al. 1983; Reinecke accelerometers would help to quantify the et al. 1989). We typically regard the former birds’ activities at local and micro-habitat region as “breeding grounds”, yet some scales, which in turn would improve our Mallard remain there during winter. knowledge of the functional values of Although some resources (e.g. agriculture) in habitats frequented by waterfowl. all these geographic regions get exploited by (4) Remoteness and difficulty in accessing Mallard, basing habitat selection on a habitats. Inhospitable conditions and cohort of a species in one region may not remoteness of habitats pose challenges to reflect important resource components studying birds such as sea ducks (Silverman elsewhere in the species’ range. Thus, what et al. 2013) and other arctic-nesting cohorts of birds should be studied? waterfowl. Establishing true habitat selection

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 50 Autumn–winter habitat use by waterfowl among sea ducks in remote environments, These and related metrics are useful for especially when trying to link movements or understanding cross-seasonal carry-over habitat use in relation to food, is particularly effects (Harrison et al. 2011; Sedinger & problematic. Researchers hypothesise that Alisauskas 2014), but difficulty lies in the serious challenges face wintering sea ducks, fact that autumn staging and migration including marine (boat) traffic, wind-power immediately follow the breeding season, development and aquaculture practices and are temporally furthest from the (Skov et al. 2011; Silverman et al. 2013). next breeding season. Hence, “back-dating” Despite inherent difficulties in investigating and identifying resources used by birds birds and habitats in marine environments, following their arrival on the breeding recent research has greatly advanced grounds, in relation to previous habitat use, understanding of non-breeding ecology of are paramount needs. For example, if body sea ducks, albeit continued efforts are condition of a cohort of Mallard in essential to sustain these birds (Faulkner Nebraska in late March was known and 2013; Silverman et al. 2013). these birds were subsequently sampled (5) Cumulative resource use. Lastly, there on the breeding grounds, linking March exists a lack of understanding of how condition and breeding success seems cumulative use of resources during the reasonable (i.e. Devries et al. 2008). non-breeding period may influence However, how should we consider body reproduction and recruitment (i.e. Heitmeyer condition in relation to future fitness & Fredrickson 1981; Kaminski & Gluesing prospects in a cohort of birds examined 1987). Indeed, body condition is an months earlier, during autumn–winter? important factor in waterfowl survival and No doubt, fitness is partly a result of fitness. For example, Devries et al. (2008) some cumulative use of resources during found that female Mallard which arrive in an animal’s annual cycle. The greatest better condition on breeding grounds in the uncertainty seems to be in understanding at Canadian prairie-parklands hatched eggs 15 what point in the non-breeding phase of the days earlier than those in relatively poor cycle a potential shortfall (or indeed windfall) condition. Guillemain et al. (2008) also of resources might influence future fitness observed more juveniles during autumn in prospects. There are likely bottlenecks or southern France when body condition of thresholds related to resource use during the females was greater at the end of the year which could impose disproportionate previous winter. Gunnarsson et al. (2005) impacts on subsequent fitness; these may used stable-carbon isotopes to demonstrate vary considerably between years and across that Black-tailed Godwits Limosa limosa species, and deserve further investigation. wintering in high quality sites in Europe were As an alternative to indexing body more likely to use higher-quality breeding condition or some other fitness metric, habitats and have greater reproductive perhaps coordinated inter-regional aerial success than birds using poorer-quality transect surveys of waterfowl during habitats (see also Norris & Marra 2007). autumn–spring migration could be

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 51 conducted (sensu Pearse et al. 2008) to Accurso, L.M. 1992. Distribution and abundance determine “hot spots” of waterfowl use, of wintering waterfowl on San Francisco Bay, thereby identifying and characterising 1988–1990. M.Sc. thesis, Humboldt State complexes of wetlands and uplands used by University, Arcata, California, USA. the majority of waterfowl (Pearse et al. 2012). Ackerman, J.T., Takekawa, J.Y., Orthmeyer, D.L., Fleskes, J.P., Yee, J.L. & Kruse, K.L. 2006. Aerial survey data could be incorporated with Spatial use by wintering greater white- GIS layers to illustrate habitat features and fronted geese relative to a decade of habitat describe high and low priority habitats for change in California’s Central Valley. Journal of North American waterfowl during winter and Wildlife Management 70: 965–976. migration (e.g. Pearse 2007), analogous to the Ackerman, J.T., Herzog, M.P., Yarris, G.S., “thunderstorm maps” used by waterfowl Casazza, M.L., Burns, E. & Eadie, J.M. breeding ground JV programmes (Loesch 2014. Chapter 5: Waterfowl ecology and et al. 2012). Clearly, we must be creative in management. In P.B. Moyle, A. Manfree & engaging diverse human expertise and reliable P.L. Fiedler (eds.), Suisun Marsh: Ecological technologies to understand the ecology of History and Possible Futures, pp. 103–132 and waterfowl throughout their annual cycle and maps 10 & 11. University of California Press, range, then use this knowledge to conserve Berkeley, California, USA. Adair, S.E., Moore, J.L. & Kiel, W.H., Jr. 1996. important habitats for birds across the Winter diving duck use of coastal ponds: an Holarctic region and worldwide. analyses of alternative hypotheses. Journal of Acknowledgements Wildlife Management 60: 83–93. Afton, A.D. & Anderson, M.G. 2001. Declining We are indebted to Joe Lancaster and Justyn scaup populations: a retrospective analysis of Foth, both Ph.D. students, and J. Czarnecki, long-term population and harvest survey data. Mississippi State University (MSU), for Journal of Wildlife Management 65: 781–796. dedicating days to help us complete review Alerstam, T. & Lindström, Å. 1990. Optimal bird and references for this manuscript. We migration: the relative importance of time, thank the Forest and Wildlife Research energy and safety. In E. Gwinner (ed.), Bird Center (FWRC) and Department of Migration: Physiology and Ecophysiology, pp. Wildlife, Fisheries and Aquaculture, MSU 331–351. Springer, Berlin, Germany. Alisauskas, R.T. 1998. Winter range expansion for supporting J.B. Davis and R.M. and relationships between landscape and Kaminski during writing of the manuscript. morphometrics of midcontinent Lesser This manuscript has been approved for Snow Geese. Auk 115: 851–862. publication by the FWRC as FWRC/WFA Alisauskas, R.T. 2002. Arctic climate, spring publication 395. nutrition, and recruitment in midcontinent Lesser Snow Geese. Journal of Wildlife References Management 66: 181–193. Abraham, K.F., Jefferies, R.L. & Alisauskas, R.T. Alisauskas, R.T. & Ankney, C.D. 1992. Spring 2005. The dynamics of landscape change and habitat use and diets of midcontinent adult snow geese in mid-continent North America. Lesser Snow Geese. Journal of Wildlife Global Change Biology 11: 841–855. Management 57: 43–54.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 52 Autumn–winter habitat use by waterfowl

Alisauskas, R.T. & Hobson, K.A. 1993. Baldassarre, G.A. 2014. Ducks, Geese, and Swans of Determination of Lesser Snow Goose diets North America. Johns Hopkins University and winter distribution using stable isotope Press, Baltimore, Maryland, USA. analysis. Journal of Wildlife Management 57: Baldassarre, G.A. & Bolen, E.G. 1984. Field- 49–54. feeding ecology of waterfowl wintering on Alisauskas, R.T., Ankney, C.D. & Klaas, E.E. 1988. the Southern High Plains of Texas. Journal of Winter diets and nutrition of midcontinental Wildlife Management 48: 63–71. lesser snow geese. Journal of Wildlife Management Baldassarre, G.A. & Bolen, E.G. 2006. Waterfowl 52: 403–414. Ecology and Management. Krieger Publishing Anderson, J.T. & Ballard, B.M. 2006. Survival and Company, Malabar, Florida, USA. Habitat Use of Female Northern Pintails Ballard, B.M., Thompson, J.E., Petrie, M.J., Wintering Along the Central Coast of Texas. Final Chekett, M. & Hewitt, D.G. 2004. Diet report to Ducks Unlimited, Inc., Ridgeland, and nutrition of northern pintails wintering Mississippi, USA. along the southern coast of Texas. Journal of Ankney, C.D. 1996. An embarrassment of riches: Wildlife Management 68: 371–382. too many geese. Journal of Wildlife Management Ballard, B.M., James, J.D., Bingham, R.L., Petrie, 60: 217–223. M.J. & Wilson, B.C. 2010. Coastal pond use Arner, D.H. & Hepp, G.R. 1989. Beaver Pond by redheads wintering in the Laguna Madre, Wetlands: a Southern Perspective. Texas Tech Texas. Wetlands 30: 669–674. University Press, Lubbock, Texas, USA. Barras, S.C., Kaminski, R.M. & Brennan, L.A. Arzel, C., Elmberg, J. & Guillemain, M. 2006. 1996. Acorn selection by female Wood Ducks. Ecology of spring-migrating Anatidae: a Journal of Wildlife Management 60: 592–602. review. Journal of Ornithology 147: 167–184. Bateman, H.A., Joanen, T. & Stutzenbaker, C.D. Atlantic Coast Joint Venture (ACJV). 2005. 1988. History and status of midcontinental North American Waterfowl Management snow geese on their Gulf Coast winter range. Plan. Atlantic Coast Joint Venture Waterfowl In M.W. Weller (ed.), Waterfowl in Winter, pp. Implementation Plan Revision: June 2005. 495–515. University of Minnesota Press, Hadley, Massachusetts, USA. Minneapolis, Minnesota, USA. Atlantic Coast Joint Venture (ACJV). 2009. Batt, B.D. 1997. Arctic Ecosystems in Peril: Report of Atlantic Coast Joint Venture Strategic Plan: the Arctic Goose Habitat Working Group. Arctic Updated July 2009. Hadley, Massachusetts, Goose Joint Venture Special Publication. U.S. USA. Fish & Wildlife Service, Washington D.C., Audubon, J.J. 1840. The Birds of America. J.J USA and Canadian Wildlife Service, Ottawa, Audubon & J.B. Chevalier, New York, Ontario, Canada. USA. Batzer, D.P. & Baldwin, A.H. 2012. Wetland Baar, L., Matlack, R.S., Johnson, W.P. & Habitats of North America: Ecology and Barron, R.B. 2008. Migration chronology of Conservation Concerns. University of California waterfowl in the Southern High Plains of Press, Berkley, California, USA. Texas. Waterbirds 31: 394–401. Beatty, W.S., Kesler, D.C., Webb, E.B., Raedeke, Babineau, F.M. 2004. Winter ecology of A.H., Naylor, L.W. & Humburg, D.D. 2013. Trumpeter Swans in southern Illinois. Quantitative and qualitative approaches to M.Sc. thesis, Southern Illinois University identifying migration chronology in a Carbondale, Carbondale, Illinois, USA. continental migrant. PloS ONE 8: e75673.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 53

Beekman, J.H., Nolet, B.A. & Klaassen, M. 2002. in North America, pp. 341–365. Texas Tech Skipping swans: differential use of migratory University Press, Lubbock, Texas, USA. stop-over sites in spring and autumn in Bourne, N. 1984. Clam Predation by Scoter Ducks in relation to fuelling rates. Ardea 90: 437– the Strait of Georgia, British Columbia, Canada. 460. Canadian Technical Report of Fisheries and Bellebaum, J., Larsson, K. & Kube, J. 2012. Aquatic Science No. 1331. Department of Research on sea ducks in the Baltic Sea. Fisheries & Oceans Fisheries (Research Uppsala University – Campus Gotland, Visby, Branch), Pacific Biological Station, Nanaimo, Sweden. http://seaducks.hgo.se/?q=system/ British Columbia, Canada. files/dokument/Reserach%20on%20Sea%20 Brand, L.A., Takekawa, J.Y., Shinn, J., Graham, T., Ducks.pdf. Buffington, K., Gustafson, K.B., Smith, Bellrose, F. 1980. Ducks, Geese and Swans of L.M., Spring, S.E. & Miles, A.K. 2014. North America. Stackpole Books, Harrisburg, Effects of wetland management on carrying Pennsylvania, USA. capacity of diving ducks and shorebirds in a Bellrose, F.C. & Crompton, R.D. 1970. coastal estuary. Waterbirds 37: 52–67. Migrational behavior of mallards and black Brochet, A.L., Mouronval, J.B., Aubry, P., ducks as determined from banding. Illinois Gauthier-Clerc, M., Green, A.J., Fritz, H. & Natural History Survey Bulletin 30: 167–234. Guillemain, M. 2012. Diet and feeding Bellrose, F.C. & Holm, D.J. 1994. Ecology and habitats of Camargue dabbling ducks: What Management of the Wood Duck. Stackpole has changed since the 1960s? Waterbirds 35: Books, Mechanicsburg, Pennsylvania, USA. 555–576. Bergmann, C. 1847. Über die verlhältnisse der Brusendorff, A.C., Korpinen, S., Meski, L. & wärmeökonomie der thiere zu ihrer grösse. Stankiewicz, M. 2013. HELCOM actions to Göttinger Studien 3: 595–708. eliminate illegal and accidental oil pollution Birkhead, M.E. & Perrins, C.M. 1986. The Mute from ships in the Baltic Sea. In A.G. Swan. Croom Helm, London, UK. Kostianoy & O.Y. Lavrova (eds.), Oil Pollution Black, J.M. & Rees, E.C. 1984. The structure and in the Baltic Sea, pp. 15–40. Springer, behaviour of the Whooper Swan population Heidelberg, Germany. wintering at Caerlaverock, Dumfries and Caizergues, A., Guillemain, M., Arzel, C., Galloway, Scotland: an introductory study. Devineau, O., Leray, G., Pilvin, D., Lepley, Wildfowl 35: 21–36. M., Massez, G. & Schricke, V. 2011. Block, W.M. & Brennan, L.A. 1993. The habitat Emigration rates and population turnover of concept in ornithology. Current Ornithology teal Anas crecca in two major wetlands of 11: 35–91. western Europe. Wildlife Biology 17: 373–382. Bolduc, F. & Afton, A.D. 2004. Relationships Calenge, C., Guillemain, M., Gauthier-Clerc, M. between wintering waterbirds and & Simon, G. 2010. A new exploratory invertebrates, sediments and hydrology of approach to the study of the spatio-temporal coastal marsh ponds. Waterbirds 27: 333– distribution of ring recoveries: the example 341. of Teal (Anas crecca) ringed in Camargue, Bolen, E.G., Baldassarre, G.A. & Guthery, Southern France. Journal of Ornithology 151: F.S. 1989. Playa lakes. In L.M. Smith, R.L. 945–950. Pederson & R.M. Kaminski (eds.), Habitat Central Valley Joint Venture. 2006. Central Valley Management for Migrating and Wintering Waterfowl Joint Venture Implementation Plan-Conserving Bird

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 54 Autumn–winter habitat use by waterfowl

Habitat. U.S. Fish & Wildlife Service, Cowardin, L.M., Carter, V., Golet, F.C. & LaRoe, Sacramento, California, USA. E.T. 1979. Classification of Wetlands and Chabreck, R.H., Joanen, T. & Paulus, S.L. 1989. Deepwater Habitats of the United States. U.S. Southern coastal marshes and lakes. In L.M. Department of the Interior, Fish & Wildlife Smith, R.L. Pederson & R.M. Kaminski (ed.), Service, Washington D.C., USA. Habitat Management for Migrating and Wintering Cox, R.R., Jr. & Afton, A.D. 1996. Evening Waterfowl in North America, pp. 249–277. flights of female Northern Pintails from a Texas Tech University Press, Lubbock, major roost site. The Condor 98: 810–819. Texas, USA. Cox, R.R., Jr. & Afton, A.D. 1997. Use of Chesapeake Bay Program. 2007. Bay trends and habitats by female Northern Pintails wintering indicators. http://www.chesapeakebay.net/ in southwest Louisiana. Journal of Wildlife indicators.htm. Management 61: 435–443. Christopher, M.W., Hill, E.P. & Steffen, D.E. Dalby, L. 2013. Waterfowl duck distributions and 1988. Use of catfish ponds by waterfowl a changing climate. Ph.D. thesis, Aarhus wintering in Mississippi. In M.W. Weller (ed.), University, Aarhus, Denmark. Waterfowl in Winter, pp. 413–418. Minnesota Dandelot, S., Verlaque, R., Dutartre, A. & Press, Minneapolis, Minnesota, USA. Cazaubon, A. 2005. Ecological, dynamic Clausen, K.K., Clausen, P., Fælled, C.C. and taxonomic problems due to Ludwigia & Mouritsen, K.N. 2012. Energetic (Onagraceae) in France. Hydrobiologia 551: consequences of a major change in habitat 131–136. use: endangered Brent Geese Branta bernicla Dänhardt, J. & Lindström, Å. 2001. Optimal hrota losing their main food resource. Ibis departure decisions of songbirds from 154: 803–814. an experimental stopover site and the Clausen, K.K. & Clausen, P. 2014. Forecasting significance of weather. Animal Behaviour 62: future drowning of coastal waterbird habitats 235–243. reveals a major conservation concern. Darwin, C. 1859. The Origin of Species. Penguin Biological Conservation 171: 177–185. Books, New York, USA. Coates, P.S., Casazza, M.L., Halstead, B.J. & Davis, B.E. & Afton, A.D. 2010. Movement Fleskes, J.P. 2012. Relative value of managed distances and habitat switching by female wetlands and tidal marshlands for wintering Mallards wintering in the Lower Mississippi Northern Pintails. Journal of Fish and Wildlife Alluvial Valley. Waterbirds 33: 349–356. Management 3: 98–109. Delnicki, D. & Reinecke, K.J. 1986. Mid-winter Conomos, T., Smith, R. & Gartner, J. 1985. food use and body weights of Mallards and Environmental setting of San Francisco Bay. Wood Ducks in Mississippi. Journal of Wildlife Hydrobiologia 129: 1–12. Management 50: 43–51. Conover, M.R. & Kania, G.S. 1994. Impact of Devries, J.H., Brook, R.W., Howerter, D.W. & interspecific aggression and herbivory by Anderson, M.G. 2008. Effects of spring Mute Swans on native waterfowl and aquatic body condition and age on reproduction in vegetation in New England. Auk 111: 744–748. Mallards (Anas platyrhynchos). Auk 125: Cornelius, S.E. 1977. Food and resource utilization 618–628. by wintering redheads on lower Laguna Dias, M.P. 2009. Use of salt ponds by wintering Madre. Journal of Wildlife Management 41: 374– shorebirds throughout the tidal cycle. 385. Waterbirds 32: 531–537.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 55

Dickson, D.L. 2012. Movement of King Eiders from diet of Greater White fronted Geese. Journal breeding grounds on Banks Island, Northwest of Wildlife Management 75: 78–91. Territories, to moulting and wintering areas. Ely, C.R., Sladen, W.J.L., Wilson, H.M., Savage, Technical Report Series No. 516, Canadian S.E., Sowl, K.M., Henry, B., Schwitters, M. & Wildlife Service, Edmonton, Alberta, Canada. Snowdon, J. 2014. Delineation of Tundra Dingle, H. 1996. Migration: The Biology of Life on the Swan Cygnus c. columbianus populations in Move. Oxford University Press, New York, North America: geographic boundaries and USA. interchange. Wildfowl 64: 132–147. Dingle, H. & Drake, V.A. 2007. What is Engelhardt, K.A., Kadlec, J.A., Roy, V.L. & Powell, migration? Bioscience 57: 113–121. J.A. 2000. Evaluation of translocation criteria: Dirksen, S. & Beekman, J.H. 1991. Population case study with Trumpeter Swans Cygnus size, breeding success and distribution of buccinator. Biological Conservation 94: 173–181. Bewick’s Swans, Cygnus columbianus bewickii, Faulkner, H. 2013. Influence of aquaculture wintering in Europe 1986–87. Wildfowl on winter sea duck distribution and (Supplement No. 1): 120–124.. abundance in south Puget Sound. M.Sc. Dubovsky, J. & Kaminski, R.M. 1992. Waterfowl thesis, The Evergreen State College, and American coot habitat associations with Olympia, Washington, USA. Mississippi catfish ponds. Proceedings of the Feaga, J.S. 2013. Winter waterbird use and food Annual Conference of Southeastern Association of resources of aquaculture lands in Mississippi. Fish and Wildlife Agencies 46: 10–17. M.Sc. thesis, Mississippi State University, Eadie, J.M., Elphick, C.S., Reinecke, K.J. & Miller, Mississippi State, Mississippi, USA. M.R. 2008. Wildlife values of North Fleskes, J. 2012. Climate change impacts on ecology American ricelands. In S. Manley (ed.), and habitats of Central Valley waterfowl and Conservation of Ricelands of North America, pp. other waterbirds. http://www.werc.usgs.gov/ 8–90. Ducks Unlimited, Inc., Memphis, Project.aspx?ProjectID=204. Tennessee, USA. Fleskes, J.P., Yee, J.L., Casazza, M.L., Miller, M.R., Elmberg, J., Hessel, R., Fox, A.D. & Dalby, L. Takekawa, J.Y. & Orthmeyer, D.L. 2005. 2014. Interpreting seasonal range shifts in Waterfowl Distribution, Movements and Habitat migratory birds: a critical assessment of Use Relative to Recent Habitat Changes in the ‘short-stopping’ and a suggested terminology. Central Valley of California: A Cooperative Project Journal of Ornithology 155: 1–9. to Investigate Impacts of the Central Valley Joint Elphick, C.S. 2000. Functional equivalency Venture and Changing Agricultural Practices on the between rice fields and seminatural wetland Ecology of Wintering Waterfowl. U.S. Geological habitats. Conservation Biology 14: 181–191. Survey, Western Ecological Research Center, Elphick, C.S., Parsons, K.C., Fasola, M. & Dixon, California, USA. Mugica, L. 2010. Ecology and conservation Fleskes, J.P. & Yee, J.L. 2007. Waterfowl distribution of birds in rice fields: a global review. and abundance during spring migration in Waterbirds 33 (Special Publication 1): 246 pp. southern Oregon and northeastern California. Ely, C.R. 1992. Time allocation by Greater White- Western North American Naturalist 67: 409– fronted Geese: influence of diet, energy 428. reserves and predation. The Condor 94: 857–870. Fleskes, J.P., Yee, J.L., Yarris, G.S., Miller, M.R. & Ely, C.R. & Raveling, D.G. 2011. Seasonal Casazza, M.L. 2007. Pintail and Mallard variation in nutritional characteristics of the survival in California relative to habitat,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 56 Autumn–winter habitat use by waterfowl

abundance, and hunting. Journal of Wildlife Gale, R.S., Garton, E.O. & Ball, I.J. 1987. The Management 71: 2238–2248. History, Ecology and Management of the Rocky Fleskes, J.P., Skalos, D.A. & Farinha, M.A. 2013. Mountain Population of Trumpeter Swans. U.S. Fish Changes in types and area of postharvest & Wildlife Service, Montana Cooperative flooded fields available to waterbirds in Wildlife Research Unit, Missoula, Montana, Tulare Basin, California. Journal of Fish and USA. Wildlife Management 4: 351–361. Ganter, B., Prokosch, P. & Ebbinge, B.S. 1997. Foth, J. R., Straub, J., Kaminski, R.M., Davis, J.B. Effect of saltmarsh loss on the dispersal & Leininger, T. In press. Aquatic invertebrate and fitness parameters of Dark bellied abundance and biomass in Arkansas, Brent Geese. Aquatic Conservation: Marine and Mississippi and Missouri bottomland Freshwater Ecosystems 7: 141–151. hardwood forests during winter. Journal of Ganter, B., Larsson, K., Syroechkovsky, E.V., Fish and Wildlife Management 00: 000–000. Litvin, K.E., Leito, A. & Madsen, J. 1999. Foster, M.A., Gray, M.J. & Kaminski, R.M. 2010. Barnacle Goose Branta leucopsis: Russia/Baltic. Agricultural seed biomass for migrating and In J. Madsen, G. Cracknell & A.D. Fox (eds.), wintering waterfowl in the southeastern Goose Populations of the Western Palearctic – A United States. Journal of Wildlife Management Review of Status and Distribution, pp. 270–283. 74: 489–495. Wetlands International Publication No. 48, Fox, A.D., Ebbinge, B.S., Mitchell, C., Heinicke, T., National Environmental Research Institute, Aarvak, T., Colhoun, K., Clausen, P., Dereliev, Ronde, Denmark. S., Faragó, S. & Koffijberg, K. 2010. Current Garðarsson, A. 1991. Movements of Whooper estimates of goose population sizes in western Swans Cygnus cygnus neckbanded in Iceland. Europe, a gap analysis and an assessment of Wildfowl (Special Supplement No. 1): 189– trends. Ornis Svecica 20: 115–127. 194. Frederick, R.B., Clark, W.R. & Takekawa, J.Y. Gayet, G., Guillemain, M., Fritz, H., Mesleard, F., 1992. Application of a computer simulation Begnis, C., Costiou, A., Body, G., Curtet, L. & model to migrating White-fronted Geese in Broyer, J. 2011. Do Mute Swan (Cygnus olor) the Klamath Basin. In D.R. McCullough & grazing, swan residence and fishpond nutrient R.H. Barrett (eds.), Wildlife 2001: Populations, availability interactively control macrophyte pp. 696–706. Elsevier Applied Science, New communities? Aquatic Botany 95: 110–116. York, USA. Gilmer, D.S., Miller, M.R., Bauer, R.D. & Fredrickson, L.H. 2005. Contemporary LeDonne, J.R. 1982. California’s Central bottomland hardwood systems: structure, Valley wintering waterfowl: concerns and function and hydrologic condition resulting challenges. Transactions of the North American from two centuries of anthropogenic Wildlife and Natural Resources Conference 47: activities. In L.M. Smith, R.L. Pederson & 441–452. R.M. Kaminski (eds.), Ecology and Management Gordon, D.H., Gray, B.T., Perry, R.D., Prevost, of Bottomland Hardwoods Systems: The State of M.B., Strange, T.H. & Williams, R.K. 1989. our Understanding, pp. 19–35. University of South Atlantic coastal wetlands. In L.M. Missouri-Columbia, Puxico, Missouri, USA. Smith, R.L. Pederson & R.M. Kaminski (eds.), Fretwell, S.D. 1972. Populations in a Seasonal Habitat Management for Migrating and Wintering Environment. Princeton University Press, Waterfowl in North America, pp. 57–92. Texas Princeton, New Jersey, USA. Tech University Press, Lubbock, Texas, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 57

Gordon, D.H., Gray, B.T. & Kaminski, R.M. Guillemain, M., Fuster, J., Lepley, M., Mouronval, 1998. Dabbling duck-habitat associations J.B. & Massez, G. 2009. Winter site fidelity is during winter in coastal South Carolina. higher than expected for Eurasian Teal Anas Journal of Wildlife Management 62: 569–580. crecca in the Camargue, France. Bird Study 56: Goudie, R., Brault, S., Conant, B., Kondratyev, 272–275. A., Petersen, M. & Vermeer, K. 1994. The Guillemain, M., Devineau, O., Brochet, A.-L., status of sea ducks in the North Pacific rim: Fuster, J., Fritz, H., Green, A.J. & Gauthier- toward their conservation and management. Clerc, M. 2010a. What is the spatial unit for a Transactions of the North American Wildlife and wintering Teal Anas crecca? Weekly day roost Natural Resources Conference 59: 27–49. fidelity inferred from nasal saddles in the Gourlay-Larour, M.-L., Pradel, R., Guillemain, M., Camargue, southern France. Wildlife Biology Santin-Janin, H., L’hostis, M. & Caizergues, 16: 215–220. A. 2013. Individual turnover in common Guillemain, M., Elmberg, J., Gauthier-Clerc, M., pochards wintering in western France. Journal Massez, G., Hearn, R., Champagnon, J. & of Wildlife Management 77: 477–485. Simon, G. 2010b. Wintering French Mallard Green, A.W. & Krementz, D.G. 2008. Mallard and Teal are heavier and in better body harvest distributions in the Mississippi and condition than 30 years ago: effects of a Central Flyways. Journal of Wildlife Management changing environment? Ambio 39: 170–180. 72: 1328–1334. Guillemain, M., Van Wilgenburg, S.L., Legagneux, Greer, D.M., Dugger, B.D., Reinecke, K.J. & P. & Hobson, K.A. 2014. Assessing Petrie, M.J. 2009. Depletion of rice as food geographic origins of Teal (Anas crecca) of waterfowl wintering in the Mississippi through stable-hydrogen (δ2H) isotope Alluvial Valley. Journal of Wildlife Management analyses of feathers and ring-recoveries. 73: 1125–1133. Journal of Ornithology 155: 165–172. Guglielmo, C.G., O’Hara, P.D. & Williams, Gunnarsson, T.G., Gill, J.A., Newton, J., Potts, T.D. 2002. Extrinsic and intrinsic sources of P.M. & Sutherland, W.J. 2005. Seasonal variation in plasma lipid metabolites of free- matching of habitat quality and fitness in a living Western Sandpipers (Calidris mauri). migratory bird. Proceedings of the Royal Society of Auk 119: 437–445. London Series B 272: 2319–2323. Guillemain, M., Fritz, H. & Duncan, P. 2002. The Hagy, H.M. & Kaminski, R.M. 2012. Winter importance of protected areas as nocturnal waterbird and food dynamics in autumn- feeding grounds for dabbling ducks wintering managed moist-soil wetlands of the in western France. Biological Conservation 103: Mississippi Alluvial Valley. Wildlife Society 183–198. Bulletin 36: 512–523. Guillemain, M., Sadoul, N. & Simon, G. Hagy, H.M., Straub, J.N., Schummer, M.L. & 2005. European flyway permeability and Kaminski, R.M. 2014. Annual variation in abmigration in Teal Anas crecca, an analysis food densities and factors affecting wetland based on ringing recoveries. Ibis 147: 688–696. use by waterfowl in the Mississippi Alluvial Guillemain, M., Elmberg, J., Arzel, C., Johnson, Valley. Wildfowl (Special Issue 4): 436–450. A. & Simon, G. 2008. The income–capital Hake, M., Månsson, J. & Wiberg, A. 2010. A breeding dichotomy revisited: late winter working model for preventing crop damage body condition is related to breeding success caused by increasing goose populations in in an income breeder. Ibis 150: 172–176. Sweden. Ornis Svecica 20: 225–233.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 58 Autumn–winter habitat use by waterfowl

Haramis, G., Nichols, J., Pollock, K. & Hines, J. hardwoods influence Mallard recruitment? 1986. The relationship between body mass Transactions of the North American Wildlife and and survival of wintering Canvasbacks. Auk Natural Resources Conferences 46: 44–57. 103: 506–514. Heitmeyer, M.E., Connelly, D.P. & Pederson, R.L. Haramis, G.M. 1991a. Canvasback Aythya valisineria. 1989. The Central, Imperial and Coachella In S. Funderburk, S. Jordan, J. Mihursky & D. Valleys of California. In L.M. Smith, R.L. Riley (eds.), Habitat Requirements for Chesapeake Pederson & R.M. Kaminski (eds.), Habitat Bay Living Resources, Revised 2nd Edition, pp. Management for Migrating and Wintering Waterfowl 17.1–17.10. Chesapeake Research Consortium, in North America, pp. 475–505. Texas Tech Solomons, Maryland, USA. University Press, Lubbock, Texas, USA. Haramis, G.M. 1991b. Redhead Aythya americana. Heitmeyer, M.E., Cooper, R.J., Dickson, J.G. & In S. Funderburk, S. Jordan, J. Mihursky Leopold, B.D. 2005. Ecological relationships and D. Riley (eds.), Habitat Requirements for of warm-blooded vertebrates in bottomland Chesapeake Bay Living Resources, 2nd ed. revised, hardwoof ecosystems. In L.H. Fredrickson, pp. 18.1–18.10. Chesapeake Research S.L. King & R.M. Kaminski (eds.), Ecology and Consortium, Solomons, Maryland, USA. Management of Bottomland Hardwood Systems: the Harmon, B.G. 1962. Mollusk as food of lesser State of Our Understanding, pp. 281–307. scaup along the Louisiana coast. Transactions University of Missouri-Columbia, Gaylord North American Wildlife and Natural Resource Memorial Laboratory Special Publication Conference 27: 132–138. No. 10, Puxico, Missouri, USA. Harrison, X.A., Blount, J.D., Inger, R., Norris, Heitmeyer, M.E. 2006. The importance of winter D.R. & Bearhop, S. 2011. Carry-over effects floods to mallards in the Mississippi Alluvial as drivers of fitness differences in animals. Valley. Journal of Wildlife Management 70: Journal of Animal Ecology 80: 4–18. 101–110. Haukos, D.A. & Smith, L.M. 2003. Past and Hénaux, V., Powell, L.A., Vrtiska, M.P. & future impacts of wetland regulations on Hobson, K.A. 2012. Establishing winter playa ecology in the Southern Great Plains. origins of migrating Lesser Snow Geese Wetlands 23: 577–589. using stable isotopes. Avian Conservation and Haukos, D.A. 2004. Analyses of selected mid-winter Ecology 7: 5. waterfowl survey data (1955–2004). U.S. Fish & Hindman, J.L. & Stotts, V.D. 1989. Chesapeake Wildlife Service, Albuquerque, New Mexico, Bay and North Carolina Sounds. In L.M. USA. Smith, R.L. Pederson & R.M. Kaminski Hearn, R. & Skov, H. 2011. A brief overview (eds.), Habitat Management for Migrating and of the status of European seaducks and Wintering Waterfowl in North America, pp. actions required for their conservation. 27–55. Texas Tech University Press, www.wwt.org.uk/blog/wp-content/uploads/ Lubbock, Texas, USA. 2011/12/Seaducks-briefing-note-Dec-11.pdf. Hindman, L.J. & Tjaden, R.L. 2014. Awareness Heinz, G.H., Hoffman, D.J. & Gold, L.G. 1989. and opinions of Maryland citizens toward Impaired reproduction of mallards fed an Chesapeake Bay Mute Swans Cygnus olor and organic form of selenium. Journal of Wildlife management alternatives. Wildfowl 64: Management 53: 418–428. 167–185. Heitmeyer, M. & Fredrickson, L.H. 1981. Do Hindman, L.J., Harvey, W.F & Conley, L.E. 2014. wetland conditions in the Mississippi Delta Spraying corn oil on Mute Swan Cygnus olor

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 59

eggs to prevent hatching. Wildfowl 64: 186– Johnson, D.H. 1980. The comparison of usage 196. and availability measurements for evaluating Hobaugh, W.C., Stutzenbaker, C.D. & Flickinger, resource preference. Ecology 61: 65–71. E.L. 1989. The rice prairies. In L.M. Smith, Johnson, F.A. & Montalbano, F. 1984. Selection R.L. Pederson & R.M. Kaminski (eds.), of plant communities by wintering waterfowl Habitat Management for Migrating and Wintering on Lake Okeechobee, Florida. Journal of Waterfowl in North America, pp. 367–383. Wildlife Management 48: 174–178. Texas Tech University Press, Lubbock, Johnson, F.A. & Montalbano, F. 1989. Southern Texas, USA. reservoirs and lakes. In L.M. Smith, R.L. Hohman, W.L. & Rave, D.P. 1990. Diurnal time- Pederson & R.M. Kaminski (eds.), Habitat activity budgets of wintering Canvasbacks in Management for Migrating and Wintering Louisiana. Wilson Bulletin 102: 645–654. Waterfowl in North America, pp. 93–116. Texas Hohman, W.L., Woolington, D.W. & Devries, J.H. Tech University Press, Lubbock, Texas, USA. 1990. Food habits of wintering Canvasbacks Johnson, M.D. 2007. Measuring habitat quality: a in Louisiana. Canadian Journal of Zoology 68: review. The Condor 109: 489–504. 2605–2609. Jorde, D.G., Krapu, G.L., Crawford, R.D. & Hay, Hothem, R.L., Lonzarich, D.G., Takewaka, J.E. M.A. 1984. Effects of weather on habitat & Ohlendorf, H.M. 1998. Contaminants selection and behavior of Mallards wintering in wintering Canvasbacks and Scaups from in Nebraska. The Condor 86: 258–265. San Francisco Bay, California. Environmental Kadlec, J.A. & Smith, L.M. 1989. The great basin Monitoring and Assessment 50: 67–84. marshes. In L.M. Smith, R.L. Pederson & Huner, J.V., Jeske, C.W. & Norling, W. R.M. Kaminski (eds.), Habitat Management for 2002. Managing agricultural wetlands for Migrating and Wintering Waterfowl in North waterbirds in the coastal regions of Louisiana, America, pp. 451–474. Texas Tech University USA. Waterbirds (Special Publication 2): Press, Lubbock, Texas, USA. 66–78. Kaminski, R.M. & Davis, J.B. 2014. Evaluation of Hurrell, J.W., Kushnir, Y., Ottersen, G. & the Migratory Bird Habitat Initiative: Report of Visbeck, M. 2003. An Overview of the findings. Forest and Wildlife Research Center, North Atlantic Oscillation. In J.W. Hurrell, Research Bulletin WF391, Mississippi State Y. Kushnir, G. Ottersen & M. Visbeck University, Mississippi, USA. (eds.), The North Atlantic Oscillation: Climatic Kaminski, R.M. & Elmberg, J. 2014. An significance and nvironmental Impact, pp. 1–35. introduction to habitat use and selection by American Geophysical Union, Washington waterfowl in the northern hemisphere. D.C., USA. Wildfowl (Special Issue 4): 9–16. Jefferies, R., Rockwell, R. & Abraham, K. 2004. Kaminski, R.M. & Gluesing, E.A. 1987. Density- The embarrassment of riches: agricultural and habitat-related recruitment in mallards. food subsidies, high goose numbers and Journal of Wildlife Management 51: 141–148. loss of Arctic wetlands a continuing saga. Kaminski, R.M. & Weller, M.W. 1992. Breeding Environmental Reviews 11: 193–232. habitats of nearctic waterfowl. In B.D.J. Batt, Jensen, R.A., Wisz, M.S. & Madsen, J. 2008. A.D. Afton, M.G. Anderson, C.D. Ankney, Prioritizing refuge sites for migratory geese D.H. Johnson, J.A. Kadlec & G.L. Krapu to alleviate conflicts with agriculture. (eds.), Ecology and Management of Breeding Biological Conservation 141: 1806–1818. Waterfowl pp. 568–589. University of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 60 Autumn–winter habitat use by waterfowl

Minnesota Press, Minneapolis, Minnesota, mallards as determined by satellite telemetry. USA. Journal of Fish and Wildlife Management 3: Kaminski, R.M., Alexander, R.W. & Leopold, 238–251. B.D. 1993. Wood duck and mallard winter Kross, J.P., Kaminski, R.M., Reinecke, K.J. & microhabitats in Mississippi hardwood Pearse, A.P. 2008. Conserving waste rice for bottomlands. Journal of Wildlife Management wintering waterfowl in the Mississippi 57: 562–570. Alluvial Valley. Journal of Wildlife Management Kaminski, R.M., Davis, J.B., Essig, H.W., 72: 1383–1387. Gerard, P.D. & Reinecke, K.J. 2003. True Lack, D. 1944. The problem of partial migration. metabolizable energy for wood ducks from British Birds 37: 122–130. acorns compared to other waterfowl foods. Lack, D. 1966. Population Studies of Birds. Journal of Wildlife Management 67: 542–550. Clarendon Press, Oxford, UK. Kear, J. 1963. The history of potato-eating by LaMontagne, J.M., Jackson, L.J. & Barclay, R.M. wildfowl in Britain. Wildfowl Trust Annual 2003. Compensatory growth responses of Report 14: 54–65. Potamogeton pectinatus to foraging by migrating Kinney, S.D. 2004. Estimating the Population of Trumpeter Swans in spring stop over areas. Greater and Lesser Scaup during Winter in Aquatic Botany 76: 235–244. off-Shore Louisiana. M.Sc. thesis, Louisiana Lancaster, J.D. 2013. Survival, habitat use and State University, Baton Rouge, USA. spatiotemporal use of wildlife management Kirk, M.K., Esler, D. & Boyd, W.S. 2007. areas by female mallards in Mississippi’s Foraging effort of Surf Scoters (Melanitta alluvial valley. M.Sc. thesis, Mississippi State perspicillata) wintering in a spatially and University, Mississippi State, Mississippi, temporally variable prey landscape. Canadian USA. Journal of Zoology 85: 1207–1215. Larsson, K. & Tydén, L. 2005. Effects of oil Klimas, C., Murray, E., Foti, T., Pagan, spills on wintering Long-tailed Ducks J., Williamson, M. & Langston, H. 2009. Clangula hyemalis at Hoburgs bank in central An ecosystem restoration model for the Baltic Sea between 1996/97 and 2003/04. Mississippi Alluvial Valley based on Ornis Svecica 15: 161–171. geomorphology, soils and hydrology. Laubek, B. 1995. Habitat use by Whooper Swans Wetlands 29: 430–450. Cygnus cygnus and Bewick’s Swans Cygnus Korschgen, C.E. 1989. Riverine and deepwater columbianus bewickii wintering in Denmark: habitats for diving ducks. In L.M. Smith, R.L. increasing agricultural conflicts. Wildfowl, 46: Pederson & R.M. Kaminski (eds.), Habitat 8–15. Management for Migrating and Wintering Laubek, B., Nilsson, L., Wieloch, M., Koffijberg, Waterfowl in North America, pp. 157–180. K., Sudfelt, C. & Follestad, A. 1999. Texas Tech University Press, Lubbock, Distribution, numbers and habitat choice of Texas, USA. the NW European Whooper Swan Cygnus Krapu, G.L., Facey, D.E., Fritzell, E.K. & cygnus population: results of an international Johnson, D.H. 1984. Habitat use by migrant census in January 1995. Vogelwelt 120: 141– Sandhill Cranes in Nebraska. Journal of 154. Wildlife Management 48: 407–417. Legagneux, P., Inchausti, P., Bourguemestre, F., Krementz, D.G., Asante, K. & Naylor, L.W. 2012. Latraube, F. & Bretagnolle, V. 2009. Effect of Autumn migration of Mississippi Flyway predation risk, body size, and habitat

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 61

characteristics on emigration decisions in Lovvorn, J.R. & Baldwin, J.R. 1996. Intertidal Mallards. Behavioral Ecology 20: 186–194. and farmland habitats of ducks in the Puget Lehikoinen, A., Jaatinen, K., Vähätalo, A.V., Sound region: a landscape perspective. Clausen, P., Crowe, O., Deceuninck, B., Biological Conservation 77: 97–114. Hearn, R., Holt, C.A., Hornman, M. & Lovvorn, J.R., De La Cruz, S.E., Takekawa, Keller, V. 2013. Rapid climate driven shifts in J.Y., Shaskey, L.E. & Richman, S.E. 2013. wintering distributions of three common Niche overlap, threshold food densities and waterbird species. Global Change Biology 19: limits to prey depletion for a diving duck 2071–2081. assemblage in an estuarine bay. Marine Ecology Lehnen, S.E. & Krementz, D.G. 2013. Use of Progress Series 476: 251–268. aquaculture ponds and other habitats by Luigujõe, L., Kuresoo, A., Keskpaik, J., Ader, A. autumn migrating shorebirds along the lower & Leito, A. 1996. Migration and staging of Mississippi River. Environmental Management the Bewick’s swan (Cygnus columbianus) in 52: 417–426. Estonia. Gibier Faune Sauvage 13: 451–461. Lewis, T.L., Esler, D. & Boyd, W.S. 2007. Luo, H.-R., Smith, L.M., Allen, B. & Haukos, Foraging behaviors of Surf Scoters and D.A. 1997. Effects of sedimentation on playa White-winged Scoters during spawning wetland volume. Ecological Applications 7: of Pacific herring. The Condor 109: 216– 247–252. 222. MacDonald, P.O., Frayer, W.E., Clauser, J.K. & Lewis, T.L., Esler, D. & Boyd, W.S. 2008. Forsythe, S.W. 1979. Documentation, Chronology Foraging behavior of Surf Scoters (Melanitta and Future Projections of Bottomland Hardwood perspicillata) and White-winged Scoters (M. Habitat Loss in the Lower Mississippi Alluvial fusca) in relation to clam density: inferring Plain. Ecological Services, U.S. Department food availability and habitat quality. Auk 125: of the Interior, Fish & Wildlife Service, 149–157. Vicksburg, Mississippi, USA. Lewis, T.L., Ward, D.H., Sedinger, J.S., Reed, A. & Madsen, J. 1980. Occurrence, habitat selection Derksen, D.V. 2013. Brant (Branta bernicla). In and roosting of the pink-footed goose at A.Poole (ed.), The Birds of North America Tipperne, Western Jutland, Dermark, 1972– Online. Cornell Laboratory of Ornithology, 1978. Dansk Ornitologisk Forenings Tidsskrift Ithaca, New York, USA. 74: 45–58. Lindberg, M.S. & Walker, J. 2007. Satellite Madsen, J. 1998. Experimental refuges for telemetry in avian research and management: migratory waterfowl in Danish wetlands. II. sample size considerations. Journal of Wildlife Tests of hunting disturbance effects. Journal Management 71: 1002–1009. of Applied Ecology 35: 398–417. Link, P.T., Afton, A.D., Cox, R.R., Jr. & Davis, B.E. Manley, S.W., Kaminski, R.M., Reinecke, K.J. & 2011. Daily movements of female Mallards Gerard, P.D. 2004. Waterbird foods in wintering in southwestern Louisiana. Waterbirds winter-managed ricefields in Mississippi. 34: 422–428. Journal of Wildlife Management 68: 74–83. Loesch, C.R., Reynolds, R.E. & Hansen, L.T. Martínez Cedillo, I., Carmona, R., Ward, D.H. & 2012. An assessment of re-directing breeding Danemann, G.D. 2013. Habitat use patterns waterfowl conservation relative to predictions of the Black Brant Branta bernicla nigricans of climate change. Journal of Fish and Wildlife (Anseriformes: Anatidae) in natural and Management 3: 1–22. artificial areas of Guerrero Negro, Baja

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 62 Autumn–winter habitat use by waterfowl

California Sur, Mexico. Revista de Biologia Miller, M.R. 1986. Northern pintail body Tropical 61: 927–935. condition during wet and dry winters in the Marty, J.R. 2013. Seed and waterbird abundances Sacramento Valley, California. Journal of in ricelands in the Gulf Coast Prairies of Wildlife Management 50: 189–198. Louisiana and Texas. M.Sc. thesis, Mississippi Miller, M.R., Orthmeyer, D.L., Casazza, M.L., State University, Mississippi State, Mississippi, McLandress, M.R. & Connelly, D.P. 1993. USA. Survival, Habitat Use and Movements of Female Mathiasson, S. 1991. Eurasian Whooper Swan Northern Pintails Radio-marked in the Suisun Cygnus cygnus migration, with particular Marsh, California. National Fish & Wildlife reference to birds wintering in southern Foundation and California Waterfowl Sweden. Wildfowl (Special Supplement No. 1): Association, Sacramento, California, USA. 201–208. Miller, M.R., Fleskes, J.P., Orthmeyer, D.L., McKelvey, R.W. & MacNeill, A.C. 1981. Newton, W.E. & Gilmer, D.S. 1995. Survival Mortality factors of wild swans in Canada. of adult female northern pintails in In G.V.T. Matthews & M. Smart (eds.), Sacramento Valley, California. Journal of Proceedings of the Second International Swan Wildlife Management 59: 478–486. Symposium, pp. 312–318. International Miller, M.R., Garr, J.D. & Coates, P.S. 2010. Waterfowl and Wetland Research Bureau, Changes in the status of harvested rice Slimbridge, UK. fields in the Sacramento Valley, California: Meineri, E., Rodriguez Perez, H., Hilaire, S. implications for wintering waterfowl. Wetlands & Mesleard, F. 2014. Distribution and 30: 939–947. reproduction of Procambarus clarkii in relation Miller, O.D., Wilson, J.A., Ditchkoff, S.S. & to water management, salinity and habitat Lochmiller, R.L. 2000. Consumption of type in the Camargue. Aquatic Conservation: agricultural and natural foods by waterfowl Marine and Freshwater Ecosystems 24: 312–323. migrating through central Oklahoma. Michot, T., Woodin, M. & Nault, A. 2008. Food Proceedings of the Oklahoma Academy of Science habits of Redheads (Aythya Americana) 80: 25–31. wintering in seagrass beds of coastal Mitchell, C.A., Custer, T.W. & Zwank, P.J. 1994. Louisiana and Texas, USA. Acta Zoologica Herbivory on shoalgrass by wintering Academiae Scientiarum Hungaricae 54: 239– redheads in Texas. Journal of Wildlife 250. Management 58: 131–141. Michot, T.C. 1996. Marsh loss in coastal Mitchell, C.D. & Eichholz, M.W. 2010. Trumpeter Louisiana: implications for management of swan (Cygnus buccinator). In A.Poole (ed.), The North American Anatidae. Gibier Faune Birds of North America Online. Cornell Sauvage 13: 941–957. Laboratory of Ornithology, Ithaca, New Michot, T.C., Woodin, M.C., Adair, S.E. & Moser, York, USA. E.B. 2006. Diurnal time activity budgets of Mitsch, W.J. & Gosselink, J.G. 2007. Wetlands. Redheads (Aythya americana) wintering in John Wiley and Sons, Inc., Hoboken, New seagrass beds and coastal ponds in Louisiana Jersey, USA. and Texas. Hydrobiologia 576: 113–128. Moon, J.A. & Haukos, D.A. 2006. Survival of Migratory Bird Data Center. 2009. http:// female northern pintails wintering in the mbdcapps.fws.gov/mbdc/databases/afsos/ Playa Lakes Region of northwestern Texas. aboutafsos.html. Journal of Wildlife Management 70: 777–783.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 63

Moon, J.A., Haukos, D.A. & Smith, L.M. 2007. Nichols, F.H., Cloern, J.E., Luoma, S.N. & Declining body condition of northern Peterson, D.H. 1986. The modification of an pintails wintering in the Playa Lakes Region. estuary. Science 231: 567–573. Journal of Wildlife Management 71: 218–221. Nichols, J.D., Reinecke, K.J. & Hines, J.E. 1983. Moore, J.E., Colwell, M.A., Mathis, R.L. & Black, Factors affecting the distribution of Mallards J.M. 2004. Staging of Pacific flyway Brant wintering in the Mississippi Alluvial Valley. in relation to eelgrass abundance and Auk 100: 932–946. site isolation, with special consideration Nolet, B.A., Langevoord, O., Bevan, R.M., of Humboldt Bay, California. Biological Engelaar, K.R. Klaassen, M., Mulder, R.J.W. Conservation 115: 475–486. & Van Dijk, S. 2001. Spatial variation in tuber Moore, J.E. & Black, J.M. 2006. Historical depletion by swans explained by differences changes in Black Brant Branta bernicla nigricans in net intake rates. Ecology 82: 1655–1667. use on Humboldt Bay, California. Wildlife Nolet, B.A., Bevan, R.M., Klaassen, M., Biology 12: 151–162. Langevoord, O. & Van Der Heijden, Y.G.J.T. Morrison, M.L., Marcot, B.G. & Mannan, R.W. 2002. Habitat switching by Bewick’s Swans: 1992. Wildlife-Habitat Relationships: Concepts maximization of average long-term energy and Applications. Island Press, Washington gain? Journal of Animal Ecology 71: 979– D.C., USA. 993. Moser, T.J. 2006. The 2005 North American Norris, D.R. & Marra, P.P. 2007. Seasonal Trumpeter Swan Survey. Division of Migratory interactions, habitat quality, and population Bird Management, U.S. Fish & Wildlife dynamics in migratory birds. The Condor 109: Service, Denver, Colorado, USA. 535–547. Moulton, D.W., Dahl, T.E. & Dall, D.M. 1997. North American Waterfowl Management Plan. Texas Coastal Wetlands; Status and Trends, Mid- 1986. North American Waterfowl Management 1950s to Early 1990s. U.S. Fish & Wildlife Plan. A Strategy for Cooperation. U.S. Department Service, Albuquerque, New Mexico, USA. of the Interior, Washington D.C., USA and Mulhern, J.H., Nudds, T.D. & Neal, B.R. 1985. Environment Canada, Ottawa, Ontario, Wetland selection by Mallards and Blue- Canada. winged Teal. Wilson Bulletin 97: 473–485. North American Waterfowl Management Plan Munro, R.E. 1981. Field feeding by Cygnus (NAWMP). 2012. People conserving waterfowl columbianus columbianus in Maryland. In G. V. and wetlands. http://nawmprevision.org/. T. Mathews and M. Smart (eds.), Second North American Waterfowl Management Plan International Swan Symposium, pp. 261–272. Committee. 2004. North American Waterfowl IWRB, Sapporo, Japan. Management Plan 2004. Implementation Nagel, J. 1965. Field feeding of Whistling Framework: Strengthening the Biological Foundation. Swans in northern Utah. Condor 67: 446– Canadian Wildlife Service, Ottawa, Ontario, 447. Canada, U.S. Fish & Wildlife Service, Neu, C.W., Byers, C.R. & Peek, J.M. 1974. A Washington D.C, USA and Secretaria de technique for analysis of utilization- Medio Ambiente y Recursos Naturales, availability data. Journal of Wildlife Management Mexico City, Mexico. 38: 541–545. O’Neal, B.J., Stafford, J.D. & Larkin, R.P. 2010. Newton, I. 2007. The Migration Ecology of Birds. Waterfowl on weather radar: applying Elsevier, London, UK. ground truth to classify and quantify bird

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 64 Autumn–winter habitat use by waterfowl

movements. Journal of Field Ornithology 81: associations between wintering dabbling 71–82. ducks and wetland complexes in Mississippi. Obenberger, S.M. 1982. Numerical response of Wetlands 32: 859–869. wintering waterfowl to macrohabitat in the Percival, S., Sutherland, W. & Evans, P. 1998. Southern High Plains of Texas. M.Sc. thesis, Intertidal habitat loss and wildfowl numbers: Texas Tech University, Lubbock, Texas, applications of a spatial depletion model. USA. Journal of Applied Ecology 35: 57–63. Olsen, R.E. & Cox, R.R., Jr. 2003. Body size and Perry, M. & Deller, A. 1995. Waterfowl population condition of male Mallards during mid- trends in the Chesapeake Bay area. In P. Hill & winter in North Dakota, USA. Waterbirds 26: S. Nelson (eds.), Proceedings of the 1994 449–456. Chesapeake Research Conference. Toward a Olsen, R.E., Cox, R.R., Jr., Afton, A.D. & Sustainable Coastal Watershed: The Chesapeake Ankney, C.D. 2011. Diet and gut morphology Experiment, pp. 490–500. Chesapeake Research of male Mallards during winter in North Consortium, Edgewater, Maryland, USA. Dakota. Waterbirds 34: 59–69. Perry, M.C., Wells-Berlin, A.M., Kidwell, D.M. & Ottersen, G., Planque, B., Belgrano, A., Post, E., Osenton, P.C. 2007. Temporal changes of Reid, P.C. & Stenseth, N.C. 2001. Ecological populations and trophic relationships of effects of the North Atlantic oscillation. wintering diving ducks in Chesapeake Bay. Oecologia 128: 1–14. Waterbirds 30: 4–16. Owen, M. & Williams, G. 1976. Winter Petersen, I.K., Christensen, T.K., Kahlert, J., distribution and habitat requirements of Desholm, M. & Fox, A.D. 2006. Final Results Wigeon in Britain. Wildfowl 27: 83–90. of Bird Studies at the Offshore Wind Farms at Owen, M. 1980. Wild Geese of the World: their Life Nysted and Horns Rev, Denmark. National History and Ecology. Batsford, London, UK. Environmental Research Institute, Ministry Owen, M., Atkinson-Willes, G.L. & Salmon, D.G. of Environment, Copenhagen, Denmark. 1986. Wildfowl in Great Britain – Second edition. Peterson, C.H. 1982. Clam predation by whelks Cambridge University Press, Cambridge, UK. (Busycon spp.): experimental tests of the Pacific Coast Joint Venture. 2004. Strategic Plan. importance of prey size, prey density and https://projects.atlas.ca.gov/frs/download. seagrass cover. Marine Biology 66: 159–170. php/15218/PCJV_StrategicPlan_2004.pdf. Peterson, C.H., Summerson, H. & Duncan, P. Pearse, A.T. 2007. Design, evaluation, and 1984. The influence of seagrass cover on applications of an aerial survey to estimate population structure and individual growth abundance of wintering waterfowl in rate of a suspension-feeding bivalve, Mercenaria Mississippi. Ph.D. dissertation, Mississippi mercenaria. Journal of Marine Research 42: 123–138. State University, Mississippi State, Peterson, T.G. 2014. Wintering waterfowl use of Mississippi, USA. Delta National Forest, Mississippi. M.Sc. Pearse, A.T., Dinsmore, S.J., Kaminski, R.M. & thesis, Mississippi State University, Mississippi Reinecke, K.J. 2008. Evaluation of an aerial State, Mississippi, USA. survey to estimate abundance of wintering Petrie, M., Vest, J. & Smith, D. 2013. ducks in Mississippi. Journal of Wildlife Intermountain West Joint Venture Management 72: 1413–1419. Implementation Plan-Waterfowl. http://iwjv. Pearse, A.T., Kaminski, R.M., Reinecke, K.J. & org/resource/2013-implementation-plan- Dinsmore, S.J. 2012. Local and landscape chapter-4-waterfowl.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 65

Petrie, M., Brasher, M. & James, D. 2014. and geographic location. Ibis 139: 337– Estimating the Biological and Economic Contributions 352. that Rice Habitats Make in Support of North Reinecke, K.J., Kaminski, R.M., Moorehead, D.J., American Waterfowl. The Rice Foundation, Hodges, J.D. & Nassar, J.R. 1989. Mississippi Stuttgart, Arkansas, USA. Alluvial Valley. In L.M. Smith, R.L. Pederson Petrie, S.A. & Francis, C.M. 2003. Rapid increase & R.M. Kaminski (eds.), Habitat Management in the lower Great Lakes population of feral for Migrating and Wintering Waterfowl in North mute swans: a review and a recommendation. America., pp. 203–247. Texas Tech University Wildlife Society Bulletin 31: 407–416. Press, Lubbock, Texas, USA. Pilarczyk, B., Tomza-Marciniak, A., Pilarczyk, R., Reinecke, K.J., Brown, M.W. & Nassar, J.R. 1992. Kavetska, K., Rza˛d, I., Hendzel, D. & Evaluation of aerial transects for counting Marciniak, A. 2012. Selenium status in sea wintering mallards. Journal of Wildlife ducks (Melanitta fusca, Melanitta nigra and Management 56: 515–525. Clangula hyemalis) wintering on the southern Rendón, M.A., Green, A.J., Aguilera, E. & Baltic coast, Poland. Marine Biology Research 8: Almaraz, P. 2008. Status, distribution and 1019–1025. long-term changes in the waterbird Plattner, D.M., Eichholz, M.W. & Yerkes, T. 2010. community wintering in Doñana, south-west Food resources for wintering and spring Spain. Biological Conservation 141: 1371–1388. staging black ducks. Journal of Wildlife Richman, S.E. & Lovvorn, J.R. 2003. Effects of Management 74: 1554–1558. clam species dominance on nutrient and Playa Lakes Joint Venture. 2014. Priority habitats. energy acquisition by spectacled eiders in the http://pljv.org/about/habitats. Bering Sea. Marine Ecology Progress Series 261: Raveling, D.G. 2004. Waterfowl of the World: A 283–297. Comparative Perspective. University of Missouri- Richman, S.E. & Lovvorn, J.R. 2004. Relative Columbia, Gaylord Memorial Laboratory, foraging value to Lesser Scaup ducks of Puxico, Missouri, USA. native and exotic clams from San Francisco Raveling, D.G. & Heitmeyer, M.E. 1989. Bay. Ecological Applications 14: 1217–1231. Relationships of population size and Robertson, G.J. & Cooke, F. 1999. Winter recruitment of pintails to habitat conditions philopatry in migratory waterfowl. Auk 116: and harvest. Journal of Wildlife Management 53: 20–34. 1088–1103. Rodway, M.S. 2007. Timing of pairing in Rees, E.C. 1987. Conflict of choice within pairs waterfowl II: Testing the hypotheses with of Bewick’s Swans regarding their migratory Harlequin Ducks. Waterbirds 30: 506–520. movement to and from the wintering Rogers, P.M. & Hammer, D.A. 1998. Ancestral grounds. Animal Behaviour 35: 1685–1693. breeding and wintering ranges of the Rees, E.C. 2006. Bewick’s Swan. T. & A.D. Poyser, Trumpeter Swan (Cygnus buccinator) in the London, UK. Eastern United States. Bulletin of The Rees, E.C. & Beekman, J.H. 2010. Northwest Trumpeter Swan Society 27: 13–29. European Bewick’s Swans: a population in Roshier, D.A., Klomp, N.I. & Asmus, M. 2006. decline. British Birds 103: 640–650. Movements of a nomadic waterfowl, Grey Rees, E.C., Kirby, J.S. & Gilburn, A. 1997. Teal Anas gracilis, across inland Australia – Site selection by swans wintering in Britain results from satellite telemetry spanning and Ireland; the importance of habitat fifteen months. Ardea 94: 461–475.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 66 Autumn–winter habitat use by waterfowl

Ross, P.G., Luckenbach, M.W., Wachapreague, and urban areas of a lowland river system. V. & Bowman, T. 2009. Distribution, Habitat Wildfowl 40: 88–98. Characteristics, Prey Abundance and Diet of Surf Sedinger, J.S. & Alisauskas, R.T. 2014. Cross- Scoters (Melanitta perspicillata) and Long-tailed seasonal effects and the dynamics of Ducks (Clangula hyemalis) in Polyhaline Wintering waterfowl populations. Wildfowl (Special Habitats in the Mid-Atlantic Region: a Comparison Issue 4): 000–000. of Shallow Coastal Lagoons and Chesapeake Bay Serie, J.R. & Bartonek, J.C. 1991. Population Environs. U.S. Fish & Wildlife Service, Sea status and productivity of Tundra Swans, Duck Joint Venture, Anchorage, Alaska, USA. Cygnus columbianus in North America. Wildfowl Roy, C.L., Herwig, C.M. & Doherty, P.F. 2013. (Special Supplement No. 1): 172–177. Mortality and refuge use by young Shaughnessy, F.J., Gilkerson, W., Black, J.M., ring necked ducks before and during hunting Ward, D.H. & Petrie, M. 2012. Predicted season in north central Minnesota. Journal of eelgrass response to sea level rise and its Wildlife Management 77: 947–956. availability to foraging Black Brant in Pacific Salewski, V. & Bruderer, B. 2007. The coast estuaries. Ecological Applications 22: evolution of bird migration – a synthesis. 1743–1761. Naturwissenschaften 94: 268–279. Silverman, E.D., Saalfeld, D.T., Leirness, J.B. & Schummer, M.L., Petrie, S.A. & Bailey, R.C. Koneff, M.D. 2013. Wintering sea duck 2008. Interaction between macroinvertebrate distribution along the Atlantic Coast of the abundance and habitat use by diving ducks United States. Journal of Fish and Wildlife during winter on northeastern Lake Ontario. Management 4: 178–198. Journal of Great Lakes Research 34: 54–71. Skalos, D.A. 2012. Evaluating body condition Schummer, M.L., Kaminski, R.M., Raedeke, A.H. and predicting lipid mass of wintering Pacific & Graber, D.A. 2010. Weather related indices greater white-fronted geese (Anser albifrons of autumn–winter dabbling duck abundance frontalis). M.Sc. thesis, University of in middle North America. Journal of Wildlife California-Davis, Davis, California, USA. Management 74: 94–101. Skov, H., Heinänen, S., Žydelis, R., Bellebaum, J., Schummer, M.L., Cohen, J., Kaminski, R.M., Bzoma, S., Dagys, M., Durinck, J., Garthe, S., Brown, M.E. & Wax, C.L. 2014. Atmospheric Grishanov, G., Hario, M., Kieckbusch, J.J., teleconnections and Eurasian snow cover Kube, J., Kuresoo, A., Larsson, K., Luigujoe, as predictors of a weather severity index L., Meissner, W., Nehls, H.W., Nilsson, L., in relation to Mallard Anas platyrhynchos Petersen, I.K., Mikkola-Roos, M., Pihl, S., autumn–winter migration. Wildfowl (Special Sonntag, N. & Stipniece, A. 2011. Waterbird Issue 4): 451–469. Populations and Pressures in the Baltic Sea. Scott, D.A. & Rose, P. 1996. Atlas of Anatidae Nordic Council of Ministers, Copenhagen, Populations in Africa and Western Eurasia. Wetland Denmark. International Publication No. 41. Wetlands Smeins, F.E., Diamond, D.D. & Hanselka, C.W. International, Waginingen, Netherlands. 1991. Coastal prairie. In R.T. Coupland (ed.), Sea Duck Joint Venture. 2004. Sea duck Ecosystems of the World 8A. Natural Grasslands: information series. www.seaduckjv.org/ Introduction and Western Hemisphere, pp. 269– infoseries/toc.html. 290. Elsevier Science, New York, USA. Sears, J. 1989. Feeding activity and body Smith, L.M., Vangilder, L.D. & Kennamer, R.A. condition of Mute Swans Cygnus olor in rural 1985. Foods of wintering brant in eastern

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 67

North America. Journal of Field Ornithology 56: Takekawa, J., Miles, A., Tsao-Melcer, D., 286–289. Schoellhamer, D., Fregien, S. & Athearn, Smith, L.M. & Sheeley, D.G. 1993. Factors N. 2009. Dietary flexibility in three affecting condition of northern pintails representative waterbirds across salinity and wintering in the Southern High Plains. Journal depth gradients in salt ponds of San of Wildlife Management 57: 62–71. Francisco Bay. Hydrobiologia 626: 155–168. Smith, L.M. 2003. Playas of the Great Plains. Tamisier, A. & Grillas, P. 1994. A review of habitat University of Texas Press, Austin, Texas, changes in the Camargue: an assessment of USA. the effects of the loss of biological diversity Söderquist, P., Gunnarsson, G. & Elmberg, J. on the wintering waterfowl community. 2013. Longevity and migration distance Biological Conservation 70: 39–47. differ between wild and hand-reared Tate J., Jr. & Tate, D.J. 1966. Additional records of Mallards Anas platyrhynchos in Northern Whistling Swans feeding in dry fields. The Europe. European Journal of Wildlife Research Condor 68: 398–399. 59: 159–166. Thomas, G. 1976. Habitat usage of wintering St. James, E.A., Schummer, M.L., Kaminski, R.M., ducks at the Ouse Washes, England. Wildfowl Penny, E.J. & Burger, L.W. 2013. Effect of 27: 148–151. weekly hunting frequency on duck abundances Thompson, J.D. & Baldassarre, G.A. 1990. in Mississippi wildlife management areas. Carcass composition of nonbreeding Blue- Journal of Fish and Wildlife Management 4: winged Teal and Northern Pintails in 144–150. Yucatan, Mexico. The Condor 92: 1057–1065. Stafford, J.D., Kaminski, R.M., Reinecke, K.J. & Tombre, I.M., Høgda, K.A., Madsen, J., Griffin, Manley, S.W. 2006. Waste rice for waterfowl L.R., Kuijken, E., Shimmings, P., Rees, E. & in the Mississippi Alluvial Valley. Journal of Verscheure, C. 2008. The onset of spring and Wildlife Management 70: 61–69. timing of migration in two arctic nesting Stafford, J.D., Janke, A.K., Anteau, M.J., Pearse, goose populations: the pink-footed goose A.T., Fox, A.D., Elmberg, J., Straub, J.N., Anser brachyrhynchus and the barnacle goose Eichholz, M.W. & Arzel, C. 2014. Spring Branta leucopsis. Journal of Avian Biology 39: migration of waterfowl in the northern 691–703. hemisphere: a conservation perspective. Toral, G.M. & Figuerola, J. 2010. Unraveling the Wilfowl (Special Issue 4): 70–85. importance of rice fields for waterbird Stenseth, N.C., Mysterud, A., Ottersen, G., populations in Europe. Biodiversity and Hurrell, J.W., Chan, K.-S. & Lima, M. 2002. Conservation 19: 3459–3469. Ecological effects of climate fluctuations. Tourenq, C., Bennetts, R.E., Kowalski, H., Vialet, Science 297: 1292–1296. E., Lucchesi, J.L., Kayser, Y. & Isenmann, P. Stott, R.S. & Olson, D.P. 1973. Food-habitat 2001. Are ricefields a good alternative to relationship of sea ducks on the New natural marshes for waterbird communities Hampshire coastline. Ecology 54: 996– in the Camargue, southern France? Biological 1007. Conservation 100: 335–343. Systad, G.H., Bustnes, J.O. & Erikstad, K.E. U.S. Department of Agriculture. 2010. Catfish 2000. Behavioral responses to decreasing day 2010. Part I: Reference of Catfish Health and length in wintering sea ducks. Auk 117: Production Practices in the United States, 2009. 33–40. U.S. Department of Agriculture – Animal

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 68 Autumn–winter habitat use by waterfowl

and Plant Health Inspection Service, Fort and climate on population dynamics. Global Collins, Colorado, USA. Change Biology 11: 869–880. Väänänen, V.-M. 2001. Hunting disturbance and Weaver, K.H. 2013. Tundra Swan (Cygnus the timing of autumn migration in Anas columbianus columbianus) habitat selection species. Wildlife Biology 7: 3–9. during the nonbreeding period. M.Sc. thesis, Varner, D.M. 2008. Survival and foraging ecology The University of Western Ontario, London, of Interior Population trumpeter swans. Ontario, Canada. M.Sc. thesis, Southern Illinois Universy, Weller, M.W. 1964. Distribution and migration of Carbondale, Illinois, USA. the redhead. Journal of Wildlife Management 28: van Eerden, M.R. (ed.). 1997. Patchwork: patch 64–103. use, habitat exploitation and carrying capacity Wetlands International 2014. Waterbird Population for water birds in Dutch freshwater wetlands. Estimates. http://wpe.wetlands.org. Ph.D. thesis, Rijksuniversiteit Groningen and Wiens, J.A. 1973. Pattern and process in grassland Rijkswaterstaat Directie IJsselmeergebied, bird communities. Ecological Monographs 43: Lelystad, Netherlands. 237–270. Venne, L.S., Anderson, T.A., Zhang, B., Smith, Wiens, J.A. 1989. The Ecology of Bird Communities. L.M. & McMurry, S.T. 2008. Organochlorine Volume 1. Foundations and Patterns. Cambridge pesticide concentrations in sediment and University Press, Cambridge, UK. amphibian tissue in playa wetlands in the Williams III, S.O. & Chabreck, R.H. 1986. Southern High Plains, USA. Bulletin of Quantity and quality of waterfowl habitat in Environmental Contamination and Toxicology 80: Louisiana. Research Report No. 8. Louisiana 497–501. State University, Baton Rouge, Louisiana, Vest, J.L., Kaminski, R.M., Afton, A.D. & USA. Vilella, F.J. 2006. Body mass of lesser scaup Wiseman, A.J., Kaminski, R.M., Riffell, S., during fall and winter in the Mississippi Reinecke, K.J. & Larson, E.J. 2010. Ratoon flyway. Journal of Wildlife Management 70: grain sorghum and other seeds for waterfowl 1789–1795. in sorghum croplands. Proceedings of the Vickery, J., Sutherland, W., Watkinson, A., Southeast Association of Fish Wildlife Agencies 64: Rowcliffe, J. & Lane, S. 1995. Habitat 106–111. switching by dark-bellied Brent Geese Branta Wooten, D.E. & Werner, S.J. 2004. Food habits b. bernicla (L.) in relation to food depletion. of Lesser Scaup Aythya affinis occupying Oecologia 103: 499–508. baitfish aquaculture facilities in Arkansas. Voslamber, B., Knecht, E. & Kleijn, D. 2010. Journal of the World Aquaculture Society 35: 70– Dutch Greylag Geese Anser anser: migrants or 77. residents. Ornis Svecica 20: 207–214. Zammuto, R.M. 1986. Life histories of birds: Waldeck, P. & Larsson, K. 2013. Effects of clutch size, longevity, and body mass among winter water temperature on mass loss in North American game birds. Canadian Journal Baltic Blue Mussels: implications for foraging of Zoology 64: 2739–2749. sea ducks. Journal of Experimental Marine Zink, R.M. 2011. The evolution of avian Biology and Ecology 444: 24–30. migration. Biological Journal of the Linnean Ward, D.H., Reed, A., Sedinger, J.S., Black, J.M., Society 104: 237–250. Derksen, D.V. & Castelli, P.M. 2005. North Zipkin, E.F., Gardner, B., Gilbert, A.T., American Brant: effects of changes in habitat O’Connell, A.F., Jr., Royle, J.A. & Silverman,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 Autumn–winter habitat use by waterfowl 69

E.D. 2010. Distribution patterns of communities in the Baltic Sea. Wilson Bulletin wintering sea ducks in relation to the 117: 133–141. North Atlantic Oscillation and local Žydelis, R.M., Bellebaum, J., Österblom, H., environmental characteristics. Oecologia 163: Vetemaa, M., Schirmeister, B., Stipniece, A., 893–902. Dagys, M., van Eerden, M. & Garthe, S. Žydelis, R.M. & Ruskete, D.R. 2005. Winter 2009. Bycatch in gillnet fisheries – an foraging of Long-tailed Ducks (Clangula overlooked threat to waterbird populations. hyemalis) exploiting different benthic Biological Conservation 142: 1269–1281.

Photograph: A spring of Green-winged and Blue-winged Teal wintering in Louisiana, by Charlie Hohorst.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 17–69 70

Spring migration of waterfowl in the northern hemisphere: a conservation perspective

JOSHUA D. STAFFORD1*, ADAM K. JANKE2, MICHAEL J. ANTEAU3, AARON T. PEARSE3, ANTHONY D. FOX4, JOHAN ELMBERG5, JACOB N. STRAUB6, MICHAEL W. EICHHOLZ7 & CÉLINE ARZEL8

1South Dakota Cooperative Fish and Wildlife Research Unit, U.S. Geological Survey, Department of Natural Resource Management, South Dakota State University, Brookings, South Dakota, USA. 2Department of Natural Resource Management, South Dakota State University, Brookings, South Dakota, USA. 3U.S. Geological Survey, Northern Prairie Wildlife Research Center, Jamestown, North Dakota, USA. 4Department of Bioscience, Aarhus University, DK-8410 Rønde, Denmark. 5Aquatic Biology and Chemistry, Kristianstad University, SE-291 88 Kristianstad, Sweden. 6Center for Earth and Environment Science, State University of New York-Plattsburgh, Plattsburgh, New York, USA. 7Cooperative Wildlife Research Laboratory, Center for Ecology, Southern Illinois University Carbondale, Carbondale, Illinois, USA. 8Section of Ecology, Department of Biology, University of Turku, 20014 Turku, Finland. *Correspondence author. E-mail: [email protected]

Abstract Spring migration is a key part of the annual cycle for waterfowl populations in the northern hemisphere, due to its temporal proximity to the breeding season and because resources may be limited at one or more staging sites. Research based on field observations during spring lags behind other periods of the year, despite the potential for fitness consequences through diminished survival or cross-seasonal effects of conditions experienced during migration. Consequently, conservation strategies for waterfowl on spring migration are often only refined versions of practices used during autumn and winter. Here we discuss the current state of knowledge of habitat requirements for waterfowl at their spring migratory sites and the intrinsic and extrinsic factors that lead to variability in those requirements. The provision of plant foods has become the main conservation strategy during spring because of the birds’ energy requirements at this time, not only to fuel migration but to facilitate early clutch formation on arrival at the breeding grounds. Although energy sources are important to migrants, there is little evidence on the extent to which the availability of carbohydrate-based food is limiting for many migratory waterfowl populations.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 71

Such limitation is relatively unlikely among populations that exploit agricultural grain during migration (e.g. arctic-nesting geese), suggesting that conservation strategies for these populations may be misplaced. In general, however, we found few cases in which an ecological understanding of spring-migrating waterfowl was sufficient to indicate true resource limitation during migration, and still fewer cases where conservation efforts ameliorated these limitations. We propose a framework that aims to address knowledge gaps and apply empirical research results to conservation strategies based on documented limitations and associated fitness impacts on migrating waterfowl. Such a strategy would improve allocation of scarce conservation resources during spring migration and greatly improve ecological understanding of migratory waterfowl and their habitats in the northern hemisphere.

Key words: conservation, limitations, lipids, nutrients, spring migration, waterfowl.

Spring is a critical phase of the annual cycles resources temporarily unavailable (Trautman of waterfowl Anatidae sp. in the northern et al. 1939; Newton 2006, 2007). Further, hemisphere because of the physiological migratory movements themselves can be and environmental conditions encountered dangerous and energetically costly, requiring during migration, and the co-occurrence of individuals and flocks to exploit habitats and pre-breeding life-history events. Maintenance foods that promote survival (sensu Fretwell or acquisition of nutrient reserves at staging 1972; Kaminski & Elmberg 2014). The areas is generally necessary in order to choice of migratory strategy therefore complete migration, and is often also a represents important trade-offs with lasting prerequisite for successful breeding (Ankney consequences for individual fitness and et al. 1991; Jenni & Jenni-Eirmann 1998). population dynamics, which may be sensitive Individuals often experience diminished food to management strategies used by availability as they await the thaw of wetland conservation organisations along migratory habitats or because of food depletion by corridors in the northern hemisphere. autumn-migrating birds (Stafford et al. Habitat conditions encountered during 2006; Greer et al. 2009; Straub et al. spring migration also have potential to 2012). Moreover, in addition to migration, influence waterfowl populations through many species are undertaking energetically cross-seasonal (or carry-over) impacts on expensive activities such as courtship, pair- individual reproduction (Davis et al. 2014, bond maintenance and moulting into Sedinger & Alisauskas 2014). The seminal breeding plumage at this time (Heitmeyer works of Weller (1975), Fredrickson and 1988; Lovvorn & Barzen 1988; Richardson & Drobney (1977), Ankney and MacInnes Kaminski 1992; Hohman et al. 1997; Barras et (1978), and others (e.g. Heitmeyer & al. 2001; Anteau et al. 2011a). Adverse and Fredrickson 1981; Kaminski & Gluesing unpredictable weather can kill birds directly 1987) prompted research on the nature and or lead to starvation by making food mechanisms for cross-seasonal effects on

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 72 Waterfowl conservation during spring migration waterfowl in particular and migratory birds in of waterfowl at spring staging areas and help general (Drent & Daan 1980; Harrison et al. to set research priorities for improving 2011; Sedinger & Alisauskas 2014). This our understanding of migratory species. work has shown evidence for various cross- Specifically, our objectives are to: 1) review seasonal relationships in waterfowl, for the general requirements of waterfowl populations both in Europe and in North on spring migration and discuss intrinsic America, and ranging across species from and extrinsic factors that influence these those using capital breeding strategies (e.g. requirements, 2) discuss inter-specific Lesser Snow Goose Chen c. caerulescens; differences in the requirements of migrating Alisauskas 2002) to those which mainly waterfowl and the limitations imposed acquire the food resources needed for egg- by habitat conditions encountered during laying on or near the breeding territories spring, and 3) propose a framework for (“income breeders”; e.g. Eurasian Teal Anas evaluating limitations on waterfowl during crecca crecca; Guillemain et al. 2008). Although spring migration, and for implementing spring migration is widely recognised as being habitat management and conservation that an important time both for individual may alleviate or mitigate these limitations. survival and for subsequent breeding success, it remains largely understudied in comparison General requirements of spring- with other stages in the annual cycle (Arzel migrating waterfowl et al. 2006), and management strategies Body reserves which can be converted into during this period are often refinements of metabolic energy are the most recognised practices intended for breeding or wintering currency for avian migration (Jenni & Jenni- populations (Soulierre et al. 2007). Eirmann 1998) and are also necessary In this paper we synthesise published for subsequent reproduction in many information on the habitat requirements of waterfowl species (Ankney et al. 1991). waterfowl during spring migration and Lipids provide the most efficient means of discuss potential applications of the storing energy for migration, and lipid knowledge for conservation initiatives at metabolism therefore is considered a key migratory stopover areas. Arzel et al. (2006) factor influencing onwards migration and has made a comprehensive review of the selection of stopover sites. Individuals the current state of literature on spring- are expected to choose habitats where migrating waterfowl, so we do not intend energy sources are readily available during to repeat their work here. Rather, we migration and avoid energetically expensive endeavour to assess available information staging areas (Bauer et al. 2008; Mini & Black and consider gaps in knowledge that have 2009; Brasher 2010). Waterfowl gain energy the potential to diminish the efficacy of and build lipid reserves from seeds, other conservation strategies aimed at enhancing plant material and invertebrates, with habitat conditions for waterfowl on spring the relative contribution of each to the migration. Identifying knowledge gaps can diet varying considerably among species. inform the management and conservation Similarly, the relative distribution of plant

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 73 and invertebrate foods varies across suggesting that foraging decisions may be different foraging habitats, which has influenced by the specific amino acids to be species-specific implications for food found in food items (Heitmeyer 1988). availability at each site (Straub et al. 2012). However, our understanding of the role of Many North American species use abundant specific nutrients (particularly at the waste agricultural seeds in croplands as essential amino acids and fatty acids level) a carbohydrate source during spring for maintaining body condition at different migration (Krapu et al. 1995; Anteau et al. times of year is still in its infancy. 2011b, Pearse et al. 2011). Others rely on The most basic requirement for all invertebrates to build lipid reserves, which waterfowl (and indeed for most living are likely to be more variable in abundance organisms) is water. Water is gained primarily and distribution, and also are apparently by drinking, but it can also be acquired in the declining in some regions (Anderson 1959; diet or derived through metabolic pathways. Wilson et al. 1995; Anteau & Afton 2008a,b; Wetlands provide not only a water source Anteau et al. 2011c; Straub et al. 2012). but are important for a range of functions Nutrients other than lipids are also most notably foraging, roosting, pair required by waterfowl during spring, most formation (Anderson & Titman 1992), safety notably protein, essential amino acids and from predators, isolation from disturbance minerals (e.g. calcium). Evidence from and protection from inclement weather Snow Geese suggests that some waterfowl conditions (LaGrange & Dinsmore 1989; mobilise protein reserves gained during Havera et al. 1992; Zimmer et al. 2010). migration for subsequent reproduction Research in Europe revealed that many (Ankney & MacInnes 1978; Gauthier et al. geese migrate toward breeding areas 2003). Moreover, moulting birds require through agricultural regions, but that bird protein to synthesise feather tissue distributions may be constrained within a (Heitmeyer 1988; Barras et al. 2001); some radius of a safe body of water or ice that can species consume protein-rich foods during function as a predator-free overnight roost contour feather moult (Fox et al. 1998; (e.g. the spring migration of Pink-footed Anderson et al. 2000; Anteau et al. 2011a), Geese Anser brachyrhynchus within Britain is whereas protein reserves may be related to thus confined to particular areas; Bell 1988; contour feather moult intensity in other Fox et al. 1994). Similar patterns have been species (Lovvorn & Barzen 1988). Protein is shown with migrating Mallard Anas also required for repairing muscles injured platyrhynchos (LaGrange & Dinsmore 1989) or catabolised during flight, similar to the and geese (Anteau et al. 2011b) in agricultural way in which fat reserves are consumed and landscapes in central North America. replenished during migration (Guglielmo et al. 2001; Piersma 2002). Earlier work Factors influencing waterfowl has established the importance of a diverse requirements during spring migration diet for maintaining body condition Although these are generally universal during winter (Loesch & Kaminski 1989), for waterfowl during spring migration,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 74 Waterfowl conservation during spring migration the relative importance of each varies strategies leads to considerable variation within and among species in response in the conditions required by waterfowl to conditions encountered en route (e.g. throughout migration. Birds expected to weather, disturbance) and in accordance adhere primarily to a capital breeding with their migration and/or breeding strategy (e.g. arctic nesting geese) need more strategies. Weather can influence individual resources from stopover locations than requirements during migration, particularly those using an income-breeding or local- among early migrants that may encounter capital strategy (sensu Klaassen et al. 2006), in physiologically demanding conditions on which they acquire most breeding resources reaching high latitudes before the ice and and nutrients from breeding habitats. snow has melted in the northern part of Variation in migration strategies among their range. For example, LaMontagne et al. species invoking an income-breeding (2001) reported differences in foraging strategy further differentiates requirements activity among spring-migrating Trumpeter throughout migration. Income migrant Swans Cygnus buccinator in response to cold waterfowl (e.g. Eurasian Teal; Arzel et al. temperatures encountered during migration. 2007) rely especially on lipids acquired at It is likely that early migrants exposed staging sites to fuel subsequent flights, to wide variations in temperature and whereas other species may carry reserves to precipitation during spring would exhibit facilitate onward migratory movement similar weather-dependent foraging and (Krapu et al. 1995; Pearse et al. 2011). Across roosting behaviours, such as hyperphagia or this gradient, from capital breeding species seeking thermal cover. Weather affects to income migrants, considerable variation habitat conditions along the migration in nutrient accumulation and storage rates route, and generally influences the have been documented throughout spring availability of food and other resources migration. For example, Garganey Anas throughout the year, as discussed further querquedula in southern France effectively below. forgo nutrient reserve accumulation during Disturbance is another important factor stopover (Guillemain et al. 2004), whereas influencing the relative importance of Greater White-fronted Geese Anser albifrons habitat requirements for migrating in the Rainwater Basin of Nebraska waterfowl (Madsen 1995), as it may affect accumulate 11–22 g (dry) mass/day in the the timing of migration strategies or staging areas (Krapu et al. 1995). individual body condition during stopover European geese provide an example of (Drent et al. 2003; Feret et al. 2003; variable requirements during migration that Pearse et al. 2012). Variation in predation manifest as a result of variable migration pressure during spring migration also may strategies. These populations rely upon the influence foraging ecology or the ability to new growth of grasses and sedges at higher exploit resources necessary for migration latitudes following the emergence of the (Guillemain et al. 2007). “green wave” of above-ground production Variation in breeding and migration following spring thaw (Drent et al. 1978; van

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 75 der Graaf et al. 2006; van Wijk et al. 2012). Distinguishing between requirements Tracking this green wave is more easily and limitations during spring undertaken in a series of relatively short migration flights, as is the case of Greater White- fronted Geese Anser albifrons in continental There is considerable variation between Europe. In contrast, Greenland White- waterfowl of the northern hemisphere in the fronted Geese A. a. flavirostris make long conditions that best match their social, overseas flights from Britain and Ireland to ecological and physiological requirements at staging areas in Iceland, and from there to different stages of migration, creating many breeding areas in west Greenland (Fox et al. challenges for research and conservation 2003). Such stepping-stone migrants have to along the flyway. However, in many cases the take calculated risks when moving onwards populations’ requirements are met through to staging areas, perhaps without adequate large-scale habitat use and selection cues to predict meteorological conditions processes (i.e. adherence to flyways with and the advancement of spring phenology necessary resources). For example, no further ahead (Fox et al. 2006; Tombre et al. studies of mid-continent goose populations 2008). Variation in the availability of spring in North America have documented nutrient staging areas has considerable effects on deficiencies during migration or upon arrival nutrient acquisition strategies adopted by on the breeding grounds, despite the the species with the same or similar body importance of nutrient reserves to these structure but in different parts of its range. populations having been established. In this The Greenland White-fronted Goose may case, adherence to the central flyway, where deplete 800–900 g of fat when flying from the availability of waste agricultural seed winter quarters to spring staging areas in exceeds population needs, ensures nutrient Iceland, and there it must acquire similar accumulation and maintenance during fat stores for the onward journey to migration and increases the likelihood of breeding areas in west Greenland (Fox successful reproduction. Thus, although et al. 2003). Remarkably, the Greenland energy is the most important nutritional population now leaves the wintering areas requirement among these migrating geese on average three weeks earlier than 25 years during spring, its availability (at least in the ago (Fox & Walsh 2012), but because of a form of carbohydrates derived from lack of warmer springs in Greenland it agricultural seeds) is not a limiting factor remains longer in Iceland (Fox et al. 2012) during migration (Krapu et al. 1995; Jefferies and fattens at a slower rate to arrive on the et al. 2004; Stafford et al. 2006; Foster et al. breeding areas at the same time as recorded 2010; Anteau et al. 2011b). This example in the 1860s (Fox et al. 2014). Such raises the need for a distinction between behaviour suggests considerable phenotypic requirements (resources that sustain plasticity in migration behaviour and migration and subsequent breeding) and ability to acquire fat stores in a fluctuating limitations (resources that are not provided environment. in sufficient supply to meet fully the needs of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 76 Waterfowl conservation during spring migration individuals at the time and/or thereafter) In the former case, management may for spring migrating populations, which in be ineffective at improving population turn should guide current and future productivity through spring migration, conservation strategies developed for whereas in the latter, large-scale efforts to these populations. Such a distinction abate food limitation would likely have is fundamental for the effective population-level implications. Therefore, implementation of conservation throughout research during spring migration should the annual cycle, but is not yet explicitly seek to identify limitations at appropriate recognised in conservation strategies for spatial scales, through intensive study of spring-migrating waterfowl. This is likely due migrant habitat use and behaviour, so to the uncertainty surrounding mechanisms as to identify important limitations on regulating populations during the period and populations. This knowledge can then be the aforementioned tendency to adapt applied to the development and delivery of wintering conservation strategies (e.g. conservation strategies for spring-migrating provision of energy) for spring-migrating waterfowl. Additionally, identification of waterfowl. Misguided conservation strategies habitat limitations during spring migration based only on requirements, rather than a might be investigated and used cautiously limiting resource, may lead to ineffective and insightfully to reduce populations of conservation. burgeoning species, such as Lesser Snow In populations where the availability of a Geese, whilst ensuring no impacts to other necessary resource is limiting, observations waterfowl species. of habitat use and distribution patterns for Evidence for habitat limitations in other the birds during migration are likely to periods of the annual cycle exists. describe these limitation(s), which may be Availability of suitable breeding habitat in driving cross-seasonal effects on population the North American prairies is a well- productivity. The opposite is also likely; documented driver of annual population when resources are not in limited supply, dynamics for a range of waterfowl (Johnson populations may be freed from constraints & Grier 1988; Bethke & Nudds 1995; on production originating during spring Reynolds et al. 2001), and population-level migration (Jefferies et al. 2004). Limitations implications of food shortages on the of waterfowl during migration vary spatially wintering grounds has similarly been and temporally, and the scale at which documented for various species (Heitmeyer limitations are assessed is important and & Fredrickson 1981; Kaminski & Gluesing should be determined by management 1987; Raveling & Heitmeyer 1989). objectives. For example, food depletions at However, few studies have been conducted local scales may or may not be consequential at an appropriate spatial and temporal scale for waterfowl populations, but regionally to document population-level limitations depressed food resources from drought or during spring. Here we highlight a case other impacts could influence annual where research has been conducted at recruitment (e.g. Davies & Cooke 1983). appropriate scales to document limitations

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 77 with the potential to inform conservation Research on wetlands used as stopovers strategies for Lesser Scaup Aythya affinis on during migration in the Midwest indicated spring migration. Considerable research has that the availability of amphipods, an been conducted on the ecology of the important, lipid-rich food item for migrating Lesser Scaup in spring following range-wide and pre-breeding Lesser Scaup (Arts et al. population declines throughout North 1995; Lindeman & Clark 1999; Anteau & America since the 1980s. We have therefore Afton 2009b), had declined in the region in taken this body of work as an example of conjunction with the documented declines the benchmark necessary for achieving a in body condition (Anteau & Afton 2006, reasonable understanding of populations- 2008a, b; Anteau et al. 2011c). Lesser Scaup level limitations during migration, which select habitats with abundant amphipods may be applied to the conservation of other (Lindeman & Clark 1999; Anteau & Afton species during spring migration. 2009b); however, they likely use proximate Research conducted along the Lesser cues (e.g. turbidity) to identify wetlands Scaup’s mid-continent spring migration previously rich in amphipods, but which are corridors indicated that the females’ lipid now less numerous due to land use changes reserve levels had declined throughout the and invasions of fish into traditionally upper Midwest since the 1980s (Anteau & fish-free habitats (Anteau & Afton Afton 2004, 2009a), but not on wintering 2008a, 2009b). Spring is an energetically areas in the southern Mississippi Flyway and nutritionally costly period for Lesser (Anteau & Afton 2004; Vest et al. 2006). This Scaup; thus, they clearly require both work led to the spring-condition hypothesis, nutrient- and energy-rich foods. However, which predicted that females were unable to their adaptation to consuming an animal- acquire energy or nutrients required during based diet and the evidence reviewed above spring migration from stopover habitats suggests that Lesser Scaup are likely limited through the Midwest. The limitation was by lipid availability during migration in the predicted to result in decreased survival Midwest. Further research in the region or diminished productivity from poor suggested that amphipod abundance in condition upon arrival in the breeding areas, stopover habitats may be subject to reduced breeding propensity, delayed management (Anteau & Afton 2008a; breeding or a combination of these (Anteau Anteau et al. 2011c) and could focus on & Afton 2004, 2009a). Further research regions with high annual Lesser Scaup use in demonstrated that females were catabolising the Midwest (Anteau & Afton 2009b). lipid reserves at spring migration stopover Accordingly, implications of the research areas throughout the Midwest where they were to focus on identifying these key were expected to be storing lipids, lending Lesser Scaup migration habitats in the support to the proposed link between region and to undertake work to improve habitat conditions during migration and the availability, quality and productivity of diminished condition prior to the breeding amphipods and other invertebrate food season (Anteau & Afton 2011). sources through wetland conservation

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 78 Waterfowl conservation during spring migration practices, such as the implementation of requirements for spring-migrating Lesser upland vegetation buffers and manipulating Snow Geese are energy and protein, this fish densities (Anteau et al. 2011c). example illustrates that the habitat factor This example with Lesser Scaup illustrates most appropriate for management is the the role of large-scale research on ecology distribution of the primary limiting resource, and habitat use of migrating waterfowl, wetlands. for identifying limitations and focusing management of those limitations at relevant Conclusions scales, with a view to improving conditions Conservation planning during spring has encountered during migration. Large- traditionally focused on ensuring adequate scale relationships between population food energy at stopover locations because productivity and limitations encountered of the well-established importance of during spring migration are one case in energy during migration and the importance which an explicit focus on limitations in of lipid reserves for subsequent breeding conservation is appropriate. A similar focus (cf., Ankney et al. 1991). For example, the has applications at finer spatio-temporal North American Waterfowl Management scales for improving local management Plan, through its Joint Ventures (i.e. public efforts and the value of habitat reserves and private partnerships that plan and intended for use by spring migrants. For implement conservation activities), has example, detailed studies of Lesser Snow typically adopted an approach of estimating Goose stopover ecology in the Rainwater the energetic carrying capacity (ECC) for a Basin of central Nebraska has shown that region based on estimated waterfowl fine-scale habitat features rather than energy population levels and goals during the non- requirements during migration drive the breeding periods of the annual cycle. In this birds’ use of space at staging sites (Pearse scenario, management activities during et al. 2010; Anteau et al. 2011c; Sherfy et al. spring migration target provision of food 2011). Cornfields dominate the landscape in resources to meet the energetic needs of the Rainwater Basin, resulting in an waterfowl, typically through wetland estimated 10-fold net surplus of energy for creation, enhancement or management to Lesser Snow Geese and other migratory produce carbohydrate-rich plant foods. waterfowl (Bishop et al. 2008), but their The ECC approach relies on the critical location in relation to wetland roosts appears assumption that energy derived from to be more important than variability wetland habitats is the main requirement between fields in the availability of waste limiting waterfowl populations in spring. grain. Changes in the distribution and area of However, this assumption is largely untested wetlands therefore would likely have the for the spring migration period, and the greatest influence on space use by Lesser importance of energy may not equate with Snow Geese and other waterfowl in the limitation should food availability exceed the region (Vrtiska & Sullivan 2009; Webb et al. requirements of the population. We 2010). Although the main nutritional contend that the assumption that energy is

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 79 the primary limiting factor for waterfowl focused on the ecology of relevant species populations in spring may be untenable for would lead to more efficient allocation of many species that supplement their diet with resources and be more likely to affect residual agricultural food sources. Rather, measurable impacts to populations. any of the requirements of spring-migrating Habitat use and selection, along with diet waterfowl discussed herein, or perhaps and behavioural studies, can provide the those yet undocumented, could limit spring foundation for determining what might be migrants and have annual implications for limiting a certain species at a certain staging survival or population productivity at area, if at all (Callicutt et al. 2011; Hagy & various temporal and spatial scales. Kaminski 2012). Some studies of this nature In some cases, energy availability to date have identified cross-seasonal effects motivates habitat use and appears to limit related to spring limitations, but many population growth (e.g. the Lesser Scaup questions remain and adoption of novel example described above). However, in research will be necessary to resolve them. other cases, energy during spring may not be Telemetry and other spatially explicit limiting and can be in surplus (e.g. for mid- individual-based studies and local-scale continent geese in North America). We surveys of waterfowl concentrations would suggest that the framework of ECC help identify factors associated with models could be reconsidered and perhaps improved individual fitness in response to restructured to evaluate whether energy is conditions experienced during spring and limiting for a species or guild within a identify key habitats for foraging and conservation region. This would require a roosting, respectively. Similarly, such studies more comprehensive ECC that assumes may yield insights into the relative role that all sources of energy are equal if they of specific stopovers during migration are available to the species/group (e.g. and assist in prioritising further research agricultural versus wetland) and would and conservation across the expansive require detailed information on the birds’ landscapes transited between wintering and diet and foraging behaviour given that all breeding areas. food sources – agricultural foods, wetland Many factors drive hierarchical resource seeds and plants and invertebrates – must be selection and knowledge of these factors considered as sources of energy. If careful can inform conservation strategies (Johnson evaluation indicates that available energy 1980). For example, intensive research on exceeds requirements for a given population Sandhill Cranes Grus canadensis during spring or region, a focus on identifying or migration at a major stopover area in central managing other possible limiting factors, North America revealed that access to if they exist, would be prudent. Such protein constituents of their diet was a an approach may change the focus of strong driver of fine-scale habitat selection conservation and management for some during staging, despite accounting for only c. organisations (e.g. Joint Ventures, resource 3% of their diet in the region (Krapu et al. agencies); however, such an endeavour re- 1984; Reinecke & Krapu 1986). This

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 80 Waterfowl conservation during spring migration research suggests that habitat selection improved this manuscript considerably. Any and time investments among migrating use of trade, product or firm names is for waterbirds can be considerable in the descriptive purposes only and does not imply acquisition of an apparently rare but endorsement by the U.S. Government. important resource (i.e. a limiting resource). References Similarly detailed studies for ducks and geese could assist in identifying other Alisauskas R.T. 2002. Arctic climate, spring potential limiting factors. nutrition, and recruitment in mid-continent lesser snow geese. Journal of Wildlife Accomplishing a revised focus on Management 66: 181–193. limitations during spring conservation will Anderson, H.G. 1959. Food habits of migratory require knowing the precise demands of the ducks in Illinois. Illinois Natural History Survey birds along the route of staging sites at Bulletin 27: 289–344. different times during migration, and Anderson, M.G. & Titman, R.D. 1992. Spacing delivering the appropriate energy, protein, patterns. In B.D.J. Batt, A.D. Afton, M.G. water and other resources such that birds Anderson, C.D. Ankney, D.H. Johnson, J.A. may access them under a range of Kadlec & G.L. Krapu (eds.), Ecology and conditions (e.g. land use or climate change). Management of Breeding Waterfowl, pp. 251–289. Recognising the opportunities of expanded, University of Minnesota Press, Minneapolis, individual-based cross-seasonal studies USA. opens up a portfolio of research objectives Anderson, J.T., Smith, L.M. & Haukos, D.A. 2000. Food selection and feather molt by that asks what birds need during spring, and nonbreeding American green-winged teal in how can we provide them most effectively, Texas playas. Journal of Wildlife Management 64: in a way that enhances condition, survival 222–230. and preparation for the breeding season, Ankney, C.D. & MacInnes, C.D. 1978. Nutrient regardless of species. The challenge for reserves and reproductive performance of conservation will be to provide resources in female Lesser Snow Geese. Auk 95: 459–471. adequate quantities to confer benefits to Ankney, C.D., Afton, A.D. & Alisauskas, R.T. individuals of targeted populations and 1991. The role of nutrient reserves in limiting species. Such a task would be difficult given waterfowl reproduction. Condor 93: 1029– the dearth of information on factors truly 1032. limiting some waterfowl populations during Anteau, M.J. & Afton, A.D. 2004. Nutrient spring. Nonetheless, until limiting factors reserves of lesser scaup during spring migration in the Mississippi Flyway: a test of are identified (or ruled out) it would be the Spring Condition Hypothesis. Auk 121: difficult to design and implement truly 917–929. effective conservation programmes for Anteau, M.J. & Afton, A.D. 2006. Diet shifts of populations of conservation concern. lesser scaup are consistent with the Spring Condition Hypothesis. Canadian Journal of Acknowledgements Zoology 84: 779–786. We thank R.M. Kaminski, D.C. Kesler and Anteau, M.J. & Afton, A.D. 2008a. Amphipod an anonymous reviewer for comments that densities and indices of wetland quality

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 81

across the upper-Midwest, USA. Wetlands 28: Arts, M., Ferguson, M.E., Glozier, N.E., Robarts, 184–196. R.D. & Donald, D.B. 1995. Spatial and Anteau, M.J. & Afton, A.D. 2008b. Diets of temporal variability in lipid dynamics of spring migrating lesser scaup throughout the common amphipods: assessing the potential upper-Midwest are consistent with the for uptake of lipophilic contaminants. Spring Condition Hypothesis. Waterbirds 31: Ecotoxicology 4: 91–113. 97–106. Barras, S.C., Kaminski, R.M. & Brennan, Anteau, M.J. & Afton, A.D. 2009a. Lipid reserves L.A. 2001. Effect of winter-diet restriction of lesser scaup (Aythya affinis) migrating on prebasic molt in female wood ducks. across a large landscape are consistent with Proceedings of the Annual Conference of the the Spring Condition Hypothesis. Auk 126: Southeastern Association of Fish and Wildlife 873–883. Agencies 55: 506–516. Anteau, M.J. & Afton, A.D. 2009b. Wetland use Bauer, S., Van Dinther, M., Høgda, K., Klaassen, and feeding by lesser scaup during spring M. & Madsen, J. 2008. The consequences of migration across the upper Midwest, USA. climate-driven stop-over sites changes on Wetlands 29: 704–712. migration schedules and fitness of Arctic Anteau, M.J. & Afton, A.D. 2011. Lipid geese. Journal of Animal Ecology 77: 654–660. catabolism of invertebrate predator indicates Bell, M.V. 1988. Feeding behaviour of wintering widespread wetland ecosystem degradation. Pink-footed and Greylag Geese in north-east PLoS One 6: e16029. Scotland. Wildfowl 39: 43–53. Anteau, M.J., Anteau, A.C.E. & Afton, A.D. Bethke, R.W. & Nudds, T.D. 1995. Effects 2011a. Testing competing hypotheses for of climate-change and land-use on duck chronology and intensity of lesser scaup abundance in the Canadian prairie-parklands. molt during winter and spring migration. Ecological Applications 5: 588–600. Condor 113: 298–305. Bishop, A.A. & Vrtiska, M. 2008. Effects of the Anteau, M.J., Sherfy, M.H. & Bishop, A. 2011b. Wetlands Reserve Program on waterfowl Location and agricultural practices influence carrying capacity in the Rainwater Basin spring use of harvested corn fields by cranes Region of south-central Nebraska. Natural and geese in Nebraska. Journal of Wildlife Resource Conservation Service, U.S. Management 75: 1004–1011. Department of Agriculture, USA. Accessible Anteau, M.J., Afton, A.D., Anteau, A.C.E. & online at http://www.nrcs.usda.gov/technical/ Moser, E.B. 2011c. Fish and land use NRI/ceap/library.html (last accessed 20 influence Gammarus lacustris and Hyalella azteca August 2014). (Amphipoda) densities in large wetlands Brasher, M.G. 2010. Duck use and energetic across the upper Midwest. Hydrobiologia 664: carrying capacity of actively and passively 69–80. managed wetlands in Ohio during autumn Arzel, C., Elmberg, J. & Guillemain, M. 2006. and spring migration. Ph.D. thesis, Ohio Ecology of spring-migrating Anatidae: a State University, Columbus, USA. review. Journal of Ornithology 147: 167–184. Callicutt, J.T., Hagy, H.M. & Schummer, M.L. Arzel, C., Elmberg, J. & Guillemain, M. 2007. A 2011. The food preference paradigm: a flyway perspective of foraging activity in review of autumn-winter food use by North Eurasian Green-winged Teal, Anas crecca American dabbling ducks (1900–2009). Journal crecca. Canadian Journal of Zoology 85: 81–91. of Fish and Wildlife Management 2: 29–40.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 82 Waterfowl conservation during spring migration

Davies, J.C. & Cooke, F. 1983. Annual nesting those marked in Lancashire. Bird Study 41: productivity in snow geese: prairie droughts 221–234. and arctic springs. Journal of Wildlife Fox, A.D., Kahlert, J. & Ettrup, H. 1998. Diet and Management 47: 291–296. habitat use of moulting Greylag Geese Anser Davis, J.B., Guillemain, M., Kaminski, R.M., anser on the Danish island of Saltholm. Ibis Arzel, C., Eadie, J.M. & Rees, E.C. 2014. 140: 676–683. Habitat and resource use by waterfowl in the Fox, A.D., Glahder, C.M. & Walsh, A.J. 2003. northern hemisphere in autumn and winter. Spring migration routes and timing of Wildfowl (Special Issue No. 4): 17–69. Greenland white-fronted geese – results Drent R.H., Ebbinge B.S. & Weijand B. 1978. from satellite telemetry. Oikos 103: 415–425. Balancing the energy budgets of arctic- Fox, A.D., Francis, I.S. & Bergersen, E. 2006. breeding geese throughout the annual Diet and habitat use of Svalbard Pink-footed cycle: a progress report. Verhandlungen der Geese Anser brachyrhynchus during arrival and Ornithologischen Gesellschaft Bayern 23: 239–264. pre-breeding periods in Adventdalen. Ardea Drent, R.H. & Daan, S. 1980. The prudent 94: 691–699. parent: energetic adjustments in avian Fox, A.D., Boyd, H., Walsh, A.J., Stroud, D.A., breeding? Ardea 68: 225–252. Nyeland, J. & Cromie, R. 2012. Earlier spring Drent, R., Both, C., Green, M., Madsen, J. & staging in Iceland amongst Greenland White- Piersma, T. 2003. Pay-offs and penalties of fronted Geese Anser albifrons flavirostris competing migratory schedules. Oikos 103: achieved without cost to refuelling rates. 274–292. Hydrobiologia 697: 103–110. Féret, M., Gauthier, G., Béchet, A., Giroux, J.F. & Fox, A.D., Weegman, M.D., Bearhop, S. Hilton, Hobson, K.A. 2003. Effect of a spring hunt G., Griffin, L., Stroud, D.A. & Walsh, A.J. on nutrient storage by greater snow geese in 2014. Climate change and contrasting southern Quebec. Journal of Wildlife plasticity in timing of passage in a two-step Management 67: 796–807. migration episode of an arctic-nesting avian Foster, M.A., Gray, M.J. & Kaminski, R.M. 2010. herbivore. Current Zoology 00: 000–000. Agricultural seed biomass for migrating and Fredrickson, L.H. & Drobney, R.D. 1977. Habitat wintering waterfowl in the Southeastern utilization by postbreeding waterfowl. In United States. Journal of Wildlife Management T.A. Bookhout (ed.), Proceedings 1977 74: 489–495. Symposium, Madison, Wisconsin, pp. 119– Fox, A.D. & Walsh, A.J. 2012. Warming winter 131. The Wildlife Society, Madison, Wisconsin, effects, fat store accumulation and timing of USA. spring departure of Greenland White- Fretwell, S.D. 1972. Populations in a Seasonal fronted Geese Anser albifrons flavirostris from Environment. Princeton University Press, New their winter quarters. Hydrobiologia 697: Jersey, USA. 97–102. Gauthier, G., Bêty, J. & Hobson, K.A. 2003. Are Fox, A.D., Mitchell, C., Stewart, A., Fletcher, J.D., Greater Snow Geese capital breeders? New Turner, J.V.N., Boyd, H., Shimmings, P, evidence from a stable isotope model. Ecology Salmon, D.G., Haines, W.G. & Tomlinson, C. 84: 3250–3264. 1994. Winter movements and site-fidelity Greer, D.M., Dugger, B.D., Reinecke, K.J. & of Pink-footed Geese Anser brachyrhynchus Petrie, M.J. 2009. Depletion of rice as food ringed in Britain, with particular emphasis on of waterfowl wintering in the Mississippi

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 83

Alluvial Valley. Journal of Wildlife Management Herrmann, K.K. & Sorensen, R.E. 2009. 73: 1125–1133. Seasonal dynamics of two mortality-related Guglielmo, C.G., Piersma, T. & Williams, T.D. trematodes using an introduced snail. Journal 2001. A sport-physiological perspective on of Parasitology 95: 823–828. bird migration: evidence for flight induced Hohman, W.L., Manley, S.W. & Richard, D. 1997. muscle damage. Journal of Experimental Biology Relative costs of prebasic and prealternate 204: 2683–2690. molts for male Blue-winged Teal. Condor 99: Guillemain, M., Fritz, H., Klaassen, M., Johnson, 543–548. A.R. & Hafner, H. 2004. Fuelling rates of Jefferies, R.L., Rockwell, R.F. & Abraham, K.F. garganey (Anas querquedula) staging in the 2004. Agricultural food subsidies, migratory Camargue, southern France, during spring connectivity and large-scale disturbance in migration. Journal of Ornithology 145: 152–158. Arctic coastal systems: a case study. Integrative Guillemain, M., Arzel, C., Legagneux, P., Elmberg, and Comparative Biology 44: 130–139. J., Fritz, H., Lepley, M., Pin, C., Arnaud, A. & Jenni, L. & Jenni-Eiermann, S. 1998. Fuel supply Massez, G. 2007. Predation risk constrains the and metabolic constraints in migrating birds. plasticity of foraging behaviour in teals, Anas Journal of Avian Biology 29: 521–528. crecca: a flyway-level circumannual approach. Johnson, D.H. 1980. The comparison of usage Animal Behaviour 73: 845–854. and availability measurements for evaluating Guillemain, M., Elmberg, J., Arzel, C., Johnson, resource preference. Ecology 61: 65–71. A.R. & Simon, G. 2008. The income-capital Johnson, D.H. & Grier, J.W. 1988. Determinants breeding dichotomy revisited: late winter of breeding distributions of ducks. Wildlife body condition is related to breeding success Monographs 100: 1–37. in an income breeder. Ibis 150: 172–176. Kaminski, R.M. & Elmberg, J. 2014. An Hagy, H.M. & Kaminski, R.M. 2012. Apparent introduction to habitat use and selection by seed use by ducks in moist-soil wetlands of waterfowl in the northern hemisphere. the Mississippi Alluvial Valley. Journal of Wildfowl (Special Issue No. 4): 9–16. Wildlife Management 76: 1053–1061. Kaminski, R.M. & Gluesing, E.A. 1987. Density- Harrison, X.A., Blount, J.D., Inger, R., Norris, and habitat-related recruitment in Mallards. D.R. & Bearhop, S. 2011. Carry-over effects Journal of Wildlife Management 51: 141–148. as drivers of fitness differences in animals. Klaassen, M., Abraham, K.F., Jefferies, R.L. & Journal of Animal Ecology 80: 4–18. Vrtiska, M. 2006. Factors affecting the site of Havera, S.P., Boens, L.R., Georgi, M.M. & Shealy, investment, and the reliance on savings for R.T. 1992. Human disturbance of waterfowl arctic breeders: the capital-income dichotomy on Keokuk Pool, Mississippi River. Wildlife revisited. Ardea 94: 371–384. Society Bulletin 20: 290–298. Krapu, G.L., Facey, D.E., Fritzell, E.K. & Heitmeyer, M.E. 1988. Protein costs of the Johnson, D.H. 1984. Habitat use by migrant prebasic molt of female mallards. Condor 90: sandhill cranes in Nebraska. Journal of Wildlife 263–266. Management 48: 407–417. Heitmeyer, M.E. & L.H. Fredrickson. 1981. Do Krapu, G.L., Reinecke, K.J., Jorde, D.G. & wetland conditions in the Mississippi Delta Simpson, S.G. 1995. Spring-staging ecology hardwoods influence mallard recruitment? of midcontinent greater white-fronted Trans. North Am. Wildl. Nat. Resour. Conf. 46: geese. Journal of Wildlife Management 59: 736– 44–57. 746.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 84 Waterfowl conservation during spring migration

Lagrange, T.G. & Dinsmore, J.J. 1989. Habitat Pearse, A.T., Krapu, G.L. & Cox, R.R., Jr. use by mallards during spring migration 2012. Spring snow goose hunting influences through central Iowa. Journal of Wildlife body composition of waterfowl staging in Management 53: 1076–1081. Nebraska. Journal of Wildlife Management 76: LaMontagne, J.M., Barclay, R.M.R. & Jackson, 1393–1400. L.J. 2001. Trumpeter swan behaviour at Piersma, T. 2002. Energetic bottlenecks and other spring-migration stopover areas in southern design constraints in avian annual cycles. Alberta. Canadian Journal of Zoology 79: Integrative and Comparative Biology 42: 51–67. 2036–2042. Raveling, D.G. & Heitmeyer, M.E. 1989. Lindeman, D.H. & Clark, R.G. 1999. Amphipods, Relationships of population size and land-use impacts, and lesser scaup (Aythya recruitment of pintails to habitat conditions affinis) distribution in Saskatchewan wetlands. and harvest. Journal of Wildlife Management 53: Wetlands 19: 627–638. 1088–1103. Loesch, C.R. & Kaminski, R.M. 1989. Winter- Reinecke, K.J. & Krapu, G.L. 1986. Feeding body weight patterns of female Mallards ecology of sandhill cranes during spring fed agricultural seeds. Journal of Wildlife migration in Nebraska. Journal of Wildlife Management 53: 1081–1088. Management 50: 71–79. Lovvorn, J.R. & Barzen, J.A. 1988. Molt in the Reynolds, R.E., Shaffer, T.L., Renner, R.W., annual cycle of Canvasbacks. Auk 105: 543– Newton, W.E. & Batt, B.D.J. 2001. Impact 552. of the conservation reserve program on Madsen, J. 1995. Impacts of disturbance on duck recruitment in the US Prairie Pothole migratory waterfowl. Ibis 137: S67–S74. Region. Journal of Wildlife Management 65: Mini, A.E. & Black, J.M. 2009. Expensive 765–780. traditions: energy expenditure of Aleutian Richardson, D.M. & Kaminski, R.M. 1992. Diet geese in traditional and recently colonized restriction, diet quality, and prebasic molt in habitats. Journal of Wildlife Management 73: female mallards. Journal of Wildlife Management 385–391. 56: 531–539. Newton, I. 2006. Can conditions experienced Sedinger, J.S. & Alisauskas, R.T. 2014. Cross- during migration limit the population levels seasonal effects and the dynamics of in birds? Journal of Ornithology 147: 146– waterfowl populations. Wildfowl (Special 166. Issue No. 4): 277–304. Newton, I. 2007. Weather related mass-mortality Sherfy, M.H., Anteau, M.J. & Bishop, A.A. 2011. events in migrants. Ibis 149: 453–467. Agricultural practices and residual corn Pearse, A.T., Krapu, G.L., Brandt, D.A. & Kinzel, during spring crane and waterfowl migration P.J. 2010. Changes in agriculture and in Nebraska. Journal of Wildlife Management 75: abundance of snow geese affect carrying 995–1003. capacity of sandhill cranes in Nebraska. Soulliere, G.J., Potter, B.A., Coluccy, J.M., Journal of Wildlife Management 74: 479–488. Gatti, R.C., Roy, C.L., Luukkonen, D.R., Pearse, A.T., Alisauskas, R.T., Krapu, G.L. & Brown, P.W. & Eichholz, M.W. 2007. Cox, R.R., Jr. 2011. Changes in nutrient Upper Mississippi River and Great Lakes dynamics of midcontinent greater white- Region Joint Venture Waterfowl Habitat fronted geese during spring migration. Journal Conservation Strategy. U.S. Fish and Wildlife of Wildlife Management 75: 1716–1723. Service, Fort Snelling, Minnesota, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 Waterfowl conservation during spring migration 85

Stafford, J.D., Kaminski, R.M., Reinecke, K.J. & peaks of temperature acceleration during Manley, S.W. 2006. Waste rice for waterfowl spring migration. Oikos 121: 655–664. in the Mississippi Alluvial Valley. Journal of Vest, J.L., Kaminski, R.M., Afton, A.D. & Vilella, Wildlife Management 70: 61–69. F.J. 2006. Body mass of lesser scaup during Straub, J.N., Gates, R.J., Schultheis, R.D., Yerkes, fall and winter in the Mississippi Flyway. T., Coluccy, J.M. & Stafford, J.D. 2012. Journal of Wildlife Management 70: 1789–1795. Wetland food resources for spring-migrating Vrtiska, M.P. & Sullivan, S. 2009. Abundance and ducks in the Upper Mississippi River and distribution of lesser snow and Ross’s geese Great Lakes Region. Journal of Wildlife in the Rainwater Basin and central Platte Management 76: 768–777. River Valley of Nebraska. Great Plains Tombre, I.M., Høgda, K.A., Madsen, J., Griffin, Research 19: 147–155. L.R., Kuijken, E., Shimmings, P., Rees, E. & Webb, E.B., Smith, L.M., Vrtiska, M.P. & Verscheure, C. 2008. The onset of spring and LaGrange, T.G. 2010. Effects of local and timing of migration in two arctic nesting landscape variables on wetland bird habitat goose populations: the pink-footed goose use during migration through the Rainwater Anser brachyrhynchus and the barnacle goose Basin. Journal of Wildlife Management 74: 109– Branta leucopsis. Journal of Avian Biology 39: 119. 691–703. Weller, M.W. 1975. Migratory waterfowl: a Trautman, M.B., Bills, W.E. & Wickliff, E.L. hemispheric perspective. Publicaciones Biologicas 1939. Winter losses from starvation and Instituto de Investigaciones Cientificas, U.A.N.L. 1: exposure of waterfowl and upland game 89–130. birds in Ohio and other Northern States. The Wilson, D.M. Naimo, T.J., Wiener, J.G., Anderson, Wilson Bulletin 51: 86–104. R.V., Sandheirich, M.B. & Sparks, R.E. 1995. van der Graaf A.J., Stahl, J., Klimkowska, A., Declining populations of fingernail clam Bakker, J.P. & Drent, R.H. 2006. Surfing on a Musculium transversum in the upper Mississippi green wave – how plant growth drives spring River. Hydrobiologia 304: 209–220. migration in the Barnacle Goose Branta Zimmer, C., Boos, M., Petit, O. & Robin, J.P. leucopsis. Ardea 94: 567–577. 2010. Body mass variations in disturbed van Wijk, R.E., Kölzsch, A., Kruckenberg, H., mallards Anas platyrhynchos fit to the mass- Ebbinge, B.S., Müskens, G.J.D.M. & Nolet, dependent starvation-predation risk trade- B.A. 2012. Individually tracked geese follow off. Journal of Avian Biology 41: 637–644.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 70–85 86

Nest site selection by Holarctic waterfowl: a multi-level review MICHAEL W. EICHHOLZ1* & JOHAN ELMBERG2

1Mailcode 6504, Cooperative Wildlife Research Laboratory, Center for Ecology, Department of Zoology, Southern Illinois University Carbondale, Carbondale, Illinois, USA. 2Division of Natural Sciences, Kristianstad University, SE-291 88, Kristianstad, Sweden. *Correspondence author. E-mail: [email protected] Abstract Because of birds’ mobility, behaviour and many species’ migratory nature, they select repeatedly and spatially among habitats and have been central figures in studies of avian breeding habitat selection during the 20th and 21st centuries. The scientific literature on habitat use by breeding waterfowl has origins dating back to the writings of Charles Darwin in The Voyage of the Beagle, wherein he described the distribution and habitat differences of two species of geese on the Falkland Islands. Since that time, waterfowl ecologists have gone from descriptive studies of nest site characteristics used for planning waterfowl conservation and management to comparing nest site use in relation to potential habitat availability and determining selection for a wide array of ecological correlates. Waterfowl ecologists most recently have been investigating the adaptive significance of nest site selection by associating the latter with individual fitness and demographic measurements to assess the birds’ adaptability under environmental conditions at multiple scales of selection. While little direct assessment of 1st and 2nd order nest site selection has occurred (sensu Johnson 1980), available information is most consistent with the hypothesis that selection at these scales is driven by food availability. At the 3rd and 4th order of selection, data are consistent with hypotheses that both food availability and predator avoidance drive nest site selection, depending on the species and type of nesting aggregation. We also identify understudied areas of nest site selection important for the conservation and management of waterfowl and suggest that the large-scale influence of current anthropogenic and natural effects on the environment indicates that greater emphasis should be directed toward understanding waterfowl nest site selection at the 1st and 2nd orders of selection and how nesting habitat selection interfaces with community ecology of sympatric breeding waterfowl. Moreover, because habitat selection of pre- fledging waterfowl is inherently linked to breeding habitat selection, we suggest an updated review of brood habitat selection should ensue from our synthesis here.

Key words: hierarchal habitat selection, nest, nesting habitat, nest site selection, waterfowl.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 87

Elucidation of use and selection of habitat (Leopold 1933), only sporadic reports of by animals for breeding and other life- waterfowl nest sites can be found in the history segments during their annual cycle is scientific literature until a synthesised essential for answering basic ecological and description of the nesting habitats for applied questions at the individual, waterfowl in North America was published population and community levels (Fretwell by Bent (1923). Although Bent reviewed the & Lucas 1969; Lack 1971; Cody 1981). available information, data were lacking on Because of birds’ mobility and many species-specific use of nesting habitats. This species’ migratory nature, they select lack of understanding was also apparent repeatedly and spatially among habitats and in Pirnie’s (1935) book on waterfowl have been central to studies of breeding ecology and management in Michigan. He habitat selection during the 20th and 21st recommended planting junipers Juniperus sp. centuries (Grinnell 1914; Lack 1933; Hildén and pine Pinus sp., or other evergreens and 1965; Cody 1985; Block & Brennan 1993). other shrubby vegetation, as nesting cover for The scientific literature on habitat use by upland game birds including Mallard Anas breeding waterfowl has origins dating back to platyrhynchos, and other ducks nesting in the writings of Charles Darwin in The uplands in Michigan. He later noted, however, Voyage of the Beagle wherein he described an abundance of successfully nesting ducks in distribution and habitat differences of the flooded meadows that provided dense “Upland Goose” Anas magellanica and the herbaceous cover. Likely in response to the “Rock Goose” Anas antarctica, although it decreasing abundance of waterfowl in North appears he was describing what are now America during the drought stricken 1930s considered subspecies of upland geese and early 1940s, numerous descriptions of Chloephaga picta leucoptera and C. p. picta on the waterfowl nesting habitat were published in Falkland Islands. Oberholser & McAtee the mid 20th century (e.g. Bennet 1938; Low (1920) reported on population declines in 1941; Gross 1945; Leitch 1951). Indeed, waterfowl in North America, exhorted the recognition became prevalent in the 1950s benefits of eliminating spring harvest in that factors including habitat type, social North America for its resultant increase avoidance and attraction, and predator in waterfowl populations. Displaying avoidance may influence species-specific nest considerable insight for their time, they site selection (e.g. Earl 1950; Kossack 1950; recognised the importance of habitat Glover 1956). As biologists recognised that by suggesting that agricultural land nesting habitat use varied by species and even development was negatively affecting duck within species, and that habitat choices populations through decreased availability influenced reproductive success, considerable of crucial wetlands and other habitats. effort was spent during the 1960s–1980s Although early conservationists were describing species-specific variation in nesting beginning to recognise the importance of habitat and how such variation was influenced preserving breeding habitat for the by predation pressure (e.g. Keith 1961; conservation of wildlife populations Duebbert & Lokemoen 1976; Weller 1979;

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 88 Nest site selection by waterfowl in the northern hemisphere

Livezey 1981; McLandress 1983; Kaminski & have recognised these distinctions and Weller 1992). These studies emphasised the accordingly often tend to focus their effort applied aspects of the data in identifying toward managing habitat that influences habitat use patterns for conservation and reproductive success for dabbling and management purposes, but rarely used the diving ducks while managing non-breeding data specifically to advance understanding of habitat and harvest for sea ducks, geese and causative drivers of waterfowl habitat swans. selection (cf. Kaminski & Prince 1984; Cody Thus, because reproductive success has 1985) despite the fact that the early habitat strong influence on population dynamics, selection models were developed and widely understanding breeding habitat use tested at that time (Fretwell & Lucas and selection is fundamental for wise 1969; Fretwell 1972). Furthermore, because decisions regarding habitat conservation dabbling and diving ducks tend to breed and management. This recognition has led in temperate, lower latitude regions than to numerous detailed studies of how arctic nesting swans, geese and sea ducks, management actions influence habitat their habitat has undergone increased selection and nest success of boreal and anthropogenic modifications. However, even temperate nesting dabbling and diving the more northern boreal regions that ducks. Additionally, possibly because of were once relatively immune to human influential writings by Romseburg (1981) developments are now influenced by both and Walters (1985), waterfowl biologists direct (forest harvest, mining, extraction of have begun to view breeding habitat fossil fuels and wind power development) and selection in basic and applied ecological indirect (atmospheric warming, airborne contexts (e.g. Nudds 1983; Clark & Shutler pollution and eutrophication) anthropogenic 1999; Fast et al. 2007). This change forces. Indeed, the rate of anthropogenic coincided with recognition that spatial scale modification has recently increased is an important factor in the habitat dramatically in some boreal regions in the selection process (Johnson 1980; Hutto northern hemisphere (Murphy & Romanuk 1985; Wiens 1989). Thus, issues of spatial 2014). and temporal scale have been influential in In general, factors that influence framing and articulating theories on which waterfowl reproduction are thought to have research about breeding habitat selection in the greatest influence on populations of waterfowl is based (Johnson 1980; Hutto generally r-selected species (dabbling and 1985; Forbes & Kaiser 1994). Selection of diving ducks; Flint et al. 1998; Hoekman et al. habitat by waterfowl during breeding and 2002; Coluccy et al. 2008), while factors that other annual cycle events may be viewed as influence post-fledging survival tend to have a hierarchal process. For example, selection a greater influence on population dynamics of nest sites is a fine-grained hierarchical of more K-selected species (sea ducks, process with individuals utilising large-grain geese, and swans; Nichols et al. 1976; Cooch characteristics for initial (first order) et al. 2001; Schamber et al. 2009). Managers selections of geographic regions, then using

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 89 characteristics that can be distinguished at questions associated with diversity, more reduced grain size as the selection distribution, and abundance of organisms at process continues (Johnson 1980; Wiens the continental scale. Brown & Maurer 1989; Kristan 2006). Grain size for nesting (1989) coined the term “macrocecolgical waterfowl generally will be smaller than approach” to describe ecological approaches other habitat components of the annual of this scale. They stated, “Our goal is to cycle, except perhaps individual foraging understand the assembly of continental and rest sites, because nests occupy biotas in terms of how the physical space microhabitats and their location is such an and nutritional resources of large areas are important aspect of reproductive success divided among diverse species.” Two areas of (O’Neila et al. 1986; Wiens 1989). study that often use a macroecological In this contribution, we review the approach to address associations among characteristics waterfowl use to make space and resource availability and species decisions at various spatial levels leading distribution, abundance and diversity are downward to nest site selection, the selective migration and community ecology; thus, pressures that affect those characteristics, these should provide insight into 1st order and the resulting fitness of those decisions. nest site selection of waterfowl are migration We review how research on waterfowl and community ecology (e.g. Davis et al. nest site selection has influenced our 2014). understanding of basic ecological theory Most waterfowl species occurring in the and how this information is used when Holarctic are long- to medium-distance making management decisions. To facilitate migrants. A general pattern among these comparison of cross taxa variation, this waterfowl is they nest in temperate, sub- paper is partitioned into four scales of arctic and arctic regions too inhospitable to selection: 1st order – general region or support them during the non-breeding latitude, 2nd order – landscape type (biome) period, and then spend the non-breeding within a region, 3rd order – location, period in locations with more moderate wetland, or upland within a landscape and climates (Bellrose 1980; Kaminski & Weller 4th order – specific nest site within the 1992). Thus, if we assume pursuit of nesting location (Kaminski & Elmberg 2014). By habitat begins when birds begin to transition organising the review in this manner, we from non-breeding to breeding periods, the hope to articulate benefits of life-history initial consideration for 1st order selection is comparisons. We conclude by identifying really a question of migration (i.e. Does areas where information is inadequate to an individual breed at the same location it answer basic ecological questions and make spent the non-breeding period or move reliable conservation decisions. to a different location?). Considerable debate exists whether migration evolved First-order selection from northern and temperate breeding Habitat selection at this scale is best populations migrating south during the non- addressed by considering studies asking breeding season as winter ensued or from

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 90 Nest site selection by waterfowl in the northern hemisphere sub- and tropical breeding locations as including grassland, temperate rain forest, species expanded their range into more tundra, taiga, eastern deciduous forest and seasonal, higher latitudes after Pleistocene desert when adequate water is present glaciation (Cox 1968; Chesser & Levey 1998; (Bellrose 1980; NAWMP 2012). The Alerstam et al. 2003). Regardless of origin northern range limit for species is most and mechanistic stimuli of migration, the likely driven by the length of the ice-free Cordillerin and Laurentide ice sheets in period being too limited to provide adequate North America and the Scandinavian ice time for reproduction (Schmidt et al. 2011), sheet in Europe during the most recent while the factors that influence the southern glaciation (from c. 120,000–8,500 years ago) range limits are generally undefined for generally limited the distribution of Holarctic waterfowl. waterfowl to regions that serve primarily as As the unique association between wintering areas under current climatic species richness and latitude in Holarctic conditions (e.g. Hawkins & Porter 2003; waterfowl demonstrates, 1st order selection Hortal et al. 2011). Thus, most species of for Holarctic waterfowl seems a question Holarctic waterfowl since glaciation have of migratory behaviour, because it can evolved strategies to migrate north to breed. influence species distribution which can A number of hypotheses have been drive spatial variation in species richness. proposed to explain northern latitude Thus, an additional ecological concept to breeding, including migration as a method gain insight into the selective forces of 1st to exploit rich seasonal food resources, order selection and the potential driving avoid predation, reduce exposure to disease force behind the southern range limit and parasites, exploit latitudes with long day in Holarctic waterfowl is the concept lengths during the growing season, or of community ecology. Nudds (1992) reduce intra- or interspecific competition thoroughly reviewed theories on species (Cox 1985; Fretwell 1980; Alerstam & richness in waterfowl communities. In Högstedt 1982; Piersma 1997; Chesser & general, the discussion has changed little Levey 1998; Rappole & Jones 2002). since that review. Species richness is The tendency for Holarctic waterfowl to dependent on three processes: speciation, migrate from southern wintering regions to extinction and dispersal. The speciation and more northern breeding regions has created extinction process influence variation in a somewhat unique pattern of species richness among clades over evolutionary richness. While species richness of most time and potentially large scale (continental) organisms, including most bird taxa, spatial variation, while the dispersal process decreases with increasing latitude, species likely has the greatest influence in ecological richness of Holarctic waterfowl tends to time and on more local (regions and less) peak between 40°–65° latitude, declining variation for mobile organisms such as rapidly north and south of that range (Dalby waterfowl (Hulbert & Stegen 2014). Because et al. 2014). Within this range of latitudes, of the close association between species waterfowl inhabit all the major biomes distribution, migration and species richness,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 91 some of the same selective forces posited to Total habitat area and heterogeneity also influence migratory behaviour (i.e. disease, have been found to be a strong predictor predation pressure, resource availability and and causative agent of species richness for intra- and interspecific competition) as well some taxa (Roth 1976; Elmberg et al. 1993). as other factors such as the species-area If total wetland area or wetland relationship, habitat heterogeneity and time heterogeneity were driving latitudinal for or rate of diversification have been variation in waterfowl species distribution, proposed as primary mechanisms for spatial we would predict greater wetland area and variation in species richness (Arrhenius heterogeneity in more northern latitudes. 1921; Willson 1976; Wright 1983; Evans et We were unable to locate latitudinal data al. 2005). While none of these mechanisms reflecting wetland heterogeneity, but the have been excluded, the roles of time, wetland trends data provide an estimate of habitat heterogeneity and resource total wetland acreage for each of the United availability have been accepted as the most States (Dahl 2011). If wetland area is driving likely mechanisms. the latitudinal variation in waterfowl species A commonly proposed mechanism to richness, we predict an increase in wetland explain the decline in species richness with area with latitude. When we regressed total increasing latitude observed in most groups wetland acreage for each state in the United of organisms other than waterfowl is that States against its geographic midpoint, we the former have had a longer period of time found no relationship between latitude and to diversify, diversify at a faster rate and have total wetland acreage for all states (r2 = 0.18, greater niche conservatism in the tropics (i.e. P = 0.24) or those states between 80–105° long-term stability of the environment longitude where most nesting waterfowl conserving the niche); thus, the tropics occur (r2 = 0.15, P = 0.29). Furthermore, if support greater species richness (Brown we assume the relationship between wetland 2014). Using similar logic, more recent large heterogeneity and wetland area at regional scale climatic factors such as glaciation have scales is similar to the relationship at more also been proposed to explain patterns of local scales (Elmberg et al. 1993, 1994), there species richness at the continental scale should be a strong correlation between total within the Holarctic (Hawkins & Porter wetland abundance and wetland diversity. 2003; Hortal et al. 2011). If time were the Thus, if wetland heterogeneity is driving primary mechanism driving the observed the relationship between species richness relationship between Holarctic waterfowl and latitude, we should again detect a species richness and latitude, we would positive relationship between latitude and predict greater species richness in the more overall wetland acreage, a relationship that southern latitudes, given that vast areas of contradicts our observation. the more northern latitudes were glaciated as The final explanation of latitudinal recently as 8,000 years ago. This prediction is patterns in waterfowl species richness is opposite of the present pattern, indicating resource availability. Resource availability time is an unlikely explanation. has been found to be strongly correlated to

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 92 Nest site selection by waterfowl in the northern hemisphere species richness at multiple scales for a leafy vegetation, seeds, tubers and agricultural variety of taxa. Latitudinal variation in per seeds produced during the previous growing capita resource availability could be due to season, as well as aquatic invertebrates. The reduced intra- or inter-specific competition high concentration of waterfowl exploiting in more northern latitudes (Ashmole 1963), these resources likely reduces their availability more wetlands (or other requisite habitat) in to waterfowl as winter progresses and more northern latitudes, or higher transitions to spring (Davis et al. 2014). Thus, productivity in more northern latitudes. it’s possible the exploitation of resources by Ashmole (1963) argued that because wintering waterfowl reduces the resources to the inhospitable winter climate of more migrating and breeding waterfowl, reducing seasonable environments limited the the species richness and abundance in more number of birds breeding in temperate southern latitudes. regions, there was greater per capita Alternatively, a similar amount of resource availability. This argument does not wetlands in more northern latitudes could appear to hold true for waterfowl, however, provide more resources if vegetation in in that both abundance and richness peak at more northern regions provides greater higher latitudes. benefit to secondary consumers (Coley et al. In recent large scale studies, resource 1985; Moles et al. 2011; Morrison & Hay availability seems to be the best predictor of 2012). Vegetation with greater nutritional species richness for the majority of taxa value would directly influence resource (Hawkins et al. 2003; Hulbert & Haskell availability for geese which are herbivorous 2003; Evans et al. 2005). As the analysis throughout their annual cycle and produce described above indicates, relative surface young that require nutrient rich plants area of wetlands, assuming a correlation during early post-hatch growth (Coley et al. between surface area and heterogeneity, 1985; Sedinger 1992). Nutritional quality of wetland heterogeneity does not appear to vegetation could also influence ducks that increase with latitude in North America but are primarily carnivorous during the our analyses currently are restricted to this breeding season by providing substrates continent. Thus, if resource availability is and food for aquatic invertebrates that the driving mechanism for latitudinal ducks consume (Krapu & Reinecke 1992). variation in waterfowl species richness, a Because most vegetation in higher latitude latitudinal trend in per area resource wetlands continues to grow until freezing availability may be the driving mechanism. temperatures cause senescence, annual rate Resource availability per area may vary due of plant decomposition is much lower to greater exploitation by waterfowl in more in more northern wetlands (Webster & southern wetlands or greater per area Benfield 1986; Magee 1993; Holt 2008). productivity in higher latitude wetlands. Thus, as opposed to more southern During winter, waterfowl congregate in wetlands where plant decomposition and regions that provide predictable food nutrient turnover continues throughout the availability (i.e. remain unfrozen) exploiting winter with little remaining by spring, there

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 93 is considerable organic material and a Although other mechanisms for large food base for invertebrates when explaining the more northern latitudinal spring arrives at more northern latitudes. peak in species richness of waterfowl cannot Moreover, seeds and invertebrates are not be excluded, per capita nutrient availability exposed to predation by birds during the appears to have the greatest level of support winter, providing returning breeders with a (Dalby et al. 2014). While at a more regional less depleted food base than is the case in scale, waterfowl richness appears to increase more southern latitudes where wetlands with both nutrient availability and habitat are used year round by resident and heterogeneity (2nd order level selection; seasonally occurring species of waterbirds. Elmberg et al. 1993). Although the nutrient Additionally, a number of studies have availability hypothesis appears to be most demonstrated that invertebrates prefer to consistent with data currently available, forage on plant material in more northern direct tests with empirical data have not been latitudes (e.g. Pennings et al. 2001, 2007). conducted. For example, no one has tested Some have suggested this strategy is due to for general latitudinal variation in aquatic fewer chemical defences from plants from invertebrate biomass (cf. Arzel et al. 2009 more northern latitudes while others have for a one-species all-flyway example), and suggested it is due to higher concentrations although evidence exists indicating of nitrogen (Coley et al. 1985; Moles et al. vegetation from more northern areas may 2011; Morrison & Hay 2012). Regardless be preferred by invertebrates over the of the mechanism, vegetation from same species of vegetation from more more northern latitudes appears to be of southern latitudes, studies have not isolated greater nutritional value to invertebrates confounding effects of chemical or physical and herbivorous waterfowl, potentially defence and concentration of nutrients for producing greater nutritional resources. the more northern vegetation. Finally, as Willson (1976) suggested the shorter growing season in more northern Second-order selection latitudes may itself lead to an increase in Often waterfowl have multiple options for standing biomass of aquatic invertebrates. choice of biome (e.g. grassland, forest, The briefer period for reproduction may tundra) to select for nest sites even after cause more species of invertebrates to selecting a geographic latitude in which to reproduce simultaneously, leading to a breed. Unlike latitudinal variation in species higher spike in overall invertebrate biomass richness, there appears no clear pattern in in more seasonal northern environments. species richness across various northern High-latitude wetlands may thus offer a high hemispheric biomes. In fact, although some standing biomass and higher per capita food species appear to be habitat specialists in resource availability when waterfowl nest as that most individuals are found almost a result of higher productivity, concurrent exclusively in one biome (e.g. Blue-winged food peaks or a long period without food Teal Anas discors in grassland and Green- depletion (Danell & Sjöberg 1977). winged Teal/Eurasian Teal A. crecca in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 94 Nest site selection by waterfowl in the northern hemisphere boreal environs), most species appear to be ranges overlap between Blue- and Green- habitat generalists, selecting specific nest site winged Teal or Gadwall A. strepera and characteristics regardless of the biome and American Wigeon A. americana, species that thus can be found in multiple biomes (e.g. breed in different biomes but appear closely Nicolai et al. 2005; Safine & Lindberg 2007). related genetically and ecologically, is due to Furthermore, with the exception of cavity variation in habitat requirements or nesting species, which largely require competition may help elucidate questions forested habitat, or sea ducks that specialise associated with 2nd order selection. in coastal waters for brood rearing, almost all species can be found nesting in multiple Third-order selection biomes. Even species like the Common Third-order selection is the level at which Eider Somateria mollissima, which depend on they select a specific local habitat(s) within a marine environments for foraging and are biome. A number of nest site selection considered a tundra nesting species in North characteristics discussed here also could be America, will nest in island forests when considered 4th order characteristics. We such forests are in close proximity to brood consider them 3rd order characteristics for habitat (Öst et al. 2008a). the sake of this discussion due to the grain Species that demonstrate a clear size at which the characteristic may have been preference also have likely adapted to measured. For example, if a characteristic specialise on certain characteristics of was measured at a scale that was relevant to habitat (Mulhern et al. 1985). The two more than one female (i.e. a field or patch of selection factors that appear to be acting trees), we considered it 3rd order selection; most at the scales of 3rd and 4th order whereas, if a characteristic was measured at a selection are food availability and predator scale relevant to one female (e.g. density of avoidance, making them good candidates vegetation surrounding a specific prospective for proximate cues that may drive 2nd order nest or size in the opening of a tree cavity), selection. While nest predation pressure we considered it 4th order selection. In the potentially varies with latitude (Hanski et al. following sections, we review potential 1991; Elmberg et al. 2009), it doesn’t appear ecological, environmental or social influences to vary among biomes so selection of biome of 3rd order selection of nest habitats. may be dictated by food availability (Grand & Flint 1997; Fournier & Hines 2001; Predator and coexisting prey densities Walker et al. 2005; Schamber et al. 2009; cf. Nest predation is the primary cause of nest Elmberg et al. 2009). Currently, increasing failure in waterfowl (Sargeant & Raveling exploitation of natural resources of the 1992; Stephens et al. 2005). Egg predators tundra and boreal regions emphasises the are distributed heterogeneously across need to better understand the requirements landscapes concentrating in habitats that of habitat specialists for future management provide efficient foraging (Kuehl & Clark and conservation. Studies addressing the 2002; Phillips et al. 2003; Elmberg & question of whether the limited breeding Gunnarsson 2007; Klug et al. 2009) and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 95 protection from higher trophic level Researchers have long assumed nest site predators (Crabtree et al. 1989; Dion et al. selection was influenced by predator 2000; Chalfoun et al. 2002). High nest avoidance, but only recently has there been predation and heterogeneous distribution of evidence that birds could assess local predators invoke selective pressure for predator density and modify their behaviour waterfowl to adapt strategies of selecting accordingly. Fontaine & Martin (2006) found nest habitats and sites with fewer predators numerous species of passerines modify their or more coexisting prey than other potential reproductive investment by increasing their sites, leading to a reduction in nest predation feeding behaviour when predator abundance (Holt 1977; Ackerman 2002; Eichholz et al. was reduced, but provided no explanation as 2012). to the mechanism parents used to assess Predator avoidance appears to be the predator density. Similarly, Dassow et al. primary selective force for colonial nesting (2012) found evidence that ground nesting species, apparently having less impact on ducks modify reproductive investment based dispersed nesting species (Schmutz et al. on density of predators (cf. Duebbert & 1983; Bousfield & Syroechkovskiy 1985; Kantrud 1974). A number of studies have Fox et al. 2009). This inconsistency may be now provided evidence that birds are able to related to the level of feeding that occurs by use various mechanisms to assess predator females during incubation and the distance abundance and modify reproductive young can travel after hatch. Most colonial strategies (e.g. Lima 2009; Eichholz et al. nesting species of waterfowl, with the 2012; Forsman et al. 2012). Additionally, possible exception of Black Brant Branta researchers conducting predator exclusion bernicla nigricans and Ross’s Goose Chen rossii, and reduction studies have observed feed little, if any, during incubation and increases in nesting densities in areas where often travel long distances from nests to predators were reduced. Although greater brood rearing locations; thus, there appears adult philopatry and an increase in the to be little pressure to nest near high abundance of breeding yearlings associated concentrations of food for females and with increased nest success and natal goslings. Although Black Brant often travel philopatry are typically proposed as the a substantial distance from nest sites to mechanisms for this increase (Duebbert & brood rearing areas, females spend as Kantrud 1974; Duebbert & Lokemoen much as of 20% of the incubation period 1980; Garrettson & Rohwer 2001), these off the nest in maintenance activities such results also are consistent with the idea as feeding (Eichholz & Sedinger 1999; that waterfowl select sites with reduced Sedinger et al. 2004). Brant colonies typically predator abundance, thus immigrating into are located near the coast where nutritious experimental areas. Although results from foods are available and Arctic Fox Vulpes studies consistent with the idea that lagopus numbers are reduced due to fall waterfowl have developed a mechanism for flooding from storm surges (Mickelson assessing predator abundance and avoiding 1975; Raveling 1989). areas of high predator density, further

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 96 Nest site selection by waterfowl in the northern hemisphere empirical evidence is needed to test this not been tested empirically. Furthermore, hypothesis. while a pattern of increased nesting effort In addition to predator density, coexisting and success in years of greater small mammal prey density may play a role in nest site abundance is well established in arctic and selection. A number of studies have found a subarctic regions, the relationship is less clear correlation between coexisting prey in temperate regions, potentially due to abundance and nest success in tundra, taiga greater abundance of generalist predators and temperate grasslands (Pehrsson 1986; (Hanski et al. 1991). Perhaps an increase in Summers et al. 1994; Ackerman 2002; Brook abundance of coexisting prey, such as other et al. 2008; Iles et al. 2013). For some tundra waterfowl or bird eggs, arthropods or small nesting species, the relationship between nest mammals, would produce a functional success and coexisting prey appears adequate response by satiating or decreasing to cause certain populations of waterfowl to movement of predators, thus decreasing modify nesting distribution or forgo nesting susceptibility of nests to predation (Crabtree in years when prey abundance is low & Wolfe 1988; Crabtree et al. 1989; Larivière (Underhill et al. 1993; Sittler et al. 2000; & Messier 2001; Ackerman 2002). In Quakenbush et al. 2004). Researchers cannot contrast, increased abundance of coexisting explain whether decreased reproductive prey may produce a numerical response by investment is due to a lack of coexisting prey concentrating predators into areas of high (Bêty et al. 2001, 2002; Gauthier 2004; Iles et abundance of coexisting prey, decreasing al. 2013) or “protective umbrella” species – waterfowl nest success (Holt & Lawton species of predatory birds that inadvertently 1994). In the only known experimental study defend other birds’ nest from mammalian of nesting ducks belonging to different guilds predators while defending their own nest (i.e. tree cavity versus ground nesters), Elmberg (Dyrcz et al. 1981). The mechanism(s) by & Pöysä (2011) found that adding ground which waterfowl assess abundance of nests near cavity nests did not increase coexisting prey is also unclear. A potential predation risk for the latter in an area where mechanism may be use of mammalian urine the main nest predator (Pine Marten Martes similar to that demonstrated for raptors. martes) was a genuine generalist. Clearly, the Evidence consistent with the hypothesis interrelationships between waterfowl nest that predatory birds use UV light reflecting success, predator abundance and coexisting off phosphorous in mammal urine to prey abundance are complex and unresolved. locate areas of high prey density is well A lack of consistent results between nest documented (e.g. Viitala et al. 1995; Koivula & success and coexisting prey may be due to the Viitala 1999; Probst et al. 2002). Ducks can variability in the balance between the strength also see into the UV light spectrum (Jane & of functional and numerical responses Bowmaker 1988) and may use a technique associated with varying species of predators similar to that described by Eichholz et al. and abundance of coexisting prey (Ackerman (2012) to assess indirectly coexisting prey 2002; Brook et al. 2008), making predictions abundance. This hypothesis, however, has uncertain about how coexisting prey

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 97 distribution should impact nest site selection Interspecific Associations by waterfowl. Multiple studies have found evidence that Food availability subarctic and arctic nesting species nest in Non-breeding or abandonment of association with large avian predators even reproductive attempts have been observed in though the same predators prey on young Northern Pintail Anas acuta (Derksen & waterfowl (Young & Titman 1986; Underhill Eldridge 1980), Mallard (Krapu et al. 1983) et al. 1993; Summers et al. 1994; Quakenbush and Lesser Snow Geese Chen caerulescens et al. 2004; van Kleef et al. 2007). In the (Ankney & MacInnes 1978), indicating that process of deterring mammalian predators securing adequate resources is an important from their own nests, these avian predators component of reproductive success. Thus, inadvertently deter mammalian predators in contrast to colonial nesting species, a from nearby waterfowl nests. Thus, number of studies have found that dispersed waterfowl selecting nests within a protective nesting waterfowl nest in areas where food umbrella of predatory birds may gain for adults during incubation and post-hatch benefits of egg protection that outweigh young is available (Swanson et al. 1974; potential predation of hatchlings (Vermeer Derksen & Eldridge 1980; Haszard & Clark 1968; Young & Titman 1986; Bird & 2007; Fox et al. 2009). This phenomenon Donehower 2008). Some have suggested the may be because dispersed nesting waterfowl extent of this protection is so important for tend to be smaller bodied than colonial some species that certain individuals will nesting species, thus are required to feed forgo breeding in years when predatory more during incubation, or tend to nest amid birds are not present to provide protection more structurally complex vegetation that (Underhill et al. 1993; Summers et al. 1994; limits overland movement of young. For Quakenbush et al. 2000). In addition to example, a number of studies have found nesting in association with predatory reduced survival associated with increased species, smaller waterfowl may enjoy fitness overland movement of dispersed nesting benefits by nesting near large waterfowl that females and broods (Rotella & Ratti 1992; deter small and medium-sized predators Pearse & Ratti 2004; Simpson et al. 2005; (McLandress 1983; Baldwin et al. 2011). For Krapu et al. 2006; Davis et al. 2007); however, example, Canada Geese are known to other studies have found no relationship reduce predation on and increase species (Talent et al. 1983; Dzus & Clark 1997; Pöysä richness of co-nesting ducks (Fabricius & & Paasivaara 2006). In contrast, colonial Norgren 1987; Allard & Gilchrist 2002). To nesting species, such as Snow and Barnacle date, such relationships between predatory Geese have adapted to travel long distances birds, colonial birds and dispersed nesting from nest sites to brood sites to maximise waterfowl have been reported mainly from fitness, indicating little cost to overland travel arctic biota, but Fabricius & Norgren (1997) (Larsson & Forslund 1991; Sedinger 1992; observed diving and dabbling ducks nesting Aubin et al. 1993; Cooch et al. 1993). close to geese on islets in archipelagos in the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 98 Nest site selection by waterfowl in the northern hemisphere temperate biome. We see no reason why this 1990; Rolland et al. 1998). The most recent should not be a widespread phenomenon, discussions of the evolution of colonial suggesting it should be investigated more nesting suggest it evolves through: (1) a thoroughly. “limitation of breeding site” framework where a lack of nesting sites force individuals Nesting congregations to nest in aggregation with no net benefit Lack (1965) suggested that birds have (Wittenberger & Hunt 1985), (2) an evolved two primary forms of nest “economic” framework where a cost-benefit distribution, colonial and dispersed nesting. tradeoff of specific habitat conditions In waterfowl, however, there appears to be a favour coloniality (Alexander 1974; gradient from dense to loose colonies for Wittenberger & Hunt 1985; Sachs et al. some species to species generally considered 2007), or (3) a “by-product” framework dispersed nesters, but congregate into nest where individual habitat selections or sexual “clumps” or nest in high densities on selection leads to aggregation and colonial islands. Here, we partition the discussion breeding results from these individual of nesting congregations into three sections: selection decisions not as a direct result of (1) coloniality – which pertains to species being aggregated (Wagner et al. 2000; Wagner that generally congregate when nesting, & Danchin 2003; Sachs et al. 2007). (2) clump nesting – pertaining congregations To our knowledge, no studies have been of typically dispersed nesters in contiguous conducted directly to address theories on the upland habitat, and (3) island congregations evolution of coloniality in waterfowl; – typically dispersed nesters are found however, a number of studies appear to be nesting in congregations on islands. consistent with factors described under the “economic” framework. For Holarctic Coloniality waterfowl, evolution of coloniality has been Coloniality appears to be the evolved trait limited to species that generally breed in from the ancestral condition of dispersed open tundra, although admittedly it is a nesting (Coulson & Dixon 1979; matter of definition whether intra- and Wittenberger & Hunt 1985; Rolland et al. interspecific aggregation of nests on islands 1998) and may have evolved multiple times in prairies and archipelagos in temperate and due to a variety of selective pressures (Siegel- boreal regions should be construed as Causey & Kharitonov 1990; Rolland et al. colonial nesting. Because open tundra makes 1998). One reason this topic has garnered nest concealment difficult, this observation substantial attention is the few measured appears consistent with the hypothesis that benefits (advantages linked to predation and open habitat favours coloniality over enhanced food finding) relative to costs of dispersed nests as a means of predator colonial nesting (competition for food, nest avoidance. Indeed, positive correlations sites and mates, increased conspicuousness, between colony size and density and transfer of disease and parasites, nest success appear consistent with this cannibalism; Siegel-Causey & Kharitonov hypothesis (Bousfield & Syroechkovskiy

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 99

1985; Raveling 1989). Furthermore, the the strategy of dispersed nesting, because species of waterfowl that commonly nest nests become concentrated, allowing for colonially generally feed little during possible functional or numerical responses incubation; thus, shared information of by predators (Tinbergen et al. 1967; feeding sites appears unlikely to exert Holt 1977). Studies have found evidence considerable selective pressure for coloniality for negative density dependence, no (Milne 1976; Korshgen 1977; Parker & density dependence, and positive density Holms 1990; Erikstad & Tveraa 1995). dependence of nests in both artificial and Food availability during brood rearing, natural nest studies (Duebbert & Lokemoen however, may be influential for evolution of 1976; Andrén 1991; Major & Kendal 1996; coloniality in waterfowl. A number of Larivière & Messier 1998; Sovada et al. 2000; studies have now documented reduced Ackerman et al. 2004; Gunnarsson & growth rate of goslings with increased Elmberg 2008). Inconsistency in results colony size, observations consistent with a among studies likely is due to variation in the cost associated with colonial nesting (Cooch numerical (Holt 1977) and functional et al. 1991; Larsson & Forslund 1991; response behaviour of predatory species Sedinger et al. 1998). In the case of Black (Holling 1965; Tinbergen et al. 1967), Brant, however, grazing pressure from high variations in the response behaviour of densities of colonial geese appeared to prey, in the scale of the studies and in the maintain quality grazing lawns; thus, colonial habitat condition in which predators and behaviour appears to impact nutrient prey exist (Grand & Flint 1997; Flint et al. availability positively for young brant 2006; Ringelman et al. 2012). Thus, the (Person et al. 2003; Nicolai et al. 2008). adaptive costs and benefits of nesting within Hence, with currently available data, factors close proximity of heterospecifics and associated with the “economic” framework conspecifics are not well understood. seem most likely to explain the evolution of Clump nesting may be adaptive and due to coloniality in waterfowl; however, the multiple individuals selecting nest sites in specific mechanism(s) is still unclear and locations with fewer predators (Eichholz et may vary among species. al. 2012; Forsman et al. 2012), selection by multiple individuals of a yet unidentified nest Clump nesting site characteristic that provides appropriate With exception of a few species of sea ducks nest microclimate or safe habitat such as sites and geese, waterfowl are generally dispersed that more adequately disperse scent of nests nesters (Anderson & Titman 1992). An and hens (Conover 2007), rate of homing by unusual phenomenon often described by successful hens (Greenwood 1982; Hepp & researchers, but yet to be explained, is the Kennamer 1992; Blums et al. 2003; Öst et al. clumping of nests in relatively uniform 2011), natal philopatry of young (Hines & habitat (Duebbert & Lokemoen 1976; Hines Mitchell 1983; Lindberg & Sedinger 1997; & Mitchell 1983; Fowler et al. 2004; Fowler Coulton et al. 2011), or social attraction 2005). In theory, this behaviour contradicts and transfer of information (Hines &

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 100 Nest site selection by waterfowl in the northern hemisphere

Mitchell 1983; Pöysä 2006; Valone 2007). widely dispersed nests or the inability of Alternatively, in the case of negative density early nesters to defend territories and dependence and survival, clumped nesting maintain dispersed nests because of an may even be maladaptive behaviour, being overwhelming drive of individuals to nest due to environmental change outpacing the on islands (Mack et al. 2003). The latter ability of birds to adapt, creating a false explanation appears most likely based on signal for appropriate nest site selection the extreme number of pursuit flights (Dessborn et al. 2011). Because clumping emanating from islands during early nesting behaviour may appear obtuse evolutionarily (Duebbert 1966). An additional likely and with regard to conservation prerequisite for dense nesting congregations ramifications, a better understanding of the on island is adequate food resources to mechanistic characteristics of this behaviour support high densities of adults and young. and its adaptive significance is needed. Wetland proximity Natural islands and island Primary and secondary productivity during congregations summer dry seasons in the Holarctic is often A number of species prefer islands as nest concentrated around wetlands (Greenwood sites (Ryder 1972; Gosser & Conover 1999; et al. 1995; Larivière & Messier 2000). Traylor et al. 2004; Öst et al. 2011). Island Greater primary productivity within and nesting is thought to be beneficial because immediately adjacent to local complexes of important nest predators such as skunks, wetlands is thought to increase secondary badgers and foxes generally avoid water productivity, concentrating higher trophic (Ryder 1972; Mickelson 1975; Thompson & organisms, including predators, near Raveling 1987; Petersen 1990; Zoellick et al. wetlands (Greenwood et al. 1995; Larivière & 2004). An interesting aspect of island Messier 2000). In theory, birds should nest nesting is that a number of species of ducks away from wetlands, where predators are less and geese tend to nest at densities as much abundant (Robb & Bookhout 1995; as two orders of magnitude greater than Pasitchniak-Arts et al. 1998a; Phillips et al. densities observed on the mainland 2003; Traylor et al. 2004; cf. Keith 1961). For (Hammond & Mann 1956; Dwernychuk & wetland dependent precocial species, such as Boag 1972; Duebbert et al. 1983; Willms waterfowl, shorter travel distance from the & Crawford 1989). This occurrence is nest to brooding habitats may induce especially surprising for species such as selective pressure to nest close to wetlands Canada geese that typically are extremely (Duncan 1987). Furthermore, most ducks territorial, maintaining territories as large as make daily or multiple feeding bouts per day, ≥ 100 m around the nest on mainland. The leading to an additional energetic constraint mechanism(s) allowing extremely high for nesting far from wetlands (Shutler et al. nesting density in territorial species may be 1998). Thus, most waterfowl species face a due to decreased predation pressure leading trade-off between nesting farther from to fewer individuals attempting to maintain wetlands, where hatching success may be

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 101 increased, with nesting closer to wetlands for an adequate period of time (Clark & where duckling survival is maximised Shutler 1999). Because of benefits (Dzubin & Gollop 1972; Ball et al. 1975; associated with widely distributed nests, a Duncan 1987; Pöysä et al. 1999). The nest generally accepted paradigm is large patches site distance from wetlands that maximises of habitat are better for production and fitness likely varies among species with fitness of birds than small patches (Ball different life history traits (Duncan 1987). et al. 1995; Greenwood et al. 1995; Reynolds This prediction is supported by various et al. 2001). Smaller patches are thought to studies finding Northern Shoveler Anas be less productive because they increase clypeata and Blue-winged Teal nesting closer foraging efficiency of predators by to water than other species or random sites providing proportionally more edge habitat, while Northern Pintails nest farther from increase density of foraging predators water than other duck species (Dzubin & (i.e. concentration of enemies hypothesis), Gollop 1972; Ball et al. 1975; Livezey 1981; force birds to nest in greater density, Shutler et al. 1998). Studies comparing impact species composition of predator nesting distance from water between communities, provide more homogeneous mainland and islands also have found results vegetation facilitating movement of consistent with this tradeoff. Individuals predators, or increase dispersal inhibiting nesting on islands secluded from mammalian maintenance of higher concentrations and predators selected nest sites nearer to water, more intact communities (Higgins 1977; suggesting the threat of predation associated Clark & Nudds 1991; Stephens et al. 2004; with different landscapes and nest substrates Bayard & Elphic 2010). may affect the distance that females build Selection of larger patches should lead to nests to water (Kellet & Alisauskas 1997; a positive relationship between nest density Bentzen et al. 2009). and patch size, termed area sensitivity (Robbins et al. 1989; Bender et al. 1998; Habitat fragmentation Conner et al. 2000). When considering Historically, pristine nesting landscapes for breeding density and patch size relationships waterfowl, whether in temperate forests, for a wide diversity of fauna, a neutral grasslands, or sub-arctic boreal and tundra, relationship due to equilibrium theory of were vast mosaics of upland and wetlands. biogeography tends to be most supported However, agriculture, forestry, damming (Bender et al. 1998; Conner et al. 2000). For for hydroelectric power and other human avifauna, area sensitivity due to resource development have fragmented these concentration (more resources in larger landscapes, especially temperate uplands, patches) or concentration of enemies (higher making habitat patch size a potentially recent concentration of predators in smaller evolutionary nest site selection characteristic patches) often garners greatest support (Clark & Nudds 1991; Reynolds et al. 2001). (Raupp & Denno 1979; Conner et al. 2000). To become a selected trait, patch size For example, Ribic et al. (2009) reviewed would need to influence fitness predictably statistically rigorous studies of 32 species of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 102 Nest site selection by waterfowl in the northern hemisphere obligate grassland passerines and noted supporting the relationship among birds evidence for area sensitivity in half of those nesting in mixed and short grass prairies, species. The review by Ribic et al. (2009), and where most temperate upland nesting work of other researchers, suggested that waterfowl nest, remains inconsistent area sensitivity is strongest for forest (Sargeant et al. 1984; Krasowski & Nudds dwelling species relative to grassland species 1986; Clark & Nudds 1991; Horn et al. 2005). (Conner et al. 2000; Bayard & Elphick 2010), In short and mixed grass prairie grasslands, and proposed that area sensitivity appears to patch edges tend to be very distinct, with ignore the potential negative impacts of no transitional zones, and perhaps these negative density dependence on nest success abrupt edges don’t attract predators discussed earlier. (Pasitchniak-Arts & Messier 1998; Phillips et For upland nesting ducks breeding in the al. 2003; Horn et al. 2005). Additionally, Prairie Pothole Region in North America, grasslands tend to support both core and where habitat fragmentation has been most edge predators, making predation more dramatic, available evidence is inconsistent distributed across the landscape (Bergin et al. with the area sensitivity hypothesis (Clark & 2000; Chalfoun et al. 2002; Winter et al. 2006). Nudds 1991; Arnold et al. 2007; Haffele An alternative mechanism for area 2012). There are a number of reasons why sensitivity in grassland birds is the one may not expect to observe a nest “concentration of enemies” hypothesis density-patch size relationship for upland (Conner et al. 2000). This mechanism nesting ducks when considering mechanisms hypothesises that predators respond typically invoked to explain area sensitivity in numerically or behaviourally to passerines. In forested landscapes and tall fragmentation in the landscape, leading to a grass prairies adjacent to forests, edges positive relationship between patch size and typically result in an intermediate scrub- fitness of birds. Smaller patches may allow a shrub habitat that decreases nest success by predator to modify its behaviour by concentrating predators; thus, birds should concentrating foraging effort in remaining select nest sites farther from habitat edges grassland, increasing the likelihood a (Johnston & Odum 1956; Root 1973; Gates predator encounters a nest regardless of nest & Gysel 1978; Vickery et al. 1992). Because density. Additionally, smaller patches may be larger more uniform patches would allow more attractive to predators, thus increasing birds to select nest sites farther from edge, their abundance and concentration within larger patches are thought to be beneficial the patch (Root 1973; Sovada et al. 2000; (Gates & Gysel 1978; Pasitchniak-Arts & Kuehl & Clark 2002). Although this theory Messier 1996; Clark & Shutler 1999). appears intuitive, empirical support for the Although edge effects on nest success hypothesis is conflicting, with results of appear intuitive and have been well studies being both consistent (Fritzell 1975; documented in some landscapes (Root 1973; Oetting & Dixon 1975; Cowardin et al. 1985; Gates & Gysel 1978; Whitcomb et al. 1981; Johnson & Shaffer 1987) and inconsistent Sliwinski & Koper 2012), empirical evidence with this hypothesis (Duebbert & Lokemoen

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 103

1976; Livezey 1981; Vickery et al. 1992; Esler Greenwood et al. 1995). Such an outcome & Grand 1993; Phillips et al. 2004). may be especially relevant to upland nesting The remaining alternative explanations of waterfowl in grasslands and many authors area sensitivity in grassland birds are have suggested this mechanism has led to a associated with greater resource availability in decline in reproductive success of these birds larger patches. Resource availability is often (Greenwood et al. 1995; Sovada et al. 2000; found to influence habitat selection and Reynolds et al. 2001). Assuming wetland breeding distribution of wildlife (Stephens & resources drive waterfowl distribution, even if Krebs 1986). Studies often find evidence for waterfowl have adapted to select large area sensitivity with no corollary relationship patches, fragmentation and loss of grasslands in productivity, suggesting that resource have potentially forced them to nest in availability is the probable mechanism (Van remnant patches near wetlands that may not Horn 1983; Bock & Jones 2004; Winter et al. be the preferred patch size, obscuring any 2006). While this explanation is intuitive for relationship between patch size and nest passerines, which acquire most resources density. from upland landscapes, most resources for Finally, habitat fragmentation may not ducks are gleaned from wetlands during have been enacting selective pressures on breeding seasons, whereas grassland and ducks for an adequate period of time to other uplands merely provide cover for nests. evolve an adaptation. The majority of the In fact, for grassland nesting species of fragmentation of prairie grassland has waterfowl, which may not be as susceptible to occurred since the 1950s and fragmentation edge effects and acquire resources from of boreal forest is even more recent. wetlands, inverse area sensitivity often is Although selective pressures that directly predicted between nesting density and patch influence species demographics tend to size (MacArthur et al. 1972; Pasitchniak-Arts evolve quickly, 60 generations may not be et al. 1998; Sovada et al. 2000; Donovan & adequate time for a behavioural adaptation Lamberson 2001). Similar to the mechanism such as the selection of large patches to proposed for predators in the concentration occur especially considering that breeding of enemies hypothesis, habitat fragmentation individuals may not be reproductively hypothetically forces birds to nest densely in successful annually during their longevity. remnant cover, potentially leading to In summary, unlike forest birds, evidence “ecological traps” (Clark & Nudds 1991; Ball for area sensitivity or a positive relationship et al. 1995; Reynolds et al. 2001). This between reproductive success and patch size hypothesis assumes bird population for grassland birds in short or mixed grass abundance is at some level independent of prairies is equivocal. Furthermore, the few grassland abundance, so when grasslands are studies that have tested for a breeding reduced, birds modify distributions but do density:patch size relationship were not not exhibit an isometric decline in consistent with the prediction of area abundance, leading to inverse area sensitivity sensitivity or inverse area sensitivity typically (Braun et al. 1978; Clark & Nudds 1991; proposed in studies associated with habitat

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 104 Nest site selection by waterfowl in the northern hemisphere fragmentation. The limited number of study durations (2–3 years) preventing studies that address this question appears inference regarding short- versus long-term surprising considering patch size plays such variation (Clark & Nudds 1991). Additionally, an important role in habitat conservation dramatic anthropogenic changes to natural and restoration. environments since the 20th century may have caused some nest sites selection Fourth-order selection characteristics to become maladaptive traits This level of selection has been the focus of as environmental change outpaced adaptive most waterfowl nesting studies (Kaminski & ability of birds (Dwernychuk & Boag 1972; Weller 1992). While predator avoidance and van Riper 1984; Schlaepfer et al. 2002). food availability for adults and young likely Herein, we review characteristics proposed play a strong role in 1st–3rd orders of nest as 4th order nest site characteristics of site selection, predator avoidance seems waterfowl and their level of empirical the predominant selective force for 4th support. order selection. Nest site selection likely influences predation rate and most female Vegetation structure annual mortality occurs during the breeding Evidence supports the notion that physical season (Ricklefs 1969; Southwood 1977; structure of the vegetation is an important Hoekman et al. 2002). criterion of nest site selection for almost all Nest-site characteristics are well species of ground and over-water nesting documented for many species of birds waterfowl (McLandress 1983; Miller et al. including waterfowl and are often found to 2007; Safine & Lindberg 2008; Haffele et al. be significantly different from characteristics 2013). Two characteristics that have been at randomly selected locations (Bellrose 1980; identified as important, especially for many Clark & Shutler 1999). This non-random ground-nesting ducks, are height and density distribution is usually assumed to be caused of vegetation. In general, they appear to by habitat preference and thus adaptive select taller and thicker vegetation (Schrank (Martin 1998; Clark & Shutler 1999). 1972; Martin 1993; Clark & Shutler 1999; Preferred nest site characteristics have been Haffele et al. 2013). Cover height appears to described in numerous studies, but evidence be most important when the primary of their adaptive value is inconsistent and predators are avian but also can be important limited (Hines & Mitchell 1983; Crabtree et al. when mammalian predators such as Striped 1989; Clark & Shutler 1999; Durham & Skunks Mephitis mephitis are the primary Afton 2003). This inconsistency is at least predator (Hines & Mitchell 1983; Crabtree et partially due to species-specific variation; al. 1989), whereas cover density is most geographic, temporal and design variation of important when the primary predators are studies; focus on restricted components of mammalian (Schrank 1972; Bowman & the ecological community (i.e. only waterfowl Harris 1980; Livezey 1981; Hines & Mitchell and predators, waterfowl and vegetation, 1983; Rangen et al. 2000). Height and density or predators and vegetation); and short of cover may directly impact success by

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 105 obstructing movement of predators, for selection of intermediate nest cover increasing abundance of alternative prey, height to occur in ground nesting birds providing visual obstruction, or obstructing (Hines & Mitchell 1983; Crabtree et al. 1989; distribution of scent. While a large number Durham & Afton 2003; Haffele et al. 2013). of studies have found waterfowl select Level of reproductive investment varies specific structural characteristics of among species based on the likelihood of vegetation for nesting, studies testing for the future productivity (Fontaine & Martin 2006; adaptive benefit have been inconsistent with Dassow et al. 2012). Theoretically, factors only a few studies finding a positive that influence 4th order nest site selection, relationship with nest success (Glover 1956; such as vegetation height, should vary Crabtree et al. 1989; Clark & Shutler 1999; among bird species based on life-history Durham & Afton 2003). The inconsistency traits leading to interspecific variation (Grant of results among studies likely is due to & Shaffer 2012). This theoretical relationship variation in methodology and scale, the is supported among closely related species of complex relationship between life history ducks; shorter lived species that invest more characteristics, cover characteristics, types of into current reproduction tend to nest in predators, availability of alternative prey or a more dense cover (Keith 1961; Duncan combination of these other factors (Clark & 1986; Greenwood et al. 1995; Haffele 2012). Nudds 1991; Horn et al. 2005; Haffele 2012). This interspecific variation in nesting cover Some authors have proposed a tradeoff requirements should be recognised when between concealment of the eggs from planning and implementing conservation predation (current reproductive investment) and restoration of waterfowl nesting habitats and escaping from predation (future and not establish only uniform dense nesting reproductive investment) leading to covers (Keith 1961; Livezey 1981; Lokemoen selection of cover of an intermediate height et al. 1990; Greenwood et al. 1995). and density (Götmark et al.1995; Traylor et al. 2004; Miller et al. 2007; McRoberts et al. Specific composition of cover 2012); however, evidence supporting the Species composition of vegetation can selection of cover height or density as a influence nest site selection and success, and stabilising selective trait is equivocal (Keith preferred vegetative composition varies 1961; Duncan 1986; Clark & Shutler 1999; among closely related species (Keith 1961; Haffele et al. 2013). Additionally, for some Duncan 1986; Crabtree et al. 1989; Lokemoen species, likely because cover > 45 cm tends et al. 1990). Crabtree et al. (1989) found to shade out shorter vegetation and thereby nest success to be greatest in vegetation reduces density at ground level, nest success composed of grasses and forbs and appears to be greatest when cover is at an surmised this combination provided greater intermediate height (16–45 cm; Crabtree et concealment by grasses at lower levels and al. 1989; Haffele et al. 2013); thus, the forbs at higher levels. Although physical perceived tradeoff between current and characteristics of vegetative cover likely have future reproductive success is not necessary the greatest influence on nest site selection by

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 106 Nest site selection by waterfowl in the northern hemisphere birds (Schrank 1972; Gilbert et al. 1996), prospecting for nest sites. An issue related to factors influenced by vegetative species this is the potential mismatch between composition itself also may directly or nesting timetable of waterfowl and indirectly influence nest site selection. For vegetation development that may arise as a example, although measures of height and consequence of global climate change density did not discern differences between (Drever & Clark 2007). If these processes forbs and grasses, Blue winged Teal and become uncoupled, decreased nest success Northern Shoveler tended to select sites with may result, which may ultimately affect more grasses while Mallard and Gadwall population trajectories (Drever et al. 2012). selected sites with more forbs (Livezey 1981; Lokemoen et al. 1990; Clark & Shutler 1999). Vegetative litter and remnant down These researchers could not discern if these Aldo Leopold (1933) first proposed leaf differences in selection were due to smaller litter was an important component of nest bodied teal being constrained by robust forbs, site selection and success for ground nesting a difference in unmeasured characteristics birds. The overall importance of litter such as overhead cover, or a direct selection depth in selection of nest-sites has been for preferred species of vegetation. For documented for certain grassland songbirds example, certain species of vegetation being (Winter 1999; Davis 2005; Fisher & Davis preferred or avoided as nest sites due to their 2010), but the relative importance for ability to disrupt scent plumes of birds and waterfowl is largely unknown. Leaf litter nests (Aylor et al. 1993; Conover 2007). and remnant down may be important Regardless of the cause, because of the for providing the appropriate thermal continuous loss of nesting cover in temperate microenvironment and concealing eggs from regions of North America and replacement predation, especially for early nesting species of native vegetation with exotic species, prior to emergence of new vegetation (Bue et understanding the degree of specificity or al. 1952; Duebbert 1969; Fast et al. 2010; level of plasticity in this trait may be Haffele 2012). Alternatively, depth of important for determining the necessary remnant leaf litter may be used as a predictor composition of nesting cover that maximises for the amount of vegetation that can be benefits to waterfowl. Choosing a specific expected to grow in that location later in the plant species to provide cover is a process nesting season (Haffele 2012). Height and that also has a component of phenology and amount of leaf litter were important factors between-year variation in the advancement of for nest site selection by early (Mallard, spring. This is especially true for early nesting Northern Pintail), intermediate (Blue- species such as Mallard and Northern Pintail, winged Teal) and late nesting ducks (Gadwall which often start nest-building and egg-laying and Northern Shoveler; Haffele 2012). before the present year’s plants provide cover Haffele (2012) found strong selection by adequate for nesting. Accordingly, in early ducks for nest sites with deeper leaf litter, nesting species, evergreens and last year’s and found nest sites with more litter had vegetation may act as cues for hens lower nest mortality. Furthermore, Common

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 107

Eiders initiated nesting earlier in nest bowls may not be immediately recognised (Wilson with experimentally increased amount of & Verbeek 1995; Zicus et al. 2004; Hepp et al. down suggesting preference for nest bowls 2006; Fast et al. 2007; DuRant et al. 2011). with remnant down from the previous year’s Furthermore, the cost of incubating eggs nesting attempt. However, during heavy with less than optimal microclimates can snow years, both Cackling Geese Branta affect the adult’s ability to care for young and hutchinsii and Emperor Geese Chen canagica adult survive post hatch (Erikstad & Tveraa selected sites with more short dead 1995; Öst et al. 2003). Selecting nest sites that vegetation (Petersen 1990; Fast et al. 2010). minimise energetic costs of incubation could be especially important for smaller bodied Microclimate species nesting in temperate areas or larger In addition to shielding the nest and hen bodied species in arctic and sub-arctic from predators, nest site selection influences regions (Piersma et al. 2003; Hilton et al. physical conditions of incubation such 2004). Selecting nest sites that provide as shelter from wind, relative humidity, appropriate insulation could come at a cost precipitation and excessive solar radiation to nest concealment, leading to a tradeoff diurnally and loss nocturnally (Walsberg between appropriate microclimate and 1981). Because the vast majority of egg concealment (Shutler et al. 1998). Finally, nest mortality comes from predation with only a site micro-climate could indirectly influence minor component of embryonic mortality risk of predation by influencing incubation associated with nest microclimate, studies behaviour of adults. Less insulated clutches associating nest site selection with could be more energetically expensive to microclimate are few relative to those incubate forcing hens to leave the nest more associating nest site selection with predation. to feed, thus increasing susceptibility to For example, Gloutney & Clark (1997) egg predation (Thompson & Raveling investigated the influence of nest site 1987; Afton & Paulus 1992; Durrant et al. selection of Mallard and Blue-winged Teal 2011). Because of the limited amount of on nest microclimate and concluded interest this area has received and the selective pressure of optimising physical potential demographic impact of reduced conditions of incubation is secondary to reproduction, more studies simultaneously the selective pressures of egg and hen considering the impact of nest site selection survival, based on a combined measure of on microclimate, predator avoidance and the temperature and relative humidity. More long-term ramifications on fitness of both recently, however, evidence is accumulating adults and young seem warranted. indicating the impact of the nest microclimate is not limited to the immediate Date of ice or snow melt impact of embryonic mortality but can To breed successfully, the ice- and snow-free influence both short- and long-term fitness period must be sufficient for waterfowl to lay of the adults and young, thus benefits and incubate eggs, for young to grow and for associated with favourable microclimates young and moulting adults to attain flight and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 108 Nest site selection by waterfowl in the northern hemisphere acquire adequate nutrient reserves to sustain species, related individuals have been found them for the first stage of the fall migration. to nest in closer proximity than would be Because the amount of time to achieve these expected under a random distribution (van stages of the annual cycle is so limited, there der Jeugd et al. 2002; Fowler et al. 2004; is strong selective pressure for females to Sonsthagen et al. 2010). Clusters of related initiate nesting as soon as possible at latitudes individuals can arise due to adult and natal where the time window for successful breeding philopatry, phenotype matching, breeding may be limiting. Earlier nest and kinship associations (Ely & Scribner initiation for arctic and subarctic breeding 1994; van der Jeugd et al. 2002; Fowler et al. species has been found to influence nest 2004; Sonsthagen et al. 2010). However, survival, growth rates of young, adult body evidence is consistent with the idea that size, year of first breeding and first year individuals actively seek nest sites near survival; which are vital rates that influence closely related individuals (van der Jeugd et individual fitness (Lindholm et al. 1994; al. 2002; Sonsthagen et al. 2010; Fishman et Sedinger et al. 1995; Cooch 2002; Blums et al. al. 2011). For example, Barnacle Geese have 2005; Pilotte et al. 2014). This strong selective been observed nesting near their siblings pressure for early nest initiation has caused from the same brood on islands different the date that potential nesting sites become from their natal island, indicating nesting snow free to be an important component of proximity was not due to natal philopatry nest site selection in species breeding in arctic (van der Jeugd et al. 2002). A similar habitats (Ely & Raveling 1984; Petersen 1990; observation was not made regarding sisters Chaulk et al. 2007; Lecomte et al. 2008). from different broods (i.e. different years), a Because winter winds often sweep snow from result most consistent with the hypothesis the highest potential nest sites, the date a nest females are actively selecting nest sites near site becomes ice or snow free is often strongly known kin (van der Jeugd et al. 2002). At this correlated to the height of the nest site. Thus, point, benefits of selecting nest sites near when selecting the first snow-free sites, kin are speculative but may include waterfowl are often selecting for sites with increased willingness for cooperation in higher elevation in the area (Mickelson 1975; predator defence, joint defence of high Eisenhauer & Kirkpatrick 1977; Peterson quality food patches and relatives with 1990). Selection of higher nesting sites also which they share heredity, or decreased has been reported in more temperate areas, aggression between related neighbours (van likely decreasing the potential for flooding der Jeugd et al. 2002; Sonsthagen et al. 2010; and increasing probability for early snow melt Fishman et al. 2011). from tallest sites (Jarvis & Harris 1967; Ely & Raveling 1984; O’Neil 1988). Fidelity, experience and information sharing Kinship The costs and benefits of nest site fidelity For some semi- and colonial nesting species and complementary behaviour of dispersal or for clumps of nests of dispersed nesting are thought to be an influential component of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 109 nest site selection (Hinde 1956; Greenwood et al. 2003; Valone 2007). For waterfowl, 1982). There are a number of proposed there is some evidence of hetero- and advantages to exhibiting fidelity to a nesting conspecific attraction (Elmberg et al. 1997; site, including familiarity with food and other Pöysä et al. 1998); however, use of public resources and with neighbours decreasing information for nest sites selection is only aggressive interactions (Greenwood 1982), speculative for ground nesting birds and has all of which may ramify into to greater received limited support for cavity nesting productivity (Harvey et al. 1979; Newton & species (Pöysä 2006; Ringelman et al. 2012; Marquiss 1982; Gratto et al. 1985; Korpimäki cf. Roy et al. 2009). 1988). One particularly important factor influencing an individual’s decision to Obstacles and structures maintain breeding site fidelity or disperse In addition to vegetation, waterfowl nesting appears to be past reproductive success. in open landscapes, such as tundra, often Making nest site selection decisions based on use obstacles such as rocks or drift wood to past experiences would require some level of conceal nests from predators and increase consistency in the success of specific nest quality of the nesting microclimate (Ryder site. Although there often appears to be 1972; Noel et al. 2005; Fast et al. 2007; Öst substantial annual variability in nest success & Steele 2010). Waterfowl nesting in within a nesting location (Haffele et al. association with such obstacles have been 2013; Ringelman 2014), other studies have reported to have higher nest success and documented consistency in security of nest reduced weight loss during incubation (Kilpi sites, such as nest cavities used by Common & Lindström 1997; Öst et al. 2008b; Öst & Goldeneyes Bucephala clangula (Elmberg & Steele 2010). Similar to waterfowl nesting in Pöysä 2011). Therefore, a number of studies vegetation, however, there appears to be a have found migratory birds disperse farther tradeoff between level of concealment and when they fail in a reproductive attempt the ability of incubating hens to detect and (Weatherhead & Boak 1986; Paton & quickly escape from predators (Öst et al. Edwards 1996). This behaviour also has been 2008a; Öst & Steele 2010). observed in waterfowl, with a positive relationship between degree of fidelity and Nesting cavities fecundity (MacInnes & Dunn 1988; Hepp & A somewhat unique adaptation evolved by a Kennamer 1992; Lindberg & Sedinger 1997; number of birds including waterfowl is Öst et al. 2011). nesting in tree and artificial cavities. Birds In addition to the individual’s past likely began nesting in cavities in an effort to reproductive success, a number of studies avoid nest predation (Lack 1954; Nilsson have provided evidence that birds use 1986). Ducks are secondary cavity nesters success of neighbours (public information) that rely on tree holes excavated by other to make initial nest site selection or wildlife, natural formation through injury to determine whether to disperse or exhibit trees and subsequent decay, or nest boxes spatial fidelity (Boulinier et al. 1996; Doligez erected by humans (Bellrose & Holms 1994;

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 110 Nest site selection by waterfowl in the northern hemisphere

Nielsen et al. 2007). Although cavities were to be more important than the cavity tree once thought to be limited in North America characteristics when ducks are selecting nest for Wood Duck Aix sponsa; this may no sites (Robb & Brookhout 1995; Yetter et al. longer be the pervasive reality in most areas 1999). Cavities used by waterfowl have been (Soulliere 1988; Nielsen et al. 2007; Denton et found to be higher from the ground or water, al. 2012). Lowney & Hill (1989), however, nearer forest opening or wetlands, and have reported that densities of cavities suitable for entrances with smaller widths and heights Wood Duck nesting (i.e. adequate dimensions (Prince 1968; Dow & Fredga 1985; Yetter et and internal surface for eggs) in Mississippi al. 1999; Robert et al. 2010). All of these hardwood bottomlands were among the characteristics have been found to reduce the lowest reported for mature forests in North potential for predation. Orientation of cavity America. Additionally, in more northern entrance could also be important for nest regions of Europe and North America, areas microclimate (Gilmer et al. 1978), and while where cavities are used by Smew Mergellus distance to wetlands was important for albellus and Bucephala species, cavities are still Common Goldeneyes (Pöysä et al. 1999), it thought to limit some populations (Savard did not appear to have a strong influence on 1988; Pöysä & Pöysä 2002; Vaillancourt et al. cavity use by Wood Ducks (Robb & 2009; Robert et al. 2010), mainly because Bookhout 1995). natural forests contain many more old and hollow trees than do modern managed Conclusions and implications forests. Thus, some suggest nest box Throughout the 20th and 21st centuries, programmes are important to maintain both environmental impacts and breeding populations of cavity nesting ducks conservation-management actions to in some regions (Lowney & Hill 1989). remediate those impacts for breeding Although over-water cavities would waterfowl have occurred on characteristics intuitively appear to be more secure and nest that affect nest site selection at the local success has been found to be higher during or patch and microhabitat scales (i.e. 3rd periods of flood (Nielsen & Gates 2007), and 4th order selection), those likely there appears to be no clear selection for based almost entirely on more fine-grained over-water cavities for Wood Ducks. interactions (Hutto 1985; Kaspari et al. 2010). However, Wood Duck duckling survival was The outcomes of habitat selection decisions greatest for individuals hatched in predator- at 3rd and 4th order scales influence species protected nest structures located amidst compositions of waterfowl communities flooded scrub-shrub wetlands in Mississippi, in breeding areas, which may influence which may have concealed nest structures predator-prey relationships and competition from avian egg predators during egg laying for resources on an ecological time scale and incubation or provided near cover for and behavioural and morphological hens and broods after exodus from the nest characteristics on an evolutionary scale (Davis et al. 2007, 2009). Moreover, regarding (Morris 2003). Outcomes from these natural cavities, cavity characteristics appear intrinsic interactions likely provided selective

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 111 forces acting upon individuals which may present regional community composition. have influenced their 1st and 2nd order Nevertheless, the ecological time frame lets selection. Additionally, whether hierarchical us study range expansions and retractions habitat selection by migratory waterfowl is and abundance shifts relative to abiotic and driven by bottom-up, top-down, or both biotic conditions. Both are important agents processes in different time and space scales shaping waterfowl and other wildlife on individual survival and reproductive distributions and communities in the past success remains intriguing and worthy of and present (e.g. Schummer et al. 2010; Pearse research (Kaminski & Elmberg 2014). et al. 2010). Furthermore, the scale has Moreover, 1st order scales of selection of changed at which ecologists now recognise migratory waterfowl operate at a continental anthropogenic activities are modifying context and generally determine distribution the environment. Climate change and or range limits of species, while 2nd order the additive effects of natural resource scales of selection often determine habitat(s) exploitation and agricultural development by used within a biome(s) in which species and humans are now recognised to influence the individuals distribute themselves. Decisions environment at the scale of the biome or at the 1st and 2nd level seem based on continent; thus, understanding ecological large-grain abiotic (e.g. wetland system questions at a larger scale through characteristics, climate, and landscape macroecological studies will become configurations) and biotic characteristics (e.g. more relevant for the management and terrestrial and aquatic communities) that may conservation of waterfowl. As this review be perceived from a long distance but likely indicates, however, studies of 1st and 2nd have developed through novel or philopatric order waterfowl nest site selection are experiences at smaller scales (Hutto 1985). relatively few and often indirect, forcing us Hutto (1985) termed these large-grained to speculate based on results of 3rd and 4th characteristics “extrinsic characteristics” order selection. Thus, an increased emphasis because they are external to local habitats or on studies addressing 1st and 2nd orders of patches; thus, these characteristics may habitat selection appears warranted. Yet, we not lend themselves to manipulation by must recognise that some characteristics are management and conservation actions. changing at the granular level of 1st and 2nd Observed patterns from 1st and 2nd order order selection and may be manifested or selection are results of processes in detected in 3rd or 4th orders of selection evolutionary (ultimate factors) more than in studies through outcomes of individual ecological time frames (proximate factors) distribution, survival and reproductive and are likely maintained through adult and success. Thus, even studies addressing 1st natal philopatry (Klomp 1953; Hutto 1985). and 2nd order selection should associate the Even in migrants, like most waterfowl, outcome of those selection decisions at the glaciation and concurrent biome and level of the individuals to understand climate shifts are background long-term processes and patterns promoting fitness. influences of species pools and hence of We also recognise the process of nest site

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 112 Nest site selection by waterfowl in the northern hemisphere selection interacts strongly with the process our ability to understand the process of of community ecological development, yet nest selection, limiting management and most studies of this process are limited to conservation actions; thus, a multi-scale very few components of the community community approach of study is warranted. (e.g. vegetation characteristics and nest This approach might best be achieved success or predator abundance and nest by following the advice of Bloom et al. success) and often only one or two small and (2013) and incorporating both habitat often ambiguously defined landscape scales. selection and demographic variables in the We suggest the simplistic approach that modelling process. We also recognise has dominated past studies has limited questions of causality are best addressed

Table 1. Non-ranked priority recommendations for future research on nest site (habitat) selection in waterfowl, derived but adapted from 16 suggestions proposed by Kaminski and Weller (1992). We recognise that a number of the recommendations address recruitment, whereas this review was limited to nest site selection. Thus, our recommendations are limited to the nest site selection component of recruitment.

Priority Recommendation

1 Relate habitat selection to waterfowl survival and recruitment rate. 2 Test models and determine the effects on recruitment of possible inter and intra-specific density-dependent habitat selection. 3 Determine effects of different densities and communities of predators and prey on waterfowl habitat selection and how that process interacts with scale and community of vegetation. 4 Invoke hierarchical approaches in studies of habitat selection by individuals to obtain data on individuals’ habitat use throughout their annual cycle and range for incorporation into population models and to guide habitat conservation planning and implementation. 5 Relate long-term changes in wetland and upland composition to corresponding changes in the variety of interacting factors (e.g. vegetation, food, predators and competitors) that influence waterfowl recruitment at multiple scales. 6 Determine effects of variation and loss in availability of intermittently, temporarily, and seasonally flooded wetlands on waterfowl habitat selection, dispersal and recruitment. 7 Determine the effect of habitat fragmentation on waterfowl habitat selection and community organisation.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 113 through manipulative experiments; advancing reviews of waterfowl pre-fledging however, experimental studies are inherently ecology and summarising the current status difficult and costly to carry out at larger of information regarding 3rd and 4th order scales. This dilemma is a true challenge brood site selection (e.g. Sedinger 1992), given for waterfowl and wildlife research, that 1st and 2nd order selection occur at the conservation and management, but it time of nest site selection and were covered deserves resolution and implementation in this review. through landscape-scale cooperatives of ecological and conservation partners. Acknowledgements Kaminski & Weller (1992) produced a We would like to thank Stuart Slattery, Dave thorough review of breeding habitats of Howerter, and Tony Fox for their input in Nearctic waterfowl. At the conclusion of the oral presentation of this manuscript at their review, they identified 16 issues which the 6th North American Duck Symposium, they believed needed further study. Here, we Ecology and Conservation of North suggest that a slightly modified list of seven American Waterfowl. We also thank Rick of those 16 issues warrant priority Kaminski and Aaron Pearse for constructive consideration (Table 1). We conclude that comments and edits, both of which greatly only the first of the 16 recommendations benefitted the quality of this manuscript, has been explored with sufficient replication and Lauren Hill for transcribing the using species of adequately diverse life numerous citations for this manuscript. histories to allow for general inference and management actions, although some unique References species of conservation concern are Ackerman, J.T. 2002. Of mice and mallards: exceptions. This observation corroborates positive indirect effects of coexisting prey our assertion that the advancement in on waterfowl nest success. Oikos 99: 469– understanding the process of nest site 480. selection has been limited by approach. Ackerman, J.T., Blackmer, A.L. & Eadie, J.M. Finally, the time required to complete the 2004. Is predation on waterfowl nests density task of reviewing thoroughly the substantial dependent? – Tests at three spatial scales. literature on multi-scale nest site selection by Oikos 107: 128–140. Holarctic waterfowl has prevented us from Afton, A.D. & Paulus S.L. 1992. Incubation and addressing habitat selection by broods, a brood care. In B.D.J. Batt, A.D. Afton, M.G. second but no less important component of Anderson, C.D. Ankney, D.H. Johnson, J.A. Kadlec & G.L. Krapu (eds.), Ecology and breeding habitat selection. In fact, a recent Management of Breeding Waterfowl, pp. 62–108. Mallard study suggests potential tradeoffs University of Minnesota Press, Minneapolis, between nest site and brood habitat selection, USA. suggesting simultaneous integration of these Alerstam, T. & Högstedt, G. 1982. Bird migration may be most appropriate for future studies and reproduction in relation to habitats for (Bloom et al. 2013). We recommend survival and breeding. Ornis Scandinavica 13: undertaking the task of updating and 25–37.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 114 Nest site selection by waterfowl in the northern hemisphere

Alerstam, T., Hedenström, A. & Åkesson, S. Aylor, D.E., Wang, Y. & Miller, D.R. 1993. 2003. Long-distance migration: evolution Intermittent wind close to the ground within and determinants. Oikos 103: 247–260. a grassy canopy. Boundary-Layer Meteorology Alexander, R.D. 1974. The evolution of social 66: 427–448. behavior. Annual Review of Ecology and Baldwin, F.B., Alisauskas, R.T. & Leafloor, J.O. Systematics 5: 325–383. 2011. Nest survival and density of cackling Allard, K. & Gilchrist, H.G. 2002. geese inside and outside a Ross’s goose Kleptoparasitism of Herring Gulls taking colony. The Auk 128: 404–414. eider eggs by Canada Geese. Waterbirds 25: Ball, I.J., Eng, R.L. & Ball, S.K. 1995. Population 235–238. density and productivity of ducks on large Anderson, M.G. & Titman R.D. 1992. Spacing grassland tracts in Northcentral Montana. patterns. In B.D.J. Batt, A.D. Afton, M.G. Wildlife Society Bulletin 23: 767–773. Anderson, C.D. Ankney, D.H. Johnson, J.A. Bayard, T.S. & Elphick, C. 2010. How area Kadlec & G.L. Krapu (eds.), Ecology and sensitivity in birds is studied. Conservation Management of Breeding Waterfowl, pp. 251–289. Biology 24: 938–947. University of Minnesota Press, Minneapolis, Bellrose, F.C. 1980. Ducks, Geese and Swans of USA. North America. Stackpole Books, Harrisburg, Andrén, H. 1991. Predation: an overrated factor Pennsylvania, USA. for over-dispersion of birds’ nests? Animal Bellrose, F.C. & Holms D.J. 1994. The Ecology Behaviour 41: 1063–1069. and Management of Wood Ducks. Stackpole Ankney, C.D. & MacInnes, C.D. 1978. Nutrient Books, Mechanicsburg, Pennsylvania, reserves and reproductive performance of USA. female Lesser Snow Geese. The Auk 95: Bender, D.J., Contreras, T.A. & Fahrig, L. 1998. 459–471. Habitat loss and population decline: a meta- Arnold, T.W., Craig-Moore, L.M. & Armstrong, analysis of the patch size effect. Ecology 79: L.M. 2007. Waterfowl use of dense nesting 517–533. cover in the Canadian Parklands. The Journal of Bennett, L.J. 1938. Redheads and ruddy ducks Wildlife Management 71: 2542–2549. nesting in Iowa. Transactions of the North Arrhenius, O. 1921. Species and area. Journal of American Wildlife and Natural Resources Ecology 9: 95–99. Conference 3: 647–650. Arzel, C., Elmberg, J., Guillemain, M., Lepley, M., Bent, A.C. 1923. Life Histories of North American Bosca, F., Legagneux, P. & Nogues, J-B. 2009. Wildfowl. United States Natural History A flyway perspective on food resource Museum Bulletin No. 126. U.S. Natural abundance in a long-distance migrant, the History Museum, Washington D.C., USA. Eurasian Teal (Anas crecca). Journal of Bentzen, R.L., Powell, A.N. & Suydam, R.S. 2009. Ornithology 150: 61–73. Strategies for nest-site selection by King Ashmole, N. 1963. The regulation of numbers Eiders. Journal of Wildlife Management 73: of tropical oceanic birds. Ibis 116: 217– 932–938. 219. Bergin, T.M., Best, L.B. & Freemark, K.E. 2000. Aubin, A.E., Dzubin, A., Dunn, E.H. & Effects of landscape structure on nest MacInnes, C.D. 1993. Effects of summer predation in roadsides of a Midwestern feeding area on gosling growth in Snow agroecosystem: a multiscale analysis. Geese. Ornis Scandinvica 24: 255–260. Landscape Ecology 15: 131–143.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 115

Bêty, J., Gauthier, G., Giroux, J.-F. & Korpimäki, Bowman, G.B. & Harris, L.D. 1980. Effect E. 2001. Are goose nesting success and of spatial heterogeneity on ground-nest lemming cycles linked? Interplay between nest depredation. Journal of Wildlife Management 44: density and predators. Oikos 93: 388–400. 806–813. Bêty, J., Gauthier, G. Korpimäki, E. & Giroux, J- Braun, C.E., Harmon, K.W., Jackson, J.A. F. 2002. Shared predators and indirect & Littlefield, C.D. 1978. Management of trophic interactions: lemming cycles and National Wildlife Refuges in the United arctic-nesting geese. Journal of Animal Ecology States: its impacts on birds. Wilson Bulletin 90: 71: 88–98. 309–332. Bird, D.M. & Donehower, C.E. 2008. Gull Brook, R.W., Pasitschniak-Arts, M. & Howerter, predation and breeding success of Common D.W. 2008. Influence of rodent abundance Eiders on Stratton Island, Maine. Waterbirds on nesting success of prairie waterfowl. 31: 454–462. Canadian Journal of Zoology 86: 497–506. Block, W.M. & Brennan, L.A. 1993. The habitat Brown, J.H. 2014. Why are there so many species concept in ornithology: theory and in the tropics? Journal of Biogeography 41: 8–22. applications. Current Ornithology 11: 35–91. Brown, J.H. & Maurer, B.A. 1989. Macroecology: Bloom, P.M., Clark, R.G., Howerter, D.W. & The division of food and space among Armstrong, L.M. 2013. Multi-scale habitat species on continents. Science 243: 1145–1150. selection affects offspring survival in Bue, I.B., Blankenship, L.H. & Marshall, W.H. a precocial species. Oecologia 173: 1249– 1952. The relationship of grazing practices 1259. to waterfowl breeding populations and Blums P., Nichols, J.D., Lindberg, M.S., Hines, production on stock ponds in western South J.E. & Mednis, A. 2003. Factors affecting Dakota. Transactions of the North American breeding dispersal of European ducks on Wildlife and Natural Resources Conference 17: Engure Marsh, Latvia. Journal of Animal 396–414. Ecology 72: 292–307. Chalfoun, A.D., Thompson, F.R. & Ratnaswamy, Blums, P., Nichols, J.D. & Hines, J.E. 2005. M.J. 2002. Nest predators and fragmentation: Individual quality, survival variation and a review and meta-analysis. Conservation patterns of phenotypic selection on body Biology 16: 306–318. condition and timing of nesting in birds. Chalfoun, A.D. & Martin, T.E. 2009. Habitat Oecologia 143: 365–376. structure mediates predation risk for Bock, C.E. & Jones, Z.F. 2004. Avian habitat sedentary prey: experimental tests of evaluation: should counting birds count? alternative hypotheses. Journal of Animal Frontiers in Ecology and the Environment 2: Ecology 78: 497–503. 403–410. Chaulk, K.G., Robertson, G.J. & Montevecchi, Boulinier, T., Danchin, E. & Monnat, J.Y. 1996. W.A. 2007. Landscape feature and sea ice Timing of prospecting and the value of influence nesting common eider abundance information in a colonial breeding bird. and dispersion. Canadian Journal of Zoology 85: Journal of Avian Biology 27: 252–256. 301–309. Bousfield, M.A. & Syroechkovskiy, Y.V. 1985. A Chesser, T. & Levey, D.J. 1998. Austral migrants review of Soviet research on the Lesser and the evolution of migration in new world Snow Goose on Wrangel Island, U.S.S.R. birds: diet, habitat and migration revisited. Wildfowl 36: 13–20. The American Naturalist 152: 311–319.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 116 Nest site selection by waterfowl in the northern hemisphere

Clark, R.G. & Nudds, T.D. 1991. Habitat patch Coulson, J.C. & Dixon, F. 1979. Colonial size and duck nesting success: the crucial breeding in sea-birds. In G. Larwood. & B.T. experiments have not been performed. Rosen (eds.), Biology and Systematics of Colonial Wildlife Society Bulletin 19: 534–543. Organisms, pp. 445–458. Academic Press, Clark, R.G. & Shutler, D. 1999. Avian habitat London, UK. selection: pattern from process in nest-site Coulton, D.W., Clark, R.G. & Wassenaar, L.L. use by ducks? Ecology 80: 272–287. 2011. Social and habitat correlates of Cody, M.L. 1985. An introduction to habitat immigrant recruitment of yearling female selection in birds. In M.L. Cody (ed.), Habitat Mallards to breeding locations. Journal of Selection in Birds, pp. 3–56. Academic Press, Ornithology 152: 781–791. Orlando, Florida, USA. Cox, G.W. 1968. The Role of Competition in the Coley, P.D., Bryant, J.P. & Chapin, F.S. 1985. Evolution of Migration. Evolution 22: 180– Resource availability and plant antiherbivore 192. defense. Science 230: 895–899. Cox, G.W. 1985. The evolution of avian Coluccy, J.M., Yerkes, T., Simpson, R., Simpson, migration systems between temperate and J.W., Armstrong, L. & Davis, J. 2008. tropical regions of the New World. The Population dynamics of breeding mallards in American Naturalist 126: 451–474. the great lakes states. Journal of Wildlife Crabtree R.L. & Wolfe, M.L. 1988. Effects Management 72: 1181–1187. of alternate prey on skunk predation of Conner, E.F., Courtney, A.C. & Yoder, J.M. waterfowl nests. Wildlife Society Bulletin 16: 2000. Individuals-area relationships: the 163–169. relationship between animal population Crabtree, R.L., Broome, L.S. & Wolfe M.L. 1989. density and area. Ecology 81: 734–748. Effects of habitat characteristics on gadwall Conover, M.R. 2007. Predator–prey dynamics: the role nest predation and nest-site selection. Journal of olfaction. Taylor and Francis, Boca Raton, of Wildlife Management 53: 129–137. Florida, USA. Dahl, T.E. 2011. Status and trends of wetlands in Cooch, E.G. 2002. Fledging size and survival in the conterminous United States 2004 to snow geese: timing is everything (or is it?). 2009. U.S. Department of the Interior, Fish Journal of Applied Statistics. 29: 143–162. and Wildlife Service, Washington D.C., USA. Cooch, E.G., Lank, D.B., Dzubin, A., Rockwell, Dalby, L., McGill, B.J. & Fox, A.D. 2014. R.F. & Cooke, F. 1991. Body size variation in Seasonality drives global-scale diversity Lesser Snow Geese: environmental plasticity patterns in waterfowl (Anseriformes) via in gosling growth rates. Ecology 72: 503– temporal niche exploitation. Global Ecology 512. and Biogeography 23: 550–562. Cooch, E.G., Jefferies, R.L., Rockwell, R.F. & Danell, K. & Sjöberg, K. 1977. Seasonal Cooke, F. 1993. Environmental change and emergence of chironomids in relation to the cost of philopatry: An example in the egglaying and hatching of ducks in a restored Lesser Snow Goose. Oecologia 93: 128–138. lake (northern Sweden). Wildfowl 28: 129– Cooch, E., Rockwell, R.F. & Brault, S. 2001. 135. Retrospective analysis of demographic Darwin, C. 1972. The Voyage of the Beagle. responses to environmental change: a Lesser Bantam Books, New York, New York, USA. Snow Goose example. Ecological Monographs Dassow, J., Eichholz, M.W., Weatherhead, P.J. & 71: 377–400. Stafford, J.D. 2012. Risk taking by upland

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 117

nesting ducks relative to the probability of nest Donovan, T.M. & Lamberson, R.H. 2001. Area- predation. Journal of Avian Biology 43: 61–67. sensitive distributions counteract negative Davis, J.B., Cox, R.R., Kaminski, R.M. & effects of habitat fragmentation on breeding Leopold, B.D. 2007. Survival of Wood duck birds. Ecology 82: 1170–1179. ducklings and broods in Mississippi and Dow, H & Fredga, S. 1985. Selection of nest sites Alabama. The Journal of Wildlife Management by a hole-nesting duck, the Goldeneye 71: 507–517. Bucephala clangula. Ibis 127: 16–30. Davis, J.B., Leopold, B.D., Kaminski, R.M. & Drever, M.C. & Clark, R.G. 2007. Spring Cox, R.R. 2009. Wood duck duckling temperature, clutch initiation date and mortality and habitat implications in duck nest success: a test of the mismatch floodplain systems. Wetlands 29: 607–614. hypothesis. Journal of Animal Ecology 76: Davis, J.B., Guillemain, M., Kaminski, R.M. 139–148. Arzel, C., Eadie, J.M. & Rees, E.C. 2014. Drever, M.C., Clark, R.G. & Derksen, C. 2012. Habitat and resource use by waterfowl in the Population vulnerability to climate change northern hemisphere in autumn and winter. linked to timing of breeding in boreal ducks. Wildfowl (Special Issue No. 4): 17–69. Global Change Biology 18: 480–492. Davis, S.K. 2005. Nest-site selection patterns Duebbert, H.F. 1966. Island nesting of the and the influence of vegetation on nest Gadwall in North Dakota. The Wilson Bulletin survival of mixed-grass prairie passerines. The 78: 12–25. Condor 107: 605–616. Duebbert, H.F. 1969. High nest density and Denton, J.C., Roy, C.L., Soulliere, G.J. & Potter, hatching success of ducks on South Dakota B.A. 2012. Change in density of duck nest CAP land. Transactions of the North American cavities at forests in the north central United Wildlife and Natural Resources Conference 34: States. Journal of Fish and Wildlife Management 218–228. 3: 76–88. Duebbert, H.F. & Kantrud, H.A. 1974. Upland Derksen, D.V. & Eldridge, W.D. 1980 Drought- duck nesting related to land use and predator displacement of Pintails to the Arctic Coastal reduction. Journal of Wildlife Management 38: Plain, Alaska. Journal of Wildlife Management 257–265. 44: 224–229. Duebbert, H.F. & Lokemoen, J.T. 1976. Duck Dessborn, L., Elmberg, J. & Englund, G. 2011. nesting in fields of undisturbed grass-legume Pike predation affects breeding success and cover. Journal of Wildlife Management 40: 39–49. habitat selection of ducks. Freshwater Biology Duebbert, H.F. & Lokemoen, J.T. 1980. High 56: 579–589. duck nesting success in a predator-reduced Dion, N., Hobson, K.A. & Larivière, S. 2000. environment. Journal of Wildlife Management Interactive effects of vegetation and 44: 428–437. predators on the success of natural and Duebbert, H.F., Lokemeon, J.T. & Sharp, D.E. simulated nests of grassland songbirds. The 1983. Concentrated nesting of mallards and Condor 102: 629–634. gadwalls on Miller Lake Island, North Dakota. Doligez, B., Cadet, C. & Danchin, E. 2003. When Journal Wildlife Management 47: 729–740. to use public information for breeding Durham, R.S. & Afton A.D. 2003. Nest-site habitat selection? The role of environmental selection and success of mottled ducks on predictability and density dependence. agricultural lands in southwest Louisiana. Animal Behaviour 66: 973–988. Wildlife Society Bulletin 31: 433–442.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 118 Nest site selection by waterfowl in the northern hemisphere

Duncan, D.C. 1986. Influence of vegetation on Eisenhauer, D.I. & Kirkpatrick, C.M. 1977. composition and density of island-nesting Ecology of the Emperor goose in Alaska. ducks. Wildlife Society Bulletin 14: 158–160. Wildlife Monographs 57: 3–62. Duncan, D.C. 1987. Nest-site distribution and Elmberg, J. & Pöysä, H. 2011. Is the risk of overland brood movements of northern nest predation heterospecifically density- pintails in Alberta. Journal of Wildlife dependent in precocial species belonging to Management 51: 716–723. different nesting guilds? Canadian Journal of DuRant, S.E., Hopkins, W.A., Hawley, Zoology 89: 1164–1171. D.M. & Hepp, G.R. 2011. Incubation Elmberg, J., Nummi, P. & Pöysä, H. 1993. Factors temperature affects multiple measure of affecting species number and density of immunocompetence in young wood ducks dabbling duck guilds in North Europe. (Aix Sponsa). Biology Letters 10: 1–6. Ecography 16: 251–260. Dwernychuk, L.W. & Boag, D.A. 1972. Ducks Elmberg, J., Nummi, P. & Pöysä, H. 1994. nesting in association with gulls – an Relationships between species number, lake ecological trap? Canadian Journal of Zoology 50: size and resource diversity in assemblages of 559–563. breeding waterfowl. Journal of Biogeography 21: Dyrcz A., Okulewicz J. & Witkowski J. 1981. 75–84. Nesting of “timid” waders in the vicinity of Elmberg, J., Pöysä, H. & Sjöberg, K. 1997. “bold” ones as an antipredator adaptation. Interspecific interactions and co-existence in Ibis 123: 542–545. dabbling ducks: observations and an Dzubin, A. & Gallop, J.B. 1972. Aspects of experiment. Oecologia 111: 129–136. breeding Mallard ecology in Canadian Elmberg, J., Folkesson, K., Guillemain, M. parkland and grassland. In R.I. Smith, J.R. & Gunnarsson, G. 2009. Putting density Palmer & T.S. Baskett (eds.), Population dependence in perspective: nest density, Ecology of Migratory Birds – a Symposium, pp. nesting phenology, and biome, all matter 113–152. United States Fish and Wildlife to survival of simulated mallard Anas Service Wildlife Report No. 2. USFWS, platyrhynchos nests. Journal of Avian Biology 40: Washington D.C., USA. 317–326. Dzus, E.H. & Clark, R.G. 1997. Overland travel, Ely, C.R. & Raveling, D.G. 1984. Breeding food abundance, and wetland use by biology of Pacific White-Fronted Geese. Mallards: relationships with offspring Journal of Wildlife Management 48: 823–837. survival. Wilson Bulletin 109: 504–515. Ely, C.R. & Scribner, K.T. 1994. Genetic diversity Earl, J.P. 1950. Production of mallards on irrigated in Arctic-nesting geese: implications for land in the Sacramento Valley, California. management and conservation. Transactions of Journal of Wildlife Management 14: 332–342. the North American Wildlife and Natural Eichholz, M.W. & Sedinger, J.S. 1999. Regulation Resources Conference 59: 91–110. of incubation behavior in Black Brant. Erikstad, K.E. & Tveraa, T. 1995. Does the cost Canadian Journal of Zoology 77: 249–257. of incubation set limits to clutch size in Eichholz, M.W., Dassow, J.A., Weatherhead, P.J. Common Eiders Somateria mollissima? Oecologia & Stafford, J.D. 2012. Experimental evidence 103: 270–274. that nesting ducks use mammalian urine to Esler, D. & Grand, J.B. 1993. Factors influencing assess predator abundance. The Auk 129: depredation of artificial duck nests. Journal of 638–664. Wildlife Management 57: 244–248.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 119

Evans, K.L., Warren, P.H. & Gaston, K.J. 2005. predator feces as a cue in avian habitat Species-energy relationships at the selection decisions. Behavioural Ecology 10: 1–5. macroecological scale: a review of the Fournier, M. A. & Hines, J. E. 2001. Breeding mechanisms. Biological Review 80: 1–25. ecology of sympatric Greater and Lesser Fabricius, E. & Norgren, H. 1987. Lär känna Scaup (Aythya marila and Aythya affinis) in the Kanadagåsen. Svenska Jägareförbundet [Swedish Subarctic Northwest Territories. Arctic 54: Sportsmens’ Association], Stockholm, Sweden. 444–456. Fast, P.L., Gilchrist, G. & Clark, R.G. 2007. Fowler, A.C. 2005. Fine-scale spatial structuring Experimental evaluation of nest shelter in cackling Canada geese related to effects on weight loss in incubating common reproductive performance and breeding eiders Somateria mollissima. Journal of Avian philopatry. Animal Behaviour 69: 973–981. Biology 38: 205–213. Fowler, A.C., Eadie, J.M. & Ely, C.R. 2004. Fast, P.L.F., Grant G.H. & Clark, R.G. 2010. Relatedness and nesting dispersion within Nest-site materials affect nest-bowl use breeding populations of greater white- by Common Eiders Somateria mollissima. fronted geese. The Condor 106: 600–607. Canadian Journal of Zoology 88: 214–218. Fox, T., Eide, N.E. & Bergersen, E. 2009. Fisher, R.J. & Davis, S.H. 2010. From Wiens to Resource partitioning in sympatric arctic- Robel: A review of procedures and patterns breeding geese: summer habitat use, spatial describing grassland bird habitat selection. and dietary overlap of Barnacle and Pink- Journal of Wildlife Management 74: 265–273. footed Geese in Svalbard. Ibis 151: 122–133. Fishman, D.J., Craik, S.R., Zadworny, D. & Fretwell, S.D. 1972. Populations in a Seasonal Titman, R.D. 2011. Spatial-genetic structuring Environment. Princeton University Press, in a red-breasted merganser colony in the Princeton, New Jersey, USA. Canadian Maritimes. Ecology and Evolution 1: Fretwell, S.D. 1980. Evolution of migration in 107–118. relation to factors regulating bird numbers. In Flint, P.L., Grand, J.B. & Rockwell, R.F. 1998. A A. Keast & E.S. Morton (eds.), Migrant Birds model of northern pintail productivity and in the Neotropics: Ecology, Behavior, Distribution, population growth rate. Journal of Wildlife and Conservation, pp. 517–527. Smithsonian Management 62: 1110–1118. Institute Press, Washington D.C., USA. Flint, P.L., Grand, J.B. Fondell, T.F. & Morse, Fretwell, S.D. & Lucas, H.L. 1969. On territorial J.A. 2006. Population dynamics of Greater behavior and other factors influencing Scaup breeding on the Yukon-Kuskokwim distribution in birds. I. Theoretical Delta, Alaska. Wildlife Monographs 162: 1– development. Acta Biotheoretica 19: 1–20. 21. Fritzell, E.K. 1975. Effects of agricultural Fontaine, J. J. & Martin, T.E. 2006. Parent birds burning on nesting waterfowl. Canadian Field- assess nest predation risk and adjust their Naturalist 89: 21–27. reproductive strategies. Ecology Letters 9: 428– Garrettson, P.R. & Rohwer, F.C. 2001. Effects of 434. mammalian predator removal on production Forbes, S.L. & Kaiser, G.W. 1994. Habitat choice of upland-nesting ducks in North Dakota. in breeding seabirds: when to cross the Journal of Wildlife Management 65: 398–405. information barrier. Oikos 70: 377–384. Gates, J.E. & Gysel, L.W. 1978. Avian nest Forsman, J.T, Mönkkönen, M., Korpimäki, E. & dispersion and fledging success in field- Thomson, R.L. 2012. Mammalian nest forest ecotones. Ecology 59: 871–883.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 120 Nest site selection by waterfowl in the northern hemisphere

Gauthier, G., Bety, J., Giroux, J.F. & Rochefort, Gratto, C.L., Morrison, R.I.G. & Cooke, F. 1985. L. 2004. Trophic interactions in a high Philopatry, site tenacity and mate fidelity in arctic Snow Goose Colony. Integrative and the semipalmated sandpiper. The Auk. 102: Comparative Biology 44: 119–129. 16–24. Gilbert, D.W., Anderson, D.R., Ringelman, J.K. Greenwood, P.J. 1982. The natal and breeding & Szymczak, M.R. 1996. Response of dispersal of birds. Annual Review of Ecology nesting ducks to habitat and management and Systematics 13: 1–21. on the Monte Vista National Wildlife Greenwood, R.J., Sargeant A.B., Johnson D.H., Refuge, Colorado. Wildlife Monographs 131: 3– Cowardin L.M. & Shaffer, T.L. 1995. Factors 44. associated with duck nest success in the Gilmer, D.S., Ball, I.J. & Mathisen, J.E. 1978. prairie pothole region of Canada. Wildlife Natural cavities used by Wood Ducks in Monographs 128: 3–57. North-Central Minnesota. Journal of Wildlife Greenwood, R.J., Pietruszewski, D.G. & Management 42: 288–298. Crawford, R.D. 1998. Effects of food Gloutney, M.L. & Clark R.G. 1997. Nest-site supplementation on depredation of duck selection by mallards and blue-winged teal in nests in upland habitat. Wildlife Society Bulletin relation to microclimate. The Auk 114: 381– 26: 219–226. 395. Grinnell, J. 1914. Barriers to distribution as Glover, F.A. 1956. Nesting and production of the regards birds and mammals. The American blue-winged teal (Anas discors Linnaeus) in Naturalist 48: 248–254. northwest Iowa. Journal of Wildlife Management Gross, A.O. 1945. The black duck nesting on the 20: 28–46. outer coastal islands of Maine. The Auk 62: Gosser, A.L. & Conover, M.R. 1999. Will 620–622. the availability of Insular nesting sites Gunnarsson, G. & Elmberg, J. 2008. Density- limit reproduction in urban Canada goose dependent nest predation – an experiment populations? Journal of Wildlife Management with simulated Mallard nests in contrasting 63: 369–373. landscapes. Ibis 150: 259–269. Götmark, F. & Åhlund, M. 1988. Nest predation Haffele, R.D. 2012. Nesting ecology of ducks and nest site selection among Eiders Somateria in dense nesting cover and restored mollissima: the influence of gulls. Ibis 130: native plantings in northeastern North 111–123. Dakota. MSc. thesis. Southern Illinois Grand, J.B. & Flint, P.L. 1996. Survival of University Carbondale, Carbondale, Northern Pintail Ducklings on the Yukon- USA. Kuskokwim Delta, Alaska. The Condor 98: Haffele, R.D., Eichholz, M.W. & Dixon, C.S. 48–53. 2013. Duck productivity in restored species – Grand, J.B. & Flint, P.L. 1997. Productivity of rich native and species-poor non-native nesting spectacle Eiders on the lower plantings. PLoS ONE 8: e68603. doi:10.1371/ Kashunuk River, Alaska. The Condor 99: journal.pone.0068603. 926–932. Hammond, M.C. & Mann, G.E. 1956. Waterfowl Grant, T.A. & Shaffer, T.L. 2012. Time-specific nesting islands. Journal of Wildlife Management patterns of nest survival for ducks and 20: 345–352. passerines breeding in North Dakota. The Hanski, I., Hansson, L. & Henttonen, H. 1991. Auk 129: 319–328. Specialist predators, generalist predators and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 121

the microtine rodent cycle. Journal of Animal Hines, J.E. & Mitchell, G.J. 1983. Gadwall nest- Ecology 60: 353–367. site selection and nesting success. Journal of Harrison, S. & Bruna, E. 1999. Habitat Wildlife Management 47: 1063–1071. fragmentation and large-scale conservation: Hoekman, S.T., Mills, L.S., Howerter, D.W., what do we know for sure? Ecography 22: Devries, J.H. & Ball, I.J. 2002. Sensitivity 225–232. analysis of the life cycle of midcontinent Harvey, P.H., Greenwood, P.J. & Perrins, C.M. mallards. Journal of Wildlife Management 66: 1979. Breeding area fidelity of great tits (Parus 883–900. major). Journal of Animal Ecology 48: 305–313. Holling, C.S. 1965. The functional response of Haszard, S. & Clark, R.G. 2007. Wetland use by predators to prey density and its role in mimicry White-winged Scoters (Melanitta fusca) in the and population regulation. Entomological Society Mackenzie Delta Region. Wetlands 27: 855– of Canada Memoirs S45: 5–60. 863. Holt, R.D. 1977. Predation, apparent competition, Hawkins, B.A., Field, R. & Cornell, H.V. 2003. and the structure of prey communities. Energy, water, and broad-scale geographic Theoretical and Population Biology 12: 197–229. patterns of species richness. Ecology 84: Holt, R.D. 2008. Theoretical perspectives on 3105–3117. resource pulses. Ecology 89: 671–681. Hawkins, B.A. & Porter, E.E. 2003. Relative Holt, R.D. & Lawton, J.H. 1994. The ecological influence of current and historical factors on consequences of shared natural enemies. mammal and bird diversity patterns in Annual Review of Ecology and Systematics 25: deglaciated North America. Global Biology & 495–520. Biogeography. 12: 475–481. Horn, D.J., Phillips, M.L. & Korford, R.P. Hepp, G.R. & Kennamer, R.A. 1992. 2005. Landscape composition, patch size and Characteristics and consequences of nest-site distance to edges: interactions affecting duck fidelity in wood ducks. The Auk 109: 812– reproductive success. Ecological Applications 818. 15: 1367–1376. Hepp, G.R, Kennamer, R.A & Johnson, M.H. Hortal, J., Diniz-Filho, J.A., Bini, L.M., Rodríguez, 2006. Maternal effects in wood ducks: M.Á., Baselga, A., Nogués-Bravo, D., Rangel, incubation temperature influences incubation T.F., Hawkins, B.A. & Lobo, J.M. 2011. Ice age period and neonate phenotype. Functional climate, evolutionary constraints and diversity Ecology 20: 307–314. patterns of European dung beetles. Ecological Higgins, K.F. 1977. Duck nesting in intensively Letters 14: 741–748. farmed areas of North Dakota. Journal of Hurlbert, A.H. & Stegen, J.C. 2014. When should Wildlife Management 41: 232–242. species richness be energy limited, and how Hildén, O. 1965. Habitat selection in birds: a would we know? Ecology Letters 17: 401–413. review. Annales Zoologici Fennici 2: 53–75. Hurlbert, A.H. & Haskell, J.P. 2003. The effect of Hinde, R.A. 1956. The biological significance of energy and seasonality on avian species the territories of birds. Ibis 98: 340–369. richness and community composition. The Hilton, G.M., Hansell, M.H., Ruxton, G.D., Reid, American Naturalist 161: 83–97. J.M. & Monaghan, P. 2004. Using artificial Hutto, R.L. 1985. Habitat selection by nests to test importance of nesting material nonbreeding, migratory land birds. In M.L. and nest shelter for incubation energetics. Cody (ed.), Habitat Selection in Birds, pp. 455– The Auk 121: 777–787. 476. Academic Press, Orlando, Florida, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 122 Nest site selection by waterfowl in the northern hemisphere

Iles, D.T., Rockwell, R.F. & Matulonis, P. 2013. Keith, L.B. 1961. A study of waterfowl ecology Predators, alternative prey and climate on small impoundments in southeastern influence annual breeding success of a long- Alberta. Wildlife Monographs 6: 3–88. lived sea duck. Journal of Animal Ecology 82: Kellett, D.K. & Alisauskas, R.T. 1997. Breeding 683–693. biology of King Eiders nesting on Karrak Jane, S.D. & Bowmaker, J.K. 1988. Tetrachromatic Lake, Northwest Territories. Arctic 50: 47– colour vision in the duck (Anas platyrynchos L.): 54. Microspectophotometry of visual pigments Klomp, H. 1953. De Terreinkeus van de Kievit, and oil droplets. Journal of Comparative Vanellus vanellus (L.). Ardea 41: 1–139. Physiology A 162: 225–235. Klug, P.E., Wolfenbarger, L.L. & McCarty, J.P. Jarvis, R.L. & Harris, S.W. 1967. Canada goose 2009. The nest predator community of nest success and habitat use at Malheur grassland birds responds to agroecosystem Refuge. Murrelet 48: 46–51. habitat at multiple scales. Ecography 32: Johnson, D.H. 1980. The comparison of usage 973–982. and availability measurements for evaluating Koivula, M. & Viitala, J. 1999. Rough-legged resource preference. Ecology 61: 65–71. Buzzards use of vole scent marks to assess Johnson, D.H. & Shaffer. T.L. 1987. Are mallards hunting areas. Journal of Avian Biology 30: declining in North America? Wildlife Society 329–332. Bulletin 15: 340–345. Korshgen, C.E. 1977. Breeding stress of female Johnston, D.W. & Odum, E.P. 1956. Breeding bird eiders in Maine. Journal of Wildlife Management populations in relation to plant succession 41: 360–373. on the piedmont of Georgia. Ecology 37: 50– Korpimäki, E. 1988. Effects of territory quality 62. on occupancy, breeding performance and Kaminski, R.M. & Elmberg, J. 2014. An breeding dispersal in Tengmalm’s Owl. introduction to habitat use and selection Journal of Animal Ecology 57: 97–108. by waterfowl in the northern hemisphere. Kossack, C.W. 1950. Breeding habits of Canada Wildfowl (Special Issue No. 4): 9–16. geese under refuge conditions. American Kaminski, R.M. & Prince, H.H. 1984. Dabbling Midland Naturalist 43: 627–649. duck-habitat associations during spring in Krapu, G.L. & Reinecke, K.J. 1992. Foraging Delta Marsh, Manitoba. Journal of Wildlife ecology and nutrition. In B.D.J. Batt, A.D. Management 48: 37–50. Afton, M.G. Anderson, C.D. Ankney, D.H. Kaminski, R.M. & Weller, M.W. 1992. Breeding Johnson, J.A. Kadlec & G.L. Krapu (eds.), habitats of Nearctic waterfowl. In B.D.J. Batt, Ecology and Management of Breeding Waterfowl, A.D. Afton, M.G. Anderson, C.D. Ankney, pp. 1–29. University of Minnesota Press, D.H. Johnson, J.A. Kadlec & G.L. Krapu Minneapolis, USA. (eds.), Ecology and Management of Breeding Krapu, G.L., Klett. A.T. & Jorde, G.D. 1983. The Waterfowl, pp. 568–589. University of effect of variable spring water conditions on Minnesota Press, Minneapolis, USA. Mallard reproduction. The Auk 100: 689–698. Kaspari, M., Stevelson, B.S. & Shik, J. 2010. Krapu, G.L., Pietz, P.J., Brandt, D.A. & Cox, R.R., Scaling community structure: how bacteria, Jr. 2006. Mallard brood movements, wetland fungi, and ant taxocenes differentiate along use, and duckling survival during and a tropical forest floor. Ecology 91: 2221– following a prairie drought. Journal of Wildlife 2226. Management 70: 1436–1444.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 123

Krasowski, T. P. & Nudds, T.D. 1986. Canada. Transactions of the North American Microhabitat structure of nest sites and Wildlife and Natural Resources Conference 16: nesting success of diving ducks. Journal of 94–99. Wildlife Management 50: 203–208. Leopold, A. 1933. Game Management. University of Kristan, W.B. 2006. Sources and expectations for Wisconsin Press, Madison, Wisconsin, USA. hierarchical structure in bird-habitat Lima, S.L. 2009. Predators and the breeding bird: associations. The Condor 108: 5–12. behavioral and reproductive flexibility under Kuehl, A.K. & Clark, W.R. 2002. Predator activity the risk of predation. Biological Reviews 84: related to landscape features in northern Iowa. 485–513. Journal of Wildlife Management 66: 1224–1234. Lindberg, M.S. & Sedinger, J.S. 1997. Ecological Lack, D. 1933. Habitat selection in birds consequences of nest site fidelity in Black with special reference to the effects of Brant. The Condor 99: 25–38. afforestation on the breckland avifauna. Lindholm, A., Gauthier, G. & Desrochers, A. Journal of Animal Ecology 2: 239–262. 1994. Effects of hatch date and food supply Lack, D. 1954. The Natural Regulation of Animal on gosling growth in Arctic-nesting greater Numbers. Clarendon Press, London, UK. snow geese. The Condor 96: 898–908. Lack, D. 1965. Evolutionary ecology. Journal of Livezey, B.C. 1981. Location and success of duck Animal Ecology 34: 223–231. nests evaluated through discriminant analysis. Lack, D. 1971. Ecological Isolation in Birds. Wildfowl 32: 23–27. Blackwell Scientific Publishing, Oxford, UK. Lokemoen, J.T., Duebbert, H.F. & Sharp, D.E. Larivière S. & Messier, F. 1998. Effect of density 1990. Homing and reproductive habits of and nearest neighbours on simulated mallards, gadwalls, and blue-winged teals. waterfowl nests: can predators recognize Wildlife Monographs 106: 1–28. high density nesting patches. Oikos 83: 12–20. Lowney, M.S. & Hill, E.P. 1989. Wood Duck nest Larivière, S. & Messier. F. 2000. Habitat selection sites in bottomland hardwood forests of and use of edges by striped skunks in the Mississippi. Journal of Wildlife Management 53: Canadian prairies. Canadian Journal of Zoology 378–382. 78: 366–372. Low, J.B. 1941. Nesting of the ruddy duck in Larivière, S. & Messier, F. 2001. Space-use Iowa. The Auk 58: 506–517. patterns by female striped skunks exposed MacArthur, R.H., Diamond, J.M. & Karr, J.R. to aggregations of simulated duck nests. 1972. Density compensation in island faunas. Canadian Journal of Zoology 79: 1604–1608. Ecology 53: 330–342. Larsson, K. & Forslund P. 1991. Environmental MacInnes, C.D. & Dunn, E.H. 1988. Components induced morphological variation in the of clutch size variation in arctic-nesting Barnacle Goose, Branta leucopsis. Journal of Canada Geese. The Condor 90: 83–89. Environmental Biology 4: 619–636. Mack, G.G., Clark, R.G. & Howerter, D.W. 2003. Lecomte, N., Gauthier, G. & Giroux, J.F. Size and habitat composition of female 2008. Breeding dispersal in a heterogeneous mallard home ranges in the prairie-parkland landscape: the influence of habitat and nesting region of Canada. Canadian Journal of Zoology success in greater snow geese. Oecologia 155: 81: 1454–1461. 33–41. Magee, P.A. 1993. Detrital accumulation and Leitch, W.G. 1951. Saving, maintaining and processing in wetlands. In D.H. Cross & P. developing waterfowl habitat in western Vohs (eds.), Waterfowl Management Handbook,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 124 Nest site selection by waterfowl in the northern hemisphere

pp. 3–14. Fish and Wildlife Leaflet No. 13. Murphy, G.E.P. & Romanuk, T.N. 2014. A meta- U.S. Department of the Interior, Fish and analysis of declines in local species richness Wildlife Service, Washington D.C., USA. from human disturbances. Ecology and Major, R.E. & Kendal, C.E. 1996. The Evolution 4: 91–103. contribution of artificial nest experiments to Newton, I. & Marquiss, M. 1982. Fidelity to understanding avian reproductive success: a breeding area and mate in Sparrowhawks review of methods and conclusions. Ibis 138: Accipiter nisus. Journal of Animal Ecology 51: 298–307. 327–341. Martin, T.E. 1998. Are microhabitat preferences Nicolai, C.A., Flint, P.L. & Wege, M.L. 2005. of coexisting species under selection and Annual survival and site fidelity of Northern adaptive? Ecology 79: 656–670. pintail banded on the Yukon-Kuskokwim McLandress, M.R. 1983. Temporal changes in Delta, Alaska. Journal of Wildlife Management habitat selection and nest spacing in a colony 69: 1202–1210. of Ross’ and lesser snow geese. The Auk 100: Nicolai, C.A., Sedinger, J.S. & Wege, M.L. 2008. 335–343. Differences in growth of Black Brant McRoberts, J.T., Quintana, N.T. & Smith, A.W. goslings between a major breeding colony 2012. Greater Scaup, Aythya marila, nest and outlying breeding aggregations. The site characteristics on grassy islands, New Wilson Journal of Ornithology 120: 755– Brunswick. Canadian Field-Naturalist 126: 766. 15–19. Nichols, J.D., Conley, W., Batt, B. & Tipton, A.R. Mickelson, P.G. 1975. Breeding biology of 1976. Temporally dynamic reproductive cackling geese and associated species on the strategies and the concept of R- and K- Yukon-Kuskokwin Delta, Alaska. Wildlife Selection. The American Naturalist 110: 995– Monographs 45: 3–35. 1005. Miller, D.A., Grand, J.B., Fondell, T.F. & Anthony Nielsen, C.L. & Gates, R.J. 2007. Reduced R.M. 2007. Optimizing nest survival and female nest predation of cavity-nesting wood survival: consequences of nest site selection ducks during flooding in a bottomland for Canada geese. Condor 109: 769–780. hardwood forest. The Condor 109: 210– Milne, H. 1976. Body weights and carcass 215. composition of the Common Eider. Wildfowl Nielsen, C.L., Gates, R.J. & Zwicker, E.H. 2007. 27: 115–122. Projected availability of natural cavities for Moles, A.T., Bonser, S.P. & Poore, A.G.B. 2011. Wood Ducks in Southern Illinois. Journal of Assessing the evidence for latitudinal Wildlife Management. 71: 875–883. gradients in plant defence and herbivory. Nilsson, S.G. 1986. Evolution of hole-nesting in Functional Ecology 25: 380–388. birds: on balancing selection pressures. The Morris, D.W. 2003. Toward an ecological synthesis: Auk 103: 432–435. a case for habitat selection. Oecologia 136: 1–13. North American Waterfowl Management Plan, Morrison, W.E. & Hay, M.E. 2012. Are lower- Plan Committee. 2012. North American latitude plants better defended? Palatability of Waterfowl Management Plan 2012. People freshwater macrophytes. Ecology 91: 65–74. conserving waterfowl and wetlands. Mulhern, J.H., Nudds, T.D. & Neal, B.R. 1985. Canadian Wildlife Service, U.S. Fish and Wetland selection by Mallards and Blue- Wildlife Service, Secretaria de Medio winged teal. Wilson Bulletin 97: 473–485. Ambiente y Recursos Naturales.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 125

Nudds, T.D. 1983 Niche dynamics and breeding dispersal in a female-philopatric organization of waterfowl guilds in variable species. Oecologia 166: 327–336. environments. Ecology 64: 319–330. Pasitschniak-Arts, M. & Messier, F. 1996. Nudds, T.D. 1992. Patterns in breeding duck Predation on artificial duck nests in a communities. In B.D.J. Batt, A.D. Afton, fragmented prairie landscape. Ecoscience 3: M.G. Anderson, C.D. Ankney, D.H. Johnson, 436–441. J.A. Kadlec & G.L. Krapu (eds.), Ecology and Pasitischniak-Arts, M., Clark, R.G. & Messier, F. Management of Breeding Waterfowl, pp. 540–567. 1998. Duck nesting success in a fragmented University of Minnesota Press, Minneapolis, prairie landscape: is edge effect important? USA. Biological Conservation 85: 55–62. Oberholser, H.C. & McAtee, W.L. 1920. Paasitischniak-Arts, N. & Messier F. 1998. Waterfowl and their food plants in the Effects of edges and habitats on small sandhill region of Nebraska. United States mammals in a prairie ecosystem. Canadian Department of Agriculture Bulletin No. 794. Journal of Zoology 76: 2020–2025. U.S. Department of Agriculture, Washington Parker, H. & Holm, H. 1990. Patterns of nutrient D.C., USA. and energy expenditure in female common Oetting, R.B. & Dixon, C.D. 1975. Waterfowl nest eiders nesting in the high Arctic. The Auk densities and success at Oak Hammock Marsh, 107: 660–668. Manitoba. Wildlife Society Bulletin 3: 166–171. Paton, P.W.C. & Edwards, T.C. 1996. Factors O’Neill, R.V., D.L. Deangelis, D.L., Wade, J.B. & affecting interannual movements of snowy Allen, T.F.H. 1986. A Hierarchical Concept of plovers. The Auk 113: 534–543. Ecosystems. Princeton University Press, Pearse, A.T., Krapu, G.L., Brandt, D.A. & Kinzel, Princeton, New Jersey, USA. P.J. 2010. Changes in agriculture and O’Neil, T.A. 1988. Controlled pool elevation and abundance of snow geese affect carrying its effects on Canada goose productivity and capacity of sandhill cranes in Nebraska. nest location. The Condor 90: 228–232. Journal of Wildlife Management 74: 479–488. Öst, M. & Steele, B.B. 2010. Age-specific nest- Pearse, A.T. & Ratti, J. T. 2004. Effects of predator site preference and success in eiders. Oecologia removal on Mallard duckling survival. Journal 162: 59–69. of Wildlife Management 68: 342–350. Öst, M., Ydenberg, R., Kilpi, M. & Lindström, Pehrsson, O. 1986. Duckling production of the K. 2003. Condition and coalition formation Oldsquaw in relation to spring weather and by brood-rearing common eider females. small rodent fluctuations. Canadian Journal of Behavioural Ecology 14: 311–317. Zoology 64: 1835–l841. Öst, M., Wickman, M., Matulionis, E. & Steele, B. Pennings, S.C., Siska, E.L. & Bertness, M.D. 2008a. Habitat-specific clutch size and cost 2001. Latitudinal differences in plant of incubation in eiders reconsidered. palatability in Atlantic Coast salt marshes. Oecologia 158: 205–216. Ecology 82: 1344–1359. Öst, M., Smith, B.D. & Kilpi, M. 2008b. Social Pennings, S.C., Zimmer, M. & Dias, N., Sprung, and maternal factors affecting duckling M., Davé, N., Ho, C.K., Kunza, A., McFarlin, survival in eiders Smateria mollissima. Journal of C., Mews, M., Pfauder, A. & Salgado, C. Animal Ecology 77: 315–325. 2007. Latitudinal variation in plant-herbivore Öst, M., Lehikoinen, A., Jaatinen, K. & Kilpi, M. interactions in European salt marshes. Oikos 2011. Causes and consequences of fine-scale 116: 543–549.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 126 Nest site selection by waterfowl in the northern hemisphere

Person, B.T., Herzog, M.P. & Ruess, R.W. Pöysä H. & Pöysä, P. 2002. Nest-site limitation 2003. Feedback dynamics of grazing lawns: and density dependence of reproductive coupling vegetation change with animal output in the common goldeneye Bucephala growth. Oecologia 135: 583–592. clangula: implications for the management of Petersen, M.R. 1990. Nest-site selection by cavity-nesting birds. Journal of Applied Ecology emperor geese and cackling Canada geese. 39: 502–510. Wilson Bulletin 102: 413–426. Pöysä, H., Elmberg, J. & Sjöberg, K. 1998. Phillips, M.L., Clark, W.R. & Sovada, M.A. 2003. Habitat selection rules in breeding mallards Predator selection of prairie landscape (Anas platyrhynchos): a test of two competing features and its relation to duck nest success. hypotheses. Oecologia 114: 283–287. Journal of Wildlife Management 67: 104–114. Pöysä, H., Milonoff, M., Ruusila, V. & Virtanen, J. Phillips, M.L., Clark, W.R. & Nusser, S.M. 1999. Nest site selection in relation to habitat 2004. Analysis of predator movement in edge: experiments in the common goldeneye. prairie landscapes with contrasting grassland Journal of Avian Biology 30: 79–84. composition. Journal of Mammalogy 85: 187– Pöysä, H., Elmberg, J., Sjöberg, K. & Nummi, P. 195. 2000. Nesting mallards (Anas platyrhynchos) Piersma, T. 1997. Do global patterns of habitat forecast brood-stage food limitation when use and migration strategies co-evolve with selecting habitat: experimental evidence. relative investments in immunocompetence Oecologia 122: 582–586. due to spatial variation in parasite pressure? Prince, H.H. 1968. Nest sites used by Wood Ducks Oikos 80: 623–631. and Common Goldeneyes in New Brunswick. Piersma, T., Lindstrom, A., Drent, R.H., Tulp, Journal of Wildlife Management 32: 489–500. I., Jukema, J., Morrison, T.I.G, Reneerkens, Probst, R., Pavlicev, M. & Viitala. J. 2002. UV J., Schekkerman, H. & Visser G.H. 2003. reflecting vole scent marks attract a passerine, High daily energy expenditure of incubating the Great Grey Shrike Lanius excubitor. Journal shorebirds on high arctic tundra: a of Avian Biology 33: 437–440. circumpolar study. Functional Ecology 17: Quakenbush, L., Suydam, R. & Deering, M. 2004. 356–362. Breeding biology of Steller’s Eiders near Pirnie, M.D. 1935. Michigan Waterfowl Management. Barrow, Alaska, 1991–99. Arctic 57: 166–182. Franklin DeKleine Company, Lansing, Rangen, S.A., Clark, R.G. & Hobson. K.A. 2000. Michigan, USA. Visual and olfactory attributes of artificial Pilotte, C., Reed, E.T., Rodrigue, J. & Giroux, J.F. nests. The Auk 117: 136–146. 2014. Factors influencing survival of Canada Rappole, J.H. & Jones, P. 2002. Evolution of old geese breeding in southern Quebec. The and new world migration systems. Ardea 90: Journal of Wildlife Management 78: 231–239. 925–937. Pöysä, H. 2006. Public information and Raupp, M.J. & Denno, R.F. 1979. The influence conspecific nest parasitism in goldeneyes: of patch size on a guild of Sapfeeding insects targeting sage nests by parasites. Behavioural that inhabit the Salt Marsh Grass Spartina Ecology 10: 459–460. patens. Entomological Society of America 8: 412– Pöysä, H. & Paasivaara, A. 2006. Movements and 417. mortality of common goldeneye Bucephala Raveling, D.G. 1989. Nest-predation rates in clangula broods in a patchy environment. relation to colony size of Black Brant. Journal Oikos 115: 33–42. of Wildlife Management 53: 87–90.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 127

Reynolds, R.E., Shaffer, T.L., Renner, R.W. 2001. Root, R.B. 1973. Organization of a plant- Impact of the conservation reserve program arthropod association in simple and diverse on duck recruitment in the U.S. prairie habitats: the fauna of collards (Brassica pothole region. Journal of Wildlife Management oleracea). Ecological Monographs 43: 95–124. 65: 765–780. Rotella J.J. & Ratti, J.T. 1992. Mallard brood Ribic, C.A., Koford, R.R., Herkert, J.R., Johnson, movements and wetland selection in D.H., Niemuth, N.D., Naugle, D.E, Bakker, southwestern Manitoba. Journal of Wildlife K.K., Sample, D.W. & Renfrew, R.B. 2009. Management 56: 508–515. Area sensitivity in North American grassland Roth, R.R. 1976. Spatial heterogeneity and bird birds: patterns and processes. The Auk 126: species diversity. Ecology 57: 773–782. 233–244. Roy, C., Eadie, J.M. & Schauber, E.M. 2009. Ricklefs, R.E. 1969. An analysis of nesting Public information and conspecific nest mortality in birds. Smithsonian Contributions to parasitism in wood ducks: does nest density Zoology 9: 1–48. influence quality of information? Animal Ringelman, K.M. 2014. Predator foraging Behaviour 77: 1367–1373. behavior and patterns of avian nest success: Ryder, J.P. 1972. Biology of nesting Ross’ Geese. What can we learn from an agent-based Ardea 60: 185–215. model? Ecological Modeling 272: 141–149. Sachs, J.L., Hughes, C.R., Neuchterlein, G.L. & Ringelman, K.M., Eadie, J.M. & Ackerman, J.T. Buitron, D. 2007. Evolution of coloniality in 2012. Density-dependent nest predation in birds: a test of hypotheses with the Red- waterfowl: the relative importance of nest necked Grebe (Podiceps grisegena). The Auk density versus nest dispersion. Oecologia 169: 124: 628–642. 695–702. Safine, D.E. & Lindberg, M.S. 2008. Nest habitat Robb, J.R. & Bookhout, T.A. 1995. Factors selection of White-winged scoters on Yukon influencing Wood Duck use of natural Flats, Alaska. The Wilson Journal of Ornithology cavities. Journal of Wildlife Management 59: 120: 582–593. 372–383. Sargeant A.B. & Raveling D.G. 1992. Mortality Robbins, C.S., Dawson, D.K. & Dowell, B.A. during the breeding season. In B.D.J. Batt, 1989. Habitat area requirements of breeding A.D. Afton, M.G. Anderson, C.D. Ankney, forest birds of the middle Atlantic states. D.H. Johnson, J.A. Kadlec & G.L. Krapu Wildlife Monographs 103: 3–34. (eds.), Ecology and Management of Breeding Robert, M., Vaillancourt, M. & Drapeau, P. 2010. Waterfowl, pp. 396–422. University of Characteristics of nest cavities of Barrow’s Minnesota Press, Minneapolis, USA. Goldeneyes in eastern Canada. Journal of Sargeant, A.B., Allen, S.H. & Eberhardt, R.T. Field Ornithology 81: 287–293. 1984. Red fox predation on breeding ducks Rolland, C., Danchin, E. & de Fraipoint, M. 1998. in midcontinent North America. Wildlife The evolution of coloniality in birds in Monographs 89: 3–41. relation to food, habitat, predation, and life- Savard, J.P. 1988. Use of nest boxes by Barrow’s history traits: a comparative analysis. The Goldeneyes: nesting success and effect on American Naturalist 151: 514–529. the breeding population. Wildlife Society Romesburg, H.C. 1981. Wildlife science: gaining Bulletin 16: 125–132. reliable knowledge. Journal of Wildlife Schamber, J.L., Flint, P.L. & Grand, J.B. 2009. Management 45: 293–313. Population dynamics of Long-tailed Ducks

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 128 Nest site selection by waterfowl in the northern hemisphere

breeding on the Yukon-Kuskokwin Delta, Brant (Branta bernicla nigricans) into the Alaska. Arctic 62: 190–200. breeding population. The Auk 121: 68–73. Schmidt, J.H., Lindberg, M.S., Johnson, D.S. & Shutler, D., Gloutney, M.L. & Clark, R.G. 1998. Verbyla, D.L. 2011. Season length influences Body mass, energetic constraints, and duck breeding range dynamics of trumpeter swans nesting ecology. Canadian Journal of Zoology Cygnus buccinator. Wildlife Biology 17: 364–372. 76: 1805–1814. Schmutz, J.K., Robertson, R.J. & Cooke, F. 1983. Siegel-Causey, D. & Kharitonov, S.P. 1990. The Colonial nesting of the Hudson Bay eider evolution of coloniality. Current Ornithology 7: duck. Canadian Journal of Zoology 61: 2424– 285–330. 2433. Sittler, B., Olivier, G. & Berg, T.B. 2000. Low Schlaepfer, M.A., Runge, M.C. & Sherman, P.W. abundance of king eider nests during low 2002. Ecological and evolutionary traps. lemming years in Northeast Greenland. Trends in Ecology and Evolution 17: 474–480. Arctic 53: 53–60. Schrank, B.W. 1972. Waterfowl nest cover and Simpson, J.W., Yerkes, T.J., Smith, B.D. & Nudds, some predation relationships. Journal of T.D. 2005. Mallard duckling survival in the Wildlife Management 36: 182–186. Great Lakes region. The Condor 107: 898–909. Schummer, M.L., Kaminski, R.M., Raedeke, Sliwinski, M.S. & Koper, N. 2012. Grassland bird A.H. & Graber, D.A. 2010. Weather-related responses to three edge type in a fragmented indices for autumn-winter dabbling duck mixed grass prairie. Avian Conservation and abundance in middle North America. Journal Ecology 7: 6. of Wildlife Management 94: 94–101. Sonsthagen, S.A., Talbot, S.L., Lanctot, R.B. & Sedinger, J.S. 1992. Ecology of prefledging Mccracken, K.G. 2010. Do common eiders waterfowl. In B.D.J. Batt, A.D. Afton, M.G. nest in kin groups? Microgeographic genetic Anderson, C.D. Ankney, D.H. Johnson, J.A. structure in a philopatric sea duck. Molecular Kadlec & G.L. Krapu (eds.), Ecology and Ecology 19: 647–657. Management of Breeding Waterfowl, pp. 109–127. Sovada, M.A., Zicus, M.C. & Greenwood, R.J. University of Minnesota Press, Minneapolis, 2000. Relationships of habitat patch size to USA. predator community and survival of duck Sedinger, J.S. & Flint, P.L. 1991. Growth rate is nests. Journal of Wildlife Management 64: 820– negatively correlated with hatch date in black 831. brant. Ecology 72: 496–502. Soulliere, G.J. 1988. Density of suitable wood Sedinger, J.S., Flint, P.L. & Lindberg, M.S. 1995. duck nest cavities in a Northern Hardwood Environmental influence on life-history Forest. Journal of Wildlife Management. 52: traits: growth, survival, and fecundity in 86–89. Black Brant (Branta bernicla). Ecology: 76: Southwood, T.R.E. 1977. Habitat, the templet for 2404–2414. ecological strategies? Journal of Animal Ecology Sedinger, J.S., Lindberg, M.S., Person, B.P., 46: 377–365. Eichholz, M.W., Herzog, M.P. & Flint, P.L. Stephens, D.W. & Krebs, J.R. 1986. Foraging Theory. 1998. Density dependent effects on growth, Princeton University Press, New Jersey, body size, and clutch size in Black Brant. The USA. Auk 115: 613–620. Stephens, S., Koons, D.N. & Rotella, J.J. 2004. Sedinger, J.S., Herzog, M.P. & Ward, D.H. 2004. Effects of habitat fragmentation on avian Early environment and recruitment of Black nesting success: a review of the evidence at

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 Nest site selection by waterfowl in the northern hemisphere 129

multiple spatial scales. Biological Conservation eastern boreal forest of Quebec, Canada. 115: 101–110. Forest Ecology and Management 255: 2272–2285. Stephens, S.E., Rotella, J.J., Lindberg, M.S., Taper, Valone, T.J. 2007. From eavesdropping on M.L. & Ringelman, J.K. 2005. Duck nest performance to copying the behavior of survival in the Missouri Coteau of North others: a review of public information use. Dakota: landscape effects at multiple spatial Behavioral Ecology and Sociobiology 62: 1–14. scales. Ecological Applications 15: 2137–2149. van der Jeugd, H.P., van der Veen, I.T. & Summers, R.W., Underhill, L.G., Syroechkovski, Larsson, K. 2002. Kin clustering in barnacle E.E., Lappo, H.G., Pryˆs-Jones, R.P. & geese: familiarity or phenotype matching? Karpov, V. 1994. The breeding biology of Behavioural Ecology 13: 786–790. Dark-bellied Brent Geese Branta b. bernicla Van Horne, B. 1983. Density as a misleading and King Eiders Somateria spectabilis on the indicator of habitat quality. Journal of Wildlife northeastern Taimyr Peninsula, especially in Management 47: 893–901. relation to Snowy Owl Nyctea scandiaca nests. van Kleef, H.H., Willems, F. & Volkov, A.E. Wildfowl 45: 110–118. 2007. Dark-bellied brent geese breeding near Swanson, G.A., Meyer, M.I. & Serie, J.R. 1974. snowy owl nest lay more and larger eggs. Feeding ecology of breeding blue-winged Journal of Avian Biology 38: 1–6. teal. Journal of Wildlife Management 38: 396– van Riper, C., III. 1984. The influence of nectar 407. resources on nesting success and movement Talent, L.G., Jarvis, R.L. & Krapu, G.L. 1983. patterns of the Common Amakihi. The Auk Survival of Mallard broods in south-central 101: 38–46. North Dakota. The Condor 85: 74–78. Vermeer, K. 1968. Ecological aspects of ducks Thompson, S.C. & Raveling, D.G. 1987. nesting in high densities among Larids. The Incubation behavior of emperor geese Wilson Bulletin 80: 78–83. compared with other geese: interactions of Vickery P.D., Hunter, M.L. & Wells, J.V. 1992. predation, body size, and energetics. The Auk Evidence of incidental nest predation and its 104: 707–716. effects on nests of threatened grassland Tinbergen, N., Impekoven, M. & Franck, D. 1967. birds. Oikos 63: 281–288. An experiment on spacing-out as a defence Viitala, J., Korpimäki, E., Palokangas, P. & against predation. Behavior 38: 307–321. Koivula, M. 1995. Attraction of kestrels to Traylor, J.T. Alisauskas, R.T. & Kehoe, F.P. 2004. vole scent marks visible in ultraviolet light. Nesting ecology of White-winged scoters, at Nature 373: 425–427. Redberry Lake, Saskatchewan. The Auk 121: Wagner, R.H. 1993. The pursuit of extrapair 950–962. copulations by female birds: A new Underhill, L.G., Pryˆs-Jones, R.P. & hypothesis of colony formation. Journal of Syroechkovski, E.E. 1993. Breeding of Theoretical Biology 163: 333–346. waders and Brent Geese Branta bernicla Wagner, R.H. & Danchin, E. 2003. Conspecific bernicla at Pronchishcheva Lake, northeastern copying: A general mechanism of social Taimyr, Russia, in a peak and a decreasing aggregation. Animal Behaviour 65: 405–408. lemming year. Ibis 135: 277–292. Walker, J., Lindberg, M.S. & MacCluskie, M.C. Vaillancourt, M., Drapeau, P., Gauthier, S. & 2005. Nest survival of Scaup and other ducks Robert, M. 2008. Availability of standing in the Boreal Forest of Alaska. Journal of trees for large cavity-nesting birds in the Wildlife Management 69: 582–591.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 130 Nest site selection by waterfowl in the northern hemisphere

Walsberg, G. E. 1981. Nest-site selection and the Wilson, S.E. & Verbeek, N.A.M. 1995. Patterns radiative environment of the warbling vireo. of Wood Duck nest temperatures during The Condor 83: 86–88. egg-laying and incubation. The Condor 97: Walters, C.J. 1986. Adaptive Management of Renewable 963–969. Resources. McGraw-Hill, New York, USA. Winter, M., Johnson, D.H. & Shaffer, J.A. 2006. Weatherhead P.J. & Boak, K.A. 1986. Site Patch size and landscape effects on density infidelity in song sparrows. Animal Behaviour and nesting success of grassland birds. The 34: 1299–1310. Journal of Wildlife Management 70: 158–172. Webster, J.R. & Benfield, E.F. 1986. Vascular Wittenberg, J.F. & Hunt, G.L. 1985. The adaptive plant breakdown in freshwater ecosystems. significance of coloniality in birds. In D.S. Annual Review of Ecology and Systematics 17: Farner, J.R. King, & K.C. Parkes (eds.), Avian 567–594. biology. Vol. VIII, pp. 1–78. Academic Press, Weller, M.W. 1979. Density and habitat New York, New York. relationships of blue-winged teal nesting in Wright, D.H. 1983. Species-energy theory: an northwestern Iowa. Journal of Wildlife extension of species-area theory. Oikos 41: Management 43: 367–374. 495–506. Whitcomb, R.F., Robbins, C.S., Lynch, J.F., Yetter, A.P., Havera, S.P. & Hine, C.S. 1999. Whitcomb, B.L., Klimkiewicz, K. & Bystrak, Natural-cavity use by nesting Wood Ducks in D. 1981. Effects of forest fragmentation on Illinois. Journal of Wildlife Management 63: avifauna of the eastern deciduous forest. In 630–638. R.L. Burgess & D.M. Sharpe (eds.), Forest Young, A.D. & Titman, R.D. 1986. Cost and Island Dynamics in Man-dominated Landscapes, pp. benefits of Red-breasted Mergansers nesting 125–205. Springer-Verlag, New York, USA. in tern and gull colonies. Canadian Journal Wiens, J.A. 1989. Spatial scaling in ecology. Zoology 64: 2339–2343. Functional Ecology 3: 385–397. Zicus, M.C., Rave, D.P. & Riggs, M.R. 2004. Willms, M.A. & Crawford, R.D. 1989. Use of Factors influencing incubation egg-mass loss earthen islands by nesting ducks in North for three species of waterfowl. The Condor Dakota. Journal of Wildlife Management 53: 106: 506–516. 411–417. Zoellick, B.W., Ulmschneider, H.M. & Cade, B.S. Willson, M.F. 1976. The breeding distribution of 2004. Isolation of Snake River islands and North American migrant birds: a critique of mammalian predation of waterfowl nests. MacArthur (1959). Wilson Bulletin 88: 582–587. Journal of Wildlife Management 68: 650–662.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 86–130 131

Waterfowl habitat use and selection during the remigial moult period in the northern hemisphere

ANTHONY D. FOX1*, PAUL L. FLINT2, WILLIAM L. HOHMAN3 & JEAN-PIERRE L. SAVARD4

1Department of Bioscience, Aarhus University, Kalø, Grenåvej 14, DK-8410 Rønde, Denmark. 2USGS, 4210 University Drive, Anchorage, AK 99508-4626, USA. 3USDA/NRCS, National Wildlife Team, 501 W. Felix St. Bldg. 23, Fort Worth, Texas 76115, USA. 4Scientist Emeritus, Environment Canada, 801-1550 Avenue d’Estimauville, Québec, G1J 0C3, Canada. *Correspondence author. E-mail: [email protected]

Abstract This paper reviews factors affecting site selection amongst waterfowl (Anatidae) during the flightless remigial moult, emphasising the roles of predation and food supply (especially protein and energy). The current literature suggests survival during flightless moult is at least as high as at other times of the annual cycle, but documented cases of predation of flightless waterfowl under particular conditions lead us to infer that habitat selection is generally highly effective in mitigating or avoiding predation. High energetic costs of feather replacement and specific amino-acid requirements for their construction imply adoption of special energetic and nutritional strategies at a time when flightlessness limits movements. Some waterfowl meet their energy needs from endogenous stores accumulated prior to remigial moult, others rely on exogenous supply, but this varies with species, age, reproductive status and site. Limited evidence suggests feather proteins are derived from endogenous and exogenous sources which may affect site selection. Remigial moult does not occur independently of other annual cycle events and is affected by reproductive investment and success. Hence, moult strategies are affected by age, sex and reproductive history, and may be influenced by the need to attain a certain internal state for the next stage in the annual cycle (e.g. autumn migration). We know little about habitat selection during moult and urge more research of this poorly known part of the annual cycle, with particular emphasis on identifying key concentrations and habitats for specific flyway populations and the effects of disturbance upon these. This knowledge will better inform conservation actions and management actions concerning waterfowl during moult and the habitats that they exploit.

Key words: Anatidae, energy balance, feather synthesis, moult, predation, protein, survival.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 132 Habitat use during remigial moult

Normally annual replacement of avian cycle, these species exploit aquatic systems remiges (hereafter flight feathers) is that expose them to relatively low predation essential, because damage, abrasion and risk within and outside the flightless remigial exposure to ultra-violet light degrade such moult period. Furthermore, the moult of tissues (Stresemann & Stresemann 1966; males of several species of waterfowl from Bergman 1982). Allometric relationships a bright breeding plumage (the alternate underlying the production of feather tissues plumage according to Palmer 1972 and proportionally prolong replacement of Weller 1980, basic according to Pyle 2005) flight feathers in larger (> 300 g) birds to a cryptic eclipse plumage prior to the (Rohwer et al. 2009). For this reason, large wing feather moult (the basic plumage bird species extend flight feather moult over according to Palmer 1972 and Weller 1980, two or more seasons (e.g. Bridge 2007) or, as the alternate according to Pyle 2005) is is the case amongst birds that do not rely on likely, at least in part, an adaptation to the powers of flight for feeding, undergo a reduce conspicuousness during this simultaneous moult of all flight feathers highly vulnerable period. Loss of the that renders them flightless temporarily, powers of flight also reduces foraging including species of the Alcidae, Anatidae, opportunities, so a likely determinant Anhingidae, Bucerotidae, Gavidae, Gruidae, of habitat selection during remigial moult Heliornithidae, Jacanidae, Pelecanoididae, is the need to derive sufficient energy Phoenicopteridae, Podicipedidae, Rallidae and nutrients to satisfy maintenance and Scolopacidae (Stresemann & requirements and supplementary needs Stresemann 1966; Marks 1993). With the of feather replacement (Hanson 1962; notable exception of the Bucerotidae, where Hohman et al. 1992). breeding females undergo simultaneous For example, the massive aggregations of moult whilst sealed inside nest cavities by Eared Grebes Podiceps nigricollis that moult males (Stonor 1937; Moreau 1937), the on the Great Salt Lake, Utah, USA common feature of all these avian families is encapsulate elements of wetland selection their occupancy of aquatic or marine that likely characterise a moulting site for habitats. Indeed, of the avian families that waterbirds during remigial moult, namely forage primarily on or under water the flat trophic structure of a hyper-saline throughout the annual cycle (i.e. excluding lake with little emergent or submersed seabirds that forage on land or on the wing; vegetation means aquatic or aerial predators Fregatidae, Laridae, Procellariiformes and are rare or non-existent, and the protein Sternidae), members of the Anatidae, content of the super-abundant and highly Alcidae, Gavidae, Pelecanoididae and accessible prey (in this case the Brine Podicipedidae replace their flight feathers Shrimp Artemia franciscana) in the lake synchronously whilst in wetlands, most provides optimal conditions for obtaining often on water. To minimise the energy and nutrients for feather growth (Jehl omnipresent mortality risk posed by aquatic 1990; Wunder et al. 2012). Flightless and aerial predators throughout the annual waterbirds constrained to a wetland

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 133 minimise depredation by diving in the case habitat selection in relation to strategies that of ducks or, as in the case of dabbling fulfil the needs for energy and protein ducks, geese and swans, escaping terrestrial during the remigial moult, while minimising mammals by swimming into vegetation or risks to survival posed by flightlessness. open water distant from shore. Prior to Those few studies that exist suggest that remigial moult, birds must select habitats safety from predators is paramount in the that fulfil their nutritional needs, specifically selection of habitats for moult but food those for energy and protein with specific resources and availability are also likely to be amino acids (Murphy & King 1984a,b). critical. Nonetheless, numerous questions Given the need to meet normal energy and remain. What general features can we protein requirements, plus the extra needs distinguish about moulting sites that are of wing moult, and avoidance of predation similar or different from habitats used at whilst unable to fly, how do moulting other times of the annual cycle? How can waterfowl select their moulting habitats? identification of these features help us Specifically, with regard to improving understand the potential process of habitat conservation and management options to selection by moulting waterfowl? Can protect and enhance such habitats, what are existing variation within and among species the key features of these habitats? be used to address survival and possible Answering these questions is difficult, fitness consequences of selecting specific because few studies have ever specifically habitats? Lastly, how do answers to these examined habitat use and selection by questions provide insight into how the moulting waterfowl, in the sense that a conservation and management of moulting particular feature or features are selected habitats might be improved for indigenous over others. Early examples attempted to waterfowl in the northern hemisphere? describe the physical, floristic, and Intuitively, moult is not independent of invertebrate features within sites that other annual cycle events and is highly attracted, for example, moulting Green- dependent on reproductive investment and winged Teal Anas crecca (Kortegaard 1974), outcome (i.e. whether an individual is paired but rarely have researchers investigated or unpaired, a breeder or non-breeder, a habitat use before, during and following the successful or unsuccessful breeder). Thus, remigial moult to enlighten factors affecting we also attempt to assess how this may habitat choice at that critical period. Thus, influence habitat use during remigial moult while we can describe habitats used by and the manner in which moult is moulting waterfowl, there is insufficient completed. Finally, we consider the future literature available to relate use to available key policy, conservation, and research needs habitats to infer preferences or selection in this arena. We start by considering the (sensu Kaminski & Weller 1992; Kaminski & evidence for selection of habitats based on Elmberg 2014). safety from predation and assess the specific In this review, we adopt a comparative energetic and nutritional needs of remigial approach among waterfowl taxa to examine moult. The focus is on remigial moult rather

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 134 Habitat use during remigial moult than body plumage moult, because both Gifford, in litt.). Interestingly, both situations occur simultaneously during the summer relate to expanding Palearctic populations of moult into the eclipse plumage in the geese moulting in newly colonised areas, northern hemisphere (Weller 1980) and the where local terrestrial moulting habitats period of flightlessness while remiges are have become saturated as a result of lost and regrown and associated habitat and increasing local density. These examples resource use by waterfowl are most crucial may therefore be atypical in the sense that to individual survival and future fitness these populations have yet to reach fitness prospects. equilibrium with regard to colonisation of formerly unoccupied territory and exposure Role of predation during to novel predators. In the case of Mottled remigial moult Duck, drying wetland conditions in peninsular Florida in late summer The limited evidence available suggests that concentrate wildlife, including alligators survival during the post-breeding and Alligator mississippiensis, in and around remigial moult is the same, if not greater, remaining wetlands, resulting in alligator than during other periods in the annual cycle depredation of post-breeding moulting for such species as Mallard Anas platyrhynchos Mottled Duck (Bielefeld & Cox 2006). (Kirby & Cowardin 1986), Mottled Duck A. An exception to “normal” survival rates fulvigula (Bielefeld & Cox 2006), Black Duck during moult is the periodic occurrence of A. rubripes (Bowman 1987), Wood Duck Aix mass mortality due to diseases frequently sponsa (Thompson & Baldassarre 1988; occurring during moult and summer drought Davis et al. 2001), Harlequin Duck periods (Bellrose 1980). Dabbling ducks Histrionicus histrionicus (Iverson & Esler 2007), exposed to botulism during the post- Barrow’s Goldeneye Bucephala islandica breeding period may suffer severe mortality (Hogan et al. 2013a) and scoters Melanitta sp. (e.g. > 350,000 Northern Pintail Anas acuta in (Anderson et al. 2012). However, survival prairie Canada, Miller & Duncan 1999 and probability by itself does not imply that < 12,000 Mallard, Fleskes et al. 2010; predation risk has no role in shaping habitat Evelsizer et al. 2010), as have Redhead use by post-breeding moulting waterfowl as Aythya americana (> 3,000 at one site, Wobeser it may simply indicate that waterfowl have & Leighton 1988) and sea ducks have mitigated this risk through adaptations. suffered mortality related to virus exposure Indeed, there are reports of flightless (Hollmén et al. 2003), although other die-off Pink-footed Geese Anser brachyrhynchus being events may have been the result of depredated by Walrus Odobenus rosmarus contaminant exposures (Henny et al. 1995). rosmarus in Svalbard (Fox et al. 2010) and The flocking behaviour of moulting flightless moulting Common Eider Somateria waterfowl may facilitate disease spread mollissima and Greylag Geese Anser anser within groups, and large mass mortalities are being pursued by Killer Whales Orcinus orca likely to have population level effects (Reed in the Shetland Islands, Scotland (D. & Rocke 1992), but these events are rare and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 135 are unlikely to represent a major factor in (1996) reported social interactions tend to moult ecology and site selection. synchronise moult timing in captive female Mallard and that this may have important Factors that mitigate or reduce survival advantages. The group response of mortality during moult moulting waterfowl to disturbance events suggests that one of the advantages to Moult migration moulting in flocks is the increased detection Ecologists have hypothesised that an of potential predators (Kahlert 2003). Thus, advantage of moult migration is selection of flocking behaviour may have survival habitats with lower predation pressure or advantages in spite of the potential for less diverse predator communities increased competition for food and more (Salomonsen 1968). In many species of rapid pathogen transmission. ducks and geese, moult migrations are to Behavioural activity also affects group higher latitude locations where predator size. Foraging flocks of moulting waterfowl populations may be numerically, or more are usually much smaller than roosting seasonally, constrained (Yarris et al. 1994). flocks which can aggregate thousands of For sea duck species, moult migrations may individuals (Reed 1971; Jepsen 1973; or may not represent an increase in latitude, Joensen 1973). Roosting in dense flocks but almost always represent a movement to likely has important bioenergetic and marine areas where predators are fewer. predator protection advantages. Moulting Thus, moult migration itself may represent Surf Scoter Melanitta perspicillata (O’Connor an adaptation to minimise mortality during 2008) and King Eider Somateria spectabilis the flightless period. (Frimer 1994) stop feeding when disturbed and regroup in large flocks offshore, a Flocking behaviour behaviour common to most moulting sea Factors affecting abundance of moulting ducks and divers and some dabbling ducks birds at a given site have not been explored. (Oring 1964). Similarly, resting flocks of Most, if not all, waterfowl species moult in flightless Red-breasted Merganser Mergus flocks that vary in size from a few birds to serrator are larger than foraging flocks (JPLS thousands of individuals. Solitary moulting unpublished data). The tendency of flocked does not seem to have evolved in any birds (diving and sea ducks) to dive waterfowl species, although northern synchronously to feed may be an anti- breeding female ducks completing brood predator or a foraging efficiency strategy rearing late will moult alone. In general, (Schenkeveld & Ydenberg 1985). Certainly, birds in groups are less susceptible to simultaneous diving is a typical response to predation than solitary birds and safety is, to avian predators (PLF pers. obs.). Some some degree, proportional to group size and species, especially dabbling ducks, rely on to position within the group (Petit & Bilstein cover for protection and disperse in 1987; Elgar 1989; Tamisier & Dehorter vegetation when disturbed. In some species 1999; cf. Davis et al. 2007). Leafloor et al. of sea ducks, females often moult in smaller

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 136 Habitat use during remigial moult groups than males and may use different with presence of emergent vegetation habitats (Gauthier & Bédard 1976; Jepsen (Kortegaard 1974; Fleskes et al. 2010). 1973; Diéval 2006; Diéval et al. 2011). Moulting diving ducks escape danger by Dabbling ducks seem to rely more on cover diving and dispersing and select wetlands of during the remigial moult, and unlike some sufficient depth to avoid aerial predation. sea ducks, tend to form smaller flocks Thus, most dabbling ducks select habitats during remigial moult than before or after with emergent vegetation to moult whereas this period (Kortegaard 1974). moulting diving and sea ducks generally avoid them and favour open areas (Oring Selection for escape habitat 1964). However, predation on flightless Moulting birds may show stronger selection Common Eider by Killer Whales has been for escape mechanisms as opposed to documented (Smith 2006; Booth & Ellis foraging habitat. For example, most 2006) indicating bird vulnerability to marine northern geese moult in treeless areas and predators. Most sea ducks moulting in moulting concentrations can reach coastal waters forage at shallower water thousands of birds, where terrestrially depths when flightless, perhaps linked to feeding birds rely on adjacent open-water their impaired diving capacities as most use rivers, lakes, and other wetlands for escape their wings underwater (Comeau 1923). from predators (Derksen et al. 1982; Madsen Selection of shallow water may reduce the & Mortensen 1987). For these arctic geese, amount of time spent foraging and possibly the primary predator influencing forage minimise heat loss during submergence in behaviour was likely Arctic Fox Alopex cold waters. Further, use of shallow waters lagopus. Fox & Kahlert (2000) and Kahlert may be a strategy to minimise predation as (2003) found that moulting Greylag Geese water depth may limit exposure to some only foraged in close proximity to water (i.e. marine mammals (e.g. Killer Whales). In the escape habitat) even though abundant extreme, some moulting Kerguelen Pintail unexploited forage of equal quality was Anas eatoni apparently moult in a cave on the available in other locations. Thus, these Kerguelen archipelago to escape predation geese seem to be selecting for open habitats (Buffard 1995). with good visibility but restrict their habitat use based on access to escape habitat from Behavioural modifications the local predators, although predators Numerous studies have also documented actually were absent during the study general behavioural shifts associated with (Kahlert et al. 1996). In contrast, moulting moult that might influence predation risk ducks spend most of their feeding and (Hohman et al. 1992). In general, moulting roosting times on water where predators waterfowl reduce active behaviours and may harass and attack them from the air or spend more time roosting (e.g. Döpfner under water. Dabbling ducks rely on et al. 2009; Portugal et al. 2011). The degree emergent vegetation for concealment to which this behavioural modification and escape and select lakes and marshes is adopted is likely linked to energy

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 137 balance, body mass dynamics and ability Role of food of birds to store reserves prior to moult. Nonetheless, by minimising time in Is there a high energetic cost of potentially less safe habitats (i.e. foraging remigial moult? habitats) and selecting relatively safe Replacing feathers is energetically costly, so roosting habitats, moulting birds may be waterfowl face potentially increased energy minimising predation risk. demands to meet the costs of feather growth (Payne 1972; Thompson & Boag Minimisation of flightless period 1976; Dolnik & Gavrilov 1979; Qian & Xu Several studies have suggested that body 1986; Portugal et al. 2007), estimated at 1.3 mass dynamics of moulting waterfowl times basal metabolic rate in Mallard (Prince represent an adaptive mechanism to 1979). However, most species moult at the minimise length of the flightless period warmest point in the annual cycle and other (Douthwaite 1976; Brown & Saunders 1998; factors suggest that the costs of feather Owen & Ogilvie 1979). The logic is that replacement are not necessarily difficult to wing loading ultimately determines ability to meet from external sources. For example, fly. Therefore, mass loss during moult many waterfowl engage in the moult of reduces wing loading and allows birds to body feathers synchronously with remiges regain flight before primaries are fully (Weller 1980; Taylor 1995; Howell et al. grown. For many species, mass loss may 2003). Further, many species restrict food allow birds to gain flight when primaries are intake during moult compared to other about 70% of ultimate length (Taylor 1995; times of the year. Finally, while many species Howell 2002; Flint et al. 2003; Dickson lose mass during moult which could indicate 2011). If flightlessness itself increases an inability to balance energy intake with mortality risk, then this adaptation to demand, this is not the case for all species or minimise the flightless period would be an sites (Lewis et al. 2011a,b; Dickson 2011). adaptation to reduce mortality. Interestingly, Accordingly, moult is potentially a time there are some species (e.g. scoters) which period of energetic constraint, yet it appears have protracted flightless periods and that waterfowl have adapted to mitigate this corresponding high survival (Anderson et al. cost. 2012). These species have slow rates of feather growth and no mass loss while Mass loss, what does it mean? flightless (Dickson et al. 2012). As such, There has been debate as to whether mass there appears to be little selective pressure to loss during remigial moult in waterfowl is: minimise the flightless period for scoters because they use habitats with adequate (1) an adaptive trait, whereby fat stores food and encounter little apparent risk of provide an endogenous source of predation, although we cannot ignore the energy to regrow feathers as rapidly as alternative hypothesis that these ducks need possible whilst reducing reliance on full wing length to fly. external energy sources, access to which

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 138 Habitat use during remigial moult

potentially increases predation risk (as energy budgets during the flightless period suggested for some geese; Fox & of remigial feather re-growth. In other Kahlert 2005); words, both sexes of Mallard showed prior mass gain (mostly fat stores) to fuel energy (2) a simple reflection of the elevated demands during wing moult, just as energetic costs of feather synthesis migratory populations accumulate such which birds meet by catabolism of body stores to fuel migration. Male and non- “reserves” (the energetic stress breeding female Mallard could meet up to hypothesis; Hohman 1993) rather than 82% of all energy expenditure whilst body “stores” acquired exogenously flightless from energy stores alone, and prior to the moult (sensu van der Meer & Pochard Aythya ferina could derive up to 92% Piersma 1994); of such energy demands (Fox & King 2011). (3) due to predation risk, that imposes Fondell et al. (2013) also provided evidence cryptic feeding behaviour and against (2) as Black Brant Branta bernicla exploitation of habitats where foraging nigricans with access to the most nutritious is less effective, such foraging forage lost the most mass. However, Lewis et constraints necessitate exploitation of al. (2011b) emphasised that the adaptive fat (which does not necessarily preclude relationship described in (1) is not fixed, as pre-moult accumulation of fat stores; the overall rates of mass loss declined across Panek & Majewski 1990); or several decades for moulting Black Brant. Many species show mass loss during (4) an adaptive trait to reduce the length of flightlessness (see review in Hohman et al. the flightless period because lighter 1992 and references therein), including body mass enables Anatidae to regain temperate moulting Greylag Geese. On the flight earlier on incompletely re-grown Danish island of Saltholm, their mass loss flight feathers earlier than if heavier equated to depletion of fat stores (Douthwaite 1976; Owen & Ogilvie accumulated prior to moult and which again 1979; Brown & Saunders 1998). could support a large proportion of the For at least one population of energy expenditures during moult if the Mallard, Fox et al. (2013) showed (using geese opted not to move between safe supplementary feeding) that there was no resting areas during daylight and their night- support for (2) and (3) above and that (4) time feeding grounds (Kahlert 2006a). alone was not the primary factor that shaped Through hyperphagia, male Northern weight loss. Rather they considered that the Shoveler Anas clypeata accumulated reserves accumulation and subsequent depletion of prior to moulting and used stored resources fat stores, together with reductions in energy to grow feathers, an adaptation to declining expenditure, enables Mallard to re-grow cladoceran availability in mid-summer feathers as rapidly as possible by exploiting (DuBowy 1985). However, even within a habitats that offer safety from predators, but species, not all populations lose mass at the do not necessarily enable them to balance same rate implying that, while overall mass

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 139 loss may be adaptive, there is some influence selection for the most highly digestible and of local feeding conditions (Fox et al. 1998; protein-rich species available; as the quality Fox & Kahlert 2005; Fondell et al. 2013). of this forage declined, the diet became increasingly diverse (Fox et al. 1998). Factors that mitigate or However, that same study showed that birds influence energetic costs and exploited the best quality forage that was mass loss closest to open water to which birds could escape when threatened; leaving food Meeting protein needs during remigial resources distant from the water’s edge moult unexploited, strongly suggested that Flight feathers comprise c.0.7% of the total predation risk during flightlessness was body mass of a Greylag Goose (A.D. Fox, more important than simply food supply unpubl. data) and 0.2% of a female Mallard (Fox & Kahlert 2000, even in this case, (Heitermeyer 1988) and remiges some 25% where predators are absent, geese still of feather mass, so while a substantial part responded vehemently to predator-like of overall plumage, the absolute mass of stimuli, Kahlert 2003, 2006b). However, in flight feathers is relatively not that great. the case of moulting Greylag Geese on However, the simultaneous replacement of Saltholm, stable isotope data evidenced that the largest feathers of most waterfowl over geese used protein accumulated on the a relatively short period necessitates access mainland in Sweden for feather synthesis to food that provides the basic nutrients for (Fox et al. 2009). Proteins were released from their synthesis. This includes amino acids organs which change in size during moult containing sulphur for β-keratin synthesis (Fox & Kahlert 2005), and excretion of (Hohman et al. 1992), which are generally nitrogen in the form of urea and uric acid less common in avian tissues and diet than nearly ceased during the middle part of in feathers (Murphy & King 1982, 1984a,b). moult suggesting considerable physiological The extent to which protein invested in mechanisms that reduced reliance on feather tissue is derived from exogenous external sources of nitrogen during wing versus endogenous sources in moulting moult (Fox & Kahlert 1999). waterfowl remains unclear, but the only study (Fox et al. 2009) suggests both Meeting lipid needs during moult sources are used. Recent studies suggest a Lipids accumulated prior to moult are progressive change in isotopic composition, primarily used to meet elevated energy shifting from largely endogenous sourced demands during moult to offset the protein to protein derived from the diet temporarily increased demands of feather along the length of the feather as moult synthesis (Young & Boag 1982; Fox & migrant geese come into equilibrium with a Kahlert 2005; Fox et al. 2008). Many new isoscape (S. Rohwer pers. comm.). dabbling ducks (e.g. DuBowy 1985; Sjöberg Studies of the diet of moulting Greylag 1988; Panek & Majewski 1990; Moorman et Geese on Saltholm showed there was al. 1993; King & Fox 2012) and diving ducks

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 140 Habitat use during remigial moult

(e.g. Fox & King 2011) lose mass during the the environment in which it moults to flightless moult (cf. Fox et al. 2008; Dickson meet needs of maintenance and feather 2011; Hogan 2012; Hogan et al. 2013b), replacement. likely due to consumption of body fat Other species, such as some arctic stores. As in the case of Saltholm moulting moulting geese, unconstrained by feeding Greylag Geese (Fox & Kahlert 2005), Folk et restrictions because of diel light conditions al. (1966) reported mass loss in moulting during summer (Fox et al. 1999), or sea Mallard, and Young & Boag (1982) ducks, such as the Common Scoter Melanitta documented a reduction in fat stores in nigra that live on protein and energy-rich Mallard through moult, although they food and occupy habitat that subsequently is asserted there was no overall change in total the source of its winter food supply (Fox et carcass lipids, total proteins, or total body al. 2008), show no such accumulation of fat mass. However, Panek & Majewski (1990) stores in advance of wing moult and appear showed 12% declines in body mass amongst entirely able to supply their energy males and females through moult. Taylor expenditure from exogenous sources during (1993) reported that moulting Black Brant moult. Indeed, Canvasback Aythya valisineria lost 71–88% of stored lipid reserves and males can actually accumulate mass in the ended moult with only structural lipids form of lipid stores from exogenous remaining (i.e. 2–4% of fresh body mass). sources toward the end of remige growth So why do some Anatids lose mass during in preparation for autumn moult into moult, while others do not? Not only can breeding plumage, presumably because they using fat stores potentially free moulting undertake remigial moult in habitats with waterfowl from feeding or at least as high food quality (Thompson & Drobney intensively as would otherwise be necessary 1996). Surf and White-winged Scoter (e.g. Fox & King 2011) but there is also Melanitta deglandi moulting in coastal British evidence from the difference in energy Columbia and Alaska also gained weight stores between moulting individuals that during their remigial moult. Barrow’s energy stores may affect the rate of feather Goldeneye moulting on arctic wetlands lost growth. In Barrow’s Goldeneye, van der weight during remigial moult (van de Wetering & Cooke (2000) found that Wetering & Cooke 2000) but those moulting remigial growth rate in recaptured in northern Alberta gained weight (Hogan individuals was positively correlated with 2013b) suggesting high variability in moult size-adjusted body mass at initial capture ecology within species, perhaps linked to the and that the daily rate of body mass loss was time constraints imposed on birds in these greater amongst birds that started moult in different biogeographic settings. better condition. Hence, the lipid status in Furthermore, the evidence presented which an individual starts moult may above suggest Mallard under certain have considerable implications for rate of circumstances do not lose mass during feather growth (and hence duration of moult (e.g. Young & Boag 1982), and there is flightlessness) and may also be a function of evidence that Greylag Geese in the north of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 141 their range in Iceland and northern Norway during moult (Fondell et al. 2013, see Table 1 lose far less mass than do those on Saltholm for an overall summary). Subjectively, forage (Arnor Sigfusson, Carl Mitchell and Arne varied across these three groups, with little Follestad, pers. comm.). Rates of mass loss forage available near the nesting colony for in Black Brant have varied through time brood rearing birds, intermediate levels of such that birds moulting on the same lakes forage available to failed breeding birds on now lose less mass compared to several the Yukon Delta, and abundant food decades ago, suggesting that this trait is available near Teshekpuk Lake for the moult highly adaptive and may depend on local migrant functional non-breeders. However, circumstances associated with the specific adults rearing young did not lose mass moult site and individual status (Lewis et al. during moult. Failed breeding birds on the 2011b). Hence, moult mass dynamics Yukon Delta lost intermediate amounts of appear to vary within species with site and weight during moult and those at Teshekpuk potentially reproductive status. Lake lost the most. In these cases, mass loss Data from moulting Black Brant from was negatively correlated with apparent three different areas (i.e. brood-rearing forage availability, yet mass at the onset of flocks, failed breeding birds on the Yukon moult also varied across these three areas Delta, and failed- or non-breeding birds and was linked with the forage available at from Teshekpuk Lake) show major each area. So successful breeding birds differences in mass dynamics before and started moult at a relatively low mass and

Table 1. Summary of moult behaviour of discrete elements of the brant population from the Yukon-Kuskokwim Delta in Alaska, showing the location, duration of flightlessness, mass dynamics and food quality of the habitats used for brood-rearing parents, failed- and non-breeding birds (Singer et al. 2012; Fondell et al. 2013).

Population Moult site Flightless Feather Starting Food Mass loss segment location period growth rate body quality during (days) (mm/day) mass moult

Brood At colony 30 5 Low Low None rearing (few km) parents Failed Yukon Delta 21 7.5 Inter- Inter- Little breeders (many km) mediate mediate Non- Teshekpuk 21 7.5 High High Most breeders (950 km)

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 142 Habitat use during remigial moult maintained body mass, whereas moult species that showed fat store accumulation migrants started at a higher mass and lost prior to moult, where these stores were most of that during moult. Thus, all groups depleted as feather regrowth proceeded. converged on a similar mass at the end of This strategy may be an adaptation toward moult. This convergence would suggest that energetic independence from exogenous mass loss is adaptive in that there is some energy-rich foods in situations where optimal mass that reduces wing loading and foraging brings accompanying predation facilitates flight and dispersal. These data risks. However, evidence indicated that the suggest that there would appear to be some rate of depletion of fat stores varied with selection pressure to minimise the length of remigial growth rate and initial mass in some the flightless period. It is also interesting species, suggesting endogenous food stores that moult migrants start at the highest mass could accelerate feather growth and reduce implying that they are more than able to duration of flightlessness and predation make up for the energetic costs of migration risk. Hence, the presence of lipid body and perhaps bring fat stores as a hedge stores does not necessarily imply that against uncertain food resources at the moulting birds have reduced need for ultimate moult destination. exogenous energy, because there are many Black Brant rearing broods delay the examples of moulting birds selecting food- onset of moult until about 16 days after rich environments. For instance, at the hatch, thus regaining flight at about the Ismaninger Teichgebiet, a complex of fish same time as goslings fledge (Singer et al. ponds in Bavaria, Germany, where 3 out of 2012). So for waterfowl that stay with their 30 impoundments were left fishless, these broods, there may be little selective pressure ponds each attracted, on average, about to reduce the length of the flightless period 2,000 birds, mostly moulting Eared Grebes, because there is little advantage to adults Gadwall Anas strepera, Mallard, Pochard, flying before their young. Failed- and non- Tufted Duck Aythya fuligula, Red-crested breeding birds utilise available forage to gain Pochard Netta rufina and Coot Fulica atra, mass prior to moult. They can use compared to <100 moulting waterbirds on supplementary body fat stores potentially to the remaining lakes stocked with Carp invest in more rapid growth of flight Cyprinus carpio (Köhler & Köhler 1998). The feathers (compared to the brood-rearing implication of this extraordinary difference adults) to allow an early return to flight. in moulting waterbird density was that the However, female ducks raising broods face a abundance of macroinvertebrates and algae trade-off between protecting the brood and in the ponds without carp provided departing with sufficient time to complete improved feeding conditions over those remigial moult in appropriate habitats. ponds with carp. Intriguingly, Northern Shoveler was the only species to show Forage resources similar (but low) moulting densities on In previous sections on the energetic stocked and fishless lakes, likely because it is dynamics of moult, we described a range of a pelagic dabbling duck specialising on filter

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 143 feeding of macro- and micro-invertebrates conditions. Even where this is not the case, (Ankney & Afton 1988). Amongst Gadwall the physical constraint on movements likely at this site, both sexes lost relatively little enhances local depletion of food resources. body mass, which they recovered before the Feeding ecology may also impose end of wing moult (Gehrold & Köhler constraints on the size of moulting flocks. 2013). This contrasts with the consistent In general, species feeding on shellfish loss of mass amongst both sexes of the moult in larger flocks than species feeding same species at a moulting site in southeast on mobile invertebrates, fish or vegetation. England (King & Fox 2012). The unplanned For shellfish feeding species, foraging in Ismaninger Teichgebiet “experiment” groups is not limited to the moulting season strongly suggested that locally high densities but occurs during non-breeding seasons of moulting waterfowl may react to a when birds are gregarious. Common Eider combination of factors in moulting habitats, moult in mussel rich habitats, such as in the but the high density of food in the fishless Danish Wadden Sea and Kattegat where ponds may have overridden any anti- aggregations reach tens of thousands, while predator function. scoters moult in areas with large sandy The density of moulting Barrow’s subtidal areas rich in bivalve shellfish Goldeneye on 21 ponds of the Old Crow resources (Joensen 1973; Laursen et al. Flats, Canada, was positively correlated with 1997). These numbers likely represent total phosphorous levels (van de Wetering abundance at moulting locations as foraging 1997) suggesting a positive relationship flocks at moulting sites tend to be smaller between goldeneye abundance and than those later in the winter (Follestad et al. primary productivity of the wetlands. In 1988; Rail & Savard 2003). Mergansers feed northwest North America, large moulting on fish and moulting flocks are usually concentrations of waterfowl are associated smaller than those of scoters and eiders: with highly productive large shallow boreal Red-breasted Merganser flocks moulting in and sub-boreal lakes (Munro 1941; van de the coastal waters of Anticosti Island in the Wetering 1997; Hogan et al. 2011), large St Lawrence, Canada averaged 39 shallow boreal (Bailey 1983a,b) and arctic birds/flock with a range of 1–322 birds wetlands (King 1963, 1973) and coastal (Craik et al. 2009, 2011); Common estuaries (Flint et al. 2008), while in Iceland, Merganser M. merganser flocks moulting on numbers of moulting Barrow’s Goldeneye fresh and salt water appear to be in a similar and Red-breasted Merganser were range of size (Kumari 1979; Pearce et al. correlated positively with food abundance 2009). Goldeneyes forage on invertebrates (Einarsson & Gardarsson 2004). Loss of the and also moult in smaller groups than ability to fly not only affects the ability of scoters or eiders (Jepsen 1973; van de birds to avoid predation, it also spatially Wetering 1997). Freshwater diving and limits the ability to gather food during the dabbling ducks that consume aquatic flightless period, which likely restricts the vascular plant material may rely on large ability to exploit the best foraging beds of such plants to support them

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 144 Habitat use during remigial moult through remigial moult, where larger conserve energy by not flying, which is the feeding resources attract dense flocks of most expensive activity of all avian energy moulting birds (Bailey 1983a,b). expenditure, usually estimated at 12–15 Flock behaviour for moulting may times basal metabolic rate in waterfowl (e.g. actually enhance forage quality and quantity Prince 1979; Madsen 1985). In moulting via positive feedback. This hypothesis is Common Eider, Guillemette et al. (2007) particularly true for goose grazing systems showed that daily and resting metabolic where regular foraging tends to maintain rates increased by 9 and 12%, but also that plant productivity at a higher biomass and flightlessness reduced daily and resting prolong the peak of nitrogen content metabolic rate by 6 and 14%, respectively, (Cargill & Jefferies 1984). Fox & Kahlert helping to balance energy demands of (2000, 2003) demonstrated that moulting feather synthesis. Indeed, many authors Greylag Geese maintained both protein and concur that moulting waterfowl are much biomass production at the greatest possible more secretive, but also far more quiescent levels by frequent re-grazing of the sward. than in the period prior to and following There are no similar documented positive flightlessness, reducing activity substantially feedback relationships for other forage (e.g. Adams et al. 2000; Döpfner et al. 2009) systems (e.g. invertebrates) that would be even in captive waterfowl fed ad libitum relevant for other waterfowl; however, (Portugal et al. 2010), although these studies behavioural modifications like simultaneous collected activity data only during daylight. diving may enhance feeding efficiency. Thus, Hogan et al. (2013a,b) reported that for some species of waterfowl, flocking moulting Barrow’s Goldeneye foraged behaviour during the flightless period may primarily at night on one lake and actually increase forage quantity, quality, diurnally on another, possibly in relation foraging efficiency, or a combination of to prey behaviour and availability these. Moulting waterfowl therefore seem to suggesting adaptability to local conditions concentrate in areas of high forage but perhaps also due to greater vulnerability abundance, hardly surprising as large to predation on the smaller lake where aggregations of birds would require a they fed nocturnally, although they could concentrated food source, from which they not quantify this risk. Hence, by their lack may gain an additional advantage from the of flight and generally reduced activities, decrease in predation risk to individual moulting waterfowl can substantially birds. reduce their energy expenditure during flightlessness. Behavioural modifications In contrast to the general observations One means of rebalancing an energy budget that flightless moulting waterfowl are less burdened by the additional energy demands active whilst replacing feathers, an of major feather synthesis is to reduce other alternative strategy would be for birds to forms of energy and other nutrient feed more or on more energy dense or expenditure. Birds replacing flight feathers nutritious foods to meet elevated needs of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 145 moult. This hypothesis has rarely been response to food supply (Piersma et al. 1993) tested because of the paucity of studies in and during refuelling episodes on migration this field, but has certainly been advanced (Piersma et al. 1999a,b) and grebes adjust for Mallard by Hartman (1985) and seems to organ sizes to meet the energetic costs of occur in moulting Steller’s Eider which feed migration (Jehl & Henry 2010, 2013) and on more energetically rich prey during the muscle mass for those that moult (Piersma remigial moult (Petersen 1981). 1988). Maintenance costs consume the vast Strategies invoked during remigial moult majority of normal energetic expenditures may also differ between species (Döpfner et of an organism, and most internal al. 2009): Mute Swans Cygnus olor reduced organs (but particularly the liver and swimming activities but increased foraging gastrointestinal tract) are energetically costly during moult; Red-crested Pochard Netta to maintain in a larger state than is rufina increased locomotion and feeding functionally necessary (Ferrell 1988). Hence, activities; and Gadwall and Tufted Duck the level of downsizing of organs that spent less time foraging. However, nocturnal occurs in the Greylag Goose during moult activities were not monitored during this on the Danish island of Saltholm is likely study and it is known that Redhead and associated with energetic savings over and Red-crested Pochard feed primarily at night above the alternative explanation under the during remigial moult (Bailey 1981; van use-disuse hypothesis (Fox & Kahlert 2005). Impe 1985), which has been suggested to That study found reduction of 41% in provide an exogenous source of energy intestine mass and 37% reductions in liver to offset thermoregulatory costs during and heart mass during moult, although these the coolest part of the 24 hour cycle. were increasingly reconstructed as birds Future study of the diurnal and nocturnal progressed toward completion of moult. In activities and energy budgets of moulting a study of high arctic Brant, Ankney (1984) waterfowl would be extremely valuable found no such changes in liver or intestine to enlightening our understanding of mass through moult, but this phenomenon habitat use and selection by moulting may be because arctic geese are more able to waterfowl. meet their energy demands from herbivory during moult than are Greylag Geese Physiological modifications moulting farther south. Dramatic changes in An alternative strategy would be for digestive organ size have been described in moulting waterfowl to reconstruct body other waterfowl undergoing remigial moult parts to help meet the energetic needs of (e.g. DuBowy 1985; Thompson & Drobney remigial moult, an aspect of phenotypic 1996), but this aspect of energy plasticity. In this way, organs or muscles that conservation and the degree to which such are costly to maintain are reduced in size to plasticity in organs can contribute to minimise energy consumption. Shorebirds reducing energy expenditure has been are well-known for making radical and rapid rarely studied. adjustments to the digestive apparatus in There is also evidence of changes in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 146 Habitat use during remigial moult muscle size during moult. In Mallard (Young groups if disturbed (van de Wetering 1997). & Boag 1982), Long-tailed Duck Clangula These associations seem based on selection hyemalis (Howell 2002), Black Brant (Taylor of similar habitat by moulting species rather 1993), Northern Shoveler and Blue-winged than attraction to sites used by a given Teal Anas discors (DuBowy 1985) flight species. However, species likely benefit from muscle mass decreased and leg muscle mass each other from protection in numbers. In increased consistently with the predictions northern Greenland, Common Eider and arising from the hypothesis of phenotypic King Eider moult at similar locations but plasticity (see also Ankney 1979, 1984; Fox use different habitats with King Eider & Kahlert 2005). Despite major changes in foraging in deeper waters (Frimer 1995). In muscle architecture, these tend not to northern Alaska, four species of geese contribute to major overall changes in body utilise the large thaw lakes north of mass during flightless moult in most studied Teshekpuk Lake (Flint et al. 2008). However, Anatidae (e.g. Ankney 1979, 1984; there is some spatial segregation with Thompson & Drobney 1996), so such different species using somewhat different changes are less easy to dissociate from the areas. Even in cases where multiple species simple hypothesis of use/disuse. are moulting within the same watershed, we do not know if they compete for the same Competition within and among forage (Lewis et al. 2011b). species Selection of moulting habitats by sea Some moulting sites attract single species ducks is likely based on food resources and (e.g. Harlequin Duck or Steller’s Eider; on the presence of congeners. Thousands Boertmann & Mosbech 2002; P. Flint of eiders, scoters and Long-tailed unpubl. data) whereas others are used by Duck may use a given moulting site. several species of waterfowl including Approximately 30,000 Long-tailed Duck dabbling, diving and even sea ducks. For moult along 531 km of coastline in the example Ohtig Lake (5.5 x 2.5 km) in Alaska Beaufort Sea, dispersed among 73 areas supports 20,000 moulting ducks of at least supporting from a few to 2,500 birds, 10 species, as does Takslesluk Lake (19 x 6 depending on local habitat configuration km) which attracts 10,000 moulters (King (e.g. presence of islands, sand spits; river 1963, 1973). When disturbed, birds form deltas; Gollop & Richardson 1974). mixed species flocks and swim toward However, within that site they form several natural sanctuary habitats. These lakes foraging flocks of various sizes. Flint et al. obviously fulfil the diverse needs of these (2004) studied radio-tagged Long-tailed species, combining vast areas of shallow Duck moulting in lagoon systems and open water with dense shoreline cover. Even demonstrated that while flocks were smaller lakes may attract several species. consistently observed in the same locations, Moulting Barrow’s Goldeneye often form there was considerable turnover of loose groups with moulting Canvasback and individuals within some flocks. As such, scaup Aythya sp. when resting and tighter aggregations could simply be the result

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 147 of numerous individuals sampling and movements are resident species, female selecting habitat rather than benefitting Anatids that successfully rear a brood, and from flocking behaviour. Hagy and some species such as scoter, King Eider, Kaminski (2012) report for wintering Northern Pintail, and some Aythya species dabbling ducks that they continue to sample that moult on or near the wintering grounds and forage in emergent wetlands in spite of (Sheaffer et al. 2007; Luukkonen et al. 2008; food depletion in the wetlands. Because Oppel et al. 2008). Ducks, geese and swans waterfowl may not be able to assess food also aggregate in biologically productive abundance without sampling, they continue areas in the arctic, such as major river deltas to forage in patches to assess resource to regrow flight feathers, although such abundance. Forage sampling may be an aggregations are also known from temperate adaptive strategy for waterfowl as a regions (e.g. the Volga Delta for Northern consequence of temporal and spatial Pintail; Dobrynina & Kharitonov 2006). dynamics of wetlands used by Anatids. Indeed, although great variability exists Finally, almost nothing is known about among Anatid moulting strategies social intra- and inter-specific interactions throughout the northern hemisphere, during moult, especially of interaction habitat use and selection have evolved to between males and females and between promote individual survival and as rapid adults and sub-adults. as possible growth of remiges to regain flight. Habitat selection For many sexually immature, unpaired or otherwise non-breeding birds, the spring Moult migration: selection of moult migration is functionally a moult migration, location at macro-scales especially for males, because moulting is the Given available food and safety from only major annual-cycle physiological predation during the flightless moult period, process experienced by these non-breeding the default setting for waterfowl ought to be individuals during spring and summer. In to moult on summer areas, typically within species with delayed maturation, sub-adult the breeding range, to conserve energy females often return to their natal grounds expended in migrating. Yet despite this before moult (Eadie & Gauthier 1985; expectation, many waterfowl show a well- Pearce & Petersen 2009). North American developed moult migration (Salomonsen Aythyini pair during spring, but all females 1968), and the review by Hohman et al. probably return to their natal sites whereas (1992) revealed a wide and bewildering males follow mates or use unfamiliar range of moult migration strategies among breeding areas hoping to invest in and within species and populations. In reproduction. Functionally non-breeding North America, there is a general northward sub-adult geese may associate with parents movement into boreal forest and tundra before moving to other areas to moult. For biomes and associated offshore areas, the males of many dabbling and diving duck although notable exceptions to northward species, they are able to exploit the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 148 Habitat use during remigial moult abundance of food in the northern spring females and ducklings. Scoters breed on and summer to undertake a moult migration inland lakes but moult in coastal waters, away from the nesting areas early in the frequenting habitats similar to those used season after having acquired necessary body during winter. Some populations even moult stores. and winter in the same locations (Fox et al. Reproductive success has a major 2008) but most do not (Bordage & Savard influence on the timing of moult migration 1995; Savard et al. 1998). Even species that and moult, so while successful breeding moult on or near breeding areas use geese moult with their offspring on natal different habitats. Moulting scaups and grounds, male and female ducks rarely if goldeneyes use larger lakes for moulting ever do. Breeding failure is typically greatest than during the breeding season. Colonial during egg laying, early incubation, and early Common Eider that frequent coastal waters brood rearing in ducks, so the onset of the during their entire life cycle segregate brood “post-breeding” period may be variable, rearing and moulting habitats (Dieval et al. depending on possible re-nesting effort and 2011). Mergansers and Harlequin Duck that success and species-, sex-, and age-specific breed on rivers, typically avoid these habitats variation. Although individual waterbirds for moulting and regroup on larger lakes are flightless for 3–7 weeks (Dickson et al. (Common Merganser) or in coastal waters 2012), not all individuals moult at the same (mergansers and Harlequins Duck; Pearce et time (Jepsen 1973; Austin & Fredrickson al. 2009; Robertson & Goudie 1999; Mallory 1986; Hohman et al. 1992), so at some & Metz 1999). Adult female Common moulting locations flightless birds can be Merganser may moult with offspring in present over a period of 2–3 months rivers in Alaska, but it is likely a rare event (Dickson et al. 2012). Amongst ducks, sub- resulting from late nest initiation and adults and unpaired adult males generally success (J. Pearce, in litt.). moult first, followed by paired males, One of the earliest studies of waterfowl unsuccessful breeding females and moult migration was by Sven Ekman successful breeding females (Jepsen 1973; (1922) who described movement of Lesser Joensen 1973; Savard et al. 2007; Hogan White-fronted Geese Anser erythropus uphill 2012). Breeding status also influences the within the same area to moult. The timing and duration of moult as was shown same phenomenon was evident amongst in the example of moulting Black Brant in Greenland White-fronted Geese Anser Alaska above (Table 1). albifrons flavirostris in west Greenland, which An early departure from nesting or exploited lowland wetlands during spring brooding females, and an ability to acquire arrival and subsequent breeding period, but body stores and migrate to moulting successively moved to plateau areas to moult habitats, enables male ducks to exploit (Fox & Stroud 1981). In both cases, this productive habitats (not necessarily suitable behaviour likely was a response to the for brood rearing) whilst avoiding intra- successive delay in growth of plants at a specific competition from brood rearing higher altitude, because the Greenland

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 149

White-fronted Geese finally fed on the Svalbard, where the delay in thaw is along a north facing slopes, the site of the most west-east axis, with non- and failed breeders delayed plant growth around the lakes. In travelling to Edgeøya and moulting further the west Greenland study area, adequate east of the core nesting range (Glahder et al. water bodies existed at all altitudes (i.e. 2007). offering safety from predation), but As migration often has some genetic movement of foraging geese uphill basis (Berthold et al. 2003), the same likely is suggested that food availability and quality true for moult migration, albeit mediated by were fundamental in choice of their moult individual reproductive outcome and status. habitat. For example, translocated Canada Geese For most northern breeding geese, the Branta canadensis established in Yorkshire, period of moult occurs simultaneously with England undertook an apparently innate, brood rearing period. Adequate moulting northward moult migration to Scotland habitat exists in close proximity to the (Dennis 1964), perhaps as their conspecifics nesting grounds, yet non-breeding and failed did on the North American continent. breeding Greenland White-fronted Geese However, learning may be important as well, often moult in the same habitats but away as demonstrated by the selection of new from families (Fox & Stroud 1981). When moulting locations. For example, the such habitat is limiting within an area of the Yorkshire Canada Geese subsequently breeding grounds, food limitation may started to moult locally (Garnett 1980), as influence non- or failed-breeders to disperse Canada Geese have started to moult in the to alternative moulting areas. Amongst salt marshes of the St. Lawrence Estuary in Greenland White-fronted Geese, this recent years (Canadian Wildlife Service circumstance involves the non-breeding or unpublished data) and urban sites have been failed nesting birds moving uphill within used by moulting Canada Geese for decades their mountainous summer range to exploit creating management problems (Breault & the delayed thaw and plant growth at McKelvey 1991; Moser et al. 2004). Factors higher altitudes, thereby reducing direct affecting the selection of a moulting competition from breeders and broods (Fox location appear complex. Birds often fly & Stroud 1981). A similar shift but on a far thousands of kilometres to distant moulting greater spatial scale is also evident amongst areas, although they could moult in suitable the Pink-footed Geese that breed in Iceland, nearby areas (Brodeur et al. 2002; Robert et where the non-breeding individuals al. 2008, Chubbs et al. 2008; Savard & undertake a moult migration to north Robert 2013). Some female Common eastern Greenland (Taylor 1953, Mitchell et Eider breeding in the Gulf of St. Lawrence al. 1999). In this case, travelling north to moult along Anticosti Island, an important exploit the delayed arrival of spring growth, moulting location for about 30,000 eiders, an area where the growing season is too located about 100 km from the colony but short to support breeding birds. The same others from the same colony moult in Maine species shows an analogous shift in about 800 km away. Such discrepancies may

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 150 Habitat use during remigial moult reflect reproductive status of the females, or well, for example, a moulting adult male the time at which they abandoned their White-winged Scoter, captured in the St. attempt to breed, not least because the Lawrence Estuary, moulted on the Labrador timing of arrival to moult sites may affect coast the following year and in Hudson Bay the degree to which local density affects the during the third year, whereas five others decision to settle or not or could have some returned to their previous moulting area genetic basis. Nevertheless, how can these in the St. Lawrence (JPLS, unpubl. data). two vastly different strategies have equal Likely, age, sex, pairing status, reproductive fitness benefits, especially given the success and body condition may all interact differential energetic investment needed to and affect selection of a moulting location move between these different areas? (Jepsen 1973; Petersen 1981). In eastern Most waterfowl exhibit site fidelity to North America, Barrow’s Goldeneye moult their moulting location and even their in a variety of habitats from inland lakes to moulting site within a location (Szymczak & estuarine and even marine wetlands (Robert Rexstad 1991; Bollinger & Derksen 1996; et al. 2002; Savard & Robert 2013). An adult Bowman & Brown 1992; Breault & Savard female Barrow’s Goldeneye moulted one 1999; Flint et al. 2000; Phillips & Powell year on an inland lake near James Bay, 930 2006; Knoche et al. 2007), suggesting km from her nesting area and the following local knowledge about conditions during year in the St. Lawrence Estuary, only 132 flightlessness may be an advantage and km from her breeding area (Savard & therefore also a factor in habitat selection. Robert 2013) indicating plasticity in choice Equally, the reliability of appropriate of moulting location and habitat. conditions, such as food resources and lack In many cases, moulting birds occupy of predation, is likely to favour return to habitats that are unsuitable for breeding specific areas (Salomonsen 1968). However, birds because some critical component is some individuals are known to change missing (e.g. nesting habitat nearby) or moult location between years. Female season length is inadequate. However, in breeding philopatry, combined with winter other cases, the habitat seems suitable, but pairing, means that drake waterfowl are breeding birds are geographically separated likely to find themselves in very different from moulting non-breeders. Generally locations annually at the end of egg laying family groups are behaviourally dominant (Peters et al. 2012). For example, a drake over non-breeders and arctic nesting Northern Pintail might pair with a hen that ganders, for instance, will aggressively nests in the prairies in one year and with one displace non-breeding geese, despite being that nests in Alaska the next, confronting numerically outnumbered. But growing that individual with radically different moult goslings require an abundance of high migration conditions (e.g. geographical and quality forage and broods cannot defend nutritional) among years even when their entire home range from competitive reproductive investment does not vary. This foragers. Given that moulting locations are possibility is borne out by observation as traditionally used and individuals show high

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 151 site fidelity, non-breeding birds may have returning to a previously used lake to moult. displaced breeders from these “optimal” As such, site selection can have a fidelity habitats. This hypothesis avoids the “group component where birds show a preference selection” argument that is required to for locations across years, but by explain why moulting birds voluntarily prospecting multiple sites each year, they are abandon breeding habitats, if those able to detect potentially new high quality breeding habitats are the best available. sites allowing them to adapt to habitat change (Flint et al. 2014). Selection of the moult location at the intermediate scale Selection of moult location at the very Given that birds have selected habitat at the fine scale large scale (i.e. via moult migration or not), When birds have selected a wetland they next select habitat at an intermediate complex, how they utilise habitat within scale in terms of actual moulting location. these areas during the flightless period Functionally, this level of choice represents represents their balance between nutrient selection of the watershed or wetland acquisition for maintenance and moult and complex and is likely to be the unit scale to survival. Settlement at a local scale is more which an individual shows high levels of likely to reflect annual habitat and ambient inter-annual site fidelity. The scale of this conditions. For example hydrology, air and selection is ultimately determined by water temperatures, wind exposure, extent wetland size, complexity, and continuity of escape cover, and food availability affect (Lewis et al. 2011a,b). Multiple studies have habitat selection at the fine scale, which as a shown that individuals have generally high result may show lesser levels of inter-annual rates of fidelity to specific wetland site fidelity. Fox & Kahlert (2000) showed complexes, implying generally consistent that food may be broadly distributed within conditions across years; nevertheless, there moulting sites, but birds only utilised forage are several examples where moulting birds in close proximity to escape habitat. Lewis et have shifted distributions or colonised new al. (2011a) did not measure forage areas over time (Flint et al. 2008). availability, but showed that moulting Black Lewis et al. (2010a, b) examined pre-moult Brant used a home range that was a patterns of movements for Black Brant that functional strip of foraging habitat along had undergone a short distance moult shorelines. Further, Lewis et al. (2011a) migration. In that study, individual brant found a relationship between initial body used a range of wetlands over a broad area mass and home range size such that birds before ultimately selecting a specific moult with increased body mass had decreased location. They concluded that patterns home ranges. This pattern fits the notion of movements were consistent with that stored reserves are primarily used to birds functionally prospecting for moult minimise activity during the moult. Thus, locations. In most cases, birds visited a range moulting locations can range from a of potential moult locations, before restricted locality of a few hectares to over

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 152 Habitat use during remigial moult hundreds of square kilometres, depending relatively low levels of human disturbance on body condition of moulting birds, the could cause moulting geese to abandon a size and dispersion of suitable habitats wetland, but uncertainty exists as to what (Joensen 1973; Gilliland et al. 2002), and the extent such disturbance and displacement scale of resolution (i.e. Ungava Bay vs. might have on rates of site fidelity among Labrador coast; southern vs. eastern coast of years and individual survival. However, Ungava Bay; “inlet A” vs. “inlet B”; etc.). intentional regular disturbance precluded Harlequin Duck moult along rocky birds from moulting on specific wetlands in coastlines and feed within a few metres of an urban environment (Castelli & Sleggs the tide line (Robertson & Goudie 1999). 2000). Thus, disturbance may play a role in The configuration of their foraging habitat habitat selection by moulting waterfowl with is basically linear imposing a limit on the demonstrated displacement within years and sizes of home range and moulting flocks potential for displacement among years. The (usually < 50 birds). In contrast, moulting thresholds and stimuli involved should be scoters and eiders forage mostly in subtidal the focus of research attention in the future zones at depths < 10 m, which vary greatly to improve our ability to undertake impact in area; thus, their home range size is likely assessments and provide management determined by bathymetry and moulting recommendations. flocks can reach thousands of birds. Adaptation to change Effects of disturbance There is little evidence that moulting It has been noted that most species of habitats may be limiting, but habitat loss waterfowl tend to moult in relatively could potentially result in limits on moulting undisturbed locations. While flocks were habitat availability. We have mentioned the observed to respond strongly to disturbance expanding populations of geese that have stimuli, Lacroix et al. (2003) found no clear fully occupied the available freshwater effect of a localised seismic survey on lochs as potential moult sites and have displacement of moulting Long-tailed commenced moulting in marine waters Duck. In cases of persistent harassment, where they are exposed to novel predation such disturbance may lead to drowning, by marine mammals (Glahder et al. 2007). considering flightless birds with incomplete This phenomenon suggests exposure to a plumage are less efficient divers, have lower new source of mortality as a result of intra- thermal efficiency, and probably less specific competition and density dependent buoyant than fully feathered birds. Comeau processes on land. We know little evidence (1923) reported fishermen harassing to support the idea that inter-specific flightless birds until they drowned. Derksen competition could also impose limitations et al. (1982) reported that moulting geese on moulting birds. The only evidence comes formed tight flocks and immediately ran to from two studies of interactions between water when disturbed. Further, Madsen goose species moulting in the same arctic (1984) and Derksen et al. (1982) noted that areas. The first is the study of Madsen &

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 153

Mortensen (1987), which showed that Moulting waterfowl are also faced with allopatric Barnacle Geese and Pink-footed broadscale ecological changes as dictated by Geese in Greenland fed on the same global climate or more localised weather graminoid plants, but more moss occurred dynamics (e.g. Pacific Decadal Oscillation, in the diet under sympatry, especially North Atlantic Oscillation; Schummer et al. amongst Barnacle Geese. Barnacle Geese 2014). Such changes may allow for range spent more time feeding in sympatry than expansion or invoke range contraction. For when feeding alone, while Pink-footed example, distributional shifts of moulting Geese showed no change. The other case is geese along the Arctic Coastal Plain may the study of the interactions between the have been caused by climatic ecological endemic Greenland White-fronted Geese in changes (Flint et al. 2008; Lewis et al. 2011b; west Greenland and recently colonised Tape et al. 2013). Further, recent surveys Canada Geese Branta canadensis interior to this document a substantial range expansion for region. In this case, both species fed on moulting Black Brant in this area (Flint et al. similar plants in allopatry, but where they 2014). Ground observations document an moulted together, Canada Geese showed no increase in grazing lawns along the coast and diet shift, whilst White-fronted Geese fed experimental manipulations demonstrate on lower quality forage, such as moss, and that longer, warmer summers likely tended to feed on the periphery of areas increase forage plant productivity. Thus, where they would feed in isolation even environmental changes appear to have though little overt aggression was witnessed influenced the distribution and abundance between the species (Kristiansen & Jarrett of forage plants, and moulting geese have 2002). Although such a shift in diet is not expanded their range accordingly. evidence of competitive interactions, such a Habitat loss and degradation are likely to mechanism may have fitness consequences, become an increasing challenge, for given that feeding at the study site was example, as a result of loss of boreal forest restricted to 200 m of the edge of water habitat, changes in precipitation, oil and gas where food may have become limited exploration. Yet we know almost nothing because geese were reluctant to forage away about how waterfowl adapt when wetlands from water to which they would retreat and food resources used for moult are no when threatened by predators such as Arctic longer available. Fox. Other studies of the same species, in extensive wetlands where feeding limitations Post-moult requirements may not be so manifest, found little We have already discussed the influence of evidence for such shifts in diet and no age, sex, body condition, breeding success, evidence of changes in local abundance of and environmental resources on habitat White-fronted Geese in the face of selection during the flightless period. increasing Canada Geese (Levermann & Further, wing moult does not appear to be a Raudrup undated; Boertmann & Egevang period of particularly high mortality, nor 2002). does it necessarily expose birds to unusually

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 154 Habitat use during remigial moult high energetic constraints. Nonetheless, Black Brant functionally exhausted all moulting physiological ecology may have available stored lipids, brood rearing birds cross-seasonal effects on subsequent may not finish moult in worse condition. annual-cycle events. Here we consider the Thus, brood rearing birds may finish moult potential effects of post-moult requirements in comparable conditions, but at a later date on habitat selection and body mass leaving less time to rebuild reserves for dynamics during the moult. Following wing migration. Therefore, the primary cost of moult, the next critical period in the life wing moult for breeding birds is likely cycle of waterfowl is completion of the related to time between completion of wing body moult (Pyke 2005) and fall migration moult and onset of fall migration. and associated staging. Fifty years ago, Given time constraints, some portion of Harold Hanson (1962) speculated that habitat selection during moult may be “…there could be little doubt that stress of dictated by habitat needs immediately remigial moult is particularly heavy on following wing moult. Lewis et al. (2010a,b) females, following the energy demands of showed that Black Brant departed inland egg laying and care of young. It seems moulting lakes and moved to nearby coastal possible that the apparent differential stress estuaries upon resumption of flight. of moult may be a primary reason for the Importantly, this shift occurred before flight preponderance of males in populations of feathers were fully grown. Thus, apparent adult waterfowl”. Hanson’s general point is habitat suitability changes when birds regain likely highly relevant for many female ducks flight. They then stage in these coastal that invest heavily (especially somatic lipids) estuaries for several weeks before initiating in a clutch, incubation and rearing offspring, autumn migration. This phenomenon raises followed by the energetic challenges of the possibility that habitat selection during remigial moult and hyperphagia in the moult was influenced by proximity of preparation for autumn migration. This has suitable staging areas or individual ability to recently been confirmed in Mallard, where exploit moult habitats that facilitate foraging radio tracking of brood rearing females and body condition to enable flight to more showed “…reduced survival of females that distant post-moult staging areas without raise broods presumably resulted from fitness costs. Several studies have also insufficient time to moult and prepare for demonstrated that while birds lose mass fall migration” (Arnold & Howerter 2012). during the moult they begin to gain mass as The key point here is likely the “time” soon as they regain flight (Brown & Saunders element. For Black Brant, brood rearing and 1998; Howell 2002). The behaviour of birds failed or non-breeding birds moult on about following wing feather moult seems variable the same date (9 July) but brood rearing with some moving out of the moulting area birds have slower feather growth resulting in (Lewis et al. 2010a,b) and others remaining at a longer flightless period (Taylor 1995; the same location until autumn migration Singer et al. 2012). Given that Taylor (1993) (Brodeur et al. 2002; Robert et al. 2002). Some demonstrated that non-breeding moulting waterfowl also complete body moult at or

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 155 near their wing moult location taking reduce their general level of activity, yet advantages of productivity of these sites. We whilst many authors speculate on the fitness know very little of this period of the life costs of various behaviours during moult, cycle of waterfowl. We believe that cross- none have actually shown the magnitude of seasonal effects are important factors the relative costs and benefits of these in determining habitat use by moulting different strategies to survival and fulfilling waterfowl, but we suggest that it is the period the life-cycle. If a moulting waterbird immediately following moult that likely has drastically reduces feeding activity, what is the greatest effect. the cost (in terms of reduced energetic intake) versus the gain (in terms of reduced Conclusions energy expenditure and elevated survival Post-breeding waterfowl appear to select probability)? With the exception of Kahlert moult habitats on the basis of avoiding: (i) (2006a, which could not address survival predation whilst unable to fly, and (ii) the probability) there have been few attempts to nutritional stress associated with flight empirically demonstrate the effects of feather replacement (especially amongst adopting such strategies. brood rearing female ducks). Indeed, We should also consider that perhaps a habitats will be selected to reduce risk of very important factor in choosing a moult mortality to an absolute minimum; this site is not merely that the individual emerges outcome is borne out by studies that from remigial moult with a new set of flight confirm high survival during moult. Despite feathers but also attains improved body the apparent wealth of references cited here, composition, organ status, and condition this part of the annual cycle remains poorly for onward movement to the next stage in studied and even less well understood, the annual cycle. We also suggest that particularly with respect to habitat selection. researchers reconsider the interpretation of Waterfowl show a remarkable array of mass loss during moult. Generally, we find structural and metabolic changes associated support for the hypothesis that mass loss is with moult which must confirm this period an adaptation to minimise the length of the as critical to completion of the life cycle, flightless period. However, this review also because waterfowl remiges and plumages demonstrates that local conditions influence must be replaced annually. Different rates of mass loss. Mass loss is variable waterfowl species adopt a wide range of within and among species, among years, potential mechanisms to acquire adequate locations, ages, sexes and breeding success. nutrients and energy to survive flight feather In cases where forage is abundant, mass loss moult, presumably as rapidly as possible, may be behaviourally driven, stimulating without compromise to the quality of the birds to forage and complete moult at an feather structures (although this is not increased rate. Conversely, when forage is known) that must support the bird inadequate, mass loss may be compulsory, throughout the coming year. Increasingly, but the associated early return to flight observations suggest moulting Anatidae allows birds to seek more favourable

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 156 Habitat use during remigial moult conditions. Accordingly, we suggest that phenotypic plasticity and the ways they may mass loss is a mechanism used by moulting go about reducing energy expenditure, waterfowl or a consequence of moulting whilst others (such as non-breeding arctic and environmental conditions to adapt to geese) have little difficulty in balancing variable constraints. As such, mass loss energy budgets during the replacement of cannot be used to infer habitat limitation. remigial feathers. Yet even within species, An especially exciting research prospect individuals may adopt different strategies at would be to use manipulative experiments different stages of their own lives, to see how movement from selected habitats dependent on breeding status, confirming to unoccupied ones might affect mass loss that not all individuals or species have the during wing moult. Samples of flightless same goals through to the end of moult. birds could be captured and moved Faced with the challenges of moult, to similar but unoccupied habitats to waterfowl may take multiple routes to compare survival, body mass dynamics, and achieve the same end. An individual behaviour with those left to remain at selecting between these alternatives is likely selected habitats. Such approaches might to do so according to its internal state and also challenge the notion that social the environmental factors that it encounters, dynamics combined with site fidelity so we need to better understand this part of may have reduced the possibilities of the process before we can define concrete colonization of potentially suitable moulting research/management options for specific habitats not currently exploited. species in specific situations. Finally, our narrow focus on remigial Faced with such massive variation and moult in this review has perhaps had the uncertainty, but confronted by an urgent effect of emphasising loss of flight (and the need to think about developing associated survival risks and foraging conservation policy to address the needs of challenges) at the expense of nutritional waterfowl moult, we suggest some priorities aspects of body moult in the Anatidae for actions in the immediate future. The first generally. Completion of body moult (i.e. is to define moult habitat for each other than flight feathers) in other regions population flyway and identify the larger or and other times of year also has important most sensitive concentrations (but beware consequences for performance of other the likelihood of rapid changes in feeding annual cycle events (not least for attracting a ecology and use of moulting resorts, e.g. mate) and therefore fitness and deserves far Nilsson et al. 2001). Such a simple inventory greater research attention. exercise, to know “how many and where”, would provide a framework for establishing Conservation and management site-safeguard networks and potentially implications identify critical habitat types to ensure What is clear from this synthesis is that there these figure prominently in the design of is no “one size fits all”. We see different site safeguard networks as well as in land- species with contrasting patterns in use planning and environmental impact

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 157 assessment where these habitats are susceptibility to human activities that threatened. The second priority is to increasingly encroach on moulting habitats. attempt to define the moult strategies for However, we still know very little about each population, for example determining if threshold levels that can effect change, such the various subsets of a species are moult as shifts in moulting location or strategies migrants, depend upon fat stores for energy between seasons as a result of disturbance. supply or on local resources. Particular focus We know that this happens, because there should be placed upon breeding females, are documented cases of Greylag Goose because of the nutritional and energetic moulting grounds being completely stresses on them following incubation and abandoned from one year to another (e.g. brood rearing, and the specific implications Nilsson et al. 2001). These aspects need to of their moult patterns for habitat be the subject of much greater study to protection, protection from disturbance establish management prescriptions for and hunting. This will affect whether more effective management of moulting site safeguard and/or protection from sites where human disturbance is a factor in disturbance is more important to be affecting carrying capacity. There are some implemented on the moulting grounds or indications that flightless sea ducks are on the post-breeding pre-moulting habitats extremely vulnerable to disturbance, where fat stores are accumulated. It is also especially because of their reduced diving important to: 1) assess the degree of capacities (Comeau 1923), thus moulting wetland loss and global change (including sites may need special consideration. The climate) that affect moult behaviour and exploitation of moulting flocks for distribution, 2) determine the effects of subsistence occurred in the north (King internal state specificity on the strategies 1973) but its current level is unknown as is adopted by different individuals at different its impact at the population level. Little work stages of the life cycle, 3) design site has been done on the carrying capacity of safeguard and habitat management moulting locations and its annual variability, programmes to optimise the availability and yet such knowledge is essential if we are to suitability of moult habitat along the flyways be able to provide management advice in of the northern hemisphere but that can generic or specific case of conflict. Gill net also accommodate change, and 4) define fisheries at important moult sites should be relationships between breeding, moult and managed to avoid by-catch casualties. Also wintering locations (pooling present and contingency plans are needed at all future knowledge from satellite telemetry important moulting locations to minimise studies for example). the impact of possible oil spills. Northern moulting locations both inland and coastal Management implications and policy have not all been identified. As northern needs development is likely to increase as a result Moulting waterfowl are sensitive to of climate warming it is urgent to identify all disturbance and therefore show enhanced sites of particular significance to flyway

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 158 Habitat use during remigial moult populations. In particular, there are Hupp, J.W. 2012. Predation rates, timing, and increasing pressures on boreal forest, near- predator composition for scoters (Melanitta shore, and arctic biomes posed by mining, spp.) in marine habitats. Canadian Journal of oil and gas exploration and exploitation, Zoology 90: 42–50. agriculture, timber harvest, direct and Ankney, C.D. 1979. Does the wing moult cause nutritional stress in lesser snow geese? Auk indirect (e.g. transport, contaminants) 96: 68–72. impacts of oil, disease and contaminants, Ankney, C.D. 1984. Nutrient reserve dynamics of and the general increase in accessibility to breeding and moulting brant. Auk 101: these areas that brings associated levels of 361–370. disturbance and recreational pressures. Ankney, C.D. & Afton, A.D. 1988. Bioenergetics of breeding Northern Shovelers: diet, Acknowledgements nutrient reserves, clutch size and incubation. We are deeply grateful to a great many Condor 90: 459–472. colleagues and contacts for their Arnold, T.W. & Howerter, D.W. 2012. Effects of contributions to discussions about Anatidae radio transmitters and breeding effort on harvest and survival rates of female mallards. moult throughout our careers that have Wildlife Society Bulletin 36: 286–290. contributed to this review and we especially Austin, J.E. & Fredrickson, L.H. 1986. Molt of acknowledge the contributions of Hugh female Lesser Scaup immediately following Boyd, Fred Cooke, Johnny Kahlert, Roy breeding. Auk 103: 293–298. King, Myrfyn Owen, Steve Portugal and Bailey, R.O. 1981. The postbreeding ecology of the Sievert Rohwer. Thanks to Rick Kaminski Redhead Duck (Aythya americana) on Long Island for the invitation to contribute to the Bay, Lake Winnipegosis, Manitoba. Ph.D. thesis, ECNAW conference in Memphis, for McGill University, Montréal, Canada. engineering the production of proceedings Bailey, R.O. 1983a. Distribution of postbreeding from that meeting, encouraging our diving ducks (Aythyini and Mergini) on contribution and contributing significantly southern boreal lakes in Manitoba. Canadian to the quality of this review. Our grateful Wildlife Service Progress Notes 136: 1–8. Bailey, R.O. 1983b. Use of southern boreal lakes thanks go to John Pearce and Jane Austin as by moulting and staging diving ducks. In H. well as the editor and referees Johnny Boyd (ed.), Proceeding of the First Western Kahlert and Matthieu Guillemain for Hemisphere Waterfowl and Waterbird Symposium, improving earlier drafts. pp. 54–59. International Wetlands and Waterbird Research Bureau, Slimbridge, UK. References Bellrose, F.C. 1980. Ducks, Geese and Swans of Adams, P.A., Robertson, G.J. & Jones, I.L. 2000. North America. Stackpole Books, Harrisburg, Time-activity budgets of Harlequin Ducks USA. molting in the Gannet Islands, Labrador. Bergman, G. 1982. Why are the wings of Larus f. Condor 102: 703–708. fuscus so dark? Ornis Fennica 59: 77–83. Anderson, E.M., Esler, D., Boyd, W.S., Evenson, Berthold, P., Gwinner, E. & Sonnenschein, E. J.R., Nysewander, D.R., Ward, D.H., Dickson, (eds.). 2003. Avian Migration. Springer-Verlag, R.D., Uher-koch, B.D., van Stratt, C.S. & Berlin, Germany.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 159

Bielefeld, R.R. & Cox, R.R. 2006. Survival and Bridge, E.S. 2007. Influence of morphology and cause-specific mortality of adult female behavior on wing-moult strategies in Mottled Ducks in East-central Florida. seabirds. Marine Ornithology 34: 7–19. Wilson Society Bulletin 34: 388–394. Brodeur, S., Savard, J.-P.L., Robert, M., Laporte, Boertmann, D. & Egevang, C. 2002. Canada P. Lamothe, P., Titman, R.D., Marchand, S., Geese Branta canadensis in West Greenland: in Gilliland, S. & Fitzgerald, G. 2002. Harlequin conflict with Greenland White-fronted Geese Duck Histrionicus histrionicus population Anser albifrons flavirostris? Ardea 90: 335–336. structure in eastern Nearctic. Journal of Avian Boertmann, D. & Mosbech, A. 2002. Moulting Biology 33: 127–137. Harlequin Ducks in Greenland. Waterbirds 25: Brown, R.E. & Saunders, D.K. 1998. Regulated 326– 332. changes in body mass and muscle mass in Bollinger, K.S. & Derksen, D.V. 1996. moulting Blue-winged Teal for an early return Demographic characteristics of moulting to moult. Canadian Journal of Zoology 76: 26–32. Black Brant near Teshekpuk Lake, Alaska. Buffard, E. 1995. Anti-predator behaviour of Journal of Field Ornithology 67: 141–158. flightless Kerguelen Pintail Anas eatoni Booth, C. J. & Ellis, P. 2006. Common Eiders and moulting in a cave on the Kerguelen Common Guillemots taken by killer whales. archipelago. Wildfowl 46: 66–68. British Birds 99: 533–535. Cargill, S.M. & Jefferies, R.L. 1984. The effects of Bordage, D. & J.-P.L. Savard. 1995. Black scoter grazing by Lesser Snow Geese on the (Melanitta nigra). In A. Poole & F. Gill (eds.), vegetation of a sub-arctic salt marsh. Journal The Birds of North America, No. 177. The of Applied Ecology 21: 669–686. Academy of Natural Sciences, Philadelphia Castelli, P.M., & Sleggs, S.E. 2000. Efficacy of and the American Ornithologists Union, border collies to control nuisance Canada Washington DC, USA. Geese. Wildlife Society Bulletin 28: 385–392. Bowman, T.D. 1987. Ecology of male Black Ducks Chubbs, T.E., Trimper, P.G., Humphries, G.W., moulting in Labrador. M.Sc. thesis, University Thomas, P.W., Elson, L.T. & Laing, D.K. of Maine, Orono, Maine, USA. 2008. Tracking seasonal movements of adult Bowman, T.D. & Brown, P.W. 1992. Site fidelity male Harlequin Ducks from central Labrador of Black Ducks to a moulting area in using satellite telemetry. Waterbirds 31(Special Labrador. Journal of Field Ornithology 63: Publication 2): 173–182. 32–34. Comeau, N.A. 1923. Notes on the diving of Breault, A. & McKelvey, R.W. 1991. Canada Geese loons and ducks. Auk 40: 525. in the Fraser Valley: A Problem Analysis. Craik, S.R., Savard, J.-P.L. & Titman, R.D. 2009. Technical Report Series No. 133. Canadian Wing and body molts of male Red-breasted Wildlife Service, Pacific and Yukon Region, Mergansers in the Gulf of St. Lawrence, Vancouver, Canada. Canada. Condor 111: 71–80. Breault, A.M. & Savard, J.-P.L. 1999. Philopatry Craik, S.R., Savard, J.-P.L. Richardson, M.J. & of Harlequin Ducks moulting in southern Titman, R.D. 2011. Foraging ecology of British Columbia. In R.I. Goudie, M.R. flightless male red-breasted Mergansers in Petersen & G.J. Robertson (eds.), Behaviour the Gulf of St. Lawrence, Canada. Waterbirds and Ecology of Sea Ducks, pp. 41–44. Canadian 34: 280–288. Wildlife Service Occasional Paper No. 100. Davis, J.B., Kaminski, R.M., Leopold, B.D. & Canadian Wildlife Service, Ottawa, Canada. Cox, R.R., Jr. 2001. Survival of female Wood

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 160 Habitat use during remigial moult

Ducks during brood rearing in Alabama and & D.A. Stroud (eds.), Waterbirds Around Mississippi. Journal of Wildlife Management 65: the World. The Stationery Office, Edinburgh, 738–744. UK. Davis, J.B., Cox, R.R., Jr., Kaminski, R.M. & Dolnik, V.R. & Gavrilov, V.M. 1979. Leopold, B.D. 2007. Survival of Wood Duck Bioenergetics of moult in the chaffinch ducklings and broods in Mississippi and (Fringilla coelebs). Auk 96: 253–264. Alabama. Journal of Wildlife Management 71: Döpfner, M., Quillfeldt, P. & Bauer, H.-G. 2009. 507–517. Changes in behavioral time allocation of Dennis, R. 1964. Capture of moulting Canada waterbirds in wing-moult at Lake Constance. Geese in the Beauly Firth. Wildfowl Trust Waterbirds 32: 559–571. Annual Report 15: 71–74. Douthwaite, R.J. 1976. Weight changes and wing Derksen, D.V., Eldridge, W.D. & Weller, M.V. moult in the Red-billed Teal. Wildfowl 27: 1982. Habitat ecology of Pacific Black Brant 123–27. and other geese moulting near Teshekpuk DuBowy, P.J. 1985. Seasonal organ dynamics in Lake, Alaska. Wildfowl 33: 39–57. post-breeding male Blue-winged Teal and Dickson, R.D. 2011. Postbreeding ecology of Northern Shovelers. Comparative Biochemistry White-winged Scoters (Melanitta fusca) and and Physiology 82A: 899–906. Surf Scoters (M. perspicillata) in western Eadie, J.M. & Gauthier, G. 1985. Prospecting for North America: wing moult phenology, body nest sites by cavity-nesting ducks of the mass dynamics and foraging behaviour. Genus Bucephala. Condor 87: 528–534. M.Sc. thesis, Simon Fraser University, Einarsson, A. & Gardarsson, A. 2004. Moulting Vancouver, Canada. diving ducks and their food supply. Aquatic Dickson, R.D., Esler, D., Hupp, J.W., Anderson, Ecology 38: 297–307. E.M., Evenson, J.R. & Barrett. J. 2012. Ekman, S. 1922. Djürvärldens Utbredningshistoria på Phenology and duration of remigial moult in den Skandinaviska halvön. Bonnier, Stockholm, Surf Scoters (Melanitta perspicillata) and Sweden. White-winged Scoters (Melanitta fusca) on the Elgar, M.A. 1989. Predation, vigilance and group Pacific coast of North America. Canadian size in mammals and birds: a critical review Journal of Zoology 90: 932–944. of the empirical evidence. Biological Reviews Diéval, H. 2006. Répartition de l’Eider à duvet 64: 13–33. pendant les périodes d’élevage des jeunes et Evelsizer, D.D., Clark, R.G. & Bollinger, T.K. de mue des adultes le long du fleuve Saint- 2010. Relationships between local carcass Laurent. M.Sc. thesis, University of Québec density and risk of mortality in molting in Montréal, Montréal, Canada. mallards during avian botulism outbreaks. Diéval, H., Giroux, J.-F. & Savard, J.-P.L. 2011. Journal of Wildlife Diseases 46: 507–513. Distribution of common eiders Somateria Ferrell, C.L. 1988. Contribution of visceral mollissima during the brood-rearing and organs to animal energy expenditures. Journal moulting periods in the St. Lawrence of Animal Science 66: 23–34. Estuary. Wildlife Biology 17: 124–134. Fleskes, J.P., Mauser, D.M., Yee, J.L., Blehert, Dobrynina, I.N. & Kharitonov, S.P. 2006. D.S. & Yarris, G.S. 2010. Flightless and The Russian waterbird migration atlas: post-moult survival and movements of temporal variation in migration routes, pp. female Mallards moulting in Klamath Basin. 582–589. In G.C. Boere, C.A. Galbraith Waterbirds 33: 208–220.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 161

Flint, P.L., Petersen, M.R., Dau, C.P., Hines, J.E. Fox, A.D. & Kahlert, J. 1999. Adjustments to & Nichols, J.D. 2000. Annual survival and site nitrogen metabolism during wing moult in fidelity of Steller’s eiders molting along the greylag geese. Functional Ecology 13: 661–669. Alaska Peninsula. Journal of Wildlife Fox, A.D. & Kahlert, J. 2000. Do moulting Management 64: 261–268. Greylag Geese Anser anser forage in proximity Flint, P.L., Reed, J.A., Franson, J.C., Hollmén, to water in response to food availability T.E., Grand, J.B., Howell, M.D., Lanctot, and/or quality? Bird Study 47: 266–274. R.B., Lacroix, D.L. & Dau, C.P. 2003. Fox, A.D. & Kahlert, J. 2003. Repeated grazing of Monitoring Beaufort Sea Waterfowl and Marine a salt marsh grass by moulting greylag geese Birds. U.S. Geological Survey Report, OCS Anser anser – does sequential harvesting Study MMS 2003–037. USGS, Alaska optimise biomass or protein gain? Journal of Science Center, Anchorage, Alaska, USA. Avian Biology 34: 89–96. Flint, P.L., Lacroix, D.L., Reed, J.A. & Lanctot, Fox, A.D. & Kahlert, J. 2005. Changes in body R.B. 2004. Movements of flightless Long- organ size during wing moult in non- tailed Ducks during wing molt. Waterbirds 27: breeding Greylag Geese on Saltholm, 35–40. Denmark. Journal of Avian Biology 36: 538– Flint, P.L., Mallek, E., King, R.J., Schmutz, 548. J.A., Bollinger, K.S. & Derksen, D.V. Fox, A.D. & King, R. 2011. Body mass loss 2008. Changes in abundance and spatial amongst moulting Pochard Aythya ferina and distribution of geese molting near Tufted Duck A. fuligula at Abberton Teshekpuk Lake, Alaska: interspecific Reservoir, South-east England. Journal of competition or ecological change? Polar Ornithology 152: 727–732. Biology 31: 549–556. Fox, A.D. & Stroud, D.A. 1981. Report of the Flint, P.L., Meixell, B.W. & Mallek, E.J. 2014. 1979 Greenland White-fronted Goose Study High fidelity does not preclude colonization: Expedition to Eqalungmiut Nunaat, West range expansion of molting Black Brant on Greenland. Greenland White-fronted Goose the Arctic coast of Alaska. Journal of Field Study (GWGS), Aberystwyth, UK. Ornithology 85: 75–83. Fox, A.D., Kahlert, J. & Ettrup, H. 1998. Diet and Folk, C., Hudec, K. & Toufar, J. 1966. The weight habitat use of moulting Greylag Geese Anser of the mallard, Anas platyrhynchos, and its anser on the Danish island of Saltholm. Ibis changes in the course of the year. Zoologicke 140: 676–683. Listy 15: 249–250. Fox, A.D., Kahlert, J., Walsh, A.J., Stroud, D.A., Follestad, A., Larsen, B.H., Nygard, T. & Rov, N. Mitchell, C., Kristiansen, J.N. & Hansen, E.B. 1988. Estimating numbers of moulting 1999. Patterns of body mass change during eiders Somateria mollissima with different flock moult in three different goose populations. size and flock structure. Fauna Norvegica Series Wildfowl 49: 45–56. C, Cinclus 11: 97–99. Fox, A.D., Hartmann, P. & Petersen, I.K. 2008. Fondell, T.F., Flint, P.L., Schmutz, J.A., Schamber, Body mass and organ size change during J.L. & Nicolai, C.A. 2013. Variation in body wing moult in common scoter. Journal of mass dynamics among moult sites in Brant Avian Biology 39: 35–40. Geese Branta bernicla nigricans supports Fox, A.D., Hobson, K.E. & Kahlert, J. 2009. adaptivity of mass loss during moult. Ibis Isotopic evidence for endogenous protein 155: 593–604. contributions to Greylag Goose Anser anser

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 162 Habitat use during remigial moult

flight feathers. Journal of Avian Biology. 40: moulting areas along the Beaufort Sea coast 108–112. from Prudhoe Bay, Alaska to Shingle Point, Fox, A.D., Fox, G.F., Liaklev, A. & Gerhardsson, Yukon Territory, July, 1973. Arctic Gas N. 2010. Predation of flightless pink-footed Biological Report, Serial 26(1): 1–61. geese Anser brachyrhynchus by Atlantic Guillemette, M., Pelletier, D., Grandbois, J.M. & walruses Odobenus rosmarus rosmarus in Butler, P.J. 2007. Flightlessness and the southern Edgeøya, Svalbard. Polar Research energetic cost of wing moult in a large sea 29: 455–457. duck. Ecology 88: 2936–2945. Fox, A.D., King, R. & Owen, M. 2013. Wing Hagy, H.M. & Kaminski, R.M. 2012. Winter moult and mass change in free-living Mallard waterbird and food dynamics in autumn- Anas platyrhynchos. Journal of Avian Biology 44: managed moist-soil wetlands in the 1–8. Mississippi Alluvial Valley. Wildlife Society Frimer, O. 1994. The behaviour of moulting King Bulletin 36: 512–523. Eiders Somateria spectabilis. Wildfowl 45: 176–187. Hanson, H.C. 1962. The dynamics of condition Frimer, O. 1995. Comparative behaviour of factors in Canada geese and their relation to sympatric moulting populations of Common seasonal stresses. Arctic Institute of North Eider Somateria mollissima and King Eider America Technical Paper 12: 1–68. Somateria spectabilis in central West Greenland. Hartman, G. 1985. Foods of male Mallard before Wildfowl 46: 129–139. and during moult, as determined by faecal Garnett, M.G.H. 1980. Moorland breeding and analysis. Wildfowl 36: 65–71. moulting of Canada Geese in Yorkshire. Bird Heitermeyer, M.E. 1988. Protein costs of the Study 27: 219–226. prebasic moult of female Mallards. Condor Gauthier, J. & Bédard, J. 1976. Les déplacements 90: 263–266. de l’Eider commun (Somateria mollissima) dans Henny, C.J., Rudis, D.D., Roffe, T.J. & Robinson- l’estuaire du Saint-Laurent. Canadian Wilson, E. 1995. Contaminants and sea Naturalist 103: 261–283. ducks in Alaska and the circumpolar region. Gehrold, A. & Köhler, P. 2013. Wing-moulting Environmental Health Perspectives 103 (Suppl. waterbirds maintain body condition under No. 4): 41–49. good environmental conditions: a case study Hogan, D. 2012. Postbreeding ecology of Barrow’s of Gadwalls (Anas strepera). Journal of goldeneyes in northwestern Alberta. M.Sc. Ornithology 154: 783–793. thesis, Simon Fraser University, Burnaby, Gilliland, S.G., Robertson, G.J., Robert, M., Canada. Savard, J.-P.L., Amirault, D., Laporte, P. Hogan, D., Tompson, J.E., Esler, D. & Boyd, W.S. & Lamothe, P. 2002. Abundance and 2011. Discovery of important postbreeding distribution of Harlequin Ducks moulting in sites for Barrow’s Goldeneye in the boreal eastern Canada. Waterbirds 25: 333–339. transition zone of Alberta. Waterbirds 34: Glahder, C.M., Fox, A.D., O’Connell, M., 261–388. Jespersen, M. & Madsen, J. 2007. Eastward Hogan, D., Tompson, J.E. & Esler, D. 2013a. moult migration on non-breeding Svalbard Survival of Barrow’s Goldeneyes during Pink-footed Geese (Anser brachyrhynchus) in remigial molt and fall staging. Journal of Svalbard. Polar Research 26: 31–36. Wildlife Management 77: 701–706. Gollop, M.A. & Richardson, W.J. 1974. Inventory Hogan, D., Tompson, J.E. & Esler, D. 2013b. and habitat evaluation of bird breeding and Variation in body mass and foraging effort of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 163

Barrow’s Goldeneyes (Bucephala islandica) (Bucephala clangula) in Denmark. Danish Review during remigial molt. Auk 130: 313–322. of Game Biology 8(6): 1–23. Hohman, W.L. 1993. Body composition Joensen, A.H. 1973. Moult migration and wing- dynamics of ruddy ducks during wing feather moult of seaducks in Denmark. moult. Canadian Journal of Zoology 71: Danish review of Game Biology 8(4): 1–42. 2224–2228. Kahlert, J. 2003. The constraint on habitat use in Hohman, W.L., Ankney, C.D. & Gordon, D.H. wing-moulting Greylag Geese Anser anser 1992. Ecology and management of caused by anti-predator displacements. Ibis postbreeding waterfowl. In B.D.J. Batt, A.D. 145: E45–E52. Afton, M.G. Anderson, C.D. Ankney, D.H. Kahlert, J. 2006a. Effects of feeding patterns on Johnson, J.A. Kadlec & G.L. Krapu (eds.), body mass loss in moulting Greylag Geese Ecology and Management of Breeding Waterfowl, Anser anser. Bird Study 53: 20–31. pp. 128–189. University of Minnesota Press, Kahlert, J. 2006b. Factors affecting escape Minneapolis, USA. behaviour in moulting Greylag Geese Anser Hollmén, T.E., Franson, J.C., Flint, P.L., Grand, anser. Journal of Ornithology 149: 567–577. J.B., Lanctot, R.B., Docherty, D.E. & Wilson, Kahlert, J., Fox, A.D. & Ettrup, H. 1996. H.M. 2003. An adenovirus linked to Nocturnal feeding in moulting Greylag mortality and disease in Long-tailed Ducks Geese Anser anser – an anti-predator (Clangula hyemalis) in Alaska. Avian Diseases 47: response? Ardea 84: 15–22. 1434–1440. Kaminski, R.M. & Weller, M.W. 1992. Howell, M.D. 2002. Molt dynamics of male long-tailed Breeding habitats of Nearctic waterfowl. ducks on the Beaufort Sea. M.Sc. thesis, Auburn In B.D.J. Batt, A.D. Afton, M.G. Anderson, University, Alabama, USA. C.D. Ankney, D.H. Johnson, J.A. Kadlec Howell, M.D., Grand, J.B. & Flint, P.L. 2003. & G.L. Krapu (eds.), Ecology and Body molt of male Long-tailed Ducks in the Management of Breeding Waterfowl, pp. 568–589. near-shore waters of the North Slope, University of Minnesota Press, Minneapolis, Alaska. Wilson Bulletin 115: 170–175. USA. Iverson, S.A. & Esler, D. 2007. Survival of female Kaminski, R.M. & Elmberg, J. 2014. An Harlequin Ducks during wing moult. Journal introduction to habitat use and selection of Wildlife Management 71: 1220–1224. by waterfowl in the northern hemisphere. Jehl, J.R. 1990. Aspects of the moult migration. Wildfowl (Special Issue No. 4): 9–16. In E. Gwinner (ed.), Bird Migration: Physiology King, J.G. 1963. Duck banding in arctic Alaska. and Ecophysiology, pp. 102–113. Springer, Journal of Wildlife Management 27: 356–362. Berlin, Germany. King, J.G. 1973. A cosmopolitan duck moulting Jehl, J.R. & Henry, A.E. 2010. The postbreeding resort; Takslesluk Lake Alaska. Wildfowl 24: migration of Eared Grebes. Wilson Journal of 103–109. Ornithology 122: 217–227. King, R. & Fox, A.D. 2012. Moulting mass Jehl, J.R. & Henry, A.E. 2013. Intra-organ dynamics of female Gadwall Anas strepera flexibility in the eared grebe Podiceps nigricollis and male Wigeon A. penelope at Abberton stomach: A spandrel in the belly. Journal of Reservoir, South East England. Bird Study 59: Avian Biology 44: 97–101. 252–254. Jepsen, P.U. 1973. Studies of the moult migration Kirby, R.E. & Cowardin, L.M. 1986. Spring and and wing-feather moult of the Goldeneye summer survival of female mallards from

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 164 Habitat use during remigial moult

north central Minnesota. Journal of Wildlife following flightless molt. Wilson Journal of Management 50: 38–43. Ornithology. 122: 484–493. Knoche, M.J., Powell, A.N., Quakenbush, L.T., Lewis, T.L., Flint, P.L., Schmutz, J.A. & Derksen, Wooller, M.J. & Phillips, L.M. 2007. Further D.V. 2010b. Pre-moult patterns of habitat evidence for site fidelity to wing molt use and moult site selection by Brent Geese locations by King Eiders: Intergrating stable Branta bernicla nigricans: Individuals prospect isotope analyses and satellite telemetry. for moult sites. Ibis 152: 556–568. Waterbirds 30: 52–57. Lewis, T.L., Flint, P.L., Derksen, D.V. & Schmutz, Köhler, P. &. Köhler, U. 1998. Considerable J.A. 2011a. Fine scale movements and habitat increase of moulting waterfowl in fishponds use of Black Brant during the flightless wing without carp in the Ismaninger Teichgebiet molt in Arctic Alaska. Waterbirds 34: 177–185. (Bavaria, Germany). Sylvia 34: 27–32. Lewis, T.L., Flint, P.L., Derksen, D.V., Schmutz, Kortegaard, L. 1974. An ecological outline of a J.A., Taylor, E.J. & Bollinger, K.S. 2011b. moulting area of teal, Vejlerne, Denmark. Using body mass dynamics to explain long- Wildfowl 25: 134–142. term shifts in habitat use of arctic molting Kristiansen, J.N. & Jarrett, N.S. 2002. Inter- geese: evidence for ecological change. Polar specific competition between Greenland Biology 34: 1751–1762. White- fronted Geese Anser albifrons flavirostris Luukkonen, D.R., Prince, H.H. & Mykut, R.C. and Canada Geese Branta canadensis interior 2008. Movements and survival of molt moulting in West Greenland. Ardea 90: 1–13. migrant Canada Geese from southern Kumari, E. 1979. Moult and moult migration of Michigan. Journal of Wildlife Management 72: waterfowl in Estonia. Wildfowl 30: 90–98. 449–462. Lacroix, D.L., Lanctot, R.B., Reed, J.A. & Madsen, J. 1984. Study of the possible impact of oil McDonald, T.L. 2003. Effect of underwater exploration on goose populations in Jameson seismic surveys on molting male Long-tailed Land, East Greenland. A progress report. Ducks in the Beaufort Sea, Alaska. Canadian Norsk Polarinstitutt Skrifter 181: 141–151. Journal of Zoology 81: 1862–1875. Madsen, J. 1985. Relations between change in Laursen, K., Pihl, S., Durinck, J., Hanse, M., Skov, spring habitat selection and daily energetics H., Frikke, J. & Danielsen, F. 1997. Numbers of Pink-footed Geese Anser brachyrhynchus. and distribution of waterbirds in Denmark. Ornis Scandinavica 16: 222–228. Danish Review of Game Biology 15(1): 1–181. Madsen, J. & Mortensen, C.E. 1987. Habitat Leafloor, J.O., Ankney, C.D. & Risi, K.W. 1996. exploitation and interspecific competition in Social enhancement of wing moult in east Greenland. Ibis 129: 25–44. females Mallards. Canadian Journal of Zoology Mallory, M. & Metz, K. 1999. Common 74: 1376–1378. Merganser (Mergus merganser). In A. Poole & F. Levermann, N. & Raundrup, K. undated. Do the Gill (eds.), The Birds of North America, No. Greenland white-fronted geese stand a chance against 442. The Birds of North America, Inc., the invasive Canadians? Unpublished report, Philadelphia, USA. Biologisk Institut, Copenhagen University, Marks, J.S. 1993. Molt of bristle-thighed curlews Copenhagen, Denmark. [In Danish.] in the northwestern Hawaiian islands. Auk Lewis, T.L., Flint, P.L., Schmutz, J.A. & Derksen, 110: 573–587. D.V. 2010a. Temporal and spatial shifts in Miller, M.R. & Duncan, D.C. 1999. The habitat use by Black Brant immediately Northern Pintail in North America: Status

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 165

and conservation needs of a struggling Murphy, M.E. & King, J.R. 1984b. Dietary sulfur population. Wildlife Society Bulletin 27: 788– amino acid availability and moult dynamics 800. in White-crowned Sparrows. Auk 101: 164– Mitchell, C., Fox, A.D., Boyd, H., Sigfusson, A. & 167. Boertmann, D. 1999. Pink-footed Goose Nilsson, L., Kahlert, J. & Persson, H. 2001. Moult Anser brachyrhynchus: Iceland/Greenland. In J. and moult migration of Greylag Geese Anser Madsen, G. Cracknell & A.D. Fox (eds.), anser from a population in Scania, south Goose Populations of the Western Palearctic. A Sweden, Bird Study 48: 129–138. Review of Status and Distribution, pp. 68–81. O’Connor, M. 2008. Surf Scoter (Melanitta Wetlands International Publ. 48, Wetlands perspicillata) ecology on spring staging grounds International, Wageningen, the Netherlands. and during the flightless period. M.Sc. thesis, National Environmental Research Institute, McGill University, Montréal, Canada. Rønde, Denmark. Oppel, S., Powell, A.N. & Dickson, D.L. 2008. Moorman, T.E., Baldassarre, G.A. & Hess Jr, Timing and distance of king eider migration T.R. 1993. Carcass mass and nutrient and winter movements. Condor 110: 296–305. dynamics of Mottled Ducks during remigial Oring, L. 1964. Behavior and ecology of certain molt. Journal of Wildlife Management 57: ducks during the postbreeding period. Journal 224–228. of Wildlife Management 28: 223–233. Moreau, R.E. 1937. The comparative breeding Owen, M. & Ogilvie, M.A. 1979. Wing moult and biology of the African Hornbills weights of Barnacle Geese in Spitzbergen. (Bucerotidae). Proceedings of the Zoological Society Condor 81: 42–52. of London, Series a – General and Experimental Palmer, R.S. 1972. Patterns of molting. In D.S. 107: 331–346. Farner & J.R. King (eds), Avian Biology. Moser, T.J., Lien, R.D., VerCauteren, K.C., Volume 2, pp. 65–102. Academic Press, New Abraham, K.F., Andersen, D.E., Bruggink, York, USA. J.G., Coluccy, J.M., Graber, D.A., Leafloor, Panek, M. & Majewski, D.P. 1990. Remex growth J.O., Luukkonen, D.R. & Trost, R.E. (eds.) and body mass of Mallards during wing 2004. Proceedings of the 2003 International moult. Auk 107: 255–259. Canada Goose Symposium, 19–21 March 2003. Payne, R.B. 1972. Mechanisms and control of Madison, Wisconsin, USA. moult. In D.S. Farner & J.R. King (eds.), Munro, J.A. 1941. Studies of waterfowl in British Avian Biology. Volume 2, pp. 103–155. Columbia Greater Scaup Duck, Lesser Scaup Academic Press, New York, USA. Duck. Canadian Journal of Research D 19: Pearce, J.M. & Petersen, M.R. 2009. Post-fledging 113–138. movements of juvenile Common Murphy, M.E. & King, J.R. 1982. Amino acid Mergansers (Mergus merganser) in Alaska as composition of the plumage of the White- inferred by satellite telemetry. Waterbirds 32: crowned Sparrow. Condor 84: 435–438. 133–137. Murphy, M.E. & King, J.R. 1984a. Sulfur amino Pearce, J.M., Zwiefelhofer, D. & Maryanski, acid nutrition during moult in the White- N. 2009. Mechanisms of population crowned Sparrow. 2. Nitrogen and sulfur heterogeneity among moulting Common balance in birds fed graded levels of the Mergansers on Kodiak Island, Alaska: sulfur-containing amino acids. Condor 86: implications for genetic assessments of 324–332. migratory connectivity. Condor 111: 283–293.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 166 Habitat use during remigial moult

Peters, J.L, Bolender, K.A. & Pearce, J.M. 2012. flightless moult? Journal of Ornithology 151: Behavioural vs. Molecular sources of conflict 443–448. between nuclear and mitochondrial DNA: Portugal, S. J., Green, J.A., Piersma, T., Eichhorn, the role of male-biases dispersal in a G. & Butler, P.J. 2011. Greater energy stores Holarctic sea duck. Molecular Ecology 21: enable flightless moulting geese to increase 3562–3575. resting behaviour. Ibis 153: 868–874. Petersen, M.R. 1981. Populations, feeding Prince, H.H. 1979. Bioenergetics of post-breeding ecology and moult of Steller’s Eiders. Condor dabbling ducks. In T.A. Bookhout (ed.), 83: 256–262. Waterfowl and wetlands: an integrated review, pp. Petit, D.R. & Bildstein, K.L. 1987. Effect of 103–117. Proceedings of the 1977 Symposium group size and location within the group on of the North Central Section of The Wildlife the foraging behaviour of white ibises. Society. Madison, Wisconsin, USA. Condor 99: 602–609. Pyle, P. 2005. Molts and plumages of ducks Phillips, L.M. & Powell, A.N. 2006. Evidence for (Anatinae). Waterbirds 28: 208–219. wing moult and breeding site fidelity in King Qian, G.H. & Xu, H.F. 1986. Moult and resting Eiders. Waterbirds 29: 148–153. metabolic rate in the common teal Anas crecca Piersma, T. 1988. Breast muscle atrophy and and the shoveler A. clypeata. Acta Zoologica constraints on foraging during the flightless Sinica 32: 68–73. period of wing moulting Great Crested Rail, J.-F. & Savard J.-P.L. 2003. Identification des Grebes. Ardea 76: 96–106. Aires de Mue et de Repos au Printemps des Piersma, T., Dietz, M.W., Dekinga, A., Nebel, S., Macreuses (Melanitta sp.) et de l’Eider à Duvet van Gils, J., Battley, P.F. & Spaans, B. 1999a. (Somateria mollissima) dans l’Estuaire et le Reversible size-changes in stomachs of Golfe du Saint-Laurent. Technical Report Series shorebirds: when, to what extent, and why? No. 408. Canadian Wildlife Service, Quebec Acta Ornithologica 34: 175–181. Region, Quebec, Canada. Piersma, T., Gudmundsson, G.A. & Lilliendahl, Reed, A. 1971. Pre-dusk rafting flights of K. 1999b. Rapid changes in the size of wintering goldeneyes and other diving ducks different functional organ and muscle groups in the Province of Quebec. Wildfowl 22: during refuelling in a long-distance migrating 61–62. shorebird. Physiological and Biochemical Zoology Reed, T.M. & Rocke, T.E. 1992. The role of avian 72: 405–415. carcasses in botulism epizootics. Wildlife Piersma, T., Koolhaas, A. & Dekinga, A. 1993. Society Bulletin 20: 175–182. Interactions between stomach structure and Robert, M., Benoit, R. & Savard, J.-P.L. 2002. diet choice in shorebirds. Auk 110: 552–564. Relationship among breeding, moulting and Portugal, S.J., Green, J.A. & Butler, P.J. 2007. wintering areas of male Barrow’s Goldeneyes Annual changes in body mass and resting (Bucephala islandica) in eastern North America. metabolism in captive barnacle geese (Branta Auk 119: 676–684. leucopsis): the importance of wing moult. Robert, M., Mittelhauser, G.H., Jobin, B., Journal of Experimental Biology 210: 1391– Fitzgerald, G. & Lamothe, P. 2008. New 1397. insights on Harlequin Duck population Portugal, S.J., Isaac, R., Quinton, K.L. & structure in eastern North America as Reynolds, S.J. 2010. Do captive waterfowl revealed by satellite telemetry. Waterbirds 31 alter their behaviour patterns during their (Special Publication 2): 159–172.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 Habitat use during remigial moult 167

Robertson, G.J. & Goudie, R.I. 1999. Harlequin Atlantic flyway resident population of Duck (Histrionicus histrionicus). In A. Poole & Canada Geese, Branta canadensis. Canadian F. Gill (eds.), The Birds of North America, No. Field Naturalist 121: 313–320. 547. The Birds of North America, Inc., Singer, H.V., Sedinger, J.S., Nicolai, C.A., van Philadelphia, USA. Dellen, A.W. & Person, B.T. 2012. Timing of Robertson, G. J., Cooke, F., Goudie, R.I. & Boyd, adult remigial wing molt in female Black Brant W.S. 1998. Moult speed predicts pairing (Branta bernicla nigricans). Auk 129: 239−246. success in male Harlequin Ducks. Animal Sjöberg, K. 1988. The flightless period of free- Behaviour 55: 1677–1684. living male Teal Anas crecca in northern Rohwer, S., Ricklefs, R.E., Rohwer, V.G. & Sweden. Ibis 130: 164–171. Copple, M.M. 2009. Allometry of the Smith, W.E. 2006. Moulting Common Eiders duration of flight feather moult in birds. devoured by killer whales. British Birds 99: 264. PLoS Biol 7(6): e1000132. Stonor, C.R. 1937. On the attempted breeding of Salomonsen, F. 1968. The moult migration. a pair of Trumpeter Hornbills (Bycanistes Wildfowl 19: 5–24. buccinator) in the gardens in 1936; together Savard, J.-P.L. & Robert, M. 2013. Relationships with some remarks on the physiology of the among Breeding, Moulting and Wintering moult in the female. Proceedings of the Zoological Areas of Adult Female Barrow’s Goldeneyes Society of London, Series A – General and (Bucephala islandica) in Eastern North Experimental 107: 89–94. America. Waterbirds 36: 34–42. Stresemann, E. & Stresemann, V. 1966. Die Mauser Savard, J-P.L., Bordage, D. & Reed, A. 1998. Surf der Vögel. Journal of Ornithology 107: 3–448. Scoter (Melanitta perspicillata). In A. Poole & F. Szymczak, M.R. & Rexstad, E.A. 1991. Harvest Gill (eds.), The Birds of North America, No. distribution and survival of a gadwall 363. The Birds of North America, Inc. The population. Journal of Wildlife Management 61: Academy of Natural Sciences, Philadelphia, 191–201. USA. Tamisier, A. & Dehorter, O. 1999. Camargue Savard, J.-P.L., Reed, A. & Lesage, L. 2007. Canards et Foulques. Centre Ornithologique du Chronology of breeding and moult migration Gard. Nimes, France. in Surf Scoters (Melanitta perspicillata). Tape, K.D., Flint, P.L., Meixell, B.W. & Gaglioti, Waterbirds 30: 223–229. B.J. 2013. Inundation, sedimentation and Schenkeveld, L.E. & Ydenberg, R.C. 1985. subsidence creates goose habitat along the Synchronous diving by Surf Scoter flocks. Arctic coast of Alaska. Environmental Research Canadian Journal of Zoology 63: 2516–2519. Letters 8: 045031. Schummer, M.L., Cohen, J., Kaminski, R.M., Taylor, J. 1953. A possible moult-migration of Brown, M.E. & Wax, C.L. 2014. Pink-footed Geese. Ibis 95: 638–641. Atmospheric teleconnections and Eurasian Taylor, E.J. 1993. Molt and bioenergetics of snow cover as predictors of a weather Pacific Black Brant (Branta bernicla nigricans) severity index in relation to Mallard on the Arctic Coastal Plain, Alaska. Ph.D. Anas platyrhynchos autumn–winter migration. thesis, Texas A&M University, College Wildfowl (Special Issue No. 4): 000–000. Station, Texas, USA. Sheaffer, S.F., Malecki, R.A., Swift, B.L., Dunn, Taylor, E.J. 1995. Molt of Black Brant (Branta J. & Scribner, K. 2007. Management bernicla nigricans) on the arctic coastal plain, implications of molt migration by the Alaska. Auk 112: 904–919.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 168 Habitat use during remigial moult

Thompson, D.C. & Boag, D.A. 1976. Effect of Goldeneye during wing molt. Condor 102: moulting on the energy requirements of 228–231. Japanese quail. Condor 78: 249–252. Van Impe, J. 1985. Mues des remiges chez la Thompson, J.D. & Baldassarre, G.A. 1988. Nette a huppe rousse Netta rufina (Pallas) en Postbreeding habitat preference of wood Espagne du Nord. Alauda 53: 2–10. ducks in northern Alabama. Journal of Wildlife Weller, M.W. 1980. 3. Molts and Plumages of Management 52: 80–85. Waterfowl. In F.C. Bellrose Ducks, Geese and Thompson, J.E. & Drobney, R.D. 1996. Swans of North America, pp. 34–38. Stackpole Nutritional implications of moult in male Books, Harrisburg, USA. Canvasbacks: variation in nutrient reserves Wobeser, G. & Leighton, T. 1988. Avian Cholera and digestive tract morphology. Condor 98: epizootic in wild duck. Canadian Veterinary 512–526. Journal 29: 1015–1016. Van der Meer, J. & Piersma, T. 1994. Wunder, M.B., Jehl, J.R. & Stricker, C.A. 2012. Physiologically inspired regression models The early bird gets the shrimp: confronting for estimating and predicting nutrient stores assumptions of isotopic equilibrium and and their composition in birds. Physiological homogeneity in a wild bird population. Zoology 67: 305–329. Journal of Animal Ecology 81: 1223–1232. Van de Wetering, D. 1997. Moult characteristics and Yarris, G.S., McLandress, R. & Perkins, A.E.H. habitat selection of postbreeding male Barrow’s 1994. Molt migration of postbreeding female Goldeneye (Bucephala islandica) in northern mallards from Suisun Marsh, California. Yukon. Technical Report Series Number 296. Condor 96: 36–45. Canadian Wildlife Service, Pacific and Yukon Young, D.A. & Boag, D.A. 1982. Changes in the Region, British Columbia. physical condition of male mallards (Anas Van de Wetering, D. & Cooke, F. 2000. Body platyrhynchos) during moult. Canadian Journal of weight and feather growth of male Barrow’s Zoology 60: 3220–3226.

Photograph: Moulting Brent Geese, by Gerrit Vyn.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 131–168 169

Drivers of waterfowl population dynamics: from teal to swans

DAVID N. KOONS1*, GUNNAR GUNNARSSON2, JOEL A. SCHMUTZ3 & JAY J. ROTELLA4

1Department of Wildland Resources and the Ecology Center, Utah State University, Logan, Utah 84322, USA. 2Division of Natural Sciences, Kristianstad University, SE-291 88 Kristianstad, Sweden. 3U.S. Geological Survey, Alaska Science Center, 4210 University Drive, Anchorage, Alaska 99508, USA. 4Department of Ecology, Montana State University, Bozeman, Montana 59717, USA. *Correspondence author. E-mail: [email protected]

Abstract Waterfowl are among the best studied and most extensively monitored species in the world. Given their global importance for sport and subsistence hunting, viewing and ecosystem functioning, great effort has been devoted since the middle part of the 20th century to understanding both the environmental and demographic mechanisms that influence waterfowl population and community dynamics. Here we use comparative approaches to summarise and contrast our understanding of waterfowl population dynamics across species as short-lived as the teal Anas discors and A.crecca to those such as the swans Cygnus sp. which have long life-spans. Specifically, we focus on population responses to vital rate perturbations across life history strategies, discuss bottom-up and top-down responses of waterfowl populations to global change, and summarise our current understanding of density dependence across waterfowl species. We close by identifying research needs and highlight ways to overcome the challenges of sustainably managing waterfowl populations in the 21st century.

Key words: climate change, demographic buffering and lability, density dependence, ducks, elasticity, environmental stochasticity, geese, population dynamics, swans.

Competition for and selection of available these topics were reviewed by the first three habitats throughout the annual cycle, trophic plenary sessions of the 6th North American interactions and associated life history trade- Duck Symposium and Workshop, “Ecology offs can all affect individual fitness (Dawkins and Conservation of North American & Krebs 1979; Stearns 1992; Manly et al. Waterfowl,” (ECNAW) in Memphis, 2002). Advances in our understanding of Tennessee in January 2013. Ultimately,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 170 Drivers of waterfowl population dynamics factors that have an impact on individual better understanding of the interplay fitness can scale up to affect population between dynamic environmental conditions dynamics via a change in mean fitness and and demography in an era of global change also through the assembly of co-occurring that, fourth, can be augmented or tempered species in a waterfowl community (Ricklefs by density-regulating mechanisms. 2008). The population remains the primary biological unit on which waterfowl Perturbation analyses across the management objectives are based (e.g. the slow–fast continuum in waterfowl North American Waterfowl Management Waterfowl have evolved a diverse array of Plan; Canadian Wildlife Service & U.S. Fish life histories: teal live a life almost as short as and Wildlife Service 2012), but a greater focus many passerines, whereas many swans and on the community of waterfowl species is geese have annual survival rates of ~0.90 starting to emerge (Péron & Koons 2013). that can lead to average life spans of 10 To set the stage for papers that review the years, and lifetimes in excess of 30 years for roles of harvest (Cooch et al. 2014) as well as the longest-lived individuals. Although some the combination of habitat and harvest waterfowl are quite long-lived, “penguins (Osnas et al. 2014) when managing and albatrosses they are not” (sensu C.D. waterfowl populations, we begin by focusing Ankney), as typical life spans in those on the basic demographic mechanisms groups are 20 years and some individuals governing waterfowl population dynamics, live past 60 (U.S.G.S. Bird Banding and how they interact with environmental Laboratory longevity records). Given their conditions. We take a comparative approach slow pace of life, long-lived species can by considering the full suite of waterfowl afford to delay reproduction and invest in species that range from those as small as teal offspring “slowly” over their lifespan while to those as large as swans, and assess the balancing life history trade-offs. In contrast, state of knowledge of waterfowl population in short-lived species a “fast” start to ecology relative to relevant theories and reproduction early in life is favoured to help to what is known for other animal taxa. ensure that individuals pass on genes to the First, we compare the functional response next generation. This inter-specific pattern of population dynamics to life-cycle of life history strategies in birds and perturbations across waterfowl life history mammals has aptly become known as the strategies. Second, we discuss how “slow–fast continuum” (sensu Harvey & resistance to perturbation in one part of the Zammuto 1985; Sæther 1988; Gaillard et al. life-cycle might in turn have a mechanistic 1989; Promislow & Harvey 1990; for an connection to other stages in the life-cycle analogous life history theory in plants, see that are “more free” to vary over time and Grime 1977; Silvertown et al. 1992). thereby can potentially make important With greater longevity comes greater contributions to observed changes in complexity in the age structure of a abundance. Third, we discuss areas that population. In turn, immediate population require additional research to provide a growth rates and long-term abundances are

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 171 more sensitive to perturbations of the stable 2012), Common Eider Somateria mollissima age distribution in long-lived species with (Gilliland et al. 2009; Wilson et al. 2012), as slow life histories (e.g. geese, swans and well as Barnacle Goose Branta leucopsis eiders) than they are in short-lived species (Tombre et al. 1998), Emperor Goose Chen with fast life histories (e.g. teal; Koons et al. canagica (Schmutz et al. 1997) and Snow 2005, 2006a, 2007). These management- Goose Chen caerulescens (Cooch et al. 2001; relevant effects of perturbed age structure Aubry et al. 2010). We used studies like these on immediate and future abundance arise and other published demographic data to through demographic processes known as construct simple matrix population models “transient dynamics” and “population (following Oli & Zinner 2001) to compare momentum” (cf. Koons et al. 2006b). The the elasticities of λ to proportional changes importance of age structure in models for in annual fertility (i.e. the rate of recruiting guiding harvest management is emphasised females to 1 year of age per adult female) by Cooch et al. (2014) in this special issue and survival after the hatch year (AHY) (see also Hauser et al. 2006), but the general across waterfowl life histories. topics of age structure and its impact on From dabbling ducks to pochard, transient dynamics and population shelduck, sea ducks, geese and swans, the momentum across waterfowl life history elasticity of the population growth rate to strategies are ripe areas for future research. changes in AHY survival increases with To the contrary, the impact of changes in generation time (i.e. the average difference in vital rates (i.e. survival and reproductive age between parents and newborn offspring) success) on the long-term population whereas that to changes in fertility decreases growth rate is better studied for many with generation time (Fig. 1), a pattern waterfowl species. One popular metric for common to all birds (Sæther & Bakke 2000; measuring the relative impact of vital rate Stahl & Oli 2006). Given current data, only perturbations is the “elasticity”, which the Blue-winged Teal Anas discors has a higher measures the effect of equal proportionate elasticity for fertility than for AHY survival changes in vital rates on the focal metric (Fig. 1). This implies that, all else being equal, of population dynamics (most often λ, the population growth rate will respond the deterministic finite rate of growth; more readily to changes in AHY survival Caswell 2001). Elasticity analyses have been than it will to changes in fertility for the published for an array of waterfowl majority of waterfowl species that have species including Mallard Anas platyrhynchos moderate or long generation times. For (Hoekman et al. 2002, 2006), Northern species with very fast life histories like teal, Pintail Anas acuta (Flint et al. 1998), Lesser however, the reverse is true. Scaup Aythya affinis (Koons et al. 2006c), Greater Scaup A. marila (Flint et al. Demographic buffering and lability in 2006), Long-tailed Duck Clangula hyemalis waterfowl life-cycles (Schamber et al. 2009), King Eider Despite the greater sensitivity of λ to Somateria spectabilis (Bentzen & Powell proportional changes in survival for

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 172 Drivers of waterfowl population dynamics

1.0 0.9 0.8 0.7 0.6 0.5 0.4

E la s ticity 0.3 0.2 0.1 0.0 1 2 3 4 5 6 7 8 9 10 11 Generation time (years)

Figure 1. The elasticity of population growth rate (λ) to changes in annual fertility (circles) and adult survival (triangles) for 19 species of Nearctic and Palearctic ducks (white), shelduck and sea ducks (grey), and geese and swans (black), in relation to generation time (i.e. the average difference in age between parents and newborn offspring). On the far left are results for Blue-winged Teal and on the far right are results for Tundra Swan Cygnus c. columbianus. Elasticity results or demographic data used to conduct an elasticity analysis were extracted from: Patterson 1982; Rohwer 1985; Kennamer & Hepp 1987, 2000; Hepp et al. 1989; Hepp & Kennamer 1993; Flint et al. 1998; Cooch et al. 2001; Hoekman et al. 2002; Oli et al. 2002; Rohwer et al. 2002 and references therein; Flint et al. 2006; Grand et al. 2006 and references therein; Koons et al. 2006c; Gilliland et al. 2009. many species, changes in the underlying contributing to changes in abundance; components of fertility can still have Davison et al. 2010). In fact, the contribution management-relevant effects on λ. It is of each vital rate to past variation in often the case that management actions population growth rate can be calculated cannot elicit the desired response in a vital analytically by using Life Table Response rate (Baxter et al. 2006; Koons et al. 2014). Experiments (LTRE), which comprise When managers have difficulty influencing multiple approaches for best addressing the the vital rate(s) with the greatest elasticity, study hypothesis and design (Caswell 2001, they might look to those that are most 2010; see Cooch et al. 2001 for a waterfowl prone to being affected by changes in example). Over time, the contribution of a environmental conditions or management vital rate to past variation in population actions (i.e. the most labile). Actual changes growth rate is proportional to the product in population abundance are not only a of its squared elasticity and the square function of the vital rates with the greatest of its coefficient of process variation elasticities, but also density-dependent (when estimates are available, vital rate feedbacks (see section below) and the correlations should also be incorporated degree to which each vital rate changes over into calculations). Thus, a vital rate can make time (with fluctuations in age structure also important contributions to changes in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 173 population growth by having a large conditions across waterfowl life histories, elasticity, a large process variance, or both it is necessary to understand the basic (eqn. 7 in Heppell et al. 1998). theory of population growth in stochastic In waterfowl, vital rate contributions to environments. From first principles, the rate retrospective variation in λ have only been of growth for any population experiencing computed for Mallard (Hoekman et al. 2002; environmental stochasticity can be Amundson et al. 2012), King Eider (Bentzen approximated as: & Powell 2012), Common Eider (Wilson et  2 al. 2012), Barnacle Goose (Tombre et al. ln  s  ln   , (1) 2 et al 1998) and Snow Goose (Cooch . 2001). – A summary of these studies indicates that where lnλ denotes the mean and σ2 the although adult survival may have the variance of annual population growth rate greatest elasticity, nesting success, pre- (Nt/Nt–1) across environmental conditions fledging survival and juvenile survival are on the log scale (Lewontin & Cohen 1969; often more labile to environmental Tuljapurkar 1982). Of key importance, conditions. As a result, variation in these increased variance in annual population fertility components can sometimes growth rate decrements the long-term rate of λ contribute more to observed variation in growth in a stochastic environment (ln s). population dynamics than do changes in Recognising this relationship, Gillespie adult survival (see Gaillard et al. 1998 for (1977), and later Pfister (1998), noted that similar results in ungulates). Yet, a larger the vital rates with the greatest potential to λ contribution of fertility components to affect mean fitness (i.e. ln s) should exhibit observed changes in λ does not necessarily the least amount of temporal variance imply that changing these vital rates will because organisms should presumably be have a greater impact on population growth selected to avoid fluctuations in vital rates than would similar-sized changes in adult that would have severe negative impacts. survival. Rather, fertility components Over the long-term, natural selection should could fluctuate enough over time to make have favoured such life history properties; a important contributions to population concept that has become known as the dynamics (Caswell 2000). Managers might Demographic Buffering hypothesis (DB; thus use LTRE and other variance Gaillard et al. 2000; Boyce et al. 2006). There decomposition methods (e.g. life-stage is general support for the DB hypothesis in simulation analysis; Wisdom et al. 2000) plants and diverse animal species (e.g. Morris as a platform for identifying the vital et al. 2008; Dalgleish et al. 2010; Rotella et al. rates that can make important contributions 2012) including birds (Schmutz 2009). Avian to population growth through their species with high adult survival elasticities natural lability to changing environmental tend to exhibit less variation in adult survival conditions or management actions. over time than do species with low adult To gain a broader insight into which vital survival elasticities. In all comparative rates are most labile to environmental studies, however, empirical fits to DB

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 174 Drivers of waterfowl population dynamics predictions are not perfect; there is often a addition, rapid anthropogenic alterations to great deal of deviation from the predicted the environment might have exceeded the negative relationship between vital rate capacity of organisms to buffer or respond elasticities and temporal variation in vital to environmental fluctuations (Schmutz rates (Jäkäläniemi et al. 2013). Among 2009). At a fundamental level, the DB waterfowl, currently available data for concept is also restricted and cannot always temporal variation in adult survival do not capture the full effects of environmental λ support the DB hypothesis (Fig. 2). stochasticity on ln s. Optimising fitness in a Mixed support for the DB hypothesis stochastic world is a balancing act of could occur for a variety of reasons. For increasing mean vital rates, which affects the example, few studies are able to study the first term on the right-hand side of eqn. 1, entire life-cycle of a population over a long but also decreasing variance in vital rates, enough period of time to attain proper which minimises the negative effect of the estimates of temporal “process variation” in second term. It has only recently been each vital rate (see Rotella et al. 2012 for an recognised that temporal variation in vital λ example of the benefits provided by such rates can have a positive impact on ln s estimates). Unfortunately, ignoring the issue when– the variation induces an increase in of vital rate variance decomposition in tests lnλ that is sufficiently large to outweigh the of the DB hypothesis will inevitably inflate negative effect of σ2 (see eqn. 1; Drake type II errors (Morris & Doak 2002). In 2005; Koons et al. 2009). This can occur

0.45 0.40

0.35

max 0.30 0.25 0.20 0.15 CV/CV 0.10 0.05 0.00 0.60 0.65 0.70 0.75 0.80 0.85 0.90 Adult survival elasticity Figure 2. The relationship (P > 0.10, n.s.) between adult survival elasticities and the relative temporal “process variation” in adult survival across 13 waterfowl populations (white for ducks and black for geese; no sea duck data were available for these analyses). The coefficient of process variation (CV) was divided by the theoretically maximum possible value of CV for a given mean survival probability (see Morris & Doak 2004). Similar results were achieved using raw values of CV. The waterfowl data were extracted from Schmutz (2009) and references therein, as was the approach to developing matrix population models for the calculation of elasticities.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 175 ) t Y ( ) t ( Y Vital rate

EnvironmentalEnvironmental Conditions conditions x ( tx)( t)

Unfavourable Favourable Figure 3. A hypothetical sigmoid relationship between an environmental variable (x) and any vital rate Y. The lower box illustrates the mean (blue bar) and temporal distribution of environmental conditions x(t). The left box indicates the resulting temporal distribution of vital rate Y(t). The solid blue bars in the main figure illustrate vital rate performance in the average environment. The dashed blue bars show vital rate performance 1 s.d. below and above the average environment. As conditions vary over time, favourable environments produce larger increases in vital rate performance than decreases experienced in equally unfavourable environments. In turn, this raises the mean of Y and the long-term stochastic growth rate (red lines). The Demographic Buffering hypothesis implicitly assumes a linear relationship between x(t) and Y(t), in which case variation in x(t) has no effect on the mean of Y, only its statistical variance. when a vital rate has a convex relationship For vital rates that are on average low, with prevailing environmental conditions such as nest success (i.e. the proportion of (Fig. 3). Although the effect of “variance” in clutches where at least one egg hatches) in the strictest sense will still be negative, many duck populations, or offspring survival variation across convex relationships with in some goose and swan populations, the environmental variables can “skew” the potential for convex relationships with niche distribution of a vital rate and increase its– axes is strong. If such relationships occur, mean value (Fig. 3), thereby increasing lnλ enhanced environmental stochasticity could et al (the interested reader should see Rice– . actually be beneficial because favourable 2011 for a decomposition of ln λ that is conditions could induce booms in vital more complete than the commonly used rates that enhance the long-term stochastic approximation shown in eqn. 1). population growth rate (Koons et al. 2009;

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 176 Drivers of waterfowl population dynamics

Walker et al. 2013). Further research on DB compound the ongoing problem of over- and lability in waterfowl is nevertheless abundant goose populations (Ankney 1996; needed to provide a better understanding of Abraham et al. 2005). how vital rates will respond to global change In addition to predicted changes in across species, and how such responses will mean climatic conditions, increases in the affect population growth rates. variability of climate could also occur in many parts of the Nearctic and Palearctic Demographic drivers in an era of (e.g. Wetherald 2009), including the prairie global change pothole region of North America (Johnson For waterfowl management to keep pace et al. 2010). This could imply greater with climate, land-use and water-use changes extremity in climate from year to year; for in the 21st century, a greater depth of instance, very dry years followed by heavy knowledge is needed of the mechanisms precipitation that leads to flooding, or balmy affecting demography and population winters followed by bitter cold ones. Other dynamics. Habitat and predation continue to than the direct effects of exposure, climate be central topics of waterfowl research and tends to affect waterfowl populations management (e.g. Duebbert & Kantrud through changes in bottom-up food 1974) but we must also prepare for changes resources, top-down predation, and density- in the dynamic interplay between abiotic dependent interactions (Nudds 1992). conditions and trophic interactions. Some Trophic ecologists in Scandinavia have regions of the Nearctic and Palearctic are shown that the shape of a predator-prey becoming warmer while others are not; some functional response can determine may receive more precipitation and others fundamentally whether increased temporal less (IPCC 2013). Many landscapes are being variation in a resource has a positive or converted to produce more ethanol, wind or negative effect on the mean vital rates of a solar power (Northrup & Wittemyer 2013; consumer (e.g. Henden et al. 2008). In other Wright & Wimberly 2013), and greater words, trophic interactions can dictate the human demands for land and water will relative advantages of DB versus lability in a continue to affect the amounts and quality of population. In other systems, climate-driven habitat available to wildlife (Fischer & Heilig pulses in primary productivity can result in 1997; Lemly et al. 2000; Pringle 2000). The counter-acting direct and time-lagged effects aforementioned changes describe shifts in on primary consumers because of complex the “mean” environmental conditions that trophic interactions such as apparent often come to mind when we speak of competition (Schmidt & Ostfeld 2008). For global change in the 21st century. Some example, in the prairie pothole region, outcomes of these changes are predictable; Walker et al. (2013) found that duck nest less water or breeding habitat would likely success in year t was positively associated lead to fewer ducks (Reynolds et al. 2001; with pond density and primary productivity Stephens et al. 2005), whereas more corn in the same year (t), but negatively related to planted for ethanol production could these variables in previous years (t–1 and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 177 t–2). Findings like these could possibly be photoperiod, plant phenology and resource attributed to the positive numerical response availability along migration paths, as well as of alternate prey to climate-induced change plant phenology and snow cover on the in primary productivity, which could swamp breeding grounds (Strong & Trost 1994; out generalist predators in the current year, Gwinner 1996; Bauer et al. 2008, Tombre et but then lead to numerical responses in al. 2008). Thus, they cannot always “match” predator populations that later have a their nesting phenology with the greening of deleterious effect on waterfowl reproductive graminoid forage on the breeding grounds success (Ackerman 2002; Brook et al. 2008; (Gauthier et al. 2013). A year of late nesting Schmidt & Ostfeld 2008; Iles et al. 2013a; relative to the phenology of graminoid Walker et al. 2013). greening can result in reduced food quality Changing mean and variance of climatic for goslings that in turn inhibits their conditions can also affect waterfowl growth, body condition and ultimately populations through differential changes in survival (Dickey et al. 2008; Aubry et al. phenology across trophic levels. For 2013). However, the trend toward earlier example, the match-mismatch hypothesis greening is not consistent from year to year predicts that a consumer should try to because the variability in growing degree “match” its life-cycle events with the timing days in spring seems to be getting larger. of maximal resource availability or quality Although the trend toward earlier greening because failure to do so results in a may force a mismatch between arctic geese “mismatch” between resource and consumer and their preferred forage, stochasticity in phenologies that reduces fitness (Visser & the near-term will still provide some years Both 2005). Research on the match- where they can match their nesting to plant mismatch hypothesis in ducks is advancing phenology (Aubry et al. 2013). (e.g. Drever & Clark 2007; Sjöberg et al. Climate-driven changes in phenology can 2011), and species with less flexible nesting also affect the intensity of top-down phenology (e.g. scaup and scoters; Gurney et predation and even result in completely al. 2011) might suffer more from climate- novel predator-prey interactions. For induced shifts in the timing of resource example, climate-driven declines in the availability compared to those that are more extent and duration of sea ice each year have flexible (e.g. Mallard; Drever et al. 2012). been linked to earlier onshore arrival of Research on the match-mismatch Polar Bears Ursus maritimus in many parts of hypothesis in geese indicates that warming the Arctic and reduced opportunities for the of the Arctic has led to, on average, earlier Polar Bears to hunt seals (Family: Phocidae greening of graminoid plants that can and Otariidae; Stirling & Derocher 1993; readily take advantage of early growing Stirling & Parkinson 2006). This has resulted degree days (van der Jeugd et al. 2009; in a mismatch for Polar Bears with their Doiron et al. 2013; Gauthier et al. 2013). preferred seal prey. Ironically, this now Some arctic geese have, however, evolved to exposes Polar Bears to a novel overlap with balance their timing of migration with the breeding seasons of many ground-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 178 Drivers of waterfowl population dynamics nesting waterfowl populations (Rockwell & bound on the capacity of a population to Gormezano 2009), from which they are grow (Turchin 2003). Ever since Malthus readily consuming eggs and offspring (1798), density dependence has therefore (Madsen et al. 1998; Noel et al. 2005; Drent been central to our way of thinking & Prop 2008; Smith et al. 2010; Iles et al. about population dynamics. Given its role 2013b; Gormezano & Rockwell 2013; in affecting population growth, harvest Iverson et al. 2014). Novel predator-prey management has often been used to interactions such as these have strong manipulate population density in an potential to affect waterfowl population attempt to optimise long-term yield from dynamics (Rockwell et al. 2011). Climate populations (e.g. maximum sustainable yield change could thus affect waterfowl theory; Walters 1986). On the other hand, populations through changes in both habitat management attempts to manipulate bottom-up and top-down interactions. Yet the carrying capacity by providing more per the strength of these interactions, as well capita resources, and thereby increasing the as those among conspecifics, may be ceiling on abundance where reproductive moderated by population density and the success and mortality balance each other out ability of individual species to respond to (Smith et al. 1989). Only recently has the novel selection pressures induced by waterfowl management begun a formal climate change. integration of these concepts, which are tied both to density dependence and to Density dependence: the ever-elusive environmental change (Runge et al. 2006; regulator of populations Mattsson et al. 2012), a topic specifically Population density affects population addressed by Osnas et al. (2014) in this dynamics through both intra- and inter- volume. For these reasons, and others, it is specific interactions. For example, density critical to improve our understanding of can influence competition for territories how density dependence operates over the and mates, competition for and depletion annual life-cycle, and across the diverse array of limited foods, rates of pathogen of waterfowl life histories (Gunnarsson et al. transmission, and functional responses of 2013). predators. Density can also act in positive Measuring the influence of density ways. In arctic geese, for example, colonial dependence on population dynamics in the nesting density and intermediate levels of wild, however, has been said to be “like a grazing have both been shown to enhance search for the holy grail” (Krebs 1995). Two vital rates (e.g. Raveling 1989; Hik & Jefferies demographic approaches have nevertheless 1990; Aubry et al. 2013). Both positive and been employed to make progress toward negative density-dependent interactions understanding density dependence: surveys scale up to affect vital rates in ways that of population abundance, and studies of life shape the pace of population growth and, history traits (Lebreton & Gimenez 2013). ultimately, negative density dependence at Time series analyses of how surveyed some point in the life-cycle places an upper abundance responds to levels of population

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 179 abundance in previous years are sensu stricto between explanatory and response variables. phenomenological (Krebs 2002) but can, That said, uncertainty in estimates of with care, be used to test for the presence of population abundance (the explanatory density dependence, measure its impact variable) can result in bias toward false relative to environmental variables, and even absence of density dependence (McArdle examine the interactive effects of variation 2003); a conservative outcome that is often in population density and environmental more favoured in science than a false variables on population dynamics (e.g. positive (Lebreton & Gimenez 2013). When Stenseth et al. 2003; Rotella et al. 2009). conducted with explicit attention to biotic The long-standing problem, however, has interactions, the life history trait approach been the lack of independence between can provide a deeper understanding of explanatory and response variables and the density dependence than can analyses of related issue of shared sampling variation abundance time series (Krebs 2002). Such (i.e. uncertainty in abundance) between the approaches have been used nicely in both axes being analysed, both of which bias observational and experimental waterfowl estimation towards greater strength (and studies to elucidate the mechanistic effects presence) of density dependence than of population density on fidelity and adult actually exists (Freckleton et al. 2006; breeding probability (Sedinger et al. 2008), Lebreton & Gimenez 2013). Unfortunately, clutch size (Cooch et al. 1989; Sedinger et al. many older published studies did not 1998), nesting success (Raveling 1989 for account for analytical problems that induce geese; Gunnarsson & Elmberg 2008; these false positives. The results of those Ringleman et al. 2012 for ducks; see studies should probably be disregarded and, Gunnarsson et al. 2013 for a review), where available data allow, the information offspring growth (Lindholm et al. 1994; should be re-analysed with modern methods Schmutz & Laing 2002; Person et al. 2003), to improve our understanding of density offspring survival (Williams et al. 1993; dependence. Modern state-space statistical Nicolai & Sedinger 2012 for geese; models for time-series data can account for Gunnarsson et al. 2006; Amundson et al. these issues (e.g. Knape & de Valpine 2012; 2011 for ducks; see Gunnarsson et al. 2013 Delean et al. 2013) and help make use of for review), subsequent post-fledging widely available monitoring data to gain survival (e.g. Schmutz 1993; Sedinger & insight into the influence of density Chelgren 2007; Aubry et al. 2013) and even dependence on waterfowl population the effects of nutrient limitation during dynamics over time (Murray et al. 2010), development on eventual adult body size space (Viljugrein et al. 2005), and across life (e.g. Cooch et al. 1991a,b; Sedinger et al. 1995; histories (Jamieson & Brooks 2004; Sæther Loonen et al. 1997). Moreover, density et al. 2008; Murray et al. 2010). dependence at one stage of the life-cycle The approach of studying the effects of (e.g. nesting) can affect population density population density on life history traits and its impact later in the life-cycle (e.g. nicely avoids the issue of dependence offspring rearing and then post-fledging)

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 180 Drivers of waterfowl population dynamics through a sequential cohort process Most sensitive Pre-fledging survival (Elmberg et al. 2005). Although its effects are complex, great Nesting success progress has been made in recent years toward understanding the role of density Post-fledging survival dependence in waterfowl (Gunnarsson et al. Natal fidelity 2013). Lack of long-term and experimental studies for many taxa and the Clutch size aforementioned issues with estimation currently prevent us from being able to Age at first reproduction make robust conclusions about patterns in Adult breeding site fidelity density dependence across waterfowl life histories. When resources are limited, Breeding probability density dependence can have strong effects Sub-adult survival on reproductive success of the most fecund (e.g. Mallard; Kaminski & Gluesing 1987) Old-age survival and even the most long-lived of waterfowl species (e.g. arctic geese and swans; Williams Least sensitive Prime-age survival et al. 1993; Nummi & Saari 2003). Rarely, however, has population density been Figure 4. A hypothesised sequence of vital rate shown to have an effect on adult survival sensitivity to density dependence over the age- (Dugger et al. 1994; Ludwichowski et al. structured life-cycle of a hypothetical waterfowl 2002; Menu et al. 2002; Sedinger et al. 2007), species. which we might expect because adult survival has high elasticity values and For a given waterfowl taxonomic group, the should thus be highly buffered against rank-order of vital rate sensitivity to density environmental changes induced by dependence can be organised as a list (see population density (see sections above). Fig. 4), but the organisation and presence of Given existing evidence, and following density dependence across the life-cycle the Eberhardt hypothesis that ungulate could shift with life history strategy (e.g. r–K ecologists have used to focus their study of selection; MacArthur & Wilson 1967; density dependence (Eberhardt 1977, 2002; Pianka 1970) or perhaps more so with Bonenfant et al. 2009), we hypothesise that “lifestyle” (e.g. diet, nest-site preference, waterfowl become more robust to the mating strategy, etc.; Dobson 2007; Sibly & effects of density dependence as they Brown 2007) and prevailing environmental develop into the prime ages of adulthood, conditions. At this point, these ideas are but might again become susceptible at nascent and are presented here for future older ages; for instance, because of waterfowl ecologists to advance or repudiate immunosenescence and density-related as science progresses. Knowing where pathogen transmission (Palacios et al. 2011). density dependence operates in the life-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 181 cycle, and how it varies over time and space energetic models, explicit studies of density will eventually help managers apply actions dependence during the staging and in the most appropriate seasons, habitats wintering periods of the life-cycle are scarce. and environmental conditions for achieving The cross-seasonal approach to studying desired population responses relative to life-cycle dynamics may be the best way to resource investments (du Toit 2010). fulfil these informational needs (see Sedinger & Alisauskas 2014). Conclusions In addition to adaptive management Our understanding of waterfowl population approaches, formal experiments are needed dynamics has come a long way since the to separate density-dependent from density- middle part of the 20th century. A strong independent processes across gradients focus on studying vital rates has helped of resource availability. Long-term develop and improve population models for observational studies additionally offer the several species, and new tools have allowed test of time, and are perhaps the best way researchers to identify differences in vital to understand the complex effects of rate contributions to population dynamics environmental change and stochasticity on across life histories. Progress has been made populations (Clutton-Brock & Sheldon in research on bottom-up and top-down 2010). Where possible, experimental and mechanisms affecting populations, and light observational approaches should be has even been shed on the once elusive combined to enhance learning within the mechanism of density dependence. adaptive management framework. There are nevertheless key gaps that need The relatively new Integrated Population to be filled in order to sustain healthy Model (IPM, not to be confused with waterfowl populations amidst the challenges “integral” population models) offers an presented by global change. For example, innovative way to combine data from significant portions of Palearctic waterfowl different research approaches to test populations occupy regions where research hypotheses and, importantly, to link and monitoring are scarce (e.g. Russia and information from detailed field studies with the boreal forest of North America), which large-scale monitoring data to guide makes it difficult to develop scientifically adaptive management (Besbeas et al. 2002). robust studies and management of In addition, by using the constraint that only populations that do not necessarily birth, immigration, death and emigration recognise geographic borders and survey can affect population abundance, IPMs can boundaries. We outlined briefly our current combine abundance data with vital rate data understanding of density dependence in in a way that can reduce bias and improve waterfowl, and presented a framework for precision in demographic estimates (Abadi organising an understanding of the key life- et al. 2010). Utilising these features, IPMs are cycle stages where density dependence has a already being used to provide a synthetic significant impact on population dynamics view of the mechanisms that shape (Fig. 4). Although often assumed in waterfowl population dynamics (Péron et al.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 182 Drivers of waterfowl population dynamics

2012), and are even being used to study the and environmental factors on Mallard demographic effects of species interactions duckling survival in North Dakota. Journal of at large scales (Péron & Koons 2012). IPMs Wildlife Management 75: 1330–1339. thus offer a way to model the dynamic Amundson, C.L., Pieron, M.R., Arnold, T.W. mechanisms that affect waterfowl & Beaudoin, L.A. 2013. The effects of predator removal on Mallard production and populations and communities at scales that population change in northeastern North are relevant to managing migratory species. Dakota. Journal of Wildlife Management 77: Waterfowl are among the most extensively 143–152. studied vertebrates on the planet, and we Ankney, C.D. 1996. An embarrassment of riches: predict that this rich tradition will contribute too many geese. Journal of Wildlife Management to great advancements in population 60: 217–223. ecology, evolution and applied natural Aubry, L.M., Rockwell, R.F. & Koons, D.N. resource management in the 21st century. 2010. Metapopulation dynamics of mid- continent Lesser Snow Geese: implications Acknowledgements for management. Human-Wildlife Interactions We thank the ECNAW scientific committee 4: 11–32. for inviting us to give our plenary talk at the Aubry, L.M., Rockwell, R.F., Cooch, E.G., 2013 symposium in Memphis, Tennessee, as Brook, R.W., Mulder, C.P. & Koons, D.N. 2013. Climate change, phenology, and well as R.G. Clark and J.S. Sedinger for habitat degradation: drivers of gosling body helpful comments on an earlier version of condition and juvenile survival in Lesser Snow the manuscript. DNK thanks R.F. Rockwell Geese. Global Change Biology 19: 149–160. for discussions about waterfowl population Bauer, S., Van Dinther, M., Høgda, K.-A., ecology over the years and NSF (DEB Klaassen, M. & Madsen, J. 2008. The 1019613) for financial support. consequences of climate-driven stop-over sites changes on migration schedules and References fitness of Arctic geese. Journal of Animal Abadi, F., Gimenez, O., Arlettaz, R. & Schaub, M. Ecology 77: 654–660. 2010. An assessment of integrated population Baxter, P.W.J., McCarthy, M.A., Possingham, models: bias, accuracy, and violation of the H.P., Menkhorst, P.W. & McLean, N. assumption of independence. Ecology 91: 2006. Accounting for management costs in 7–14. sensitivity analyses of matrix population Abraham, K.F., Jefferies, R.L. & Alisauskas, R.T. models. Conservation Biology 20: 893–905. 2005. The dynamics of landscape change and Bentzen, R.L. & Powell, A.N. 2012. Population Snow Geese in mid-continent North dynamics of King Eiders breeding in America. Global Change Biology 11: 841–855. northern Alaska. Journal of Wildlife Management Ackerman, J.T. 2002. Of mice and Mallards: 76: 1011–1020. positive indirect effects of coexisting prey Besbeas, P., Freeman, S.N., Morgan, B.J.T. & on waterfowl nest success. Oikos 99: Catchpole, E.A. 2002. Integrating mark- 469–480. recapture-recovery and census data to Amundson, C.L., & Arnold, T.W. 2011. The role estimate animal abundance and demographic of predator removal, density-dependence, parameters. Biometrics 58: 540–547.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 183

Bonenfant, C., Gaillard, J.-M., Coulson, T., Festa- Cooch, E.G., Lank, D.B., Rockwell, R.F., Dzubin, Bianchet, M., Loison, A., Garel, M., Loe, A. & Cooke, F. 1991a. Body size variation in L.E., Blanchard, P., Pettorelli, N., Owen- Lesser Snow Geese: environmental plasticity Smith, N., du Toit, J. & Duncan, P. 2009. in gosling growth rates. Ecology 72: 503–512. Empirical evidence of density-dependence Cooch, E.G., Lank, D.B., Rockwell, R.F. & in populations of large herbivores. Advances Cooke, F. 1991b. Long-term decline in body in Ecological Research 41: 314–352. size in a Snow Goose population: evidence Boyce, M.S., Haridas, C.V., Lee, C.T., & the of environmental degradation? Journal of NCEAS Stochastic Demography Working Animal Ecology 60: 483–496. Group. 2006. Demography in an increasingly Cooch, E.G., Rockwell, R.F. & Brault, S. 2001. variable world. Trends in Ecology & Evolution Retrospective analysis of demographic 21: 141–148. responses to environmental change: an Brook, R.W., Pasitschniak-Arts, M., Howerter, example in the Lesser Snow Goose. Ecological D.W. & Messier, F. 2008. Influence of rodent Monographs 71: 377–400. abundance on nesting success of prairie Cooch, E.G., Guillemain, M., Boomer, G.S., waterfowl. Canadian Journal of Zoology 86: Lebreton, J.-D. & Nichols, J.D. 2014. The 497–506. effects of harvest on waterfowl populations. Canadian Wildlife Service & U.S. Fish and Wildfowl (Special Issue No. 4): 220–276. Wildlife Service. 2012. North American Dalgleish, H.J., Koons, D.N. & Adler, P.B. 2010. waterfowl management plan. Canadian Can life-history traits predict the response of Wildlife Service and U.S. Fish and Wildlife forb populations to changes in climate Service Report, U.S. Fish and Wildlife variability? Journal of Ecology 98: 209–217. Service, Washington D.C., USA. Davison, R., Jacquemyn, H., Adriaens, D., Caswell, H. 2000. Prospective and retrospective Honnay, O., de Kroon, H. & Tuljapurkar, perturbation analyses: their roles in S. 2010. Demographic effects of extreme conservation biology. Ecology 81: 619–627. weather events on a short-lived calcareous Caswell, H. 2001. Matrix Population Models: grassland species: stochastic life table response Construction, Analysis and Interpretation, 2nd experiments. Journal of Ecology 98: 255–267. edition. Sinauer Associates Inc., Sunderland, Dawkins, R. & Krebs, J.R. 1979. Arms races Massachusetts, USA. between and within species. Proceedings of the Caswell, H. 2010. Life table response experiment Royal Society of London, B 205: 489–511. analysis of the stochastic growth rate. Journal Delean, S., Brook, B.W. & Bradshaw, C.J.A. 2013. of Ecology 98: 324–333. Ecologically realistic estimates of maximum Clutton-Brock, T. & Sheldon, B.C. 2010. population growth using informed Bayesian Individuals and populations: the role of priors. Methods in Ecology and Evolution 4: long-term, individual-based studies of 34–44. animals in ecology and evolutionary biology. Dickey, M.–H., Gauthier, G. & Cadieux, M. 2008. Trends in Ecology and Evolution 25: 562–573. Climatic effects on the breeding phenology Cooch, E.G., Lank, D.B., Rockwell, R.F. & and reproductive success of an Arctic-nesting Cooke, F. 1989. Long-term decline in goose species. Global Change Biology 14: 1973– fecundity in a Snow Goose population: 1985. evidence for density dependence. Journal of Doiron, M., Legagneux, P., Gauthier, G. & Animal Ecology 58: 711–726. Lévesque, E. 2013. Broad-scale satellite

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 184 Drivers of waterfowl population dynamics

normalized difference vegetation index Elmberg, J., Gunnarsson, G., Pöysä, H., Sjöberg, data predict plant biomass and peak date K. & Nummi, P. 2005. Within-season of nitrogen concentration in Arctic tundra sequential density dependence regulates vegetation. Applied Vegetation Science 16: breeding success in Mallards Anas platyrhynchos. 343–351. Oikos 108: 582–590. Dobson, F.S. 2007. A lifestyle view of life-history Fischer, G. & Heilig, G.K. 1997. Population evolution. Proceedings of the National Academy of momentum and the demand on land and Sciences, USA 104: 17565–17566. water resources. Philosophical Transactions of the Drake, J.M. 2005. Population effects of increased Royal Society, B 352: 869–889. climate variation. Proceedings of the Royal Society Flint, P.L., Grand, J.B. & Rockwell, R.F. 1998. A of London, B 272: 1823–1827. model of Northern Pintail productivity and Drent, R., & Prop, J. 2008. Barnacle Goose Branta population growth rate. Journal of Wildlife leucopsis survey on Nordenskiöldkysten, west Management 62: 1110–1118. Spitsbergen 1975–2007: breeding in relation Flint, P.L., Grand, J.B., Fondell, T.F. & Morse, J.A. to carrying capacity and predator impact. 2006. Population dynamics of Greater Scaup Circumpolar Studies 4: 59–83. breeding on the Yukon-Kuskokwim Delta, Drever, M.C. & Clark, R.G. 2007. Spring Alaska. Wildlife Monographs 162: 1–22. temperature, clutch initiation date and duck Freckleton, R.P., Watkinson, A.R., Green, R.E. & nest success: a test of the mismatch hypothesis. Sutherland, W.J. 2006. Census error and the Journal of Animal Ecology 76: 139–148. detection of density dependence. Journal of Drever, M.C., Clark, R.G., Derksen, C., Slattery, Animal Ecology 75: 837–851. S.M., Toose, P. & Nudds, T.D. 2012. Gaillard, J.-M., Pontier, D., Allainé, D., Lebreton, Population vulnerability to climate change J.-D. & Trouvilliez, J. 1989. An analysis of linked to timing of breeding in boreal ducks. demographic tactics in birds and mammals. Global Change Biology 18: 480–492. Oikos 56: 59–76. Duebbert, H.F. & Kantrud, H.A. 1974. Upland Gaillard, J.-M., Festa-Bianchet, M. & Yoccoz, N.G. duck nesting related to land use and predator 1998. Population dynamics of large reduction. Journal of Wildlife Management 38: herbivores: variable recruitment with constant 257–265. adult survival. Trends in Ecology & Evolution 13: Dugger, B.D., Reinecke, K.J. & Fredrickson, L.H. 58–63. 1994. Late winter survival of female Mallards Gaillard, J.-M., Festa-Bianchet, M., Yoccoz, N.G., in Arkansas. Journal of Wildlife Management 58: Loison, A. & Toigo, C. 2000. Temporal 94–99. variation in fitness components and du Toit, J.T. 2010. Considerations of scale in population dynamics of large herbivores. biodiversity conservation. Animal Conservation Annual Review of Ecology and Systematics 31: 13: 229–236. 367–393. Eberhardt, L.L. 1977. Optimal policies for Gauthier, G., Bêty, J., Cadieux, M.-C., Legagneux, conservation of large mammals, with P., Doiron, M., Chevallier, C., Lai, S., special references to marine ecosystems. Tarroux, A. & Berteaux, D. 2013. Long-term Environmental Conservation 4: 205–212. monitoring at multiple trophic levels Eberhardt, L.L. 2002. A paradigm for population suggests heterogeneity in responses to analysis of long-lived vertebrates. Ecology 83: climate change in the Canadian Arctic 2841–2854. tundra. Philosophical Transactions of the Royal

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 185

Society, B 368: 20120482. http://dx.doi.org/ R.W., Afton, A.D., Cutting, K., Warren, J.M., 10.1098/rstb.2012.0482. Fournier, M. & Koons, D.N. 2011. Time Gillespie, J.H. 1977. Natural selection for constraints in temperate-breeding species: variance in offspring numbers: a new influence of growing season length on evolutionary principle. American Naturalist reproductive strategies. Ecography 34: 628– 111: 1010–1014. 636. Gilliland, S.G., Gilchrist, H.G., Rockwell, R.F., Gwinner, E. 1996. Circannual clocks in avian Robertson, G.J., Savard, J.-P.L., Merkel, reproduction and migration. Ibis 138: 47–63. F. & Mosbech, A. 2009. Evaluating the Harvey, P.H. & Zammuto, R.M. 1985. Patterns of sustainability of harvest among northern mortality and age at first reproduction in Common Eiders Somateria mollissima borealis in natural populations of mammals. Nature 315: Greenland and Canada. Wildlife Biology 15: 319–320. 24–36. Hauser, C.E., Cooch, E.G. & Lebreton, J.-D. Gormezano, L.J. & Rockwell, R.F. 2013. What to 2006. Control of structured populations by eat now? Shifts in Polar Bear diet during the harvest. Ecological Modelling 196: 462–470. ice-free season in western Hudson Bay. Henden, J.-A., Bårdsen, B.-J., Yoccoz, N.G. & Ecology and Evolution 3: 3509–3523. Imset, R.A. 2008. Impacts of differential Grand, J.B., Koons, D.N., Arnold, J.M. & prey dynamics on the potential recovery of Derksen, D.V. 2006. Modeling the Recovery of endangered Arctic Fox populations. Journal of Avian Populations. USGS Report to the Alaska Applied Ecology 45: 1086–1093. Science Center, Anchorage, USA. Hepp, G.R., Kennamer, R.A. & Harvey, W.F. Grime, J.P. 1977. Evidence for the existence of 1989. Recruitment and natal philopatry of three primary strategies in plants and its Wood Ducks. Ecology 70: 897–903. relevance to ecological and evolutionary Hepp, G.R. & Kennamer, R.A. 1993. Effects theory. American Naturalist 111: 1169–1194. of age and experience on reproductive Gunnarsson, G., Elmberg, J., Sjöberg, K., Pöysä, performance of Wood Ducks. Ecology 74: H. & Nummi, P. 2006. Experimental 2027–2036. evidence for density-dependent survival in Heppell, S.S. 1998. Application of life-history Mallard (Anas platyrhynchos) ducklings. theory and population model analysis to Oecologia 149: 203–213. turtle conservation. Copeia 1998: 367–375. Gunnarsson, G. & Elmberg, J. 2008. Density- Hik, D.S. & Jefferies, R.L. 1990. Increases in the dependent nest predation – an experiment net above-ground primary production of a with simulated Mallard nests in contrasting saltmarsh forage grass: a test of the landscapes. Ibis 150: 259–269. predictions of the herbivore-optimization Gunnarsson, G., Elmberg, J., Pöysä, H. Nummi, model. Journal of Ecology 78: 180–195. P., Sjöberg, K., Dessborn, L. & Arzel, C. Hoekman, S.T., Mills, L.S., Howerter, D.W., 2013. Density dependence in ducks: a review Devries, J.H. & Ball, I.J. 2002. Sensitivity of the evidence. European Journal of Wildlife analyses of the life cycle of midcontinent Research 59: 305–321. Mallards. Journal of Wildlife Management 66: Gurney, K.E.B., Clark, R.G., Slattery, S.M., 883–900. Smith-Downey, N.V., Walker, J., Armstrong, Hoekman, S.T., Gabor, T.S., Petrie, M.J., Maher, L.M., Stephens, S.E., Petrula, M., Corcoran, R., Murkin, H.R. & Lindberg, M.S. 2006. R.M., Martin, K.H., DeGroot, K.A., Brook, Population dynamics of Mallards breeding in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 186 Drivers of waterfowl population dynamics

agricultural environments in eastern Canada. Kaminski, R.M. & Gluesing, E.A. 1987. Density- Journal of Wildlife Management 70: 121–128. and habitat-related recruitment in Mallards. Iles, D.T., Rockwell, R.F., Matulonis, P., Journal of Wildlife Management 51: 141–148. Robertson, G.J., Abraham, K.F., Davies, C. & Kennamer, R.A. & Hepp, G.R. 1987. Frequency Koons, D.N. 2013a. Predators, alternative and timing of second broods in Wood prey and climate influence annual breeding Ducks. Wilson Bulletin 99: 655–662. success of a long-lived sea duck. Journal of Kennamer, R.A. & Hepp, G.R. 2000. Integration Animal Ecology 82: 683–693. of long-term research with monitoring goals: Iles, D.T. Peterson, S.L., Gormezano, L.J., Koons, breeding Wood Ducks on the Savannah D.N. & Rockwell, R.F. 2013b. Terrestrial River Site. Studies in Avian Biology 21: 39–49. predation by Polar Bears: not just a wild Knape, J. & De Valpine, P. 2012. Are patterns of goose chase. Polar Biology 36: 1373–1379. density dependence in the global population Intergovernmental Panel on Climate Change dynamics database driven by observation (IPCC). 2013. Climate change 2013: the errors? Ecology Letters 15: 17–23. physical science basis. In T.F. Stocker, D. Koons, D.N., Grand, J.B., Zinner, B. & Rockwell, Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. R.F. 2005. Transient population dynamics: Boschung, A. Nauels, Y. Xia, V. Bex & P.M. relations to life history and initial population Midgley (eds.), Contribution of Working Group state. Ecological Modelling 185: 283–297. I to the Fifth Assessment Report of the Koons, D.N., Grand, J.B. & Arnold, J.M. 2006a. Intergovernmental Panel on Climate Change, 1535 Population momentum across vertebrate pp. Cambridge University Press, Cambridge, life histories. Ecological Modelling 197: 418– UK. 430. Iverson, S.A., Gilchrist, H.G., Smith, P.A., Gaston, Koons, D.N., Rockwell, R.F. & Grand, J.B. 2006b. A.J. & Forbes, M.R. 2014. Longer ice-free Population momentum: implications for seasons increase the risk of nest depredation wildlife management. Journal of Wildlife by Polar Bears for colonial breeding birds in Management 70: 19–26. the Canadian Arctic. Proceedings of the Royal Koons, D.N., Rotella, J.J., Willey, D.W., Taper, Society, B 281: 20133128. http://dx.doi.org/ M.L., Clark R.G., Slattery, S., Brook, R.W., 10.1098/rstb.2013.3128. Corcoran, R.M. & Lovvorn, J.R. 2006c. Jamieson, L.E. & Brooks, S.P. 2004. Density Lesser Scaup population dynamics: what can dependence in North American ducks. be learned from available data? Avian Animal Biodiversity and Conservation 27: 113– Conservation and Ecology 1(3): 6. [online] URL: 128. http://www. ace-eco.org/vol1/iss3/art6/. Jäkäläniemi, A., Ramula, S. & Tuomi, J. 2013. Koons, D.N., Holmes, R.R. & Grand, J.B. 2007. Variability of important vital rates challenges Population inertia and its sensitivity to the demographic buffering hypothesis. changes in vital rates and population Evolutionary Ecology 27: 533–545. structure. Ecology 88: 2857–2867. Johnson, W.C., Werner, B., Guntenspergen, G.R., Koons, D.N., Pavard, S., Baudisch, A. & Metcalf, Voldseth, R.A., Millett, B., Naugle, D.E., C.J.E. 2009. Is life-history buffering or Tulbure, M., Carroll, R.W.H., Tracy, J. & lability adaptive in stochastic environments? Olawsky, C. 2010. Prairie wetland complexes Oikos 118: 972–980. as landscape: functional units in a changing Koons, D.N., Rockwell, R.F. & Aubry, L.M. 2014. climate. BioScience 60: 128–140. Effects of exploitation on an overabundant

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 187

species: the Lesser Snow Goose predicament. Malthus. T.R. 1798. An essay on the principle of Journal of Animal Ecology 83: 365–374. population. In G. Hardin (ed.), Population, Krebs, C.J. 1995. Two paradigms of population Evolution, and Birth Control: A Collage of regulation. Wildlife Research 22: 1–10. Controversial Ideas (1964), pp. 4–16. Freeman, Krebs, C.J. 2002. Two complementary paradigms San Francisco, California, USA. for analysing population dynamics. Manly, B.F., McDonald, L., Thomas, D.L., Philosophical Transactions of the Royal Society, B McDonald, T.L. & Erickson, W.P. 2002. 357: 1211–1219. Resource Selection by Animals: Statistical Design Lebreton, J.-D. & Gimenez, O. 2013. Detecting and Analysis for Field Studies, 2nd edition. and estimating density dependence in wildlife Kluwer, Boston, Massachusetts, USA. populations. Journal of Wildlife Management 77: Mattsson, B.J., Runge, M.C., Devries, J.H., 12–23. Boomer, G.S., Eadie, J.M., Haukos, D.A., Lemly, A.D., Kingsford, R.T. & Thompson, J.R. Fleskes, J.P., Koons, D.N., Thogmartin, W.E. 2000. Irrigated agriculture and wildlife & Clark R.G. 2012. A modeling framework for conservation: conflict on a global scale. integrated harvest and habitat management of Environmental Management 25: 485–512. North American waterfowl: Case-study of Lewontin, R.C. & Cohen, D. 1969. On Northern Pintail metapopulation dynamics. population growth in a randomly varying Ecological Modelling 225: 146–158. environment. Proceedings of the National McArdle, B.H. 2003. Lines, models, and errors: Academy of Sciences, USA 62: 1056–1060. regression in the field. Limnology and Lindholm, A., Gauthier, G. & Desrochers, A. Oceanography 48: 1363–1366. 1994. Effects of hatch date and food-supply Menu, S., Gauthier, G. & Reed, A. 2002. Changes on gosling growth in Arctic nesting Greater in survival rates and population dynamics of Snow Geese. Condor 96: 898–908. Greater Snow Geese over a 30-year period: Loonen, M.J.J.E., Oosterbeek, K. & Drent, R.H. implications for hunting regulations. Journal of 1997. Variation in growth of young and adult Applied Ecology 39: 91–102. size in Barnacle Geese Branta leucopsis: Morris, W.F. & Doak, D.F. 2002. Quantitative evidence for density dependence. Ardea 85: Conservation Biology: Theory and Practice 177–192. of Population Viability Analysis. Sinauer, Ludwichowski, I., Barker, R. & Bräger, S. 2002. Sunderland, Massachusetts, USA. Nesting area fidelity and survival of female Morris, W.F. & Doak, D.F. 2004. Buffering of life Common Goldeneyes Bucephala clangula: are histories against environmental stochasticity: they density-dependent? Ibis 144: 452–460. accounting for a spurious correlation Madsen, J., Bregnballe, T., Frikke, J. & Kristensen, between the variabilities of vital rates and I.B. 1998. Correlates of predator abundance their contributions to fitness. American with snow and ice conditions and their role in Naturalist 163: 579–590. determining timing of nesting and breeding Morris, W.F., Pfister, C.A., Tuljapurkar, S., success in Svalbard Light-bellied Brent geese Haridas, C.V., Boggs, C.L., Boyce, M.S., Branta bernicla hrota. Norsk Polarinstitutt Skrifter Bruna, E.M., Church, D.R., Coulson, T., 200: 221–234. Doak, D.F., Forsyth, S., Gaillard, J.-M., MacArthur, R.H. & Wilson, E.O. 1967. The Theory Horvitz, C.C., Kalisz, S., Kendall, B.E., of Island Biogeography. Princeton University Knight, T.M., Lee, C.T. & Menges, E.S. 2008. Press, Princeton, New Jersey, USA. Longevity can buffer plant and animal

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 188 Drivers of waterfowl population dynamics

populations against changing climatic Palacios, M.G., Winkler, D.W., Klasing, K.C., variability. Ecology 89: 19–25. Hasselquist, D. & Vleck, C.M. 2011. Murray, D.L., Anderson, M.G. & Steury, T.D. Consequences of immune system aging in 2010. Temporal shift in density dependence nature: a study of immunosenescence costs among North American breeding duck in free-living Tree Swallows. Ecology 92: populations. Ecology 91: 571–581. 952–966. Nicolai, C.A. & Sedinger, J.S. 2012. Are there Patterson, I.J. 1982. The Shelduck – A Study in trade-offs between pre- and post-fledging Behavioural Ecology. Cambridge University survival in Black Brent geese? Journal of Press, Cambridge, UK. Animal Ecology 81: 788–797. Péron, G., Nicolai, C.A. & Koons, D.N. 2012. Noel, L.E., Johnson, S.R., O’Doherty, G.M. Demographic response to perturbations: the & Butcher, M.K. 2005. Common Eider role of compensatory density dependence in (Somateria mollissima v-nigrum) nest cover and a North American duck under variable depredation on central Alaskan Beaufort Sea harvest regulations and changing habitat. barrier islands. Arctic 58: 129–136. Journal of Animal Ecology 81: 960–969. Northrup, J.M. & Wittemyer, G. 2013. Péron, G., & Koons, D.N. 2012. Integrated Characterising the impacts of emerging energy modeling of communities: parasitism, development on wildlife, with an eye towards competition, and demographic synchrony in mitigation. Ecology Letters 16: 112–125. sympatric ducks. Ecology 93: 2456–2464. Nudds, T.D. 1992. Patterns in breeding duck Péron, G., & Koons, D.N. 2013. Intra-guild communities. In B.D.J. Batt, A.D. Afton, interactions and projected impact of climate M.G. Anderson, C.D. Ankney, D.H. Johnson, and land use changes on North American J.A. Kadlec & G.L. Krapu (eds.), Ecology and pochard ducks. Oecologia 172: 1159–1165. Management of Breeding Waterfowl, pp. 540–567. Person, B.T., Herzog, M.P., Ruess, R.W., University of Minnesota Press, Minneapolis, Sedinger, J.S., Anthony, R.M. & Babcock, USA. C.A. 2003. Feedback dynamics of grazing Nummi, P. & Saari, L. 2003. Density-dependent lawns: coupling vegetation change with decline of breeding success in an introduced, animal growth. Oecologia 135: 583–592. increasing Mute Swan population. Journal of Pianka, E.R. 1970. On r- and K-Selection. Avian Biology 34: 105–111. American Naturalist 104: 592–597. Oli, M.K. & Zinner, B. 2001. Partial life-cycle Pfister, C.A. 1998. Patterns of variance in analysis: a model for pre-breeding census stage-structured populations: evolutionary data. Oikos 93: 376–387. predictions and ecological implications. Oli, M.K., Hepp, G.R. & Kennamer, R.A. 2002. Proceedings of the National Academy of Sciences, Fitness consequences of delayed maturity in USA 95: 213–218. female Wood Ducks. Evolutionary Ecology Pringle, C.M. 2000. Threats to U.S. public lands Research 4: 563–576. from cumulative hydrologic alterations outside Osnas, E.E., Runge, M.C., Mattsson B.J., Austin, of their boundaries. Ecological Applications 10: J., Boomer, G.S., Clark, R.G., Devers, P., 971–989. Eadie, J. M., Lonsdorf, E.V. & Tavernia, B.G. Promislow, D.E.L. & Harvey, P.H. 1990. Living 2014. Managing harvest and habitat as fast and dying young: a comparative analysis integrated components. Wildfowl (Special of life-history variation among mammals. Issue No. 4): 305–328. Journal of Zoology 220: 417–437.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 189

Raveling, D.G. 1989. Nest-predation rates in Rotella, J.J., Link, W.A., Nichols, J.D., Hadley, G.L., relation to colony size of Black Brant. Journal Garrott, R.A., and Proffitt, K.M. 2009. An of Wildlife Management 53: 87–90. evaluation of density-dependent and density- Reynolds, R.E., Shaffer, T.L., Renner, R.W., independent influences on population growth Newton, W.E., & Batt, B.D.J. 2001. Impact rates in Weddell Seals. Ecology 90: 975–984. of the Conservation Reserve Program on Rotella, J.J., Link, W.A., Chambert, T., Stauffer, duck recruitment in the US Prairie Pothole G.E. & Garrott, R.A. 2012. Evaluating the Region. Journal of Wildlife Management 65: demographic buffering hypothesis with vital 765–780. rates estimated for Weddell Seals from 30 Rice, S.H., Papadopoulos, A. & Harting, J. 2011. years of mark–recapture data. Journal of Stochastic processes driving directional Animal Ecology 81: 162–173. evolution. In P. Pontarotti (ed.), Evolutionary Runge, M.C., Johnson, F.A., Anderson, M.G., Biology – Concepts, Biodiversity, Macroevolution Koneff, M.D., Reed, E.T. & Mott, S.E. 2006. and Genome Evolution, pp. 21–33. Springer– The need for coherence between waterfowl Verlag, Berlin, Germany. harvest and habitat management. Wildlife Ricklefs, R.E. 2008. Disintegration of the Society Bulletin 34: 1231–1237. ecological community. American Naturalist Sæther, B.-E. 1988. Pattern of covariation 172: 741–750. between life history traits of European birds. Ringelman, K.M., Eadie, J.M. & Ackerman, J.T. Nature 331: 616–627. 2012. Density-dependent nest predation in Sæther, B.-E. & Bakke, O. 2000. Avian life history waterfowl: the relative importance of nest variation and contribution of demographic density versus nest dispersion. Oecologia 169: traits to the population growth rate. Ecology 695–702. 81: 642–653. Rockwell, R.F. & Gormezano, L.J. 2009. The Sæther, B.-E., Lillegård, M., Grøtan, V., Drever, early bear gets the goose: climate change, M.C., Engen, S., Nudds, T.D. & Podruzny, Polar Bears and Lesser Snow Geese in K.M. 2008. Geographical gradients in the western Hudson Bay. Polar Biology 32: population dynamics of North American 539–547. prairie ducks. Journal of Animal Ecology 77: Rockwell, R.F., Gormezano, L.J. & Koons, D.N. 869–882. 2011. Trophic matches and mismatches: can Schamber, J.L., Flint, P.L., Grand, J.B., Wilson, Polar Bears reduce the abundance of nesting H.M. & Morse, J.A. 2009. Population Snow Geese in western Hudson Bay? Oikos dynamics of Long-tailed Ducks breeding on 120: 696–709. the Yukon-Kuskokwim Delta, Alaska. Arctic Rohwer, F.C. 1985. The adaptive significance of 62: 190–200. clutch size in prairie ducks. Auk 102: Schmidt, K.A. & Ostfeld, R. S. 2008. Numerical 354–361. and behavioral effects within a pulse-driven Rohwer, F.C., Johnson, W.P., and Loos, E.R. system: consequences for shared prey. Ecology 2002. Blue-winged Teal (Anas discors). In A. 89: 635–646. Poole (ed.), The Birds of North America Online. Schmutz, J.A. 1993. Survival and pre-fledging Cornell Laboratory of Ornithology, Ithaca, body mass in juvenile Emperor Geese. USA. Available at http://bna.birds.cornell/ Condor 95: 222–225. bna/species/625doi:10.2173/bna.625 (last Schmutz, J.A. 2009. Stochastic variation in accessed 14 March 2014). avian survival rates: life-history predictions,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 190 Drivers of waterfowl population dynamics

population consequences, and the potential Sibly, R.M. & Brown, J.H. 2007. Effects of body responses to human perturbations and size and lifestyle on evolution of mammal climate change. Environmental and Ecological life histories. Proceedings of the National Statistics 3: 441–461. Academy of Sciences, USA 104: 17707–17712. Schmutz, J.A. & Laing, K.K. 2002. Variation in Silvertown, J., Franco, M. & McConway, K. 1992. foraging behavior and body mass in broods A demographic interpretation of Grime’s of Emperor Geese (Chen canagica): evidence triangle. Functional Ecology 6: 130–136. for interspecific density dependence. Auk Sjöberg, K., Gunnarsson, G., Pöysä, H., 119: 996–1009. Elmberg, J. & Nummi, P. 2011. Born to Schmutz, J.A., Rockwell, R.F. & Petersen, M.R. cope with climate change? Experimentally 1997. Relative effects of survival and manipulated hatching time does not affect reproduction on the population dynamics duckling survival in the Mallard Anas of Emperor Geese. Journal of Wildlife platyrhynchos. European Journal of Wildlife Management 61: 191–201. Research 57: 505–516. Sedinger, J.S. & Alisauskas R.T. 2014. Cross- Smith, L.M., Pederson, R.L. & Kaminski, R.M. seasonal effects and the dynamics of (eds.) 1989. Habitat Management for Migrating waterfowl populations. Wildfowl (Special Issue and Wintering Waterfowl in North America. Texas No. 4): 277–304. Tech University Press, Lubbock, Texas, USA. Sedinger, J.S. & Chelgren, N.D. 2007. Survival Smith, P.A., Elliott, K.H., Gaston, A.J. & and breeding advantages of larger Black Gilchrist, H.G. 2010. Has early ice clearance Brant goslings: within and among cohort increased predation on breeding birds by variation. Auk 124: 1281–1293. Polar Bears? Polar Biology 33: 1149–1153. Sedinger, J.S., Flint, P.L. & Lindberg, M.S. 1995. Stahl, J.T. & Oli, M.K. 2006. Relative importance Environmental influence on life-history of avian life-history variables to population traits: growth, survival, and fecundity in growth rate. Ecological Modelling 198: 23–39. Black Brant (Branta bernicla). Ecology 76: Stearns, S.C. 1992. The Evolution of Life Histories. 2404–2414. Oxford University Press, New York, USA. Sedinger, J.S., Lindberg, M.S., Person, B.T., Stenseth, N.C., Viljugrein, H., Saitoh, T., Hansen, Eichholz, M.W., Herzog, M.P. & Flint, P.L. T.F., Kittilsen, M.O., Bølviken, E. & 1998. Density dependent effects on growth, Glöckner, F. 2003. Seasonality, density body size and clutch size in Black Brant. Auk dependence, and population cycles in 115: 613–620. Hokkaido Voles. Proceedings of the National Sedinger, J.S., Nicolai, C.A., Lensink, C.J., Academy of Sciences, USA 100: 11478–11483. Wentworth, C. & Conant, B. 2007. Black Stephens, S.E., Rotella, J.J., Lindberg, M.S., Taper, Brant harvest, density dependence, and M.L. & Ringelman, J.K. 2005. Duck nest survival: a record of population dynamics. survival in the Missouri Coteau of North Journal of Wildlife Management 71: 496–506. Dakota: landscape effects at multiple spatial Sedinger, J.S., Chelgren, N.D., Lindberg, M.S. & scales. Ecological Applications 15: 2137–2149. Ward, D.H. 2008. Fidelity and breeding Stirling, I. & Derocher, A.E. 1993. Possible probability related to population density and impacts of climatic warming on Polar Bears. individual quality in Black Brent geese (Branta Arctic 46: 240–245. bernicla nigricans). Journal of Animal Ecology 77: Stirling, I. & Parkinson, C.L. 2006. Possible 702–712. effects of climate warming on selected

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 Drivers of waterfowl population dynamics 191

populations of Polar Bears (Ursus maritimus) Viljugrein, H., Stenseth, N.C., Smith, G.W. & in the Canadian Arctic. Arctic 59: 261–275. Steinbakk, G.H. 2005. Density dependence in Strong, L.L. & Trost, R.E. 1994. Forecasting North American ducks. Ecology 86: 245–254. production of Arctic nesting geese by Visser, M.E. & Both, C. 2005. Shifts in monitoring snow cover with advanced very phenology due to global climate change: the high resolution radiometer (AVHRR) data. need for a yardstick. Proceedings of the Royal Proceedings of the Pecora Symposium 12: 425– Society, B 272: 2561–2569. 430. Walker, J., Rotella, J.J., Stephens, S.E., Lindberg, Tombre, I.M., Black, J.M. & Loonen, M.J.J.E. 1998. M.S., Ringelman, J.K., Hunter, C. & Smith, Critical components in the dynamics of a A.J. 2013. Time-lagged variation in pond Barnacle Goose colony: a sensitivity analysis. density and primary productivity affects duck In F. Mehlum, J.M. Black & J. Madsen (eds.), nest survival in the Prairie Pothole Region. Research on Arctic Geese, pp. 81–89. Proceedings Ecological Applications 23: 1061–1074. of the Svalbard Goose Symposium, Oslo, Walters, C. 1986. Adaptive Management of Renewable Norway, Norsk Polarinstitutt Skrifter 200. Resources. MacMillan, New York, USA. Tombre, I.M., Høgda, K.A., Madsen, J., Griffin, Wetherald, R. 2009. Changes of variability in L.R., Kuijken, E., Shimmings, P., Rees, E. & response to increasing greenhouse gases. Verscheure, C. 2008. The onset of spring and Part II: hydrology. Journal of Climate 22: timing of migration in two arctic nesting 6089–6103. goose populations: the pink-footed goose Williams, T.D., Cooch, E.G., Jefferies, R.L. & Anser brachyrhynchus and the barnacle goose Cooke, F. 1993. Environmental degradation, Branta leucopsis. Journal of Avian Biology 39: food limitation and reproductive output: 691–703. juvenile survival in Lesser Snow Geese. Tuljapurkar, S.D. 1982. Population dynamics Journal of Animal Ecology 62: 766–777. in variable environments. II. Correlated Wilson, H.M., Flint, P.L., Powell, A.N., Grand, environments, sensitivity analysis and J.B. & Moran, C.L. 2012. Population ecology dynamics. Theoretical Population Biology 21: of breeding Pacific Common Eiders on the 114–140. Yukon-Kuskokwim Delta, Alaska. Wildlife Turchin, P. 2003. Complex Population Dynamics. Monographs 182: 1–28. Princeton University Press, Princeton, New Wisdom, M.J., Mills, L.S. & Doak, D.F. 2000. Jersey, USA. Life-stage simulation analysis: estimating van der Jeugd, H.P., Eichhorn, G., Litvins, K.E. vital-rate effects on population growth for Stahl, J., Larsson, K., van der Graaf, A.J. & conservation. Ecology 81: 628–641. Drent, R.H. 2009. Keeping up with early Wright, C.K. & Wimberly, M.C. 2013. Recent springs: rapid range expansion in an avian land use change in the western corn belt herbivore incurs a mismatch between threatens grasslands and wetlands. Proceedings reproductive timing and food supply. Global of the National Academy of Sciences, USA 110: Change Biology 15: 1057–1071. 4134–4139.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 169–191 192

Conspecific brood parasitism in waterfowl and cues parasites use

HANNU PÖYSÄ1*, JOHN M. EADIE2 & BRUCE E. LYON3

1Finnish Game and Fisheries Research Institute, Joensuu Game and Fisheries Research, Yliopistokatu 6, FI-80010 Joensuu, Finland. 2Department of Wildlife, Fish and Conservation Biology, University of California, One Shields Avenue, Davis, California 95616, USA. 3Department of Ecology and Evolutionary Biology, University of California, Santa Cruz, California 95064, USA. *Corresponding author. E-mail: [email protected]

Abstract Conspecific brood parasitism (CBP) occurs in various insects, fishes and birds, but it is disproportionately common in waterfowl (Anatidae). Studies of CBP in Anatids therefore have helped to develop a fundamental conceptual framework with which to explain this intriguing behaviour. Yom-Tov (1980) first drew attention to CBP, and Andersson and Eriksson (1982) also hinted at the fascinating behavioural, ecological and evolutionary aspects of CBP in waterfowl. Several reviews followed these early papers, but much has been learned more recently about CBP in waterfowl. Here we aim to review the traditional conceptual framework of CBP in waterfowl and to consider empirical studies that have attempted to test related hypotheses. The survey provided support for the hypotheses that CBP allows some females to reproduce when not otherwise possible, whereas other females use parasitic egg- laying as a way to enhance their fecundity. A recently developed framework that considers CBP as part of a flexible life-history strategy could provide a useful direction for future studies of CBP. A second aim of this review is to consider the use of cues by conspecific brood parasites seeking suitable places to lay eggs parasitically. Recent studies have revealed remarkable cognitive abilities in parasitic females, but the actual mechanisms remain unknown. Clearly, breeding females are sensitive to cues such as nest site security, patterns of previous nest use or success, clutch size, and perhaps even the degree of kinship between hosts and other parasites. Indeed, additional investigations of CBP are needed to provide a better understanding of the processes and patterns of this avian reproductive strategy.

Key words: Anseriformes, brood parasitism, information use, life-history, nest predation risk.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 193

Conspecific brood parasitism (CBP) is an occurrence and evolution of CBP in birds alternative reproductive tactic in which a (Eadie et al. 1988; Rohwer & Freeman 1989; female lays eggs in the nests of other Sayler 1992; Lyon & Eadie 2008) and this conspecific individuals and leaves the paper does not intend to provide another subsequent care of the eggs and young to broad-ranging overview of the CBP the host female. CBP has been documented breeding strategy. However, empirical and in at least 234 species of birds and is theoretical research on CBP has grown since particularly prevalent in Anseriformes the original paper by Yom-Tov (1980) and where it has been reported in 76 of the 161 much of this work has focused on species (Yom-Tov 2001). waterfowl (Fig. 1). Development of theory Several comprehensive reviews have been to explain CBP grew through the 1990s published on the hypotheses for the peaking in 2001–2005. Empirical studies

2.0 (a)

1.5

1.0 p ers p er year

Pa 0.5

0.0 1986–1990 1991–1995 1996–2000 2001–2005 2006–2010 2011–

5.0 (b)

4.0

3.0

p ers p er year 2.0 Pa

1.0

0.0 1986–1990 1991–1995 1996–2000 2001–2005 2006–2010 2011– Figure 1. Number of (a) theoretical papers published on CBP in general (top panel; n = 24 papers in total), and (b) empirical papers published on CBP in waterfowl (bottom panel; n = 77) over the past three decades.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 194 Conspecific brood parasitism in waterfowl lagged and reached their highest frequency though parasitism represents a relatively in the last 5–10 years (Fig. 1). Many cheap way to gain fitness at the expense advances have been made, but there is a of other individuals, this does not mean surprising amount that is not yet known. that parasites should lay their eggs Indeed, for many species, it is still not clear indiscriminately. To the extent that fitness which females within a population pursue from parasitism can be enhanced by this behaviour, nor do we fully understand decisions that parasites make regarding the fitness consequences to parasites or where to lay their eggs, or how many eggs to their hosts. Much of the work to date has lay in a given nest, natural selection should focussed on ecological factors that correlate favour those decisions or tactics. However, with the occurrence of CBP, but parasites must be able to gather useful longitudinal studies of females are still rare. information about potential host nests that Similarly, most studies are observational, they can use to inform their laying decisions. albeit with an expanded toolkit of molecular Do they gather this information, and if so, genetic techniques which help to ascertain what cues do they use? A second goal of maternity. Experimental studies are this paper therefore is to examine the uncommon with a few notable exceptions growing body of work that is beginning (Eadie 1989; Pöysä 2003a,b; Pöysä et al. to explore the cues used by conspecific 2010; Odell & Eadie 2010). Despite these brood parasites. The focus here is on gaps, the field is now at a point where some specific cues that the parasites may use to retrospection would be valuable. The initial select host nests into which to lay eggs. This goal of this paper therefore is to review is not meant to imply that cues and decisions briefly the traditional set of hypotheses used by hosts are not important. For posed to account for CBP in waterfowl and example, hosts may desert nests in response to evaluate how existing empirical work to CBP (Eadie 1989; Jaatinen et al. 2009), meets those expectations. We then offer an resulting in the direct loss of parasitic alternative conceptual framework proposed eggs. Hosts could also influence which by Lyon & Eadie (2008) that could advance parasites gain access to their nests (Åhlund our understanding of this behaviour more 2005). effectively. Secondly, considerable growth in this Conspecific brood parasitism in field involves the information that might be waterfowl available to parasites and hosts to modulate their behaviour in an adaptive manner. How Here we provide a brief overview of the set do females choose a nest or host to of traditional hypotheses that have been parasitize? What information might be suggested to explain CBP, outlined in earlier available to females to shape their more comprehensive reviews (including behavioural decisions? Brood parasites gain Andersson 1984; Eadie et al. 1988; Sayler fitness by having other females provide 1992; Lyon & Eadie 2008), and examine the parental care for their offspring. Even evidence from the literature in support of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 195 these hypotheses for waterfowl. We then empty nests without a host female being present a revision of the framework of present. Hence, parasitic laying in these hypotheses proposed by Lyon and Eadie instances cannot be explained as a result of (2008) to help guide further research in this the laying female being “ousted” from the field. nest by the host female who incubates the eggs, since there was no host with whom to Is it parasitism or inadvertent compete. Second, Eadie (1989) conducted competition? removal experiments with Barrow’s Some early researchers considered parasitic Goldeneye Bucephala islandica and Common egg-laying (also called “egg-dumping” in the Goldeneye B. clangula and found that, when older literature) in waterfowl to be non- the host female was removed, putative adaptive, either because it reflects abnormal parasites continued to lay eggs but did not behaviour (“loss of maternal instinct”) or incubate the eggs, despite the fact that there because it is a side-effect of competition for was no female to prevent them from doing suitable nest sites in hole-nesting waterfowl so. Conversely, when the parasitic females (Erskine 1990). Semel and Sherman (2001) were removed, the hosts continued to resurrected this idea and proposed a similar lay and ultimately incubate the clutch, mechanism to account for “apparent” CBP demonstrating that the different response by in Wood Duck Aix sponsa. They proposed parasites was not simply due to an effect of that some nest sites are preferred, perhaps experimental disturbance. This suggests that because of their quality or because young parasites and hosts behave very differently females return to their natal nest on right from the outset. breeding for the first time. Contests for Finally, recent work by Åhlund (2005) has these nests ensue with more than one female demonstrated striking differences in the laying eggs in the nest, but ultimately only a behavioural tactics of parasitic and host single female incubates the clutch. The Common Goldeneye females at the nest: usurped females become de facto parasites. hosts and parasites differed in the timing of Parasitism was not the focus of their egg-laying, deposition of down, covering behaviour and arises only as an inadvertent eggs on departure, and time spent on the consequence of laying eggs and failure to nest as the egg-laying sequence progressed. establish final ownership of the nest (i.e. These observations suggest that CBP is a accidental parasitism). genuine reproductive tactic and not just a However, several recent lines of evidence consequence of nest site competition argue against the accidental parasite (Åhlund 2005). Similarly, recent hypothesis as an explanation for CBP in experimental studies have revealed waterfowl, and in other birds as well. First, sophisticated responses of parasitically several researchers (Eadie 1989; Pöysä laying Common Goldeneye females to 2003b; Odell & Eadie 2010) observed variation in nest (egg) predation risk (Pöysä frequent parasitic egg-laying in nests in 2003a; Pöysä et al. 2010), while other studies which eggs were added experimentally to have demonstrated clear fitness advantages

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 196 Conspecific brood parasitism in waterfowl of CBP for parasitically laying Common during egg-laying or early incubation may be Goldeneye females (Åhlund & Andersson able to lay some additional eggs (or may 2001). The fact that CBP occurs in such a have eggs already developing in the ovary) large number of waterfowl species, many of but are not able (or it is not worthwhile) to which do not nest in tree cavities and do not establish a new nest. This hypothesis could compete for specific nests sites, argues be classified as a form of constraint (BOBJ), against the accidental parasite hypothesis as but many researchers have discussed it as a a general explanation of this behaviour. separate mechanism (and so we list it here in These observations, together with the that form for comparison). discovery that brood parasites often make fine-tuned, adaptive egg-laying decisions in “Professional” or life-long specialist parasites. waterfowl and other species (Brown & Under this hypothesis, females never raise Brown 1991; Lyon & Everding 1996; Pöysä their own young and only lay in the nests of 1999; Lyon & Eadie 2008 and below), other females. It is argued that these females confirm that CBP is generally an adaptive have higher lifetime fitness (when rare in the alternative reproductive strategy in both population), because they are emancipated waterfowl and other birds. from the costs of parental care and so are able to invest in additional production of Traditional hypotheses eggs. Under a game theoretic version of this Adaptive hypotheses about the CBP have hypothesis, negative frequency-dependent traditionally been classified into four types selection works to stabilise the frequencies (summarised from Lyon & Eadie 2008) as of nesting and parasite females in the follows: population (a mixed evolutionarily stable strategy; ESS). Best-of-a-bad-job (BOBJ). According to this hypothesis, females lay eggs parasitically Fecundity enhancement. This hypothesis posits when they are unable to breed otherwise that nesting females also lay some additional (constraint), or when environmental eggs parasitically and, by doing so, are able conditions are unfavourable such that the to increase fitness beyond that possible prospects for successful reproduction by through nesting alone, presumably by nesting are low (restraint). A variety of bypassing some of the constraints or costs ecological and physiological factors have of raising the additional eggs/young on been proposed to influence a female’s ability their own. to nest on her own, including nest site or We were able to locate 17 studies that territory limitation, body condition, age and have attempted to test at least some of these experience. hypotheses for waterfowl (Table 1). Of the four traditional hypotheses, no support has Nest loss. A variant of the BOBJ hypothesis been found for either the nest loss focuses on nest loss as the causative factor. hypothesis (0 of 5 studies that examined this Females that lose their nest to predation hypothesis) or the life-long “professional”

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 197 2009 1992 1989 et al. et al. et al . 2011 et al . 2004 et al . 2010 et al. Source Pöysä 1999 Pöysä Briggs 1991 Åhlund 2005 Sorenson 1993 Sorenson 1991 Lank Robertson 1998 Robertson Eadie 1989, 1991 Waldeck Waldeck Waldeck Waldeck Reichart Reichart Robertson Robertson Anderholm Semel & Sherman 2001 Forslund & Larsson 1995 Forslund No No No No Yes Åhlund & Andersson 2001 No Yes No No 1991 & Lamprecht Weigmann parasite enhancement No No No No Yes No Hypotheses No No Yes Yes Yes Yes Yes Yes Yes Yes Yes/No Best-of-a-bad-job Nest loss Life-long Fecundity Summary ofapplicability ofthe empirical studies in which of traditional hypotheses conspecific brood parasitism (CBP) Common Eider Common Eider Common and Goldeneye Barrow’s Common Goldeneye (nest site limitation) Yes Common Goldeneye No (nest site limitation)Common Goldeneye Duck Ruddy No No (nest site limitation) No No Yes Redhead Duck Wood Common EiderCommon Eider No (nest site limitation) No (nest site limitation) Maned Duck Canvasback have been assessed. Tests of been assessed. Tests in parentheses. have are given a specific best-of-a-bad-job hypothesis Table 1. Table Species Goose Bar-headed Snow Goose Snow Barnacle Goose Barnacle Goose

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 198 Conspecific brood parasitism in waterfowl parasite hypothesis (0 of 6 studies) although Common and Barrow’s Goldeneyes in the later may be difficult to detect given that British Columbia (Eadie 1989; Jaatinen it would be beyond the scope of most et al. 2009, 2011). Clearly, more than studies to follow a “pure” parasite one hypothesis can apply to the same throughout her entire lifetime (requires population. How best to make sense of the detailed observational data, recaptures, range of outcomes summarised in Table 1 is genotyping of all eggs in a population, etc.). considered below. Nonetheless, in most cases where females have been followed through time, they have A revised framework for future been observed to switch between nesting research and parasitism (or to use both strategies in Lyon and Eadie (2008) pointed out that the the same year – dual nesters) suggesting that traditional set of hypotheses are potentially pure parasites, if they occur, are rare (Eadie confounded at several levels, conflating 1989; Sorenson 1991; Åhlund & Andersson what a female does (nest, parasitize, or both) 2001; Reichart et al. 2010). To date, the with fitness benefits of doing so, with hypotheses that are best supported are ecological factors influencing her decision versions of the BOBJ hypothesis (10 of 16 (nest loss, nest limitation, host availability, studies) and, to a lesser extent, the fecundity etc.), and finally with the evolutionary enhancement hypothesis (4 of 8 studies; dynamics that maintain some frequency Table 1). Thus, at least for most waterfowl, of CBP in the population (frequency- there is support for the idea that some dependent ESS). Lyon and Eadie (2008) females pursue CBP due to constraint or proposed a revision to the traditional set of restraint, whereas other females appear to hypotheses and they based this revision do so to enhance total reproductive output. partly on a conceptual framework derived Few of these studies were able to determine from Sorenson’s (1991) reproductive which females did what. Perhaps the most decision model. Under this model, the interesting are the results of Åhlund and ability to lay some eggs parasitically Andersson (2001) who showed that parasitic allows females to fine-tune reproductive Common Goldeneyes comprised both investment because without the possibility females that only laid parasitically in a given of CBP, females are faced with an all-or- season (pure parasites) and others that laid none decision to nest or not to reproduce parasitically and also had a nest of their at all. Key to this framework is the own (dual nesters). The reproductive fundamental difference between two “payoffs” varied considerably – dual nesters contexts of brood parasitism – parasitism produced 1.5 times more offspring than by non-nesting females and parasitism by non-parasitic (nesting) females and 2 times nesting females. For non-nesting females, that of pure parasites. By combining parasitism allows for an intermediate parasitism with normal nesting, some investment between no reproduction and females were able to double their nesting. Thus, females prevented from reproduction. Similar patterns occur in nesting can gain some fitness through

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 199 parasitism where otherwise none would be offspring survival, and (c) adult survival (i.e. possible. Parasitism by nesting females, in future fecundity). Of particular importance contrast, allows females to increase their is the distinction between parasites with and reproductive effort when conditions are without their own nests because these two very good without entailing a full second contexts likely involve different constraints, nest effort. Parasitic egg-laying allows these and different hypotheses may apply. females to adjust reproductive effort Accordingly, there are three questions that upwards in smaller increments to match must be addressed to understand the benefit expected returns. of CBP: (a) does the female have a nest or This conceptual framework is a useful not, (b) what fitness components and life- advance in two ways. First, it unifies all four history trade-offs play a role in leading to possible nesting options (not breeding, increased fitness benefits via parasitism, and parasitize only, nest only, parasitize and nest) (c) what ecological, social, or physiological as part of a single continuum that varies factors influence these trade-offs? Table 2 from low-to-high reproductive investment, summarises Lyon and Eadies’ (2008) revised and low-to-high expected fitness benefits. hypothesis framework. Data are not This captures the variation found both available to test these hypotheses (few within and among species of waterfowl studies have followed the life-histories (Table 1). Second, parasitism can be of individual parasites) and so we combined with nesting in various ways over cannot place current studies in this new a female’s lifetime to provide a flexible life- context. However, this is a new, more history, whereby females are able to modify integrated framework for future studies their reproductive investment and options of CBP in waterfowl and other birds. to variable ecological and social conditions. Perhaps the biggest requirement to This framework moves the field forward improve understanding of this intriguing from considering a large number of single reproductive system is to determine more independent hypotheses for each type of clearly what females are doing, both within a parasitism that intermix ecological factors, breeding season (nest, parasitize, or both) proximal influences, fitness benefits and and among breeding seasons. Once a evolutionary dynamics (the traditional female’s nesting status is determined, the framework; Table 1) into a more general life- context and suite of relevant hypotheses can history context with hypotheses that focus be analysed more carefully and thus on the specific life-history trajectories of thoroughly evaluated (Table 2). This opens a females and the expected fitness returns wide range of new and intriguing questions from pursuing those alternatives. about information use by the females which With this new framework, Lyon and pursue these alternative pathways, and it is Eadie (2008) proposed a modified perhaps here where some of our newest categorisation of hypotheses for CBP, insights on CBP have emerged. The second focusing on the three key fitness half of this review focuses on these new components: (a) current fecundity, (b) developments.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 200 Conspecific brood parasitism in waterfowl

Table 2. A framework of hypotheses on the adaptive benefits of conspecific brood parasitism, modified from Table 1 in Lyon & Eadie 2008. This framework emphasises the distinction between hypotheses that apply to females without nests (strategy a) versus parasitizing females that also nested (strategy b).

Strategy Mechanism Fitness component Traditional enhanced hypotheses

(a) Non-nesting Egg production Current fecundity Not emphasised; parasite (bypass costs of nesting could explain life- and allocate more effort long parasites into egg production) Nest/territory limitation Adult survival BOBJ (constraint, (unable to obtain a nest and/or current salvage strategy, site or territory) fecundity nest limitation) Energy/condition/ Adult survival BOBJ (restraint, experience salvage strategy, (females in poor energy limitation) condition, young) Quality of brood rearing Offspring survival Nest predation (parasites lay in high could apply quality nests; good hosts/ safe sites) (b) Nesting Nest loss Current fecundity BOBJ (constraint, parasite (loss during egg-laying) salvage strategy, nest loss) Clutch size/brood size Current fecundity Fecundity constraints and/or offspring enhancement (high quality females in survival Side-payment excellent condition Dual nesting increase egg production, bypass brood size constraints) Cost of reproduction Adult survival Not emphasised (reduce cost of care in own nest/brood to enhance future reproduction)

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 201

Cues used by parasitic breeders types of information are used when making choices about parasitism. There is a long history of discussion about Cues can reflect characteristics of the the distinction between signals and cues physical environment or the social (see Danchin et al. 2008; Wagner & environment. Types of information of Danchin 2010). Signals are traits that have particular interest in the context of CBP been designed by selection to convey are “personal information” and “social information, whereas cues incidentally information”. The former is the contain information but were not selected information obtained by an individual’s own for that purpose (Maynard Smith & Harper interaction with the environment (e.g. its 2003; Danchin et al. 2008). Information experience or history, such as the failure of a about the attributes of potential hosts or previous breeding attempt), whereas the their nests are most likely cues, and signals latter refers to information obtained by might be expected in species where kin observing other individuals (e.g. their selection plays a role in facilitating the location, phenotypic condition and occurrence of CBP. Studies of obligate reproductive performance). Danchin et al. brood parasites nicely illustrate the (2008) term social information and non- importance of information use to parasitic private personal information (i.e. the tactics. Reproductive fitness for obligate information accessible to other individuals) parasites depends entirely on the success of as “public information”. Social information their parasitic eggs, so there is strong may also be based on signals, i.e. traits selection to employ tactics that enhance the that evolve and are involved in true survival of the eggs and young. For communication between individuals (see example, obligate parasites often need to Danchin et al. 2008). However, the role of find nests of the right host species and signals in CBP has not been addressed (for collect information that enables them to interspecific brood parasitism, see Parejo & time their laying to match that of the host’s Avilés 2007). breeding cycle (Davies 2000). Although total How might conspecific brood parasites female fitness in species with conspecific use information to modify parasitic egg- parasitism does not depend nearly as laying behaviour adaptively? Three issues heavily on gains from parasitic eggs as must be considered: 1) how do researchers obligate parasites, fitness from parasitism test for evidence of cue use; 2) what can nonetheless be substantial (Åhlund behaviours might enable potential parasites & Andersson 2001). Thus, whenever to acquire information; and 3) what specific parasitism is a well-developed component of cues might be used? The last question reproduction, we expect that the use of cues further entails a nested series of choices that to select hosts will be important as well. The parasites might make, each involving a question then is whether conspecific brood distinct cue or set of cues, such as: (a) what parasites show tactics similar to those of general area to use (habitat cues), (b) which obligate brood parasites and, if so, what nests or females to parasitize, and (c) when

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 202 Conspecific brood parasitism in waterfowl to parasitize a host nest. In addition to these on parasitism (Brown & Brown 1991; choices, parasites need to decide how many Andersson & Åhlund 2000). Monte Carlo eggs to lay in a given nest and, if parasites randomisations provide one powerful also have their own nest, how many eggs to method for assessing patterns of brood allocate to nesting versus parasitism. parasitism in this context (Emlen & Wrege Information gained by parasites could 1986; Lyon 1993). McRae and Burkes’ influence any of these decisions and so it is (1996) study of Moorhens Gallinula choropus important to clarify which aspect is being highlights the value of controlling for the addressed. This in turn requires that spatial pool of available hosts. Host-parasite researchers are thoughtful about their relatedness was higher than expected at the methodology and careful in interpreting the population level but was not different from patterns observed. random expectation given the pool of hosts actually available to brood parasites. Methods used to detect cue use Controlling for the pool of potential hosts Three methods can be used to elucidate the revealed that parasites were not specifically tactics and cues used by brood parasites. targeting relatives (McRae & Burke 1996). Simplest is a comparison showing that However, neither population comparisons parasitized and non-parasitized nests differ nor contrasts that control for with respect to some attribute likely to be spatiotemporal patterns of potential hosts important to the success of the parasite, provide definitive evidence for which cues such as the quality of the territory, nest or parasites actually use to select nests. The host. However, this method provides problem is that factors that correlate with somewhat weak evidence for parasitic parasitism may not be the actual cues that tactics because the patterns may reflect the parasites use when choosing nests to outcome of host defenses and not parasitic parasitize. Only experiments manipulating tactics. It also can be possible to obtain false putative cues provide fully convincing positive evidence for non-random patterns evidence for cue use. These experiments are of parasitism. For example, spatiotemporal rarely done, but they have been conducted clustering in attributes of host nests can in a few waterfowl species (see below). result in patterns when data are analysed for the entire population but not at spatial and How do females obtain information? temporal scales that are relevant to the Studies addressing behavioural aspects of choices that individual parasites face (e.g. nest site selection in Barrow’s Goldeneye Lyon 1993; McRae & Burke 1996). To assess and Common Goldeneye have revealed that parasitic tactics properly, it may be necessary females gather information by prospecting to understand the spatial and temporal for potential nest sites prior to the next patterns of parasitism and then assess the breeding season (Eadie & Gauthier 1985; choices parasites make with respect to the Zicus & Hennes 1989; Pöysä et al. 1999). pool of hosts that are actually available Pöysä (2006) found for the latter species that given the spatial and temporal constraints this behaviour is associated with CBP: nest

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 203 sites that were visited more frequently by Prospecting activity peaks after most nests prospecting females in year t had a higher have hatched (and ducklings have left nests), probability of being parasitized in year t + 1, matching the time when cues of a successful suggesting that parasites gather information nest (eggshell membranes and fragments; through nest-site prospecting to target Fig. 2) are highly visible (J. Eadie, unpubl. parasitic laying in particular nests. data; H. Pöysä, unpubl. data). It is not

A. Wood Duck

(a) Not used (shavings, (b) Abandoned or laying (c) Active incubation (down, undisturbed) (no down) eggs covered)

(d) Active incubation/laying (e) Successful hatch (shell (f) Failed/depredated (rotten, (female, eggs, down) membranes, egg caps) broken eggs, shells)

B. Common Goldeneye

(a) Successful hatch (7 ducklings (b) Successful hatch (9 ducklings hatched hatched and left the nest) of 11 eggs and left the nest)

Figure 2. Examples of cues available in (A) Wood Duck nests and (B) Common Goldeneye nests after nesting. A: Wood Duck nests showing various stages of nesting, from not used (a) through to failed (f). B: Common Goldeneye nests showing two different examples of successful nests.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 204 Conspecific brood parasitism in waterfowl known if prospecting behaviour is conspicuous host behaviour, host activity; associated with CBP in the Barrow’s see Patten et al. 2011). While characteristics Goldeneye but, interestingly, prospecting associated with nest placement in waterfowl activity seems to be higher at nest sites that may not be as diverse as they are in had been parasitized earlier in the season passerines, the most important group of (see Fig. 2 in Eadie & Gauthier 1985). host species for interspecific brood Prospecting behaviour is not restricted to parasites, some general patterns emerge. hole-nesting species. Schamel (1977) First, CBP in waterfowl is more frequent in mentions that preferred nest-sites are visited cavity-nesting species than in species that (prospected) regularly by non-breeding nest in emergent vegetation or upland females throughout the summer (i.e. after (Rohwer & Freeman 1989; Sayler 1992; hatching) in the ground-nesting Common Eadie et al. 1998), implying that the ease in Eider Somateria mollissima, another duck in locating nest sites could play a role in CBP. which CBP is common (e.g. Robertson 1998; Support for this idea also comes from Waldeck et al. 2008). Fast et al. (2010) showed cavity-nesting Wood Ducks where highly experimentally that nest-site materials left visible nest boxes are more frequently from the previous year influence the use of parasitized than less visible nest boxes nest bowls by Common Eider females: nest (Semel et al. 1988; Roy Nielsen et al. 2006a; bowls containing down were occupied but see Jansen & Bollinger 1998 for a less earlier than control nest bowls with no clear effect). On the other hand, nest box down. As discussed by the authors, one visibility does not seem to affect the explanation could be that nest down may frequency of parasitism in another cavity- indicate previous nest success and nest-site nesting duck, the Barrow’s Goldeneye safety to females prospecting for nests. It (Eadie et al. 1998). Similarly, Åhlund (2005, would be interesting to study whether these p. 434) mentions that parasitism rate does aspects, i.e. prospecting behaviour and cues not differ between nests near the shore and indicating previous nest success, are nests further inland in the Common associated with the occurrence of CBP in Goldeneye population he studied; visibility the species. presumably differed considerably between the nest site types. In line with this, an Cues used by parasites to locate nests experiment addressing nest site selection in Cues used by parasitically laying females Common Goldeneye revealed that females to find and select suitable host nests have (potential parasites) prospect shore and been studied extensively in the context forest boxes equally, irrespective of of interspecific (obligate) avian brood differences in the visibility of the nest boxes parasitism. The main hypotheses can be (Pöysä et al. 1999), suggesting that females classified as those dealing with nest are very capable at finding nest sites. Hence, placement (e.g. nest exposure, characteristics while highly visible nest sites may be easier of the surrounding habitat) and those to locate, there must be other cues parasites dealing with host behaviour (e.g. use to locate and select nests.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 205

Early observations suggested that selection, and the design of some studies parasitic females use the activity of other does not allow a clear separation (see text on females (potential hosts) to locate nests. For methodology above). In this section we example Weller (1959) describes several consider only those studies that deal with cases in which Redhead Aythya americana the final step of the process, i.e. actual females apparently observed the nest- selection of nests by parasites, and review a building and egg-laying activities of other variety of cues that have been identified. females, leading the author to suggest that parasitic females used this behaviour to find Nest site quality or state nests. Similarly, several authors have Empirical studies have explored several suggested that Wood Duck females have a possible cues parasites may use to select a “decoying effect” on one another, leading nest (Table 3). Several studies have often to heavily parasitized nests addressed nest site characteristics while (Heusmann et al. 1980; Semel & Sherman characteristics of the host female have 1986; see also Roy Nielsen et al. 2006a). received less attention. We are aware of only Inspired by these observations Wilson one study for waterfowl, on nest-box- (1993) carried out an experiment using breeding Common Goldeneye, in which Wood Duck decoys and found evidence for both nest and host traits were considered, the hypothesis that parasitic Wood Duck and nest site characteristics turned out to be females use the presence of conspecifics as more important than those of the host a cue in the selection of nests. On the other female (Paasivaara et al. 2010). Parasitism in hand, decoy nests with experimental eggs relation to nest site quality has been well- but no host were parasitized at the same rate studied in a non-waterfowl species, the Cliff as real nests that did have a host in Common Swallow Hirundo pyrrhonota (Brown & Brown Goldeneyes (Pöysä 2003b), Barrow’s 1991). Parasitic Cliff Swallows show Goldeneyes (Eadie 1989) and Wood Ducks remarkable sophistication in their ability to (Odell & Eadie 2010), suggesting that the target host nests that are more likely to be presence of a conspecific host is not a successful than average, in part due to lower necessary cue for parasites. infestation by blood sucking nest parasites. Nest age was also identified as one cue used Cues used by parasites to select a nest by parasitic females (Brown & Brown 1991): Location of potential nests for egg laying nest age may be a reliable indicator of the is the first step in the process of nest safety of a particular nest site. selection of parasites. However, not all Nest success and nest site safety, traits of the located nests will eventually be that do not necessarily mean the same thing, parasitized (H. Pöysä, unpubl. data) have been found to be associated with the suggesting that parasites actively select occurrence of CBP in some waterfowl among potential nest sites. It is not always species but not in others (Table 3). For easy to make a clear distinction between example, in the Common Goldeneye these two steps in the process of nest parasitism in a given year occurred more

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 206 Conspecific brood parasitism in waterfowl frequently in nest sites that were not targeted by parasites. This would present an depredated (at least one duckling hatched interesting situation of conflicting cues and left the nest) during the previous nesting (predation risk or nest abandonment) attempt than in nest sites that were either and females might be predisposed to one depredated or control nest sites (Pöysä source of information (a sensory trap). 1999). A later study revealed a mechanism Alternatively, this pattern might result if by which parasitically laying females identify parasitic egg-laying in deserted nests was safe nest sites, i.e. by nest site prospecting more frequent because of a lack of host during the previous year (Pöysä 2006). defence (i.e. at tended nests, hosts may Remnants of successful hatching of a clutch prevent access whereas this would not be (see Fig. 2) thus seem to be important cues the case at untended nests). This highlights by which parasitic Common Goldeneye the difficulty of inferring cue use from females select target nests. A critical patterns of nest use and only experimental prerequisite in this hypothesis is that nest studies are likely to tease these apart. At any success is predictable between successive rate, patterns of parasitic egg-laying in breeding seasons, as found for Common Wood Duck females did not correspond to Goldeneyes in which nest depredation is the patterns of nest success. The authors main determinant of variation in breeding suggested that high nest density may have success (Pöysä 2006). Predictability of confounded the quality of information and nest success, coupled with the ability of caused parasites to make poor decisions (see parasites to assess it and lay accordingly, Roy et al. 2009). make parasitic laying an advantageous The nest success hypothesis also has been evolutionary strategy (Pöysä & Pesonen tested for Common Eider, and results 2007). suggest that parasitic females did not use Roy et al. (2009) tested the nest success nest-site safety as a cue for egg laying hypothesis for Wood Ducks and found, (Lusignan et al. 2010; Table 3). Specifically, contrary to the prediction, that previously Lusignan et al. (2010) found that nests in unsuccessful nest sites were more likely to dense woody vegetation had the highest be parasitized in the following year. These probability of survival but the lowest authors also found that previous success did frequency of CBP. On the other hand, nests not consistently predict future success. An in highly visible artificial wooden nest important difference between this study and shelters had the highest rate of parasitism Pöysä’s studies is that in the Wood Duck the and ranked second in terms of nest survival. main cause of failing was nest desertion, This finding suggests that nest visibility had probably caused by a high rate of parasitic a greater effect on parasitism rate than nest laying (Roy et al. 2009). If females are simply site safety (see Lusignan et al. 2010). evaluating cues related to nest predation (e.g. Another way that parasitically laying broken eggs) a nest with a large number of Common Eider females could use cues to deserted eggs may still indicate a safe nest choose high quality nests was suggested by with respect to predation risk, and be Ruxton (1999). He was inspired by the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 207 ther 2009 et al . et al. Source non-depredated nests in nests ofdepredation when laying they which previous experience ofdid not have success nest depredation risk depredation stopped eggclutch in the laying that had not nest, whereas parasites 2010 experienced it laid in the nest later season successful nests unsuccessful nests frequency oflowest parasitism (nest-site visibility more influential) 2010 cue use (simulation (simulation before egg laying) (simulation during egg laying) Empirical studies that have tested for evidence ofEmpirical studies that have the use of cues in the selection of females. parasitically laying nests by Species of Type cue of Type study Support for Main finding and comments comments according to original articles. Type ofType Interpretation of together cue and study (observational with main findings. or experimental) are given, results and fur Table 3. Table Common Goldeneye Nest predation Common Experimental risk Goldeneye Nest predation Experimental risk YesCommon Noin previously more frequent Parasitism Goldeneye Nest predation Common Experimental risk did not respond to simulated Parasites Goldeneye Nest predation 1999 Pöysä Experimental risk Yes 2003a Pöysä Common Yes/No with low on lakes more frequent Parasitism Goldeneye partial that experienced simulated Parasites Nest success DuckWood 2003a Pöysä Nest success Pöysä ObservationalCommon ObservationalEider Yes Nest-site Yes/No more frequent in previously Parasitism more frequent in previously Parasitism safety/visibility Observational 2006 Pöysä Roy Nohighest probability of Nests with survival had Lusignan et al .

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 208 Conspecific brood parasitism in waterfowl et al . Source 2011 number ofnumber previous nesting attempts and occupation rate 2010 used and parasitized equally parasitized as real nests with a host nesting experience (and less clearly with host body condition) but, in general, host traits less 2010 important than nest site traits success size (hatch with smaller clutch probably higher in smaller clutches) nests than did random eggs placedtheir host’s 2010 in random nests Sedinger 2011 as non-parasitized females cue use (number of(number nesting attempts, occupation rate) (nesting experience, condition) (condition) ( continued ). Species of Type cue of Type study Support for Main finding and comments Common and Barrow’s Host presenceGoldeneye ExperimentalCommon Goldeneye Host presenceCommon No ExperimentalGoldeneye Host traits a host were nests that did not have Dummy No Eadie 1989 Observational DuckWood size Clutch a host were nests that did not have Dummy Yes/No Probability of parasitism increased with host 2003b Pöysä BrantBlack Experimental et al . Paasivaara Egg size YesCommon Observational nests more frequent in simulated Parasitism Host trait Odell & Eadie Yes Observational eggs Parasitic in size from eggs differed less in & Lemons No of females were Parasitized same body mass Waldeck Table 3 Table Common Goldeneye Nest site traits Observational Yes Probability of parasitism increased with the et al . Paasivaara Eider

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 209 observation that predation rates were lower (Pöysä 2003a; Pöysä et al. 2010). Those for parasitized Common Eider nests in experiments revealed that parasitically one population (Robertson 1998). This laying females respond to varying degrees finding, coupled with observations that nest of nest depredation risk (i.e. they prefer depredation rates were correlated with laying in simulated nests that are in female attentiveness (Swennen et al. 1993), safe environments) but their response to led Ruxton (1999) to suggest that simulated nest depredation varied parasitically laying females could use depending on whether females experienced variation in the female attentiveness at the simulated partial clutch depredation (Table nest during the egg-laying period as a cue of 3). These experimental findings suggest, nest depredation risk and select target nests first, that nest depredation and nest accordingly. The hypothesis predicts that depredation risk are important cues, and there should be a positive relationship second, that both personal information and between an individual female’s attentiveness social information are used in the selection at the nest during early egg-laying and her of target nests by parasitically laying risk of parasitism (Ruxton 1999). To our Common Goldeneye females. Other knowledge this prediction has not been experimental studies have found that host tested. It is important to note that nest presence was not an important cue in nest attentiveness could also indicate female selection by parasitically laying Common quality, making it difficult to distinguish Goldeneye (Pöysä 2003b) and Barrow’s between female quality and nest site quality Goldeneye females (Eadie 1989). because of the correlation between these Parasitic Wood Duck females appear to two variables. Thus although it may be respond to variation in the number of eggs possible to correlate apparent cues (e.g. in a nest (Odell & Eadie 2010). In this study pattern of female nest attentiveness) with a choice of nests containing clutches of 5, parasitic behaviour, we need to be careful to 10, 15 or 20 experimental eggs was offered consider the underlying information that to Wood Duck females, and the number of these patterns might represent (i.e. female eggs laid in the simulated nests declined quality). The “cues” that we measure may in direct relation to the number of not necessarily be the cues that parasites experimental eggs in the nest. This finding is perceive or respond to. of particular interest because it suggests the There are only a few studies in which cues possibility that parasitically laying females affecting nest site selection and the laying are able to assess the number of eggs in a decisions of parasites have been addressed nest, a cue associated with important fitness experimentally (Table 3). In addition to consequences because large clutches often Pöysä (1999, see above), two experimental have low hatching success (Roy Nielsen et al. studies have addressed the role of nest 2006b; Odell & Eadie 2010). Lemons and depredation risk and actual (partial) nest Sedinger (2011) report a remarkable pattern depredation in affecting the laying decision for the Black Brant Branta bernicla nigricans in of parasites in the Common Goldeneye which parasitic eggs match the size of host

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 210 Conspecific brood parasitism in waterfowl eggs, suggesting that parasitically laying this remains an interesting and important females recognise host egg size and lay direction for future work. accordingly, probably to improve hatching success. How they might do so is completely Timing of host laying cycle unknown. Synchronising the timing of egg-laying with the host’s laying cycle is thought to be Territory or host quality important for interspecific brood parasites The importance of host female quality (i.e. (Davies 2000) and for conspecific brood body condition) has been addressed in parasites as well. A role for host cues that Common Goldeneyes and Common Eiders, reveal the timing of their cycle (but which and appeared not to be an important cue for have yet to be confirmed) has been found to parasitically laying females (Paasivaara et al. be important in the few studies of non- 2010; Waldeck et al. 2011). It should be waterfowl species that have examined noted that nesting schedule, a feature that these patterns while also taking random also may reflect host female quality, has been expectations into account (Emlen & Wrege found to be associated with the occurrence 1986; Lyon & Everding 1996). Brown and of CBP in many waterfowl species (i.e. Brown (1988) found that parasitic Cliff clutches laid early in the season are more Swallows, which parasitize hosts by frequently parasitized than late clutches: transferring eggs physically in their beaks, Dow & Fredga 1984; Sorenson 1991; were remarkably good at synchronising timing Robertson et al. 1992); however, other with the host’s laying and incubation period. factors were not controlled in these studies Matching the timing of egg-laying to a (see Paasivaara et al. 2010). Older females host’s own clutch is particularly important in often nest earlier and if CBP occurs more precocial birds such as waterfowl because frequently early in the season, then a pattern the young hatch synchronously and leave would emerge of older (and perhaps more the nest simultaneously within 24–48 h after experienced or higher quality) females being hatching. Mismatched timing of egg-laying parasitized disproportionately. The causative by the parasite can result in eggs failing to arrow could however be in the opposite hatch, or young hatching after the host direction: parasites might target older female has already left with her brood (see experienced females and, if older females Bellrose & Holm 1994; review in Sayler nest earlier, then CBP would be more 1992). Nonetheless, parasitic eggs are laid frequent in early nests. This would require after the onset of incubation in several careful experiments or statistical controls species (Jones & Leopold 1967; Clawson et to decouple this pattern. Monte Carlo al. 1979; Heusmann et al. 1980; Eriksson & randomisation analyses could be used to Andersson 1982; Eadie 1989; Bellrose & determine if parasites selected non- Holm 1994; Št’ovícˇek et al. 2013; review in randomly from among the host nests Sayler 1992). Sayler (1992) describes a case available, as noted above. Very few such of interspecific brood parasitism in which tests have been conducted for waterfowl and a parasitic Redhead female laid in a

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 211

Canvasback Aythya valisineria nest while it females may have no potential hosts in the contained hatched ducklings. In contrast, laying stage to parasitize when they have a Wood Ducks in some populations laid up to parasitic egg ready for laying. A similar 80% of parasitic eggs prior to host explanation may account for the fact that incubation (Clawson et al. 1979). parasitic females will often lay eggs in Matching the timing of egg-laying with deserted nests, occasionally leading to large hosts has been documented in several non- accumulations of abandoned eggs (termed waterfowl species, but more research would “dump nests” in the older literature). Sayler be useful to explore the cues and (1992) noted that parasitic Redhead females mechanisms used by parasites to fine-tune will often follow each other to nests and lay the timing of parasitic egg-laying and to a series of eggs in those nests over several understand better the constraints of doing days. Similar behaviour has been observed in so. To add further complexity, exciting new Wood Ducks (Semel & Sherman 2001) and work by Hepp and colleagues has shown Common Goldeneyes (Eadie 1989; Åhlund that even slight differences in incubation 2005). Sayler (1992) suggested that these temperature can have significant impacts on nests appear active to parasitic females given the post-natal development and survival of the presence of other females and eggs young (Hepp et al. 1990, 2006; Kennamer et being laid in the nest, even though the host al. 1990; DuRant et al. 2010, 2011, 2012a,b). has already abandoned the nest. Thus, timing of egg-laying and incubation A curious (and opposite) pattern has been efficiency could have a large impact on documented in Common Eiders. In several parasite (and host) fitness. Odell (2008) populations, researchers have found that found that eggs of parasitic Wood Ducks in parasitic eggs are often laid in nests before the California had higher levels of androgens host female begins to lay her own eggs. The than host eggs and this might accelerate the host female thus lays her eggs in a nest in development of parasitic eggs laid at the end which another egg is already present, and of the host laying period or after the then subsequently completes her clutch and initiation of incubation. Typically, host incubates the nest. Common Eiders often females spend large portions of the day on reuse nest bowls in successive years and this the nest during incubation, and yet parasitic pattern could be explained if females are females do not appear to use host presence simply competing for certain nest sites, with as a cue to avoid these nests. Possibly, the first female being ousted (the accidental parasites cannot detect accurately the stage parasite). However, Robertson (1998) of incubation and simply the presence of suggested that this phenomenon was due to another female or evidence of an active nest takeover and adoption, possibly in nest provides a sufficient incentive to response to nest predation risk. He reported induce egg-laying. Alternatively, cases of that in nests where females took over nests, mismatched timing of laying with respect to predation on the first eggs was lower than in the host’s laying cycle might be influenced comparable nests with only a single female by host availability, for instance some (i.e. no takeover). Robertson (1998) argued

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 212 Conspecific brood parasitism in waterfowl that the presence of an egg in a nest would role of hosts in determining the patterns or indicate that the site was a safe nest location outcomes of parasitic interactions, are not (since the egg had not been depredated), important. Indeed, there are a number of and therefore the nest may be attractive to decisions hosts might make to influence the another female. The benefit of obtaining a occurrence of parasitism, each involving safe nest site could outweigh the potential different cues. An analysis of host decisions cost of caring for the additional eggs. is beyond the scope of our paper, and Ruxton (1999) further suggested that limited information is available. However, it females can detect variation in the predation will be important in future studies to risk associated with different nests, and use integrate cue use and decision-making by this information to target nests with low hosts for several reasons: 1) it links predation risk as sites for laying parasitically, cognition and decision-making to parasitism a hypothesis similar to Pöysä’s (1999, 2003a, broadly; 2) decisions are linked in a game 2006; Pöysä et al. 2010) hypothesis to explain theoretic way that include both parasite and CBP in Common Goldeneyes. Waldeck host responses (e.g. Andersson & Eriksson’s and Andersson (2006), using protein 1982 clutch size model); and 3) for some fingerprinting techniques, found that decisions (e.g. who lays in a nest), it can be another female laid before the host started very difficult to determine whether the host laying in 41% of mixed clutches. Similarly, or the parasite determines the outcome (and Hario et al. (2012) found that 58% of hence who is using what cues). parasitic eggs in a population in Finland This last question is particularly germane were the first or second eggs laid. Waldeck in our efforts to understanding the relative and Andersson (2006) reported that nests roles that hosts and parasites play in that were taken over have higher early determining the patterns of parasitism. survival than other nests, consistent with the While some authors emphasise the active hypothesis that CBP in Common Eiders is role of hosts (e.g. Andersson & Åhlund driven by selection for safe nest sites. It is 2000), others suggest that hosts do not play still unclear whether “nest takeover” is a an important role (e.g. Pöysä 2004). form of CBP at all, although it is often Evidence that host-mediated facilitation presented in that context (see discussion in does not play a central role in the laying Roberston 1998). Clearly, much more work decisions of parasites comes from remains to better understand the cues that experiments in which parasitic laying has parasitic females use to fine-tune the timing been induced in simulated nests that do not of egg-laying. have a host present (Pöysä 2003a,b; Odell & Eadie 2010; Pöysä et al. 2010). It is What determines the patterns of noteworthy that laying in simulated nests is parasitism? frequent in Common Goldeneyes even We have, to this point, focused on the cues when active real nests are available (H. that parasites use. This is not intended to Pöysä, unpubl. data), indicating that this imply that cues that hosts might use, and the behaviour is not simply due to a shortage

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 213 of host nests. Pöysä (2003b) specifically costs to hosts are high, then parasites should addressed the importance of host avoid laying eggs in a relative’s nest so as not recognition by parasites in the selection of to reduce the host’s fitness, and thereby nests and found it not to be important. lower the parasite’s inclusive fitness. Testing These experiments suggest that host the kin selection hypothesis requires detailed females do not play an active role as information not only on host-parasite facilitators of parasitism. However, the role relatedness, but also on the costs to the of hosts in CBP remains uncertain. One hosts, the degree (or existence) of kin issue that has recently been of growing recognition and the extent to which interest, and is particularly relevant to the parasites versus hosts control or facilitate question of the role of hosts in facilitating CBP. CBP, is the potential influence of kinship A number of empirical studies using amongst hosts and parasites. molecular genetic techniques (DNA microsatellites and isoelectric focusing A role for kinship? of egg albumin proteins) have now The idea that hosts and parasites might be documented high host-parasite relatedness related, and hence that CBP is not a form of in waterfowl, including the Wood Duck, parasitism per se, but rather a cooperative Common Eider, Barrow’s Goldeneye and behaviour facilitated by kin selection, was Common Goldeneye (review in Eadie & suggested over 30 years ago by Andersson Lyon 2011). However, the mechanisms (1984). In waterfowl, females are the leading to high host-parasite relatedness philopatric sex so the premise of the kinship remain unknown, although kin recognition hypothesis is that females might return to and discrimination against unrelated their natal area and lay eggs in the nests of parasites by hosts have been suggested in the close kin. A central feature of the Common Goldeneye (Andersson & Åhlund mechanism is that hosts are in the driver’s 2000; but see Pöysä 2004). The finding that seat. By allowing kin to lay eggs in the host hosts and parasites are often related opens nest, the host may be facilitating the possibility that parasites or hosts could reproduction by a relative where otherwise recognise kin and that kinship could provide none would have been possible, thereby a cue in nest/host selection. Interestingly, increasing the hosts’ own inclusive fitness. Jaatinen et al. (2011b) found that the The idea was often cited but rarely tested response of parasitic Barrow’s Goldeneye until a number of new theoretical models females to relatedness depends on their revisited this idea (Zink 2000; Andersson nesting status: parasitic females that had a 2001; Lopez-Sepulcre & Kokko 2002; nest of their own (“nesting parasites”) Jaatinen et al. 2011a). The current consensus responded to relatedness by laying more eggs of these models is that kinship can facilitate with increasing relatedness to the host, while CBP, provided that costs to the host are low non-nesting parasites did not respond to and some degree of kin recognition exists relatedness. The authors discuss several (Lyon & Eadie 2000; Eadie & Lyon 2011). If possible reasons why nesting and non-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 214 Conspecific brood parasitism in waterfowl nesting parasites differed in their response to readily integrated into this new framework. relatedness but the mechanisms underlying We suggest that this new framework will this finding are currently unknown. Finally, provide a useful direction and impetus for Pöysä et al. (2014) provided experimental the next generation of studies of CBP, evidence that nest predation risk and fuelled by an increasing battery of molecular interaction between related parasites are genetic techniques and a growing array of associated with kin-biased co-parasitism technological tools to track females and (related parasites non-randomly laying in the their reproductive trajectories throughout same nest) in the Common Goldeneye. In their lifespan. reviewing the evidence to date, the initial We have also focused on an emerging, results suggest that kinship could play a role exciting area for future investigations of as a cue at least for parasites that have a CBP – namely the use of cues and nest of their own, but it remains to be information by conspecific brood parasites determined whether kinship is a central as they seek suitable places to lay their driver in the evolution of CBP, or instead parasitic eggs. Recent empirical studies of just one of several factors that influence the CBP in waterfowl have revealed remarkable frequency and occurrence of this intriguing cognitive abilities in parasitic females, behaviour (Eadie & Lyon 2011). although the actual mechanisms remain unknown in most cases. In particular, the Conclusions and future issues use of public information by parasites in Our understanding of CBP has advanced locating and selecting nests that have high considerably in the 35 years since Yom-Tov prospects of success is a promising avenue (1980) first brought it to our attention, and worth exploring to gain insight into the research on waterfowl (Anseriformes) has evolution of nest/host selection and egg contributed disproportionately to this laying decisions of parasites. Interestingly, knowledge. We now have a much better the importance of public information has understanding of the ecological and social also been stressed recently in the context of conditions under which CBP occurs, and interspecific brood parasitism; Parejo and over two dozen studies of waterfowl have Avilés (2007) suggested that parasites might tested the existing set of hypotheses eavesdrop on the sexual signals of their proposed to account for this behaviour. hosts to find high quality foster parents for Collectively, these studies support the their own offspring. The role of parental hypotheses that CBP allows some females to quality as a cue in CBP has received little reproduce when they otherwise could not, support in waterfowl but more research on while other females use parasitic egg-laying this aspect is needed. The ability of parasites as a way to enhance total fecundity. A life- to evaluate the number of eggs in a nest and history approach offers a new framework by to modify their own laying behaviour which to integrate all of these possibilities accordingly, as demonstrated with Wood into a theory of flexible life-history, and the Ducks, is intriguing and worth further set of traditional hypotheses for CBP can be exploration in other species.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 215

The role that kinship plays as a cue in the brood-parasitic duck. Animal Behaviour 70: laying decisions of parasites (and hosts) 433–440. remains a challenging task for future studies. Åhlund, M. & Andersson, M. 2001. Female ducks High host-parasite relatedness has been can double their reproduction. Nature 414: reported for several species, but it is unclear 600–601. Anderholm, S., Waldeck, P., van der Jeugd, H.P., whether direct assessment of relatedness is Marshall, R.C., Larsson, K. & Andersson, M. involved or if some unmeasured correlate 2009. Colony kin structure and host-parasite leads related females to select the same nest relatedness in the Barnacle Goose. Molecular site. For example, as suggested by Pöysä Ecology 18: 4955–4963. (2004), high natal and nest site philopatry Andersson, M. 1984. Brood parasitism within and preference of both hosts and parasites species. In C.J. Barnard (ed.), Producers and to lay in safe nest sites will also generate high Scroungers. Strategies of Exploitation and host-parasite relatedness. Experimental Parasitism, pp. 195–228. Chapman & Hall, studies and examination of the cost of New York, USA. parasitism to hosts and the ability of females Andersson, M. 2001. Relatedness and the to recognise or interact differentially with evolution of conspecific brood parasitism. kin are required to disentangle these effects. American Naturalist 158: 599–614. Andersson, M. & Åhlund, M. 2000. Host-parasite Acknowledgements relatedness shown by protein fingerprinting in a brood parasitic bird. Proceedings of the We thank the scientific committee of the National Academy of Sciences, USA 97: Ecology and Conservation of North 13188–13193. American Waterfowl symposium for the Andersson, M. & Åhlund, M. 2012. Don’t put opportunity to give the plenary talk from all your eggs in one nest: spread them and which this paper is extracted. Gary Hepp cut time at risk. American Naturalist 180: 354– served as the lead editor of this manuscript 363. and we appreciate his enduring patience and Andersson, M. & Eriksson, M.O.G. 1982. Nest useful comments on the manuscript. We parasitism in goldeneye Bucephala clangula: also thank Rick Kaminski and Eileen Rees some evolutionary aspects. American Naturalist for useful editorial comments. Support for 120: 1–16. the development of the ideas and research Bellrose F.C. & Holm D.J. 1994. Ecology and Management of the Wood Duck. Stackpole reported in this paper was provided by Books, Mechanicsburg, Pennsylvania, USA. funding from Finnish Game and Fisheries Briggs, S.V. 1991. Intraspecific nest parasitism in Research Institute (to HP), NSF (IOS- Maned Ducks Chenonetta jubata. Emu 91: 1355208 to JME; IOS- 1354894 to BEL), 230–35. and the Dennis G. Raveling Waterfowl Brown, C.R. & Brown, M.B. 1988. A new form of Professorship at U.C. Davis (to JME). reproductive parasitism in Cliff Swallows. Nature 331: 66–68. References Brown, C.R. & Brown, M.B. 1991. Selection of Åhlund, M. 2005. Behavioural tactics at nest high-quality hosts by parasitic Cliff Swallows. visits differ between parasites and hosts in a Animal Behaviour 41: 457–465.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 216 Conspecific brood parasitism in waterfowl

Clawson, E.L., Hartman, G.W. & Fredrickson, waterfowl. In B.D. Bell, R.O. Cossee, J.E.C. L.H. 1979. Dump nesting in a Missouri Flux, B.D. Heather, R.A. Hitchmough, C.J.R. Wood Duck population. Journal of Wildlife Robertson & M.J. Williams (eds.), Proceedings of Management 43: 347–355. the Twentieth International Ornithological Conference, Danchin, É., Giraldeau, L.-A. & Wagner, R.H. pp. 1031–1040. Hutcheson, Bowman & 2008. An information-driven approach to Stewart Ltd., Wellington, New Zealand. behaviour. In É. Danchin, L.-A. Giraldeau & Eadie, J.M. & Gauthier, G. 1985. Prospecting for F. Cézilly (eds.), Behavioural Ecology, pp. 97– nest sites by cavity-nesting ducks of the 130. Oxford University Press, Oxford, UK. genus Bucephala. Condor 87: 528–534. Davies, N.B. 2000. Cuckoos, Cowbirds and Other Eadie, J.M. & Lyon, B. 2011. The relative role of Cheats. T. & A.D. Poyser, London, UK. relatives in conspecific brood parasitism. Dow, H. & Fredga, S. 1984. Factors affecting Molecular Ecology 20: 5114–5118. reproductive output of the goldeneye duck Eadie, J.M., Kehoe, F.P. & Nudds, T.D. 1988. Pre- Bucephala clangula. Journal of Animal Ecology 53: hatch and post-hatch brood amalgamation in 679–692. North American Anatidae: a review. Canadian DuRant S.E., Hepp G.R., Moore I.T., Hopkins Journal of Zoology 66: 1709–1721. B.C. & Hopkins W.A. 2010. Slight differences Eadie, J.M, Sherman, P. & Semel, B. 1998. in incubation temperature affect early growth Conspecific brood parasitism, population and stress endocrinology of wood duck dynamics, and the conservation of cavity- (Aix sponsa) ducklings. Journal of Experimental nesting birds. In T. Caro (ed.), Behavioral Biology 213: 45–51. Ecology and Conservation Biology, pp. 306–340. DuRant, S.E., Hopkins, W.A. & Hepp, G.R. 2011. Oxford University Press, Oxford, UK. Embryonic developmental patterns and Emlen, S.T. & Wrege, P.H. 1986. Forced energy expenditure are affected by incubation copulation and intra-specific parasitism: two temperature in Wood Ducks (Aix sponsa). costs of social living in the White-fronted Physiological and Biochemical Zoology 84: 451–457. Bee-eater. Ethology 71: 2–29. DuRant, S.E., Hopkins, W.A., Hawley, Eriksson, M.O.G. & Andersson, M. 1982. Nest D.M. & Hepp, G.R. 2012a. Incubation parasitism and hatching success in a temperature affects multiple measures of population of goldeneyes Bucephala clangula. immunocompetence in young Wood Ducks Bird Study 29: 49–54. (Aix sponsa). Biology Letters 8: 108–111. Erskine, A.J. 1990. Joint laying in Bucephala ducks DuRant, S.E., Hopkins, W.A., Wilson, A. – “parasitism” or nest-site competition? & Hepp, G.R. 2012b. Incubation Ornis Scandinavica 21: 52–56. temperature affects the metabolic cost of Fast, P.L.F., Gilchrist, H.G. & Clark, R.G. 2010. thermoregulation in a young precocial bird. Nest-site materials affect nest-bowl use by Functional Ecology 26: 416–422. Common Eiders (Somateria mollissima). Eadie, J.M. 1989. Alternative reproductive tactics Canadian Journal of Zoology 88: 214–218. in a precocial bird: the ecology and evolution Forslund, P. & Larsson, K. 1995. Intraspecific of brood parasitism in goldeneyes. Ph.D. nest parasitism in the Barnacle Goose: thesis, University of British Columbia, behavioural tactics of parasites and hosts. Vancouver, Canada. Animal Behaviour 50: 509–517. Eadie, J.M. 1991. Constraint and opportunity Hario, M., Koljonen, M.-L. & Rintala, J. 2012. in the evolution of brood parasitism in Kin structure and choice of brood care in a

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 217

Common Eider (Somateria m. mollissima) Lank, D.B., Cooch, E.G., Rockwell, R.F. & population. Journal of Ornithology 153: Cooke, F. 1989. Environmental and 963–973. demographic correlates of intraspecific nest Hepp, G.R., Kennamer, R.A. & Harvey W.F. parasitism in Lesser Snow Geese Chen 1990. Incubation as a reproductive cost in caerulescens caerulescens. Journal of Animal female Wood Ducks. Auk 107: 756–764. Ecology 58: 29–45. Hepp, G.R., Kennamer, R.A. & Johnson M.H. Lemons, P.R. & Sedinger, J.S. 2011. Egg size 2006. Maternal effects in Wood Ducks: matching by an intraspecific brood parasite. incubation temperature influences incubation Behavioral Ecology 22: 696–700. period and neonate phenotype. Functional Lopez-Sepulcre, A. & Kokko, H. 2002. The Ecology 20: 308–314. role of kin recognition in the evolution Heusmann, H.W., Bellville, R. & Burrell, R.G. of conspecific brood parasitism. Animal 1980. Further observations on dump nesting Behaviour 64:215–222. by Wood Ducks. Journal of Wildlife Lusignan, A.P., Mehl, K.R., Jones, I.L. & Management 44: 908–915. Cloutney, M.L. 2010. Conspecific brood Jaatinen, K., Öst, M., Waldeck, P. & Andersson, parasitism in Common Eiders (Somateria M. 2009. Clutch desertion in Barrow’s mollissima): do brood parasites target safe nest Goldeneyes (Bucephala islandica) – effects on sites? Auk 127: 765–772. non-natal eggs, the environment and host Lyon, B.E. 1993. Tactics of parasitic American female characteristics. Annales Zoologici Fennici coots: host choice and the pattern of egg 46: 350–360. dispersion among host nests. Behavioral Jaatinen, K., Lehtonen, J. & Kokko, H. 2011a. Ecology and Sociobiology 33: 87–100. Strategy selection under conspecific brood Lyon, B.E. & Eadie, J.M. 2000. Family matters: kin parasitism: an integrative modeling approach. selection and the evolution of conspecific Behavioral Ecology 22: 144–155. brood parasitism. Proceedings of the National Jaatinen, K., Öst, M., Gienapp, P. & Merilä, J. Academy of Sciences USA 97: 1942–1944. 2011b. Differential responses to related hosts Lyon, B.E. & Eadie, J.M. 2008. Conspecific by nesting and non-nesting parasites in a brood parasitism in birds: a life-history brood-parasitic duck. Molecular Ecology 20: perspective. Annual Review of Ecology, 5328–5336. Evolution and Systematics 39: 343–363. Jansen, R.W. & Bollinger, E.K. 1998. Effects of Lyon, B.E. & Everding, S. 1996. A high nest-box visibility and clustering on Wood frequency of conspecific brood parasitism in Duck brood parasitism in Illinois. Transactions a colonial waterbird, the Eared Grebe of the Illinois State Academy of Science 91: (Podiceps nigricollis). Journal of Avian Biology 27: 161–166. 238–244. Jones, R.E. & Leopold, A.S. 1967. Nesting Maynard Smith, J. & Harper, D. 2003. Animal interference in a dense population of Wood Signals. Oxford University Press, Oxford, Ducks. Journal of Wildlife Management 31: UK. 221–228. McRae, S.B. & Burke, T. 1996. Intraspecific Kennamer, R.A., Harvey, W.F. & Hepp, G.R. brood parasitism in the Moorhen: parentage 1990. Embryonic development and nest and parasite-host relationships determined attentiveness of Wood Ducks during egg- by DNA fingerprinting. Behavioral Ecology and laying. Condor 92: 587–592. Sociobiology 38: 115–129.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 218 Conspecific brood parasitism in waterfowl

Odell, N.S. 2008. Female reproductive parasitism: from risk spreading to risk investment in a conspecific brood parasite. assessment. American Naturalist 169: 94–104. Ph.D. thesis. University of California, Davis, Pöysä, H., Milonoff, M., Ruusila, V. & Virtanen, USA. J. 1999. Nest-site selection in relation to Odell, N.S. & Eadie, J.M. 2010. Do Wood Ducks habitat edge: experiments in the Common use the quantity of eggs in a nest as a cue to Goldeneye. Journal of Avian Biology 30: 79–84. the nest’s value? Behavioral Ecology 21: 794– Pöysä, H., Lindblom, K., Rutila, J. & Sorjonen, J. 801. 2010. Response of parasitically laying Paasivaara, A., Rutila, J., Pöysä, H. & Runko, P. goldeneyes to experimental nest predation. 2010. Do parasitic Common Goldeneye Animal Behaviour 80: 881–886. Bucephala clangula females choose nests on the Pöysä, H., Paasivaara, A., Lindblom, K., Rutila, J. & basis of host traits or nest site traits? Journal Sorjonen, J. 2014. Co-parasites preferentially of Avian Biology 41: 662–671. lay with kin and in safe neighbourhoods: Parejo, D. & Avilés, J.M. 2007. Do avian brood experimental evidence from goldeneye ducks. parasites eavesdrop on heterospecific sexual Animal Behaviour 91: 111–118. signals revealing host quality? A review of Reichart, L.M., Anderholm, S., Munoz-Fuentes, the evidence. Animal Cognition 10: 81–88. V. & Webster, M.S. 2010. Molecular Patten, M.A., Reinking, D.L. & Wolfe, D.H. 2011. identification of brood-parasitic females Hierarchical cues in brood parasite nest reveals an opportunistic reproductive tactic in selection. Journal of Ornithology 152: 521–532. Ruddy Ducks. Molecular Ecology 19: 401–413. Pöysä, H. 1999. Conspecific nest parasitism is Robertson, G.J. 1998. Egg adoption can explain associated with inequality in nest predation joint egg-laying in Common Eiders. risk in the Common Goldeneye (Bucephala Behavioral Ecology and Sociobiology 43: 289–296. clangula). Behavioral Ecology 10: 533–540. Robertson, G.J., Watson, M.D. & Cooke, F. 1992. Pöysä, H. 2003a. Parasitic Common Goldeneye Frequency, timing and costs of intraspecific (Bucephala clangula) females lay preferentially nest parasitism in the Common Eider. Condor in safe neighbourhoods. Behavioral Ecology and 94: 871–879. Sociobiology 54: 30–35. Rohwer, F.C. & Freeman, S. 1989. The Pöysä, H. 2003b. Low host recognition tendency distribution of conspecific nest parasitism in revealed by experimentally induced parasitic birds. Canadian Journal of Zoology 67: 239–253. egg laying in the Common Goldeneye Roy, C., Eadie, J.M., Schauber, E.M., Odell, N.S., (Bucephala clangula). Canadian Journal of Zoology Berg, E.C. & Moore, T. 2009. Public 81: 1561–1565. information and conspecific nest parasitism Pöysä, H. 2004. Relatedness and the evolution of in Wood Ducks: does nest density influence conspecific brood parasitism: parameterizing quality of information? Animal Behaviour 77: a model with data for a precocial species. 1367–1373. Animal Behaviour 67: 673–679. Roy Nielsen, C.L., Gates, R.J. & Parker, P.G. Pöysä, H. 2006. Public information and 2006a. Intraspecific nest parasitism of Wood conspecific nest parasitism in goldeneyes: Ducks in natural cavities: comparisons with targeting safe nests by parasites. Behavioral nest boxes. Journal of Wildlife Management 70: Ecology 17: 459–465. 835–843. Pöysä, H. & Pesonen, M. 2007. Nest predation Roy Nielsen, C.R., Parker, P.G. and Gates, R.J. and the evolution of conspecific brood 2006b. Intraspecific nest parasitism of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 Conspecific brood parasitism in waterfowl 219

cavity-nesting wood ducks: costs and Common Eiders. Ornis Scandinavica 24: 48– benefits to hosts and parasites. Animal 52. Behaviour 72: 917–926. Wagner, R.H. & Danchin, É. 2010. A taxonomy of Ruxton, G.D. 1999. Are attentive mothers biological information. Oikos 119: 203–209. preferentially parasitised? Behavioral Ecology Waldeck, P. & Andersson, M. 2006. Brood and Sociobiology 46: 71–72. parasitism and nest takeover in Common Sayler, R.D. 1992. Ecology and evolution of Eiders. Ethology 112: 616–624. brood parasitism in waterfowl. In B.D.J. Batt, Waldeck, P., Kilpi, M., Öst, M. & Andersson, A.D. Afton, M.G. Anderson, C.D. Ankney, M. 2004. Brood parasitism in a population D.H. Johnson, J.A. Kadlec & G.L. Krapu of Common Eider (Somateria mollissima). (eds.), Ecology and Management of Breeding Behaviour 141: 725–739. Waterfowl, pp. 290–322. University of Waldeck, P., Andersson, M., Kilpi, M. & Öst, M. Minnesota Press, Minneapolis, USA. 2008. Spatial relatedness and brood Schamel, D. 1977. Breeding of the Common parasitism in a female-philopatric bird Eider (Somateria mollissima) on the Beaufort population. Behavioral Ecology 19: 67–73. Sea coast of Alaska. Condor 79: 478–485. Waldeck, P., Hagen, J.I., Hanssen, S.A. & Semel, B. & Sherman, P.W. 1986. Dynamics of Andersson, M. 2011. Brood parasitism, nest parasitism in Wood Ducks. Auk 103: female condition and clutch reduction in the 813–816. Common Eider Somateria mollissima. Journal of Semel, B. & Sherman, P.W. 2001. Intraspecific Avian Biology 42: 231–238. parasitism and nest-site competition in Wood Weigmann, C. & Lamprecht, J. 1991. Intraspecific Ducks. Animal Behaviour 61:787–803. nest parasitism in Bar-headed Geese, Anser Semel, B., Sherman, P.W. & Byers, S.M. 1988. indicus. Animal Behaviour 41: 677–688. Effects of brood parasitism and nest-box Weller, M.W. 1959. Parasitic egg laying in the placement on Wood Duck breeding ecology. Redhead (Aythya americana) and other North Condor 90: 920–930. American Anatidae. Ecological Monographs 29: Sorenson, M.D. 1991. The functional significance 333–365. of parasitic egg laying and typical nesting in Wilson, S.F. 1993. Use of Wood Duck decoys in Redhead Ducks: an analysis of individual a study of brood parasitism. Journal of Field behaviour. Animal Behaviour 42: 771–796. Ornithology 63: 337–340. Sorenson, M.D. 1993. Parasitic egg-laying Yom-Tov, Y. 1980. Intraspecific nest parasitism in in Canvasbacks – frequency, success and birds. Biological Reviews 55: 93–108. individual behavior. Auk 110: 57–69. Yom-Tov, Y. 2001. An updated list and some Št’ovícˇek, O., Kreisinger, J., Javu˚rková, V. & comments on the occurrence of intraspecific Albrecht, T. 2013. High rates of conspecific nest parasitism in birds. Ibis 143: 133–143. brood parasitism revealed by microsatellite Zicus, M.C. & Hennes, S.K. 1989. Nest analysis in a diving duck, the Common prospecting by Common Goldeneyes. Condor Pochard Aythya ferina. Journal of Avian Biology 91: 807–812. 44: 369–375. Zink, A.G. 2000. The evolution of intraspecific Swennen, C., Ursem, J.C.H. & Duiven, P. 1993. brood parasitism in birds and insects. Determinate laying and egg attendence in American Naturalist 155: 395–405.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 192–219 220

The effects of harvest on waterfowl populations

EVAN G. COOCH1*, MATTHIEU GUILLEMAIN2, G. SCOTT BOOMER3, JEAN-DOMINIQUE LEBRETON4 & JAMES D. NICHOLS5

1Department of Natural Resources, Cornell University, Ithaca, New York 14853, USA. 2Office National de la Chasse et de la Faune Sauvage, CNERA Avifaune Migratrice, La Tour du Valat, Le Sambuc, 13200 Arles, France. 3Division of Migratory Bird Management, U.S. Fish & Wildlife Service, Laurel, Maryland 20708, USA. 4Centre d’Écologie Fonctionnelle et Évolutive, UMR 5175, CNRS, 34293 Montpellier Cedex 5, France. 5Patuxent Wildlife Research Center, U.S. Geological Survey, Laurel, Maryland 20708, USA. *Correspondence author. E-mail: [email protected]

Abstract Change in the size of populations over space and time is, arguably, the motivation for much of pure and applied ecological research. The fundamental model for the dynamics of any population is straightforward: the net change in the abundance is the simple difference between the number of individuals entering the population and the number leaving the population, either or both of which may change in response to factors intrinsic and extrinsic to the population. While harvest of individuals from a population constitutes a clear extrinsic source of removal of individuals, the response of populations to harvest is frequently complex, reflecting an interaction of harvest with one or more population processes. Here we consider the role of these interactions, and factors influencing them, on the effective harvest management of waterfowl populations. We review historical ideas concerning harvest and discuss the relationship(s) between waterfowl life histories and the development and application of population models to inform harvest management. The influence of population structure (age, spatial) on derivation of optimal harvest strategies (with and without explicit consideration of various sources of uncertainty) is considered. In addition to population structure, we discuss how the optimal harvest strategy may be influenced by: 1) patterns of density-dependence in one or more vital rates, and 2) heterogeneity in vital rates among individuals within an age-sex-size class. Although derivation of the optimal harvest strategy for simple population models (with or without structure) is generally straightforward, there are several potential difficulties in application. In particular, uncertainty concerning the population structure at the time of harvest, and the ability to regulate the structure of the harvest itself, are significant complications. We therefore review the evidence of effects of harvest on waterfowl populations. Some

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 221

of this evidence has focussed on correspondence of data with more phenomenological models and other evidence relates to specific mechanisms, including density- dependence and heterogeneity. An important part of this evidence is found in the evolution of model weights under various adaptive harvest management programmes of the U.S. Fish and Wildlife Service for North American waterfowl. Overall, there is substantial uncertainty about system dynamics, about the impacts of potential management and conservation decisions on those dynamics, and how to optimise management decisions in the presence of such uncertainties. Such relationships are unlikely to be stationary over space or time, and selective harvest of some individuals can potentially alter life history allocation of resources over time – both of which will potentially influence optimal harvest strategies. These sources of variation and uncertainty argue for the use of adaptive approaches to waterfowl harvest management.

Key words: additive mortality, compensatory mortality, harvest, population structure.

Annual migrations of waterfowl have long waterfowl harvest is now through provided human populations with a regular recreational hunting. Because of their source of protein and outdoor recreation. cosmopolitan distribution over both the Their use as a valuable quarry and food has northern and the southern hemisphere, been facilitated by the great concentrations there is virtually no area around the globe of these birds at some wintering or where there are wetlands and no waterfowl migration stopover sites, the ease with harvest of any type. which their eggs could be collected, and the Throughout much of the world, and flightless moulting period making adults certainly in North America and Europe, particularly vulnerable to trapping during there is widespread recognition that late summer. Not surprisingly, waterfowl waterfowl hunting requires some form of remains are very common at prehistoric regulation. This recognition reflects the human settlement sites (e.g. Ericson & assumption that unregulated harvest has the Tyrberg 2004), and antique artefacts of potential to reduce waterfowl populations waterfowl hunting are numerous (e.g. to dangerously low levels. As a result, Egyptian paintings or Roman mosaics; see government organisations worldwide have Arnott 2007). Subsistence hunting of imposed various restrictions on the hunting waterfowl is still a traditional activity, of waterfowl, including, for example, especially in the Arctic where some duck establishment of seasons of the year and and goose species breed (Padding et al. times of day when hunting is not permitted, 2006). Commercial harvest of waterfowl is areas within which hunting is not permitted, also legal and heavily practiced in some parts daily limits to the number of birds that can of species wintering ranges (Balmaki & be harvested, restrictions on types of baits Barati 2006). However, much of current and other attractants (e.g. types of decoys)

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 222 Effects of harvest on waterfowl populations that can be used, and restrictions on the 1990), constructing islands as potential number and types of shells permitted. nesting sites with reduced access to These various restrictions represent predators (Hammond & Mann 1956; management actions designed to bring Giroux 1981), and planting dense nesting about desired outcomes with respect cover for prairie nesting species (Duebbert to harvested waterfowl populations. et al. 1981; McKinnon & Duncan 1999). Implementation of such restrictions thus Other actions, such as the Conservation pre-supposes knowledge of the Reserve Program (Reynolds et al. 1994, relationships between these regulations and 2001), are much less specific, and are harvest rates, and between harvest rates and designed to influence habitat across broad waterfowl population change. However, geographic areas. Active control of nesting both sets of relationships are characterised predators can be viewed as a form of by uncertainty. Some of this uncertainty is breeding habitat improvement. Predator likely not resolvable. For example, the exact reduction has been successfully applied harvest rates that result from a specific set of throughout the world (Garrettson & hunting regulations are always likely to be Rohwer 2001; Kauhala 2004; Whitehead et viewed as random variables arising from a al. 2008; Pieron & Rohwer 2010); however, distribution that characterises this source of in many parts of the world (e.g. North partial controllability (Johnson et al. 1993, America) controversy remains about 1997). However, the relationships between whether this action should be considered. harvest rates and both waterfowl survival Management actions affecting migration rates and population change are represented and wintering habitat have also been by competing hypotheses and thus by identified and implemented for waterfowl uncertainty that is potentially resolvable by (Gilmer et al. 1982; Smith et al. 1989). In evidence. These hypotheses are the focus of summary, a variety of potential management this review, as we consider available actions exists, and integrated programs of evidence and ways to provide further waterfowl management should include resolution as a means of improving future consideration of multiple actions (including management of waterfowl resources. harvest regulations) in order to achieve Despite the relatively narrow focus of our programme objectives (Runge et al. 2006). review, we remind the reader that harvest regulation is one of a relatively large number History of waterfowl harvest of potential actions that can be used to management manage waterfowl populations. For In North America prior to the mid-1800s, example, a variety of management actions waterfowl were viewed as extremely has been developed to improve habitat on abundant and accordingly were hunted for waterfowl breeding grounds. Some of these sale and recreation throughout the year actions are very specific and local, such as (Phillips & Lincoln 1930; Day 1949). erecting nesting structures for cavity nesting Population declines in the late 1800s and species (Hawkins & Bellrose 1940; Bellrose early 1900s led to concerns about effects of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 223 harvest and to the beginning of government conclusion that historical data did not intervention. The United States government provide strong support for the premise that was granted authority to regulate waterfowl had guided Mallard harvest for the prior 50 harvest, and the Migratory Bird Treaty Act years, that changes in Mallard harvest rates of 1918 specified that hunting would be had produced corresponding changes in permitted only when deemed compatible Mallard survival and population size. This with protection and maintenance of landmark study introduced structural populations. During the period 1930–1950, uncertainty to North American waterfowl the perception of declines and low harvest management; that is, uncertainty in populations led to restrictions in hunting hypotheses about how changes in waterfowl regulations (U.S. Fish and Wildlife Service harvest translate into changes in population 1988) and to the initiation of monitoring dynamics. Subsequent efforts to resolve this programmes designed to assess waterfowl uncertainty and manage waterfowl harvest population status (Martin et al. 1979; Smith in the face of it include a period (1979– et al. 1989; Nichols 1991a). Over the next 25 1985) of stabilised hunting regulations years, these monitoring programmes were (McCabe 1987) and a subsequent period expanded and improved, and resulting data (1985–1990) of risk-aversive conservatism were employed to develop population (Sparrowe & Patterson 1987; U.S. Fish and models for use in establishing harvest Wildlife Service 1988). However, neither of regulations for key species (Crissey 1957; these approaches led to resolution of the Geis et al. 1969). During this period (1951– uncertainty, nor to a widely accepted 1975), these models and conventional approach for dealing with it. wisdom led to restriction of hunting In the early 1990s, members of the Office regulations during years when breeding of Migratory Bird Management, U.S. Fish grounds were dry and population sizes and Wildlife Service (USFWS), began to low, producing disagreements about the give serious consideration to implementing effectiveness of such restrictions and an adaptive approach to harvest the perceived lack of consideration of management. Although the central ideas the desires of the hunting public underlying adaptive management had been (Nichols 2000). However, these political described and developed by Walters disagreements were not well-grounded in (1986), the approach had never been fully science, and the management of waterfowl implemented on even a small scale. In 1992, hunting in North America was generally Fred Johnson of the USFWS assembled an viewed as a good example of the scientific ad hoc working group of state and federal management of animal populations waterfowl biologists to discuss alternative (Nichols et al. 1995). approaches for waterfowl harvest In the early 1970s, analyses of Mallard management. The ideas of adaptive harvest Anas platyrynchos ringing and recovery data, management (AHM) were discussed, and using newly developed inference methods, the group decided to develop this approach, led Anderson & Burnham (1976) to the becoming the interagency working group

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 224 Effects of harvest on waterfowl populations for AHM. The proposed approach to AHM legal in some countries. Because of the was outlined in Johnson et al. (1993) and number of different countries in Europe, it formally adopted by the USFWS for mid- is more difficult to reach international continent Mallard in 1995 (Nichols et al. agreements, and national waterfowl 1995; Williams & Johnson 1995; Johnson et management policies have sometimes al. 1997). AHM was especially attractive developed towards different systems and at because it provided a means of different paces. The main current legal simultaneously reducing uncertainty while framework is the ‘Bird Directive’ (adopted managing in the face of it. The AHM by the European Commission in 1979), programme for mid-continent Mallard which limits in particular the periods of the is still used each year to establish year during which birds can be harvested recommended hunting regulations. Its anywhere along their flyways, and the success has led to the development of AHM African-Eurasian Waterbird Agreement programmes for other Mallard populations (AEWA), which aims to coordinate and other waterfowl species in North research, monitoring and policy at the America (e.g. Atlantic Flyway Canada Geese Palearctic flyway scale (including beyond the Branta canadensis; Hauser et al. 2007). European Union). The European waterfowl Inferences reviewed in this paper about the management policy is therefore far less relationship of hunting regulations and developed than the American system, harvest rates to waterfowl populations are although the eventual set-up of a proper based both on specific analyses and on the international adaptive harvest management results of AHM programmes. scheme is a goal for the future (Elmberg et The general pattern of increasing al. 2006). In fact, an adaptive management protection of waterfowl and regulation of programme for the Svalbard Pink-footed harvest has occurred in Europe as well, Goose Anser brachyrynchus population is where waterfowl were also considered as an under current development (e.g. Johnson et almost infinite resource until the end of the al. 2014). 19th Century, and were commercially exploited as such. Duck decoys, in Life history characteristics of particular, were used to trap birds at their waterfowl wintering and migration stopover sites, Despite the general similarities in sometimes in an industrial manner (33,000 morphology and behavioural habits, teal were caught in a single season on one waterfowl form a very diverse family of island in the North Sea, leading to the birds when it comes to body size, with up to building of a duck canning factory; Phillips a 32-fold difference in body mass between a 1923). Such commercial harvests gradually 330g Green-winged Teal Anas carolinensis lost popularity and were abandoned and a 10.5 kg Trumpeter Swan Cygnus throughout Europe during the 20th buccinator. Such differences in body mass century, although trade of legally-harvested have obvious consequences in terms of, for waterfowl by recreational hunters is still instance, energy needs (Miller & Eadie

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 225

2006), which translate into very different life breeding grounds to fuel their reproduction. history strategies. These differences have In contrast, the larger size of geese and long been recognised, as exemplified in the swans permits the storage of lipid and following statement from the mid-20th protein reserves well before reproduction, century: “Another inference from the often as early as on the wintering grounds, general observation that geese are bigger leading to their characterisation as “capital than ducks is that large waterfowl survive breeders”. Such differences in life history better, and produce fewer offspring, than strategies have profound consequences for small ones” (Boyd 1962). population structure and, hence, the Waterfowl therefore can be broadly modelling of population dynamics. Duck organised along a “fast-slow” gradient, with populations are generally considered as being faster duck species having short life relatively simple in structure, with little need expectancies but a high annual reproductive to incorporate age structure beyond the first- output, as opposed to slower geese and year/adult dichotomy (Devineau et al. 2010). swans surviving much longer but producing Conversely, the low reproductive rate and fewer offspring per breeding attempt long survival of geese, swans and many sea (Gaillard et al. 1989). In line with ducks ducks lead to more complex populations, producing more offspring which survive with delayed age of first breeding, extended more poorly, density-dependent feedback age-specificity, etc. The greater heterogeneity on individual survival is thought to be more among individuals within such populations common in ducks, at least during some often requires the use of more structured stages of their life cycle (Gunnarsson et al. models. 2013), while this is not so much the case in geese and swans. Life history variation exists Modelling considerations: even within ducks (subfamily Anatinae). For structure & heterogeneity example, among North American ducks, Patterson (1979) characterised Mallards, Models are used in harvest management to Blue-winged Teal Anas discors, and Northern allow us to predict the numerical response Pintail Anas acuta as relative “r-strategists” of a population subjected to a certain level (faster life histories) and Redhead Aythya of harvest. However, all population models, americana, Canvasback Aythya valisineria and regardless of their application (e.g. harvest Scaup Aythya marila as “K-strategists” management), represent approximations to (slower life histories). reality which can never be fully specified. The constraint imposed by their smaller Utility of the model in the context size prevents ducks from carrying substantial of harvest management is primarily body reserves along long migratory flights determined by the degree to which the (Klaassen 2002), leading to their model correctly represents the functional characterisation as “income breeders”; sensu form relating the management control Drent & Daan (1980), in that they mostly option (say, varying harvest pressure rely on the energy available at or near their through legislative action), and the response

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 226 Effects of harvest on waterfowl populations of the population to harvest. Much step t to t + 1 requires inference about a literature has focussed on the question of binomial parameter based on a single whether harvest mortality in waterfowl is Bernoulli trial. Based on whether the “additive” or “compensatory” to natural individual is alive or dead at time t + 1, we mortality (see below), and the impacts of must somehow estimate the underlying that distinction on optimal harvest probability of survival, and there is simply management. Second, the ability of a not enough information in this single population model to reflect accurately the observation to allow us to do this well (see dynamics of a population will be strongly Cohen 1986). As such, we are generally left influenced by the degree to which the model constructing a model which represents a structure adequately accounts for important compromise between a simple scalar model, differences among individuals, both in terms and a fully individually-based model. Such of underlying vital rates (survival, fertility), “intermediate” models are based on the but also potentially in the functional reasonable idea that much (if not all) response of those vital rates to perturbation of the variation among individuals (i.e. harvest). At one extreme, simple scalar can be explained by one or more models assume that all individuals have factors (demographic, genotypic, spatial, identical latent probabilities of survival and developmental), which can be used to reproduction – we refer to such models as structure a population into (generally) scalar projection models. As noted in the discrete classes of individuals grouped preceding section, for many duck together by sharing one or more of these populations, such simple scalar or near- factors (we note that in some cases, scalar models are often sufficient. At the discretization is a mathematically convenient other extreme, we imagine a model approximation to the continuous state- containing sufficient structure to model the space). For many waterfowl species, dynamics of each individual in the particularly longer-lived swans and geese, population. We refer to such models as and many sea ducks, there is significant individually-based projection models. It is clear variation in both survival and fertility as a that this latter class of models represents the function of the age of the individual. closest approximation to full reality. Such a A model which differentiates among model would completely account for individuals based on differences in such heterogeneity among individuals in the factors is known generally as a structured population (the role of individual model. Such models are parameterised not heterogeneity will be re-visited later). only in terms of potential differences in However, a fully individually-based survival and fertility among classes of population model is generally intractable, individual, but also in terms of the both in terms of construction, analysis and probability of making transitions among application in a management context. For classes (due to aging, growth or movement). example, estimation of a time-specific This section addresses the impact of survival probability for individual i over time population structure on the projected

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 227 impacts of that harvest on waterfowl equilibrium harvest would be 50 individuals population dynamics; we defer consideration at each time step. of the role of different functional forms relating harvest to the numerical response to Equilibrium harvest for structured the next section. Here, we describe the populations conditions under which harvest would lead Now we consider a structured population. to change in population abundance, and the While there are a number of different structural factors influencing the magnitude classes of structured models, we will focus and time course of such change. on the use of matrix-based models, which are canonical models for discrete-time Equilibrium harvest for scalar population dynamics, where individuals populations are classified (grouped) into discrete We introduce some of the basic “(st)ages” (Caswell 2001; Lebreton 2005). considerations in model-based harvest To simplify the presentation, we will management through a numerical example. consider deterministic models, with no We consider first a simple deterministic density-dependence. We’ll assume the scalar population in discrete time, without minimal structured model with 2 age classes density-dependence, where the population (juveniles and adults, where the adult class size N at time t + 1 is given as the product consists of all individuals ≥ 1 years of age). of the current population size at time t and Fertility and survival transitions for our a scalar multiplier, λ. example population are given in the life- cycle graph shown in Fig. 1 (constructed N t +1 = N t . (1) assuming a post-breeding census; Caswell As long as λ > 1, then N > N (i.e. the t + 1 t 2001). Assuming S = 0.65, S = 0.5, and F = population will grow). In the absence of A J 0.8, then the projection matrix model A can harvest or some other “control” measure, be constructed directly from the life-cycle the population will grow without bound. In graph as such cases, we focus on calculating the   equilibrium harvest rate, E, which S JF S JF  0.40 0.52  A =   =   , (3) 0.50 0.65 represents the maximum harvest which does  S J S A    not lead to the increase or decline of the population over time (i.e. the harvest from which we derive the following standard metrics for projected growth (λ), condition under which Nt + 1/Nt = 1): stable age proportions (wi), and age-specific N t +1 = ()  E . (2) reproductive values (vi; for details see N t Caswell 2001): λ Thus, for example, a population with =  0.3478   1.0  1.05 is projected to grow at 5% per time  = 1.05, w =  , v =   . 0.4348 1.5 step. The equilibrium harvest rate then is     simply E = (1.05–1) = 5%. In terms of Thus, in the absence of harvest and absolute numbers, if Nt = 1,000, then the assuming time-invariance and no density-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 228 Effects of harvest on waterfowl populations

SFA the harvest of a structured population, and harvest of a simple scalar population. Reproductive value is a well-known concept in evolutionary biology (Stearns 1992 1 2 provides a general review) and has been S J identified as affecting the optimal age- or stage-specific harvest of a population

SFJ S A (MacArthur 1960; Grey & Law 1987; Brooks & Lebreton 2001; Kokko 2001; Lebreton Figure 1. Life-cycle graph and structure of the 2005; Hauser et al. 2006). Harvest of underlying life history for the 2-age class (adults, juveniles) example. The life-cycle graph is based individuals of higher reproductive value will, on a post-breeding census. Node 1 is the number generally, have a greater proportional impact of juveniles (offspring) in the population, and on population dynamics than harvest of node 2 is the number of adults (age ≥ 1 year). individuals with lower reproductive value The arcs connecting the nodes reflect survival (although the relative value of individuals (left-to-right) and fertility (right-to-left). SA and SJ may change following a perturbation are the survival probabilities for adults and (Caswell 2001; Cameron & Benton 2004) juveniles, respectively. F is the reproductive rate, and is a function of whether or not the and is assumed to be invariant with age for age population is increasing or decreasing at the > 1 year. time of harvest (Mertz 1971)). Thus, the inclusion of structure adds extra dimensions dependence, the population is projected to of uncertainty, but also additional flexibility grow without bound at 5% per time step, and opportunities, to harvest management. eventually stabilising at a juvenile:adult ratio Most obviously, a structured model may of 4:5. As indicated by the reproductive require a structured harvest to reach an value vector v (Fisher 1930), each adult in the optimal harvest objective. population is worth 1.5 juveniles to future population growth, suggesting that harvest Constant harvest of an adult individual is potentially of We can demonstrate the role of structure greater impact than harvest of a juvenile and age-specific reproductive value by individual. (Here, we have normalised v, and means of a simple numerical example. w so that v’w = 1. It is customary to express Suppose at time t the population consists of v such that v1 = 1, so that the reproductive ~1,000 individuals. Based on a simple scalar value of each stage is compared to that of model, a projected growth rate of λ = 1.05 the first stage, and w such that the elements (as per the preceding example) implies that wi sum to 1, so they represent the proportion we could harvest at most 5% of the of the population in each stage class). population each time step. Given, say, 1,000 The difference in reproductive value individuals in the population at the time of between adults and juveniles represents a key harvest, this would correspond to a constant consideration which differentiates modelling harvest of 50 individuals.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 229

But, which 50 individuals? Juveniles, period (here, we assume a specific, constant adults, or some of each? Consider the 2-age- number of individuals harvested for each class structured model introduced earlier age class. We consider proportional harvest (eqn. 3), and the consequences of later). Under equilibrium harvest for a harvesting 50 juveniles, versus harvesting 50 discrete-time projection model, Nt + 1 = Nt adults from a starting population of ~1,000 = N*. Thus, eqn. (4) can be rearranged to individuals (assuming for the moment that show that: 1 we could selectively harvest a particular age N* = ()A  I . (5) class – the implications of violating this If A is primitive (which is generally the assumption are considered later). If the case for population projection models, population structure at the time of the which are generally positive and square), harvest is proportional to the equilibrium then the unharvested population will age structure (i.e. consisted of 444 juveniles, eventually grow as: and 555 adults), then the population – and N  ()v'N tw . thus each age class – is projected to increase t 0  at 5% per time step. Under these conditions, Following Hauser et al. (2006), if the it might seem reasonable to assume that difference between the harvested population harvesting 50 adults, or 50 juveniles, or any and the equilibrium state after some time vector summing to 50 total individuals (e.g. t is: t 25 adults and 25 juveniles), would have the N t – N*~[]v'()N 0 – N *  w , (6) same effect on long-term dynamics (namely, then the equilibrium harvest vector E is no change in population size between now given as: and the next time step following harvest). v'E = v'N () 1  . (7) However, this is not the case – in fact, the  0  direction and magnitude of the change Now (λ–1) is the long-term proportional in the population is determined by the increase of the unharvested population (e.g. relative proportions of each age class in the if λ = 1.05, then the population will increase harvest. by (1.05–1) = 5% per year in the long term). Since such a result might seem counter- The reproductive value of this ‘excess’ intuitive, it is useful to evaluate the correct proportion of the initial population N0 (the equilibrium harvest conditions for a right-hand side of eqn. (7) must be equal to structured population. Let the dynamics of the reproductive value of the harvest (the a structured population subjected to a left-hand side of eqn. (7). This ensures that constant harvest be given by: the harvest is sustainable and that the population will approach a steady state over N +1 = AN  E , (4) t t time. where A is the matrix projection model, and Returning to our numerical example, if

E is the harvest vector where the ith element the initial population N0 is known, then we represents the number of individuals in find harvest vectors E that satisfy the stage i that is harvested during each time equilibrium condition (eqn. 7). The initial

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 230 Effects of harvest on waterfowl populations population may have the stable stage 50 adults only, and no juveniles, is above the distribution of the unharvested system, e.g. line, leading to a population decrease. In ′ N0 = (444,555) . Then: contrast, a harvest of 50 juveniles only, and   no adults, is below the line, leading to a   E J  1.0 1.5   population increase.  E A  Note that the structure and size of this  444  steady state population (i.e. at equilibrium) = (1.05 1)  ,  1.0 1.5   are dependent on the particular harvest  555  vector E that is used. For example, if we leading to: choose to harvest only adults then the +1.5 = 63.825 , E J E A equilibrium population size is calculated to where EJ and EA are the numbers of be > 1,000 (i.e. above the starting population juveniles and adults harvested per time step, size): respectively. The equilibrium condition  0.00   464.2  (eqn. 7) is that a particular linear E =   and N* =   = 1005.7 .  42.55   541.5  combination of the harvest taken from each class is held at a constant. It is important to If instead we choose to harvest only note that the coefficients for harvest from a juveniles, the equilibrium population size is i.e. class (i.e. EJ, EA) are the reproductive values calculated to be < 1,000 ( below the starting population size), then: of individuals in those classes (i.e. v1 = 1.0 for EJ, v2 = 1.5 for EA). The constant (in  63.83   406.2  this case 63.825) is dependent on the initial E =   and N* =   = 986.4 .  0.00   580.2  population size and structure N0. The particular solution to the equilibrium Three important points should be noted harvest equation for our present example, here. First, as mentioned earlier, the ′ where N0 = (444,555) , is shown in Fig. (2). coefficients for harvest from a given age- ′ If the harvest E = (EJ, EA) that is actually class (i.e. EJ, EA) are the reproductive values taken falls below this line, then the of individuals in those classes (i.e. v1 = 1.0 population will eventually increase. If the for EJ, v2 = 1.5 for EA). In other words, the harvest falls above this line then the equilibrium harvest of juveniles only population will eventually decline. Harvest (63.825) would need to be 1.5 times larger that falls on this line (i.e. satisfying the than the equilibrium harvest of adults only equilibrium condition) will cause the (42.550; 63.825/42.550 = 1.5). This is population to stabilise over time to a because the harvest of a single adult population size and structure given by from the population is demographically equation (7). We introduced this example by equivalent to the harvest of 1.5 juveniles. claiming that a total harvest of 50 This linear relationship between the individuals would cause the population to reproductive value vector and the increase or decrease over the long term. We equilibrium harvest vector is not limited to see clearly from this figure that a harvest of simple 2-age class models – for any number

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 231

50 ) A 40 – (Population decrease)

30

20 + 10 (Population increase) N u mb er of adults harvested ( E

0 0 10 20 30 40 50 60

Number of juveniles harvested (EA) Figure 2. Equilibrium harvest for constant harvest of a fixed number of adults (vertical axis) and juveniles (horizontal axis), for the population projection model described by eqn. (3). The equilibrium harvest is specific to the initial population size and structure, which here we assume to be 1,000 ′ individuals in the stable age proportions (i.e. N0 = (444,555) ). Harvest at any point below the equilibrium (shaded area) will cause the population to increase, whereas harvest at any point above the equilibrium will cause the population to decrease. Adapted from Hauser et al. (2006). of age classes (≥ 2), the equilibrium the structure of the harvest itself, is an expression is a k-dimensional plane, for k important consideration addressed later (see age classes in the model (Hauser et al. 2006; also Koons et al. 2014a). for most long-lived waterfowl species, Finally, it is possible that the population k ≥ 5). The coefficients of equilibrium does not stabilise at equilibrium abundance, solution are the reproductive values of the but instead grows unbounded, even if the corresponding age classes. harvest satisfies eqn. (7). This can occur if in The second point is that the population some time steps the number of individuals size at equilibrium is smaller for a juvenile- that needs to be removed for a given age only harvest (986.4), and more skewed class is larger than the number of towards adults, compared to an adult- individuals existing in that age class at that only harvest (1,005.7, 46% juvenile). This time. In such cases, the full “equilibrium dependence of the final size and structure harvest” cannot be taken, and the of the population on the structure of the population grows unbounded (Hauser et al. population at the time of the harvest, and 2006). (This issue does not occur if we

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 232 Effects of harvest on waterfowl populations implement a proportional harvest, as We solve for K in a straightforward way. developed in the next section). Using the matrix A from our 2-age-class model, then from eqn. (9): Proportional harvest 0.7  0.3K J 0.6  0.6K J In the preceding, we considered a constant = 0 , 0.5  0.5 0.35  0.65 harvest, where a constant number of K A K A individuals is harvested (from a given age which can be simplified to: class) at each harvest decision. Such a 1181K J scenario is arguably unrealistic for waterfowl = , K A 151 21 harvest, where harvest regulations are based K J on an assumed relationship between various which is a non-linear expression in KA and regulatory options and the proportional KJ (although over the limited range of ≤ ≤ probabilities of mortality due to harvest plausible values of 0 Ki 1 for this (i.e. kill rate, Kt, the probability of being problem, the equilibrium solution appears harvested during time interval t). Under linear; Fig. 3). If the harvest chosen falls on proportional harvest, the same proportion of this curve then the population will approach individuals is removed from each age class a steady state over time. If the harvest each time period (although the proportion chosen falls below the curve then the may differ among age classes). The harvest population will grow geometrically over model becomes: time. If the harvest chosen falls above the curve then the population will decline Nt +1 = (I  K)AN t , geometrically. where K = diag(K ,K ,…,K ), and 0 ≤ K ≤ 1 1 2 k i Two important differences should be is the proportion of age class i to be noted compared to the constant harvest harvested. If harvest occurs immediately scenario introduced earlier. First, the after reproduction (generally the case for equilibrium condition (eqn. 9) is not waterfowl), then: dependent on N0 (although the equilibrium N = AI() K N . (8) t +1 t population vector N* is; eqn. 10), and is not To find the equilibrium condition under a simple function of reproductive values proportional harvest, we again set Nt + 1 = (although it is clear from the equilibrium Nt = N*, and solve for N*: condition that a higher harvest rate for

N* = AI() K N * . (9) juveniles is required in order to achieve equivalence with a given harvest rate for That is, we choose harvest vector K so that adults, consistent with the interpretation of 1 is the dominant eigenvalue of (I – K)A. the relative value of juveniles and adults We denote the corresponding right and under a constant harvest model). Second, left eigenvectors as v and w , respectively. K K because we are dealing here with a Then the population under harvest K will proportional harvest, K , where 0 ≤ K ≤ 1, approach the equilibrium: i i then the situation where a full equilibrium N* = (v' N0 )w . (10) K K harvest cannot be taken – as was the case

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 233

0.08

) – A (Population decrease)

( H 0.06

0.04 +

A dult harvest rate (Population increase) 0.02

0.00 0.00 0.02 0.04 0.06 0.08 0.10 0.12

Juvenile harvest rate (HJ) Figure 3. Equilibrium harvest for harvest of a fixed proportion of adults (vertical axis) and juveniles (horizontal axis), for the population projection model described by eqn. (3). The equilibrium harvest is specific to the initial population size and structure, which here we assume to be 1,000 individuals in the ′ stable age proportions (i.e. N0 = (444,555) ). Harvest at any point below the equilibrium (shaded area) will cause the population to increase, whereas harvest at any point above the equilibrium will cause the population to decrease. Adapted from Hauser et al. (2006). for a constant harvest if the equilibrium “Structure”, by any other name… harvest for a particular age class was larger than the number of individuals in that age In the preceding, we focussed exclusively on class – cannot occur. As such, under a age structure (i.e. where the age of the proportional harvest, there is always a individual was the only determinant of bounded equilibrium. However, as was variation in survival or fertility among the case for a constant harvest, the individuals specified in the model). For size and structure of the equilibrium some species of waterfowl, especially population varies depending on both the longer-lived swans, geese and many sea structure of the harvest vector K, and the ducks, age structure is clearly an important structure of the population at the time of consideration. harvest, N0 (Hauser et al. 2006). Since However, while age structure may waterfowl harvest management is almost generally be a less important consideration universally based on proportional harvest, for many short-lived duck species, there we do not discuss constant harvest models are other forms of structure which may further. be important considerations in model

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 234 Effects of harvest on waterfowl populations construction, some especially for ducks, and or sex, or location) based on a finite set others for waterfowl generally (e.g. spatial of “classes” of individuals (this is strictly structure). Brooks & Lebreton (2001) analogous to the practice of using describe a simple application of the finite mixture models to approximate methods described in the preceding to a heterogeneity in analysis of mark-encounter metapopulation, where the value of the data; e.g. Pledger et al. 2003). For example, individual is conditioned on both age and consider a population where individuals can location. Further, the models presented be characterised as either “high quality” or above considered individuals of only one “low quality”, based on their total latent sex. While this may be appropriate in some survival probability (where a “low quality” cases, it may not always be the case, individual is one with a lower probability of especially for species where the dynamics of survival). In theory, we could reconfigure a the population are influenced by the form of matrix model which assumes that all the mating system (e.g. polygamy), or more individuals within a given structural class commonly, where survival rates differ have identical latent vital rates (e.g. Fig. 1), to significantly between the sexes. In such account for two discrete “quality” classes cases, models where the dynamics of the within that structural class. For example, two sexes are linked by the pair bond and if we assume that “quality” differences maternity function are more appropriate. influence only juvenile and adult survival, Such sex-linked models can generate rather but not fertility, then we might restructure complex dynamics (Caswell & Weeks 1986; the 2 age-class matrix model we considered Lindström & Kokko 1998). earlier (Fig. 1) to reflect the unequal Finally, all waterfowl populations (and contributions of individuals of different wild populations in general) are quality to population growth (Fig. 4). characterised by differences among The projection matrix model A can be individuals that extend beyond the sources constructed directly from this life-cycle of variation already discussed. Even when graph as: models structurally separate males and   S J ,h F +(1–  )S J ,l S A,h FSA,l F females, or young and old, or one location   versus another, within a given “node” (say, A =  S J ,h S A,h 0  , females of age 2, that are in location X),  (1–  ) 0   S J ,l S A,l  there are remaining differences among individuals. These differences are commonly where Sx,q is the latent survival probability of referred to as reflecting “individual individuals of age class x, and quality class q heterogeneity”. While it is quite likely that (where q = high or low), F is the fertility rate, these differences vary continuously among and π is the probability that a juvenile is (or individuals, as a first approximation, we will become) a “high” quality individual. can consider modelling this additional In this sense, heterogeneity models are (at heterogeneity (i.e. differences beyond those least in simple, discrete form) structurally explained by structural elements such as age, equivalent to models structured based on

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 235

SF A,h For some taxa (either sessile organisms, organisms that are temporarily sessile at πSFJ,h 2 the time of harvest – e.g. fish in the net S π J,h S in a commercial fishery, or where the 1 A,h mechanism of harvest is state-selective – e.g. (1- π)S J,l use of nets of specific mesh size or other (1-π)SF J,l 3 gear to capture only certain size-classes of S A,l fish), it is possible to harvest an individual SF A,l selectively on the basis of its individual state Figure 4. Life-cycle graph and structure of the (e.g. you keep the big ones, and toss back the underlying life history for a 2-age class (adults, small ones). For most waterfowl, however, juveniles) model, with discrete (finite mixture) there will often be considerable uncertainty heterogeneity in juvenile and adult survival. The in establishing the “state” of an individual at life-cycle graph is based on a post-breeding the time of harvest, and optimisation of census. Node 1 is the number of juveniles harvest based on state structure of the (offspring) in the population, and represents the population will be only partially controllable. combined fertility contributions of high quality Even in cases where the extent of ≥ (node 2) and low quality (node 3) adult (age 1 uncertainty about system state is reduced year) individuals, where differences in quality are (for example, if harvest occurs at locations characterised by lower juvenile and adult survival where the targeted individuals represent rates among lower quality individuals. The arcs specific age- or sex-classes), waterfowl connecting the nodes reflect survival (left-to- harvest in many cases is simply a random right) and fertility (right-to-left). SA,q and SJ,q are the survival probabilities for adults and juveniles, selection of individuals with differential respectively, for quality class q. The parameter π vulnerability to the harvest. determines the probability that a new juvenile (node 1) becomes a high quality adult. F is the Consequences of hunting for reproductive rate, and is assumed to be invariant waterfowl populations with age and quality for age > 1 year. As noted above, a key to making wise age, location, or gender, and we can apply management decisions is to be able to make the same methods discussed earlier to predictions about system response to evaluate the dynamics of the population potential management actions. For hunted described by this “heterogeneity” model to populations, predictions will usually include evaluate the relative value of harvest of the hunting mortality rate that will result individuals of low or high quality to from any prescribed set of regulations. those dynamics (i.e. reproductive value Studies contrasting ring recovery rates in conditioned on both age and quality). The years of differing hunting regulations have possible role of unspecified heterogeneity provided evidence supporting the inference on numerical response to harvest is of higher harvest rates in years of more considered in a later section. liberal hunting regulations (see reviews of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 236 Effects of harvest on waterfowl populations

Nichols & Johnson 1989; Nichols 1991; where E denotes expected value, St is the Johnson & Moore 1996, and more recent probability that a bird alive at the beginning species-specific evidence of Johnson et al. of the hunting season in year t survives and 1997; Francis et al. 1998; Calvert & Gauthier is alive at this same time the next year (t + 1),

2005; Alisauskas et al. 2011; Peron et al. 2012; S0 is the probability that a bird alive at the Iverson et al. 2014). Of course numerous beginning of the hunting season in year t variables (e.g. environmental conditions and would survive to that same time the resulting migration timing, and regional following year t + 1 in the complete absence hunter activity) in addition to hunting of any hunting mortality, Kt is the regulations are expected to influence probability that a bird alive at the beginning hunting mortality rates. As a result, hunting of the hunting season in year t would die mortality rates predicted to correspond to a from hunting causes before that same time specific set of hunting regulations are best the following year t + 1 in the complete characterised as a probability distribution. absence of any non-hunting mortality, β Examples of such distributions estimated and is the slope parameter relating St for mid-continent Mallard and Black Duck and Kt. are provided by Johnson et al. (1997) and Equation (11) is very general and can be USFWS (2013), respectively. used to model a variety of relationships In addition to predicting the hunting between hunting and survival depending on mortality rate, predictions are required for the value of β, with β = 1 and β = 0 the changes in survival rates (probability of indicating plausible models with maximal surviving all mortality sources), reproductive and minimal effects of hunting on survival, rates, and the rate of movement in and respectively. S0 and Kt are each defined as out of the focal population expected to applying when the other mortality source is accompany this level of hunting mortality. not operating. As such, they are referred to Much of the uncertainty in waterfowl as net rates in the literature of competing management involves these relationships, mortality risks (e.g. Berkson & Elveback and most of the North American 1960; Chiang 1968; see below). S0 is not waterfowl programmes in adaptive harvest defined as time-specific (it is not subscripted management include multiple models (and by t), for consistency with historical corresponding hypotheses) as a means of development, but time-specificity is dealing with this uncertainty. We consider certainly possible conceptually. these relationships below. Because S0 and Kt are net rates, they cannot usually be estimated directly. Instead, Survival without extra information about the timing The most direct influence of hunting of the different mortality sources, we mortality should be on the total (all sources) are usually restricted to estimation of so- mortality rate. A general form for this called ‘crude’ rates (sensu Chiang 1968). relationship can be expressed as (eqn. 11): Specifically, a crude, source-specific

E(S t ) = S 0(1 K t ) , (11) mortality rate is the probability of dying

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 237 from that source in the presence of all other the common situation (for North American mortality sources. In the case of modelling waterfowl) of ringing occurring just before ′ and inference about duck populations, Kt is the hunting season. defined as the crude hunting mortality rate, or the probability that a bird alive at the Additive mortality hypothesis beginning of the hunting season in year t Anderson & Burnham (1976) used equation would die from hunting causes during the (11) to define two endpoint hypotheses hunting season of year t in the presence of designed to bracket the possible non-hunting mortality that occurs during relationships between total survival and ′ this period. Even Kt cannot typically be hunting mortality. They used the term estimated directly, but instead requires “additive” for the situation where β = 1, information from multiple sources. Ring corresponding to a model in which hunting recovery data are the common source of and non-hunting mortality sources are information about hunting mortality, and viewed as independent competing risks corresponding models permit direct (Berkson & Elveback 1960; Chiang 1968). estimation of ring recovery rates, ft, the The term “additive” is applicable, as the probability that a ringed bird alive at the instantaneous risks associated with the two beginning of the hunting season of year t is mortality sources are added in order to shot and retrieved by a hunter during the obtain the total (both sources) probability of hunting season of year t and its ring dying during an interval in which both correctly reported (see Brownie et al. 1985; sources apply. This additive mortality model

Williams et al. 2002). Define ct as the is commonly used in fisheries management probability that a bird shot by a hunter (Beverton & Holt 1957; Ricker 1958; during the hunting season of year t is Hilborn & Walters 1992) and is intuitively retrieved by the hunter (1 – ct denoting reasonable. λ ‘crippling loss’), and t as the probability Although equation (11) applies to net that a retrieved bird is reported. Then: rates of hunting and non-hunting mortality, regardless of their temporal patterns of f t = K t c t t . (12) occurrence, the intuition underlying this If probabilities associated with ring expression is perhaps most apparent when reporting and retrieval are constant over the two mortality sources are completely ′ time, then Kt is related to ft by a separated in time. So assume that only proportionality constant, making ft a hunting mortality occurs during the hunting ′ reasonable index of Kt (see Anderson & season and that all non-hunting mortality is Burnham 1976; Burnham & Anderson restricted to the period following the 1984; Burnham et al. 1984). And because hunting season. Then, equation (11) simply non-hunting mortality during the hunting states that the probability of surviving the season is often thought to be small relative year is the product of first surviving hunting ′ to hunting mortality, Kt , and thus ft, are mortality during the hunting season and thought to be reasonable indices to Kt for then surviving non-hunting mortality

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 238 Effects of harvest on waterfowl populations sources during the rest of the year. Because to exhibit more additive versus more survival is a multiplicative process in time, compensatory mortality. Species with with the survival probability for one period relatively “slow” life histories have higher being applied to the survivors of the survival rates and are expected to exhibit previous period, equation (11) should less ability to compensate for hunting losses correspond to intuition. North American than species with faster life histories waterfowl management was based on the characterised by much lower survival rates. additive mortality hypothesis prior to 1976 At a minimum, total annual survival rates (see Geis et al. 1969; Nichols 2000). Finally, and net survival rates from non-hunting note that equation (11) specifies a linear sources impose a constraint on the decrease in total annual survival, St, as net maximum value of a compensation ≤ hunting mortality, Kt, increases (Anderson & threshold, C 1 – S0. Conroy & Krementz Burnham 1976; Nichols et al. 1984; Fig. 5a). (1990) thus noted that a variety of hypotheses about hunting-survival Compensatory mortality hypothesis relationships exists between the two Anderson & Burnham (1976) used equation endpoint hypotheses of additivity and (11) to specify a compensatory mortality complete compensation. They referred to hypothesis, under which β = 0 for some these hypotheses as “partial compensation” ≤ range of values of Kt, specifically for Kt C and noted that they are characterised by where C is a threshold subject to the 0 < β < 1 in equation (11) below some ≤ ≤ inequality, C 1 – S0. So St = S0 for Kt C, threshold C (Fig. 5c). that is changes in kill rate below the Possible mechanisms for compensation. The threshold induce no variation in total family of hypotheses defined by equation survival, which remains equal to net survival (11) thus covers the full range of possible from non-hunting sources only. Because of relationships between hunting mortality and this complete lack of influence of hunting total survival, ranging from complete mortality on total survival, at least for a additivity, to partial compensation to range of hunting mortality rates, this basic complete compensation (Fig. 5). However, a model is sometimes referred to as depicting cost of this flexibility is that the model is “complete compensation” (Conroy & phenomenological, in the sense that it Krementz 1990; Fig. 5b.). Kill rates greater provides no hint of plausible mechanisms than the threshold necessarily result in that might underlie most of the possible declines in total annual survival (linear hypotheses. Additivity, with β = 1, is decline under Anderson & Burnham 1976). consistent with intuition about how Conroy & Krementz (1990) noted that different mortality sources might interact. the diversity of life history characteristics An individual can only die of one source among waterfowl (e.g. with respect to fast- and each death translates to fewer survivors slow and r–K variation) should lead to at the end of any time period (e.g. 1 year). predictions about the degree to which However, hypotheses reflecting some particular species would be expected degree of compensation, β < 1, are not

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 239

1 a. Additive underlying compensatory mortality

S0 (Anderson & Burnham 1976; Nichols 1991). The usual explanation under density- S dependence is that population size at the t end of the hunting season is a determinant of subsequent survival during the portion of the year without hunting. In years where 0 K 1 t hunting mortality is large, abundance at the 1 b. Compensatory end of the hunting season is reduced, and each individual alive at this time has an

S0 increased probability of surviving the rest of the year. In years of low hunting mortality, St abundance at the end of the hunting season is increased, and each individual has a lowered chance of surviving the rest of the 0 C K 1 t year. Although we can fit equation (11) to 1 c. Partially compensatory data that derive from this mechanism, the

S0 actual mechanism is that the probability of surviving the hunting season is 1 – Kt, and

St the magnitude of S0,t (the year-specific probability of surviving non-hunting mortality following the hunting season) now 0 C 1 depends on abundance at the end of the Kt hunting season, and thus on Kt. Figure 5. Additive (a), compensatory (b), and Johnson et al. (1993) suggested a more partially compensatory (c) mortality hypotheses. mechanistic model designed to incorporate S represents the net probability of surviving 0 the above thinking about density-dependent non-hunting mortality sources, which is also the theoretical survival rate in the absence of harvest. non-hunting survival. The model consists of C is the threshold beyond which kill rate K affects the following 2 expressions: ≤ S most strongly (where C 1 – S0). Adapted from E(S t ) = S 0,t (1 K t ) , (13) Conroy & Krementz (1990). and + (1 ) e a bN t K t necessarily intuitive and require some sort 0, = , (14) S t + (1 ) of underlying mechanism. Most of the 1+ e a bN t K t discussions about compensatory mortality where Nt is the population size at the have involved one of two mechanisms, beginning of the hunting season in year t, density-dependence and individual and a and b are parameters specifying the heterogeneity. exact nature of the relationship between Density-dependence is the most the probability of surviving non-hunting frequently cited mechanistic hypothesis mortality and population size. Thus,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 240 Effects of harvest on waterfowl populations equation (13) simply expresses the standard of the hunting season. Despite substantive competing risks relationship that total research on food resources during winter, survival is the product of the probability of clear linkages with subsequent survival surviving two risks, the probability of implications are difficult to discern. surviving hunting mortality (in the complete A second mechanism that can underlie absence of non-hunting mortality) and the compensatory responses (i.e. responses in probability of surviving non-hunting which β < 1 in equation 11) involves mortality (in the complete absence of heterogeneity among individual birds in hunting mortality). Equation (14) then underlying probabilities of surviving both specifies that the probability of surviving hunting and non-hunting mortality sources non-hunting mortality risks is a function of (Johnson et al. 1986, 1988; Nichols 1991; the expected abundance at the end of the Lebreton 2005; Sedinger & Herzog 2012; hunting season, Nt (1 – Kt). We note that Lindberg et al. 2013). Arguments about the strict application of this model with the relevance of individual heterogeneity to above definitions assumes that only hunting animal population dynamics can be traced mortality (no non-hunting mortality) occurs back at least as far as Errington (1943, during the hunting season. However, if 1967), who wrote about predation (one most of the mortality occurring during the mortality source) and the fact that predated hunting season results directly from individuals would likely not have survived ′ hunting, then we can use Kt as an index of other sources had they survived predation. Kt in the above expression. Of course there Writing about Muskrats Ondatra zibethicus are other functional forms than the linear- that suffered predation, he wrote, “…they logistic relationship of equation (14), and usually represented wastage, and, from the any such relationship would be more standpoint of the population biology of the mechanistic than equation (11). species, it did not matter much what befell Characterisation of a model as them” (Errington 1967: 155), and “the phenomenological or mechanistic is of predation is centred upon….what is course subjective, and in reality these terms identifiable as the more biologically apply to regions along a continuum of expendable parts of the population” mechanistic detail. For example, our use of (Errington 1967: 225). Errington’s population size at the end of the hunting arguments about predation could be season as the determinant of the probability relevant to human hunting of waterfowl as of surviving the rest of the year represents a well, if the segment of the population that simplification (see Lebreton 2005). Density- experiences the higher hunting mortality dependence in ecological relationships rate also experiences the higher probability typically involves some resource that is of dying from non-hunting mortality potentially in short supply, such that a more sources (Johnson et al. 1986, 1988; Lebreton mechanistic depiction for equation (14) 2005; Lindberg et al. 2013). With respect would be to substitute for Nt the number of to the above modelling discussion of animals per unit of limiting resource at the end structured populations, hunting mortality is

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 241 largest for the individuals of low quality given that it is alive at the end of the with the lowest reproductive values. hunting season. For simplicity and ease of presentation The average probability of surviving non- assume that a population of ducks is hunting mortality sources for a randomly characterised by heterogeneity of survival selected individual alive at the end of the probabilities that can be approximated as a hunting season is: 2-point (finite) mixture (see “structure”  (1 1 ) section above, page 233), with the groups 1 t K t 0, = S t S 0,t 1 2 labelled as g = 1, 2. For ease of  t (1 K t )+(1 t )(1 K t ) interpretation, also assume that the (1 )(1 2 ) 2 t K t + . (15) anniversary date each year is the beginning S 0,t 1 2  t (1 K )+(1 t )(1 K ) of the hunting season. Only hunting t t mortality occurs during the hunting season, Unlike the average probability of and only non-hunting mortality occurs surviving hunting mortality, this population following the hunting season. This temporal level average includes not only the initial π partitioning of mortality sources is perhaps probabilities of group membership, t and π not a bad approximation for waterfowl and 1 – t, but also the relative probabilities of leads to views of seasonal hunting and non- surviving the hunting season. Thus, the hunting survival and mortality rates as net terms in large parentheses reflect the rates (rates that occur when only the focal expected proportions of the population at mortality source is operating). the end of the hunting season in each of the Define the following parameters: groups 1 and 2, respectively. The key concept in considering the influence of π = probability, at the beginning of year t heterogeneity on population level effects of t, that a randomly selected individual in hunting is that the composition of the the population is a member of group 1, heterogeneous population changes (as g 1 – Kt = probability that a bird in group g reflected in these proportions) over time. If (g = 1 or 2) survives exposure to probabilities of surviving both hunting and hunting mortality during the hunting non-hunting mortality are greater for one season of year t, group than another, then this high survival – group will increase in representation during 1 – K = π (1 – K 1) + (1 – π)(1 – K 2) = t t t t the hunting season. This will lead to a probability that a randomly selected greater average probability of surviving member of the population survives non-hunting sources than if both groups exposure to hunting mortality during had experienced similar hunting mortality the hunting season of year t (mean net rates. hunting survival), and The population in the above example g S0,t = probability that a bird in group g (g consists of two groups of birds, and each = 1 or 2) survives exposure to non- group is characterised by its own hunting mortality sources in year t, probabilities of surviving hunting and non-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 242 Effects of harvest on waterfowl populations hunting mortality sources. If we assume change in composition that occurs during additive mortality within each group of the hunting season. The lower probability of birds, then the probability that a bird in surviving non-hunting mortality sources group g survives exposure to all mortality produces a lower annual survival rate for the sources in year t can be written as: homogeneous population (Table 1). These g g g = (1 ) . (16) differences between heterogeneous and S t S 0,t K t – homogenous populations are attributable to Define S as the probability that an individual t the change in composition of the randomly selected from the population at heterogeneous population (also see Vaupel the beginning of the year survives exposure & Yashin 1985; Johnson et al. 1986), and are to all mortality sources in year t (mean total consistent with the basic mechanism survival at the population level). For the 2- underlying Errington’s (1967) ideas of the group population considered above, we can “doomed surplus”. write this total survival as: The emphasis in this section has been on 1 1 2 2 =  (1 )+(1 ) (1 ) . (17) S t t S 0,t K t t S 0,t K t heterogeneity in survival probabilities, and Consider two hypothetical groups of the example in Table 1 suggests that individuals in a heterogeneous population substantial differences in survival among with corresponding probabilities of individuals do not necessarily produce large surviving hunting and non-hunting differences in total annual survival. Thus, mortality sources given in Table 1. both density-dependence and heterogeneous Individuals of group 1 experience higher survival can mediate the effects of hunting probabilities of surviving both hunting and on populations, but both processes are non-hunting mortality sources than limited in their ability to compensate for individuals of group 2. Indeed annual hunting losses. Heterogeneous vital rates can probabilities of survival, computed using also include reproduction. If the individuals equation (17), are over twice as high for that are better able to survive hunting and individuals of group 1, so the heterogeneity non-hunting mortality sources are also the is substantial (Table 1). Average annual better reproducers, then heterogeneity offers survival at the population level is computed even greater potential for compensatory using equation (16) as 0.53. The final row of effects. Indeed, Lindberg et al. (2013) Table 1 assumes a homogeneous population provided evidence that female Pacific Black in which each individual is characterised by Brant Branta bernicla nigricans exhibit probabilities of surviving hunting and non- heterogeneous survival and recruitment hunting mortality sources computed as probabilities that lead to increased weighted averages of the group-specific population growth rates, relative to growth π π vital rates, with weights t and 1 – t. The of hypothetical homogeneous populations. probability of surviving non-hunting mortality sources differs from the average Evidence for a heterogeneous population, because the Anderson & Burnham (1976) specified the homogeneous population value ignores the two extreme hypotheses, additive mortality

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 243

Table 1. Survival rates in a heterogeneous population comprised of two groups of individuals with different survival rates. Mortality sources are restricted to seasons, with hunting mortality occurring first, followed by non-hunting mortality. Rates include net probabilities of surviving hunting (1–K) and non-hunting (S0) mortality sources, as well as total annual survival (S). Rates are presented for individuals in group 1 and group 2. Average rates are based on the proportions of the populations in each group to which each rate applies. The homogeneous rates correspond to a population in which each individual experiences source-specific survival rates that are simple averages of those for the two groups.

Group (g) Proportion of population 1–KS0 S

1 0.5 0.95 0.75 0.71 2 0.5 0.75 0.45 0.34 Average 0.85 0.62a 0.53 Homogenous 0.85 0.60 0.51 a Conditional on the expected population composition at the beginning of the season during which non-hunting mortality applies (computed using equation 15). and compensatory mortality, and then Burnham (1976), as did efforts to apply analysed extensive ringing and recovery data these various methods to other waterfowl for Mallard in North America to draw species. These methods and results inferences about which hypothesis seemed constitute a substantial literature that has to correspond most closely to these birds. been reviewed periodically (Nichols et al. They took advantage of the ring recovery 1984; USFWS 1988; Nichols 1991; Nichols models that had just been developed (Seber & Johnson 1996). The most recent reviews 1970; Brownie & Robson 1976; Brownie et (Nichols 1991; Nichols & Johnson 1996) al. 1978) to estimate annual survival rates show a mixed bag of results, with a number and hunting mortality (kill) rates, and then to of studies providing evidence favouring the use these estimates with various analytic compensatory mortality hypothesis, some approaches for inference about hunting favouring the additive mortality hypothesis, effects. They concluded that the Mallard and many providing equivocal results. This data largely supported the compensatory uncertainty led Nichols & Johnson (1996) to mortality hypothesis, and that Mallard conclude that an adaptive approach to experienced hunting mortality rates that harvest management would be useful for the were typically below threshold levels. purposes of both managing harvest and A variety of improvements in analytic learning about harvest effects (see the methods for testing these hypotheses adaptive harvest management section, page followed the publication of Anderson & 255 below).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 244 Effects of harvest on waterfowl populations

Pöysä et al. (2004) reviewed previous (Francis et al. 1998). Conroy et al. (2002) studies on Mallard in North America, developed various models to assess habitat contrasting earlier work with the more and density-dependent effects on Black recent efforts by Smith & Reynolds (1992, Duck survival. Model weights indicated also see Sedinger & Rexstad 1994; Smith & support for models that reflected the Reynolds 1994). Based primarily on the additive mortality hypothesis. Rice et al. Smith & Reynolds (1992) inferences from a (2010) found no evidence that Pintail more recent period (1979–1989), Pöysä et al. survival rates varied among groups of (2004) suggested that Mallard populations years characterised by different hunting may have experienced a change in response regulations, but concluded that serious to harvest over time, where hunting effects evaluation of effects of hunting was beyond became additive, at least to some degree. their scope of investigation. Peron et al. Sedinger & Herzog (2012) argued that the (2012) developed an integrated population conclusions of Pöysä et al. (2004) were model for Redhead that used ringing and unwarranted, but Pöysä et al. (2013) noted recovery data as well as information about that despite some relevant points, the abundance from the Waterfowl Breeding criticisms of Sedinger & Herzog (2012) did Population and Habitat Survey and about not cause them to change their conclusions. harvest age and sex ratios from the USFWS Studies of compensatory versus additive Harvest Survey. They found no evidence mortality appearing after the review by that Redhead survival varied in response Nichols & Johnson (1996) include four to either daily bag limit or recovery papers on ducks and several papers on rate, providing some support for the goose species. The life history differences compensatory mortality hypothesis. between ducks and geese (see above) lead to Alisauskas et al. (2011) conducted an the observation that geese tend to have extensive analysis of Lesser Snow Goose higher annual survival probabilities than Chen caerulescens caerulescens population ducks, and thus less potential to compensate responses to increased harvest associated for hunting losses. This observation leads to with special conservation measures the expectation that many goose species will designed to reduce abundances. They found exhibit additive mortality, whereas ducks some evidence of decreased annual survival are more likely to exhibit possible rates associated with the additional harvest compensatory mechanisms. Francis et al. pressure for southern nesting populations, (1998) analysed ringing and recovery data but no such evidence from the much larger for American Black Duck Anas rubripes over northern population segments. However, three groupings of years characterised by Alisauskas et al. (2011) estimated much increasingly restrictive hunting regulations. smaller increases in harvest rates associated They found evidence of increases in with the conservation measures than had survival rates, some consistent with the been hoped. An analysis of the southern La additive mortality hypothesis and some Pérouse Bay population of Lesser Snow smaller than expected under this hypothesis Geese led Koons et al. (2014b) to conclude

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 245 that young females exhibited evidence of history and associated high survival rates. compensation for an early time period, but Studies of ducks have yielded period- and evidence favoured additivity for young species-specific results, with some studies females in later years and adult females for supporting the additive mortality hypothesis the entire period of study. Two analyses of and others the partial and completely ringing and recovery data from G. compensatory mortality hypotheses. Many Gauthier’s long-term study of Greater Snow of the papers reporting these analytic results geese Chen caerulescens atlantica have focussed ended with caveats and recommendations. on effects of hunting on survival. Gauthier The caveats virtually all involved the et al. (2001) provided evidence of additivity typically correlative nature of the efforts for the decade preceding a spring to study effects of hunting, and the conservation harvest designed to reduce acknowledgement that weak inferences abundance. Calvert & Gauthier (2005) are a likely result of this restriction. analysed ring recovery data for the initial The recommendations were for either years of the spring conservation harvest and experimentation or adoption of an adaptive found evidence of decreased survival approach to harvest management as (consistent with additive mortality) for adult potential approaches to yielding stronger Greater Snow Geese but not juveniles. inferences. Adaptive approaches have been Iverson et al. (2014) analysed ring recovery adopted for some species and do permit data for a population of Canada Geese in additional inferences about effects of Ontario. Their analysis stratified individuals hunting (see below). by reproductive status, and they found evidence of additivity for breeding adults, Some methodological challenges but not for non-reproductive birds. Sedinger A variety of approaches exists for drawing et al. (2007) investigated variation in survival inferences about effects of hunting. With and recovery rates of Black Brant over the the development of ring recovery models period 1950–2003 and found evidence of a (Seber 1970; Brownie & Robson 1976; decrease in recovery rate estimates over Brownie et al. 1978, 1985), waterfowl ringing time, from early to more recent decades. programmes now permit estimation of ring These decreases in recovery rates were recovery rates (indices to both crude and – accompanied by an increase in annual for pre-season ringing – net hunting survival rates, until recent decades when mortality rates, see page 236) and total recovery rates became very small. annual survival rates. One straightforward Results of studies summarised in approach to inference about hunting effects previous reviews and the more recent work is to contrast recovery rates and annual cited above provide mixed results. The survival rates for years of differing hunting majority of studies of effects of hunting on regulations. If recovery rates indeed differ geese have provided at least some support as predicted by the regulations changes, then for the additive mortality hypothesis, as an expectation under the additive mortality predicted based on their typically slow life hypothesis is that annual survival rates are

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 246 Effects of harvest on waterfowl populations reduced in years with more liberal hunting under complete compensation. This regulations (e.g. larger daily bag limits or approach was used for Canada Geese in longer seasons). Ontario by Iverson et al. (2014). Another seemingly straightforward This basic approach of drawing approach is to correlate ring recovery rates inferences about effects of hunting by and annual survival rates estimated from ring directly estimating the β of equation (11) or recovery data. However, as noted by by estimating covariation of recovery and Anderson & Burnham (1976) when these annual survival rates is appropriate for rate parameters are estimated from the same situations in which recovery rates are set of ring recovery data, the estimates are reasonable indices to net hunting mortality characterised by non-negligible sampling rates. We noted above that estimates of covariances. Because of these sampling retrieval rates for hunter-killed birds and covariances, simple correlation analyses ring reporting rates are needed to translate using, for example, point estimates of time- ring recovery rates to crude hunting or area-specific survival and recovery mortality rates. Retrieval rates are typically rates will yield inferences that confound assumed to be approximately time-invariant, true process covariation and sampling but these rates have received little study. covariation, yielding correlation statistics that Ring reporting rates have been studied, and cannot be interpreted as pertaining strictly to if these vary over time and/or space, then the underlying mortality processes. they can and should be incorporated directly As a means of dealing with sampling into inferences about hunting mortality variation, Burnham et al. (1984) proposed rates. Reporting rate estimates can be the direct fitting of the equation (11) (and a incorporated into ring recovery analysis as related power function model) to ring constants (with or without sampling recovery data using a deterministic variation), or the raw data (e.g. recoveries ultrastructural model. This approach to from reward-ringed birds) used to estimate inference has been used in several waterfowl reporting rate can be incorporated into the analyses (Barker et al. 1991; Smith & analyses via joint likelihoods that include, Reynolds 1992; Rexstad 1992). Otis & White for example, both standard and reward (2004) developed a random effects approach rings. We also noted that crude hunting to this kind of modelling by considering mortality rates (rates obtained in the recovery rate as a random effect that presence of other (non-hunting) mortality) covaries to some degree with annual survival are most useful for inferences about hunting because of possible effects of harvest on when they are estimated from ringing that survival. This approach permits direct occurs just before the hunting season. When estimation of the process correlation (not ringing occurs at other times of the year (e.g. confounded with sampling correlation) post-season only), resulting data are not as between recovery and survival rates, with a useful for drawing inferences about hunting negative correlation expected under (e.g. see Nichols & Hines 1987) absent additivity and no correlation predicted additional assumptions, as non-hunting

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 247

′ survival becomes a potentially important correlation with St. For example, if Kt is ′ source of variation in estimates of recovery constant, but (1 – S0,t) varies, then rt will be rates and crude hunting mortality rates. larger when non-hunting mortality is smaller A variant (e.g. Sedinger et al. 2010) on this and total survival larger. Consider the approach of investigating the correlation numerical example of Table 2 in which we between ring recovery rates and annual model additive mortality and again make the survival substitutes Seber’s (1970) reporting simplifying assumption that only hunting parameter, rt, for ring recovery rate, ft, where mortality occurs during the hunting season ′ ft = rt (1 – St). Use of a random effects (thus Kt = Kt) and only non-hunting approach, similar to that of Otis & White mortality occurs following the hunting ′ (2004), permits direct estimation of the season (thus 1 – S0,t = (1 – Kt)(1 – S 0,t ); i.e. a process correlation absent any confounding bird must survive hunting mortality in order with sampling variation. However, process to be exposed to non-hunting mortality). We correlations between St and rt predicted consider two years in Table 2, with constant under the compensatory and additive rates of kill, retrieval, and ring reporting for mortality hypotheses are not as both years, and differences only in the net straightforward as those expected for St and probabilities of dying from non-hunting ft. In analyses where ring recoveries are causes, 1 – S0,t. As a result of the variation in restricted to hunting recoveries, we can write non-hunting mortality, both annual survival Seber’s reporting rate parameter as: and the Seber (1970) reporting parameter are larger for the year of lower non-hunting = K t  rt c t t 1 S t mortality (year 2). So the process correlation between rt and annual survival would be K t = c t t . (18) positive and possibly interpreted as evidence +(1 ) K t S 0,t against additive mortality, whereas in reality The term in brackets is the probability additive mortality governed the survival that a bird that died during year t died as a process for both years. Note that there is no result of hunting. The remaining terms are such indication of a positive correlation the probabilities of retrieval and ring between ring recovery rate, ft, and annual reporting defined for equation (12). We have survival. In addition, note that Seber’s already indicated that rates of retrieval and (1970) reporting parameter, rt, while ring reporting can either be estimated or else incorporating what we have termed ring λ are frequently assumed to be constant over reporting rate, t, is a very different quantity time. So the question for selecting ft versus rt (eqn. 18) that attains very different values ′ is whether we prefer Kt (see eqn. 12), or (Table 2). ′ ′ ′ Kt /[Kt + (1 – S0, t)] (eqn. 18), as an index of Another basic approach to inference net hunting mortality rate, Kt. about additive and compensatory mortality The difficulty in using rt is that it is is based on the relationship between potentially influenced by non-hunting net probabilities of experiencing hunting mortality in a manner that leads to a positive and non-hunting mortality, where “net”

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 248 Effects of harvest on waterfowl populations

Table 2. Numerical example illustrating the process correlation between Seber’s (1970) reporting parameter, rt, and total annual survival, St, under the additive mortality hypothesis. In the absence of variation in net hunting mortality, Kt, higher net probabilities of surviving non-hunting mortality, S0,t, lead to higher rt and St. Other definitions: ct = probability that a λ bird shot in year t is retrieved, t = probability that the ring of a bird shot and retrieved in ′ the hunting season of year t is reported, ft = ring recovery rate, 1 – S0,t = crude non-hunting θ mortality rate, and t = probability that a bird which died during year t died as a result of hunting.

′ λ a ′ b θ c d e Year (t) Kt = Kt ct t ft S0,t 1 – S0,t t St rt

1 0.200 0.800 0.750 0.120 0.600 0.320 0.385 0.480 0.231 2 0.200 0.800 0.750 0.120 0.800 0.160 0.556 0.640 0.333

 K   a b c t d f t = K tc t t ; 1 S 0, t = (1 K t )(1 S 0,t );  t =  ; S t = (1 K t )S 0,t ;  K t +(1 S 0, t )  K   e t rt =  c t t =  t c t t  K t +(1 S 0, t ) indicates a mortality probability that would population, leaving relatively more high apply in the absence of any other mortality survival group individuals, and thus greater source (Chiang 1968). Both proposed average non-hunting survival. In summary, mechanisms underlying compensatory both density-dependence and heterogeneity mortality hypotheses suggest a change in lead to the prediction of a negative one net mortality probability (non-hunting) correlation between the net survival as a function of variation in another net probabilities associated with the two mortality probability (hunting). Density- mortality sources. Specifically, lower hunting dependence results in lower net non- mortality is associated with higher non- hunting survival in years where hunting hunting mortality and vice versa. mortality is low, as many birds are alive at the If only hunting mortality occurs during end of the hunting season (other things the waterfowl hunting season and only non- being equal). Competition for resources is hunting mortality occurs after the hunting then hypothesised to result in lower net season, then this temporal separation non-hunting survival of the survivors, than permits direct estimation of both net if hunting mortality had been larger. Under mortality rates via a ringing programme that the heterogeneity hypothesis, the larger the includes ringing at two times of the year, the hunting mortality rate the greater the change beginning and end of the hunting season. in composition of the heterogeneous Inference based on a single ringing period

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 249 each year is more difficult, however. Schaub changing net rates. This was noted by & Pradel (2004) and Schaub & Lebreton Schaub & Lebreton (2004) and Servanty et (2004) used multistate models with recovery al. (2010) who refer to this component data based on annual ringing to estimate of process correlation as an “intrinsic bias” separate survival rates associated with two (Peron 2013 labelled it “competition bias”). different mortality sources. However, it is To demonstrate this natural negative important to recognise that these are correlation between the two crude rates “crude” rates (sensu Chiang 1968), as they are under the additive mortality hypothesis (i.e., conditional on the deaths occurring as a in the absence of any correlation between result of the other mortality source. Above net rates), consider the following example of ′ the crude hunting mortality rate, Kt , is temporally separated hunting and non- defined as the probability that a bird alive at hunting mortality. Assume that hunting the beginning of the hunting season in year mortality occurs first, and that this mortality ′ t would die from hunting causes during the is larger in year t1 than in year t2, Kt1 = K t1 ′ hunting season of year t in the presence of > Kt2 = K t2 (note that because of the timing the non-hunting mortality that occurs of mortality, these are both crude and net during this period. Similarly, the crude rates). Assume that net non-hunting survival ′ mortality rate (1 – S0,t) could be defined as is the same in the 2 years; that is, survivors the probability that a bird alive at the of the hunting season have equal chances of beginning of the hunting season in year t surviving non-hunting mortality in the two would die from non-hunting causes years, S0,t1 = S0,t2. Thus there is no throughout the year t in the presence of the association between the net rates associated hunting mortality that occurred earlier in the with non-hunting and hunting mortality in year. the two years. However, because of the There are two important difficulties to greater hunting mortality in year 1, the crude consider when using estimates of these non-hunting mortality rate will be smaller in ′ ′ crude rates, Kt and 1 – S0,t, to draw year 1 than in year 2: inferences about effects of hunting. The (1 S 0, t 1 ) = (1 S 0,t 1 )(1 K t 1 ) < (1 S 0, t 2 ) first difficulty is similar to that noted above ( )( )( )( ) = (1 )(1 ) . for direct inferences about the relationship S 0,t 2 K t 2 between annual survival and recovery rates, Smaller numbers of birds are available to and involves the sampling covariance be exposed to non-hunting mortality between time-specific estimates of these sources in year 1; hence fewer die from these crude rates. These can be dealt with using sources, despite equal net non-hunting rates either frequentist or Bayesian approaches in the 2 years. So the crude rates covary (see Schaub & Lebreton 2004). The other negatively, but this covariance is induced by difficulty is that these crude rates of hunting variation in only one net rate and has and non-hunting mortality are expected to nothing to do with the covariance between exhibit negative process covariation for net rates that underlies compensatory reasons that have nothing to do with mortality.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 250 Effects of harvest on waterfowl populations

Although mortality sources were throughout the year. Even though she was separated in time for our example, for ease able to separately estimate these annual risks of presentation, we note that the issue without having to guess at their magnitude, remains regardless of the timing of source- she noted both the restrictive nature of her specific rates. Even when non-hunting and assumptions about temporal constancy of hunting mortality occur simultaneously, it is risks and the high sampling covariance still true that more birds removed by one between the estimates of risk, and source (e.g. hunting) will leave fewer birds abandoned this approach to inference about available to die of the competing source effects of hunting. (non-hunting). Schaub & Lebreton (2004: Finally, we note that radio-telemetry data 83) noted this problem when attempting to are well suited to estimate net source-specific draw inferences about compensation or mortality rates (the rates from which strong additivity using source-specific mortality inferences about compensation can be most data, but their important caveat does not readily obtained) directly. Time-specific seem to be appreciated by all others who deaths from sources other than a focal source have used this approach. Servanty et al. can be immediately censored, providing (2010) dealt with this issue by eliciting direct inference about the net mortality rate expert opinion about the timing and of the focal source (Heisey & Fuller 1985; magnitude of non-hunting mortality, and Heisey & Patterson 2006). Sandercock et al. incorporating these opinions using a (2011) provide a nice example of this Bayesian inference framework. In summary, approach to inference about compensation we emphasise that attempts to use estimates using a non-waterfowl species. of crude source-specific mortality rates (based on Schaub & Pradel 2004; Schaub & Reproduction Lebreton 2004) to draw inferences about While assessment of the impact of harvest hunting effects require careful analysis and on waterfowl dynamics has generally (and interpretation. intuitively) focussed on the direct It is interesting that in work that is closely relationship between harvest and survival related in some ways to that of Servanty et (preceding section), harvest can potentially al. (2010), but carried out 35 years previous, drive population dynamics in other ways, Brownie (1974) developed an approach to through the indirect influence of removal estimate the instantaneous risks (these of individuals on other components of translate directly into net mortality rates) fitness. In this and the following section, we associated with hunting and non-hunting briefly consider the effects of harvest using the extra information about the date on components of reproduction, and on of recovery of each hunting season ring migration and movements. recovery. She was able to estimate these risks Harvest can potentially influence directly, but had to assume that each was reproduction in several ways. First, and time-constant; the hunting risk throughout perhaps most obviously, harvest clearly the hunting season and the non-hunting risk removes both the immediate and future

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 251

(residual) contributions of a harvested decisions involving annual harvest individual to population growth. However, regulations. For example, a reduction in total the typical accounting of the reproductive number of breeding individuals may not value of an individual, reflecting current and necessarily result in increased production at future reproductive contributions lost to the population level, if the harvested harvest (sensu MacArthur 1960), assumes individuals are lower quality birds which do that such individual contributions to not contribute significantly to annual population growth represent independent production. Uncertainty about the events (i.e. that removal of one individual functional form relating production to does not change the reproductive value of changes in abundance or density can be any other individual). However, in many particularly important (Kokko 2001; Runge cases per capita reproduction is influenced & Johnson 2002). Similarly, the influence of by the number (abundance or density) and breeding population size on various structure (age, sex, spatial) of conspecifics. components of post-laying fitness (i.e. Density-dependence in population growth events that might occur following the has been demonstrated at large spatial and primary production of the clutch) may be temporal scales (Viljugrein et al. 2005; more difficult to predict. For example, Sæther et al. 2008; Murray et al. 2010), increased density of nesting birds may and for many species of waterfowl increase nest survival (Ringleman et al. (especially shorter lived ducks), variation in 2014), but might lead to increased reproductive output is a dominant driver of competition among juveniles for limiting annual population dynamics. Thus, it is resources, leading to reduced juvenile reasonable to assume that any activity which growth and survival (Cooch et al. 1991; reduces “density” (e.g. harvest) has the Sedinger et al. 1995). It is probably true that potential to increase population growth, by for many species, the impact of population increasing reproductive productivity. abundance (density) on both pre- and However, this conclusion is arguably post-laying components of reproductive overly simplistic, in at least a couple of fitness will reflect a complex interaction of respects. First, as noted by Lebreton (2009; both frequency- and density-dependent also see above), estimates of density or effects, strongly influenced by various population size are only a proxy for what environmental effects which can affects demographic performance – a significantly modify the relationships correlation between density or population (Lebreton 2009). size and reproductive performance is Second, harvest not only has the potential phenomenological, and does not generally to change the size of the population, but its indicate the important mechanisms structure (age, sex, spatial, heterogeneity) as underlying the observed relationship. This well. Any harvest which changes the increases the uncertainty in projecting structure of a population will influence the impacts of harvest on population reproductive output, if those structural production, which is an important factor in elements themselves influence one or more

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 252 Effects of harvest on waterfowl populations components of reproductive fitness, breeding, whether an individual bred at all independently or in addition to potential that year, the timing and pattern of density-dependent effects. For example, for migration, number of offspring produced, species with polygynous mating systems, or physiological condition following differential harvest of males and females breeding). might result in a skew to the sex-ratio in the breeding population, which could Evidence potentially influence the probability of any Much of the preceding is couched in terms individual female laying a clutch, or the of “potential impacts” of harvest on probability of that clutch surviving to reproductive output, at either the individual fledging (in species where males and females or population level. In general, there seems play different roles in nest guarding and to us to be a lack of consideration of the brood rearing), or both. Or, for species processes underlying the relationship where breeding propensity may be based on between harvest and reproduction in relative proportions of individuals of waterfowl, beyond the obvious and logically different age or breeding experience at the trivial observation that a harvested start of breeding, differences in harvest individual has no future reproductive vulnerability between younger and older potential. Even seemingly simple (yet quite birds could potentially be strong drivers for important) relationships between (say) annual variation in the proportion of harvest vulnerability and reproductive individuals breeding, which in turn would output are poorly quantified. For example, have a strong impact on population despite the long-held belief amongst many production. Changes to the structure of the goose biologists that adults with young will population by harvest are not unexpected, have different vulnerability to harvest than since harvest is generally not random with adults without young, there are few rigorous respect to individual contributions to attempts to quantify this relationship. population growth. This non-randomness The relative paucity of empirical studies can be intentional (say, for example, due to a to date on the role of harvest on male only season or greater bag limit for reproduction likely reflects several factors. males than females, or a region- or season- First, there has been a general tendency to specific regulatory programme which allows focus on impacts of harvest on survival, for different harvest as a function of since: (i) the relationship of harvest and location and time of year), or an artefact survival is potentially (and presumed to be) of the interaction of non-specific (i.e. more accessible to management action, and presumed random) harvest with structural (ii) the impacts of survival on overall differences in susceptibility to harvest. Such population growth are often higher than differences could reflect differences in possible impacts on reproduction, and thus vulnerability due to heterogeneity in are arguably more important to quantify. individual reproductive performance (e.g. as This is especially true for longer-lived a function of differences in timing of species, where it has been shown repeatedly

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 253

(to the point of redundancy, perhaps) that to this general difficulty in studying nesting projected population growth is more ducks are provided by cavity-nesting species, sensitive to changes in adult survival than which can attain high nesting densities and any other single demographic parameter (in exhibit high philopatry relative to most fact, this result is a logical necessity for any ground-nesting ducks. In Wood Ducks Aix species with a generation length greater than sponsa, for example, high nesting densities ~4–5 years; Lebreton & Clobert 1991; have been associated with reductions in Caswell 2001; Niel & Lebreton 2005). reproductive parameters such as breeding Second, evaluating the impacts of harvest probability, nesting and hatching success, as on components of reproductive output (e.g. well as with increased nest abandonment egg laying, nesting success, breeding (Haramis & Thompson 1985; review in proportions, recruitment) or population Nichols & Johnson 1990). With this structure (e.g. sex ratio, spatial distribution) exception of cavity-nesters, most of the often requires intensive data from breeding empirical tests of the effect of density on ground studies, with adequate samples components of reproductive rate for ducks of marked, known-aged individuals for relate to nest survival (e.g. Prop & Quinn estimating some parameters. At present, 2003; Ringelman et al. 2014 and references there are a number of such studies involving therein). However, large-scale aggregate breeding populations of geese, where high measures of reproductive rate (e.g. age ratios density nesting and strong natal philopatry in autumn) have been related to population lend themselves to collection of extensive, size and density in prairie-nesting Mallard of detailed demographic data. For example, North America, providing some evidence of Sedinger & Nicolai (2011) and Lindberg negative density-dependence (see Anderson et al. (2013) have provided compelling 1975; Brown et al. 1976; Kaminski & evidence from their long-term study of Gluesing 1987; Johnson et al. 1997; also see Pacific Black Brant that harvest has both page 260). direct and indirect impacts on reproduction. Even when a relationship between Several near-replicate studies of a number of harvest and one or more components different goose species have clearly shown of reproductive performance has been the negative impacts of density on clutch established, the larger consequences on the size (Cooch et al. 1989; Sedinger et al. 1998) dynamics (and management) of the and post-hatch growth and survival and population have not generally been recruitment of goslings (Cooch et al. 1991; considered. For example, while there have Sedinger et al. 1995, 1998; Lepage et al. 1999). been a number of studies of the role of mate In contrast, there are relatively few loss on reproductive success for several comparable studies involving breeding waterfowl species (Cooke et al. 1981; Martin ducks, where natal dispersal and difficulty in et al. 1985; Forslund & Larsson 1991; capturing and marking broods makes Manlove et al.1998; Lercel et al. 1999; Hario et analysis of variation of many reproductive al. 2002), to our knowledge, there has been parameters much more difficult. Exceptions no rigorous assessment of the role of sex-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 254 Effects of harvest on waterfowl populations specific differences in harvest vulnerability reared Mallard in Legagneux et al. 2009). on population dynamics in waterfowl, However, protected areas may be attractive although such studies are quite common for to birds not only because they are free other taxa (see Milner et al. 2007). Similarly, a from hunting, but also because they number of studies have suggested that lower receive specific waterfowl-friendly habitat quality individuals (based on various criteria, management. Several studies nevertheless such as body condition) are likely more have demonstrated that local redistribution of vulnerable to harvest (Hepp et al. 1986; waterfowl towards protected areas was Dufour et al. 1993; Heitmeyer et al. 1993; genuinely linked with hunting. For example, Pace & Afton 1999). Such non-random Madsen (1998b) established experimental selection clearly has the potential to alter the reserves within a Danish fjord, which were structure and production of the breeding moved from year to year. He documented population. However, there has been little concomitant changes in the annual consideration of such selective harvesting distribution of hunted waterfowl species, and resulting changes in population structure which matched the movement of the on overall dynamics and management – the reserves over time. Protected species recent paper by Lindberg et al. (2013) is an conversely did not adjust their distribution to important exception. that of the reserves. Within a given year, Cox & Afton (1997) also recorded redistribution Migration/movement of female Northern Pintail depending on Rates of movement are often ignored in whether hunting was or was not occurring: discussions of population dynamics, yet the ducks increased their diurnal use of they are important vital rates that bring protected areas during two successive about changes in population size at various hunting periods, while this use decreased scales. As with the other components of during the periods pre-hunting, post-hunting, their population dynamics, there is and during the time between the two split convincing evidence that hunting can affect hunting periods. It should be noted, though, the movement rates of waterfowl. that Link et al. (2011) did not obtain the same Escape flights of waterfowl facing result in Mallard, which may indicate hunters are a common field observation and differential susceptibility of the different lead to local redistribution of the birds species to hunting disturbance and/or towards hunting-free refuges (Béchet et al. differences in habitat selection processes 2004). Ample demonstrations have been related with differential food availability. published of local increases in waterfowl Beyond such local scale effects, hunting numbers after reserve creation (Bellrose has also been shown to affect waterfowl 1954; Madsen & Fox 1995; Fox & Madsen movement at the regional level (review in 1997; Madsen 1998a). Some authors Madsen & Fox 1995). Ebbinge (1991) demonstrated that such local movements documented changes in the number of geese were due to greater emigration rates from at the national scale in some countries after areas with more hunting pressure (e.g. hand- hunting was banned or, conversely,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 255 reintroduced in neighbouring countries. To our knowledge, the only species in Madsen & Jepsen (1992) also reported that which the consequences of hunting for both increased shooting pressure (together individual movement and population with changes in farming practices) led to dynamics have been demonstrated is the an earlier departure of the Pink-footed Greater Snow Goose. Spring hunting of this Goose population from Denmark to the species at a migration area was introduced as Netherlands during autumn migration. In a means to reduce the population and the ducks, Väänänen (2001) showed that Finnish problems caused to northern breeding Anas species move towards protected areas habitats by these superabundant birds. The at the onset of the hunting season, and that implementation of spring hunting changed their numbers also decline at the regional the regional migration movements of these scale after this date, which may indicate that birds, with fewer unidirectional movements hunting precipitates fall migration in these and more westward and reverse movements species. This is consistent with Cox & Afton (Béchet et al. 2003). These changed (2000) recording greater emigration rates of movement patterns, a greater proportion of female Northern Pintail from Louisiana time spent alert and in flight, and greater use (largely towards other regions further north of less profitable habitats, resulted in poorer in the alluvial valley of the Mississippi River) body condition of the geese after the during hunting periods than before, between initiation of the spring hunts (Béchet et al. or after these periods. 2004). Lower breeding output was in turn The potential effect of hunting on eventually recorded (Mainguy et al. 2002). waterfowl movement rates is therefore Such an effect of hunting for population clearly established. However, whether this dynamics via changes in movement rates is later translates into lower survival or also possible in other waterfowl species, but breeding output at the population level is has yet to be demonstrated. not so clear. Gill & Sutherland (2000) and Gill et al. (2001) have already commented on Adaptive harvest management the difficulty to predict the consequences of Overview such disturbance at the population level based on observations of local responses of As noted in the above historical review, individuals (either through change in local AHM was implemented by the USFWS for time-activity budgets or in movement rates). mid-continent Mallard in 1995 as a means of Indeed, the local behaviour of individuals is simultaneously managing in the face of dependent upon their respective abilities to uncertain harvest–population relationships respond locally (i.e. based on individual and reducing this uncertainty. AHM body condition, and/or the availability of programmes have now been implemented alternative sites to the disturbed area), and for other Mallard populations as well as for the consequent effect at the population level other waterfowl species (USFWS 2013). may largely depend upon density-dependent Here we briefly describe AHM components processes. and the AHM process and then review what

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 256 Effects of harvest on waterfowl populations has been learned through its application to habitat management, for example), current different Mallard populations. deliberations are considering the “Adaptive management” has taken on modification of existing AHM programmes many different meanings (Williams et al. to incorporate additional objectives and 2009; Williams & Brown 2012), and our additional kinds of actions (Runge et al. focus here will be on the passive process 2006; Osnas et al. 2014). used by the USFWS in their various AHM Models are required in order to predict programmes. This process requires some population responses to management key elements: objectives, potential actions, actions. Optimisation essentially entails a model(s), a monitoring programme and a comparison of these predictions in order to decision algorithm. Harvest management select the action that is expected to do objectives are expressed as an objective the best job of achieving management function which typically is to maximise objectives. Models used in AHM are not average annual harvest over a long time simply models of waterfowl population horizon. The time horizon places value on processes, but are tailored to the specific harvests from future waterfowl populations purpose of predicting the consequences of and thus serves to promote conservation. actions to population change. Frequently, The potential actions are typically sets of there is substantial uncertainty about regulations packages that specify season population responses and associated lengths and daily bag limits. At least one models, and we refer to this as structural package is relatively restrictive (short uncertainty. We attempt to incorporate this seasons and small bag limits), at least one is uncertainty into the models in one of two relatively liberal (long seasons and large ways. The most commonly used approach to bags), and at least one package is moderate, date has been the use of a discrete model set where ideas about liberal and restrictive are containing multiple models of system typically population- and species-specific. responses. Relative degrees of belief, or An alternative approach to discrete model weights, are associated with each regulations packages would be to specify model. These weights determine the a target harvest rate, treating this rate as influence of each model in the optimisation a continuous control variable in the and are updated through time based on a optimisation. However, because the comparison of model-based predictions and regulations themselves are the actions observations from monitoring programmes. selected by the USFWS, and because there The evolution of model weights over time is substantial uncertainty about the reflects our learning. The other approach to relationship between regulations and dealing with model uncertainty is to employ harvest rate (partial controllability), the a very general model and express the USFWS has focussed on these discrete sets uncertainty as the variance of a key model of regulations. In an effort to integrate parameter. In this case also, monitoring data the various potential actions available to are used to update estimates of the manage waterfowl populations (including parameter and hopefully reduce its variance.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 257

Monitoring programmes provide data that adaptive management can still be that play several roles in the management carried out, and learning can still occur, even process. They provide estimates of system when sub-optimal actions are taken. state (e.g. population size) in order to Programmes of adaptive management make state-dependent decisions. State- typically require a deliberative or set-up dependence simply refers to the fact that phase during which the above components we will typically take very different are all developed (e.g. Williams et al. 2007). At management actions when our focal the first decision point, the optimisation population far exceeds a desired level as algorithm is used to produce optimal state- when our population is below the desired dependent actions, based on the objectives, level. Estimates of system state are also used the available actions, and the model(s). to see how well we are doing at meeting These recommended actions are sometimes management objectives. Estimates of referred to as a policy matrix or decision system state are essential to the process of matrix. The current state of the system is learning, in providing an estimate of truth estimated via the monitoring programme against which to compare our model-based and these estimates are used with the predictions, thus providing a way to update decision matrix to select the appropriate model weights (discrete model set) or better action for that decision point. The action is estimate key model parameters that taken and drives the system to a new state, characterise our uncertainty. which is estimated by the monitoring Finally, some sort of decision algorithm programme. In the case of multiple models, must use these various components to the new estimate of system state is determine an optimal action for each compared with the predictions of the possible system state. Harvest management different models, and the degrees of belief programmes are typically recurrent decision or weights associated with the models are problems. For example, we usually make modified using Bayes formula (see Williams decisions about hunting regulations et al. 2002). In the case of a single model annually. Because optimal actions are state- characterised by uncertainty in one or more specific and because an action taken at time key parameters, these parameter estimates t influences the state of the system, and thus are updated with the new data. This the optimal decision, at time t + 1, the completes the first step in the adaptive optimal decision at time t depends on management process. projections of future system states and At this point, the process enters the decisions. Dynamic optimisation algorithms iterative phase, and the next step is to are thus needed, with stochastic dynamic consider the action to take at the next programming (Bellman 1957; Williams et al. decision point. In the case of “passive” 2002) being the method of choice for adaptive management (Nichols & Williams problems for which dimension is not 2013), which has been used for most North too large. Although optimisation is American waterfowl management, the recommended for decision making, we note optimisation algorithm is run with the new

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 258 Effects of harvest on waterfowl populations model weights or updated parameter following two sections, we briefly review the estimates. This produces a new decision competing models used in these two matrix characterising the updated state of programmes and show the evolution of knowledge of system dynamics. This matrix support (or lack of it) for these models. is used with the estimate of current system state to select the optimal action for the Mid-continent and eastern Mallard upcoming decision point. This action is then The mid-continent Mallard stock includes taken, the system is driven to a new state birds breeding in the traditional survey area identified by monitoring, information about of the waterfowl breeding population and that new state is used to update model habitat survey (WBPHS) strata 13–18, weights and/or parameter estimates, and the 20–50, and 75–77, plus birds observed in process proceeds iteratively in this manner. the states of Michigan, Minnesota, and In the case of waterfowl harvest, the Wisconsin (Fig. 6; USFWS 2013). Harvest decision points occur annually. After some policies derived for the Mississippi and experience (e.g. several years) with the Central Flyways are based on this population. process, there may be cause to reconsider The eastern Mallard stock includes birds some of its components. This has been breeding in the WBPHS strata 51–54 and 56 referred to as double-loop learning in the provinces of Ontario and Quebec plus (Williams et al. 2007) and entails moving birds breeding in the eastern states of back into the deliberative phase to revisit Virginia northward into New Hampshire any of the process components, from which are monitored through the Atlantic objectives through models and monitoring. Flyway Breeding Waterfowl Survey. Harvest Indeed, this process of revisiting policies derived for the Atlantic Flyway are components is occurring now for some of based on this population. the processes developed for adaptive For each Mallard stock, a set of four harvest management of North American models representing different combinations waterfowl, with special attention focussed of recruitment and survival relationships are on objectives and actions (e.g. incorporation used to represent structural uncertainty and of habitat management into decisions; predict Mallard population responses to Runge et al. 2006). environmental changes and harvest This basic process has been used to regulations. For mid-continent Mallard, establish annual hunting regulations for strong and weak density-dependent mid-continent Mallard for nearly 19 years relationships are used to predict recruitment and for eastern Mallard for 14 years. As as a function of breeding population size these are the longest-running programmes and the number of Canadian ponds for North American waterfowl harvest observed in the WBPHS (Fig. 7); while management, they provide the best the eastern Mallard strong and weak opportunities to learn about effects of recruitment models are based on a non- harvest and to begin to discriminate among linear relationship that predicts annual competing models of such effects. In the recruitment as a function of breeding

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 259

Figure 6. Distribution of eastern, mid-continent, and western Mallard stocks in North America, as described in the USFWS’s adaptive harvest management programme. population size only. Each model set also strong density-dependent recruitment includes two survival sub-models to (SaRs); compensatory hunting mortality and represent uncertainty in the relationship weak density-dependent recruitment (ScRw); between harvest mortality and survival. additive hunting mortality and weak These models are based on an ultra- density-dependent recruitment (SaRw); and structural formulation (see eqn. 11) that compensatory hunting mortality and strong predicts annual survival as a function of the density-dependent recruitment (ScRs). The survival rate in the absence of harvest and mid-continent and eastern Mallard model the kill rate expected under a particular set sets were last updated in 2002 (Runge et al. of harvest regulations. Under the additive 2002) and 2012 (USFWS 2013), respectively. model, survival rates decline linearly with Under current AHM protocols, the increasing harvest rates, while under the relative belief we have in each model is compensatory model, survival rates remain quantified with an individual model weight. unchanged until a threshold harvest rate Because these weights sum to 1, they serve as ≤ (C 1 – S0) has been exceeded and then individual measures of relative credibility. decline linearly with increasing kill rates When AHM programmes were first (see Fig. 7). Combining the recruitment implemented for these two populations, and survival sub-models results in four individual prior model weights were set equal models: additive hunting mortality and (0.25), reflecting equal confidence in the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 260 Effects of harvest on waterfowl populations

Reproduction Survival

2.0 1.0 Weak density dependence Additive Strong density dependence Compensatory 0.8

1.5 0.6

A g e ratio 0.4

1.0 Male s urvival rate

0.2

0.5 3.36 Million ponds 0.0

02468 10 0.0 0.2 0.4 0.6 0.8 1.0

Breeding population in millions Kill rate

Figure 7. Reproductive and survival sub-models used to represent different relationships for recruitment and male survival in the mid-continent Mallard model set. Predicted levels of recruitment assume 3.36 million Canadian ponds. ability of each model to predict Mallard The evolution of model weights over population sizes. As observed population time resolves structural uncertainty and estimates were compared to individual represents learning in adaptive management model predictions, model weights and our (Williams et al. 2002). Current model relative beliefs in each model have been weights suggest strong evidence for weak updated with Bayes theorem. For mid- density-dependent reproduction (96% and continent Mallard, model weights remained 68%) versus strong density-dependent essentially unchanged until 1999, when each reproduction (4% and 32%) in the mid- model predicted a population decline in the continent and eastern Mallard populations, face of a significant population increase (Fig. respectively. For both the mid-continent and 8). Because the weak density-dependent eastern Mallard stocks, model weights favor model predictions were closer to the the additive mortality model (67% and 70%) observed population estimates, they gained compared to the compensatory mortality more weight and credibility. Since 1999, model (33% and 30%), suggesting support models SaRw and ScRw (weak density- for the additive harvest mortality hypothesis. dependent recruitment) have continued As the number of comparisons of model to gain credibility as weights for models predictions to observed population estimates SaRs and ScRs (strong density-dependent has increased over time, evidence has recruitment) have declined. For eastern accumulated indicating that predictions from Mallard, changes in model weights have been model SaRw are more reliable compared to slower, with the additive mortality weak other predictions from the model set. density-dependent model (SaRw) gradually However, it must be recognised that accumulating more weight over time (Fig. 8). conclusions based on an interpretation of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 261

A 9 10 8 67 Mid-continent population

B Model wei g ht s 0.0 0.2 0.4 0.6 C

Observed Model SaRw Model ScRw Model SaRs 0.9 1.0 1.1 1.2 Model ScRs 8 Model average E a s tern population 0.7 0.

D Model wei g ht s 0.0 0.2 0.4

1995 2000 2005 2010 Year Figure 8. Panels A and C: population estimates of mid-continent and eastern Mallards (in millions) compared to predictions of each member of the mid-continent and eastern Mallard model set (SaRw = additive mortality and weakly density-dependent reproduction, ScRw = compensatory mortality and weakly density-dependent reproduction, SaRs = additive mortality and strongly density-dependent reproduction, ScRs = compensatory mortality and strongly density-dependent reproduction). The grey shading represents 95% confidence intervals for observed population estimates. For each model set, the arrow represents a weighted mean annual prediction. Panels B and D: annual changes in model weights for each member of the mid-continent and eastern Mallard model sets; weights were assumed to be equal in 1995 and 2002, respectively. For the eastern Mallard population, model weights were not updated in 2013 because breeding population estimates were not available. relative credibility measures (model weights) spurious relationship between harvest and are conditional on the set of hypotheses annual survival rates because density- represented in the entire model set and that dependent mortality is not explicitly ecological interpretations are limited by the considered in the AHM model set. They dependency of regulations on population base part of their argument on results from status (Johnson et al. 2002; Sedinger & Conn & Kendall (2004), who demonstrated Herzog 2012). Recently, Sedinger & Herzog through simulation, that the model weight (2012) questioned whether current AHM updating procedures used in AHM may model weights provide support for the result in model weights that support additive mortality hypothesis, suggesting the additive mortality hypothesis in cases that this outcome may result from a where the true underlying dynamics were

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 262 Effects of harvest on waterfowl populations generated with a density-dependent, efforts are underway to evaluate the sub- compensatory survival model. We note that model performance of the eastern and mid- the mid-continent Mallard model set and the continent Mallard AHM protocols through prediction variance estimates that governed the double-loop learning process. The the model weight updating procedures continued successful application of adaptive simulated by Conn & Kendall (2004) were harvest management requires that model completely revamped during the 2002 sets are updated with the most recent revisions (for details see Runge et al. 2002). information to ensure that key demographic Simulation experiments using the current relationships are properly represented and a AHM protocols have verified that model full suite of population responses to weights converge on the correct model when management actions is considered. the true generating model is included in the AHM model set (G.S. Boomer, unpubl. data). Summary and conclusions We acknowledge that the true A defensible argument could be made that relationship between harvest and survival there has been and continues to be more may not be represented well by the interest (and pages published) related to current model set and also note that harvest of biotic resources than perhaps any model predictions may be consistent other subject in biology (studies related to with observations without accurately disease dynamics being the only likely representing the demographic relationships competitor). Our intent in this review was to that determine population dynamics provide a reasonably complete review of (Johnson et al. 2002). The current some of the current and historical interest in formulations of survival sub-models were harvest, and harvest management, as included in the AHM model set to account pertains to waterfowl. However, beyond the for uncertainty in the relationship between usual difficulties of reviewing such a large harvest mortality and annual survival, literature, we were faced with the additional specifying endpoints on a range of possible challenge presented by the fact that “ducks responses to harvest mortality. Certainly, an aren’t geese, and geese aren’t ducks…” (C.D. alternative harvest-survival model may Ankney, pers. comm.). We have proposed perform better than each survival model in that differences in life histories between the current model set. The consideration of most duck and goose species (reviewed alternative beliefs in decision making is a briefly at the start of this paper) result in hallmark of adaptive management, which important differences in estimation and provides a rigorous process to evaluate the management of the impacts of harvest on ability of alternative models to predict the their respective population dynamics. consequences of management actions. The Perhaps the most obvious and important current implementation of AHM provides difference between modelling goose and an ideal framework to consider alternative duck dynamics (with or without harvest) is models describing population responses to the presence of significant “age structure” harvest management. We note that current for goose species, and several sea ducks.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 263

Specifically, for geese, we need to adequately often estimated annually, modelling and account for significant age-specific management of structured models also differences in survival and fertility in the ideally requires an estimate of the structural model. While the presence of greater composition of the population at each time degrees of age-structure for geese (and step. In cases where harvest can be selective probably most sea ducks) is important, all of individuals in different structural classes, species of waterfowl are likely “structured” an optimal harvest strategy may prescribe a to some degree, independent of possible structured harvest (e.g. so many big ones, so age differences. For example, differences in many small ones, etc.), with the actual survival or reproductive output as a function structure of the harvest depending on of location, or amongst individuals of harvest objectives. differing “quality”, are both forms of However, for waterfowl species, we are “structure” which are likely common to usually limited in both our ability to most waterfowl populations. We note that characterise structure at the time of harvest recent work by F.A. Johnson (pers. comm.) (i.e. structure is often only partially has extended this idea to multispecies observable), and to select the age or stage of management, where the structural the individuals that we harvest (i.e. harvest of component involves relative proportions of specific classes of individuals is often only different species, and how one or more under partial control). More often than not, species can be managed when harvest might actual harvest is a function of an interaction be in part a function of the structure and of: (i) the harvest regulatory option(s), and dynamics of the focal species. When (ii) the relative vulnerability to harvest of populations are structured, the harvest different classes of individuals in the required to achieve a particular management population (young, old, male, female, etc.). In objective can be described as a vector principle, adaptive harvest management (specifying the number or proportion of approaches can be applied here, since they each structural class in the harvest), the can explicitly account for such uncertainties. elements of which are determined by: (i) the The larger technical challenge with applying number of structural classes, and (ii) the AHM to structured populations is that the reproductive value vector of individuals in state of the population at the time of the each of those structural classes at the time annual harvest decision is only partially of harvest. The inclusion of structure adds observable in many cases, and may need extra dimensions of uncertainty to harvest to be reconstructed in some form. management. Most obvious is the need for Alternatively, you might accept that you are estimation of an increased number of harvesting an unknown mixture, and that demographic rates and functional forms your expected returns will likely be reduced relating one or more vital rates to various because of this uncertainty. To date, there is intrinsic (population density, population limited experience with application of AHM structure, etc.) and extrinsic (e.g. climate to such partially observable structured variables) factors. While population size is systems.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 264 Effects of harvest on waterfowl populations

The differences between ducks and geese (which typically is one of several criteria have important ramifications beyond the by which a species might be considered as structure of population models. Of “over abundant”). Moreover, even if duck particular interest in the management populations increased significantly relative to context is that the management objectives historical numbers, it is difficult for most can and occasionally do differ significantly biologists and managers to imagine duck between ducks and geese. For most ducks, populations causing major negative impacts the management objective can be stated to the environment (save, perhaps, for simply as “we want enough birds in the the increased likelihood of disease with population to keep it viable, in the presence increased density). In contrast, many goose of sustained mortality impacts from sport species, primarily because of the significant harvest”. While this is generally also true for differences they exhibit in foraging geese, for some populations of some goose behaviour and bill morphology, and little species, there is the unique additional evidence they have approached their overall objective relating to limiting or reducing the carrying capacity, have been increasingly size of the population. For example, light observed to have strongly negative impacts geese (primarily Lesser Snow Geese and on their habitat (both winter and breeding), Ross’s Geese Chen rossii) in several parts lead to detrimental collateral impacts on of North America are demonstrably other wildlife species, and in some cases overabundant (Ankney 1996), and the significant economic liability for agricultural current management objective for these crops (Abraham & Jefferies 1997; Batt 1997, populations is to reduce numbers to a point 1998; Moser 2001). where habitat damage is mitigated, and Second, the “biocontrol” problem forces (importantly) where normal harvest pressure consideration of the underlying assumption can serve to keep the population in check that harvest, and therefore harvest (Batt 1997; Rockwell et al. 1997). What is management, is an efficient and effective interesting in this situation is that the tool to facilitate change in abundance. “biocontrol” objective for light geese Traditionally, waterfowl management was represents a clear paradigm shift for premised on the belief that small changes in waterfowl management, in a couple of bag limits would lead to desired and respects. First, it is entirely counter to the detectable numerical responses in the target usual management perspective that ‘more is populations. However, despite special better’. Among waterfowl managers, this legislation passed to encourage very large view is probably heavily influenced by the increases in light goose harvest, there is perspective that there is “no such thing as strong empirical evidence to suggest that too many ducks”. The prevailing view is that goose populations have not been controlled, because ducks are strongly limited by specific and continue to increase (Alisauskas et al. habitat requirements (e.g. ponds in breeding 2011). There are at least two proposed wetland areas), there is little potential for explanations for the failure of harvest to duck populations to exceed carrying capacity significantly reduce goose numbers and limit

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 265 further population growth. One is that the degree to which harvest can influence number of geese was simply too large for waterfowl population dynamics. Under an the current number of hunters to possibly objective of population control, optimal control. While this is undoubtedly true to management would attempt to maintain some degree, this problem is likely population size below the threshold at compounded by: 1) a functional response by which population growth exceeds the geese becoming more wary and thus less capacity for control by harvest (see Hauser et vulnerable, or 2) some aspect of hunter al. 2007 for an example applied to the behaviour, such as an attenuation in Atlantic Flyway population of Canada effort/hunter or general willingness to Geese). Thus, for fixed (frequently at pursue Snow Geese after an initial interest maximally liberal levels for control for doing so during the beginning of the objectives) hunting regulations, goose conservation order. Both a functional populations appear to respond in a positive behavioural response by geese and changes density-dependent manner. This is quite in hunter behaviour were observed different than typical duck management, following implementation of liberalisations where optimal strategies are conditioned by in Greater Snow Goose harvest regulations the expectation of some level of negative (Béchet et al. 2003; Calvert et al. 2007). density-dependent feedback over the range Additional focus on understanding these of population sizes typically encountered. aspects of partial controllability and hunter Much of our presentation has focussed behaviour would be useful. on “models” – specifically, population Regardless of whether the underlying projection models which allow us to make mechanisms involve hunter behaviour or predictions about the numerical trajectory bird behaviour, the result for management is of populations over time and space, and the a need to modify model components degree to which those trajectories might be relating hunting regulations to harvest rate. influenced by perturbation, natural or Rather than model harvest rate resulting anthropogenic (in the context of this review, from any fixed set of hunting regulations as meaning “harvest”). The consideration of characterised by a single probability models focuses on two key points. distribution (Johnson et al. 1997; USFWS Models represent canonical and (relatively) 2013), it will be necessary to consider a transparent representations about what we family of distributions associated with think we know concerning the factors which different population sizes. Above some determine the dynamics of populations. threshold population size, average harvest Second, we can, and often do, use these rates for any fixed set of hunting regulations models in applications. Here, we have are hypothesised to decrease as a function of focussed on models where dynamics are abundance. Incorporation of such density- potentially impacted by harvest. These dependent harvest rates into our harvest models are used to project the management models highlights the fact that consequences to waterfowl population there are real and practical limits on the dynamics of specified changes in rates

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 266 Effects of harvest on waterfowl populations of survival, fecundity, and in some studies. These latter difficulties led to our cases movement. The major sources description of some commonly used of uncertainty in waterfowl harvest techniques for such inferences, with management involve the translation of warnings about possible pitfalls. Effects hunting regulations (the actions that we of changes in hunting mortality on take) into various hunting mortality rates, reproductive output generally entail density- and the subsequent translation of specific dependent responses, and there is evidence hunting mortality rates into changes in total of negative density-dependent reproduction rates of survival and fecundity. in a number of waterfowl species. Finally, With respect to the first source of there is also good evidence that hunting can uncertainty, hunting mortality rates are influence waterfowl movement rates at stochastically predictable for sets of hunting spatial scales ranging from local to regional, regulations with which we have experience. but the population dynamic consequences of Uncertainty about translation of hunting these movements are not well understood. mortality rates into population level effects Traditionally, the use of such models (say, is greater, but potentially reducible. The in harvest management) has been presented additive mortality hypothesis and the as something distinct from efforts to use compensatory mortality hypothesis represent models as a basis for discriminating among endpoints in which effects of hunting associated hypotheses. However, we submit mortality on total survival are maximal and that this dichotomy is a false one, and that minimal, respectively. The additive mortality use of models within a structured, adaptive hypothesis corresponds to the idea of management framework not only serves as a independent competing risks, whereas the transparent, defensible mechanism to compensatory mortality hypothesis is evaluate harvest objectives, but the process thought to require mechanisms such as of harvest in turn provides a powerful density-dependent mortality or individual experimental framework in which the heterogeneity in survival probabilities. various hypotheses about system dynamics A summary of evidence based on analyses inherent in the population models can for various waterfowl species leads to the be tested. Adaptive management was general inference that the additive hunting developed to aid decision making for mortality hypothesis represents a reasonable recurrent decision problems characterised by approximation to reality for most goose potentially reducible structural uncertainty. species. Results are quite varied The establishment of annual hunting for duck species to the point that regulations is certainly a recurrent decision generalisations are not really possible. It is problem, and the various hypotheses about not clear whether this inability to generalise population dynamic effects of hunting is attributable to underlying processes represent reducible structural uncertainty. varying across species, locations and even A programme of adaptive harvest time periods, or instead to the difficulties of management (AHM) was established for drawing inferences from observational mid-continent Mallard in 1995 and has been

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 267 used to by the USFWS to develop hunting much emphasis has been placed on the regulations for these birds since that time. consideration of alternative harvest AHM for eastern Mallard was initiated by the management objectives and possible USFWS in 2000. AHM programmes have linkages to habitat management programmes since been developed for other species and with formal connections to the human populations as well, but mid-continent and dimensions of waterfowl management eastern Mallard provide our longest running (Runge et al. 2006; NAMWP 2014; Osnas et examples. Evolution of model weights al. 2014). In addition to updating AHM (measures of model credibility) shows that model sets, the double-loop learning process predictions of the models including weakly of AHM has also offered an opportunity to density-dependent reproduction and additive think critically about large scale system hunting mortality have performed best over change and how decision frameworks will the respective periods of operation of these need to be adjusted to cope with this new two AHM programmes. In addition to this form of uncertainty (Nichols et al. 2011). As reduction in uncertainty (learning), AHM the interest in AHM has expanded, so has provides a basis for making decisions in the the variety of populations and systems for face of such uncertainty about processes which it has been proposed (e.g. Hauser et al. governing population responses to 2007; Johnson et al. 2014). In developing management actions. these new applications, managers and While there was probably no serious waterfowl biologists are forced to identify expectation that application of an AHM important differences in the systems in framework to harvest management of mid- question (e.g. system dimension and continent Mallard would unequivocally come alternative harvest management objectives), down as favouring one hypothesis or the and how well or directly experiences other, results to date have formed the basis accumulated to date with (primarily) single for much new thinking about not only the species near-scalar duck models, apply to underlying mechanisms of compensation other systems, including (in particular) geese. (e.g. the possible role of individual heterogeneity or alternative models to Acknowledgements represent density-dependent survival), but We thank the organisers of the North also the policy elements of AHM. Perhaps America Duck Symposium for inviting this one of the greatest benefits of AHM is that submission. In particular, we thank Dave it provides a structured decision making Koons and Rick Kaminski for keeping us on approach that allows one to agree to a formal track for what was intended to be a “brief process while explicitly considering review” of the subject. They are not to be alternative beliefs (and disagreements) about blamed, however, for what is a clear failure management outcomes (Johnson & Case on our part to attain anything approaching 2000). As the United States waterfowl brevity. The paper benefitted from harvest management community has entered comments by Dave Koons, Mark Lindberg, the double-loop learning phase of AHM, Eric Osnas, Eileen Rees and Gary White.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 268 Effects of harvest on waterfowl populations

References Publication. U.S. Fish and Wildlife Service, Washington D.C., USA, and Canadian Abraham, K.F. & Jefferies, R.L. 1997. High goose Wildlife Service, Ottawa, Canada. populations: causes, impacts and implications. Batt, B.D.J. (ed.). 1998. The Greater Snow Goose: In B.D.J. Batt, (ed.), Arctic Ecosystems in Peril: Report of the Arctic Goose Habitat Working Report of the Arctic Goose Habitat Working Group. Arctic Goose Joint Venture Special Group, pp. 7–72. Arctic Goose Joint Venture Publication. U.S. Fish and Wildlife Service, Special Publication. U.S. Fish and Wildlife Washington D.C., USA, and Canadian Service, Washington D.C., USA, and Wildlife Service, Ottawa, Canada. Canadian Wildlife Service, Ottawa, Canada. Béchet, A., Giroux, J-F., Gauthier, G., Nichols, Alisauskas, R.T., Rockwell, R.F., Dufour, K.W., J.D. & Hines, J.E. 2003. Spring hunting Cooch, E.G., Zimmerman, G., Drake, K.L., changes the regional movements of Leafloor, J.O., Moser, T.J. & Reed, E.T. 2011. migrating greater snow geese. Journal of Effect of population reduction efforts on Applied Ecology 40: 553–564. harvest, survival, and population growth of Béchet, A., Giroux, J-F. & Gauthier, G. 2004. The mid-continent lesser snow geese. Wildlife effects of disturbance on behaviour, habitat Monographs 179: 1–42. use and energy of spring staging snow geese. Anderson, D.R. 1975. Population ecology of the Journal of Applied Ecology 41: 689–700. mallard: V. Temporal and geographic Bellrose, F.C. 1954. The value of waterfowl estimates of survival, recovery and harvest refuges in Illinois. Journal of Wildlife rates. USFWS Resource Publication No.125. Management 18: 160–169. U.S. Fish and Wildlife Service, Washington Bellrose, F.C. 1990. History of wood duck D.C., USA. management in North America. In L. Anderson, D.R. & Burnham, K.P. 1976. Fredrickson, G.V. Burger, S.P. Havera, D.A. Population ecology of the mallard: VI. The Gruber, R.E. Kirby & T.S. Taylor (eds.), effect of exploitation on survival. USFWS Proceedings of the 1988 North American Wood Resource Publication No. 128. U.S. Fish and Duck Symposium, pp. 13–20. Gaylord Wildlife Service, Washington D.C., USA. Memorial Laboratory, St. Louis, USA. Ankney, C.D. 1996. An embarrassment of riches: Bellman, R.E. 1957. Dynamic Programming. too many geese. Journal of Wildlife Management Princeton University Press, Princeton, USA. 60: 217–223. Berkson, J. & Elveback, L. 1960. Competing Arnott, W.G. 2007. Birds in the Ancient World from exponential risks, with particular reference to A to Z. Routledge, Oxford, UK & New York, the study of smoking and lung cancer. Journal USA. of the American Statistical Association 55: 415–428. Balmaki, B. & Barati, A. 2006. Harvesting status Beverton, R.J.H., & Holt, S.J. 1957. On the of migratory waterfowl in northern Iran: a Dynamics of Exploited Fish Populations. case study from Gilan Province. In G.C. Chapman and Hall, London, UK. Boere, C.A. Galbraith & D.A. Stroud (eds), Bischof, R., Mysterud, A. & Swenson, J.E. 2008. Waterbirds around the World, pp. 868–869. The Should hunting mortality mimic the patterns Stationery Office, Edinburgh, UK. of natural mortality? Biology Letters 4: 307–310. Batt, B.D.J. (ed.). 1997. Arctic Ecosystems in Peril: Black, J.M. 1996. Partnerships in Birds – the Study of Report of the Arctic Goose Habitat Working Monogamy. Oxford University Press, Oxford, Group. Arctic Goose Joint Venture Special UK.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 269

Boyd, H. 1962. Population dynamics and the empirical investigation using soil mites. exploitation of ducks and geese. In E.D. Le Journal of Animal Ecology 73: 996–1006. Cren & M.W. Holdgate (eds.), The Exploitation Caswell, H. 2001. Matrix Population Models: of Natural Animal Populations. A Symposium of Construction, Analysis and Interpretation, 2nd the British Ecological Society, Durham 28th–31st Edition. Sinauer Associates, Sunderland, March 1960, pp. 85–95. John Wiley & Sons, USA. New York, USA. Caswell, H. & Weeks, D.E. 1986. 2-sex models – Brooks, E.N. & Lebreton, J-D. 2001. Optimizing chaos, extinction, and other dynamic removals to control a metapopulation: consequences of sex. American Naturalist 128: application to the yellow legged herring gull 707–735. (Larus cachinnans). Ecological Modelling 136: Charnov, E.L. 1993. Life History Invariants. 269–284. Oxford University Press, Oxford, UK. Brown, G.M., Jr., Hammack, J. & Tillman, M.F. Chiang, C.L. 1968. Introduction to Stochastic Processes 1976. Mallard population dynamics and in Biostatistics. John Wiley, New York, USA. management models. Journal of Wildlife Cohen, J.E. An uncertainty principle in Management 40: 542–555. demography and the unisex issue. The Brownie, C. 1974. Estimation of Instantaneous American Statistician 40: 32–39. Mortality Rates from Band Return Data When Conn, P.B. & Kendall, W.L. 2004. Evaluating ‘Within-season Date of Kill’ is Also Available. mallard adaptive management models with Biometrics Unit Report No. BU-536-M. time series. Journal of Wildlife Management 68: Biometrics Unit, Cornell University, Ithaca, 1065–1081. New York, USA. Conroy, M.J. & Krementz, D.G. 1990. A review Brownie, C. & Pollock, K.H. 1985. Analysis of the evidence for the effects of hunting of multiple capture-recapture data using on American black duck populations. band-recovery methods. Biometrics 41: 411– Transactions of the North American Wildlife 420. and Natural Resources Conference 55: 501–517. Buebbert, H.F. 1981. Breeding birds on Conroy, M.J., Miller, M.W. & Hines, J.E. 2002. waterfowl production areas in northeastern Identification and synthetic modeling of North Dakota. Prairie Naturalist 13: 19–22. factors affecting American black duck Bunnefeld, N., Reuman, D.C., Baines, D. & populations. Wildlife Monographs 150. Milner-Gulland, E.J. 2011. Impact of Cooch, E.G., Lank, D.B., Rockwell, R.F. & unintentional selective harvesting on the Cooke, F. 1989. Long-term decline in population dynamics of red grouse. Journal of fecundity in a snow goose population – Animal Ecology 80: 1258–1268. evidence for density dependence. Journal of Burnham, K.P. & Anderson, D.R. 1984. Tests of Animal Ecology 58: 711–726. compensatory vs additive hypotheses of Cooch, E.G., Lank, D.B., Rockwell, R.F. & mortality in mallards. Ecology 65: 105–112. Cooke, F. 1991. Long-term decline in body Burnham, K.P., White, G.C. & Anderson, D.R. size in a snow goose population – evidence 1984. Estimating the effect of hunting on of environmental degradation. Journal of annual survival rates of adult mallards. Animal Ecology 60: 483–496. Journal of Wildlife Management 48: 350–361. Cooke, F., Bousfield, M.A. & Sadura, A. 1981. Cameron, T.C. & Benton, T.G. 2004. Stage- Mate change and reproductive success structured harvesting and its effects: an in the lesser snow goose. Condor 83: 322–327.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 270 Effects of harvest on waterfowl populations

Cox, R.R. & Afton, A.D. 1997. Use of habitats by Akademiens Handlingar, Antikvariska serien 45. female northern pintails wintering in Almqvist & Wiksell International, Stockholm, southwestern Louisiana. Journal of Wildlife Sweden. Management 61: 435–443. Errington, P.L. 1943. An analysis of mink Crissey, W.F. 1957. Forecasting waterfowl harvest predation upon muskrats in north-central by flyways. Transactions of the North American United States. Iowa Agricultural Research Wildlife and Natural Resources Conference 22: Bulletin 320: 728–924. 256–268. Errington, P.L. 1967. Of Predation and Life. Iowa Day, A. 1949. North American Waterfowl. Stackpole State University Press, Ames, USA. Boolks, Harrisburg, USA. Fox, A.D. & Madsen, J. 1997. Behavioural and Devineau, O., Guillemain, M., Johnson, A.R. & distributional effects of hunting disturbance Lebreton, J-D. 2010. A comparison of green- on waterbirds in Europe: Implications for winged teal Anas crecca survival and harvest refuge design. Journal of Applied Ecology 34: between Europe and North America. Wildlife 1–13. Biology 16: 12–24. Gaillard, J.M., Pontier, D., Allaine, D., Lebreton, Drent, R.H. & Daan, S. 1980. The prudent J.D., Trouvilliez, J. & Clobert, J. 1989. An parent: energetic adjustments in avian analysis of demographic tactics in birds and breeding? Ardea 68: 225–252. mammals. Oikos 56: 59–76. Duebbert, H.F., Jacobson, E.T., Higgins, K.F. Garrettson, P.R., & F.C. Rohwer. 2001. Effects of & Podoll, E.B. 1981. Establishment of Sseeded mammalian predator removal on upland- Grasslands for Wildlife Habitat in the Prairie Pothole nesting duck production in North Dakota. Region. U.S. Fish and Wildlife Service Special Journal of Wildlife Management 65: 398–405. Scientific Report: Wildlife No. 234. U.S. Fish Gauthier, G. & Lebreton J-D. 2004. Population and Wildlife Service, Washington D.C., USA. models for Greater Snow Geese: a Dufour, K.W., Ankney, C.D. & Weatherhead, P.J. comparison of different approaches to 1993. Condition and vulnerability to hunting assess potential impacts of harvest. Animal among mallards staging at Lake St. Clair, Biodiversity and Conservation 27: 503–514. Ontario. Journal of Wildlife Management 57: Geis, A.D., Martinson, R.K & Anderson, D.R. 209–215. 1969. Establishing hunting regulations and Ebbinge, B.S. 1991. The impact of hunting on allowable harvest of mallards in United States. mortality-rates and spatial-distribution of Journal of Wildlife Management 33: 848–859. geese wintering in the western Palearctic. Gill, J.A. & Sutherland, W.J. 2000. Predicting the Ardea 79: 197–209. consequences of human disturbance from Elmberg, J., Nummi, P., Pöysä, H., Sjöberg, K., behavioural decisions. In L.M. Gosling & W.J. Gunnarsson, G., Clausen, P., Guillemain, M., Sutherland (eds.), Behaviour and Conservation, Rodrigues, D. & Väänänen, V-M. 2006. The pp. 51–64. Cambridge University Press, scientific basis for a new and sustainable Cambridge, UK. management of migratory European ducks. Gill, J.A., Norris, K. & Sutherland, W.J. 2001. Wildlife Biology 12: 121–128. Why behavioural responses may not reflect Ericson, P.G.P. & Tyrberg, T. 2004. The early the population consequences of human history of the Swedish avifauna. A review of disturbance. Biological Conservation 97: 265–268. the subfossil record and early written sources. Gilmer, D.S., Miller, M.R., Bauer, R.D. & Kungl. Vitterhets Historie och Antikvitets Ledonne, J.R. 1982. California’s Central

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 271

Valley Wintering Waterfowl – Concerns and in presence of competing risks. Journal of Challenges. Transactions of the North American Wildlife Management 70: 1544–1555. Wildlife and Natural Resources Conference 47: Heitmeyer, M.E., Fredrickson, L.H. & Humburg, 441–452. D.H. 1993. Further evidence of biases Giroux, J-F. 1981. Ducks nesting on artificial associated with hunter-killed mallards. Journal islands during drought. Journal of Wildlife of Wildlife Management 57: 733–740. Management 45: 783–786. Hepp, G.R., Blohm, R.J., Reynolds, R.E., Hines, Grey, D.R. & Law, R. 1987. Reproductive values J.E. & Nichols, J.D. 1986. Physiological and maximum yields. Functional Ecology 1: condition of autumn-banded mallards and 327–330. its relationship to hunting vulnerability. Gunnarsson, G., Elmberg, J., Pöysä, H., Nummi, Journal of Wildlife Management 50: 177–183. P., Sjöberg, K., Dessborn, L. & Arzel, C. Hilborn, R. & Walters, C.J. 1992. Quantitative Fisheries 2013. Density-dependence in ducks: a review Stock Assessment: Choice, Dynamics and Uncertainty. of the evidence. European Journal of Wildlife Chapman and Hall, New York, USA. Research 59: 305–321. Hutchings, J.A. & Baum, J.K. 2005. Measuring Hammond, M.C. & Mann, G.E. 1956. Waterfowl marine fish biodiversity: temporal changes in nesting islands. Journal of Wildlife Management abundance, life history and demography. 20: 345–352. Philosophical Transactions of the Royal Society of Hario, M., Hollmén, T.E., Morelli, T.L. & London. Series B, Biological Sciences 360: 315–338. Scribner, K.T. 2002. Effects of mate removal Johnson, D.H., Burnham, K.P. & Nichols, J.D. on the fecundity of common eider Somateria 1984. The role of heterogeneity in animal mollissima females. Wildlife Biology 8: 161–168. population dynamics. Proceedings of the Haramis, G.M. & Thompson, D.Q. 1985. International Biometrics Conference 13: 1–15. Density-production characteristics of box- Johnson, D.H., Nichols, J.D., Conroy, M.J. & nesting wood ducks in a northern green-tree Cowardin, L.M. 1988. Some considerations impoundment. Journal of Wildlife Management in modeling the mallard life cycle. Pp. 9–20 in 49: 429–436. M. W. Weller, ed. Waterfowl in Winter. University Hauser, C.E., Cooch, E.G. & Lebreton, J-D. of Minnesota Press, Minneapolis, USA. 2006. Control of structured populations by Johnson, F.A. & Case, D.J. 2000. Adaptive harvest. Ecological Modelling 196: 462–470. regulation of waterfowl harvests: lessons Hauser, C.E., Runge, M.C., Cooch, E.G., learned and prospects for the future. Johnson, F.A. & Harvey, W.F. 2007. Optimal Transactions of the North American Wildlife and control of Atlantic population Canada geese. Natural Resources Conference 65: 94–108. Ecological Modelling 201: 27–36. Johnson, F.A., Williams, B.K., Nichols, J.D., Hawkins, A.S. & Bellrose, F.C. 1940. Wood duck Hines, J.E, Kendall, W.L., Smith, G.W. & habitat management in Illinois. North American Caithamer, D.F. 1993. Developing an Wildlife Conference Transactions 5: 392–395. adaptive management strategy for harvesting Heisey, D.M. & Fuller, T.K. 1985. Evaluation of waterfowl in North America. Transactions of survival and cause-specific mortality rates the North American Wildlife and Natural using telemetry data. Journal of Wildlife Resources Conference 58: 565–583. Management 49: 68–674. Johnson, F.A., Moore, C.T., Kendall, W.L., Heisey, D.M. & Patterson, B.R. 2006. A review of Dubovsky, J.A., Caithamer, D.F., Kelley, J.T. methods to estimate cause-specific mortality & Williams, B.K. 1997. Uncertainty and the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 272 Effects of harvest on waterfowl populations

management of mallard harvests. Journal of Lebreton, J.-D. 2009. Assessing density-dependence: Wildlife Management 61: 203–217. where are we left? In D.L. Thomson, E.G. Johnson, F.A., Kendall, W.L. & Dubovsky, J.A. Cooch & M.J. Conroy (eds.), Modeling 2002. Conditions and limitations on learning Demographic Processes in Marked Populations, in the adaptive management of mallard pp. 19–42. Environmental and Ecological harvests. Wildlife Society Bulletin 30: 176–185. Statistics, Springer, New York, USA. Johnson, F.A., Jensen, G.H., Madsen, J. & Lebreton, J.-D. & Clobert, J. 1991. Bird population Williams, B.K. 2014. Uncertainty, robustness dynamics, management, and conservation: and the value of information in managing an the role of mathematical modelling. In C.M. expanding Arctic goose population. Ecological Perrins, J. Lebreton & G.J. Hirons (eds.), Bird Modelling 273: 186–199. Population Studies, pp. 105–125. Oxford Kaminski, R.M. & Gluesing, E.A. 1987. Density- University Press, Oxford, UK. and habitat-related recruitment in mallards. Legagneux, P., Inchausti, P., Bourguemestre, Journal of Wildlife Management 51:141–148. F., Latraube, F. & Bretagnolle, V. 2009. Kauhala, K. 2004. Removal of medium-sized Effect of predation risk, body size predators and the breeding success of ducks and habitat characteristics on emigration in Finland. Folia Zoologica 53: 367–378. decisions in mallards. Behavioral Ecology 20: Keyfitz, N. 1971. On the momentum of 186–194. population growth. Demography 8: 71–80. Lercel, B.A., Kaminski, R.M. & Cox, R.R. 1999. Klaassen, M. 2002. Relationships between Mate loss in winter affects reproduction of migration and breeding strategies in arctic mallards. Journal of Wildlife Management 63: breeding birds. In P. Berthold, E. Gwinner & 621–629. E. Sonnenschein (eds.), Avian Migration, pp. Lepage, D., Desrochers, A. & Gauthier, G. 1999. 237–249. Springer-Verlag, Berlin, Germany. Seasonal decline of growth and fledging Kokko, H. 2001. Optimal and suboptimal use of success in snow geese Anser caerulescens: an compensatory responses to harvesting: effect of date or parental quality? Journal of timing of hunting as an example. Wildlife Avian Biology 30: 72–78. Biology 7: 141–152. LeSchack, C.R., Afton, A.D. & Alisauskas, Koons, D.N., Rockwell, R.F. & Grand, J.B. 2006. R.T. 1998. Effects of male removal on Population momentum: Implications for female reproductive biology in Ross’ and wildlife management. Journal of Wildlife lesser snow geese. Wilson Bulletin 110: 56–64. Management 70: 19–26. Lindberg, M.S., Sedinger, J.S. & Lebreton, J.-D. Koons, D.N., Gunnarsson, G., Schmutz, J.A. & 2014. Individual heterogeneity in black brant Rotella, J.J. 2014a. Drivers of waterfowl survival and recruitment with implications population dynamics: from teal to swans. for harvest dynamics. Ecology and Evolution 3: Wildfowl (Special Issue No. 4): 169–191. 4045–56. Koons, D.N., Rockwell, R.F. & Aubry, L.M. 2014b. Lindström, J. & Kokko, H. 1998. Sexual Effects of exploitation on an overabundant reproduction and population dynamics: the species: the lesser snow goose predicament. role of polygyny and demographic sex Journal of Animal Ecology 83: 365–374. differences. Proceedings of the Royal Society B – Lebreton, J.-D. 2005. Dynamical and statistical Biological Sciences 265: 483–488. models for exploited populations. Australian Link, P.T., Afton, A.D., Cox, R.R. & Davis, B.E. & New Zealand Journal of Statistics 47: 49–63. 2011. Daily movements of female mallards

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 273

wintering in southwestern Louisiana. Martin, K., Cooch, F.G. & Rockwell, R.F. 1985. Waterbirds 34: 422–428. Reproductive performance in lesser snow MacArthur, R.H. 1960. On the relation between geese: are two parents essential? Behavioral reproductive value and optimal predation. and Ecological Sociobiology 17: 257–263. Proceedings of the National Academy of Sciences of McCabe, R.E 1987. Results of stabilized duck the United States of America 46: 143–145. hunting regulations. Transactions of the North Madsen, J. 1998a. Experimental refuges for American Wildlife and Natural Resources migratory waterfowl in Danish wetlands. I. Conference 52: 177–326. Baseline assessment of the disturbance McKinnon, D.T. & Duncan, D.C. 1999. effects of recreational activities. Journal of Effectiveness of dense nesting cover for Applied Ecology 35: 386–397. increasing duck production in Saskatchewan. Madsen, J. 1998b. Experimental refuges for Journal of Wildlife Management 63: 382–389. migratory waterfowl in Danish wetlands. II. Mertz, D.B. 1971. Life history phenomena in Tests of hunting disturbance effects. Journal increasing and decreasing populations. In E.C. of Applied Ecology 35: 398–417. Pielou & W.E. Waters (eds.), Statistical ecology, Madsen, J. & Jepsen, P.U. 1992. Passing the buck. Vol. II. Sampling and Modeling Biological Need for a flyway management plan for the Populations and Population Dynamics, pp. 361– Svalbard Pink-footed Goose. In M. van 399. Pennsylvania State University Press, USA. Roomen & J. Madsen (eds.), Waterfowl and Miller, M.R. & Eadie, J. McA. 2006. The Agriculture: Review and Future Perspective of the allometric relationship between resting Crop Damage Conflict in Europe, pp. 109–110. metabolic rate and body mass in wild IWRB Special Publications No. 21. IWRB, waterfowl (Anatidae) and an application to Slimbridge, UK. estimation of winter habitat requirements. Madsen, J. & Fox, A.D. 1995. Impacts of hunting Condor 108: 166–177. disturbance on waterbirds – a review. Wildlife Milner, J.M., Nilsen, E.B. & Andreassen, H.P. Biology 1: 193–207. 2007. Demographic side effects of selective Mainguy, J., Béty, J., Gauthier, G. & Giroux, J.F. hunting in ungulates and carnivores. 2002. Are body condition and reproductive Conservation Biology 21: 36–47. effort of laying Greater Snow Geese affected Moser T.J. (ed.). 2001. The Status of Ross’s Geese. by the spring hunt? Condor 104: 156–161. Arctic Goose Joint Venture Special Manlove, C.A. & Hepp, G.R. 1998. Effects of Publication. U.S. Fish and Wildlife Service, mate removal on incubation behavior and Washington D.C., USA, and Canadian reproductive success of female wood ducks. Wildlife Service, Ottawa, Canada. Condor 100: 688–693. Murray, D.L., Anderson, M.G. & Steury, T.D. Martin, F.W., Pospahala, R.S. & Nichols, 2010. Temporal shift in density-dependence J.D. 1979. Assessment and population among North American breeding duck management of North American migratory populations. Ecology 91: 571–581. birds. In J. Cairns, Jr., G.P. Patil & W.E. Waters Nichols, J.D. 1991a. Extensive monitoring (eds.), Environmental Biomonitoring, Assessment, programmes viewed as long-term population Prediction, and Management – Certain Case Studies studies: the case of North American and Related Quantitative Issues. Statistical waterfowl. Ibis 133 suppl. 1: 89–98. Ecology, Vol. 11. pp. 187–239, International Nichols, J.D. 1991b. Science, population ecology, Cooperative. Publishing House, Fairland, USA. and the management of the American black

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 274 Effects of harvest on waterfowl populations

duck. Journal of Wildlife Management 55: populations from incomplete data. Conservation 790–799. Biology 19: 826–835. Nichols, J.D. 2000. Evolution of harvest Osnas, E.E., Runge, M.C., Mattsson, B.J., Austin, management for North American waterfowl: J., Boomer, G.S., Clark, R.G., Devers, P. Selective pressures and preadaptations for Eadie, J.M., Lonsdorf, E.V. & Tavernia, B.G. adaptive harvest management. Transactions of 2014. Managing harvest and habitat as the North American Wildlife and Natural integrated components. Wildfowl (Special Resources Conference 65: 65–77. Issue No. 4): 305–328. Nichols, J.D. & Hines, J.E. 1987. Population Ecology Pace, R.M. & Afton, A.D. 1999. Direct recovery of the Mallard: VIII. Winter Distribution Patterns rates of lesser scaup banded in the Northwest and Survival Rates of Winter-banded Mallards. Minnesota: sources of heterogeneity. Journal U.S. Geological Survey Resource Publication of Wildlife Management 63: 389–395. No. 162, USGS, Washington D.C., USA. Padding, P.I., Gobeil, J.F. & Wentworth, C. 2006. Nichols, J.D. & Johnson, F.A. 1990. Wood duck Estimating waterfowl harvest in North population dynamics: a review. In L.H. America. In G.C. Boere, C.A. Galbraith & Fredrickson, G.V. Burger, S.P. Havera, D.A. D.A. Stroud (eds.), Waterbirds Around the Graber, R.E. Kirby & T.S. Taylor (eds.), World, pp. 849–852. The Stationery Office, Proceedings of the 1988 North American Wood Edinburgh, UK. Duck Symposium, pp. 83–105. St. Louis, USA. Patterson, J.H. 1979. Can ducks be managed Nichols, J.D. & Williams, B.K. 2013. Adaptive by regulation? Experience in Canada. management. In A.H. El-Shaarawi & W.W. Transactions of the North American Wildlife Piegorsch (eds.), Encyclopedia of Environmetrics, and Natural Resources Conferences 44: 130– 2nd edition, pp. 1–6. John Wiley, New York, 139. USA. Peron, G., Nicolai, C.A. & Koons, D.N. 2012. Nichols, J.D., Conroy, M.J., Anderson, D.R. Demographic response to perturbations: the & Burnham, K.P. 1984. Compensatory role of compensatory density-dependence in mortality in waterfowl populations – a review a North American duck under variable of the evidence and implications for research harvest regulations and changing habitat. and management. Transactions of the North Journal of Animal Ecology 81: 960–969. American Wildlife and Natural Resources Peron, G. 2013. Compensation and additivity of Conference 49: 535–554. anthropogenic mortality: life-history effects Nichols, J.D., Johnson, F.A. & Williams, B.K. and review of methods. Journal of Animal 1995. Managing North American waterfowl Ecology 82: 408–417. in the face of uncertainty. Annual Review of Phillips, J.C. 1923. A Natural History of the Ducks. Ecology and Systematics 26: 177–199. Volume II : The Genus Anas. Houghton Mifflin Nichols, J.D., Koneff, M.D., Heglund, P.J., Co., Boston, USA. Knutson, M.G., Seamans, M.E., Lyons, J.E., Phillips, J., and Lincoln, F.C. 1930. American Morton, J.M., Jones, M.T., Boomer, G.S. Waterfowl: Their Present Situation and Outlook for & Williams, B.K. 2011. Climate change, the Future. Houghton Mifflin, Boston, USA. uncertainty and natural resource management. Pieron, M.R. & Rohwer, F.C. 2010. Effects of Journal of Wildlife Management 75: 6–18. large-scale predator reduction on nest Niel, C. & Lebreton, J.-D. 2005. Using demographic success of upland nesting ducks. Journal of invariants to detect overharvested bird Wildlife Management 74: 124–132.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 Effects of harvest on waterfowl populations 275

Pledger, S., Pollock, K.H. & Norris, J.L. 2003. Rockwell, R. F., Cooch, E.G. & Brault, S. 1997. Open capture-recapture models with Dynamics of the midcontinent population of heterogeneity: I. Cormack-Jolly-Seber model. lesser snow geese: projected impacts of Biometrics 59: 786–794. reduction in survival and fertility on Pöysä, H., Elmberg, J., Gunnarsson, G., Nummi, population growth rates. In B.D.J. Batt (ed.), P. & Sjöberg, K. 2004. Ecological basis of Arctic Ecosystems in Peril: Report of the Arctic sustainable harvesting: is the prevailing Goose Joint Venture Habitat Working Group, pp. paradigm of compensatory mortality still 73–100. Arctic Goose Joint Venture Special valid? Oikos 104: 612–615. Publication. U.S. Fish and Wildlife Service, Pöysä, H., Dessborn, L., Elmberg, J., Washington D.C., USA and Canadian Gunnarsson, G., Nummi, P., Sjöberg, K., Wildlife Service, Ottawa, Canada. Suhonen, S. & Söderquist, P. 2013. Harvest Runge, M.C. & Johnson, F.A. 2002. The mortality in North American mallards: A importance of functional form in optimal reply to Sedinger and Herzog. Journal of control solutions of problems in population Wildlife Management 77: 653–654. dynamics. Ecology 83: 1357–1371. Prop, J. & Quinn, J.L. 2003. Constrained by Runge, M.C., Johnson, F.A., Dubovsky, J.A., available raptor hosts and islands: density- Kendall, W.L., Lawrence, J. & Gammonley, J. dependent reproductive success in red- 2002. A Revised Protocol for the Adaptive Harvest breasted geese. Oikos 102: 571–580. Management of Mid-continent Mallards. U.S. Fish Pulliam, H.R. 1988. Sources, sinks, and population and Wildlife Service, U.S. Department of the regulation. American Naturalist 132: 652–661. Interior Technical report, Washington D.C., Reynolds, R.E., Shaffer, T.L, Sauer, J.R. & USA. Peterjohn, B.G. 1994. Conservation Reserve Runge, M.C., Johnson, F.A., Anderson, M.G., Program: benefit for grassland birds in the Koneff, M.D., Reed, E.T. & Mott, S.E. 2006. northern plains. Transactions of the North The need for coherence between waterfowl American Wildlife and Natural Resources harvest and habitat management. Wildlife Conference 59: 328–336. Society Bulletin 34: 1231–1237. Reynolds, R.E., Shaffer, T.L., Renner, R.W., Sæther, B-E., Engen, S., Grøtan, V., Bregnballe, T., Newton, W.E. & Batt, B.D.J. 2001. Impact of Both, C., Tryjanowski, P., Leivits, A., Wright, the Conservation Reserve Program on duck J., Møller, A.P., Visser, M.E. & Winkel, W. recruitment in the U.S. prairie pothole region. 2008. Forms of density regulation and (quasi-) Journal of Wildlife Management 65: 765–780. stationary distributions of population sizes in Ricker, W.E. 1954. Stock and recruitment. Journal birds. Oikos 117: 1197–1208. of Fish Research Board of Canada 11: 559–623. Sandercock, B.K., Nilsen, E.B., Broseth, H. & Rice, M.B., Haukos, D.A., Dubovsky, J.A. & Pedersen, H.C. 2011. Is hunting mortality Runge, M.C. 2010. Continental survival and additive or compensatory to natural recovery rates of northern pintails using mortality? Effects of experimental harvest band-recovery data. Journal of Wildlife on the survival and cause-specific mortality Management 74: 778–787. of willow ptarmigan. Journal of Animal Ringelman, K.M., Eadie, J.M. & Ackerman, J.T. Ecology 80: 244–258. 2014. Adaptive nest clustering and density- Schaub, M. & Lebreton, J.-D. 2004. Testing the dependent nest survival in dabbling ducks. additive versus compensatory hypothesis of Oikos 123: 239–247. mortality from ring recovery data using a

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 276 Effects of harvest on waterfowl populations

random effects model. Animal Biodiversity and future directions. Transactions of the North Conservation 27: 73–85. American Wildlife and Natural Resources Schaub, M. & Pradel, R. 2004. Assessing the Conference 52: 320. relative importance of different sources of U.S. Fish and Wildlife Service. 2013. Adaptive mortality from recoveries of marked animals. Harvest Management: 2013 Duck Hunting Season. Ecology 85: 930–938. U.S. Department of Interior, Washington Sedinger, J.S. & Herzog, M.P. 2012. Harvest and D.C., USA. dynamics of duck populations. Journal of Väänänen, V.-M. 2001. Hunting disturbance and Wildlife Management 76: 1108–1116. the timing of autumn migration in Anas Sedinger, J.S. & Nicolai, C.A. 2011. Recent trends in species. Wildlife Biology 7: 3–9. first-year survival for Black Brant breeding in Vaupel, J.W. & Yashin, A.I. 1985. Heterogeneity southwestern Alaska. Condor 113: 511–517. ruses – some surprising effects of selection Sedinger, J.S., Flint, P.L. & Lindberg, M.S. 1995. on population-dynamics. American Statistician Environmental influence on life-history traits – 39: 176–185. growth, survival and fecundity in Black Brant Viljugrein, H., Stenseth, N.C., Smith, G.W. & (Branta Bernicla). Ecology 76: 2404–2414. Steinbakk, G.H. 2005. Density-dependence Sedinger, J.S., Lindberg, M.S., Person, B.T., in North American ducks. Ecology 86: Eichholz, M.W., Herzog, M.P. & Flint, P.L. 245–254. 1998. Density-dependent effects on growth, Walters, C.J. 1986. Adaptive Management of body size and clutch size in Black Brant. Auk Renewable Resources. McGraw Hill, New York, 115: 613–620. USA. Sedinger, J.S., White, G.C., Espinosa, S., Partee, E.T. Whitehead, A.L., Edge, K.A., Smart, A.F., Hill, & Braun, C.E. 2010. Assessing compensatory G.S. & Willans, M.J. 2008. Large scale versus additive harvest mortality: an example predator control improves the productivity using greater sage-grouse. Journal of Wildlife of a rare New Zealand riverine duck. Management 74: 326–332. Biological Conservation 141: 2784–2794. Servanty, S., Choquet, R., Baubet, E., Brandt, S., Williams, B.K. & Johnson, F.A. 1995. Adaptive Gaillard, J.-M., Schaub, M., Toïgo, C., management and the regulation of waterfowl Lebreton, J.-D., Buoro, M. & Gimenez, harvests. Wildlife Society Bulletin 23: 430–436. O. 2010. Assessing whether mortality is Williams, B.K., Nichols, J.D & Conroy, M.J. 2002. additive using marked animals: a Bayesian Analysis and Management of Animal Populations. state-space modeling approach. Ecology 91: Academic Press, San Diego, USA. 1916–1923. Williams, B.K., Szaro, R.C. & Shapiro, C.D. 2007. Smith, L.M., Pederson, R.L. & Kaminski, R.M. Adaptive Management: The U.S. Department (eds.). 1989. Habitat Management for Migrating of the Interior Technical Guide. Adaptive and Wintering Waterfowl in North America. Texas Management Working Group, U.S. Tech University Press, Lubbock, USA. Department of the Interior, Washington Stearns, S. 1992. The Evolution of Life Histories. D.C., USA. Oxford University Press, Oxford, UK. Williams, B.K. & Brown, E.D. 2012. Adaptive Sparrow, R.D. & Patterson, J.H. 1987. Management: The U.S. Department of the Interior Conclusions and recommendations from Applications Guide. Adaptive Management studies under stabilized duck hunting Working Group, U.S. Department of the regulations, management implications and Interior, Washington D.C., USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 220–276 277

Cross-seasonal effects and the dynamics of waterfowl populations

JAMES S. SEDINGER1* & RAY T. ALISAUSKAS2

1Department of Natural Resources and Environmental Science, University of Nevada Reno, 1664 N. Virginia St., Reno, Nevada 89557, USA. 2Environment Canada, 115 Perimeter Road, Saskatoon, Saskatchewan S7N 0X4, Canada. *Correspondence author. E-mail: [email protected]

Abstract Cross-seasonal effects (CSEs) on waterfowl populations link together events and habitats that individuals experience as carry-over effects (COEs) throughout the annual cycle. The importance of CSEs has been recognised since at least the 1950s. Studies of nutrient dynamics beginning in the 1970s, followed by regression analyses that linked production of young to winter habitat conditions, confirmed the importance of CSEs. CSEs have been most apparent in large-bodied waterfowl, but evidence for CSEs in much smaller passerines suggests the potential for CSEs in all waterfowl. Numerous studies have established effects of winter weather on body condition and reproduction in both ducks and geese. Additionally, the ubiquitous use (during laying and incubation) of nutrients stored previously during spring migration suggests that such nutrients commonly influence reproductive success in waterfowl. Carry-over effects from the breeding season to autumn and winter are less well understood, although nutrition during the growth period in geese has been widely demonstrated to influence subsequent survival and reproduction. Only a few studies have examined effects of breeding on reproduction in later years. Because pathogens and parasites can be carried between seasonal habitats, disease represents an important potential mechanism underlying CSEs; so far, however, this role for diseases and parasitism remains poorly understood. CSEs were originally of interest because of their implications for management of seasonal habitats and CSEs represent a fundamental rationale for the habitat joint ventures in North America. Substantial research examining the role of COEs in individual fitness and of CSEs on population dynamics has now been conducted. New techniques (e.g. stable isotopes, geolocators) developed over the last decade, combined with more traditional marking programmes have created opportunities to understand CSEs more fully and to inform the management of seasonal habitats for waterfowl.

Key words: carry-over effect, climate, cross-seasonal effect, fitness, population, reproduction, survival.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 278 Cross-seasonal effects on waterfowl

The concept that variation in habitat quality role of habitats used by waterfowl outside can influence population dynamics of the breeding season in governing their waterfowl has been established for some subsequent breeding activity and thus their time (Lynch 1952). For example, reduced population dynamics. Ankney & MacInnes snow accumulation and spring rains in mid- (1978) and Raveling (1979) demonstrated continent prairie-parkland ecosystems result that female geese stored nutrients acquired in fewer ponds, which reduces annual before nesting and used these nutrient production of young ducks (Pospahala reserves during egg laying and incubation. et al. 1974). Kaminski & Gluesing (1987) Ankney & MacInnes (1978) further showed extended this notion to include cross- that clutch size of Lesser Snow Geese Chen seasonal effects (CSEs) when they showed caerulescens caerulescens was directly related to that the influence of breeding habitat their levels of endogenous nutrients. Shortly availability on autumn age-ratios (an index of thereafter, Heitmeyer & Fredrickson (1981) the breeding success of the population) was used regression approaches to illustrate that modified by population density in Mallard winter wetland indices and winter Anas platyrhynchos. Such density-dependence precipitation predicted age-ratios in Mallard in the reproductive process is incorporated harvested the next autumn in the Mississippi into models of population dynamics used to Flyway, suggesting that winter conditions manage harvest rates (Johnson et al. 1997). influenced production of young at the For ducks, the number of breeding population level. Raveling & Heitmeyer individuals in the population is influenced by (1989) used similar approaches to show that the production of young from the previous winter habitat conditions in California were breeding season and the number of important predictors of productivity in individuals surviving through winter, so that Northern Pintail Anas acuta, but that density-dependent effects on reproduction importance of winter habitat also depended represent a CSE. Newton (2006) showed on population size and breeding habitat how density-dependent mortality operating conditions. In 1982, 40 waterfowl scientists at different segments of the annual met in Puxico, Missouri (Anderson & Batt cycle interacted with density-dependent 1983) to identify key aspects of the winter reproduction to regulate population size. ecology of waterfowl that were still largely Before the 1970s, attempts to understand unstudied and not well understood. waterfowl population dynamics in North Specifically, this group focused on the America focused largely on habitat potential role of CSEs in governing conditions in the mid-continent prairie- dynamics of waterfowl populations and parkland breeding areas and on the role of proposed fundamental questions and harvest as the major drivers of population approaches toward an improved trends (Johnson et al. 1992). The interests of understanding of CSEs. There was strong a number of prominent waterfowl scientists support for hypothesis testing, as well as in the late 1970s to early 1980s coincided, comparative and experimental approaches stimulating discussion and thought on the to research combined with long-term

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 279 studies. Participants identified the need to COEs and CSEs may be modified further by understand a number of basic aspects of population size and density-dependence waterfowl biology including the extent and (Harrison et al. 2011). We exclude from basis for philopatry, potential of winter CSEs those environmental or maternal habitats to limit waterfowl populations, the effects experienced early in life on traits ecology of winter habitats, the bioenergetics such as body size, even though they of winter moult, the role of flocking may influence subsequent survival and behaviour in winter ecology and other reproduction by individuals (e.g. Sedinger et biological processes. A number of these al. 1995; Cooch 2002), because such effects questions have been addressed successfully, are relatively permanent and not necessarily but more importantly, the questions subject to modification of individual fitness themselves have evolved as our by events occurring later in life. We allow for understanding of CSEs has increased. COEs on both survival and reproduction Our goals in this paper are to: 1) provide with potential lagged effects over multiple a functional definition of CSEs; 2) discuss seasons, although we exclude direct effects how body size and life-histories affect the of behaviour or habitat on fitness during the potential for CSEs; 3) describe mechanisms same season. For example, we do not view underlying CSEs; 4) discuss the role of reduced survival of female ducks associated CSEs in management of waterfowl; and with nesting (Arnold et al. 2010) as a COE 5) propose approaches for detecting, but we do consider reduced survival of understanding and measuring the magnitude breeding individuals over the next winter of CSEs. (e.g. Daan et al. 1996). Similarly, we do not consider reduced within-winter survival Cross-seasonal effects associated with occupancy of a particular We largely subscribe to the definitions of habitat (Fleskes et al. 2007) as a COE, but do CSEs developed by Norris (2005) and include the potential lagged effect of lower Harrison et al. (2011). These definitions survival during spring migration associated attribute lagged population processes in one with use of a particular winter habitat as a season to conditions during a previous COE. season. For example, breeding propensity by Generally, we are interested in COEs that Mallard in prairie-parkland ecosystems influence breeding success as a result of might depend on habitat conditions, feeding previous events during the same annual opportunities and climate experienced by cycle (e.g. Alisauskas 2002; Bêty et al. 2003). individuals during the previous winter in the Presumably, a COE may extend from one Mississippi Alluvial Valley. We distinguish breeding season to the next, although effects between CSEs at the population level and extending over a full annual cycle may be carry-over effects (COEs) that operate at mediated by events during intervening the individual level, but the two are linked as seasons (Fig. 1). We are also concerned with population CSEs can result from COEs factors that affect survival in seasons experienced by individuals. The influence of subsequent to use of particular habitats or

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 280 Cross-seasonal effects on waterfowl

Breed Breed Breed

a b c Frequency Frequency Frequency

0 50 100 150 200 0 50 100 150 200 0 50 100 150 200 Residual mass Residual mass Residual mass

Habitat management

Figure 1. Hypothetical links between habitat availability, nutritional status (measured as body mass) and the probability of breeding. We show increases in mean mass in response to increases in wintering habitat but two different responses to increased mass. In scenario (b), threshold mass required for breeding does not change in response to increased mass, while in scenario (c) threshold mass increases in response to habitat availability, resulting in a similar proportion of individuals breeding before and after habitat enhancement. Under scenario (c), individuals benefit nutritionally from increased food availability but other attributes (potentially latent) prevent many individuals from translating improved nutritional status into increased reproductive success. We do not advocate scenario (c) as the expected response to improved habitat conditions. In fact, numerous examples suggest that individuals respond positively to increased food availability. The intention, here, is to indicate that complexities, including latent attributes of individuals, may influence population level responses to habitat management. investments in reproduction. In this paper, another (Alisauskas & Ankney 1992), we consider COEs that have the potential to which results in part from lower mass- influence both reproduction and survival. specific metabolic rate in larger species We also explore how COEs that operate at (Kleiber 1975). Use of previously stored the individual level as precursors may scale (endogenous) nutrients for egg formation up to CSEs at the population level. Such and incubation generally increases with scaling up brings population level processes body size (Alisauskas & Ankney 1992; such as density-dependence into play. Meijer & Drent 1999; Alisauskas & DeVink 2015), although details about use of Body size, life-histories and the endogenous nutrients for breeding may vary potential for cross-seasonal as a function of the ecology or life-histories effects of individual species (e.g. Alisauskas & Ankney 1992; Alisauskas & DeVink 2015). Body size is clearly related to the ability to Because greater reliance on endogenous store nutrients in one season for use in nutrients should increase the potential for

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 281

COEs, one might expect that COEs are Teal Anas discors and Green-winged Teal more common in larger-bodied waterfowl A. crecca may not be paired until later in species (but see below). In fact, the largest the year (MacKinney 1986). Additional numbers of examples of CSEs exist for energy demands of males from larger larger bodied waterfowl, especially geese species associated with maintenance of (Table 1), although this may reflect to some pair bonds throughout autumn and winter extent the greater ease with which might increase the potential that males individuals of larger-bodied species can be of larger species are more influenced by monitored (e.g. Ebbinge & Spaans 1995). habitat quality in winter than is the case for Other aspects of life-histories covary smaller bodied species. Alternatively, costs with body size and also might be predicted of delayed pairing by females of larger to influence the potential for COEs. For bodied species, and associated reduced example, duration of maintenance of social access to food resources from intra- and bonds within families increases with body interspecific competition, might interact size; most geese and swans normally show disadvantageously with habitat quality to lifelong monogamy and their young remain create greater potential for COEs in females in family units through the first winter (with of larger-bodied species. their parents) and sometimes through a Despite the clear logic underlying the second winter (either with parents or solely linkage between body size and potential for in sibling groups, e.g. Prevett & MacInnes COEs, there may be species-specific 1980; Scott 1980). Because size of family deviations from this general pattern in groups influences social status (Boyd 1953; waterfowl. Examples from non-waterfowl Raveling 1970; Poisbleau et al. 2006), species indicate that some caution is breeding success in one year affects social warranted when considering such status in the following winter. Social status relationships. Variation in quality of related to family size affects dominance and wintering habitat, and its influences on aggressive defence of food resources. reproductive performance by American Presumably, enhanced access to food Redstarts Setophaga ruticilla is a very clear improves daily efficiency of nutrient storage illustration of COEs (Marra et al. 1998), and during spring hyperphagia as geese travel there exist numerous other examples from north to breeding areas. Thus, a series of small passerines and shorebirds (Norris & sequential COEs from the arctic summer, Marra 2007). Body size is probably too small through the following winter and spring in these passerine and shorebird examples migration may influence reproductive for substantial mass to be carried between investment by individual females in a sequentially used seasonal habitats. So, the subsequent breeding season. existence of COEs in such small birds Timing of pair formation in ducks is a suggests that mechanisms underlying COEs function of body size. For example, larger may involve complexities beyond nutritional bodied Mallard form pair bonds in early dynamics. We address some of these autumn, while small-bodied Blue-winged potential mechanisms below.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 282 Cross-seasonal effects on waterfowl

Cross-seasonal effects in capture-mark-recapture (CMR) studies waterfowl which demonstrate that quality of wintering locations influences breeding probability (e.g. Evidence of CSEs transmitted from Sedinger et al. 2011) provide additional wintering areas to breeding is extensive evidence for the impact of winter habitats on (Table 1). Studies based on regression of production of offspring. Evidence for CSEs population age ratios in post-breeding currently appears stronger for geese than for censuses against precipitation in the previous ducks (Table 1), but we suspect that this in winter, demonstrate ecological links between part represents the greater ability to monitor landscapes separated sometimes by lengthy individual geese throughout the annual cycle time lags of several months or great than is the case for ducks. distances of thousands of kilometres. For Evidence for CSEs linking spring example, winter precipitation can influence migration and breeding are especially strong habitat, and thus feeding conditions for (Table 1). Studies using stable isotopes wintering waterfowl that in turn affect the demonstrate that nutrients acquired away production of young at the population level from breeding areas contribute to egg as indexed by age ratios during the following production in numerous species of ducks hunting season (Heitmeyer & Fredrickson and geese (Table 1). In some cases 1981; Raveling & Heitmeyer 1989). Boyd et contributions of endogenous nutrients to al. (1982) found that winter precipitation in eggs were relatively modest (< 30% of the the U.S. along the Gulf of Mexico affected total). Such modest contributions to egg the number of adult Lesser Snow Geese and formation may, however, provide essential their breeding success to a greater extent supplements to dietary nutrients if the latter than spring temperatures and precipitation in are insufficient to meet the daily needs of the Dakotas and southern Manitoba, which females for egg production. Such needs can are used during spring migration. More be substantial during early clutch formation, recently, winter climate has been related to when yolk and albumin formation overlap age-ratios in Barnacle Geese Branta leucopsis (Alisauskas & Ankney 1992). Drent & Daan (Trinder et al. 2009), pre-breeding body mass (1980) considered variation in reliance on of Common Eiders Somateria mollissima nutrient reserves for egg formation along a (Descamps et al. 2010), as well as their capital-income continuum. Some studies breeding propensity (Jónsson et al. 2009) and have characterised egg formation as breeding success (Lehikoinen et al. 2006). income-based if more than half of egg Correlations between expansion of cereals nutrients supplied were exogenous. and other agricultural crops in both North However, the pervasiveness of nutrient America and Europe, and the growth of reserve use (particularly of fat) in waterfowl goose populations on both continents, (Ankney & Alisauskas 1991; Alisauskas & support inferences from earlier regression- Ankney 1992) suggests to us that usage of based studies about the importance of food nutrient reserves might be largely obligatory, supply to population growth. Finally, or at least highly adaptive, and so any usage

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 283 e . 1984; Cooch et al . et al . 1984; Cooch 1991a; Aubry et al . 2013 Reference Species Examples of These cross-seasonal and carry-over examples are restricted to cases where one of effects in waterfowl. two staging area clutches isotopes associated with increased corn production size clutch isotopes stable isotopes offspring Females with greatermigration fat before spring Females earlier and larger laid Endogenous reserves > 20% of provided egg based on stable nutrients Goose Greater Snow Bêty et al . 2003 Increased lipid stores on spring staging area and population increase Goose Greater Snow Gauthier et al . 2003 correlationendogenous and between lipid and protein on arrival Positive Goose Greater Snow Gauthier et al . 2005 Goose Lesser Snow 55% ofEndogenous provided nutrients in eggs nutrients based on stable 1978 & MacInnes Ankney Goose Lesser Snow 25–36% ofEndogenous provided nutrients in eggs nutrients based on Hobson et al . 2011 goslingsEarlier hatching are larger and are recruitedat a higher rate Goose Lesser Snow Sharp et al . 2013 Goose Lesser Snow Cooke Table 1. Table demograph state in one season and individual ic individual demonstrated a direct link between met: 1) investigators conditions was state at th habitat conditions or average demonstrated a relationship between rates in a subsequent season; or 2) investigators Carry-over effect effects level Individual hunting on in springs with size lower Breeding propensity and clutch Goose Greater Snow Mainguy et al . 2002 population level and population demographicpopulation level season. rates in a subsequent Migration strategy and condition during spring related to production of Pink-footed Goose Madsen 2001

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 284 Cross-seasonal effects on waterfowl et al . 2012 et al . 1995; et al . 2011 et al . 2003 Sedinger Sedinger & Chelgren 2007 Schamber Dzus & Clark 1998 Reference Black BrantBlack Mallard Sedinger Dark-bellied Brent Spaans et al . 2007 Species Barnacle Goose1989 & Black Owen ( continued ). year reproductive success reproductive to nests attentive propensity stable isotopes Breeding affected quality of wintering area Use of particular brood rearing areas affects breeding probability the next Brant Black earlier are recruited that hatch at a higher rate Females Nicolai & Sedinger 2012 Greater nutrient reservesGreater nutrient related to and earlier migration positively to breeding area laid more eggs females on arrival Heavier more and were Barnacle Goose BarnacleGoose with greater reservesnutrient in spring produced more offspringFemales Endogenous used during egg nutrients and incubation laying et al . 2012 Prop Tombre Dark-bellied Brent goslingsEarlier hatching survive better and are recruited at a higher rate Ebbinge1995 Spaans & Brant Black Increased sea surface temperatures on wintering area reduced breeding Winter location affected breeding propensity and timing of nesting Brant Black Sedinger & Flint 1991; Sedinger et al . 2006 Brant Black Sedinger et al . 2011; Larger goslings survive autumn migration at a higher rate Table 1 Table Carry-over effect More than 50% of in eggs nutrients from endogenous sources based on Canada Goose Sharp et al . 2013

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 285 et al . 2005 et al . 2004 et al . 2008 et al . 1982 Hobson Reference Devries Black DuckBlack Redhead 2002 Barboza and Jorde Species ( continued ). broods and had low lymphocyte levels had lower return rates the next year had lower lymphocyte levels broods and had low next year stable isotopes staging areas weather propensity, nested earlier and laid larger clutches propensity, Immunosuppression by low body mass females. Females that abandoned that Females body mass females. low by Immunosuppression Common Eider size the function and clutch Increased incubation demand reduced immune Common Eider Hanssen et al . 2003 Wintering location related to timing of size to clutch nesting and weakly Endogenous contributed significantly to egg nutrients production based on Hanssen et al . 2005 King Eider King Eider effects level Population Increased commodity acreagecorrelated with increased population size Oppel et al . 2010 Mehl et al . 2004 related to mean body mass and fat in spring age-ratios positively Autumn breeders in years with drought on spring young fewer Smaller clutches Goose Lesser Snow Goose Lesser Snow Abraham Alisauskas 2002 ofEstimated number correlated adults and young with previous winter Goose Lesser Snow 1983 & Cooke Davies Goose Lesser Snow Boyd Females fasted in winter were heavier but laid fewer eggs but laid fewer heavier winter were fasted in Females >30% of protein in eggs from endogenous source in one of three yearsEndogenous used for maintenance during egg nutrients laying Lesser Scaup Cutting et al . 2011 Table 1 Table Carry-over effect with greater reserves nutrient had greater on arrival Females breeding Mallard

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 286 Cross-seasonal effects on waterfowl 1981 Reference Heitmeyer & Fredrickson fronted Goose Species Barnacle Goose et al . 2009 Trinder Common Eider Jónsson et al . 2009 Common Eider Jónsson et al . 2009 ( continued ). age-ratios breeding following wintering areas mass of and subsequent age adult females at hatch, ratios nesting populations Increased age-ratios with increased availability ofIncreased age-ratios with increased availability agricultural foods on Greenland White- et al . 2005 Fox Table 1 Table Carry-over effect reservesIncreased disturbance reduced nutrient during spring staging and Pink-footed Goose Drent et al . 2003 Age ratios positively correlatedAge with winter temperature ratios positively related to age ratios the next autumnWinter precipitation positively related to age ratios the next autumnWinter precipitation positively Plumagepair formation quality during among moulting areas varied Mallard Winter index of correlated North with body Atlantic Oscillation positively Northern Pintail Common Eider warm-wet winters More females nested following Green-winged Teal separate related to prenesting body condition in two positively NAO & Heitmeyer 1989 Raveling Legagneaux et al . 2012 et al . 2006 Lehikoinen warm-wet winters More females nested following Common Eider Descamps et al . 2010

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 287 might be properly considered as a capital individuals outside the breeding season and breeding strategy. the difficulty of monitoring non-breeding Numerous studies have shown that individuals that typically migrate away from nutritional state in spring affects success in breeding areas to moult. Nevertheless, there producing offspring (Table 1). Similarly, are a priori reasons to expect that such CSEs earlier migrants typically are more likely to exist and empirical evidence exists for such produce offspring (e.g. Prop et al. 2003). CSEs. Because these studies were without Female waterfowl invest substantial experimental manipulation and so primarily nutrients into both egg formation and observational, we cannot rule out the incubation (Alisauskas & Ankney 1992; potential for spurious correlation between Afton & Paulus 1992; Alisauskas & DeVink condition, or timing of migration, and 2015) and these nutrients must be reproductive success; that is, other replenished before autumn migration unmeasured variables may have been (Sedinger & Bollinger 1987). Post-breeding associated with both nutritional status and storage of nutrients occurs simultaneously production of offspring, but without the with brood rearing and moult. Increased former variate directly influencing the latter. vigilance and attentiveness of adults toward Two kinds of experiments, however, suggest their offspring may be at the expense of that nutritional status can have an important reduced foraging effort (Schindler & influence on reproductive success in Lamprecht 1987; Sedinger & Raveling 1990; waterfowl. First, implementation of a spring Sedinger et al. 1995), potentially reducing hunt on a major spring staging area for nutrient intake. The simultaneous need to Greater Snow Geese C. caerulescens atlanticus, grow feathers and restore nutrient reserves reduced nutritional status of staging geese, has the potential to reduce feather quality, associated with increased disturbance (Féret rate of restoration of nutrient reserves or et al. 2003), and reduced the proportion of both. Schmutz & Ely (1999) and Eichholz females that attempted to nest and (2001) each found that autumn survival was reproductive investment by those females related to nutritional status, so the ability of that did nest (Mainguy et al. 2002; Bêty et al. adults to restore nutrients after hatch has 2003). Second, reduced availability of implications for fitness later in the year. agricultural habitats for Pink-footed Geese Breeding success in geese can also Anser brachyrhynchus during spring migration influence events during the non-breeding in Norway reduced both nutritional status of season because family cohesion persists for migrating geese and their subsequent a full year after hatch and family groups are reproductive success (Drent et al. 2003). dominant to pairs without young or singles CSEs on adults from breeding to autumn (Raveling 1970; Prevett & MacInnes 1980; or winter are generally less well understood Lamprecht 1986; Poisbleau et al. 2006). than those linking other seasons. Social status is related to position in foraging Limitations in our understanding reflect the flocks and food intake (Boyd 1953; Gregoire difficulty of monitoring highly mobile & Ankney 1990; Black et al. 1992); thus,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 288 Cross-seasonal effects on waterfowl successful reproduction has the potential to relative initiation date of nests that produced influence nutrient dynamics the following King Eider S. spectabilis ducklings, with winter through its effect on social status. equivocal support for an additive negative Consistent with this hypothesis, Sedinger et influence of spring thaw date on the age that al. (2011) showed that Black Brant B. bernicla ducklings were recruited as breeders. Both nigricans that bred tended to shift to higher studies support the notion that events early quality wintering areas the following winter. in life can influence recruitment and life- Similar mechanisms could occur in ducks if time reproductive success. more “attractive” individuals pair earlier in Investments in breeding by female autumn and experience enhanced social waterfowl may result in lagged trade-offs status as a result. Gregoire & Ankney (1990) with subsequent breeding or with other also noted that family structure varied by life-history traits. Viallefont et al. (1995) winter habitat, so that any COEs of demonstrated that Lesser Snow Geese nutrition related to previous breeding breeding for the first time were less likely to success and resulting family size during breed the following year, suggesting the winter on subsequent breeding may interact existence of costs associated with these first with the types of winter habitats occupied. breeding attempts. Specific brood-rearing Finally, numerous studies have areas used by female Black Brant influenced demonstrated COEs based on experience their likelihood of breeding the next year during their first summer for young (Nicolai & Sedinger 2012). Assessment waterfowl, especially geese. Growth of breeding costs is complicated by conditions during the first summer heterogeneity in individual quality (van influence survival during the next autumn Noordwijk & de Jong 1986) and variation in (Owen & Black 1989; Hill et al. 2003; the probability of breeding among Sedinger & Chelgren 2007; Aubry et al. individuals (Sedinger et al. 2008; Hoye et al. 2013) and ultimately, recruitment into the 2012). Additionally, as noted above, breeding population (Cooke et al. 1984; successful breeding by geese enhances their Sedinger et al. 2004). An important social status (Poisbleau et al. 2006), which determinant of survival probability for may compensate for some of the costs of goslings over the year after they fledge is breeding (e.g. Sedinger et al. 2011). Leach & phenology of hatch, both within years and Sedinger (unpubl. data) manipulated brood between years (Cooch 2002; Slattery & size and found that removing broods can Alisauskas 2002; Aubry et al. 2013). Slattery negatively affect probability of breeding the & Alisauskas (2002) found that dispersal following year, but increasing brood size can distance by goslings after hatching also reduce breeding the following year for influenced their subsequent survival larger initial brood sizes. These results are probability over the next year. Dzus & Clark generally consistent both with the social (1998) showed that hatch date influenced advantages of breeding in geese, and the recruitment in Mallard, and Alisauskas & idea that successful breeding can induce Kellett (2014) found a negative effect of costs for future breeding.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 289

Drake (2006) experimentally manipulated Geese constituting the northernmost nest success of Ross’s Geese C. rossii and vanguard of the mid-continent Lesser Snow found that survival of successful nesters was Goose population during spring migration consistently lower than those with destroyed were fatter than conspecifics farther south nests, and that the strength of this effect at the same calendar date. Earlier arrival varied among years, consistent with COEs during migration increases the potential for of successful breeding on subsequent protracted residency at each staging area, survival. However, we note that there is a which may permit greater nutrient general paucity of studies of CSEs on the acquisition (Drent et al. 2003). The result is survival of breeding adults, and suggest that that individuals that store high levels of the lack of such results reflects logistical or nutrients early at any staging area tend to technical difficulties in assessing the survival maintain this advantage throughout spring of adults in breeding versus non-breeding migration, resulting in higher probability of states. Because non-breeding waterfowl early arrival onto nesting areas, earlier typically undergo moult migrations to areas breeding and larger clutches (Prop et al. remote from breeding habitats (Hohman et 2003; Drent et al. 2003; Bêty et al. 2003). al. 1992), it is uncommon that samples of It is unlikely that the same nutrients stored both breeders and non-breeders are in winter or early in migration actually monitored simultaneously. As illustrated by contribute directly to reproduction because Drake (2006), this can be remedied to a the substantial energy costs of long-distance certain extent by experimentation. The migratory flight result in a major turnover limited evidence for absence of evidence for of nutrients throughout spring migration. CSEs on survival should, thus, be viewed as Although most assessments of the an absence of appropriate data, rather than contribution of endogenous nutrients to egg an absence of such CSEs. production distinguish between body reserves acquired by the female before arrival on the Mechanisms for cross-seasonal breeding grounds (i.e. “capital” breeders) and effects food acquired locally on or near the breeding territories (for “income” breeders: Drent & The most direct mechanism underlying Daan 1980; Gauthier et al. 2003; Hobson et al. CSEs in waterfowl, especially those linking 2004; Cutting et al. 2011), they do not assign a winter and spring migration areas to geographic origin to endogenous nutrients. As breeding, relates directly to nutritional state, suggested by Klaassen et al. (2006), it is and is, therefore, related to individual important to distinguish between distant versus abilities to store more nutrients earlier than local capital breeding (i.e. whether intensive others (Prop et al. 2003; Drent et al. 2003). feeding in preparation for breeding is a few Individual variation in schedules of nutrient kilometres from the breeding areas or at more storage induces and governs variation in distant sites), in addition to determining departure phenology. For example, whether the birds use capital or income Alisauskas (1988) found that Lesser Snow breeding strategies.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 290 Cross-seasonal effects on waterfowl

The substantial investments required for al. 2004) influence feather colour in egg-laying by waterfowl create the potential passerines. Legagneaux et al. (2012) for COEs, yet few examples of such COEs demonstrated that the iridescent colouration exist. We exclude costs of breeding, such as in Green-winged Teal was influenced by the reduced survival, that occur during the location where individuals moulted, breeding season (Madsen et al. 2002; Arnold consistent with the hypothesis that nutrient et al. 2010) because such trade-offs, while availability influenced feather quality and real, do not represent COEs. Clearly, the perhaps attractiveness to mates. Feathers of ability of breeding adults to restore yearling Black Brant during their second nutrients depleted during breeding may summer are highly degraded relative to influence fitness during seasons following those of adults (J. Sedinger pers. obs.). breeding. Several studies document a Because development of these feathers relatively slow replenishment of nutrients occurred during the growth period, when before the post-breeding moult (e.g. Ankney nutrient availability is typically limiting tissue & MacInnes 1978; Ankney 1984; Fox et al. production (e.g. Sedinger 1984; Sedinger et 2013). Non- and failed-breeding Black Brant al. 2001), the poor quality of feathers on gained more mass before moult than did yearling geese is consistent with the notion adults with broods (Fondell et al. 2013), that nutrient availability could influence suggesting the possibility that mass gain feather quality, as is the case in other birds following nesting might be constrained by (Butler et al. 2008). Absence of data on competing demands of tending broods. feather quality and the relationship between Virtually all species lose mass during moult feather quality and fitness in waterfowl have (Hohman et al. 1992), and while there is hindered assessment of linkages between debate about whether such mass loss is feather quality and fitness (COE) or adaptive (e.g. Fox & Kahlert 2005), adult population dynamics (CSE) of waterfowl. waterfowl do not restore nutrient reserves at We suggest that such research is warranted. this time of year. Thus, investments in Immune function and associated disease reproduction, combined with constraints on status both provide mechanisms that could the ability to replenish these nutrients facilitate CSEs. Long-distance migrants following breeding, have the potential to appear to be more exposed, or more influence fitness in seasons following susceptible, to parasitic infection (Figuerola breeding. & Green 2000). Two studies demonstrate Waterfowl undergo complete remigial that parasite load during brood-rearing has a moult and become flightless at the same negative influence on the survival of young time that they are restoring nutrients after fledging (Slattery & Alisauskas 2002; depleted during breeding, which has the Souchay et al. 2013). Nematode levels were potential to influence feather quality as well negatively associated with lipid levels for as somatic nutrient levels. Both moult Lesser Snow Geese during spring migration location (Norris et al. 2007) and competing (Shutler et al. 2012). The authors could not nutritional demands during moult (Norris et differentiate between the hypotheses that:

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 291

1) nematode infestation inhibited lipid migration habitat is limited (Galsworthy et deposition; 2) some individuals were al. 2011; Gaidet et al. 2012). chronically infected, which reduced the size Investment in reproduction can reduce of nutrient stores; and 3) individuals that immune function in waterfowl (Hanssen et invested more in lipid storage and less in al. 2003, 2005) and other birds (Deerenberg immune function were more susceptible et al. 1997; Schmidt-Hempel 2003), with the to nematode infection. The first two potential for lower survival or reduced hypotheses would be consistent with COEs reproductive success in the future. Immune associated with disease. challenges in Mallard (post-fledging) Cholera Pasturella multicida has been influences colour preference and response known for some time to cause substantial to novel environments (Butler et al. 2012), as mortality on both breeding and wintering well as the ability to mount an immune areas (Rosen & Bischoff 1949; Wobeser response as adults (Butler & McGraw et al. 1983). The dynamics of Cholera 2012). Although Norris & Evans (2000) transmission are not well understood but demonstrated that investment in breeding evidence suggests that individuals serve as affected immune function, which in turn carriers (Samuel et al. 1999, 2004, 2005), influenced infection rate and reduced creating the potential that exposure of fitness, such studies are rare. The presence individuals in one seasonal habitat can lead of complex sequential causative events to disease transmission and mortality in relating individual variation in immune another season, with possible population function to variation in fitness could level consequences. Sub-lethal roles of translate to population level effects if disease beyond immediate mortality effects pervasive. We believe studies that address are also not well understood. Infection by such mechanisms underlying CSEs in low-pathogenic avian influenza has been waterfowl may be relevant to a full associated with less efficient feeding and understanding of population dynamics. delayed migration in Bewick’s Swans Cygnus columbianus bewickii (van Gils et al. 2007). Implications of cross-seasonal Latorre-Margalef et al. (2009) found that effects for management Mallard infected with avian influenza had lower body mass during migration, which The potential for CSEs was implicit in the could influence timing of migration and development of winter habitat joint investment in reproduction. The association ventures in North America during the between the timing of migration and 1980s, as suggested by references to specific reproductive success (see above) suggests links between habitat and waterfowl that infection could cause lower population objectives in joint venture productivity. Such sub-lethal effects are documents (e.g. Central Valley Habitat Joint likely to be of broader importance in Venture 1990). In some cases, specific waterfowl, especially when individuals mechanisms behind assumed CSEs were not are concentrated because wintering or made explicit until sometime later (Koneff

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 292 Cross-seasonal effects on waterfowl

2006 unpublished letter to the North influence survival rates (e.g. Slattery & American Waterfowl Management Plan Alisauskas 2002) and migration behaviour evaluation team). Establishment of the (van Gils et al. 2007; Latorre-Margalef et al. joint ventures stimulated substantial new 2009). Thus, availability of winter habitat research into dynamics of food availability not only has the potential to influence CSEs, in both wetlands (Naylor 2002; Hagy & but habitat management on the wintering Kaminski 2012a,b; Leach et al. 2012) and grounds could influence disease processes agricultural habitats (Stafford et al. 2010; and so affect the dynamics of waterfowl Foster et al. 2010), which typically populations. demonstrated the potential for food to Harvest pressure influences behaviour become limiting by December and January (Fox & Madsen 1997; Webb et al. 2011) and (Naylor 2002; Foster et al. 2010). These habitat use (Béchet et al. 2003; Moore & modern studies built on earlier estimates of Black 2006), which in turn can affect food abundance and carrying capacity of nutritional status and investment in winter habitats for waterfowl in Europe breeding, at least in geese (Mainguy et al. (Ebbinge et al. 1975; Owen 1977) and North 2002). Direct effects of harvest on mortality America (Givens et al. 1964). Demonstration rates and population dynamics vary as a of winter food depletion, combined with function of body size (Rexstad 1992; increases in many goose populations Sedinger et al. 2007; Sedinger & Herzog associated with increased availability of 2012; Péron et al. 2012). Indirect effects of agricultural foods (Abraham et al. 2005; Fox harvest acting through nutrient dynamics et al. 2005; Gauthier et al. 2005), indicate that are poorly understood for ducks and most agroecosystems or direct management of populations of geese (but see Pearse et al. winter habitats for waterfowl has great 2012). Nevertheless, examples cited here potential to influence dynamics of their indicate that the effects of harvest are likely populations. Refining winter habitat to be more complex in all waterfowl than are management will require improved currently assumed. understanding of the interplay among Density-dependent effects on winter food abundance, winter population recruitment are well established in both density, COEs, availability of breeding ducks (Kaminski & Gleussing 1987) and habitat and breeding population density (e.g. geese (Loonen et al. 1997; Lake et al. 2008) Runge & Marra 2005; Norris & Marra although mechanisms are less well 2007). We contend that methods now exist understood for ducks. In geese, a substantial to improve our understanding of all of proportion of density-dependence occurs these questions (see below). through food limitation during growth Disease transmission increases at higher (Cooch et al. 1991b; Sedinger et al. 1998) and densities (Rosen & Bischoff 1949; Gaidet et reduced survival during the first year al. 2012), so the availability of winter habitat associated with small body size (Sedinger can potentially influence rates of infection & Chelgren 2007; Sedinger & Nicolai at the population level. Infection may 2011). Thus, density-dependent effects on

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 293 recruitment in geese reflect a CSE/COE cohorts and per capita food abundance. Thus, effect on each year’s cohort of young. A it seems that populations may include a number of years of predator removal in the relatively fixed proportion of “low quality” North American prairies has resulted in individuals, incapable of response to substantially increased nest success (Pieron density-dependent habitat conditions. & Rohwer 2010). Increased production of Assessment of most habitat joint ducklings, however, has not produced any ventures has been based on waterfowl increase in local densities (Amundson et al. response in the season for which a specific 2013). A reasonable hypothesis for the lack habitat is managed. For winter habitat joint of a population response to increased ventures, effectiveness is often measured in duckling production is that increased brood predicted “use-days”, based on measures of densities, resulting from higher nest success, food abundance and models of nutritional increased competition for food, and slower requirements (e.g. Stafford et al. 2006). growth and lower post-fledging survival of Alternatively, nest success is used to assess ducklings, similar to mechanisms reported habitat programmes on breeding areas (e.g. for geese. Stephens et al. 2005). These approaches cannot assess the true value of habitat in the Remaining issues and context of the complete annual cycle of approaches to estimating waterfowl, however, because they do not cross-seasonal effects account for CSEs, which may modulate the response of individual birds to Much of our understanding of CSEs is concurrent habitat conditions. Moreover, an based on individual responses to an understanding of the role that CSEs play in interplay between physiological states and the dynamics of waterfowl populations environmental conditions (Table 1). An requires knowledge about proportional understanding of how individual effects habitat use by specific populations at any translate into population level responses will point in the year, and how this varies over require an appreciation of how population the annual cycle. Although challenging, such density feeds back on individual responses information would permit an assessment (Runge & Marra 2005; Norris & Marra of the reproductive success of individuals 2007). Higher density could, for instance, in relation to management actions on induce shifts in the mean and variance of previously used wintering areas. The individual states (Fig. 1). An important situation may be reinforced by individuals determinant of whether individual response that fledged in a particular breeding habitat has relevance to population level effects is migrating to, and also wintering in, the same the degree to which heterogeneity exists in habitats that had a role in their production. response to population density. For It is unclear how such COEs may have example, Lindberg et al. (2013) did not influenced the evolution of both breeding detect a relationship between the proportion and winter philopatry, but the adaptiveness of “low quality” individuals in Black Brant of occupying high quality habitats which

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 294 Cross-seasonal effects on waterfowl enables pre-breeding ducks to improve their from an interplay of winter food condition probably also has relevance for distribution and density-dependent effects the survival of the new recruits to the expressed on the lifelong morphology of population and the likelihood of them geese from events experienced as growing joining the breeding cohort in due course. goslings in the arctic. Such a scenario may have played a role in the Studies of geese have been especially evolution of persistent family bonds in successful in identifying cross-seasonal and geese. cross-ecosystem linkages because of the As mentioned above, spring phenology relative ease of marking and observing them of arctic snow melt governs the timing of at multiple sites throughout the year. Ross’s nesting, hatch and gosling growth for all Geese and Snow Geese are highly arctic-nesting geese and also for many of gregarious, very abundant and occur at very the sea duck species. High plasticity in high densities throughout the year, which growth rates is an apparent adaptation to poses great challenges in marking sufficient high variability in nutritional supply to numbers so that individuals can be detected goslings both within and among years, and studied. Ring recoveries provide a permitting nutritional flexibility in years of mechanism for linking together breeding, nutritional stress, although at the cost of migration and wintering habitats (e.g. reduced survival and adult body size when Alisauskas et al. 2011). We believe that many nutrients are limited during the growth waterfowl scientists assume similar period. Such intraspecific variation in body approaches are not possible for diving or size and morphology influences not only dabbling ducks (genera Aythya and Anas) for subsequent survival, but can influence a number of reasons, including insufficient habitats and broad-scale landscapes site fidelity and the difficulty of using occupied during the following winter. markers that can be detected at a distance. In Alisauskas (1998) found that Lesser Snow our view, this is too pessimistic. Several Geese in their traditionally used coastal studies of marked ducks on breeding areas marsh habitats during winter were larger demonstrated excellent success at re- than conspecifics from inland agricultural encountering breeding females marked in landscapes associated with rice agriculture, earlier years (Sowls 1955; Arnold & Clark both of which, were larger than Lesser 1996; Anderson et al. 2001; Blums et al. Snow Geese wintering farther north near 2002), indicating substantial fidelity to the Missouri River Valley of Iowa and relatively small, well-defined breeding areas. Nebraska. Alisauskas (1998) suggested that We acknowledge it may be difficult or perhaps large scale expansion of winter impossible to encounter individuals from range associated with the correlation these kinds of studies on migration or at between body size and bill morphology may wintering sites, as has been undertaken for have been coupled with density-dependent geese. The emergence of new technologies morphological change on breeding areas. should, however, improve the ability to Thus winter range dynamics appear to result identify links between breeding, staging and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 295 wintering areas. Stable isotopes in tissues strategies. We are beyond the demonstration grown during known time periods provide phase; we suggest that it is now time to one mechanism for assigning individuals to move waterfowl biology forward. habitats remote from the breeding range (Hobson et al. 2004; Clark et al. 2006), or Acknowledgements winter areas (Mehl et al. 2005). This Sedinger was supported by the College of approach can take advantage of specific Agriculture, Biotechnology and Natural feather tracts moulted outside the breeding Resources. His long-term studies of Black season (Smith & Sheeley 1993; Combs & Brant have been supported by the National Fredrickson 1996) or toe nails which are Science Foundation (grants OPP 92 14970, grown continuously (Clark et al. 2006). DEB 98 15383, OPP 99 85931, OPP 01 Geolocators, which record the timing of 96406, DEB 0743152, DEB 12 52656), sunrise and sunset, to allow reconstruction Alaska Science Center, U.S. Geological of latitude and longitude, provide an Survey, Ducks Unlimited, the Black Brant alternative approach (e.g. Eichhorn et al. Group (Morro Bay, California), Phil Jebbia 2006). Both geolocators and stable isotopes (in memory of Marnie Shepherd), and require that individuals be captured but the Migratory Bird Management (Region 7) of studies cited demonstrate that large the U.S. Fish and Wildlife Service. numbers of breeding females can be Alisauskas’ work has been supported by captured during nesting. The addition of the California Department of Fish & either geolocators or stable isotope methods Game, the Polar Continental Shelf Project, to traditional marking and recapture on Delta Waterfowl, Ducks Unlimited Canada, long-term study sites could enhance the University of Saskatchewan, the substantially knowledge of CSEs for species Central and Mississippi Flyway Councils, other than geese. Environment Canada and the U.S. Fish and Our understanding of the role of CSEs Wildlife Service. The manuscript benefitted in both the dynamics of populations and from comments by D. Koons and C. fitness of individuals has increased Nicolai. dramatically over the past three decades. References This information and new technologies are now sufficiently developed to allow more Abraham, K.F., Jefferies, R.L. & Alisauskas, R.T. detailed assessments of the role of seasonal 2005. The dynamics of landscape change and habitats in the dynamics of specific snow geese in mid-continent North America. Global Change Biology 11: 841–855. populations. We believe it is now possible to Afton, A.D. & Paulus, S.L. 1992. Incubation and use modern methods in the context of brood care. In B.D.J. Batt, A.D. Afton, M.G. CSEs to improve and prioritise management Anderson, C.D. Ankney, D.H. Johnson, J.A. of seasonal habitats. We also believe it is Kadlec & G.L. Krapu (eds.), Ecology and now possible to incorporate events and Management of Breeding Waterfowl, pp. 62–108. decisions throughout the complete annual University of Minnesota Press, Minneapolis, cycle into our understanding of life-history Minnesota, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 296 Cross-seasonal effects on waterfowl

Alisauskas, R.T. 1988. Nutrient reserves of Lesser removal on mallard production and population Snow Geese during winter and spring change in northeastern North Dakota. Journal migration. Ph.D. thesis, University of of Wildlife Management 77: 143–152. Western Ontario, London, Canada. Anderson, M.G. & Batt, B.D.J. 1983. Workshop Alisauskas, R.T. 1998. Winter range expansion on the ecology of wintering waterfowl. and relationships between landscapes and Wildlife Society Bulletin 11: 22–24. morphometrics of midcontinent Lesser Anderson, M.G., Lindberg, M.S. & Emery, R.B. Snow Geese. Auk 115: 851–862. 2001. Probability of survival and breeding Alisauskas, R.T. 2002. Arctic climate, spring for juvenile female Canvasbacks. Journal of nutrition, and recruitment in midcontinent Wildlife Management 65: 385–397. lesser snow geese. Journal of Wildlife Ankney, C.D. & Alisauskas R.T. 1991. Nutrient- Management 66: 181–193. reserve dynamics and diet of breeding Alisauskas, R.T. & Ankney, C.D. 1992. The cost female Gadwalls. Condor 93: 799–810. of egg laying and its relationship to nutrient Ankney, C.D. & MacInnes, C.D. 1978. Nutrient reserves in waterfowl. In B.D.J. Batt, A.D. reserves and reproductive performance of Afton, M.G. Anderson, C.D. Ankney, D.H. female Lesser Snow Geese. Auk 95: 459–471. Johnson, J.A. Kadlec, & G.L. Krapu (eds.), Arnold, T.W. & Clark, R.G. 1996. Survival and Ecology and Management of Breeding Waterfowl, philopatry of female dabbling ducks in pp. 30–61. University of Minnesota Press, southcentral Saskatchewan. Journal of Wildlife Minneapolis, Minnesota, USA. Management 65: 385–397. Alisauskas, R.T. & DeVink, J.-M. 2015. Dealing Arnold, T.W., Roche, E.A., Devries J.H. & with deficits: breeding costs, nutrient Howerter, D.W. 2010. Costs of reproduction reserves and cross-seasonal effects in in breeding female mallards: predation risk seaducks. In D.V. Derksen, J.-P. Savard, J. during incubation drives annual mortality. Eadie & D. Esler (eds.), Ecology and Avian Conservation and Ecology 7: 1. Conservation of North American Sea Ducks, pp. Aubry, L.M., Rockwell, R.F., Cooch, E.G., 000–000. Studies in Avian Biology No. 00, Brooks, R.W., Mulder, C.P.H. & Koons, D.N. Cooper Ornithological Society, c/o 2013. Climate change, phenology, and habitat Ornithological Societies of North America, degradation: drivers of gosling body Waco, Texas, USA. condition and juvenile survival in lesser snow Alisauskas, R.T. & Kellett, D.K. 2014. Age- geese. Global Change Biology 19: 149–160. specific in situ recruitment of female King Barboza, P.S. & Jorde, D.G. 2002. Intermittent Eiders estimated with mark-recapture. Auk fasting during winter and spring affects body 131: 129–140. composition and reproduction of a Alisauskas, R.T., Rockwell, R.F., Dufour, K.W., migratory duck. Journal of Comparative Cooch, E.G., Zimmerman, G., Drake, K.L., Physiology B 172: 419–434. Leafloor, J.O., Moser, T.J. & Reed, E.T. Béchet, A., Giroux, J.-F., Gauthier, G., Nichols, 2011. Harvest, survival and abundance of J.D. & Hines, J.E. 2003. Spring hunting midcontinent lesser snow geese relative to changes the regional movements of population reduction efforts. Wildlife migrating greater snow geese. Journal of Monographs 179: 1–42. Applied Ecology 40: 553–564. Amundson, C.T., Pieron, M.R., Arnold, T.W. & Bêty, J., Gauthier, G. & Giroux, J.-F. 2003. Beaudoin, L.A. 2013. The effects of predator Body condition, migration, and timing of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 297

reproduction in snow geese: a test of the (dD, d13C, d15N, d34S) composition of condition-dependent model of optimal feathers and claws from lesser scaup and clutch size. American Naturalist 162: 110–121. northern pintail: implications for the studies Black, J.M., Carbon, C., Wells, R.L. & Owen, M. of migratory connectivity. Canadian Journal of 1992. Foraging dynamics in goose flocks: the Zoology 84: 1395–1401. cost of living on the edge. Animal Behaviour Combs, D.L. & Fredrickson, L.H. 1995. Molt 44: 41–50. chronology of male Mallards wintering in Blums, P., Nichols, J.D., Hines, J.E. & Mednis, A. Missouri. Wilson Bulletin 107: 359–365. 2002. Sources of variation in survival and Cooch, E.G. 2002. Fledging size and survival in breeding site fidelity in three species of snow geese: timing is everything (or is it?). European ducks. Journal of Animal Ecology 71: Journal of Applied Statistics 29: 143–162. 438–450. Cooch, E.G., Lank, D.B., Dzubin, A., Rockwell, Boyd, H. 1953. On encounters between R.F. & Cooke, F. 1991a. Body size variation wild white-fronted geese in winter flocks. in Lesser Snow Geese: seasonal variation in Behaviour 5: 85–129. gosling growth rate. Ecology 72: 503–512. Boyd, H., Smith, G.E.J. & Cooch, F.G. 1982. The Cooch, E.G., Lank, D.B., Rockwell, R.F. & Lesser Snow Geese of the eastern Canadian Cooke, F. 1991b. Long-term decline in body arctic. Canadian Wildlife Service Occasional Paper size in a snow goose population: evidence of No. 46. Canadian Wildlife Service, Ottawa, environmental degradation? Journal of Animal Canada. Ecology 60: 483–496. Butler, L.K., Rohwer, S. & Speidel, M.G. 2008. Cooke, F., Findlay, C.S. & Rockwell, R.F. 1984. Quantifying structural variation in contour Recruitment and the timing of reproduction feathers to address functional variation and in Lesser Snow Geese (Chen caerulescens life history trade-offs. Journal of Avian Biology caerulescens). Auk 101: 451–458. 39: 629–639. Cutting, K.A., Hobson, K.A., Rotella, J.J., Warren, Butler, M.W. & McGraw, K.J. 2012. J.M., Wainwright-de la Cruz, S.E. & Takekawa, Developmental immune history affects adult J.Y. 2011. Endogenous contributions to egg immune function but not carotenoid-based protein formation in lesser scaup Aythya affinis. ornamentation in mallard ducks. Functional Journal of Avian Biology 42: 505–513. Ecology 26: 406–415. Daan, S., Deerenberg, C. & Dijkstra, C. 1996. Butler, M.W., Toomey, M.B., McGraw, K.J. Increased daily work precipitates natural & Rowe, M. 2012. Ontogenetic immune death in the kestrel. Journal of Animal Ecology challenges shape adult personality in mallard 65: 539–544. ducks. Proceedings of the Royal Society B 279: Davies, J.C. & Cooke, F. 1983. Annual nesting 326–333. productivity in snow geese: prairie droughts Central Valley Habitat Joint Venture. 1990. Central and arctic springs. Journal of Wildlife Valley Habitat Joint Venture Implementation Management 47: 291–296. Plan. Sacramento, USA. Accessible at Deerenberg, C., Apanius, V., Daan, S. & Bos, N. http://www.centralvalleyjointventure.org/ 1997. Reproductive effort decreases antibody assets/pdf/cvjv_implementation_plan.pdf responsiveness. Proceedings of the Royal Society B (last accessed on 15 March 2014). 264: 1021–1029. Clark, R.G., Hobson, K.A. & Wassenaar, L.I. Descamps, S., Yoccoz, N.G., Gaillard, J.-M., 2006. Geographic variation in the isotopic Gilchrist, H.G., K. Erikstad, K., Hanssen,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 298 Cross-seasonal effects on waterfowl

S.A., Cazelles, B., Forbes, M.R. & Bêty, J. Féret, M., Gauthier, G., Béchet, A., Giroux, J.-F. 2010. Detecting population heterogeneity in & Hobson, K.A. 2003. Effect of a spring effects of North Atlantic Oscillations on hunt on nutrient storage by greater snow seabird body condition: get into the rhythm. geese in southern Quebec. Journal of Wildlife Oikos 119: 1526–1536. Management 67: 796–807. Devries, J.H., Brook, R.W., Howerter, D.W. & Figuerola, J. & Green, A.J. 2000. Haematozoan Anderson, M.G. 2008. Effects of spring body parasites and migratory behavior in condition and age on reproduction in Mallards waterfowl. Evolutionary Ecology 14: 143–153. (Anas platyrhynchos). Auk 125: 618–628. Fleskes, J.P., Yee, J.L., Yarris, G.S., Miller, M.R. & Drake, K.L. 2006. The role of dispersal Casazza, M.L. 2007. Pintail and mallard in population dynamics of breeding survival in California relative to habitat, Ross’s Geese. Ph.D. thesis, University of abundance, and hunting. Journal of Wildlife Saskatchewan, Saskatchewan, Canada. Management 71: 2238–2248. Drent, R.H. & Daan, S. 1980. The prudent Fondell, T.F., Flint, P.L., Schmutz, J.A., Schamber, parent: energetic adjustments in avian J.L. & Nicolai, C.A. 2013. Variation in body breeding. Ardea 68: 225–252. mass dynamics among sites in Black Brant Drent, R., Both, C., Green, M., Madsen, J. & Branta bernicla nigricans supports adaptivity of Piersma, T. 2003. Pay-offs and penalties of mass loss during moult. Ibis 155: 593–604. competing migratory schedules. Oikos 103: Foster, M.A., Gray, M.J. & Kaminski, R.M. 2010. 274–292. Agricultural seed biomass for migrating and Dzus, E.H. & Clark, R.G. 1998. Brood survival wintering waterfowl in the southeastern and recruitment of Mallards in relation to United States. Journal of Wildlife Management wetland density and hatching date. Auk 115: 74: 489–495. 311–318. Fox, A.D. & Kahlert, J. 2005. Changes in body Ebbinge, B.S., Canters, K. & Drent, R. 1975. mass and organ size during wing moult in Foraging routines and estimated daily food non-breeding greylag geese Anser anser. intake in Barnacle Geese wintering in the Journal of Avian Biology 36: 538–548. northern Netherlands. Wildfowl 26: 5–19. Fox, A.D. & Madsen, J. 1997. Behavioural and Ebbinge, B.S. & Spaans, B. 1995. The importance distributional effects of hunting disturbance of body reserves accumulated in spring on waterbirds in Europe: implications for staging areas in the temperate zone for refuge design. Journal of Applied Ecology 34: breeding in dark bellied brent geese Branta 1–13. bernicla bernicla in the high arctic. Journal of Fox, A.D., Madsen, J., Boyd, H., Kuijken, Avian Biology 26: 105–113. E., Norriss, D.W., Tombre, I.M. & Stroud, Eichholz, M.W. 2001. The implications of D.A. 2005. Effects of agricultural change agriculture in interior Alaska for population on abundance, fitness components and dynamics of Canada Geese. Ph.D. thesis, distribution of two arctic-nesting goose University of Alaska Fairbanks, Alaska, USA. populations. Global Change Biology 11: 881–893. Eichhorn, G., Afanasyev, V., Drent, R.H. & van Fox, A.D., King, R. & Owen, M. 2013. Wing moult der Jeugd, H.P. 2006. Spring stopover and mass change in free-living mallard Anas routines in Russian Barnacle Geese Branta platyrhynchos. Journal of Avian Biology 44: 1–8. leucopsis tracked by resightings and Gaidet, N., Caron, A., Cappelle, J., Cumming, geolocation. Ardea 94: 667–678. G.S., Balanc¸ G., Hammoumi, S., Cattoli, G.,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 299

Abolnik, C., Servan de Almeida, R., Gil, P., Hanssen, S. A., Folstad, I. & Erikstad, K.E. 2003. Fereidouni, S.R., Grosbois, V., Tran, A., Reduced immunocompetence and cost of Mundava, J., Fofana, B., Ould El Mamy, A.B., reproduction in common eiders. Oecologia Ndlovu, M., Mondain-Monval, J.-Y., Triplet, 136: 457–464. P., Hagemeijer, W., Karesh, W.B., Newman, Hanssen, S.A., Hasselquist, D., Folstad, I. & S.H. & Dodman, T. 2012. Understanding the Erikstad, K.E. 2005. Cost of reproduction in ecological drivers of avian influenza virus a long-lived bird: incubation effort reduces infection in wildfowl: a continental-scale immune function and future reproduction. study across Africa. Proceedings of the Royal Proceedings of the Royal Society B 272: 1039– Society B 279: 1131–1141. 1046. Galsworthy, S.J., ten Bosch, Q.A., Hoye, Harrison, X.A., Blount, J.D., Inger, R., Norris, B.J., Heesterbeek, J.A.P., Klaassen, M. & D.R. & Bearhop, S. 2011. Carry-over effects Klinkenberg, D. 2011. Effects of infection- as drivers of fitness differences in animals. induced migration delays on the epidemiology Journal of Animal Ecology 80: 4–18. of avian influenza in wild mallard populations. Heitmeyer, M. & Fredrickson, L.H. 1981. Do PLoS ONE 6: e26118. wetland conditions in the Mississippi Delta Gauthier, G., Bèty, J. & Hobson, K.A. 2003. Are hardwoods influence mallard recruitment? Greater Snow Geese capital breeders? New Transactions of the North American Wildlife and evidence from a stable-isotope model. Natural Resources Conference 46: 44–57. Ecology 84: 3250–3264. Hill, M.R.J., Alisauskas, R.T., Ankney, C.D. & Gauthier, G., Giroux, J.-F., Reed, A., Béchet, A. & Leafloor, J.O. 2003. Influence of body size Bélanger, L. 2005. Interactions between land and condition on harvest and survival of use, habitat use and population increase juvenile Canada geese. Journal of Wildlife in greater snow geese: what are the Management 67: 530–541. consequences for natural wetlands? Global Hobson, K.A., Atwell, L., Wassenaar, L.I. & Change Biology 11: 856–868. Yerkes, T. 2004. Estimating endogenous Givens, L.S., Nelson, M.C. & Ekedahl, V. 1964. nutrient allocations to reproduction in Farming for waterfowl. In J.P. Linduska (ed.), redhead ducks: a dual isotope approach using Waterfowl tomorrow, pp. 599–610. U.S. δD and δ13C measurements of female and Department of Interior, Washington, USA. egg tissues. Functional Ecology 18: 737–745. Gregoire, P.E. & Ankney, C.D. 1990. Agonistic Hobson, K.A., Sharp, C.M., Jefferies, R.L., behavior and dominance relationships Rockwell, R.F. & Abraham, K.F. 2011. among Lesser Snow Geese during winter and Nutrient allocation strategies to eggs by spring migration. Auk 107: 550–560. Lesser Snow Geese (Chen caerulescens) at a Hagy, H.M. & Kaminski, R.M. 2012a. Apparent sub-arctic colony. Auk 128: 156–165. seed use by ducks in moist-soil wetlands of Hohman, W.L., Ankney, C.D. & Gordon, D.H. the Mississippi Alluvial Valley. Journal of 1992. Ecology and management of Wildlife Management 76: 1053–1061. postbreeding waterfowl. In B.D.J. Batt, A.D. Hagy, H.M. & Kaminski, R.M. 2012b. Winter Afton, M.G. Anderson, C.D. Ankney, D.H. waterbird and food dynamics in autumn- Johnson, J.A. Kadlec & G.L. Krapu (eds.), managed moist-soil wetlands in the Ecology and Management of Breeding Waterfowl, Mississippi alluvial valley. Wildlife Society pp. 128–189. University of Minnesota Press, Bulletin 36: 512–523. Minneapolis, Minnesota, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 300 Cross-seasonal effects on waterfowl

Hoye, B.J., Hahn, S., Nolet, B.A. & Klaassen, M. Latorre-Margalef, N., Gunnarsson, G., Munster, 2012. Habitat use throughout migration: V.J., Fouchier, R.A.M., Osterhaus, A.D.M.E., linking individual consistency, prior breeding Elmberg, J., Olsen, B., Wallensten, A., success and future breeding potential. Journal Haemig, P.D., Fransson, T., Brudin, L. & of Animal Ecology 81: 657–666. Waldenström, J. 2009. Effects of influenza A Johnson, D.H., Nichols, J.D. & Schwartz, M.D. virus infection on migrating mallard ducks. 1992. Population dynamics of breeding Proceedings of the Royal Society B 276: 1029– waterfowl. In B.D.J. Batt, A.D. Afton, M.G. 1036. Anderson, C.D. Ankney, D.H. Johnson, J.A. Leach, A.G., Straub, J.N., Kaminski, R.M., Ezell, Kadlec & G.L. Krapu (eds.), Ecology and A.W., Hawkins, T.S. & Leininger, T.D. 2012. Management of Breeding Waterfowl, pp. 446–485. Apparent seed use by ducks in moist-soil University of Minnesota Press, Minneapolis, wetlands of the Mississippi Alluvial Valley. Minnesota, USA. Journal of Wildlife Management 76: 1519–1522. Johnson, F.A., Moore, C.T., Kendall, W.L., Legagneaux, P., Clark, R.G., Guillemain, M., Dubovsky, J.A., Caithamer, D.F., Kelley Jr., Eraud, C., Théry, M. & Bretagnolle, V. 2012. J.T. & Williams, B.K. 1997. Uncertainty and Large-scale geographic variation in iridescent the management of mallard harvests. Journal structural ornaments of a long-distance of Wildlife Management 61: 203–217. migratory bird. Journal of Avian Biology 43: Jónsson, J.E., Gardarsson, A., Gill, J.A., Petersen, 355–361. A. & Gunnarsson, T.G. 2009. Seasonal Lehikoinen, A., Kilpi, M. & Öst, M. 2006. Winter weather effects on the Common Eider, a climate affects subsequent breeding success subarctic capital breeder, in Iceland over 55 of common eiders. Global Change Biology 12: years. Climate Research 38: 237–248. 1355–1365. Kaminski, R.M. & Gluesing, E.A. 1987. Density- Loonen M.J.J.E., Oosterbeek, K. & Drent, R.H. and habitat-related recruitment in mallards. 1997. Variation in growth of young and adult Journal of Wildlife Management 51: 141–148. size in Barnacle Geese Branta leucopsis: Klaassen, M., Abraham, K.E., Jefferies, R.L. & evidence for density dependence. Ardea 85: Vrtiska, M. 2006. Factors affecting the site of 177–192. investment, and the reliance on savings for Lynch, J.J. 1952. Escape from mediocrity. arctic breeders: the capital–income Unpublished memorandum, U. S. Fish and dichotomy revisited. Ardea 94: 371–384. Wildlife Service, Abbeville Louisiana, USA. Kleiber, M. 1975. The Fire of Life an Introduction MacKinney, F. 1986. Ecological factors to Animal Energetics. Robert E. Krieger influencing the social systems of migratory Publishing Co., Huntington, New York, USA. dabbling ducks. In D.I. Rubenstein & R.W. Lake, B.C., Schmutz, J.A., Lindberg, M.S., Ely, Wrangham (eds.), Ecological Aspects of Social C.R., Eldridge, W.D. & Broerman, F.J. 2008. Evolution, pp. 153–174, Princeton University Body mass of prefledging Emperor Geese Press, Princeton, New Jersey, USA. Chen canagica: large-scale effects of Madsen, J. 2001. Spring migration strategies in interspecific densities and food availability. Pink-footed Geese Anser brachyrhynchus and Ibis 150: 527–540. consequences for spring fattening and Lamprecht, J. 1986. Structure and causation of the fecundity. Ardea 89: 43–55. dominance hierarchy in a flock of Bar-headed Madsen, J., Frederiksen, M. & Ganter, B. 2002. Geese (Anser indicus). Behaviour 96: 28–48. Trends in annual and seasonal survival of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 301

Pink-footed Geese Anser brachyrhynchus. Ibis Norris, D.R., Marra, P.P., Kyser, T.K., Ratcliffe, 144: 218–226. L.M. & Montgomerie, R. 2007. Continent- Mainguy, J., Bêty, J., Gauthier, G. & Giroux, J.-F. wide variation in feather colour of a 2002. Are body condition and reproductive migratory songbird in relation to body effort of laying Greater Snow Geese affected condition and moulting locality. Biology Letters by the spring hunt? Condor 104: 156–161. 3: 16–19. Marra, P.P., Hobson, K.A. & Holmes, R.T. 1998. Norris, D.R., Marra, P.P., Montgomerie, R., Linking winter and summer events in a Kyser, T.K. & Ratcliffe, L.M. 2004. migratory bird by using stable-carbon Reproductive effort, molting latitude and isotopes. Science 282: 1884–1886. feather color in a migratory songbird. Science Mehl, K.R., Alisauskas, R.T., Hobson, K.A. & 306: 2249–2250. Kellett, D.K. 2004. To winter east or west? Norris, K. & Evans, M.R. 2000. Ecological Heterogeneity in winter philopatry in a immunology: life history trade-offs and central-arctic population of King Eiders. immune defense in birds. Behavioral Ecology Condor 106: 241–251. 11: 19–26. Mehl, K.R., Alisauskas, R.T., Hobson, K.A. & Oppel, S., Powell, A.N. & O’Brien, D.M. 2010. Merkel, F.R. 2005. Linking breeding and King eiders use an income strategy for egg wintering areas of king eiders: making use of production: a case study for incorporating polar isotopic gradients. Journal of Wildlife individual dietary variation into nutrient Management 63: 1297–1404. allocation research. Oecologia 164: 1–12. Moore, J.E. & Black, J.M. 2006. Historical Owen, M. 1977. The role of wildlife refuges on changes in black brant Branta bernicla nigricans agricultural land in lessening the conflict use on Humboldt Bay, California. Wildlife between farmers and geese in Britain. Biology 12: 151–162. Biological Conservation 11: 209–222. Naylor, L.W. 2002. Evaluating moist-soil seed Owen, M. & Black, J.M. 1989. Factors affecting production and management in Central the survival of barnacle geese on migration Valley wetlands to determine habitat needs from the breeding grounds. Journal of Animal for waterfowl. M.Sc. thesis, University of Ecology 58: 603–617. California Davis, Davis, California, USA. Pearse, A.T., Krapu, G. L. & Cox, R.R. Jr. 2012. Newton, I. 2006. Can conditions experienced Spring snow goose hunting influences body during migration limit the population levels composition of waterfowl staging in of birds? Journal of Ornithology 147: 146–166. Nebraska. Journal of Wildlife Management 76: Nicolai, C.A. & Sedinger, J.S. 2012. Trade-offs 1393–1400. between offspring fitness and future Péron, G., Nicolai, C.A. & Koons, D.N. 2012. reproduction of adult female black brent. Demographic response to perturbations: the Journal of Animal Ecology 81: 798–805. role of compensatory density dependence Norris, D.R. 2005. Carry-over effects and habitat in a North American duck under variable quality in migratory populations. Oikos 109: harvest regulations and changing habitat. 178–186. Journal of Animal Ecology 81: 960–969. Norris, D.R. & Marra, P.P. 2007. Seasonal Pieron, M.R. & Rohwer, F.C. 2010. Effects of interactions, habitat quality, and population large-scale predator reduction on nest dynamics in migratory birds. Condor 109: success of nesting ducks. Journal of Wildlife 535–547. Management 74: 124–132.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 302 Cross-seasonal effects on waterfowl

Poisbleau, M., Fritz, H., Valeix, M., Perroi, P.-Y., Worlds the Ecology and Evolution of Migration, Dalloyau, S. & Lambrechts, M.M. 2006. pp. 375–389, Johns Hopkins University Social dominance correlates and family status Press, Baltimore, Maryland, USA. in wintering dark-bellied brent geese, Branta Samuel, M.D., Takekawa, J.Y., Samelius, G. & bernicla bernicla. Animal Behaviour 71: 1351– Goldberg, D.R. 1999. Avian cholera mortality 1358. in lesser snow geese nesting on Banks Island, Pospahala, R.S., Anderson, D.R. & Henny, C.J. Northwest Territories. Wildlife Society Bulletin 1974. Population Ecology of the Mallard: Breeding 27: 780–787. Habitat Conditions, Size of the Breeding Populations, Samuel, M.D., Shadduck, D.J. & Goldberg, D.R. and Production Indices. U.S. Department of the 2004. Are wetlands the reservoir for avian Interior, U.S. Fish & Wildlife Service, cholera? Journal of Wildlife Diseases 40: 377–382. Washington DC, USA. Samuel, M.D., Shadduck, D.J., Goldberg, D.R. & Prevett, J.P. & MacInnes, C.D. 1980. Family and Johnson, W.P. 2005. Avian cholera in other social groups in snow geese. Wildlife waterfowl: the role of lesser snow and Ross’s Monographs 71: 1–46. geese as disease carriers in the Playa Lakes Prop, J., Black, J.M. & Shimmings, P. 2003. Travel region. Journal of Wildlife Diseases 41: 48–57. schedules to the high arctic: Barnacle Geese Schindler, M. & Lamprecht, J. 1987. Increase of trade-off the timing of migration with parental effort with brood size in a nidifigous accumulation of fat deposits. Oikos 103: bird. Auk 104: 688–693. 403–414. Schmidt-Hempel, P. 2003. Variation in immune Raveling, D.G. 1970. Dominance relationships defence as a question of evolutionary ecology. and agonistic behavior of Canada geese in Proceedings of the Royal Society B 270: 357–366. winter. Behaviour 37: 291–319. Schmutz, J.A. & Ely, C.R. 1999. Survival of Raveling, D.G. 1979. The annual cycle of body greater white-fronted geese: effects of year, composition of Canada Geese with special season, sex and body condition. Journal of reference to control of reproduction. Auk Wildlife Management 63: 1239–1249. 96: 234–252. Scott, D.K. 1980. Functional aspects of Raveling, D.G. & Heitmeyer, M.E. 1989. prolonged parental care in Bewick’s Swans. Relationships of population size and Animal Behaviour 28: 938–952. recruitment of pintails to habitat conditions Sedinger, J.S. 1984. Protein and amino acid and harvest. Journal of Wildlife Management 53: composition of tundra vegetation in relation 1088–1103. to nutritional requirements of geese. Journal Rexstad, E.A. 1992. Effect of hunting on annual of Wildlife Management 48: 1128–1136. survival of Canada geese in Utah. Journal of Sedinger, J.S. & Bollinger, K.S. 1987. Autumn Wildlife Management 56: 297–305. staging of Cackling Canada Geese on the Rosen, M.N. & Bischoff, A.I. 1949. The 1948–49 Alaska Peninsula. Wildfowl 38: 13–18. outbreak of fowl cholera in birds in the San Sedinger, J.S. & Chelgren, N.D. 2007. Survival Francisco Bay area and surrounding counties. and breeding advantages of larger Black California Fish and Game 35: 185–192. Brant goslings: within and among cohort Runge, M.C. & Marra, P.P. 2005. Modeling variation. Auk 124: 1281–1293. seasonal interactions in the population Sedinger, J.S. & Flint, P.L. 1991. Growth rate is dynamics of migratory birds. In R. negatively correlated with hatch date in Black Greenburg & P.P. Marra (eds.), Birds of Two Brant. Ecology 72: 496–502.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 Cross-seasonal effects on waterfowl 303

Sedinger, J.S. & Herzog, M.P. 2012. Harvest and Sharp, C.M., Abraham, K.F., Hobson, K.A. & dynamics of duck populations. Journal of Burness, G. 2013. Allocation of nutrients to Wildlife Management 76: 1108–1116. reproduction at high latitudes: insights from Sedinger, J.S. & Raveling, D.G. 1990. Parental two species of sympatrically nesting geese. behavior of cackling Canada Geese during Auk 130: 171–179. broodrearing: division of labor within pairs. Shutler, D., Alisauskas, R.T. & McLaughlin, J.D. Condor 92: 174–181. 2012. Associations between body Sedinger, J.S., Flint, P.L. & Lindberg, M.S. 1995. composition and helminths of lesser snow Environmental influence on life-history geese during winter and spring migration. traits: growth, survival and fecundity in Black International Journal for Parasitology 42: Brant (Branta bernicla). Ecology 76: 2404–2414. 755–760. Sedinger, J.S., Lindberg, M.S., Person, B.T., Slattery, S.M. & Alisauskas, R.T. 2002. Use of the Eichholz, M.W., Herzog, M.P. & Flint, P.L. Barker model in an experiment examining 1998. Density dependent effects on growth, covariate effects on first-year survival in body size and clutch size in Black Brant. Auk Ross’s geese (Chen rossii): a case study. Journal 115: 613–620. of Applied Statistics 29: 497–508. Sedinger, J.S., Herzog, M.P., Person, B.T., Kirk, Smith, L.M. & Sheeley, D.G. 1993. Molt patterns M.T., Obritchkewitch, T., Martin, P.P. & of wintering northern pintails in the Stickney, A.A. 2001. Large-scale variation in southern high plains. Journal of Wildlife growth of Black Brant goslings related to Management 57: 229–238. food availability. Auk 118: 1088–1095. Souchay, G., Gauthier, G. & Pradel, R. 2013. Sedinger, J.S., Herzog, M.P. & Ward, D.H. 2004. Temporal variation of juvenile survival in a Early environment and recruitment of Black long-lived species: the role of parasites and Brant into the breeding population. Auk 121: body condition. Oecologia 173: 151–160. 68–73. Sowls, L. K. 1955. Prairie Ducks: a Study of their Sedinger, J.S., Ward, D.H., Schamber, J.L., Butler, Behavior, Ecology, and Management. University W.I., Eldridge, W.D., Conant, B., Voelzer, J.F., of Nebraska Press, Lincoln, Nebraska, USA. Chelgren, N.D. & Herzog, M.P. 2006. Effects Spaans, B., van’t Hoff, C.A., van der Veer, W. & of El Niño on distribution and reproductive Ebbinge, B.S. 2007. The significance of performance of Black Brant. Ecology 87: female body stores for egg laying and 151–159. incubation in Dark-bellied Brent Geese Sedinger, J.S., Chelgren, N.D., Lindberg, M.S. & Branta bernicla bernicla. Ardea 95: 3–15. Ward, D.H. 2008. Fidelity and breeding Stafford, J.D., Kaminski, R.M., Reinecke, K.J. & probability related to population density and Manley, S.W. 2006. Waste rice for waterfowl individual quality in black brent geese (Branta in the Mississippi Alluvial Valley. Journal of bernicla nigricans). Journal of Animal Ecology 77: Wildlife Management 70: 61–69. 702–712. Stafford, J.D., Kaminski, R.M. & Reinecke, Sedinger, J.S., Schamber, J.L., Ward, D.H., K.J. 2010. Avian foods, foraging and habitat Nicolai, C.A. & Conant, B. 2011. Carryover conservation in world rice fields. Waterbirds effects associated with winter location affect 33: 133–150. fitness, social status, and population Stephens, S.E., Rotella, J.J., Lindberg, M.S., Taper, dynamics in a long distance migrant. American M.L. & Ringelman, J.K. 2005. Duck nest Naturalist 178: E110–E123. survival in the Missouri Coteau of North

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 304 Cross-seasonal effects on waterfowl

Dakota: landscape effects at multiple spatial Van Noordwijk, A.J. & De Jong, G. scales. Ecological Applications 15: 2137–2149. 1986. Acquisition and allocation of Tombre, I.M., Erikstad, K.E. & Bunnes, V. 2012. resources: their influence on variation in State-dependent incubation behaviour in the life-history traits. American Naturalist 128: high arctic barnacle geese. Polar Biology 35: 137–142. 985–992. Viallefont, A., Cooke, F. & Lebreton, J.-D. 1995. Trinder, M.N., Hassell, D. & Votier, S. 2009. Age-specific costs of first-time breeding. Reproductive performance in arctic-nesting Auk 112: 67–76. geese is influenced by environmental Webb, E.B., Smith, L.M., Vrtiska, M.P. & conditions during the wintering, breeding and Lagrange, T.G. 2011. Factors influencing migration seasons. Oikos 118: 1093–1101. behavior of wetland birds in the Rainwater van Gils, J.A., Munster, V.J., Radersma, R., Basin during spring migration. Waterbirds 34: Liefhebber, D., Uchier, R.A. & Klaassen, M. 457–467. 2007. Hampered foraging and migratory Wobeser, G., Kerbes, R. & Byersbergen, G.W. performance in swans infected with low 1983. Avian cholera in Ross’ and lesser snow pathogenic avian influenza A virus. PLoS geese in Canada. Journal of Wildlife Diseases ONE 2: e184. 19: 12.

Photograph: Mallard benefitting from wetlands in Maryland, by Ron Nichols/USDA Natural Resources Conservation Service.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 277–304 305

Managing harvest and habitat as integrated components

ERIK E. OSNAS1*, MICHAEL C. RUNGE1, BRADY J. MATTSSON2, JANE AUSTIN3, G. SCOTT BOOMER4, ROBERT G. CLARK5, PATRICK DEVERS4, JOHN M. EADIE6, ERIC V. LONSDORF7 & BRIAN G. TAVERNIA1

1U.S. Geological Survey, Patuxent Wildlife Research Center, Laurel, Maryland, USA. 2University of Natural Resources and Life Sciences, Vienna, Austria. 3U.S. Geological Survey, Northern Prairie Wildlife Research Center, Jamestown, North Dakota, USA. 4U.S. Fish and Wildlife Service, Population and Habitat Assessment Branch, Laurel, Maryland, USA. 5Environment Canada, Prairie and Northern Wildlife Research Center, Saskatoon, Saskatchewan, Canada. 6University of California, Davis, California, USA. 7Franklin and Marshall College, Lancaster, Pennsylvania, USA. *Correspondence author. E–mail: [email protected]

Abstract In 2007, several important initiatives in the North American waterfowl management community called for an integrated approach to habitat and harvest management. The essence of the call for integration is that harvest and habitat management affect the same resources, yet exist as separate endeavours with very different regulatory contexts. A common modelling framework could help these management streams to better understand their mutual effects. Particularly, how does successful habitat management increase harvest potential? Also, how do regional habitat programmes and large-scale harvest strategies affect continental population sizes (a metric used to express habitat goals)? In the ensuing five years, several projects took on different aspects of these challenges. While all of these projects are still on-going, and are not yet sufficiently developed to produce guidance for management decisions, they have been influential in expanding the dialogue and producing some important emerging lessons. The first lesson has been that one of the more difficult aspects of integration is not the integration across decision contexts, but the integration across spatial and temporal scales. Habitat management occurs at local and regional scales. Harvest management decisions are made at a continental scale. How do these actions, taken at different scales, combine to influence waterfowl population dynamics at all scales? The second lesson has been that consideration of the interface of habitat and harvest

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 306 Integrating harvest and habitat management

management can generate important insights into the objectives underlying the decision context. Often the objectives are very complex and trade-off against one another. The third lesson follows from the second – if an understanding of the fundamental objectives is paramount, there is no escaping the need for a better understanding of human dimensions, specifically the desires of hunters and non- hunters and the role they play in conservation. In the end, the compelling question is how to better understand, guide and justify decisions about conservation investments in waterfowl management. Future efforts to integrate harvest and habitat management will include completion of the species-specific case-studies, initiation of policy discussions around how to integrate the decision contexts and governing institutions, and possible consideration of a new level of integration – integration of harvest and habitats management decisions across waterfowl stocks.

Key words: decision analysis, habitat management, harvest management, integration, objectives, population models, yield curve.

Waterfowl management in North America efforts are designed for the same seeks the joint goals of providing hunting populations, they are administered under opportunity and conserving waterfowl different and independent regulatory populations by regulating harvest through contexts with goals that are possibly not the Migratory Bird Treaty Act of 1918 and consistent (Runge et al. 2006). The NAWMP of protecting and improving habitats population goals were developed in through the North American Waterfowl reference to waterfowl populations and Management Plan (NAWMP). Over time, habitat conditions of the 1970s, a decade waterfowl harvest management has evolved of above-average habitat conditions, and into one of the best applications of adaptive without reference to a specific harvest resource management in the world, policy. Harvest policy does not explicitly give where annual hunting regulations are set a population goal, but instead is designed to based on quantitative population models, achieve sustained harvests over a very long data collected annually, and optimisation timeframe, which implicitly strives for an methods designed to make the best possible average population size to support that decision in the face of many sources of harvest. Because expected population size is uncertainty to insure large sustainable linked to harvest rate and habitat conditions, harvests (Nichols et al. 1995, 2007). Habitat interpreting the population goals stated in management is organised around a series of the NAWMP is impossible without public-private partnerships (Joint Ventures) reference to specific habitat conditions and that endeavour to protect and improve harvest policy (Runge et al. 2006). In other waterfowl habitat to meet specific words, population goals specified under population goals for each of several species. harvest and habitat management plans are While both harvest and habitat management not currently coherent, and recognition of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 307 this problem has catalysed several efforts to context for habitat and population goals; reconcile these plans. harvest managers needed to be able to The first effort was the formation of a translate NAWMP accomplishments into Joint Task Group (JTG) sanctioned by harvest opportunity. To accomplish this, a NAWMP and the International Association theoretical assessment framework was of Fish and Wildlife Agencies “to explore formulated around a population model that options and recommend preferred solutions included density-dependent relationships in to reconciling the use of [NAWMP] survival and reproduction, which could then population objectives for harvest and be translated into ecological concepts of habitat management” (Anderson et al. carrying capacity and a sustainable yield 2007). NAWMP partners needed a shared curve (Fig. 1, Appendix 1). The JTG

4

Population goal

3 Improved habitat

2 Current conditions Equilibrium harvest (yield) 1 Worse habitat

0 K

024681012

Equilibrium population size

Figure 1. Using yield curves to understand the relationship between harvest and habitat management. The open circles represent a “right-shoulder strategy” – a harvest policy that maintains a yield less than the maximum sustained yield (the choice of the position on the right shoulder, here 80%, is a policy determination). Improved or worsened habitat is shown by an expanded or contracted yield curve (the curves shown are based on a 25% increase or decrease both in intrinsic growth rate, r, and in carrying capacity, K ). A population goal (dashed line) is said to be “coherent” if it falls at the intersection of the desired harvest strategy and the desired habitat conditions.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 308 Integrating harvest and habitat management explored various interpretations of size for all harvests (Fig. 1). Of course the NAWMP population goals in terms of environment and harvest rates are not sustainable harvest levels in light of the constant and populations do not behave above-average habitat conditions of the deterministically. The environment can 1970s. The main proposal was that fluctuate on short time scales (annual population goals of the NAWMP are best variation in precipitation) or shift to new interpreted as the expected population size long-term average conditions (multi-year found when a harvest strategy that seeks less changes in agriculture policy, climate than maximum sustained yield (a “shoulder change, or permanent habitat loss or strategy”) is overlain with a yield curve that improvements). Short-term fluctuations in reflects the desired long-term average the environment or population size are not habitat conditions. reflected in the yield curve; instead, the yield A sustainable yield curve (Fig. 1, curve can be viewed as an average over these Appendix 1) provides the conceptual fluctuations for a large population. If long- framework needed to integrate harvest and term shifts change demographic rates, habitat management (Runge et al. 2006; perhaps through changes in the strength of Anderson et al. 2007). A sustainable yield density-dependence, then the yield curve curve shows all the equilibrium points of a itself will expand or contract, so that a new deterministic density-dependent population equilibrium population size is realised for model, such as the familiar logistic growth the same harvest rate, or a new harvest rate model (Fig. 1). This model posits that birth is required to realise the same population rate declines and/or death rate increases size (Fig. 1, Appendix 1). This is the crux of with population density, for which there is integrating habitat and harvest management good evidence in waterfowl (Vickery & – we can now ask to what extent Nudds 1984; Cooch et al. 1989; Sedinger et demographic rates need to shift due to al. 1995; Johnson et al. 1997; Sedinger et al. habitat management in order to obtain a 2001; Conroy et al. 2002; Runge & Boomer desired population size and harvest rate. 2005; Viljugrein et al. 2005). If this is the Conversely, we can ask what combination of case and both environmental conditions and harvest rate and population size (a harvest harvest rate remain constant, then the policy) we want to achieve under current population size and structure approaches a habitat conditions. stable equilibrium, where birth and death The realisation among the waterfowl rates are equal. In the absence of harvest, management community that population the population moves toward carrying goals of the NAWMP only make sense in capacity, the maximum equilibrium light of a particular harvest strategy has population size determined by the density- brought the human component of dependent effects. With harvest, the management to the forefront. Hunter population will reach a stable equilibrium satisfaction underlies both NAWMP goals less than carrying capacity, and the yield and harvest policy but little is known curve describes this equilibrium population about their relative importance to hunter

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 309 satisfaction, which harvest managers need habitat management at scales ranging from to know to set a desirable harvest and Joint Ventures (JVs) to the continent, habitat policy. If we are below a population determining the impact of net landscape goal because of a shrinking habitat base, changes on waterfowl demography, and do we reduce harvest to maintain the enhancing the ability to target financial population at goal? Or do we maintain investment in different regions of North harvest (by increasing the per capita harvest America to advance NAWPM objectives. A rate), thus driving the population even comprehensive review of population and lower, in order to satisfy harvest desires habitat objectives was also advocated. of hunters? As long as populations Addressing these recommendations alone remain large, relative to demographic creates significant challenges, both and environmental stochasticity, so that conceptually and operationally. For instance, extinction is not likely or ecological services habitat management typically occurs locally, are not jeopardised, these seem like fair with investments intended to return long- policy alternatives. Yet we know little about term benefits. On the other hand, harvest how hunters and other stakeholders value management decisions are made in short population sizes and harvests. The JTG time steps, usually annually, with broad-scale recognised this and called for an in-depth impacts on birds. Thus, a central problem study of human dimensions underlying was to provide mechanisms allowing the waterfowl management. integration of harvest and habitat objectives While the JTG was demonstrating in ways that could reveal how management conceptual and empirical approaches for decisions could influence demographic rates providing a formal integration of harvest within JVs or ecologically-related regions, and habitat management objectives, a and in turn scale up to affect continental parallel process was underway to evaluate, population dynamics. Importantly, such an for the first time, the NAWMP’s approach would provide new insights into effectiveness in achieving its biological goals optimal allocation of limited conservation (Assessment Steering Committee 2007). resources to where they would have This assessment unearthed the many potential for greatest impacts. strengths and weaknesses inherent in In 2012 the NAWMP community reached planning and delivering large-scale consensus that a third goal, addressing the conservation programmes, and produced benefits of waterfowl populations and wide-ranging recommendations related to habitats to people, should be included NAWMP planning, adaptive processes, explicitly in the NAWMP (NAWMP 2012a), implementation strategies, institutional and agreed that efforts to define objectives issues, integration among bird groups for hunters, conservation supporters and and funding. Most relevant here are the general public should be continued. The recommendations that focused on conceptual and technical advances required developing improved ways of linking to integrate these linked NAWMP goals (i.e. demographic and population responses to for habitat, harvest and people) have been

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 310 Integrating harvest and habitat management explored in case studies focused on Black scepticism by some who asked what the Duck Anas rubripies, Northern Pintail Anas concern was about. The JV system is a highly acuta and scaup (Aythia marila and A. affinis) successful model of collaborative landscape populations. These initiatives were expected conservation, delivering waterfowl habitat to reveal new knowledge about relationships management at a local and regional scale between population processes and (Williams et al. 1999; Assessment Steering management decisions, and also trade-offs Committee 2007). Adaptive Harvest at multiple scales. Furthermore, all three Management (AHM) has provided a stable species are of high conservation concern and transparent way of setting harvest due to their population status, trends in regulations (Nichols et al. 2007). Both habitat numbers, and their importance to hunters in and harvest management appeared to be North American flyways. Each initiative has working well, so why make things more pursued a different analytical approach in complicated? Was integration technically addressing unique problems encountered in feasible? And even if it was, would it matter integrating harvest and habitat management. to the decisions being made? To some Importantly, each initiative has also engaged extent, we cannot yet fully answer these the waterfowl community in consultation questions, as the various efforts to integrate processes, to gain critical insights from habitat and harvest management are still managers and scientists alike on the works-in-progress, but the work of the provision of guidance about biological last five years has changed much of the models that link demographic rates to thinking about waterfowl management in regional habitat planning and harvest, which North America. These advances include were used to inform JV management learning about potential consequences decisions. Future implementation of the of habitat and harvest management NAWMP will be shaped in part by results of decisions given inherent system dynamics these case studies. Indeed, because there (i.e. partial controllability), identifying critical has been no clear route paving the way information gaps (i.e. factors that seem most to successful integration of different important for population dynamics but for components of the plan, there is high which new information is required) and expectation that pilot projects will help to demonstrating that integration is entirely unveil how integration may be achieved feasible and potentially useful. Thus, the (NAWMP 2012b). case studies presented below give compelling reasons for integration, Insights gained to date especially by providing a better In 2006, Runge et al. (2006) raised significant understanding of population dynamics and questions about the disconnected nature of in justifying decisions about investing scarce waterfowl harvest and habitat management, conservation resources in order to achieve suggesting the two programmes needed to the desired outcomes. Several important be more coherent. The “coherence paper” lessons about integration are emerging, as (Runge et al. 2006) was received with described below.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 311

Integrating models links local habitat conditions with the birds’ migratory behaviour, and an estimation of The first lesson concerns the efforts to build the parameters (sometimes difficult to integrated models that can represent the establish) that describe that process. In effects of both harvest and habitat the pintail case study, for example, the management, thus making underlying challenge of parameter estimation includes assumptions of habitat and harvest drawing inference about intermediate-scale management explicit. One of the most processes that have not been observed difficult aspects of this task is not the directly in the field. integration of habitat and harvest in a Building mathematical models is a means theoretical sense (see Appendix 1), but the to make assumptions explicit and to integration across spatial scales and the understand the consequences of actions. estimation of relevant parameters. The key Once hypotheses have been explicitly stated, question is this: how do habitat actions at they should be rigorously tested when the local scale translate into population possible if the claim of science-based responses at the continental scale? Habitat management is to have any merit. There is management occurs at a local scale where almost 20 years of doing just this under land managers work to bring about long- adaptive harvest management for waterfowl term habitat change in a region, such as in a (Cooch et al. 2014) but a similar process JV. Bird populations that are the target of for habitat management is less developed. habitat management are moving among Stating and testing assumptions about habitats within a region and across regions. habitat management is an important These movements act to average or dampen recommendation of the NAWMP the effects of any one habitat effect, be it assessment efforts (Assessment Steering local or regional, and also make it very Committee 2007), and the efforts toward difficult to demonstrate empirically an effect integration have made major advances in of habitat improvements on population building models that represent assumptions demographic rates. Harvest management, about how habitat is linked to demographic on the other hand, occurs at the continental rates. For example, the hypothesis that scale, where season length and bag limit are winter survival is driven by energy limitation set annually on the basis of population goals underlies many decisions about habitat and an underlying model of the extent management, yet demonstrating a causal to which density-dependence influences link between energy limitation and survival population dynamics. Integration therefore in wintering waterfowl has been elusive. requires linking dynamics across the local, Given that some studies show relatively high regional and continental scales. In the survival after the hunting season (e.g. Integrated Waterbird Management & Dugger et al. 1994; Fleskes et al. 2007) Monitoring (IWMM) example given below, and that energy surpluses are evoked to this integration requires an understanding of explain increasing goose populations the spatial structure of the landscape that (Ankney 1996), testing this assumption and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 312 Integrating harvest and habitat management considering alternatives seems reasonable, in the original NAWMP (NAWMP 2012a). including refinements to existing energy At the regional and local scales, issues hypotheses and other mechanisms. Efforts of population distribution and equity of of integration, discussed below, provide a hunter and non-hunter access arise. This framework to do just this. can then lead to conflicts between objectives: should conservation resources Integrating objectives be allocated to a region if it can be shown The second lesson follows from the first. that habitat improvement in the region Habitat management is done at local or has little value for improving continental regional scales to meet local and regional populations? Simply asking this question in objectives for populations, habitat and the context of the decision reveals that harvest, but ultimately habitat management other objectives, perhaps not yet fully is meant to scale up to affect continental articulated, exist. population objectives. Because the entire initiative to integrate habitat management Integrating human desires and with harvest management has been framed governance as a decision, this has led the waterfowl The third lesson follows directly from management community to explore and considering objectives in a decision context. articulate these objectives, to discover new When we realise that decisions are ultimately objectives, and eventually realise that trade- grounded in the objectives, we have to ask offs between objectives will be necessary. where those objectives come from. What are Thus, the work to integrate habitat and our hopes for waterfowl habitat and harvest management has led the waterfowl populations, and why? What do hunters management community to focus on and want, and how does their conservation role articulate objectives more clearly and to use affect the achievement of the other a more structured approach (NAWMP objectives? How do we structure governing 2012a), which in itself is a major institutions to best meet these objectives? accomplishment. This leads to focusing on human More complex objectives and trade-offs dimensions of waterfowl management and emerge as we consider the entire range of on the structure of governing institutions scales from local to continental. At the because the human component of continental scale, the simplest of these waterfowl management is important, trade-offs is that one cannot have higher perhaps most important, for understanding average harvest rate without a subsequent objectives, making optimal decisions and decrease in expected population size if designing institutions that can facilitate harvest mortality is not fully compensated optimal decision making. This is especially by reduced natural mortality (Anderson & the case when trade-offs between objectives Burnham 1976; Cooch et al. 2014; Appendix or scales become necessary. The waterfowl 1). This has led to an examination and re- management community must know how interpretation of the population objectives objectives rank in importance across

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 313 stakeholder groups and this can only come hunters and retention of current hunters? from focused research on these groups to This is currently unknown, but by focusing determine their values. For example, much on the decision context, human dimensions of traditional waterfowl management and have moved to the forefront of the current harvest regulation is based on integration of habitat management and the assumption that hunter satisfaction harvest. These issues are especially clear in increases with larger harvests. This may be the scaup example, given below, where the true, but how is hunter satisfaction related hunter population was included in the scaup to continental population size? If larger population model, just as in models of harvests are valued, how do hunters value classic predator-prey systems. However, all season length and bag limit combinations the integrated models in the world won’t that might lead to the same harvest? These get us anywhere if different agencies, questions can be extended to a broader authorities, administrations or nations are set of stakeholders. For instance, how not willing to integrate and coordinate does non-hunter (or even anti-hunting) policy and programmes that affect satisfaction depend on continental conservation and management at the population size? How does it depend on continental level. local hunting activity? The waterfowl Expanding the range of objectives under management community knows very little consideration raises questions about the about the desires of hunters, and probably governance structures that support even less about the desires of non-hunters. management. Government agencies have a If the general assumption is that hunter public trust responsibility to include input satisfaction increases with harvest and from a broad array of stakeholders, the population size, is it reasonable to assume hunting and non-hunting public alike. But that non-hunting stakeholders’ satisfaction the strongest “stakeholder” input for is neutral to these metrics? Do these government agencies comes from their management choices affect the political statutory mandates. For waterfowl and economic support hunters or non- management, U.S. federal agencies must hunters give to waterfowl and wetlands adhere to the Migratory Bird Treaty Act, conservation? Finally, waterfowl hunters which puts a primary emphasis on migratory have historically had a strong tie to bird populations and their habitats, with conservation of wetlands and waterfowl, only secondary consideration given to but hunter numbers are declining to the consumptive and non-consumptive use. extent that continued participation in Thus, other entities (e.g. state agencies and waterfowl hunting is itself a major concern non-government organisations (NGOs)) (Vrtiska et al. 2013). If decisions about may be better enabled to pursue objectives harvest and habitat management affect like hunter satisfaction, but the current hunter satisfaction can these decisions also governance of habitat and harvest through be used to increase waterfowl hunter the flyway system might not yet provide an numbers through recruitment of new effective structure for inclusion of these

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 314 Integrating harvest and habitat management broader aims. These governance issues, life histories. Few biological models exist to along with the complicated governance address problems at such a large scale. issues associated with habitat protection Furthermore, the challenge is amplified by and management by a diverse array of the fact that biological processes do not landholders, point to the need for align with administrative programme thoughtful consideration of the institutional boundaries and successful management relationships that underlie waterfowl harvest for these species depends on linking and habitat management. management decisions at multiple spatial In short, the emerging efforts to integrate scales. At each scale, habitat management harvest and habitat management, which decisions involve allocating resources originated from a simple suggestion that efficiently, in light of financial and the NAWMP population objectives might personnel limitations and the need for need revision (Runge et al. 2006), have public accountability, to maximise the precipitated a much deeper examination of benefits for waterbird populations. Decision the decision structures used for habitat and analytic techniques hold promise for harvest management and the technical tools addressing these challenges in a structured used to support them. It has been both a and transparent manner (Wilson et al. 2007; significant challenge, with many technical McDonald-Madden et al. 2008; Thogmartin issues yet to address, as well as a unique et al. 2009). opportunity to push the boundaries of our The IWMM seeks to provide decision current approach. Success is not assured, support tools at multiple scales to aid but we will learn much even if our waterbird habitat managers across the aspirations are not fully realised at the Atlantic and Mississippi Flyways. IWMM operational level. The following case studies represents a joint initiative of conservation describe these efforts in more detail. partners, including the USFWS, U.S. Geological Survey, state agencies and Case studies Ducks Unlimited. To date, IWMM has provided standardised waterbird and habitat “Integrated Waterbird Management monitoring protocols and a common and Monitoring Initiative” (IWMM) database with reporting tools to participants The U.S. Fish and Wildlife Service across the two flyways, and coordinated (USFWS), Joint Ventures (JVs), and the pilot data collection has been underway for Flyway Councils work to conserve more than three years. Through the migratory waterbird populations by application of structured decision analytic informing and implementing habitat techniques (Gregory et al. 2012), IWMM and other management actions. The identified pressing waterbird management conservation of migratory waterbird decisions at multiple spatial scales. At a populations and their habitats is an flyway scale, decisions must be made about inherently challenging proposition given the habitat acquisitions and restorations within geographic and temporal scope of species’ the context of budgetary constraints. At a

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 315 local scale, managers annually determine represents an important advance in guiding how to manage habitat within a wetland or land acquisitions by explicitly linking collection of wetlands to maximise long- hypotheses about waterbird biology to term benefits for waterbird populations. alternative acquisition and restoration IWMM recognises that these habitat decisions. The model can also help local management decisions are naturally linked managers understand the importance of across scales. their wetlands for specific waterbird guilds IWMM’s technical team is developing or species in a larger flyway context. For models for decision support at multiple example, the flyway model can be scales. To address habitat acquisition and incorporated into a structured decision restoration decisions at a flyway scale, the making process to evaluate land acquisitions team has developed a continental scale by the National Wildlife Refuge system. simulation model to couple waterbird This framework can also allow for the survival during the migratory period to the inclusion of alternative models in an amount and distribution of energy (in the adaptive management programme to learn form of appropriate habitat) across the about population demographic rates while flyways; thus, incorporating explicit guiding refuge acquisition and management. hypotheses about energy limitation as a Although the IWMM initiative was not determinant of survival during the originally designed to integrate harvest and migratory period. The model uses geospatial habitat, it has provided valuable tools and land cover layers and land cover-specific insights about the process of integration. roosting and caloric values to represent the The development of the IWMM predictive quality of stopover sites. Portfolios of land models has faced a deep challenge: acquisition and restoration decisions are integrating habitat management and evaluated by altering flyway food energy waterbird demography across scales. This content and examining the change in requires identifying the mechanisms that survival that results from these decisions. connect local-scale influences to broad-scale The model identifies areas along the flyway outcomes, in this case, physiological that have a large benefit for the survival of energetics and behavioural adaptations to individuals within a specific guild or species the distribution of food resources (and thus relative to the cost of management or energy) across the landscape, so that acquisition. Insight can be provided to local individual and local habitat mechanisms give managers about the importance of their rise to patterns of migration at the flyway general areas for non-breeding survival scale. Processes that connect the local to the of specific guilds or species, linking continental scales are often not observed management priorities at flyway and local directly, and yet they are the crux of scales. understanding how habitat management For the eastern U.S., the flyway scale translates into demographic change. There model is in the late stages of development are significant challenges in estimating the (Eric Lonsdorf, unpubl. data). The model parameters of these processes, but formal

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 316 Integrating harvest and habitat management methods of expert judgment and modern reflect the annual cycle of breeding, autumn Bayesian hierarchical methods are bearing migration and winter through spring fruit. The important lesson, perhaps, is that migration. Breeding areas are separated into we cannot shy away from understanding the two spatial components to reflect regional complex processes that link dynamics across differences – Alaska and the prairie potholes scales; indeed, that is one of the most and parkland, with a third “breeding” class important aspects of developing integrated used to represent drought years when pintail models. are less likely to attempt breeding and are generally less observable (Runge & Boomer Northern Pintail 2005). Wintering areas are divided into two In the integration efforts focused on pintail, regions – the California Central Valley and the central focus was to build a formal the Gulf Coast, to reflect potential mathematical framework to link habitat and differences in non-breeding season survival harvest management across spatial scales. and density-dependence (Mattsson et al. The integrated pintail model is a spatial 2012). The key aspects to the model are: version of that currently used in pintail 1) the migratory transitions that link harvest management (USFWS 2010; winter grounds to breeding grounds, and Mattsson et al. 2012). This is a spatial matrix- 2) the density-dependent relationships for projection model with an annual time step recruitment and survival in each breeding partitioned into seasonal components to and wintering region, respectively (Fig. 2).

Figure 2. Pintail metapopulation model that allows analysis of the interaction of habitat and harvest management (after Mattsson et al. 2012). Two core breeding areas (Alaska and prairie-parklands, red) and two core wintering areas (primarily the California Central Valley of the Pacific flyway and the gulf coast of the Central and Mississippi flyways, blue) are linked through autumn (red arrows) and spring (blue arrows) migratory transitions. A third breeding season state is used to represent movements of pintail in drought years when breeding effort and observability of the population is low (large light red area). Arrows represent starting and ending locations of migration and not the geographic route of migrating pintail.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 317

Using parameter values from published through the development of mechanistic literature and expert judgment, the model models, which are underway. can then be analysed to investigate the The second major accomplishment has effects of habitat improvements on the yield been to motivate in-depth discussions about curve. Efforts are now underway to use the assumptions and mechanisms of available long-term data sets on harvest, population regulation at the regional scale, band recoveries and breeding population as these are critical for translating local surveys to refine parameters for the model, habitat management into continental including estimating the form and strength demographic impacts. In several workshops, of the regional density-dependent hypothesised mechanisms of density- relationships. Just as with IWMM, there are dependence were elicited from local experts, few direct data to inform intermediate resulting in a series of conceptual models processes. The ongoing approach to that link changes in habitat to regional estimation is to use multiple data sources productivity or survival. In the prairie- linked by the matrix projection model in a parkland breeding region, competition for common hierarchical Bayesian statistical space is thought to be the leading driver of analysis, which can provide information on density-dependence. Annual variation in these hidden processes. precipitation produces variation in pond There have been two major numbers and distribution and this leads to accomplishments from work on the variation in pintail breeding effort and integration of pintail harvest and habitat distribution. However, the key element here data to date. The first is the demonstration is that, after controlling for precipitation- that it is theoretically possible to link habitat induced variation, pintail density has an and harvest across scales (Mattsson et al. additional effect. Thus in years of higher 2012). The integrated pintail model shows than average pintail numbers, more pintail how habitat improvements at the regional nest in habitats of the prairie-parkland level might increase pintail demographic region where reproductive success is lower rates in that region, and hence can increase (J.H. Devries, pers. comm.). In the wintering the continental yield curve. In the future, regions, the focus is on the relationship this model can be used to inform allocation between density and post-hunting season of conservation resources. For example, survival, through the effects of a limited preliminary analyses of the initial model food supply (and thus energy intake) on the suggests that proportional habitat-related birds’ body mass in winter and spring. The improvements on prairie-parkland breeding assumption here is that habitat managers areas could be more effective at increasing can increase pintail survival by providing the yield curve than equal proportional more nutritious food resources; in years improvements to wintering areas (Mattsson with higher post-hunting season population et al. 2012). Confirmation of this conclusion size and greater food depletion, survival is awaits formal parameter estimation and a reduced compared to years with lower better understanding of local processes populations and identical habitat. These

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 318 Integrating harvest and habitat management regional sub-models are intended to provide from the bottom up, examining long-term predictive tools that, when coupled with population and harvest data to develop the continental demographic model, will hypotheses about factors that potentially provide an analytical framework for were contributing to population decline assessing the effects of local habitat changes (Austin et al. 2000; Afton & Anderson 2001; on the continental yield curve. Austin et al. 2014). However, a model-based Two insights emerge from the pintail approach such as that used for the Northern work. First, one of the most challenging Pintail was precluded because of broad aspects of integrating habitat and harvest uncertainties, particularly the absence of management is developing demographic contemporary annual survival rates, sparse models that integrate dynamics across data on vital rates from breeding grounds spatial scales. Indeed, as with the IWMM in the boreal forest and taiga, and project, the pintail project recognised that uncertainties about the population trends the crux of the modelling effort was for each scaup species separately. It was the identification of intermediate-scale clear that the research and monitoring mechanisms of population regulation, necessary to fill these knowledge gaps, and such as the regional density-dependence to clarify the key factors affecting scaup, relationships. The second insight arises from would require substantial resources, time an ongoing challenge: these intermediate- and collaboration. At the same time, debate scale mechanisms are difficult to observe was growing about how adaptive harvest directly, and so estimating the functions and management affected the harvest of species parameters is likewise difficult. Fortunately, other than Mallard Anas platyrhynchus, and modern hierarchical statistical methods can was a particular concern for the scaup potentially be used to make inference about harvest which, like for other long-lived and such hidden processes, in this case by slower-producing ducks, is more sensitive to using observations at both the local and the duration of the hunting season (Allen et continental scales to gather insights about al. 1999). the mechanisms that link the dynamics The waterfowl management community across scales. Where possible, of course, also was coming to recognise the efforts to measure demographic rates importance of understanding the human directly and to test hypotheses underlying components of waterfowl management, habitat and harvest management should be specifically, the hunter desires and factors encouraged, but modern inferential that affect hunter participation (Case & methods provide a promising alternative. Sanders 2008). Waterfowl hunters were expressing strong concerns about fewer Scaup scaup and the loss of hunting opportunities Substantial declines in the continental scaup associated with restrictive regulations, population in the 1980s and 1990s attracted and waterfowl managers were concerned concern from biologists and hunters alike. about declining hunter numbers because of Biologists first approached the problem their important influence in conservation.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 319

Moreover, waterfowl hunters and managers hunting tradition. This last objective voiced concern about the decline of the explicitly brings people into the objectives diving duck hunting tradition – notably the and recognises the important contribution use of large decoy sets in large open water of hunters to waterfowl conservation, bodies, often using low-profile layout boats through direct financial contributions, – a practice that has changed little over the advocacy and economic activity. Thus, last century and are considered by many to sustaining the diving duck hunting tradition be among the purest of all waterfowl through maintaining the number, hunting traditions. While this concern does participation and identity of diving duck not directly relate to conservation per se, it hunters becomes an explicit conservation highlights the importance of human values objective. Although participants thought in waterfowl management that is reflected in this was fundamental, it is not to say that the NAWMP revision (NAWMP 2012a). maintaining hunter tradition and numbers is Hence, waterfowl biologists and managers equally important compared to the other were challenged to develop approaches to objectives. Determining the relative decision-making in the face of deep sources importance of each objective remains to be of uncertainty in scaup biology as well as in determined and will require input from all the objectives of the scaup conservation societal stakeholders, hunters and non- and management community. As a result, hunters included. efforts were initiated to address scaup To predict the consequences of conservation planning through the conservation actions on each objective, a principles of structured decision making coupled scaup-hunter model (i.e. predator- (Gregory et al. 2012), with a focus on how prey or consumer-resource type model) best to allocate scarce conservation was developed with explicit relationships resources among management actions on an between potential management actions and annual basis. key scaup demographic processes as well The structured decision-making approach as diving duck hunter dynamics. Hunter first required a clear, explicit statement of recruitment and retention were modelled as scaup conservation and management goals a function of scaup population levels, and and objectives. Participants in the process scaup harvest rates were driven by the quickly realised that there was one over- number of diving duck hunters. These arching goal: to conserve scaup populations linked models project scaup and hunter at levels that satisfy societal values. Under numbers forward through time, predicting this goal, a resulting objectives hierarchy the numbers of breeding scaup, numbers of established linkages among three diving duck hunters and scaup harvest. fundamental objectives (Fig. 3): 1) achieve Initial functional relationships relating continental habitat conditions capable of retention and recruitment to management supporting a target scaup population; actions were developed based largely on 2) maintain or increase the sustainable scaup expert judgement and limited knowledge harvest; and 3) sustain the diving duck from other species. Using the model,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 320 Integrating harvest and habitat management



! &
!%
 %#!
"
&!
""
!#!(
!"
&%!

&
! 
#!
-#"
&
! +

%!"
&
%
 (
 "(.
!%!"
%#
" #
 

!%  #
 &!"
" "
 %#!




 
 
  
 
  
 

&
/
&
/
Ɛ
 !
/
 !

 !
"
"
,%#
!%!"
!%!"
+
! 
! 
 ""(

 # #

2%" !
#+
/
 "(

 "(

"
&
%
""
#(
%"% 
 
 !
'" 

 "!
%#
!
&
 "
"
 #
%
%" !
!%  "
 !
&
%
%"
%"(

Figure 3. Objectives hierarchy for scaup conservation. After extensive discussions and revisions over three workshops, participants arrived at a set of three fundamental objectives under one overarching goal to conserve scaup populations at levels that satisfy societal values (black box). The fundamental objectives (medium grey boxes) identify issues of most concern. Each fundamental objective is linked to means objectives (white boxes) through simple functions that are hypothesised to affect survival and recruitment of scaup and hunters (light grey ovals). waterfowl managers can ask: what affects harvest regulations (e.g. bag limit, season each vital rate for both scaup and hunters? length), hunter access (e.g. alter amount Also, what actions could be taken to alter or of hunting habitat available through improve each vital rate? For each possible conservation or access programmes), or action, simple functional relationships were social networking (e.g. mentor programmes, developed that linked how a management web forums, community events). This action affected a landscape variable (e.g. modelling framework helps to identify amount of nesting cover on the landscape), portfolios of management actions (e.g. and how that in turn was related to a vital breeding or wintering habitat management) rate (e.g. probability of breeding success). that have the greatest impact on scaup and Possible actions affecting hunters included hunter population change. Work is ongoing

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 321 to finalise the modelling framework, Eastern Waterfowl Breeding Survey and the establish a baseline parameterisation and completion of the technical framework conduct a sensitivity analysis (Austin et al. of an international, adaptive harvest 2010). This will result in a robust platform management strategy for Black Ducks to test assumptions about scaup and hunter (BDAHM). This confluence of events population dynamics and to support allowed the BDJV to re-evaluate priority decisions about the allocation of scarce information needs for the conservation of conservation resources. the Black Duck. The BDJV decided to focus The key lessons from the work on scaup greater effort on understanding Black Duck integration have been first, that there is a habitat ecology. The BDJV also agreed to complex set of objectives involved when focus on information needs required to harvest and habitat management are determine where in the annual life-cycle considered together (see the “Integrating limited financial resources should be objectives” section above). Here participants allocated for habitat protection, restoration thought that because encouraging and and enhancement to meet four fundamental preserving hunter participation was objectives: 1) maintain Black Duck fundamental, adding hunters explicitly into abundance at levels that meet legal and the objective was not only desirable but policy mandates; 2) maintain the relative necessary. This ultimately took the form of a distribution of breeding and non-breeding mathematical model to predict the Black Ducks corresponding to the 1990– consequence of management actions on 2012 period; 3) maintain carrying capacity hunter participation, a novel innovation to support the desired population and in waterfowl management. Second, by distribution; and 4) maintain consumptive admitting hunter objectives the question and non-consumptive recreational arises as to who is the decision maker opportunities commensurate with and how can institutions be structured to population sustainability and carrying best meet these hunter objectives (see capacity. Framing information needs in this “Integrating human desires and governance” context forced the community to address above). These questions are largely the issue of integrating habitat and harvest unresolved for scaup, or for waterfowl objectives (in “Integrating objectives” management as a whole. above). To address the trade-offs between objectives 1 and 4, above (“coherence”, American Black Duck Runge et al. 2006), the desired NAWMP At the time of the first NAWMP population goal was interpreted in reference Continental Assessment (Assessment to the BDAHM strategy, to harvest the Steering Committee 2007) and release of population at 98% of maximum sustained the Joint Task Group report (Anderson et al. yield (i.e. the 98% right-shoulder strategy). 2007), the Black Duck Joint Venture (BDJV) Therefore, objective (1) can be interpreted was finishing work on two priority issues: as achieving the NAWMP population goal the development and implementation of the for the Black Duck given a 98% right-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 322 Integrating harvest and habitat management shoulder harvest strategy. A future step is to mechanism that is designed to fund projects re-evaluate, and potentially revise, the based solely on waterfowl objectives. The NAWMP population goal for Black Duck vast majority of, if not all, habitat conditioned on the BDAHM strategy and programmes are designed to achieve capacity of the habitat JVs to increase multiple objectives by providing habitat continental carrying capacity. for threatened and endangered species, The BDJV developed a conceptual waterfowl and other species. The annual life-cycle model relating Black Duck decentralised nature and multiple objectives population dynamics to habitat limiting of habitat conservation programmes create factors at the regional scale while accounting challenges regarding governance, not only for annual harvest, much like that for across the waterfowl enterprise but more Northern Pintail (Devers & Collins 2011; broadly within the wildlife conservation Mattsson et al. 2012). While progress on community. These insights about parameterising the life-cycle model and governance and sources of uncertainty have associated decision framework continues forced the BDJV community to think more (see “Integrating models” above), several broadly about the decision process, insights have emerged. The first insight including the multi-objective nature of concerns the decision context – “where” habitat programmes, and to consider a wide and “how much” habitat is needed; it array of decision tools such as dynamic became clear through this effort that there optimisation and robust decision-making are multiple habitat decisions to be made at (Lempert & Collins 2007). multiple scales. Habitat delivery on the The second insight is that, despite breeding grounds of eastern Canada is challenges related to integrated governance, independent of habitat delivery during the a decision analytic approach based on non-breeding season (e.g. the North an integrated population-habitat model American Wetland Conservation Canada allows the BDJV to make better decisions programme versus the North American regarding the allocation of limited Wetland Conservation programme in the monitoring and research funds to address U.S.). In the case of each programme, no key uncertainties and assumptions trade-off exists in terms of funds or (“Integrating models” above). For example, resources between the breeding and non- the Black Duck conceptual model assumes breeding period; this is an important habitat restoration results in increased food consideration as it establishes two separate availability (i.e. energetic carrying capacity) decision processes. This is probably true for and post-hunting season survival. However, most waterfowl species. The decision the BDJV lacks empirical data to context is also complex within countries parameterise this hypothetical relationship. because habitat decisions are made at the To address this assumption, the BDJV has national, regional and local scales using a invested resources into a two-season variety of funding mechanisms. Moreover, banding programme to estimate post- we cannot identify a single funding hunting season survival and research to

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 323 quantify the effect of restoration efforts on management. Future work in this area food availability. The results of these should use the species-specific examples to projects will be used to parameterise the life- build models that are capable of predicting cycle model and provide insight into the the consequences of large-scale landscape relationship between local scale habitat change, such as those resulting from management and population response at land use and climate changes. In addition, the continental scale. In the long-term, habitat and harvest management are not the BDJV anticipates using the life-cycle single-species endeavours. For harvest model and analytical tools to guide the management, a common framework of development of future research and hunting regulations affects many species monitoring projects. at once, including species of significant The main lessons from the work on Black conservation concern. In habitat Duck have been, much like the scaup management, decisions are rarely made in example, that the objectives become very reference to a single species, or even just complex and that governing institutions are waterfowl. Thus, future efforts to integrate not well-structured with respect to these these management decisions must embrace complex objectives. The multi-species nature a wider set of objectives, which undoubtedly of habitat investment decisions allow for will lead to more complex and difficult very little direct control to influence Black trade-offs. Duck conservation in particular. In these Integration initiatives to date have shown complex cases, the use of models becomes that the management of harvest and habitat especially important for understanding the should not continue to be viewed as consequences of decisions. separate endeavours if the waterfowl management community desires to make Conclusions optimal decisions with scarce resources. Work on the integration of harvest and These two management regimes affect the habitat management is ongoing. The same social-ecological system; thus, the species-specific initiatives must be question naturally arises as to whether the completed, followed by a dedicated effort to current governance system for waterfowl is implement these frameworks and use their in some sense sub-optimal (cf. Ostrom et al. guidance to inform management decisions. 1999; Dietz et al. 2003; Ostrom 2009) and, if This will require commitment and buy-in by so, what parts need to change. The efforts to decision makers and local managers, and integrate harvest and habitat management this can only come through continued have, if nothing else, raised this question engagement between the research and and led to an examination of waterfowl management communities, enhanced management in a broader context that understanding of the underlying objectives includes the waterfowl resource, habitat of all stakeholders, and continued critical re- ecosystems and the objectives and desires of examination of the governance structures people interacting with those systems surrounding habitat and harvest (NAWMP 2012a). Conservation decision-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 324 Integrating harvest and habitat management makers are challenged by increasingly and U.S. Geological Survey, U.S. Department complex decisions, fewer conservation and of the Interior, Washington D.C., USA. administrative resources, and changing Ankney, C.D. 1996. An embarrassment of riches: social values. In addition, conservation too many geese. Journal of Wildlife Management planners are confronted with system change 60: 217–223. Assessment Steering Committee. 2007. North brought about by agricultural commodity American Waterfowl Management Plan: markets in the short-term and climate in the continental progress assessment final report. long-term. Thus, waterfowl are only one Plan Committee, North American Waterfowl component of a complex system, and the Management Plan, Washington D.C., USA. larger hope is that what started as a simple Austin, J.E., Afton, A.D., Anderson, M.G., Clark, proposal to manage coherently hunter R.G., Custer, C.M., Lawrence, J.S., Pollard, harvest and waterfowl habitat will lead to J.B. & Ringelman, J.K. 2000. Declining stronger and more adaptable technical, scaup populations: issues, hypotheses, and conceptual and institutional structures to research needs. Wildlife Society Bulletin 28: address these larger challenges. 254–263. Austin, J.E., Boomer, G.S., Brasher, M., Clark, Acknowledgements R.G., Cordts, S.D., Eggeman, D., Eichholz, M.W., Hindman, L., Humburg, D.D., Kelley, We thank the ECNAW scientific committee J.R., Jr., Kraege, D., Lyons, J.E., Raedeke, for inviting us to present this plenary and A.H., Soulliere, G.J. & Slattery, S.M. 2010. Dave Koons and Jim Nichols for Development of scaup conservation plan. constructive comments on the manuscript. Unpublished report from the Third Scaup Workshop – Migration, Winter, and Human References Dimensions, 21–25 September 2009, Afton, A.D. & Anderson, M.G. 2001. Declining Memphis Tennessee, USA. U.S. Geological scaup populations: a retrospective analysis of Survey, Jamestown, North Dakota, USA. long-term population and harvest survey Austin, J.E., Slattery, S. & Clark, R.G. 2014. data. Journal of Wildlife Management 65: 781– Waterfowl populations of conservation 796. concern: learning from diverse challenges, Allen, G.T., Caithamer, D.F. & Otto, M. 1999. A models and conservation strategies. Wildfowl review of the status of greater and lesser (Special Issue No. 4): 470–497. scaup in North America. U.S. Fish and Beverton, R.J.H. & Holt, S.J. 1957. On the Wildlife Service, Office of Migratory Bird dynamics of exploited fish populations. Management, Washington D.C., USA. Fisheries Investigation Series 2, Vol. 19,. Anderson, M.G., Caswell, D., Eadie, J.M., Ministry of Agriculture Fisheries and Food Herbert, J.T., Huang, M., Humburg, D.D., (MAFF), London, UK. Johnson, F.A., Koneff, M.D., Mott, S.E., Case, D.J. & Sanders, S. 2008. The future of Nudds, T.D. & others. 2007. Report from the waterfowl management: Framing future Joint Task Group for clarifying North decisions for linking harvest, habitat, and American Waterfowl Management Plan human dimensions. Summary Report 10-9- population objectives and their use in harvest 08. D.J. Case and Associates, Mishawaka, management. U.S. Fish and Wildlife Service Indiana, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 325

Conroy, M.J., Miller, M.W. & J.E. Hines. 2002. Structured Decision Making: a Practical Guide Identification and synthetic modeling of to Environmental Management Choices. Wiley- factors affecting American Black Duck Blackwell, West Sussex, UK. populations. Wildlife Monographs 150: 1–64. Hilborn, R., Walters, C.J. & Ludwig, D. Cooch, E.G., Lank, D.B., Rockwell, R.F. & 1995. Sustainable exploitation of renewable Cooke, F. 1989. Long-term decline in resources. Annual Review of Ecology and fecundity in a Snow Goose population: Systematics 26: 45–67. evidence for density dependence? Journal of Johnson, F.A., Moore, C.T., Kendall, W.L., Animal Ecology 58: 711–726. Dubovsky, J.A., Caithamer, D.F., Kelley Jr., Cooch, E.G., Guillemain, M., Boomer, G.S., J.R. & Williams, B.K. 1997. Uncertainty and Lebreton, J.-D. & Nichols, J.D. 2014. The the management of mallard harvests. Journal effects of harvest on waterfowl populations. of Wildlife Management 61: 202–216. Wildfowl (Special Issue No. 4): 220–276. Lempert, R.J. & Collins, M.T. 2007. Managing Devers, P.K. & Collins, B. 2011. Conservation the risk of uncertain threshold responses: action plan for the American black duck, comparison of robust, optimum and First Edition. U.S. Fish and Wildlife Service, precautionary approaches. Risk Analysis 27: Division of Migratory Bird Management, 1009–1026. Washington D.C., USA. Mattsson, B.J., Runge, M.C., Devries, J.H., Dietz, T., Ostrom, E. & Stern, P. C. 2003. The Boomer, G.S., Eadie, J.M., Haukos, D.A., struggle to govern the commons. Science Fleskes, J.P., Koons, D.N., Thogmartin, 302(5652): 1907–1912. W.E. & Clark, R.G. 2012. A modeling Dugger, B.D., Reinecke, K.J. & Fredrickson, L.H. framework for integrated harvest and 1994. Late winter survival of female mallards habitat management of North American in Arkansas. Journal of Wildlife Management 58: waterfowl: Case-study of Northern Pintail 94–99. metapopulation dynamics. Ecological Modelling Fleskes, J.P., Yee, J.L., Yarris, G.S., Miller, M.R. & 225: 146–158. Casazza, M.L. 2007. Pintail and mallard McDonald-Madden, E., Baxter, E.W.J. & survival in California relative to habitat, Possingham, H.P. 2008. Subpopulation triage: abundance, and hunting. Journal of Wildlife how to allocate conservation effort among Management 71: 2238–2248. populations. Conservation Biology 22: 656–665. Fretwell, S.D. 1972. Populations in a Seasonal North American Waterfowl Management Plan Environment (No. 5). Princeton University (NAWMP) Committee. 2012a. North American Press, Princeton, New Jersey, USA. Waterfowl Management Plan 2012: People Fretwell, S.D. & Lucas, H.L. 1969. On territorial Conserving Waterfowl and Wetlands. Canadian behavior and other factors influencing Wildlife Service & U.S. Fish and Wildlife habitat distribution in birds. Acta Biotheoretica Service, Washington D.C., USA. 19: 16–36. North American Waterfowl Management Plan Getz, W.M. & Haight, R.G.. 1989. Population (NAWMP) Committee. 2012b. North American Harvesting: Demographic Models of Fish, Forest, Waterfowl Management Plan: a Companion and Animal Resources. Vol. 27. Princeton Document to the 2012 North American Waterfowl University Press, Princeton, New Jersey, USA. Management Plan. Canadian Wildlife Service & Gregory, R., Failing, L., Harstone, M., Long, U.S. Fish and Wildlife Service, Washington G., McDaniels, T. & Ohlson, D. 2012. D.C., USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 326 Integrating harvest and habitat management

Nichols, J.D., Johnson, F.A. & Williams, B.K. Thogmartin, W.E., Fitzgerald, J.A. & Jones, M.T. 1995. Managing North American waterfowl 2009. Conservation design: where do we go in the face of uncertainty. Annual Review of from here? Proceedings of the International Ecology and Systematics 26: 177–199. Partners in Flight Conference 4: 426–436. Nichols, J.D., Runge, M.C., Johnson, F.A. & U.S. Fish and Wildlife Service. 2010. Northern Williams, B.K. 2007. Adaptive harvest Pintail harvest strategy 2010. U.S. management of North American waterfowl Department of the Interior, Fish and populations: a brief history and future Wildlife Service, Division of Migratory Bird prospects. Journal of Ornithology 148(2): 343– Management, Washington D.C., USA. 349. Vickery, W.L. & Nudds, T.D. 1984. Detection of Ostrom, E. 2009. A general framework for density-dependent effects in annual duck analyzing sustainability of social-ecological censuses. Ecology 65: 96–104. systems. Science 325(5939): 419–422. Viljugrein, H., Stenseth, N.C., Smith, G.W. & Ostrom, E., Burger, J., Field, C.B., Norgaard, Steinbakk, G.H. 2005. Density dependence R.B. & Policansky, D. 1999. Revisiting the in North American ducks. Ecology 86: 245– commons: local lessons, global challenges. 254. Science 284(5412): 278–282. Vrtiska, M.P., Gammonley, J.H., Naylor, L.W. Runge, M.C. & Boomer, G.S. 2005. Population & Raedeke, A.H. 2013. Economic and dynamics and harvest management of the conservation ramifications from the decline continental Northern Pintail population. of waterfowl hunters. Wildlife Society Bulletin Final Report, 6 June 2005. U.S. Geological 37: 380–388. Survey, Patuxent Wildlife Research Center, Williams, B.K., Koneff, M.D. & Smith, D.A. 1999. Laurel, Maryland, USA. Evaluation of waterfowl conservation under Runge M.C., Johnson, F.A., Anderson, M.G., the North American Waterfowl Management Koneff, M.D., Reed, E.T. & Mott, S.E. 2006. Plan. Journal of Wildlife Management 63: The need for coherence between waterfowl 417–440. harvest and habitat management. Wildlife Wilson, K.A., Underwood, E.C., Morrison, E.C., Society Bulletin 34: 1231–1237. Klausmeyer, K.R., Murdoch, W.W., Reyers, Sedinger, J.S., Flint, P.L. & Lindberg, M.S. 1995. B., Wardell-Johnson, G., Marquet, P.A., Environmental influence on life-history traits: Rundel, P.W., McBride, M.F., Pressey, R.L., growth, survival, and fecundity in Black Brant Bode, M., Hoekstra, J.M. Andelman, S.J., (Branta bernicla). Ecology 76: 2404–2414. Looker, M., Rondinini, C., Kareiva, P., Shaw, Sedinger, J.S., Lindberg, M.S. & Chelgren, N.D. M.R. & Possingham, H.P. 2007. Maximizing 2001. Age-specific breeding probability in the conservation of the world’s biodiversity: Black Brant: effects of population density. What to do, where and when. PLoS Biology 5: Journal of Animal Ecology 70: 798–807. e223.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 Integrating harvest and habitat management 327

Appendix 1. A technical primer on integrating habitat and harvest: derivation of a sustainable yield curve based on density- and habitat-dependent demographic rates.

Harvest management has a long tradition of explicit quantitative demographic modelling as the basis for decisions (Beverton & Holt 1957; Getz & Haight 1989; Hilborn et al. 1995). Habitat management is implicitly based on an underlying population model (e.g. Fretwell & Lucas 1969; Fretwell 1972), but it is less common for the model to be made explicit in the context of habitat management decisions. For habitat and harvest management to be integrated in a meaningful and useful way, parameters of a population model (i.e. the demographic rates) must ultimately be functions of habitat characteristics and harvest rates. The simplest representation of a population model where density (N, numbers per unit area of space) varies over time (t), and where per capita birth (b) and death (d) rates are functions of N and habitat (H ), is: dN = b(N ,H )–d(N ,H ) Equation (1) dt The birth and death functions, b(N,H ) and d(N,H ) respectively, are also functions of time, reflecting the seasonal nature of waterfowl reproduction and mortality, but this notation has been dropped for clarity. These functions might be very complex, for example N might be a vector that refers to densities at various locations (spatial complexity), times (delay effects), or to different species (interspecific competition), and H might be a vector that refers to the area of different habitat types or the state of resources within each habitat type. In addition, b( ) or d( ) might be non-linear, such that the effect of a density or habitat manipulation is not constant across all N or H. This might be the case for d( ) with respect to N if hunting mortality is compensatory to other mortality sources (see Cooch et al. 2014, this volume, for a discussion of density-dependence and other mechanisms of compensation). While many biologists might envision very complex hypotheses about the form of these functions, practical limitations and parsimony will limit the form to fairly simple representations. The simplest form is a linear relationship in birth and death rates

b(N ,H ) = b0 – b1N + b2H + b3NH Equation (2a)

b(N ,H ) = b0 – b1N + b2H + b3NH Equation (2a) where h is the harvest rate and bi and di are parameters relating demographic rates to density and habitat. In addition, there is the constraint that b(N,H ) ≥ 0 and d(N,H ) ≥ 0. Even if reality is much more complex, this linear model can be thought of as a first approximation to a more complex model, especially if one is interested in small perturbations from a particular point of interest (i.e. current conditions). With these relationships, equation (1) can be rewritten into the familiar logistic growth form with harvest, dN N = rN 1–  – hN Equation (3) dt  K  with r = b0 – d0 + (b2 + d2)H and K = r/(b1 + d1 –(b3 + d3)H ). Equation (3) has several interesting properties from the perspective of integrating harvest and habitat management. First, the “intrinsic growth rate”, r – the growth rate as density approaches zero – is a function of the birth and death rate intercepts and the habitat coefficients. Thus, the “intrinsic growth rate” can vary with changes in habitat

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 328 Integrating harvest and habitat management

(see figures in Anderson et al. 2007). When r increases with H, this effect has been called “habitat quality” (Anderson et al. 2007; Mattsson et al. 2012, and see Fretwell & Lucas 1969 or Fretwell 1972 for an original formulation of essentially the same ideas) because this effect is independent of population density. Second, “carrying capacity” (K ) is a function of the “intrinsic growth rate”, the birth and death rate density coefficients (b1 and d1), and the habitat-by-density interaction coefficients (b3 and d3). Therefore, habitat-related changes in “carrying capacity” may come from “quality” effects or through the habitat- density interaction coefficients, which have been called “habitat quantity effects” (Anderson et al. 2007; Mattsson et al. 2012) because the effect of habitat depends on the population size by changing the quantity of resources available per individual. Regardless of how the coefficients are named, it is important to realise that changes in habitat can have both “intercept” and “slope” effects on demographic rates. A major empirical challenge for integrating habitat and harvest is estimating these effects and determining the habitat dimension(s) along which they occur. This second point is also important because the term “carrying capacity” is often used imprecisely in reference to habitat management, rather than as a specific rescaling of density- and habitat-specific demographic parameters. Without explicitly stating the functional form of birth and death rates, “carrying capacity” has little meaning other than as a dynamic equilibrium in the absence of harvest, which is never actually observed in exploited systems. Only with an explicit model (functional form) for demographic rates does “carrying capacity” have any practical utility. Third, besides the undesirable state N = 0 when h ≥ r, and the dynamic equilibrium N = K when h = 0, there is a wide range of equilibrial N for 0 < h < r that satisfy: N rN 1–  – hN = 0 Equation (4)  K  If we let Y = hN be the total sustainable harvest yield, a plot of Y versus N gives a “yield curve” (Fig. 1). Because we have explicitly written r and K as functions of habitat, we can show how the yield curve changes with habitat and how this depends on specific values of the coefficients (Fig. 1). Some points about the yield curve deserve emphasis. First, the yield curve is not directly observable. Instead it is a representation of the dynamic equilibria of a population model only reached with constant parameter values and at infinite time. It can be thought of as a long-term attractor of population size if harvest and habitat remained constant. Yearly observations of population fluctuations or demographic rates are not changes in the yield curve but are instead, stochastic realisations around an average described by the yield curve. The yield curve then serves as a summary of the consequences of equations (2a,b); and the challenge for scientists and conservation planners is to propose a functional form of equations (2a,b), estimate the relevant parameters, and then make decisions based on the consequences as summarised by the yield curve. This is not a trivial task. Second, by “habitat change” we mean long term changes such as climate, agricultural policy, or actions of conservation planners that work to shift the equilibria of the model (i.e. the yield curve). Most conservation planners seek to affect long-term shifts in habitat that change population equilibria, not short-term fluctuations around existing conditions. Third, under AHM the USFWS derives harvest regulations through optimization of a stochastic version of a population model related to equation (1). The yield curve developed from the deterministic version of that model is extraordinarily helpful in understanding the results of the stochastic dynamic optimisation. Thus, the yield curve serves as a tool to communicate the relationship between expected harvest and expected population size.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 305–328 329

Implementing the 2012 North American Waterfowl Management Plan: people conserving waterfowl and wetlands

DALE D. HUMBURG1 & MICHAEL G. ANDERSON2

1Ducks Unlimited, Inc., One Waterfowl Way, Memphis, Tennessee 38120, USA. 2Ducks Unlimited Canada, Institute for Wetland and Waterfowl Research, P.O. Box 1160, Stonewall, Manitoba R0C 2Z0, Canada. *Correspondence author. E-mail: [email protected]

Abstract The North American Waterfowl Management Plan (NAWMP) is a continental ecosystems model for wildlife conservation planning with worldwide implications. Since established in 1986, NAWMP has undergone continual evolution as challenges to waterfowl conservation have emerged and information available to support conservation decisions has become available. In the 2012 revision, the waterfowl management community revisited the fundamental basis for the Plan and placed greater emphasis on sustaining the Plan’s conservation work and on integration across disciplines of harvest and habitat management. Most notably, traditional and non- traditional users (i.e. hunters and wildlife viewers) of the resource and other conservation supporters are integrated into waterfowl conservation planning. Challenges ahead for the waterfowl management enterprise include addressing tradeoffs that emerge when habitat for waterfowl populations versus habitat for humans are explicitly considered, how these objectives and decision problems can be linked at various spatial and temporal scales, and most fundamentally how to sustain NAWMP conservation work in the face of multi-faceted ecological and social change.

Key words: conservation planning, habitat, harvest, human dimensions, hunters.

Conservation planning for waterfowl in (NAWMP). These linked features are not North America has, for nearly 30 years, new to wildlife conservation and certainly emphasised continent-scale population not to waterfowl management. As modern objectives and associated goals for waterfowl conservation was in its formative populations, habitat, and users at various stages, Fredrick Lincoln, originator of geographical scales, such as administrative the Flyways concept testified before the Flyways and Joint Ventures of the North 75th Congress relating the key elements American Waterfowl Management Plan of populations, habitat and waterfowl

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 330 Implementing the 2012 North American Waterfowl Management Plan hunting (U.S. Government Printing Office NAWMP established explicit, continental 1937): scale, numeric objectives for waterfowl populations. In a summary statement, Populations: “It is my opinion at the present NAWMP’s authors proposed: time that we have about a third of the number of ducks and geese that we had 10 or 15 years ago.” “Meeting these goals would provide opportunity for 2.2 million hunters in Canada and the Habitat: “Furthermore, I am not satisfied that United States to harvest 20 million ducks we can have the population we had 10 or 15 annually. The harvest would include 6.9 million years ago, as I am not sure we could accommodate mallards, 1.5 million pintails and 675,000 them all.” black ducks. It would also provide benefits to Hunters: “Nevertheless, I am satisfied that we millions of people interested in waterfowl for are steadily progressing toward the time when we purposes other than hunting” (U.S. can enjoy reasonable sport.” Department of the Interior and Environment Canada 1986, page 6). Efforts to develop a U.S. national waterfowl management plan during the late Concerning specific habitat goals, the 1970s and early 1980s also included a focus authors stated: on habitat, populations and recreational use “The overall aim of this continental habitat of the resource. Richard Myshak, presenting program is to maintain and manage an a summary of the emerging national appropriate distribution and diversity of high plan at the 1981 International Waterfowl quality waterfowl habitat in North America that Symposium in New Orleans (Myshak 1981), will (1) maintain current distributions of listed the goals for waterfowl management waterfowl populations and (2) under average as: 1) preserve and manage the habitat environmental conditions, sustain an abundance needed to maintain and increase waterfowl of waterfowl consistent with [population] goals numbers; 2) achieve optimum waterfowl … (U.S. Department of the Interior and population levels in relation to available Environment Canada 1986, page 13). habitat; and 3) provide optimum opportunity for people to use and enjoy waterfowl. Subsequent Plan updates continued With concerns about deteriorating habitat, evolution of the NAWMP by expanding the persistent drought in the northern plains continental partnership to include Mexico, during the 1980s, declining populations expanding habitat objectives to sustain and controversy over the effects of hunting growing waterfowl populations (NAWMP on waterfowl populations, the Canadian Committee 1994), broadening conservation government at the same time initiated strategy to regional landscapes, diversifying strategic planning for waterfowl partnerships, and managing adaptively conservation (Patterson 1985). Together, relative to environmental and human these U.S. and Canadian efforts formed the dynamics (NAWMP Committee 1998), and vanguard for negotiations that ultimately led strengthening the biological foundation of to completion of the NAWMP in 1986. The waterfowl conservation planning (NAWMP

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 Implementing the 2012 North American Waterfowl Management Plan 331

Committee 2004). Despite relatively specific the consultation process yielded three goals and assumptions outlined in the 1986 fundamental goals for North American NAWMP and continued updates to the waterfowl management: 1) abundant and Plan, ambiguity remained concerning the resilient waterfowl populations to support definition of “average environmental hunting and other uses without imperilling conditions,” the extent to which harvest habitat; 2) wetlands and related habitats management should be used to achieve sufficient to sustain waterfowl populations at population goals, and lack of an explicit desired levels, while providing places to connection between habitat management recreate and ecological services that benefit and population goals (Runge et al. 2006). society; and 3) growing numbers of Integration and efficiency were key waterfowl hunters, other conservationists themes as the waterfowl management and citizens who enjoy and actively support community strived to develop coherence waterfowl and wetlands conservation. These among habitat management, population goals are important in two ways, firstly management, and harvest (Runge et al. 2006; for the continued emphasis on healthy Anderson et al. 2007). Expanding dialogue waterfowl populations and habitat to about integrated management planning for support them and secondly, in providing the waterfowl led to the Future of Waterfowl new explicit goal for waterfowl supporters. Management Workshop in August 2008 The context of the 2012 Plan was notably (Case & Sanders 2008) where participants different than in the 1980s when a “duck agreed that work on human dimensions of crisis” was extant with record low numbers of waterfowl management should continue and breeding waterfowl and also deteriorating that the next update of NAWMP should habitat conditions. In contrast, breeding develop increasingly coherent goals for waterfowl populations in the traditional waterfowl harvest and habitat management. survey areas in North America were at record Focus on integration and reassessment levels during 2011–2013 (USFWS 2013) and of fundamental goals for waterfowl with > 15 years of liberal hunting seasons and management meant the 2012 NAWMP bag limits, the sense of urgency was less was viewed as a revision rather than as an apparent. However, mid-continent breeding update of the Plan (NAWMP Committee ground conditions aided by years of above 2012a). An extensive series of stakeholder average moisture masked the underlying workshops during 2009–2011 was designed deterioration of waterfowl habitat due to to break down administrative silos across wetland drainage and the loss of grasslands. waterfowl management public and private Additionally, growing impacts on the once sectors. The workshops identified three pristine boreal forests in Canada, water strategic foci for NAWMP 2012: 1) relevance challenges in the south and west United to contemporary society; 2) adaptable to States, and Gulf Coast marsh loss due to sea- changing ecological and social systems; and level rise and subsidence will likely soon have 3) effective and efficient with limited an effect on birds and in turn wildfowlers. funding and staff resources. Ultimately, Overall, waterfowl habitat in key North

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 332 Implementing the 2012 North American Waterfowl Management Plan

American landscapes is being lost faster than with nature through waterfowl and by it is being conserved, and threats to these highlighting environmental benefits landscapes are growing as human populations associated with waterfowl habitat increase, water quality and quantity continue conservation. to erode, energy issues often dominate land 5. Establish a Human Dimensions use decisions, and a changing climate presents Working Group to support development long-term pressures that exacerbate current of objectives for people and ensure threats. Moreover, numbers of waterfowl actions are informed by science. hunters have declined to half of 1970s levels (Vrtiska et al. 2013; Raftovich & Wilkins 2013) 6. Focus resources on important and conservation budgets are not keeping landscapes that have the greatest pace with challenges facing waterfowl. influence on waterfowl populations and Indeed, the growing detachment of North those who hunt and view waterfowl. Americans from nature (e.g. Louv 2006) is also 7. Adapt harvest management strategies a great concern for future conservation. to support attainment of NAWMP Clearly, the need for continued focus on objectives. waterfowl conservation through NAWMP is paramount. Here, we consider recommendations Priorities for implementation are found in 1–3. Building support for waterfowl > 30 key actions in the 2012 NAWMP conservation (#4 above) has become Action Plan and in the following seven primarily the responsibility of a new “Public recommendations (NAWMP Committee Engagement Team” formed under the 2012b): international NAWMP Committee. The Plan Committee and the National Flyway 1. Develop, revise or reaffirm NAWMP Council also have recently founded a new objectives so that all facets of North Human Dimensions Working Group (#5 American waterfowl management share a above) for the purpose of providing social common benchmark. science technical support and advice to 2. Integrate waterfowl management to waterfowl conservation. Efforts to focus ensure programs are complementary, resources on the most important landscapes inform resource investments and allow (#6 above) have been initiated by the managers to understand and weigh NAWMP Science Support Team; and tradeoffs among potential actions. work to adapt harvest strategies relative to revised NAWMP goals (#7 above) 3. Increase adaptive capacity so structured is pursued by the existing Harvest learning expands as part of the culture of Management Working Group chaired by waterfowl management and programme the U.S. Fish & Wildlife Service. effectiveness increases. Initial progress toward implementing the 4. Build support for waterfowl 2012 NAWMP Revision requires focus on conservation by reconnecting people recommendations 1–3 that will define

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 Implementing the 2012 North American Waterfowl Management Plan 333 actions by the waterfowl management provides perspectives on appropriateness community toward integration across of revisions in population objectives. populations, habitat, and waterfowl Substantial land-use changes have occurred supporters. Chief among these is the need in some landscapes resulting in variation in to revisit objectives established in the the capacity of habitats to support 1986 Plan. As an essential feature of waterfowl. Managers recognise the extent structured decision-making and adaptive of variation in annual environmental management (Williams et al. 2009), objective conditions and question utility of striving for setting provides context for identifying population averages. In addition, the degree management alternatives, monitoring and of management influence on population the future review of objectives. Thus, the dynamics remains uncertain. Finally, focus for initial implementation will be on managers have increased their knowledge fundamental objectives and means to and experience of the responses of birds to accomplish these objectives. habitat restoration and management and the Objectives serve three primary purposes impacts of harvest regimes. in conservation planning: 1) they operate as Numeric population objectives have been a communication and marketing tool to particularly important for habitat managers demonstrate the need for conservation; who translated resource requirements of 2) they provide a biological basis and birds into objectives for protection, planning foundation; and 3) they function as restoration and management of habitat. a performance measure for assessing Population objectives, framed as averages, conservation accomplishments. Thus, remain problematic as management targets managers must be clear about how best because of variation in wetland conditions to craft and communicate revised and other key environmental influences on objectives. Objectives should be linked at waterfowl populations. Moreover, active, administrative and implementation scales adaptive management requires sophisticated whereby tradeoffs can be identified and monitoring to track population vital rates efficiencies gained with available resources. and environmental conditions. Additionally, population objectives should also be Population objectives consistent with goals for habitat and human Objectives for waterfowl populations have use. Because these criteria frequently have remained largely unchanged since 1986. not been met, a more rigorous conceptual Benchmarks for several goose populations perspective on population status, interaction have been amended due to dramatic changes of birds with their habitat and expectations in abundance and distribution of geese; for resource use is required. however, most duck objectives have not As NAWMP population objectives are been revisited despite changes in bird reassessed, legitimate alternatives will be numbers, breeding and non-breeding considered. Among these are establishment landscapes, and the hunter population. of an objective at a relatively high level, Experience gained since the mid-1980s a minimum level below which managers are

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 334 Implementing the 2012 North American Waterfowl Management Plan concerned about sustaining populations, each in the management process is largely a “normal” operating range that dependent on subjective values placed on reflects variation in population size and numbers of birds, distribution, harvest distribution attributable to uncontrolled opportunity, viewing and ecological services environmental processes and the provided by landscapes that support birds simultaneous management of multiple and humans. There will not be a “right” species and populations. Gains in answer with respect to objectives related to management outcomes will be limited by people. The emerging question is “Whose the level of technical support required, data values matter and to what degree?” Values needed to inform decisions and the degree of waterfowl hunters, harvest managers, of complexity in the process. Although bird-watchers and landowners are different daunting, progress on these fronts has been but all are legitimate, so tradeoffs inevitably made. For example, life-cycle modelling for will be necessary. Northern Pintail (Anas acuta: Mattson et al. 2012), scaup species (Aythya affinis and A. Objectives for waterfowl habitat marila: Austin et al. 2014; Osnas et al. 2014) Protection, restoration and management of and American Black Duck (Anas rubripes: habitat are primary conservation tools Devers & Collins 2011) has already seen affecting the capacity of North American considerable progress. landscapes to support waterfowl and waterfowl enthusiasts. Substantial gains over Objectives for waterfowl supporters the period of NAWMP implementation, The 2012 NAWMP Revision explicitly estimated at nearly 7 million ha (http:// acknowledged people as fundamental to www.fws.gov/birdhabitat/NAWMP/index. the Plan. The decline in wildfowling shtm), have not necessarily kept pace with is acknowledged and integrated into net changes in landscapes, but these are management planning. Considerable poorly quantified (NAWMP Assessment changes in social structure, an aging Steering Committee 2007). When developing population and a shift to urban residence all habitat objectives, managers should take into contribute to this decline (Louv 2006; Wentz account the association between waterbird & Seng 2000). Most managers recognise the numbers and the carrying capacity of the need to increase the relevance of waterfowl landscape, as well as the influence of variable conservation to constituencies beyond environmental conditions on population hunters; however, this need is poorly demography and distribution. understood and not accepted as important Stepping-down continental objectives for by the entire waterfowl management habitat to regional or local scales is a logical community. Three interest groups are process; however, it is largely dependent specifically mentioned in the 2012 revision on selection of continental population of the Plan – waterfowl hunters, bird- objectives and an understanding of the watchers and waterfowl conservation influence of regional habitat on population supporters. The particular weight placed on processes. Thus, a key initial step for the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 Implementing the 2012 North American Waterfowl Management Plan 335 revised NAWMP is to establish population affecting waterfowl recruitment and survival objectives, despite considerable uncertainty versus those that determine engagement by about factors regulating populations of users vary considerably among regions. different species and the influence of habitats Strictly from a waterfowl perspective, in different landscapes. Efficient allocation emphasis on breeding habitat is appropriate of conservation budgets also requires because the factors primarily affecting acknowledgment of the on-going status of population growth rates occur during the habitats – whether secure or at risk in the near breeding season (see Hoekman et al. or long-term. Stepping-down revised 2002; Koons et al. 2006; Coluccy et al. population objectives will not be a trivial 2008). Human populations, however, are matter. Trends in land use and agricultural distributed differently (e.g. most reside markets worldwide represent significant outside the breeding grounds), and habitat influences on waterfowl conservation efforts, managed for users will present and sustaining habitat carrying capacity for considerations beyond the traditional continental waterfowl populations will be mission of habitat delivery for waterfowl challenging, especially with added complexity alone. Waterfowl managers therefore will be to satisfy objectives from all waterfowl challenged to formulate habitat objectives in enthusiasts. For successful waterfowl the context of consumptive and non- conservation, needs of human users of the consumptive human use plus continental resources must be considered and addressed waterfowl population objectives. using balanced strategies. To date, most habitat management From individual objectives to an partnerships have considered waterfowl integrated system – challenges at population objectives with only limited multiple scales regard for human considerations except for The 2012 NAWMP Revision accepted addressing factors directly affecting habitat that successful management of waterfowl delivery (e.g. funding for conservation and populations, conservation of waterfowl for landowners’ acceptance of programme habitat, and engagement of waterfowl options). Additionally, habitat for those users and supporters are inseparably other than traditional users (hunters) has linked components of waterfowl been considered only rarely. Complexity in conservation. To manage the different planning habitat management for the components effectively and responsively, a benefit of waterfowl will increase as management system that embraces these managers acknowledge that landscapes interrelationships should be employed. Such valuable for waterfowl also have values a coherent system will help focus on things beyond the interests of ducks and hunters. that matter most for efficient achievement Habitat objectives that integrate goals for of all NAWMP goals. waterfowl, other wildlife, and humans An integrated management system should present tradeoffs that may be quite different inform resource investment decisions by across landscapes. For instance, factors allowing managers to understand and weigh

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 336 Implementing the 2012 North American Waterfowl Management Plan tradeoffs among potential actions. This framework for waterfowl management approach will require increased adaptive decisions. They recognised a daunting capacity, and institutions and processes number of decision nodes, many decision that enable united action. Features of an makers and decision cycles operating at integrated management system should multiple spatial and temporal scales (Fig. 1), include quantifiable, coherent objectives; an and noted that analytical challenges overarching framework comprised of linked consistent with multiple objectives under models; decision tools that help inform the Plan were not independent. The team resource allocations at multiple spatial involved in preparing the Revision therefore and temporal scales; coordination among advocated instead “linked decision multiple management authorities and processes” and a continuing commitment to decision nodes; and monitoring and adaptive management. However, how to assessment to track progress and enable link various nodes and scales is not readily adaptation. apparent, and this need might vary greatly As NAWMP planners proceed with among individual management decision development of an integrated system they problems (NAWMP Action Plan 2012). face two immediate technical and process So what might comprise an integrated challenges: firstly, how will multiple management system? Certainly, coherent objectives for waterfowl management be quantifiable objectives would be one established? Can they rely on existing component, along with some concept of institutions and do they need the assistance tradeoffs amid pursuit and fulfilment of of a new entity with overarching facilitation multiple objectives. Multiple decision functions? Whatever the process, it will need processes required for management of to be iterative and adaptive. Secondly, how habitats, populations, harvest, users and will managers monitor progress toward supporters are likely to be diverse in nature, achieving expanded NAWMP objectives and we may be well-served by trying various and adapt actions to results? For instance, approaches. Several candidate approaches what technical and human resources will be have already been mentioned including needed, and who will make the many elaboration of the Joint Task Group (JTG; adaptive decisions going forward? Indeed, Anderson et al. 2007) framework, SDM, no existing entity possesses clear scenario planning, decision-criteria matrices, responsibility for all interrelated decision- resilience thinking and others (Appendix A making that will emerge in an integrated in NAWMP Committee 2012). Each has system – not the Flyway Councils, not the advantages and limitations but can provide Service Regulations Committee, not the a basis for prediction, learning and Plan Committee, and not any single country. improved decision making over time. In any During development of the 2012 case, a commitment to monitoring and Revision an ad hoc technical team tried but assessment is critical for progress in abandoned efforts to develop a singular understanding system dynamics and formal structured decision making (SDM) improving management performance.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 Implementing the 2012 North American Waterfowl Management Plan 337

Figure 1. Schematic representation of waterfowl management decisions which are made by different managers and decision-making bodies at multiple spatial and temporal scales. Linking these decisions to bring coherence to the overall management of waterfowl populations is challenging. (Illustration by John M. Eadie, University of California-Davis).

Learning how to achieve multiple Prioritizing among many monitoring and objectives simultaneously may be particularly assessment efforts will be challenging, but we challenging. Using a suite of different may find some approaches that inform conservation projects, or at least some multiple questions. Then managers must explicit tradeoffs in how individual parcels of create the commitment to undertake this vital habitat are managed, may be valuable. These adaptive management work. A necessary kinds of tradeoffs need to be addressed in related step will be to identify the main multiple places, as the nature of these coordination challenges among existing tradeoffs will vary among environmental and administrative processes and institutions and social regions and over time. ensure these are addressed in a manner that When objectives are selected, an important allows effective adaptive management for next step will be to identify main sources multiple, interrelated objectives. of uncertainty that face attainment of Linking adaptive management cycles objectives. These are likely to include matters among spatial scales (Fig. 2) would be of management control and weaknesses in advantageous. Perhaps the easiest way to our present knowledge of system dynamics. visualise this linkage is with a single suite These uncertainties may also be expressed of objectives for habitat conservation at multiple spatial and temporal scales (Fig. 2). Adding objectives for waterfowl and involve multiple human institutions. populations and users should work the same

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 338 Implementing the 2012 North American Waterfowl Management Plan

Figure 2. Links between adaptive management cycles at different spatial scales, required to ensure coherence and efficiency in waterfowl habitat management in North America. The left-hand set of links reflects the downward decision-making from continental to local scales; to the right, frequent decisions and feedback at the local scale contribute to to regional and ultimately continental decisions and outcomes. in principle although with added complexity. and among management units at equivalent Adaptive cycles should work most rapidly at scales, which should foster efficient and the smaller spatial scales where scale- effective responses of the whole system to relevant responses should be detectable changes and acquisition of new knowledge. relatively quickly. Also, existence of many Most of these linked system models are small focal areas presents opportunity for likely to be designed as decision-support innovation and experimentation in ways that tools for specific purposes and at various can accelerate learning about system scales, and no single model is likely to serve dynamics and veracity of planning the purpose for all decision-support needs. assumptions. At the continental and largest Linkage of decisions seems most important scale of interest for NAWMP, cycles of where true co-dependencies exist, such as adaptation will happen more slowly but will between harvest potential and habitat have great impact when learning and change carrying capacity or between demographic occur. Clearly, progress is made in well- metrics such as winter survival rates and connected learning organisations (Senge hunter access and success. Such linked 1990, 2006; Bennis & Biederman 1997). system models should provide a means to Therefore, we must nurture strong linkages predict consequences of management of information exchange between scales actions for attaining multiple objectives

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 Implementing the 2012 North American Waterfowl Management Plan 339 while resolving uncertainty. Some models Support Team (NSST) in 1999. The NSST, will be empirically based and rigorous, with an unfunded science-support mandate, relying on long-term data and well- struggled to generate deliverables requested documented demographic responses to by the Plan Committee. Appointments of management actions. Other models for JV science coordinators in the US and their poorly understood species or processes may part-time assignments to work on the NSST be more qualitative or hypothetical. brought much-needed capacity to bear. Coupled with the work of temporary task Increasing adaptive capacity groups like the NAWMP Continental Once objectives have been established, and Assessment team (NAWMP Assessment key decisions identified and linked, the next Steering Committee 2007), the NSST has logical step is to develop adaptive made several advancements to guide habitat frameworks and actions that will allow delivery of the Joint Ventures, but have waterfowl managers to learn from proceeded well short of their plans and management efforts (North American potential. Funding important research and Waterfowl Management Plan Committee planning activities that over-arch multiple 2012). The job of “increasing adaptive JVs has remained particularly challenging. capacity,” has at least two major Today, the broader vision of the 2012 components: 1) developing technical NAWMP Revision has moved waterfowl framework and plans to achieve increased management and the Plan Committee into a capacity, and 2) mustering political and new realm. This new vision includes science financial support and acquiring leadership to support for social and ecological sciences ensure implementation of the plan. Existing and underscores the importance of the new technical working groups should be able to Human Dimensions Working Group, the address the technical framework, but new NSST and the Harvest Management collective action seems necessary to garner Working Group. The time is rapidly resources and organise processes amongst approaching when increased, adaptive institutions so that needed adaptive loops capacity under NAWMP will be mission- actually function. critical. When waterfowl managers have With adoption of population, habitat and renewed explicit objectives to drive human goals in the new Plan, there is integrated decision frameworks, the additional need for integration of goal- adaptive capacity needed to support setting, modelling, monitoring activities and waterfowl management should become institutional support systems. The Plan both more obvious and urgent. Committee was adequately structured for its In summary, by 2016 our collective high- initial tasks of overseeing creation of the priority waterfowl management goals Joint Ventures, coordinating with the Flyway should be to: Councils, and generally guiding evolution of the 1986 Plan. However, changes began 1. Establish quantifiable objectives for with the creation of the NAWMP Science population and habitat conservation,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 340 Implementing the 2012 North American Waterfowl Management Plan

harvest opportunity and user participation Waterfowl Symposium included Ray at appropriate spatial scales and with Alisauskas, Mike Anderson, Tim Bowman, acknowledged tradeoffs among them. Bob Clark, John Eadie, Jody Enck, David Fulton, Dale Humburg, Kevin Hunt, Mark 2. Design an integrated framework for Koneff, Andrew Raedeke, Jim Ringelman, making linked harvest, habitat and user- Greg Soulliere. Contributors to the 2012 supporter management decisions where NAWMP revision and to past and future important dependencies exist among waterfowl conservation include the management objectives. countless scientists, managers and policy 3. Design and implement monitoring makers who will inform, apply and decide and evaluation programmes to track on the path forward. We thank Lisa Webb progress toward objectives and inform and Rick Kaminski for helpful comments each key decision problem. on an earlier draft of this manuscript.

4. Seek ways to fund the process. References In doing this we should recognise that we Anderson, M.G., Caswell, F.D., Eadie, J.M., are unlikely to “get it right” from the outset, Herbert, J.T., Huang, M. Humburg, D.D., so we must plan to re-plan. We would be Johnson, F.A., Koneff, M.D., Mott, S.E., Nudds, T.D., Reed, E.T., Ringelman, J.R., foolhardy to expect that a revised set of Runge, M.C. & Wilson, B.C. 2007. Report from NAWMP objectives will serve our needs for the Joint Task Group for Clarifying North American the next 28 years as have the original 1986 Waterfowl Management Plan Population Objectives objectives. This new endeavour will be and Their Use in Harvest Management. NAWMP challenging, technically and administratively Joint Task Group unpublished report, U.S. – the valuing exercises, the modelling, Fish & Wildlife Service and U.S. Geological the adaptive management frameworks, Survey, Washington D.C., USA. Accessible coordinated execution and finding fiscal at http://nawmprevision.org/sites/default/ support for the Plan will be needed to files/jtg_final_report.pdf (last accessed 10 ensure its success. False starts or dead ends July 2014). seem likely, so there may be advantages in Austin, J.E., Slattery, S., & Clark, R.G. 2014. exploring multiple options, especially at Waterfowl populations of conservation concern: learning from diverse challenges, smaller scales where relatively rapid models and conservation strategies. Wildfowl replication and learning may be most (Special Issue No. 4): 470–497. achievable. In this light, a commitment to Bennis, W. & Biederman P.W. 1997. Organizing managing adaptively may be more important Genius: The Secret of Creative Collaboration. than ever. Basic Books, New York, USA. Acknowledgments Case, D. & Sanders S. 2008. The future of waterfowl management workshop: framing Co-authors of individual presentations at future decisions for linking harvest, the 2012 NAWMP session at the Ecology habitat and human dimensions. Report and Conservation of North American 10-9-08. D.J. Case and Associates,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 Implementing the 2012 North American Waterfowl Management Plan 341

Mishawaka, Indiana, USA. Accessible In Ducks Unlimited (ed.), Proceedings of the at http://www.nawmprevision.org/sites/ Fourth International Waterfowl Symposium, New default/files/future_of_waterfowl_mgt_ Orleans, Louisiana, pp. 50–52. Ducks workshop_final_report.pdf (last accessed Unlimited, Long Grove, Illinois, USA. 10.07.14). North American Waterfowl Management Plan Coluccy, J.M., Yerkes, T., Simpson, R., Simpson, Assessment Steering Committee. 2007. J.W., Armstrong, L. & Davis, J. 2008. North American Waterfowl Management Population dynamics of breeding mallards in Plan: continental progress assessment. Final the Great Lakes states. Journal of Wildlife report. Canadian Wildlife Service, U.S. Fish Management 72: 1181–1187. & Wildlife Service, Secretaria de Medio Devers, P.K. & Collins, B. 2011. Conservation Ambiente y Recursos Naturales and Action Plan for the American Black Duck, First Secretaria de Desarrollo Social. Accessible Edition. U.S. Fish and Wildlife Service, at http://nawmprevision.org/sites/default/ Division of Migratory Bird Management, files/2007ContinentalAssessment.pdf (last Laurel, Maryland, USA. accessed 10 July 2014). Hoekman, S.T., Mills, L.S., Howerter, D.W., North American Waterfowl Management Plan DeVries, J.H. & Ball, I.J. 2002. Sensitivity Committee. 1994. 1994 update, North analyses of the life cycle of mid-continent American Waterfowl Management Plan: Mallards. Journal of Wildlife Management 66: expanding the commitment. Canadian Wildlife 883–900. Service, U.S. Fish & Wildlife Service, Secretaria Koons, D.N., Rotella, J.J., Willey, D.W., Taper, M., de Medio Ambiente y Recursos Naturales and Clark, R.G., Slattery, S., Brook, R.W., Secretaria de Desarrollo Social. Accessible at Corcoran, R.M. & Lovvorn, J.R. 2006. http://www.fws.gov/birdhabitat/NAWMP/ Lesser scaup population dynamics: what Planstrategy.shtm (last accessed 10 July can be learned from available data? Avian 2014). Conservation and Ecology 1(3): 6. Accessible North American Waterfowl Management Plan at http://www.ace-eco.org/vol1/iss3/art6/ Committee. 1998. 1998 update, North (last accessed 10 July 2014). American Waterfowl Management Plan: Louv, R. 2006. Last Child in the Woods: Saving expanding the vision. Canadian Wildlife Our Children From Nature-Deficit Disorder. Service, U.S. Fish & Wildlife Service and Algonquin Books of Chapel Hill, Chapel Instituto Nacional de Ecologia. Accessible at Hill, North Carolina. http://www.fws.gov/birdhabitat/NAWMP/ Mattsson, B.J., Runge, M.C., Devries, J.H., Planstrategy.shtm (last accessed 10 July Boomer, G.S., Eadie, J.M., Haukos, D.A., 2014). Fleskes, J.P., Koons, D.N., Thogmartin, North American Waterfowl Management Plan W.E. & Clark R.G. 2012. A modelling Committee. 2004. North American Waterfowl framework for integrated harvest and Management Plan: 2004 strategic guidance habitat management of North American strengthening the biological foundation. waterfowl: Case-study of northern pintail Canadian Wildlife Service, U.S. Fish & metapopulation dynamics. Ecological Modelling Wildlife Service and Secretaria de Medio 225: 146–158. Ambiente y Recursos Naturales. Accessible at Myshak, R.J. 1981. National waterfowl http://www.fws.gov/birdhabitat/NAWMP/ management plans: United States viewpoint. Planstrategy.shtm (last accessed 10 July 2014).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 342 Implementing the 2012 North American Waterfowl Management Plan

North American Waterfowl Management Plan Runge, M.C., Johnson, F.A., Anderson, M.G., Committee. 2012a. North American Koneff, M.D., Reed, E.T. & Mott, S.E. 2006. Waterfowl Management Plan: People The need for coherence between waterfowl conserving waterfowl and wetlands. Canadian harvest and habitat management. Wildlife Wildlife Service, U.S. Fish & Wildlife Society Bulletin 34: 1231–1237. Service and Secretaria de Medio Ambiente Senge, P.M. 1990. The Fifth Discipline: The Art y Recursos Naturales. Accessible at and Practice of the Learning Organization. http://www.fws.gov/birdhabitat/NAWMP/ Doubleday, New York, USA. Planstrategy.shtm (last accessed 10 July Senge, P.M. 2006. The Fifth Discipline: the Art and 2014). Practice of the Learning Organization. Revised North American Waterfowl Management Plan Edition. Doubleday, New York, USA. Committee. 2012b. NAWMP Action Plan: U.S. Government Printing Office. 1937. A Companion Document to the 2012 North Hearing before the Select Committee on American Waterfowl Management Plan. Conservation of Wildlife Resources, House U.S. Fish and Wildlife Service, Washington of Representatives. U.S. Government, D.C. Accessible at http://www.fws.gov/ Washington D.C., USA. birdhabitat/NAWMP/Planstrategy.shtm U.S. Department of the Interior & Environment (last accessed 10.07.14). Canada. 1986. North American Waterfowl Osnas, E.E., Runge, M.C., Mattsson, B.J., Austin, Management Plan. U.S. Department of the J.E., Boomer, G.S., Clark, R.G., Devers, P., Interior, Washington D.C., USA. Eadie, J.M., Lonsdorf, E.V. & Tavernia, B.G. U.S. Fish and Wildlife Service. 2013. Waterfowl 2014. Managing harvest and habitat as population status, 2013. U.S. Department of integrated components. Wildfowl (Special the Interior, Washington D.C., USA. Issue No. 4): 305–328. Vrtiska, M.P., Gammonley, J.H., Naylor, L.W. Pahl-Wostl, C. 2009. A conceptual framework for & Raedeke, A.H. 2013. Economic and analysing adaptive capacity and multi-level conservation ramifications from the decline learning processes in resource governance of waterfowl hunters. Wildlife Society Bulletin. regimes. Global Environmental Change-Human 37: 380–388. and Policy Dimensions 19: 354–365. Wentz, J. & Seng, P. 2000. Meeting the challenge to Patterson, J.H. 1985. The Canadian plan. In increase participation in hunting and shooting. Ducks Unlimited (ed), Proceedings of the Fifth Final Report to the National Shooting International Waterfowl Symposium, Kansas City, Sports Foundation and International Hunter Missouri, pp. 15–17. Ducks Unlimited, Long Education Association. Silvertip Productions, Grove, Illinois, USA. Ltd., Reynoldsburg, Ohio, USA. Raftovich, R.V. & Wilkins, K.A. 2013. Migratory Williams, B.K., Szaro, R.C. & Shapiro, C.D. 2009. bird hunting activity and harvest during the Adaptive Management: The U.S. Department of the 2011–12 and 2012–13 hunting seasons. U.S. Interior Technical Guide. Adaptive Management Fish and Wildlife Service, Laurel, Maryland, Working Group, U.S. Department of the USA. Interior, Washington D.C., USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 329–342 343

Wetland issues affecting waterfowl conservation in North America

HEATH M. HAGY1*, SCOTT C. YAICH2, JOHN W. SIMPSON3, EDUARDO CARRERA4, DAVID A. HAUKOS5, W. CARTER JOHNSON6, CHARLES R. LOESCH7, FRITZ A. REID8, SCOTT E. STEPHENS9, RALPH W. TINER10, BRETT A. WERNER11 & GREG S. YARRIS12

1*Illinois Natural History Survey, Forbes Biological Station – Bellrose Waterfowl Research Center, University of Illinois at Urbana-Champaign, Havana, Illinois 62644, USA. 2Ducks Unlimited, One Waterfowl Way, Memphis, Tennessee 38120, USA. 3Winous Point Marsh Conservancy, 3500 S Lattimore Road, Port Clinton, Ohio 43452, USA. 4Ducks Unlimited Mexico, Ave. Vasconcelos 209 Ote. Residencial San Agustin, Garza Garcia, Nuevo León, Mexico. 5U.S. Fish and Wildlife Service, Texas Tech University, Lubbock, Texas 79409, USA. 6Department of Natural Resource Management, South Dakota State University, Brookings, South Dakota 57007, USA. 7U.S. Fish Wildlife Service, Habitat and Population Evaluation Team, Bismarck, North Dakota 58501, USA. 8Ducks Unlimited, 3074 Gold Canal Drive, Rancho Cordova, California 95670, USA. 9Ducks Unlimited Canada, P.O. Box 1160, Stonewall, Manitoba ROC 2ZO, Canada. 10U.S. Fish and Wildlife Service, 300 Westgate Center Drive, Hadley, Massachusetts 01035, USA. 11Program in Environmental Studies, Centre College, 600 West Walnut Street, Danville, Kentucky 40422, USA. 12Central Valley Joint Venture, 2800 Cottage Way, Sacramento, California 95825, USA. * Correspondence author. E-mail: [email protected]

Abstract This paper summarises discussions by invited speakers during a special session at the 6th North American Duck Symposium on wetland issues that affect waterfowl, highlighting current ecosystem challenges and opportunities for the conservation of waterfowl in North America. Climate change, invasive species, U.S. agricultural policy (which can encourage wetland drainage and the expansion of row-crop agriculture into grasslands), cost and competition for water rights, and wetland management for non-waterfowl species were all considered to pose significant threats to waterfowl populations in the near future. Waterfowl populations were found to be faced with significant threats in several regions, including: the Central Valley of California, the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 344 Wetland issues affecting waterfowl

Playa Lakes Region of the south-central U.S., the Prairie Pothole Region of the northern U.S. and western and central Canada, the boreal forest of northern Canada, the Great Lakes region and Latin America. Apart from direct and indirect threats to habitat, presenters identified that accurate and current data on the location, distribution and diversity of wetlands are needed by waterfowl managers, environmental planners and regulatory agencies to ensure focussed, targeted and cost-effective wetland conservation. Although populations of many waterfowl species are currently at or above long-term average numbers, these populations are thought to be at risk of decline in the near future because of ongoing and predicted nesting habitat loss and wetland destruction in many areas of North America.

Key words: agriculture, climate change, dabbling duck, national wetlands inventory, playa, policy, prairie pothole.

To the casual observer, it might seem that populations and also the number of May wetland-dependent wildlife face few ponds appear to be near or above levels conservation issues at present in North observed in the early 1970s and late 1990s, America. Dahl (2006) showed a 0.3% gain in both periods thought to be the “good old deepwater and wetland area in the days” by waterfowl conservationists (Vrtiska continental United States (i.e. excepting et al. 2013). Moreover, waterfowl hunting Alaska and Hawaii) between 1998 and 2004. regulations have remained liberal since the During the early 21st century, numbers of introduction of the Adaptive Harvest breeding ducks have remained at or above Management programme in 1995, allowing their long-term average population for maximum take (regulated by bag limits) estimates, and populations of several of most species (Nichols et al. 2007; Vrtiska species (e.g. Blue-winged Teal Anas discors et al. 2013). and Northern Shoveler A. clypeata) are at all- Despite currently large waterfowl time highs (USFWS 2013). Even Lesser population sizes, many threats loom that Scaup Aythya affinis and Northern Pintail cause informed wetland and waterfowl Anas acuta populations have reversed conservationists to worry about the future. historical declines and seem to be steady or Dahl (2011) documented a loss in wetland increasing in number. The abundance of area and only modest gains in the number of ponds and wetlands containing water in May all wetlands and deepwater habitats (i.e. “May ponds”) in breeding areas combined during 2004–2009. Additionally, surveyed annually by the U.S. Fish and losses in vegetated wetlands have been Wildlife Service and the Canadian Wildlife largely offset by gains in agricultural and Service, which serves as an indicator of urban ponds and other non-vegetated wetland habitat availability and waterfowl wetlands, which likely are of less value to productivity, was 42% above the long- waterfowl, other waterbirds and other term average in summer 2013. Breeding wildlife (Weller & Fredrickson 1974; Dahl

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 345

2011). Also of concern is that wetland Recognising the ongoing and increasingly losses have not been evenly distributed significant threats to wetlands and wetland among regions and systems; for instance, wildlife, the Wetlands Working Group of losses in the Prairie Pothole Region and the the Wildlife Society held a special session lower Mississippi Alluvial Valley, which at the 6th North American Duck provide some of the most important habitat Symposium – “Ecology and Conservation for breeding and wintering waterfowl of North American Waterfowl”, to describe in North America, have been more and summarise issues affecting wetland pronounced than in other regions (Dahl conservation relating to waterfowl in North 2011; Johnston 2013). Unfortunately, we can America. Here we present topics discussed expect that current May pond abundance at this session and provide an overview of and waterfowl breeding population size are current wetland issues affecting waterfowl facing probable declines in the future conservation in North America. Our (Johnson et al. 2010; Johnston 2013). objectives are to: 1) outline the growing Agricultural policies that have long provided threats to wetlands and waterfowl in some protection for geographically-isolated North America, 2) generally highlight wetlands through the “Swampbuster” current research and management that provision in the U.S. Farm Bill now addresses these issues, and 3) provide contain reduced or increasingly ineffective recommendations for future actions that conservation provisions. Incentive-based may benefit wetland and waterfowl wetland restoration, creation and protection conservation in North America. programmes also face declining funding or elimination. Furthermore, mandates for Wetland policy ethanol production (i.e. the Renewable Fuels Standard) coupled with crop insurance The United States policies have provided incentives for In the minds of biologists, hunters and the wetland drainage in the U.S. Great Plains general public, waterfowl are stereotypically (Reynolds et al. 2006; Johnston 2013). and appropriately linked to wetlands and Reductions in federal spending and relatively other aquatic habitats. Yet, while the high waterfowl populations may dissuade waterfowl management and scientific policy makers from prioritising wetland community has dedicated substantial conservation policies in future Farm Bills. resources to population and habitat For these and many other reasons that management, there has been much less will be highlighted subsequently, we effort devoted to providing the scientific deemed it necessary to convene a forum foundation for securing policies that where scientists and conservation leaders maintain wetland habitats. The success or could discuss current wetland policy failure of these policies in maintaining the and management issues that may affect continent’s wetland habitats will ultimately waterfowl conservation efforts in the near determine the level of success achievable by future. waterfowl conservationists.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 346 Wetland issues affecting waterfowl

The series of wetland status and trends wetland losses and the incentives towards reports produced by the United States Fish maintaining and restoring wetlands were and Wildlife Service (USFWS) from the reflected in a net rate of loss 79% lower mid-1950s through to 2009 provides than that of the 1950s–1970s (Dahl evidence of the impact of policies on 2000). The trend of increasing broad and wetlands in the U.S. The first report, protective wetland policies continued examining the mid-1950s to the mid-1970s through the early 1990s, and by 2004 the net (Frayer et al. 1983), documented a loss of loss rate of wetlands most important to 113 million acres (c. 46 million ha) of waterfowl and other wildlife had declined to wetlands with net losses approaching a half- approximately 80,000 acres (c. 32,000 ha) per million acres (c. 202,000 ha) annually. year (Dahl 2006). However, implementation of the Clean However, the tide of wetland Water Act (CWA) in the mid-1970s conservation policy turned in 2001 with the provided some degree of federal protection U.S. Supreme Court in favour of the Solid to most wetlands, including the prairie Waste Agency of Northern Cook County’s potholes of the north-central United States (SWANCC) appeal against the presence of and Canada, a key region for waterfowl migratory birds being used as the sole production. The status and trends report for determinant for the U.S. Army Corps of the mid-1970s to the mid-1980s (Dahl & Engineers’ (USACE) jurisdiction over Johnson 1991) documented a slowing of the waters of the United States (SWANCC versus national rate of net wetland loss to USACE). The Supreme Court’s decision approximately one-third of pre-CWA rates. greatly narrowed the perceived jurisdiction In 1985, the Swampbuster provision of the of the Corps to regulate the drainage and federal U.S. Farm Bill, which stopped infilling of wetlands not adjacent to open agricultural subsidy payments to landowners and clearly navigable waters (Dahl 2011). who drained wetlands for farming (Dahl In response, the U.S. Environmental 2011; Johnston 2013), added another critical Protection Agency and USACE withdrew layer of protection to many wetlands at risk federal Clean Water Act protections of being drained for agricultural uses. from broad swaths of wetland categories, To complement the regulatory including so-called “geographically isolated protections of the Clean Water Act and wetlands” such as the prairie potholes, disincentives of Swampbuster, voluntary rainwater basins and playa wetlands of the incentive-based wetland conservation Great Plains (Haukos & Smith 2003). At the programmes such as the Wetland Reserve same time, funding for many of the Program, the North American Wetlands incentive-based conservation programmes Conservation Act, the Conservation peaked and has since declined. Reserve Program and the USFWS Partners The findings of the most recent for the Fish and Wildlife Program were assessment of wetland status and trends established in the late 1980s and 1990s. (Dahl 2011) mirrored this shift in Concurrently, regulatory deceleration of conservation policy. For the first time in 50

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 347 years, wetland loss accelerated, increasing by makes maintaining an adequate wetland 140% compared with 1998–2004. Policy- base to support healthy populations of based funding for wetland conservation breeding ducks impossible without wetland programmes has continued to decline, and regulations that reduce loss rates. changes to Farm Bill policy place c. 1.4 In Alberta, implementation of a new million wetlands in the Prairie Pothole Region wetland policy provides some wetland of North and South Dakota at high risk of protection and requires mitigation at a ratio being drained and lost (Reynolds et al. 2006). determined by the value of the affected wetland. However, although the new policy Canada is largely enforced for developers and the In the Prairie Pothole Region of Canada, energy sector, it is not applied consistently wetlands represent a significant obstacle to to agriculture (S. Stephens, pers. comm.). production agriculture. As a result, wetland In Saskatchewan, policy prohibits the drainage continues to occur despite growing drainage of water from wetlands from evidence of ecological goods and services an individual’s property onto another that wetlands provide, including flood landowner; however, these regulations have protection, carbon storage and groundwater been poorly enforced, resulting in conflicts recharge (Millar 1989). The jurisdiction for between neighbouring producers and Canadian wetland policy resides at the significant unauthorised drainage across provincial level (Rubec et al. 1998). As a Saskatchewan. In Manitoba, existing policy result, effective policies that protect existing protects semi-permanent and permanent wetlands must be developed for each ponds and lakes (Stewart & Kantrud 1971), provincial jurisdiction if wetlands across but shallower and more ephemeral wetlands Canadian landscapes important to remain unprotected from drainage. waterfowl, such as the Prairie Pothole Given the different stages of progress on Region, are to be protected effectively. and viewpoints regarding wetland policy High commodity prices and several years amongst the three provincial governments of above normal precipitation have resulted spanning prairie Canada, unique strategies for in high rates of wetland drainage to facilitate improving wetland policy and subsequent increased areas being put to agricultural enforcement of regulations require diverse production across the Prairie Pothole and nuanced approaches for each province. Region. For example, Ducks Unlimited Currently, conservation advocates such as Canada recently estimated that in Ducks Unlimited Canada pursue strategies Saskatchewan alone > 6,000 ha of wetlands such as building a network of grassroots were being drained on an annual basis advocates and developing an understanding (Ducks Unlimited Canada, unpubl. data). of how best to engage with those grassroots When contemporary cost estimates for advocates in the process, providing support wetland restoration are applied, the costs of to affected landowners, building coalitions restoring those drained wetlands would be with agricultural industry groups around > US$65 million. This rate of wetland loss support for wetland policy, building stronger

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 348 Wetland issues affecting waterfowl relationships with provincial staff and on regions with significant wetlands. Once ministers in key ministries, developing new in place, these inventories may prove useful science to support the economic and as the foundation for promoting subsequent ecological case for wetland regulation and conservation activities by local, state and developing a wetland monitoring system to federal governmental entities, as well as non- facilitate measuring the impact of new governmental conservation organisations. wetland policies or lack thereof. Additionally, inventories provide guidance to outside funding institutions that can help Latin America target the allocation of resources to places Latin American countries have only and activities that can generate the greatest relatively recently come to recognise the conservation return for their investment. importance and value of their wetlands To optimise wetland and waterfowl and begun to focus more attention on conservation in the Latin American and wetland conservation. In this region, earlier Caribbean region, these nations and funding public policy efforts directed at natural organisations should consider directing resource conservation focused primarily significant public policy effort toward the on establishing systems of state and development of national conservation plans federal protected areas, but wetland that include wetlands inventory data. These protection was not usually a driving force conservation plans should identify the most behind site designations. As a result, past important habitats, provide information wetland conservation tended to be largely regarding the most significant site-specific coincidental. conservation challenges, and propose More recently, the Ramsar Convention’s pragmatic actions and policies that will need initiative to identify and protect Wetlands of to be implemented to ensure long-term International Importance (“Ramsar Sites”) conservation and sustainable use of these has become an important mechanism wetlands and other wildlife habitats. The for promoting explicit recognition of continued loss and degradation of many the importance of wetlands and has important wetland ecosystems, despite the focussed additional attention on wetland existence of various international agreements conservation in Latin America. In countries and national policies, underscores the including Mexico, Colombia, Venezuela, importance of developing realistic but Argentina, Chile and Brazil, government effective conservation plans that involve and interest in the designation of Ramsar acknowledge the needs of all stakeholders in Sites has been responsible for spurring Latin America. the development of national wetlands inventories and classification systems. For Important wetlands at-risk example, Mexico has made considerable Playa wetlands progress in recording and classifying habitats across the entire country, with an Playas are dynamic, small, recharge wetlands explicit emphasis and priority being placed located in the High Plains region of the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 349 western Great Plains in the central U.S. With source and associated methodology used to ecological conditions that reflect the harsh, identify playas, with published figures unpredictable environment of the High ranging from 30,000–80,000 playas (Smith et Plains, playas form a complex system al. 2012; D. Haukos, pers. comm.). Although providing numerous ecological functions the large number of playas reported as and services, including habitat for migratory present on the landscape gives the mistaken waterfowl (Haukos & Smith 1994). Essential impression that there are sufficient to playa function is the erratic fluctuation functional playas capable of providing between wet and dry states that creates a ecological services for waterfowl, Johnson et diversity of playa conditions or habitats al. (2012) estimated that 17% of historical throughout the entire High Plains (Smith playas are no longer detectable on the et al. 2012). Inundation patterns and southern Great Plains (Oklahoma, Texas hydroperiods of playas vary annually with and New Mexico). In addition, only 0.2% of the average playa being inundated during existing playas have no wetland or watershed January once every eleven years in Texas and modification. Further, Johnson et al. (2012) New Mexico (Johnson et al. 2011a). estimated that 38.5% of historical playas Playas provide habitat for migrating, had been lost from the landscape or wintering and breeding waterfowl (Ray et al. experienced cultivation of the hydric soils, 2003; Baar et al. 2008; Haukos 2008). The which can greatly reduce or eliminate natural number of inundated playas during winter forage for waterfowl. The greatest threat determines the number of wintering to playas is unsustainable sediment waterfowl; Johnson et al. (2011a) reported accumulation (Luo 1997; Smith 2003; Tsai that the percent of inundated playas varied 2007). Combining physical wetland loss, from near zero in dry years to > 50% in wet direct wetland cultivation and fill due to years. During wet years, overwinter survival sediment accumulation results in an of Mallard Anas platyrhynchos and Northern estimated 60% of historical playas that are Pintail in the High Plains is greater than for no longer available to provide habitat for any other wintering area in North America waterfowl (Johnson et al. 2012). Of the (Bergan & Smith 1993; Moon & Haukos remaining playas on the southern Great 2006). Estimated numbers of wintering Plains, none are fully functional (Johnson ducks using southern playas during January 2011b). These impacts to playa ecosystems ranges between 200,000 and 3 million likely contributed to the 32% decline in depending on environmental conditions average body condition of Northern Pintail such as precipitation levels and winter from the mid-1980s to early 2000s (Moon temperatures (USFWS 1988; Haukos 2008). et al. 2007), with potential associated The historical number of playas is cross-seasonal effects on survival and unknown because of extensive landscape reproductive capacity (Mattson et al. 2012). alteration in the High Plains during the past Despite the acknowledged value of playas century (Smith et al. 2012). Recent estimates to waterfowl, conservation efforts have of playas vary greatly depending on the been stymied during the past three decades.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 350 Wetland issues affecting waterfowl

The vast number of playas, and the lack that these actually drive playa ecosystems. of perceptible physical differences in Relatively long temporal periods may exist their characteristics (excepting inundation between ecological states that provide high frequency) which provide value as wildlife quality habitat for waterfowl. Finally, any habitat or contribute to ecological goods conservation effort must consider the role and services, have paralysed efforts to and contribution of individual playas to the conserve these wetlands. The value of entire system when prioritising playas for playas is greatest when they are considered conservation. Despite recognition of use of in aggregate and regionally, although this playas by waterfowl, the capacity of the approach is rarely used in conservation playa system to support waterfowl is efforts (Smith et al. 2011; Johnson et al. declining (Moon & Haukos 2006; Moon et 2012). Finally, there is lack of federal and al. 2007; Smith et al. 2011). Consequently, state regulations or incentives to encourage a multifaceted approach is needed to the protection of playas, and no develop a playa conservation strategy that requirement to mitigate for any negative includes: 1) an educational effort to impacts on playa wetlands (Haukos & Smith accumulate support for playa conservation, 2003; Johnson et al. 2011b). The U.S. 2) modification of current conservation Department of Agriculture’s Conservation programmes so that playas are competitive Reserve Program, which has limited focus for funding, and 3) greatly accelerating on wetlands compared with other habitats research efforts to accumulate knowledge within the programme, is the main relative to playa ecology, management and conservation initiative affecting playas on their status across the landscape. the High Plains. Unfortunately, playas in Conservation Reserve Program watersheds Boreal forest wetlands have altered hydrology characterised by North America’s boreal forest (hereafter, reduced inundation frequency and Boreal) is part of the largest terrestrial biome hydroperiod possibly resulting from use of and unspoiled wetland and forest ecosystem non-native vegetation in CRP plantings in the world. This 600 million ha landscape (Cariveau et al. 2011; Bartuszevige et al. 2012; stretches from western Alaska to Labrador O’Connell et al. 2012). and accounts for > 35% of the continent’s Conservation efforts should be forest-cover (Wells & Blancher 2011). coordinated at larger spatial and temporal Wetlands comprise 6% of the earth’s land- scales to identify accurately the value of an cover, yet Canada alone has 25% of the individual playa. Moreover, conservation world’s wetlands (PEG 2011). Most of programmes need to be tailored specifically Canada’s wetlands (> 85%) are in the Boreal, to playas as current efforts are not effective including bogs, fens, swamps, marshes and (Bartuszevige et al. 2012; O’Connell et al. open water basins. Alaska’s Boreal has 2012). Efforts to conserve playas will > 2,000 rivers and streams that feed a water- benefit from recognition that extreme rich wetland landscape. North America’s environmental conditions are normal, and Boreal holds 25% of the freshwater and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 351

> 30% of soil carbon on the planet (PEG (> 40 ha in area) due to the melting of 2011). Despite their former isolation and permafrost, increased evaporation and vast coverage, North America’s Boreal transpiration rates, and aggregation of wetlands face increasing threats from climate floating emergent vegetation and associated change and expansion of industrial activities. inorganic sediments, resulting in regional Prairie and boreal wetlands provide decreases in surface water area (Smith et al. breeding habitat for the majority of duck 2005; Riordan et al. 2006; Roach et al. 2011). pairs across North America (Slattery et al. The extent of these changes across the 2011). Breeding season population estimates Boreal is currently unknown, but substantial for the western Boreal region alone are increases are expected. 13–15 million birds, with many species While increasing temperature may having ≥ 50% of their breeding populations represent a threat beyond the control of in the Boreal (Wells & Blancher 2011). The classic waterfowl conservation mechanisms, prairie and boreal biomes are arguably other more direct anthropogenic landscape integrated ecologically as ducks may use the changes may be more amenable to Boreal for nesting during prairie droughts sustainable development. Changes to and annual wing moult (Baldassarre & Bolen hydrology can result in long-term drying (e.g. 2006). Consequently, extensive changes to Bennett Dam on the Peace-Athabasca Delta) boreal waterfowl habitat could have or flooding (e.g. Ramparts Dam proposed for continental-level implications for waterfowl the Yukon River and also several large conservation objectives. operations in Quebec). Water pollution can The perception of a pristine Boreal has potentially reach large blocks of watersheds changed rapidly because of the wide range of because Boreal wetlands are often development activities occurring there, and hydrologically connected through subsurface development is predicted to increase flow (Smerdon et al. 2005). Timber harvest substantially into the future (Bradshaw et may increase runoff and thus local flooding, al. 2009; Wells 2011). Seven distinct and this can have a direct effect on the anthropogenic pressures threaten the North breeding success of cavity-nesting birds. American Boreal, including agricultural Road construction can impound or drain expansion, petroleum exploration and water flowing to or from wetlands. We are development, forestry, hydroelectric only just beginning to understand the impact development, mining, acid precipitation and of these factors on waterfowl and their climate change. Few regions have already and habitats, which challenges conservation are expected to experience greater changes in efforts and necessitates a cautious approach mean temperatures than the Boreal (Soja et to development and wildlife management in al. 2007; Bradshaw et al. 2009; Stocker et al. the region. 2013), yet this biome has a great influence on Protection of water quality, quantity and global temperature and carbon storage hydrologic patterns appears critical to (Bonan 2008). Impacts on Boreal wetlands conservation of waterfowl habitat within may include loss of lakes and wetlands the Boreal. Because most Boreal wetlands

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 352 Wetland issues affecting waterfowl recharge through lacustrine or riverine threaten wetland function and value for processes, those in the Alaskan boreal forest waterfowl in the Prairie Pothole Region are protected under the U.S. Clean Water (Johnston 2013; Wright & Wimberly 2013). Act, but recent Supreme Court decisions Prairie wetlands have been identified as (e.g. SWANCC versus USACE) have muddied particularly vulnerable to climate change. the jurisdictional waters for many wetlands Evidence for this conclusion has come not immediately adjacent to navigable rivers from an inter-institutional and multi- or streams. In contrast there is almost no disciplinary team of investigators which broad wetland protection in Boreal Canada, has developed and used two simulation either at the federal or provincial/territorial models, WETLAND SIMULATOR and levels, although recent legislation in Alberta WETLANDSCAPE, to project future may provide some level of protection. consequences of climate change on prairie Widespread and enforceable legislative wetlands and waterfowl (e.g. Poiani & protections are critical to ensuring that the Johnson 1991; Poiani et al. 1995, 1996; Boreal can support key North American Johnson et al. 2005, 2010; Werner et al. 2013). waterfowl populations into the future. These researchers have reached four main conclusions after 20 years of research on Prairie wetlands the subject: 1) temperature matters, 2) Wetlands potentially represent the most geography matters, 3) impacts may have critical and limiting components of the already occurred, and 4) threshold effects landscape for breeding waterfowl (Kantrud may yield future surprises. A representative & Stewart 1977). The Prairie Pothole Region simulation using weather data (1986–1989) of the north-central United States and south from the Orchid Meadows field site central Canada produces up to 75% of demonstrated the effect of increasing air waterfowl in North America (Smith et al. temperature on the length of time that water 1964; Mitsch & Gosselink 2007). Wetland stands (hydroperiod) in a semi-permanent density in this region ranges from 4–38 wetland basin (Johnson et al. 2004, 2010). potholes/km2 (Baldassarre & Bolen 2006), Raising the temperature a modest 2°C but more than half of the original wetlands shifted wetland permanence type from in the region have been lost or highly semi-permanent (not dry during the 4-year modified, principally for agriculture (Mitsch simulation) to seasonal (drying annually). A & Gosselink 2007). Moreover, conversion 4°C increase changed the wetland into one of native grassland and pastures to row- more typical of a temporary wetland that crop agriculture can have a dramatic effect dried by late spring or mid-summer each on wetland integrity by increasing sediment year. This simulation, and hundreds more and chemical runoff within the watershed that have been completed across the Prairie (Zedler 2003). A myriad of factors including Pothole Region (e.g. Poiani et al. 1996; Poiani agricultural policy, changing wetland & Johnson 2003), clearly illustrate how regulations, improved farming and land sensitive prairie wetland hydrology is to air clearing technology, and climate change temperature.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 353

The Prairie Pothole Region is of modest was conducted to determine if the change geographic area, comprising c. 800,000 km2 in climate between two 30-year periods in the U.S. and Canada. Despite its size, a (1946–1975 and 1976–2005) was sufficient strong northwest to southeast climatic to have affected wetland productivity. If so, gradient exists within the region; mean the analysis would provide evidence that annual temperature ranges from about trends for warming and drying projected 0–10°C and mean annual precipitation earlier for the mid 21st century (Johnson et from about 35–90 cm (Millett et al. 2009). al. 2005) may already have started in the late The intersection of these two climatic 20th century. The model indicated that gradients produces different sub-regional climate changes were sufficient to have climates, wetland functional dynamics and affected the wetland cover cycle, a major responsiveness to climatic change. Model indicator of wetland productivity quantified simulations using data from regional by a cover cycle index (Werner et al. 2013). weather stations with long-term records This analysis is the first to present evidence (≥ 100 years) show that the response of that climate change may already have wetlands to climate change will be highly affected wetland productivity in part of the variable geographically (Johnson et al. 2010). Prairie Pothole Region. The most favourable climate in the Prairie Climate changes that exceed ecological Pothole Region for wetland productivity thresholds can produce rapid and surprising during the 20th century is projected to shift changes in the functioning of natural eastward where there are fewer un-drained ecosystems (e.g. Holling 1973; CCSP 2009). wetland basins and much less grassland The most productive semi-permanent available as nesting habitat for waterfowl. prairie wetlands pass through three stages The naturally drier western edge of the during weather cycles: dry marsh, lake marsh Prairie Pothole Region, described as a and hemi-marsh (which includes both “boom or bust” region for waterfowl regenerating and degenerating sub-stages; production, may become largely a “bust” van der Valk & Davis 1978). Climatic should the future climate be more arid as thresholds associated with drought must be projected (Johnson et al. 2010). This possible reached and exceeded for habitats to enter future “mismatch” between the location of the dry marsh stage, as must those a productive wetland climate and functional associated with a precipitation deluge wetland basins stands as a current challenge needed to enter the lake marsh stage. for wetland managers as they develop future Between these two extremes, the most plans to allocate resources for wetland productive hemi-marsh stage is reached. conservation and management across the Ratios that produce the highest indices for Prairie Pothole Region. wetland productivity over decadal time The northwest portion of the Prairie intervals are approximately: 25:50:25 (dry, Pothole Region (west-Canadian prairies) hemi and lake, respectively). Climate warmed and dried late in the 20th Century changes that cause wetlands to be “stuck” in (Millett et al. 2009). A hindcast simulation either the lake or dry marsh extremes stop

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 354 Wetland issues affecting waterfowl vegetation cycling and decrease productivity. amendment to the 1934 “Duck Stamp Act” Because the majority of ducks produced in (legislative documents:16 U.S.C. 718-718j, 48 North America are reared in Prairie Pothole Stat.452; P.L. 85-585; 72 Stat. 486) (Loesch et Region wetlands, biologists are concerned al. 2012). The SWAP amendment authorised that wetlands responsible for past high rates that proceeds from the sale of duck stamps of waterfowl production could become and the import duties on ammunition and drier and fail in the future by never or rarely firearms should be used for the acquisition of reaching the lake marsh and hemi-marsh fee title (i.e. absolute ownership) or limited- stages of the cycle (Sorenson et al. 1998; interest title (restricted ownership) of Johnson et al. 2010). Waterfowl Production Areas, and also for Twenty years of modelling and field purchasing limited interest easements over research have found that prairie wetlands are Waterfowl Production Areas in Prairie highly sensitive to changes in climate and Pothole Region states (USFWS 2013). that they respond differently to wide- Over the past 50 years, the USFWS ranging sub-climates across the Prairie and its partners (e.g. sportsmen and Pothole Region. Moreover, they may already women, private landowners and non-profit have been negatively affected by climate conservation organisations) have acquired warming in the Canadian prairies, and may ownership of nearly 0.7 million ha in not reach water level thresholds under a National Wildlife Refuges and Waterfowl warmer climate needed to maintain historic Production Areas and easements over an dynamics and productivity. We suggest additional 1.1 million ha of Waterfowl development of an early warning system to Production Areas in the U.S. Prairie Pothole detect the onset of climate change across Region. The SWAP has thus acquired the Prairie Pothole Region by conducting easements to conserve a network of privately simulation modelling and field monitoring owned wetlands and grasslands which in tandem to provide further understanding provide nesting sites for breeding birds in of changes to date and to improve accuracy proximity to larger Waterfowl Production in predicting for future changes in Prairie Area wetland basins purchased for their Pothole Region wetlands. importance as brood-rearing habitat. During the first 35 years, habitat was acquired Prairie wetland conservation policy by USFWS biologists who applied Despite a changing climate and their knowledge of the area to prioritise anthropogenic denudation of large areas of acquisitions. More recently, spatially explicit the landscape, all is not lost in the Prairie habitat and biological data have been used to Pothole Region. To help protect critical develop statistical models used by the habitat for waterfowl, visionary waterfowl USFWS to assess the Prairie Pothole Region biologists and managers recognised the landscape (Stephens et al. 2008). Habitat importance of the region and initiated the conservation efforts are then focused toward USFWS’s Small Wetlands Acquisition areas that produce the greatest benefits for Program (SWAP) in 1958 with an migratory bird benefits, given the limited

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 355 conservation funds (Reynolds et al. 2006; be active in pursuing conservation goals Niemuth et al. 2008). in order to maintain North America’s In addition to traditional measures of waterfowl populations. Meeting existing conservation progress (e.g. money expended wetland and grassland conservation goals for land acquisitions and the number of is a daunting challenge (Doherty et al. acres protected), the success of the SWAP 2013). Only through collaborative and is assessed using measurable biological complementary efforts, which incorporate a outcomes such as the abundance of science-based approach to determining the waterfowl pairs and their breeding success. best places in the landscape for directing Through the purchase of wetland and conservation resources in the face of habitat grassland easements on private lands, the loss, will effective conservation of wetlands USFWS and its partners have secured in the Prairie Pothole Region be achieved. breeding habitat for an estimated 1.1 million waterfowl pairs across 13 species of Lower Great Lakes marshes waterfowl. The resultant effort contributes The lower Great Lakes coastal marshes are approximately 708,000 recruits annually for valuable areas for staging and wintering Mallard, Northern Pintail, Gadwall A. waterfowl and are among the most strepera, Blue-winged Teal and Northern biologically significant wetlands within the Shoveler annually (Cowardin et al. 1995; Great Lakes region. These marshes have USFWS Habitat & Population Evaluation long been recognised for their importance in Team unpubl. data). While this large providing habitat for a wide variety of flora landscape-scale approach to waterfowl and fauna, and in particular for migratory conservation has been highly successful, an birds. As an example, the coastal wetlands of additional 3.8 million ha of grassland (88% northwest Ohio alone support c. 500,000 of the remaining grassland) and 0.7 million itinerant waterfowl during autumn migration ha of wetlands (75% of the remaining (Ohio Division of Wildlife, unpubl. data). wetlands) in the U.S. Prairie Pothole Region These marshes are also subject to a have been prioritised for protection great number of anthropogenic stressors, (Ringleman 2005). In 2012, land values including dredging, nutrient/pollutant averaged across the northern plains states loading, altered hydrological regimes and the from North Dakota to Kansas were introduction of non-native species. Today, a US$5,831/ha (USDA 2012) and, if applied majority of the region’s coastal marshes and to the 4.5 million-ha goals, would require wetlands have been drained or replaced by > US$2.6 trillion in fee-title acquisition costs. shoreline development or have been further Through various partnership efforts degraded by altered hydrology and sediment including the Migratory Bird Conservation deposition. Only 5% of the original 121,000 Fund, the Land and Water Conservation ha of Lake Erie marshes and swamps in Fund and funding under the North northwest Ohio remain (Bookhout et al. American Wetlands Conservation Act, 1989), and habitat loss continues to reduce the USFWS and its partners continue to the area available for diverse wetland plant

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 356 Wetland issues affecting waterfowl communities capable of supporting wildlife are less well understood. For waterfowl populations. Habitat loss of this example, only a handful of studies have magnitude underscores the importance of shown the effects of Common Reed on maintaining the remaining habitat at the wildlife use and diversity, including studies highest level of quality possible. on turtles (Bolton & Brooks 2010), toads A wide variety of invasive species now (Greenberg & Green 2013), passerine birds dominate wetland flora in many lower Great (Meyer et al. 2010) and other wetland wildlife Lakes coastal marshes, having displaced (Schummer et al. 2012). In contrast, a native vegetation and in many cases substantial research base of Common Reed important waterfowl resources (Mills et al. biology, proliferation and management 1994; Zedler & Kercher 2004). In fact, exists in the form of peer-reviewed articles, invasive species are now considered the white papers and websites (e.g. http:// primary cause of wetland degradation in the www.greatlakesphragmites.net). Extensive region. The most abundant, widespread and research into chemical and biological harmful invasive plant species within these control measures for Purple Loosestrife wetlands include Common Reed Phragmites similarly led to the release of beetles australis, Reed Canary Grass Phalaris Galerucella sp. as a highly successful arundinacea, Curly Pondweed Potamogeton biological control during the 1990s, despite crispus, Eurasian Watermilfoil Myriophyllum a lack of data to show that the plant had spicatum, and non-native Cattail Typha negative impacts on the environment angustifolia and T. glauca. Other less (Hager & McCoy 1998; Treberg & Husband widespread but significant invasive species 1999). include European Frog-bit Hydrocharis While no single management strategy can morsus-ranae, Japanese Knotweed Polygonum be employed to treat infestations of these cuspidatum, Yellow Flag Iris Iris pseudacorus, diverse invasive plants, similarities do exist Purple Loosestrife Lythrum salicaria and among species. Most often, managers Water Chestnut Trapa natans. Invasive employ an Integrated Pest Management species outbreaks continue to occur in this (IPM) strategy that combines one or more region and relative newcomers such as techniques including mowing or harvesting, Flowering Rush Butomus umbellatus are smothering, drowning, herbicide treatments, quickly becoming established at nuisance biological control agents, controlled burns levels. and reseeding with native species In most cases, invasive plant species alter (Radosevich 2007; Holt 2009). With the the biotic and abiotic environment of exception of Purple Loosestrife, where wetlands by excluding native plants, biological control proved successful, the reducing plant diversity and modifying most effective and widely used strategies wetland processes (Drake et al. 1989; Davis typically include herbicide application within et al. 1999; Meyerson et al. 1999; Windham & the IPM strategy. As an example, the most Lathrop 1999; Rooth et al. 2003). However, effective control of Common Reed includes the indirect effects of invasive plants on a late-summer application of glyphosphate

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 357 herbicide followed by a spring burn or other management tools, implementing post- thatch removal method (J. Simpson, unpubl. treatment monitoring and research, and data; MDEQ 2008). Variations of this organising and educating landowners. method have also been applied effectively Invasive wetland plants are widespread for many other emergent invasive plants. and continually establishing across coastal Other aquatic-approved herbicides, such and inland wetlands within the Great Lakes as those formulated with imazypyr, are region. Management of invasive plants is equally effective at removing invasive plants, unavoidable in order to continue providing but residual action limits subsequent quality wetland habitat for waterfowl regeneration of native species. Submerged and other wetland-dependent species. or floating-leaved vegetation is typically Management strategies continue to be managed with granular or similar broadcast refined, tested and researched, but research herbicides in conjunction with mechanical into the biological implications of these mowing or harvesting. Ironically, some species should continue. Up-front research success has been also demonstrated by using demonstrating the negative impacts of these non-native Common Carp Cyprinus carpio to species is essential for prioritising their reduce monocultures of submersed invasive management and focusing effort on species plants (Kroll 2006), but carp often become of greatest concern. Additionally, a greater established and can remove desirable native understanding of the indirect effects of vegetation (Bajer et al. 2009). these plant species on waterfowl and In response to the logistical and financial other wetland-dependent wildlife is required hurdles associated with managing large to avoid expending exhaustive control non-native plant invasions, stakeholders measures on species whose ecological in the Great Lakes Region of the U.S. and consequences are unproven. elsewhere have united to form cooperative weed management areas. These diverse Central Valley of California groups now exist in most Great Lakes The Central Valley of California supports an states and provinces and represent a cross average of about 5.5 million wintering section of government agencies, local units waterfowl annually, making it one of the of government, non-profit conservation most important regions for waterfowl in groups, community associations and North America. However, the Central Valley individual landowners. In many cases, these has lost approximately 95% of its original associations form to address ecological, wetlands due to flood control, urbanisation social and economic problems linked to and conversion to agriculture (Fleskes 2012). vegetation management along developed During the past 20 years, conservation shorelines and within recreational sites. programmes such as the Wetlands Reserve Using private and government grant funds, Program, the North American Wetlands these cooperative weed management areas Conservation Act, the Migratory Bird have made progress by identifying and Conservation Fund and the state’s Inland prioritising treatment sites, providing Wetlands Conservation Program have

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 358 Wetland issues affecting waterfowl provided a means to protect, enhance or of contaminants and require expensive restore former and existing wetlands monitoring programmes to demonstrate throughout the Central Valley. Additionally, compliance. Additionally, wetlands and intensive management of remaining wetlands flooded rice fields are ideal environments for food production, along with flooded for methylation of mercury – the form of grain (especially rice), has helped to mitigate mercury which readily bioaccumulates and for wetland loss and allowed continued is toxic to humans and wildlife (Ackerman support of large numbers of waterfowl. & Eagles-Smith 2010). Mercury is a While partners of the Central Valley Joint legacy contaminant from the gold rush of Venture have made considerable progress the 1800s and is widespread throughout towards habitat goals of the North northern Central Valley watersheds. American Waterfowl Management Plan, Regulations restricting methylmercury changing policies and demand for limited discharge into the San Joaquin-Sacramento resources, such as water, hinders Delta may inhibit wetland restoration and management of existing wetlands and management and discourage flooding of could impair farming practices that benefit rice fields during autumn and winter. nesting and wintering waterfowl. Water Ongoing conservation planning efforts in supply for certain National Wildlife Refuges the Sacramento-San Joaquin Delta region of and State Wildlife Areas, as well as other the Central Valley emphasise the restoration wetland complexes, was required under of anadromous fish runs (e.g. salmon Salmo the provisions of the 1992 Central Valley and Oncorhynchus sp.) and other endangered Project Improvement Act. However, the full fish (e.g. Delta Smelt Hypomesus transpacificus), allocation of water required under the Act possibly at the expense of waterfowl habitat. has been achieved only once in the past 20 For example, proposed breaching of levees years (G. Yarris, pers. comm.). In-stream of some managed wetlands in the Suisun flow requirements for fish species protected Marsh to restore tidal action and provide under various state and federal plans are fish habitat will reduce managed wetlands competing pressures on the water available in the region and require the restoration for wetland management, such as winter- or creation of new managed wetlands flooding of rice fields, in the Central Valley. elsewhere to compensate for this loss. The Moreover, the Clean Water Act, which use of tidal wetlands by dabbling ducks is protects wetland resources throughout the low compared to managed wetlands (Coates United States, increases management et al. 2012). Thus, tidal restoration may complexity in certain situations. Because of reduce the waterfowl carrying capacity of the altered hydrology of the Central Valley, the Suisun Marsh, decreasing its importance most wetlands are managed with controlled for ducks in the Pacific Flyway. flooding and drainage and thus are subject Another recent constraint to wetland to the same regulations as other water management is the mosquito abatement diverters and dischargers. Current or policies of vector control districts. Because proposed regulations will limit the discharge many wetlands in California are near urban

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 359 areas, summer irrigation for waterfowl food likely coming to a close. Secondly, the fate of plant production, early autumn flooding for large scale wetland conservation lies with shorebird migration and other management private landowners – public land and areas activities that may produce mosquitoes are protected by conservation easements will discouraged. Although alternative wetland likely not sustain the current breeding management strategies are being developed populations of waterfowl in most of North in some cases (Washburn 2012), costs America. Thirdly, wetland conservation associated with mosquito control have policies and objectives must be robust to the created a disincentive to implement wetland wide variety of political, societal and management practices on both public and environmental shifts or vagaries. One private wetlands (Olson 2010). For example, such environmental factor important to mosquito control costs have tripled on State conservation priorities is changing climate, Wildlife Areas since concerns of public where simulations have shown potential exposure to West Nile Virus have come to changes in waterbird productivity and the fore (B. Burkholder, pers. comm.). impacts on wetland availability when certain Constraints, restrictions and regulations climate thresholds are exceeded. Fourthly, on wetlands and flooded agriculture in the increasing demand for water due to urban Central Valley likely will continue into the and population growth, irrigated agriculture, future as the demand for water increases. and other commercial uses (e.g. hydraulic Creative solutions to wetland restoration fracturing) combined with expected impacts and management, and especially increased of climate change will increase competition participation in policy development, will for and cost of water for managed wetlands be critical for advancing the goals of and waterfowl habitats. Fifthly, increased the Central Valley Joint Venture and wetland drainage for agriculture followed ensuring that sufficient habitat exists for all by increased crop irrigation increases wetland-dependent species in the Pacific water requirements while reducing the Flyway. opportunities for aquifer recharge. Sixthly, updating and improving existing data on Looking ahead wetland distribution and quality for waterfowl is needed but will be difficult Challenges given declining government budgets and During our session, a number of key points changes in agency priorities. Overall, and challenges to wetland conservation and managing waterfowl populations and their waterfowl management became apparent. associated habitats in the face of climate Firstly, unless there is an immediate and change, invasive species and other biotic significant change in a) wetland protection stressors will be challenging. measures, and b) agricultural policies that provide a disincentive to wetland drainage Opportunities and conversion, the recent “good old days” In spite of these challenges, there are also a of abundant wetlands for waterfowl are number of opportunities in the near future

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 360 Wetland issues affecting waterfowl that may directly or indirectly affect wetland and, in some cases, potential function for and waterfowl conservation. wetland restoration sites. For each function, wetlands providing the function at high or National wetlands inventory moderate levels are predicted based on Strategic conservation is critical to achieve certain properties included in the database. significant progress towards wetland Correlations between database features and conservation goals (Stephens et al. 2008) and functions were developed first by consulting accurate information on the location, type the literature and then by peer review and status and trends of wetlands is vital to from regional scientists. For provision of this effort. In 1974, the USFWS established waterfowl and waterbird habitat, in addition the National Wetlands Inventory Program to the high and moderate categories, a third (NWI) to provide information on the category for Wood Duck Aix sponsa habitat location, distribution and characteristics of was created because this species frequents U.S. wetlands. By late 2014, the NWI is wooded swamps along rivers and streams as expected to be complete for the lower 48 opposed to more open water wetlands (e.g. states, yet by that time much of the data marshes) occupied by most other waterfowl will be > 25 years old. While NWI maps and waterbirds. NWI+ data and the results and geospatial data showing wetland of NWI+ analyses are displayed via an types (Cowardin et al. 1979) have helped online map (NWI+ web mapper at promote wetland conservation, continual http://aswm.org/wetland-science/wetlands- updating and additional information (e.g. one-stop-mapping); NWI+ reports are also hydrogeomorphic properties) is needed to posted. This tool provides users with a first use NWI data for predicting wetland approximation of wetland functions across functions and determining more readily their large geographic areas. To date, such data value to organisms of interest (e.g. waterfowl). are available or will soon be posted for five Recognising this need, the USFWS recently entire states (CT, DE, MA, NJ and RI) while developed descriptors for landscape position, pilot or special projects are completed or are landform, water flow path and waterbody in progress for parts of other states (AK, type (LLWW descriptors; Tiner 2003, 2011) CA, MD, MS, NH, NY, PA, SC, TX, VA, VT to supplement NWI data on a case-by-case and WY). basis. When the Federal Geographic Data The NWI+ data provide a better Committee established its wetland mapping characterisation of wetlands, an expanded standard (FGDCWS 2009) for the federal geospatial database and a preliminary government, it suggested adding these landscape-level assessment of wetland attributes to increase the functionality of the functions. This information is valuable to NWI database. fish and wildlife biologists, conservation When LLWW descriptors are added to planners, ecosystem modellers, regulatory existing NWI data, a “NWI+ database” is personnel and the general public. Limited created. The NWI+ database is used to NWI funds do not allow these data to be predict 11 functions of existing wetlands produced nationwide, so NWI+ data are

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 361 project area-focused. With further budget should nonetheless have input to reductions imminent, such data, as well discussions regarding the anticipated effects as updated traditional NWI data, will of new and ongoing policies on wetlands. likely come mainly from user-funded initiatives. Other agencies/organisations Acknowledgements have produced or are producing NWI+ data We thank more than 100 wetland and for parts of many states (MI, MN, MT, NM, waterfowl professionals for participating in OR, WI), while some states (CT, DE, NY, our special session and the thousands of and PA) have funded NWI+ work in their conservationists who work daily to restore state. NWI+ data will provide new and protect wetlands for waterfowl opportunities for assessing and assigning throughout the world. We thank our functional values to wetlands at the time that Associate Editor L. Webb and Editor E. they are mapped, and have the potential to Rees for helpful comments which improved increase the efficiency of conservation this manuscript. Results, conclusions and planning for target species or groups (e.g. opinions stated in this paper do not dabbling ducks, wood ducks). necessarily represent the views of the Influencing policy USFWS, The Wildlife Society or other state and federal agencies. Scientists, wetland managers and other conservationists should not simply react to References policy shifts that influence wetland loss, but must also work to influence them. There are Ackerman, J.T. & Eagles-Smith, C.A. 2010. many opportunities to incorporate science Agricultural wetlands as potential hotspots for mercury bioaccumulation: Experimental into the policy debates that are shaping evidence using caged fish. Environmental the future of waterfowl management. Science and Technology 44: 1451–1457. Waterfowl scientists and managers can, and Baar, L., Matlack, R.S., Johnson, W.P. & Barron, must, focus increased efforts on providing R.B. 2008. Migration chronology of information that can influence the future of waterfowl in the Southern High Plains of wetland conservation policies, such as the Texas. Waterbirds 31: 394–401. Clean Water Act, that hold in the balance the Baldassarre, G.A. & Bolen, E.G. 2006. Waterfowl future of tens of millions of acres of Ecology and Management, 2nd edition. Kreiger waterfowl habitat. Moreover, waterfowl Publishing, Malabar, Florida, USA. conservationists should engage private Bajer, P.G., Sullivan, G. & Sorenson, P.W. 2009. landowners and convey to them the Effects of rapidly increasing population of importance of wetlands for waterfowl as common carp on vegetative cover and waterfowl in a recently restored midwestern well as the myriad of other functions and shallow lake. Hydrobiologia 632: 235–245. benefits that these habitats provide for Bartuszevige, A.M., Pavlacky, D.C., Jr., Burris, L. society. Although government restrictions & Herbener, K. 2012. Inundation of playa on advocacy can limit the participation of wetlands in the western Great Plains relative many scientists in policy debates, experts to landcover context. Wetlands 32: 1103–1113.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 362 Wetland issues affecting waterfowl

Bergan, J.F. & Smith, L.M. 1993. Survival rates of Dahl, T. E. 2000. Status and Trends of Wetlands in female mallards wintering in the Playa Lakes the Conterminous United States 1986 to 1997. Region. Journal of Wildlife Management 57: U.S. Department of the Interior, U.S. Fish 570–577. and Wildlife Service, Washington D.C., Bolton, R.M. & Brooks, R.J. 2010. Impact of the USA. seasonal invasion of Phragmites australis Dahl, T. E. 2006. Status and Trends of Wetlands in (common reed) on turtle reproductive the Conterminous United States 1998 to 2004. success. Chelonian Conservation and Biology 9: U.S. Department of the Interior, Fish and 238–243. Wildlife Service, Washington D.C., USA Bonan, G.B. 2008. Forests and climate change: Dahl, T. E. 2011. Status and Trends of Wetlands in forcings, feedbacks, and the climate benefits the Conterminous United States 2004 to 2009. of forests. Science 13: 1444–1449. U.S. Department of the Interior, Fish and Bookhout, T.A., Bednarik, K.E. & Kroll, R.W. Wildlife Service, Washington D.C., USA. 1989. The Great Lakes marshes. In L.M. Dahl, T. E. & Johnson, C.E. 1991. Status and Trends Smith, R.L. Pederson & R.M. Kaminski of Wetlands in the Conterminous United States, (eds.), Habitat Management for Migrating mid-1970s to mid-1980s. U.S. Department of and Wintering Waterfowl in North America, the Interior, Fish and Wildlife Service, pp. 131–156. Texas Tech University Press, Washington D.C., USA. Lubbock, Texas, USA. Davis, M.A. 2009. Invasion Biology. Oxford Bradshaw, C.J.A., Warkentin, I.G. & Sodhi, N.S. University Press, Oxford, UK. 2009. Urgent preservation of boreal carbon Doherty, K.E., Ryba, A.J., Stemler C.L., Niemuth, stocks and biodiversity. Trends in Ecology and N.D. & Meeks, W.A. 2013. Conservation Evolution 24: 541–548. planning in an era of change: state of the U.S. Cariveau, A.B., Pavlacky, D.C., Bishop, A.A. Prairie Pothole Region. Wildlife Society Bulletin & LaGrange, T.G. 2011. Effects of 37: 546–563. surrounding land use on playa inundation Drake, J.A., Mooney, H.A., Di Castri, F., Groves, following intense rainfall. Wetlands 31: 65–73. R.H., Kruger, F.J., Rejmánek, M. & Coates, P.S., Casazza, M.L., Halstead, B.J. & Williamson, M. 1989. Biological Invasions: a Fleskes, J.P. 2012. Relative value of managed Global Perspective. Wiley, Chichester, UK. wetlands and tidal marshlands for wintering Federal Geographic Data Committee Wetlands northern pintails. Journal of Fish and Wildlife Subcommittee (FGDCWS). 2009. Wetlands Management 3: 98–109. Mapping Standard. FGDC Document Number Cowardin, L.M., Carter, V., Golet, F.C. & LaRoe, FGDC-STD-015-2009. U.S. Fish and E.T. 1979. Classification of Wetlands and Wildlife Service, Washington D.C., USA. Deepwater Habitats of the United States. Report Fleskes, J.P. 2012. Wetlands of the Central Valley FWS/OBS-79/31. U.S. Fish and Wildlife of California and Klamath Basin. In D. Service, Washington D.C., USA. Batzer & A. Baldwin (eds.), Wetland Habitats Cowardin, L.M., Shaffer T.L. & Arnold, P.M. of North America: Ecology and Conservation 1995. Evaluations of Duck Habitat and Concerns, pp. 357–370. University of Estimation of Duck Population Sizes with California Press, Berkeley, California, USA. a Remote-Sensing-Based Approach. Biological Frayer, W.E., Monahan, T.J., Bowden, D.C. & Science Report No. 2. U.S. Department of Graybill, F.A. 1983. Status and Trends of Wetlands the Interior, Washington D.C., USA. and Deepwater Habitats in the Conterminous

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 363

United States, 1950s to 1970s. Colorado State hydric soils on the Southern High Plains. University, Fort Collins, Colorado, USA. Wetlands 31: 483–492. Greenberg, D.A., & Green, D.M. 2013. Effects of Johnson, L.A., Haukos, D.A., Smith, L.M. & invasive plant on population dynamics in McMurry, S.T. 2012. Physical loss and toads. Conservation Biology 27: 1049–1057. modification of Southern Great Plains Hager, H.A. & McCoy, K.D. 1998. The playas. Journal of Environmental Management implications of accepting untested hypotheses: 112: 275–283. A review of the effects of purple loosestrife Johnson, W.C., Millett, B.V., Gilmanov, T., (Lythrum salicaria) in North America. Biodiversity Voldseth, R.A., Guntenspergen, G.R. & and Conservation 7: 1069–1079. Naugle, D.E. 2005. Vulnerability of northern Haukos, D.A. 2008. Analyses of Selected Midwinter prairie wetlands to climate change. Bioscience Waterfowl Data (1955–2008) in Region 2 (Central 25: 863–872. Flyway Portion). U.S. Fish and Wildlife Service, Johnson, W.C., Werner, B., Guntenspergen, G.R., Regional Migratory Bird Office, Albuquerque, Voldseth, R.A., Millett, B., Naugle, D.E., New Mexico, USA. Tulbure, M., Carroll, R.W.H., Tracy, J. & Haukos, D.A. & Smith, L.M. 1994. Importance Olawsky, C. 2010. Prairie wetland complexes of playa wetlands to biodiversity of the as landscape functional units in a changing Southern High Plains. Landscape and Urban climate. Bioscience 60: 128–140. Planning 28: 83–98. Johnston, C.A. 2013. Wetland losses due to row Haukos, D.A. & Smith, L.M. 2003. Past and crop expansion in the Dakota Prairie Pothole future impacts of wetland regulations on Region. Wetlands 33: 175–182. playas. Wetlands 23: 577–589. Kantrud, H.A. & Stewart, R.E. 1977. Use of Holling, C.S. 1973. Resilience and stability of natural basin wetlands by breeding waterfowl ecological systems. Annual Review of Ecology in North Dakota. Journal of Wildlife Management and Systematics 4: 1–23. 41: 243–253. Holt, J.S. 2009. Management of invasive terrestrial Kroll, R.W. 2006. Status of Winous Point marsh, plants. In M.N. Clout & P.A. Williams (eds.), fall 2006. Unpublished report to the Winous Invasive Species Management: A Handbook of Point Marsh Conservancy, Port Clinton, Principles and Techniques, pp. 126–139. Oxford Ohio, USA. University Press, Oxford, UK. Loesch, C.R., Reynolds, R.E. & Hanson, L.T. Johnson, L. 2011. Occurrence, function and 2012. An assessment of re-directing breeding conservation of playa wetlands: the key to waterfowl conservation relative to predictions biodiversity of the Southern Great Plains. of climate change. Journal of Fish and Wildlife Ph.D. thesis, Texas Tech University, Management 3: 1–22. Lubbock, Texas, USA. Luo, H.R., Smith, L.M., Allen, B.L. & Haukos, Johnson, W.P, Rice, M.B., Haukos, D.A. & D.A. 1997. Effects of sedimentation on playa Thorpe, P. 2011a. Factors influencing the wetland volume. Ecological Applications 7: occurrence of inundated playa wetlands 247–252. during winter on the Texas High Plains. Mattson, B.J., Runge, M.C., Devries, J.H., Boomer, Wetlands 31: 1287–1296. G.S., Eadie, J.M., Haukos, D.A., Fleskes, J. P., Johnson, L.A., Haukos, D.A., Smith, L.M. & Koons, D.N., Thogmartin, W.E. & Clark, R.J. McMurry, S.T. 2011b. Jurisdictional loss of 2012. A prototypical modeling framework for playa wetlands caused by reclassification of integrated harvest and habitat management of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 364 Wetland issues affecting waterfowl

North American waterfowl: case-study of Niemuth N.D., Reynolds, R.E., Granfors, D.A., northern pintail metapopulation dynamics. Johnson, R.R., Wangler, B. & Estey, M.E. Ecological Modeling 225: 146–158. 2008. Landscape-level planning for Meyer, S.W., Badzinski, S.S., Petrie, S.A. & conservation of wetland birds in the U.S. Ankney, C.D. 2010. Seasonal abundance and Prairie Pothole Region. In J.J. Millspaugh & species richness of birds in common reed F.R. Thompson (eds.), Models for Planning habitats in Lake Erie. Journal of Wildlife Wildlife Conservation in Large Landscapes, pp. Management 74: 1559–1567. 533–560. Elsevier Science, Burlington Meyerson, L.A., Chambers, R.M. & Vogt, Massachusetts, USA. K.A. 1999. The effects of Phragmites O’Connell, J.L., Johnson, L.A. Smith, L.M., removal on nutrient pools in a freshwater McMurry, S.T. & Haukos, D.A. 2012. tidal marsh ecosystem. Biological Invasions 1: Influence of land-use and conservation 129–136. programs on wetland plant communities of Mitsch, W.J. & Gosselink, J.G. 2007. Wetlands. 4th the semi-arid United States Great Plains. edition. John Wiley & Sons, Inc., New York, Biological Conservation 146: 108–115. USA. Olson, B.W. 2010. An experimental evaluation of Moon, J.A. & Haukos, D.A. 2006. Survival of cost effective moist-soil management in the female northern pintails wintering in the Sacramento Valley of California. M.Sc. Playa Lakes Region of northwestern Texas. thesis. University of California, Davis, Journal of Wildlife Management 70: 777–783. California, USA. Moon, J.A., Haukos, D.A. & Smith, L.M. 2007. Pew Environmental Group (PEG). 2011. A Changes in body condition of pintails Forest of Blue: Canada’s Boreal. Seattle, wintering in the Playa Lakes Region. Journal of Washington, USA. PEG unpublished report. Wildlife Management 71: 218–221. PEG, Seattle, Washington, USA. Accessible Michigan Department of Environmental Quality at http://borealscience.org/wp-content/ (MDEQ). 2008. A Guide to the Control and uploads/2012/06/report-forestofblue.pdf Management of Invasive Phragmites, 2nd edition. (last accessed 8 February 2014). Lansing, Michigan, USA. Poiani, K.A. & Johnson, W.C. 1991. Global Millar, J.B. 1989. Perspectives on the status of warming and prairie wetlands. Bioscience 41: Canadian prairie wetlands. Freshwater Wetlands 611–618. and Wildlife 61: 821–852. Poiani, K.A., Johnson, W.C. & Kittel, T.G.F. Millett, B.V., Johnson, W.C. & Guntenspergen, 1995. Sensitivity of a prairie wetland to G.R. 2009. Climate trends of the North increased temperature and seasonal American Prairie Pothole Region. Climatic precipitation changes. Water Resources Bulletin Change 93: 243–267. 31: 283–294. Mills, E.L., Leach, L.H., Carlton, J.T. & Secor, Poiani, K.A., Johnson, W.C., Swanson, G.A. C.L. 1994. Exotic species and the integrity of & Winter, T.C. 1996. Climate change and the Great Lakes. Bioscience 44: 666–676. northern prairie wetlands: simulations Nichols, J.D., Runge, M.C., Johnson, F.A. of long-term dynamics. Limnology and & Williams, B.K. 2007. Adaptive harvest Oceanography 41: 871–881. management of North American populations: Poiani, K.A. & Johnson, W.C. 2003. Simulation of a brief history and future prospects. Journal of hydrology and vegetation dynamics of prairie Ornithology 148: S343–S349. wetlands in the Cottonwood Lake Area. In

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 365

T.C. Winter (ed.), Hydrological, Chemical and wetland conservation in Canada. Technical Biological Characteristics of a Prairie Pothole Consultation on Designing Methodologies Wetland Complex under Highly Variable Climatic to Review Laws and Institutions Relevant to Conditions – the Cottonwood Lake Area, East- Wetlands. Environment Canada unpublished Central North Dakota, pp. 95–109 U.S. report. Environment Canada, Ottawa, Geological Survey Professional Paper No. Ontario, Canada. Accessible at http:// 1675. U.S. Geological Survey, Denver, www.ramsar.org/doc/wurc/wurchbk3cs2.doc Colorado, USA. Accessible at http:// (last accessed 8 April 2014). pubs.er.usgs.gov/publication/pp1675 (last Schummer, M.L., Palframan, J., McNaughton, E., accessed 4 August 2014). Barney, T. & Petrie, S.A. 2012. Comparisons Radosevich, S.R., Holt, J.S. & Ghersa, C.M. 2007. of bird, aquatic macroinvertebrate, and plant Ecology of Weeds and Invasive Plants: Relationship communities among dredged ponds and to Agriculture and Natural Resource Management. natural wetlands habitats at Long Point, Lake John Wiley and Sons, Inc. Hoboken, New Erie, Ontario. Wetlands 32: 945–953. Jersey, USA. Slattery, S.M., Morissette, J.L., Mack, G.G. Ray, J.D., Sullivan, B.D. & Miller, H.W. 2003. & Butterworth, E.W. 2011. Waterfowl Breeding ducks and their habitats in the High conservation planning: science needs and Plains of Texas. Southwestern Naturalist 48: approaches. In J.V. Wells (ed.), Boreal Birds of 241–248. North America: a Hemispheric View of their Reynolds, R. E., Shaffer, T.L., Loesch, C.R. & Conservation Links and Significance, pp. 23– Cox, R.R. 2006. The Farm Bill and duck 40. University of California Press, Berkeley, production in the Prairie Pothole Region: USA. increasing the benefits. Wildlife Society Bulletin Smerdon, B.D., Devito, K.J. & Mendoza, C.A. 34: 963–974. 2005. Interaction of ground water and Ringelman, J.K. (ed.) 2005. Prairie Pothole Joint shallow lakes on outwash sediments in the Venture 2005 implementation plan. U.S. sub-humid Boreal Plains of Canada. Journal of Fish and Wildlife Service, Denver, Colorado, Hydrology 314: 246–262. USA. Smith, A.G., Stoudt, J.H. & Gollop, J.B. 1964. Riordan, B., Verbyla, D. & McGuive, A.A. 2006. Prairie potholes and marshes. In J. Linduska Shrinking ponds in subarctic Alaska based on (ed.), Waterfowl Tomorrow, pp. 39–50. U.S. 1950–2002 remotely sensed images. Journal of Government Printing Office, Washington Geophysical Research 111: 1–11. D.C., USA. Roach, J., Griffith, B., Verbyla, D. & Jones, J. Smith, L.C., Sheng, Y., MacDonald, G.M. & 2011. Mechanisms influencing changes in Hinzman, L.D. 2005. Disappearing arctic lake area in Alaskan boreal forest. Global lakes. Science 308: 1429. Change Biology 17: 2576–2583. Smith, L.M. 2003. Playas of the Great Plains. Rooth, J.E., Stevenson, J.C. & Cornwell, J.C. University of Texas Press, Austin, USA. 2003. Increased sediment accretion rates Smith, L.M., Haukos, D.A., McMurry, S.T., following invasion by Phragmites australis: LaGrange, T. & Willis, D. 2011. Ecosystem the role of litter. Estuaries and Coasts 26: services provided by playa wetlands in the 475–483. High Plains: potential influences of USDA Rubec, C. & Lynch-Stewart, P. 1998. Regulatory conservation programs and practices. and non-regulatory approaches for Ecological Applications 21: S82–S92.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 366 Wetland issues affecting waterfowl

Smith, L.M., Haukos, D.A. & McMurry, S. 2012. Tiner, R.W. 2011. Dichotomous keys and mapping High Plains Playas. In D. Batzer & A. codes for wetland landscape position, landform, water Baldwin (eds.), Wetland Habitats of North flow path, and waterbody type descriptors: Version America: Ecology and Conservation Concerns, pp. 2.0. U.S. Fish and Wildlife Service, National 299–311. University of California Press, Wetlands Inventory Program, Northeast Berkeley, USA. Region, Hadley, Massachusetts, USA. Soja, A.J., Tchebakova, N.M., French, N.H.F., Treberg, M.A. & Husband, B.C. 1999. Flannigan, M.D., Shugart, H.H., Stocks, B.J., Relationship between the abundance of Sukhinin, A.I., Parfenova, E.I., Chappin, F.S., Lythrum salicaria (purple loosestrife) and plant III & Stackhouse, P.W., Jr. 2007. Climate- species richness along the Bar River, Canada. induced boreal forest change: Predictions Wetlands 19: 118–125. verses current observation. Global and Tsai, J.S., Venne, L.S., McMurry, S.T. & Smith, Planetary Change 56: 274–296. L.M. 2007. Influences of land use and Sorenson, L.G., Goldberg, R., Root, T.L. & wetland characteristics on water loss rates Anderson, M.G. 1998. Potential effects of and hydroperiods of playas in the Southern global warming on waterfowl populations High Plains, USA. Wetlands 27: 683–692. breeding in the northern Great Plains. van der Valk, A. G. & Davis, C.B. 1978. The role Climatic Change 40: 343–369. of seed banks in the vegetation dynamics Stephens, S.E., Walker, J.A., Blunck, D.R., of prairie glacial marshes. Ecology 59: 332– Jayaraman, A., Naugle, D.E., Ringelman, J.K. 335. & Smith, A.J. 2008. Predicting risk of habitat Vrtiska, M.P., Gammonley, J.H., Naylor, L.W., & conversion in native temperate grasslands. Raedeke, A.H. 2013. Economic and Conservation Biology 22: 1320–1330. conservation ramifications from the decline Stewart, R.E. & Kantrud, H.A. 1971. Classification of waterfowl hunters. Wildlife Society Bulletin of natural ponds and lakes in the glaciated prairie 37: 380–388. region. U.S. Fish and Wildlife Service Resource Washburn, N.B. 2012. Experimental evaluation Publication 92. U.S. Fish and Wildlife of tradeoffs in mosquito production and Service, Washington D.C., USA. waterfowl food production in moist-soil Stocker, T.F., Qin, D., Plattner, G.K., Tignor, habitats of California’s Central Valley. M.Sc. M., Allen, S.K., Boschung, J., Nauels, A., thesis. University of California, Davis, Xia, Y., Bex, V. & Midgley, P.M. (eds.) California, USA. 2013. Climate Change 2013: The Physical Science Weller, M.W. & Fredrickson, L.H. 1974. Avian Basis. Contribution of Working Group I ecology of a managed glacial marsh. Living to the Fifth Assessment Report of the Bird 12: 269–291. Intergovernmental Panel on Climate Change. Wells, J.V. 2011. Boreal forest threats and Cambridge University Press, Cambridge, conservation status. In J.V. Wells (ed.), Boreal UK. birds of North America: a Hemispheric View of Tiner, R.W. 2003. Correlating enhanced National their Conservation Links and Significance, pp. 1–6. Wetlands Inventory data with wetland functions for University of California Press, Berkeley, watershed assessments: a rationale for northeastern USA. U.S. wetlands. U.S. Fish and Wildlife Service, Wells, J.V. & Blancher, P.J. 2011. Global role for National Wetlands Inventory Program, sustaining bird populations. In J.V. Wells Region 5, Hadley, Massachusetts, USA. (ed.), Boreal birds of North America: a

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 Wetland issues affecting waterfowl 367

Hemispheric View of their Conservation Links and Report 1949–1867. U.S. Fish and Wildlife Significance, pp. 7–22. University of California Service, Washington D.C., USA. Press, Berkeley, USA. United States Fish and Wildlife Service Werner, B.A., Johnson, W.C. & Guntenspergen, (USFWS). 1988. Playa lakes region waterfowl G.R. 2013. Evidence for 20th century climate habitat concept plan, category 24 of the warming and wetland drying in the North North American Waterfowl Management American Prairie Pothole Region. Ecology and Plan. U.S. Fish and Wildlife Service, Evolution 3: 3471–3482. Albuquerque, New Mexico, USA. Windham, L., & Lathrop, R.G. 1999. Effects of United States Fish and Wildlife Service Phragmites australis (common reed) invasion (USFWS). 2013. Annual report of lands on aboveground biomass and soil properties under the control of the U.S. Fish and in brackish tidal marsh of the Mullica River, Wildlife Service. U.S. Fish and Wildlife New Jersey. Estuaries and Coasts 22: 927–935. Service, Division of Realty, Washington D.C., United States Climate Change Science Program USA. (CCSP). 2009. Thresholds of Climate Change Zedler, J.B. 2003. Wetlands at your service: in Ecosystems. U.S. Climate Science Program reducing impacts of agriculture at the and the Subcommittee on Global Change watershed scale. Frontiers in Ecology and the Research Report, U.S. Geological Survey, Environment 1: 65–72. Reston, Virginia, USA. Zedler, J.B. & Kercher S. 2004. Causes and United States Department of Agriculture consequences of invasive plants in wetlands: (USDA). 2012. Land values 2012 summary. Opportunities, opportunists and outcomes. National Agricultural Statistics Service Critical Reviews in Plant Sciences 23: 431–452.

Photograph: Mallard and Northern Pintail in a seasonal emergent wetland in the Illinois River Valley, Illinois, USA, by Heath Hagy.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 343–367 368

Opportunities and challenges to waterfowl habitat conservation on private land

WILLIAM L. HOHMAN1*, ERIC B. LINDSTROM2, BENJAMIN S. RASHFORD3 & JAMES H. DEVRIES4

1U.S. Department of Agriculture/Natural Resources Conservation Service, National Wildlife Team, Fort Worth, Texas 76115, USA. 2Ducks Unlimited, Inc., Bismarck, North Dakota 58503, USA. 3Department of Agricultural and Applied Economics, University of Wyoming, Laramie, Wyoming 82071, USA. 4Ducks Unlimited Canada, P.O. Box 1160, Stonewall, Manitoba R0C2Z0, Canada. *Correspondence author. E-mail: [email protected]

Abstract The future of North American waterfowl populations is inseparably tied to management of private land in the United States (U.S.) and Canada. Private land ownership in major waterfowl habitat regions such as the Northern Great Plains, Lower Mississippi Alluvial Valley, Gulf Coast and California’s Central Valley generally exceeds 90%, with agriculture being the dominant land-use in these regions. Planning and implementing avian conservation on private land in a strategic manner is complicated by a wide array of social, economic, political, administrative and scientific-technical issues. Prominent among these challenges are changing economic- drivers influencing land-use decisions, integration of bird conservation objectives at various scales, reconciling differences in wildlife habitat objectives between bird conservationists and land-users, administrative impediments to conservation planning and implementation, technology and scientific information gaps, and inadequate personnel capacity and financial constraints to effectively plan and deliver conservation. Given these unprecedented challenges to waterfowl habitat conservation, the need for effective public-private partnerships and collaboration has never been greater. With the goal of advancing collaborative waterfowl conservation on private land, the broad goals of this paper are to: (1) increase stakeholder awareness of opportunities and challenges to waterfowl habitat conservation on private land, and (2) showcase examples of collaborative efforts that have successfully addressed these challenges. To accomplish these goals this paper is organised into three sections: (1) importance of agricultural policy to private land conservation, (2) habitat potential on agricultural working land, and (3) strategic approaches to waterfowl habitat conservation. U.S. Department of Agriculture conservation programmes authorised through the Conservation Title of the 1985 Food Security

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 369

Act (hereafter, Farm Bill) and subsequent farm bills have provided unequalled potential for waterfowl habitat conservation on private land. Passage of the 2014 Farm Bill provides unique opportunities and alternative approaches to promote working land conservation strategies that are economically profitable and wildlife- friendly. However, reductions in private land conservation funding will require more effective targeting to maximise resource benefits. For example, in addition to conserving and restoring traditional habitats, we must work collaboratively to identify and promote working agricultural systems that are waterfowl-friendly and provide environmental services in addition to the production of food and fibre. Cultivation of rice Oryza sativa and winter cereals described below potentially represent two such situations. For over a quarter of a century the North American Waterfowl Management Plan (NAWMP) has served as a transformative model of partnership- based, landscape-scale conservation (DOI & EC 1986). Whereas the original plan and subsequent updates established abundant waterfowl populations as the plan’s ultimate goal, the 2012 NAWMP revision seeks a formal integration of these objectives with societal needs and desires (DOI et al. 2012). The current plan recognises the critical importance of private working land; however, details are lacking, especially with respect to strategic targeting of conservation on private land. For example, the development of truly strategic plans to target waterfowl conservation on private land will require estimates of the benefits of various conservation alternatives, conservation costs, and the threat of habitat loss or conversion. We suggest development of spatially explicit models that inform landowners and managers at the field-level about the cost effectiveness of conservation and land-use options is critically needed.

Key words: agriculture, conservation, economics, environmental services, Farm Bill, habitat, private land, waterfowl.

The future of North American waterfowl regions, with ~52% of the U.S. or 900 populations is inseparably tied to the million acres (365 million ha) managed as management of private land in the United cropland, pastureland or rangeland. Thus, States (U.S.) and Canada. Approximately the overwhelming majority of land-use 70% of the conterminous U.S. is held in decisions affecting waterfowl habitats are private ownership, including > 90% of the made by agricultural producers responding land area in major waterfowl habitat regions to a multitude of factors with various social such as the Northern Great Plains (NGP), and economic motivations. Lower Mississippi Alluvial Valley (LMAV), The contemporary setting in which Gulf Coast, Playa Lakes and California’s waterfowl managers operate is complex and Central Valley (Nickerson et al. 2011). continuously changing. Global factors Agriculture is the dominant land-use in these associated with an increasing human

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 370 Private land habitat conservation population and natural resource exploitation critical funding for important wildlife and development have far-ranging impacts habitat, soil and water conservation on land-use decisions, and ultimately on programmes (Heard et al. 2000). the availability and suitability of private land Amendments to the original Farm Bill in as waterfowl habitat. Competition in 1990, 1996, 2002, 2008 and 2014 have global commodity markets, water demands, retained and expanded conservation current federal agricultural/energy policy provisions such that there are now 13 and technological advancements in agricultural conservation programmes with agriculture are fuelling agricultural a combined funding level of $28.1 billion intensification and expansion (Sohl et al. for 2014–2018 (CBO 2014). No other 2012). Moreover, the United Nation’s state or federal programme provides a Food and Agriculture Organization (FAO) comparable level of investment or impact projects that world food production will for conservation initiatives on private land. need to increase by 70% by 2050 to meet The consideration of fish and wildlife food demands for an estimated 9.1 billion (hereafter, wildlife) in the delivery of humans (FAO 2009). The FAO anticipates conservation programmes was elevated in 80% of production increases will come the 1996 Farm Bill. Wildlife currently is from increased yield and 20% from an explicit goal for the Conservation expansion of arable land; however, declines Reserve Program (CRP), the Environmental in the rate of growth in yields of major Quality Incentives Program (EQIP), cereal crops from 1960 (3.2% per year) to and components of the Agricultural 2000 (1.5% per year) suggest that FAO Conservation Easement Program (ACEP). forecasts of production increases may be Because of federal budget constraints overly optimistic and additional land may and increased demand for agricultural need to be brought under cultivation. commodities, the national cap on CRP Since passage of the 1985 Food Security acreage has been reduced from a peak of 45 Act (hereafter, Farm Bill), U.S. Department million acres (18.21 million ha) in 1990 to of Agriculture (USDA) conservation 27.5 million acres (11.13 million ha) in 2014 programmes authorised through the and 24 million acres (9.71 million ha) in Conservation Title of the Farm Bill have 2017 (Cuzio et al. 2013; Ducks Unlimited, provided unequalled potential for waterfowl unpubl. data). Declines in CRP acreage were habitat conservation on private land. This only partially offset by increases in the size complex, multi-billion dollar legislation is of the wetland easement programme typically reauthorised by Congress every (formerly the Wetland Reserve Program, five years and covers a broad range WRP) from 2.125 to 3 million acres (0.86 to of programmes for commodities, crop 1.23 million ha) through 2012. The attention insurance, farm credit, nutrition, forestry, of wildlife conservation groups has been energy and conservation. Recognised as the focused on land retirement programmes in single largest private land conservation spite of the fact that farm bill’s working initiative in the U.S., the farm bill provides lands programmes, such as EQIP, the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 371

Agricultural Land Easement (formerly the In the face of unprecedented challenges Grassland Reserve Program, GRP), the to waterfowl habitat conservation, the need Conservation Stewardship Program (CSP), for effective public-private partnerships and a “working lands” CRP concept, could and collaboration has never been greater. receive greater funding and impact a far With the goal of advancing collaborative greater area. Consequently, the full potential waterfowl conservation on private land, the for improving consideration of waterfowl broad aims of this paper are to: (1) increase and other wildlife in land-use decisions has stakeholder awareness of opportunities and yet to be realised. challenges to waterfowl habitat conservation Changes in climatic conditions (e.g. severe on private land, and (2) provide examples alterations in regional temperature and of collaborative efforts that have been precipitation patterns), correlated to successful in addressing these challenges. increasing levels of CO2 and other To accomplish these aims we have greenhouse gases in the atmosphere, are organised the paper into three sections: already affecting the nation’s natural resources, (1) importance of agricultural policy to people, communities and economies that private land conservation, (2) habitat depend on healthy, functional ecosystems and potential on agricultural working land, and the plants and animals that characterise them (3) strategic approaches to waterfowl habitat (NFWPCAS 2012). Uncertainties regarding conservation. the effects of climate change on ecosystems and associated biota, and on current land Importance of agricultural uses, pose significant challenges to both policy to private land agricultural producers and waterfowl habitat conservation managers. Planning and implementing waterfowl European settlement of North America habitat on private land is complicated by a beginning in the eighteenth century wide array of social, economic, political, produced waves of change in land forms administrative and scientific/technical issues. and vegetation (hereafter, landcover). Prominent among these challenges are Suitability of land for agriculture greatly how changing economic drivers influence influenced settlement patterns in North land-use decisions, integration of bird America (Maizel et al. 1998). As expansion conservation objectives at various scales, rapidly proceeded westward during the reconciliation of differences in wildlife 1800s and early 1900s, farms were created at objectives between bird conservationists and the population frontier; areas too wet or too land-users, administrative impediments to dry were farmed later when drainage or conservation planning and implementation, irrigation was possible. Other areas with technology and scientific information gaps, poor climate, steep slopes, or soils and constraints on the personnel and unsuitable for use as cropland, grazed finances required to plan and deliver pasture or hay fields, were either farmed conservation effectively. unsuccessfully or never farmed.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 372 Private land habitat conservation

The influence of agriculture on pre- rate of wetland conversion between the settlement landcover is especially evident in mid-1950s and 1970s was estimated at the fertile Great Plains region of North 458,000 acres/yr (185,400 ha/yr; Frayer et al. America. The vast grasslands, shrublands 1983). Extensive wetland losses occurred in and savannas that characterise the region the LMAV as bottomland hardwoods were once represented the continent’s largest cleared and drained for cultivation of ecosystem; however, conversion of agricultural crops. The rate of wetland grasslands to agricultural uses has been losses slowed somewhat (to 290,000 extensive, exceeding 99% in portions of the acres/yr or 117,400 ha/yr) during the northern tallgrass prairie region of Iowa, decade before the Emergency Wetlands Minnesota, eastern Dakotas and Manitoba Resources Act of 1986 was enacted to (Samson & Knopf 1994; Noss et al. 1995). protect wetlands (Dahl 2000). The Act Associated with landcover change in the required the U.S. Fish and Wildlife Service Great Plains came a concomitant change in (USFWS) to monitor the status and trends communities of birds and other grassland- of wetlands and report details to the dependent wildlife. For example, dramatic Congress at 10-year intervals. declines in grassland bird species since the The vast majority of inland wetland 1950s have been attributed to changes in the losses were due to agricultural conversion agricultural landscape of the region (Gerard (Dahl 1990). The 1985 Farm Bill sought to 1995). Extensive loss and degradation of stem further wetland losses by linking grasslands in the Great Plains resulted in its wetland conservation on agricultural land to designation as one of the nation’s most the landowner’s eligibility for USDA farm endangered ecosystems (Noss et al. 1995). programme benefits, a provision commonly Wetlands in the Great Plains and other referred to as “Swampbuster”. Similarly, arable regions such as the LMAV and Highly Erodible Land (HEL) provisions California’s Central Valley have been commonly referred to as “conservation similarly affected. Dahl (1990) reported that compliance” and “Sodbuster” required between the 1780s and 1980s, the U.S. producers who cultivated sensitive land to (except Hawaii and Alaska) lost 53% have fully implemented a USDA-approved of its original wetlands. In Canada, an conservation plan by 1985. Provisions for estimated 40% of wetlands within the protection of highly erodible land and Prairie Pothole Region (PPR) have been lost wetlands were retained in revisions to the to drainage since settlement (Millar 1989). farm bills through to 2008. While these Twenty-two U.S. states have lost > 50% provisions did not create wildlife habitat of their wetlands, with California having directly, they did, “… provide strong the greatest wetland loss (> 90%, Dahl motivation for producers to apply 1990). Long-term trends show freshwater conservation systems on their highly emergent wetlands, especially forested erodible land, to protect wetlands from wetlands, sustained the greatest loss of any conversion to croplands, and apply for freshwater wetland type (Dahl 2000). The enrolment in other USDA conservation

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 373 programmes, especially the Conservation ha) of native grassland were converted to Reserve and Wetland Reserve Programs” croplands in the Missouri Coteau region of (Brady 2005:5). Implementation of these North and South Dakota during 1989–2003. provisions contributed to a reduction in soil Fuelled by demand for starch-based ethanol, erosion rates between 1981 and 2001 (Brady development of drought-resistant crops, 2005). Under Swampbuster, during an era of expiration of conservation contracts, and declining wetland losses (i.e. 506,000 acres increasing commodity prices, wetland and (205,000 ha) lost in 1992–1997 vs. 281,600 grassland conversion has accelerated (Wright acres (114,000 ha) lost in 1997–2002), gross & Wimberly 2013). High commodity prices wetland losses due to agriculture declined have made farmers less reliant on USDA from 26% during 1992–1997 to 18% commodity support programmes and during 1997–2002 (USDA NRCS 2000; effectively neutralised disincentives for 2013). Wetland restorations through other habitat conversion. Specifically, new risk conservation programmes, especially CRP management tools provided by federally- and WRP, resulted in net wetland gain on subsidised crop insurance, which protect agricultural land in both 1997–2002 and those farming marginally productive land 2002–2007, although the change during from economic losses, contradict other 2002–2007 was non-significant at the 95% policies aimed at conserving grasslands or confidence level (USDA NRCS 2013). protecting highly erodible land (Wright & The contributions of farm bill Wimberly 2013). The annual wetland loss programmes to waterfowl habitat rate in the PPR of North and South Dakota conservation have been substantial (Heard (2001–2011) was 0.35% or 15,377 ac/yr et al. 2000). In the PPR, Reynolds (2000) (6,223 ha/yr, Johnston 2013). The rate of estimated that between 1992 and 1997, the grassland conversion in the Western Corn CRP contributed to a 30% improvement in Belt of North Dakota, South Dakota, duck production or 10.5 million additional Nebraska, Minnesota and Iowa from 2006 to ducks. Grassland birds likewise benefitted 2011 ranged between 1 and 5.4% annually from the CRP in the NGP (Johnson 2000) with a nearly 1.31 million acres (530,000 ha) and Midwest (Ryan 2000), as did early net decline in grass-dominated land cover successional bird species (e.g. Northern (Wright & Wimberly 2013). Since the mid- Bobwhite Colinus virginianus) in the southeast 1980s, federal and provincial programmes in U.S. (Burger 2000). prairie Canada encouraged conversion of While farm bill provisions have helped marginal cropland to perennial grassland discourage grassland and wetland conversion (typically hay fields and pasture), and to cropland and provided incentives for removal of grain transportation subsidies the establishment of perennial cover on in the mid-1990s further encouraged highly erodible land, some producers have conversion to grass-based agriculture continued to convert native grasslands to (Riemer 2005). However, despite overall croplands. For example, Stephens et al. increases in grassland during the past 25 (2008) estimated that 90,300 acres (36,540 years, the absence of native grassland and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 374 Private land habitat conservation wetland protection policies in prairie Canada how to achieve cost savings and streamline have resulted in declines in native grassland programmes to reduce the federal deficit. and wetlands of 10% and 5%, respectively, Finally, on 7 February 2014, the President during 1985−2001 (Watmough & Schmoll signed into law a new farm bill called the 2007). Agricultural Act of 2014 (hereafter, 2014 The lack of effective disincentives for Farm Bill) that reauthorised several habitat conversion in current U.S. and important conservation programmes and Canadian agricultural policies, generous risk enacted other policy reforms aimed at management tools and ongoing conversion conserving critical grassland and wetland of grasslands and wetlands to croplands habitat on private land. In addition to the pose a significant threat to waterfowl challenges and delays of getting a new farm populations in the PPR. For example, 1.4 bill passed, substantial funding reductions million temporary and seasonal wetlands of were made to conservation programmes < 1 acre in size, located in crop fields in the estimated at ~$6 billion over the next 10 eastern Dakotas and northeast Montana, are years (CBO 2014). The 2014 Farm Bill also “at risk” of drainage without effective included major reforms to commodity Swampbuster protections (R.E. Reynolds programmes, new crop insurance options and C.R. Loesch, unpubl. data). Reynolds and consolidated conservation programmes. and Loesch (unpubl. data) further indicated Given the importance of farm bill that loss of these wetlands would reduce the programmes and agricultural policy to current breeding habitat capacity by about continental waterfowl populations, resource one-third for the five most common managers and conservation planners should breeding ducks in the region. Reversing be prepared to adapt, optimise and deliver trends in habitat loss in important waterfowl targeted conservation programmes much regions will be extremely challenging and if more efficiently with significantly less current rates of habitat conversion to federal financial resources from 2014–2018. croplands continue and habitat protection rates remain at current levels, regional Re-linking conservation compliance to crop habitat conservation goals and ultimately insurance waterfowl population goals will need to be For nearly 30 years, U.S. agricultural reduced (Doherty et al. 2013). producers have agreed to minimise impacts to HEL and Swampbuster-protected The 2014 Farm Bill: reforms, wetlands in exchange for farm programme challenges and opportunities benefits primarily offered through Title The one-year extension of the 2008 Farm I commodity (e.g. direct payments, Bill expired on September 30, 2013, countercyclical payments, etc.) and other resulting in a temporary lapse in funding farm credit supports. These “conservation for farm bill programmes. Passage of a compliance” provisions were first new farm bill was delayed over a year by established in the 1985 Farm Bill to help political gridlock, as Congress debated reduce adverse effects USDA programmes

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 375 were having on environmentally-sensitive on crop insurance enrolment. For the past land by reducing soil erosion on HEL and three decades, conservation compliance has slowing wetland conversion on agricultural been very effective at conserving farmed lands. Current law allows agricultural wetlands on private agricultural land (Brady producers to farm through wetlands during 2005). According to the USDA, up to 3.3 dry periods and still retain farm programme million acres (1.3 million ha) of vulnerable benefits provided they do not modify the wetlands within or adjacent to cropland hydrology of impacted wetlands, or if were not drained because of conservation modifications were undertaken after 23 compliance policies enacted since the 1985 December 1985 steps must be taken to Farm Bill (Claassen 2012). As USDA works mitigate for equivalent wetland functions to implement newly authorised farm bill and values. Conservation compliance programmes during 2014–2018, it will be provisions also disallow USDA loans or important to retain these effective payments to producers growing annually- conservation measures. tilled commodities on HEL without a The 2014 Farm Bill eliminates several soil conservation plan having first been Title I (e.g. direct payments and counter- approved by the Natural Resources cyclical payments) programmes tied to Conservation Service (NRCS). According to conservation compliance provisions in the USDA, ~100 million acres (40.5 million the farm bill. Many groups within the ha) or 25% of all cropland in the U.S. is conservation community advocated the need considered highly erodible (Claassen 2012). to reconnect these provisions to federal crop With assistance from USDA, producers insurance benefits (Title XI). In recent years, have developed conservation plans on over many producers have opted out of Title I 140 million acres (56.7 million ha) of benefits completely, thereby allowing them farmed land and reduced soil erosion on to convert wetlands for agriculture, while still HEL by nearly 40% or 295 million tons of receiving federal crop insurance benefits soil per year (Claassen 2005). without penalty. Since 1994, federal crop From 1985 to 1995, conservation insurance has evolved to become the most compliance requirements were also tied to important and highest-funded safety net and federal crop insurance benefits, but risk management tool for agricultural Congress decoupled these requirements producers, particularly in the NGP. Indeed, from crop insurance in the 1996 Farm Bill. estimated federal outlays for the crop A large increase in crop insurance enrolment insurance programme will total nearly from 99.7 to 202.6 million acres (40.3 to $90 billion over the next 10 years (CBO 82 million ha) occurred in 1994–1995 2014). Re-linking conservation compliance following the passage of the Federal Crop provisions to crop insurance premium Insurance Reform Act of 1994; however, subsidies would help ensure that farmers ~22 million fewer acres (8.9 million ha) were maintain a strong safety net, while ensuring insured in 1996, suggesting that decoupling long-standing protections for HEL and conservation compliance had little impact farmed wetlands remain in effect. After

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 376 Private land habitat conservation being decoupled from crop insurance since Sodsaver: slowing native prairie conversion to 1996, the 2014 Farm Bill reconnected croplands conservation compliance provisions for Temperate grasslands are one of the most farmed wetlands and HEL to federal crop imperilled ecosystems on the planet, yet insurance benefits. These provisions will maintain one of the lowest habitat provide critical protections for millions of protection rates of any major terrestrial farmed wetlands on agricultural land through biome (Hoekstra et al. 2005). Native 2018; however, USDA interpretation and grasslands that support diverse wildlife implementation of this new policy will be a populations and grass-based agriculture key factor in ensuring its effectiveness. are being converted to cropland at Unlike the U.S., Canada does not maintain unprecedented rates across many parts of similar federal wetland protection policies North America. During 2012, nearly 400,000 for wetlands on private land; consequently, acres (161,900 ha) of land with no prior current laws vary significantly among cropping history was converted to crop provinces and territories (Lynch-Stewart et production across the U.S., including al. 1993). In Canada, provinces have primary >54,876 acres (22,207 ha) in Nebraska, > jurisdiction over wetland protection policies 27,128 acres (10,978 ha) in South Dakota, within their boundaries, whereas the >26,395 acres (10,682 ha) in Texas and territories generally share authority among >24,961 acres (10,101 ha) in Florida (USDA federal, territorial and native agencies. FSA 2013). At current conversion rates, over However, Canada does maintain fairly half of the native prairie remaining in robust wetland protection policies on the U.S. areas of the PPR will be lost in the federal Crown land and Environment next 34 years (Stephens et al. 2008). Canada is the primary agency responsible Agricultural policies, emerging technologies for coordinating and implementing these and economic drivers are fuelling large-scale policies (Government of Canada 1991). conversion of these rare and important Generally, provincial laws cannot bind the prairie habitats. Native grasslands provide federal Crown, which creates regional critical habitat for wildlife, including a differences and geospatial challenges when globally-significant breeding range for many trying to implement and enforce wetland waterfowl and shorebird species (Ringelman protection policies on private land across et al. 2005). These habitats also support a broad landscape. Current provincial numerous grassland-dependent songbirds, wetland protection policies are being which are experiencing a steeper population developed and/or implemented in Alberta, decline than any other avian guild in Saskatchewan, Manitoba, Ontario, New North America (Peterjohn & Sauer Brunswick and Nova Scotia. However, 1999). Additionally, native rangelands are these regional and provincial disparities fundamentally important for livestock create a significant challenge to wetland production by providing forage and conservation for waterfowl on private land resilience to drought. Ranching, recreational in Canada. hunting and ecotourism associated with the

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 377 native prairie also provide economic diversity The 2014 Farm Bill provides a new and stability to rural economies. regional Sodsaver programme that applies Today, the last remaining grassland- to Montana, South Dakota, North Dakota, dominated landscapes are largely confined Minnesota, Iowa and Nebraska. This to areas with poor soils, steep topography provision applies to the entire state, not just and climatic conditions largely unsuitable the PPR-portion, and is a mandatory for consistent crop production (Doherty requirement, in contrast to the state et al. 2013). Unfortunately, accelerated Governor opt-in programme of the 2008 grassland conversion is occurring in many of Farm Bill. This provision will not these areas, causing significant ecological completely stop native prairie conversion in and societal impacts. Further loss of native these six states, but it will provide less rangeland habitat is also an economically financial incentive for converting native costly proposition, bringing additional prairie, as the crop insurance subsidies have disaster-prone land into production, while been reduced significantly. Grassland creating significant taxpayer liabilities conversion continues to be a national issue through subsidised risk management. that plagues many grassland-dependent Sodsaver legislation enacted in the 2014 species, such as Greater Sage-grouse Farm Bill, will: 1) limit crop insurance Centrocercus urophasianus, Lesser Prairie- coverage to 65 percent of the applicable chicken Tympanuchus pallidicinctus and many transition yield (i.e. county average) for the migratory birds that depend on these rare first four years until an actual production and declining habitats across the U.S. In history is established on newly broken land; 2013, 89% of the nearly 400,000 acres 2) reduce crop insurance subsidies on (161,900 ha) of perennial cover converted newly-broken sod by 50 percentage points to cropland occurred outside of the U.S. below the premium subsidy that would PPR (USDA FSA 2013). Thus, future farm otherwise apply for the first four bill policy efforts aimed at grassland consecutive years of crop production; and protection should focus on enacting 3) make newly-broken acreage ineligible for a national Sodsaver programme that yield substitution. These provisions were applies to all states and creates other similar included as a nationwide policy in the 2013 reforms that conserve critical native Farm Bill passed by the Senate, but were habitats. Additional policy reforms such as confined to only the U.S. PPR in the Farm significantly reducing or eliminating crop Bill passed by the House of Representatives. insurance subsidies on non-arable land (i.e. As illustrated by the 2008 Farm Bill, a soil classes 6–8) should also be considered. region-only Sodsaver provision is difficult to administer and can create inequities among The future of the Conservation Reserve Program agricultural producers within and across in a changing landscape states. Instead, a national provision would The CRP is considered one of the most create a more equitable and actuarially successful USDA conservation programmes sound programme across the country. in history and its landscape-level impacts on

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 378 Private land habitat conservation reducing soil erosion, improving water million acres (2.43 million ha) of CRP will quality, sequestering carbon and enhancing remain in the U.S. PPR in 2014. This loss wildlife habitat are well-documented (see represents a substantial decrease (31%) Allen & Vandever 2012). However, growing from its peak of 8.3 million acres (3.59 global demand for commodities, escalating million ha) in 2007 and declining trends are land and cash rent values, stagnant CRP expected to continue over the next 5 years rental rates, biofuel policies, and improved (Fig. 1; USDA FSA 2013). The 2014 Farm genetics and farming technologies are Bill reduces the national CRP enrolment cap driving the loss of CRP acreage across much from 27.5 million acres (12.9 million ha) to of the U.S. particularly in the PPR. For 24 million acres (9.71 million ha) by 2017. In example, Wright & Wimberly (2013) order to achieve cost savings, the national documented conversion of 1.3 million acres enrolment cap on the CRP Farmable (0.53 million ha) of perennial grasslands (i.e. Wetland Program (FWP) will also be native prairie, tame pasture and CRP) to reduced from 1 to 0.75 million acres cropland in 2006–2011, which represents a (404,690 to 303,500 ha). However, in issuing rate of change in grassland cover not seen guidance to USDA for new CRP rule- since the “Dust Bowl” era of the 1930s. The making, the Manager’s report states “overall Farm Service Agency estimates that < 6 reduction in the maximum acres enrolled …

9,000,000 South Dakota 8,000,000 North Dakota Montana 7,000,000 Minnesota 6,000,000 Iowa

5,000,000 Acres 4,000,000

3,000,000

2,000,000

1,000,000

0

2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Figure 1. Actual (2000–2013) and projected (2014–2018) Conservation Reserve Program (CRP) enrolment area in the U.S. Prairie Pothole Region. Projected area assumes no new sign-ups and anticipated expirations based on Farm Service Agency reports for 2014–2018 (Ducks Unlimited, unpubl. data).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 379 should not serve as an indicator of declining authorised under the 2008 Farm Bill into support for CRP. The Managers intend for just 13 programmes. For example, former CRP to be implemented at authorized levels, easement programmes such as the Farm and using the statutory flexibility, and for the Ranch Lands Protection Program, GRP and program to continue as one of USDA’s key WRP were merged into the ACEP. The conservation programs in concert with ACEP establishes two separate tracks for working lands conservation efforts.” wetland reserve easements (WRE) and Despite a significant reduction in the agricultural land easements (ALE), while overall CRP acreage cap, several provisions providing > $2 billion of funding for were included in the 2014 Farm Bill to make conservation on private land over the the programme more flexible and attractive 2014–2018 period. It also allows a to producers, while promoting a “working landowner donation for ALEs as long as lands” approach. For example, the Secretary another entity matches 50% of the of Agriculture will have greater authority to: Secretary’s contribution and provides a 1) enrol newly eligible grasslands (up to waiver to pay up to 75% USDA cost- 2 million acres, or 0.81 million ha); 2) share for certain grassland conservation flexibly apply prescribed grazing, burning, easements. This provision may create new haying and other mid-contract management public-private partnership opportunities activities outside of the primary nesting among state, federal, private and NGO season; 3) provide more allowances to use partners to develop easement programmes rather than dispose of residue removed from on private land. The new farm bill also CRP land during contract maintenance and reduces the former 7-year ownership rule to management; and 4) promote expanded use 2 years to become eligible for wetland of continuous and Conservation Reserve easement enrolment. This may be an Enhancement Practices (CREP) sign-up attractive incentive for conservation buyers opportunities. Faced with these challenges looking to enrol land into the programme. and opportunities, resource managers will The 2014 Farm Bill also consolidates need to focus on making CRP more several former regional conservation economically attractive and competitive programmes (e.g. the Chesapeake Bay by updating county rental rates, increasing Watershed Program and the Cooperative land-use flexibility and management Conservation Partnership Initiative) into a allowances, maximising continuous sign-up new Regional Conservation Partnership opportunities, and working to modify the Program (RCPP). Under RCPP, projects may national Environmental Benefit Index (EBI) focus on water quality, erosion, wildlife scoring process to elevate the PPR to a habitat and other regional resource concerns, national priority area. and this new programme will create up to eight national critical conservation areas. The Other working land opportunities RCCP partnership agreements may extend The 2014 Farm Bill also consolidates and up to 5 years and the programme provides streamlines 23 conservation programmes mandatory funding of $100 million per year

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 380 Private land habitat conservation from 2014–2018. It will facilitate landscape- Indeed, exponential growth in Lesser Snow scale conservation initiatives leverage Geese Chen caerulescens populations are partnerships and enable managers to direct attributed to behavioural and morphological resources strategically towards priority adjustments that enabled birds to shift from regions for waterfowl, such as the PPR or the historical to agricultural habitats (Linscombe Gulf Coast. 1972; Alisauskas 1998). There are numerous The 2014 Farm Bill also merges EQIP other examples of waterfowl using non- and former Wildlife Habitat Incentives traditional or altered habitats, although the Program (WHIP) into one general demographic consequences of these shifts EQIP programme, but specifies that are generally unknown. Thus, in addition to “wildlife habitat development” is a defined conserving and restoring traditional habitats, programme purpose and sets a minimum 5% we must identify and work collaboratively to funding floor for wildlife habitat projects. promote working agricultural systems that The programme also requires that at least are both producer- and waterfowl-friendly 60% of the total funds be invested for and provide environmental services in livestock purposes. The EQIP is one addition to the production of food and fibre. of the highest funded conservation Cultivation of rice Oryza sativa and winter programmes in the new farm bill, providing cereals represent two such situations. an average of $1.35 to $1.75 billion per year of conservation funding. The EQIP Agricultural working land and provides cost-share for a number of wildlife- waterfowl: rice agriculture example friendly conservation (wetland development, Rice agriculture is a major component of grassland improvement, etc.) and habitat the contemporary landscapes of the Gulf management practices (brush control, weed Coastal Plain, LMAV, and Central Valley of management, prescribed grazing, forage California. Between 1985 and 2012, 2.3–3.6 stand improvement, etc.) that may be very million acres (0.93–1.46 million ha) of rice compatible with waterfowl and economically were planted annually nationwide with over attractive to livestock producers, who prefer half (60%) of this acreage located in the more short-term working land options as LMAV (Arkansas, Mississippi, Louisiana, opposed to traditional 10–15 year set-aside and Missouri), 25% in the Gulf Coast region programmes such as the CRP. (Louisiana and Texas), and 15% in the Central Valley of California (USDA NASS Habitat potential on 2014). agricultural working land Cultivation practices To the extent that waterfowl are able to adapt Rice is a warm-season crop typically planted to habitat changes, or working agricultural in the spring and harvested in summer lands retain or simulate ecological functions or autumn. Cultivation practices vary provided by historical habitats, the adverse somewhat within and among rice-growing effects of habitat loss may be dampened. regions as a consequence of differences in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 381 climate, geography, soils, topography, to facilitate feeding by wintering waterfowl surrounding land-uses, water supply, on noxious Red Rice (Smith & Sullivan disease-pest issues, rotational cropping 1980). Water seeding is also preferred by opportunities and farming traditions. In farmers that plant extensive acreages in California and the LMAV, seeding of rice is areas with high rain and is compatible with similar to seeding practices for other cereal other uses of rice fields such as crawfish crops. That is, rice seed can be drilled or aquaculture. broadcast under dry to moist conditions in Cultivation practices in water- and dry- either reduced or conventional tillage seeded fields are similar after seeding. Fields systems (“dry seeding”). In southwest are gradually (re)flooded when rice has Louisiana, rice is most commonly cultivated sprouted 4–6 inches (10–15 cm) and remain using a water-seeding system in a 3-year flooded throughout the growing season rotation with crawfish (Order: Decapoda) until rice seeds mature. Most of the and fallow or soybeans (69% water-seeded, currently grown rice varieties need ~120 31% dry seeded; J. Saichuk, pers. comm.). In days from seed germination until the grain is a water-seeded system, rice is planted aerially ready for harvest. Fields are drained 4 weeks into flooded fields in March–June (Blanche before harvest to allow combine harvesters et al. 2009). Shortly before planting (3–4 to operate in the fields. days), the seedbed is tilled rough, fertilizer is An assortment of dryland crops are applied and incorporated, and the field is rotated with rice in California and the flooded. Alternatively, rough tillage LMAV, but rotational options are limited in conducted in autumn or winter may be coastal Louisiana and Texas. Along the Gulf followed by flooding and, shortly before Coast, rice typically is not cultivated in the seeding, water-levelling (tractor pulling a same field during consecutive years because blade through the flooded rice field). Water- doing so would increase disease and weed levelling agitates the soil and water, prevalence and reduce yields. Management producing a thick slurry and level seedbed options for rice producers include when the soil settles out of the water. Water- production of a second or “ratoon” rice seeded fields typically are dewatered 24 h crop, preparing fields for winter–spring after seeding. crawfish production, or idling land for The principal advantage of water seeding fallow or dryland crop production the is that it provides an excellent cultural following spring–summer. Ratooning is the method for control of weeds, especially Red practice of harvesting grain from tillers Rice Oryza punctate (Webster & Levy 2009). originating from the stubble of a previously Red Rice is the most troublesome and harvested crop (main crop). The climatic economically damaging competitor of rice; conditions of southwest Louisiana and the annually contributing to the loss of tens of early harvest date of commonly grown rice thousands of dollars to rice producers in varieties combine to create an opportunity southern states (Webster & Levy 2009). for ratoon crop production, but weather, Some producers flood harvested rice fields planting date, quality of the first crop and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 382 Private land habitat conservation harvest conditions can all influence ratoon fields are leased for waterfowl hunting rice development and yield. In general, the (November–January) or rainfall is first crop should be harvested by 15 August inadequate to completely flood fields by to ensure adequate time for ratoon rice to January. Fields are drained in spring so that develop. Harvest of ratoon rice typically they may be tilled in preparation for rice occurs in October–November. planting as described above. The rice-crawfish-fallow (or rice- crawfish-soybean) rotational strategy Waterbird use of rice commonly deployed in the region employs A wide variety of waterbirds (waterfowl, crawfish in a rotational system of rice and shorebirds and wading birds) and some sometimes soybeans. Rice is grown and landbirds use rice fields (Taft & Elphick harvested during the summer, and crawfish 2007). Rice field use by wintering and are grown during autumn, winter and early migrating waterfowl and shorebirds is spring in the same field. Louisiana crawfish especially pronounced. Avian use is best producers rely on a forage-based system for documented in Californian rice fields where providing nourishment to growing crawfish. over 118 species representing 38 families Rice has become the standard forage crop have been recorded during winter (Eadie et for the industry because the plant exhibits al. 2008). Densities of non-breeding the desired characteristics under the long- waterfowl and shorebirds observed in term flooded condition of a crawfish pond Californian rice fields averaged 730 (peak and partly because adequate stands of count = 3,600) and 252 (2,600) birds/km2, vegetation are achievable and predictable respectively (Eadie et al. 2008). when recommended management practices The 2.5–3.75 million acres (1–1.5 million are followed. ha) of farmland in coastal Louisiana and Rice fields managed for crawfish Texas operated in rice-crawfish-fallow, rice- production are commonly fertilised and fallow, rice-pasture or rice-dryland crop irrigated to achieve a ratoon crop (re- rotational scheme simulate wet, early growth) of forage. Fields are initially successional habitats that potentially are flooded in October–December and remain highly attractive to wetland-associated flooded throughout the harvest period, wildlife. The close proximity of fields to January–June. In southwest Louisiana, fields coastal marshes, their location at the are typically fallowed following drawdown in terminus of two major migratory bird May–June, but some producers may flyways, bird-friendly cultivation practices, drawdown crawfish ponds earlier (April) to high annual rainfall, and abundant plant and plant soybeans (April–June). To control animal foods further enhance their potential weeds in fields rotating back into rice value for waterbirds. Indeed, recent shifts in cultivation, water control structures typically the distributions of waterbirds from coastal are closed in the autumn (after soybean wetlands to inland agricultural wetlands (e.g. harvest) to capture available rainfall. Fleury & Sherry 1995) coincide with the Producers may pump water onto fields if expansion of crawfish aquaculture and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 383 ongoing loss and degradation of coastal manipulation of rice straw was confounded wetlands. A minimum of 67 species of by an interaction with the depth of flooding waterbirds including 17 waterfowl, 33 (Eadie et al. 2008). shorebirds, 15 wading birds, 2 rail and 1 In the Texas and Louisiana rice fields, crane species have been observed using rice waterbird species richness/diversity was fields in coastal Texas and Louisiana (W.L. highest in fallow fields and rice crop cover Hohman, unpubl. data). Peak densities of types, greatly exceeding other crop covers non-breeding geese, ducks, shorebirds, and (W.L. Hohman, unpubl. data). Fields in 3- wading birds recorded in these rice fields year rice-fallow rotation had higher richness during winters 1996/97 or 1997/98 were and diversity scores than fields in 3- or 4- 9,300, 4,300, 1,700, and 1100 birds/km2, year rice rotations with dryland crops. respectively (W.L. Hohman, unpubl. data). Richness was decreased by grazing and Estimated seasonal use by waterbirds increased by flooding. Duck densities were (excluding geese) from October to May was affected by crop rotation scheme, shorebird 72.1 and 125.5 million use-days in 1996/97 densities were affected by crop cover and and in 1997/98, respectively (W.L. Hohman, grazing, and wader densities were affected unpubl. data). However, because waterbirds by both crop cover and rotation scheme. use rotational crops (e.g. fallow), peak and Densities of all three groups increased with seasonal use may have been underestimated flooding. by ≥50% (W.L. Hohman, unpubl. data). Louisiana rice fields also provide habitat Use of rice fields by waterbirds is for breeding waterbirds (Hohman et al. potentially influenced by factors such as 1994), at least one of which (King Rail field size, timing, duration and extent of Rallus elegans) has been given special status in flooding, crop rotation, cultivation and 12 states. Increase in the nesting density of harvest practices, grazing, height of King Rails in Louisiana’s rice fields is vegetation, stubble treatments, frequency of attributed to expansion of crawfish disturbance and surrounding landscape aquaculture. Other common to rare nesting features (e.g. land uses, cover types, amount birds include Fulvous Whistling Duck of edge, distance to water, etc.). Waterbird Dendrocygna bicolor, Purple Gallinule richness and density are greater in flooded Porphyrula martinica, Common Moorhen than unflooded rice fields in California Gallinula chloropus and Least Bittern (Eadie et al. 2008). Waterbird groups Ixobrychus exilis. responded differently to water depth, with peak species richness and conservation Opportunities for management of rice fields for value (species being indexed by their relative waterfowl abundance in North America; Elphick & Waterbirds are attracted to rice fields Oring 1998) observed at intermediate water because of the abundant foods that occur depths (10–20 cm) (Elphick 1998; Eadie et there. Potential waterbird foods include al. 2008). In Californian rice fields, however, waste grain, seeds of water tolerant (i.e. interpretation of waterbird responses to moist soil) plants, green forage and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 384 Private land habitat conservation invertebrates. Rice fields are highly dynamic to waterbirds throughout much of the year. systems and, although vegetation is highly Additionally, many coastal rice fields are left monotypic, rice fields essentially function unplanted (e.g. pasture rotation) or fallowed like early successional, seasonally-flooded every other year. Moist soil plants that grow wetlands. That is, they have high detrital (i.e. in fallowed fields produce abundant seeds straw) inputs that, when flooded, serve as that are highly preferred foods of wintering forage for production of crawfish and other waterfowl (Fredrickson & Taylor 1982). aquatic invertebrates. Further, reduced Indeed, samples taken at waterfowl feeding pesticide use in fields managed for crawfish sites in Louisiana rice fields indicated production may benefit other aquatic biomass of moist soil plant seeds in rice invertebrates (McClain et al. 2009). fields may be equivalent to that found in Waterbirds also use rice fields as resting public areas managed specifically for that areas. The general openness of the rice purpose (Hohman et al. 1996). With average agricultural landscape is attractive to many rainfall, passive management (e.g. simply species that during migration and winter closing water control structures) is likely to must remain vigilant for potential predators be sufficient to meet the diverse habitat (Elphick 2000). Rice agriculture has become needs of most waterbird species; however, especially important for Northern Pintail the productivity and attractiveness of Gulf (Anas acuta, hereafter Pintail) wintering in Coast rice fields for waterfowl and other California’s Central Valley and along the waterbirds may be further enhanced by Texas–Louisiana Gulf Coast (Miller 1987; timely manipulations of rice stubble and Cox & Afton 1997). Pintail and other flooding, precise control of water levels waterbirds may shift to rice field refuges during rice cultivation, minimising to avoid disturbance in other habitats or, disturbances in fallow fields during March– alternatively, hunting disturbance may result May, or establishment of some single in daytime avoidance of rice fields (Rave & crop ponds managed solely for crawfish Cordes 1993; Cox & Afton 1996). production (W.L. Hohman, unpubl. data). Agronomic practices typically followed Following the 2010 Deepwater Horizon during the 3-year rice-crawfish-fallow or Gulf Oil Spill, the USDA NRCS established rice-fallow rotational schemes are generally the Migratory Bird Habitat Initiative “waterbird friendly.” So in most cases, (MBHI) to provide inland waterbird habitats management of Gulf Coast rice fields for to compensate for potential oil impacts on wintering and migrating waterbirds involves coastal wetlands. Through EQIP, WHIP, only minor changes in existing management and WRP, the MBHI has provided practices. Because of high annual rainfall, incentives for private landowners in eight use of flooding for weed control, practice of states (Alabama, Arkansas, Florida, Georgia, water-levelling, water-seeding of rice, Louisiana, Mississippi, Missouri and Texas) crawfish aquaculture and the leasing of rice to enhance and increase availability of fields for waterfowl hunting, Gulf Coast rice shallow-water habitats for migrating and fields tend to be wet and therefore available wintering waterfowl, shorebirds, and other

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 385 waterbirds along the Gulf Coast and within produced and subsequently disced were the LMAV. The provision of financial especially important habitat for non- assistance and compatibility of management breeding waterbirds, if they were flooded activities with normal agronomic practices through assistance from MBHI or other and recreational use of sites contributed to means (Marty 2013). the enthusiastic response by landowners who offered almost 1 million acres Challenges to management of rice fields for (> 400,000 ha) for possible enrolment in waterfowl EQIP or WHIP. To qualify for enrolment The potential for rice agriculture to provide the proposed management activity habitat for waterfowl is substantial on the must represent a change from normal Gulf Coastal Plain, as it is in other rice agronomic practices (i.e. “enhancement”). growing regions. Challenges to the Approximately half of the offers were management of rice fields as waterfowl accepted into the programme with most habitat identified by Eadie et al. (2008) of the contracts awarded in the rice include: 1) the provision of habitat for non- growing region of southwest Louisiana. In target, undesirable or nuisance wildlife (e.g. coastal Louisiana and Texas, the primary Red-winged Blackbird, Agelaius phoeniceus, management practices implemented American Coot Fulica americana; Snow through the MBHI entailed manipulations Geese, etc.); 2) water quality concerns (e.g. of rice stubble and shallow flooding of rice release of nutrients, particulate matter in fields in early autumn or late winter. Stubble water releases from rice fields); 3) additional manipulations and early flooding were time and financial costs associated with implemented to benefit autumn-migrating management (e.g. delayed field work shorebirds which pass through the region in and costs of pumping and stubble August and September; late flooding manipulations); 4) declining rice acreage targeted spring-migrating waterfowl. The due to urban growth, farm economics or net result was that shallow-water habitats human disturbance; 5) increased habitat were available in coastal regions for an fragmentation; 6) increased harvest extended duration. Activities undertaken efficiency or changes in agronomic practices through EQIP and WHIP on agricultural (e.g. straw management, development of working land were similar, but eligibility glyphosate-tolerant rice varieties) that differences between the programmes reduce the availability of waste grain and enabled the USDA NRCS to serve a broader moist soil plant seeds; 7) decreased clientele. availability of water (e.g. conflicts caused by An evaluation of waterbird responses increased demand and use by other user to MBHI practices by researchers at groups); and 8) conservation of endangered Mississippi State University is ongoing, species. but preliminary results further substantiate Additionally, agricultural policy that the importance of rice agriculture for favours production of other crops or waterbirds. Fields in which a ratoon was restricts farmer participation in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 386 Private land habitat conservation conservation programmes may contribute Hohman, unpubl. data). Further, the value to a reduction in rice acreage. For example, of rice fields as waterbird habitat exceeded Louisiana continues to fund MBHI through the return realised by farmers for EQIP, and MBHI was expanded to include production of rice and crawfish ($208/acre, activities designed to provide nesting and or $494/ha; W.L. Hohman, unpubl. data). brood-rearing habitat for resident Knowledge of the value that rice agriculture waterbirds such as the Mottled Duck Anas provides to ecosystem services, and efforts fulvigula. Although this species has adapted to minimise mismatches between the value to survive on the wet agricultural/coastal of these services and income derived from marsh interface, it is vulnerable to urban agricultural production, should further encroachment, coastal land loss and advance stewardship of rice agriculture for conversion from “wet” agriculture (such as the conservation of waterfowl and other rice and crawfish production) to dry land wildlife. crops such as soybean, sugarcane and milo Although waterbird use of Californian rice (Hohman et al. in press). Initial interest in fields is well documented, the extent to this component of MBHI was constrained which rice fields provide a reasonable by confusion about the level of substitute for natural wetlands is unclear compensation that was to be provided for (Elphick 2000; Eadie et al. 2008). The various management scenarios. Programme functional equivalence of rice agriculture in restrictions are also limiting expansion of comparison with historical wetland habitats MBHI. Specifically, the EQIP requirement along the Texas–Louisiana coast likewise is that only allows for provision of financial unknown; nonetheless, Louisiana and Texas and technical assistance for the application have experienced extensive loss and of a new practice or activity prevents degradation of coastal wetlands. From producers from re-enrolling fields in MBHI. 1932–2000, coastal Louisiana lost > 4,900 Consequently, the acreage enrolled in MBHI km2 of land, primarily marsh, with the annual has declined because producers are rate of wetland loss estimated to be 43 km2 unwilling to bear the increased costs of between 1985–2010 (Couvillion et al. 2011). management without compensation. Continued loss of coastal wetlands, and Management of rice fields for reductions in rice acreage in coastal Texas– recreational activity and income derived Louisiana, have important implications for from hunting leases can provide a strong waterbird conservation in North America. motivation for producers to manage rice Enhanced management of agricultural fields as waterfowl habitat. The value of rice wetlands along the Gulf Coast (e.g. as fields for waterbirds in southwest Louisiana undertaken in response to the Deepwater was estimated to be > $100–1,028/acre Horizon oil spill through the NRCS’s MBHI) ($247–2,538/ ha) based on the value of may represent the best opportunity to hunting leases or the restitution value of accommodate waterbirds displaced by waterbirds using rice fields during the wetland loss associated with sea-level rises breeding and non-breeding periods (W.L. and other environmental change.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 387

Agricultural working land and since the 1970s has been the decline in waterfowl: autumn cereals example summer fallow, a practice where cropland is left uncropped for alternate growing Opportunities for management of autumn cereals seasons for moisture accumulation, nitrogen for waterfowl release and weed control (Carlyle 1997). In The glaciated PPR region of central North prairie Canada, the practice of summer America serves as the primary breeding area fallowing has declined by ~18.8 million for many of North America’s waterfowl and acres (7.6 million ha) between 1971–2011 shorebirds (Batt et al. 1989; Skagen & (Statistics Canada 2012). In its place, Thompson 2007). Historically, extensive continuous cropping under minimum and native grasslands and diverse wetlands zero-tillage practices with high nutrient and provided ideal habitat for successful pesticide inputs has prevailed. Podruzny et waterfowl reproduction in this area al. (2002) suggested that declines in (Stephens et al. 2005). Since human populations of some bird species, such as settlement, however, a majority of the PPR Pintail, may have been the result of reduced has become an important agricultural nest survival as continuous cropping production zone for small-grain, oil-seed replaced relatively safe nest sites located in and row crops. Conversion of grassland to summer fallow. annual cropland, along with drainage and Cropland conversion to grassland began degradation of wetlands, has made in the U.S. under the CRP in the late 1980s, significant alterations to the landscapes in and in Canada with removal of grain which breeding waterfowl and shorebirds transportation subsidies in 1995. Recent nest (Stephens et al. 2008). Today, this region trends and long-term projections of of North America is one of the most cropland area suggest that conversion of intensively cropped landscapes in the world, grassland to cropland is again on the with > 80% of some counties in cropland rise (Rashford et al. 2010; Wright & production (Foley et al. 2005; Statistics Wimberly 2013). Biofuel-driven agricultural Canada 2011). commodity prices are expected to increase Conversion of grasslands to cropland pressure to convert grasslands to croplands and associated alteration of predator in the foreseeable future (Wright & communities in the PPR are thought to be Wimberly 2013). While waterfowl benefited the primary reason for long-term declines in greatly from programmes such as the waterfowl production in this region CRP (Reynolds et al. 2006), these benefits (Sargeant et al. 1993; Greenwood et al. 1995; are expected to diminish as remaining Beauchamp et al. 1996; Stephens et al. 2008). grasslands are converted to cropland In addition, the intensity of cropping (Stephens et al. 2008; Rashford et al. 2011). practices has increased on existing cultivated Not all croplands are equal, however, in land in recent history. The largest and their potential to affect breeding waterfowl. most economically and environmentally While many waterfowl species can benefit significant change in agricultural land-use from croplands as a food resource during

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 388 Private land habitat conservation non-breeding periods (reviewed in Taft & documented use by nesting birds (Goelitz Elphick 2007), few crops provide relatively 1918; Earl 1950; Milonski 1958; Higgins safe nesting habitat like the grasslands they 1977; Lokemoen & Beiser 1997). This is replace. Early nesting species, such as likely because most waterfowl nesting Mallard Anas platyrhynchos and Pintail, studies historically avoided searching seeded are especially susceptible to nest failure cropland, or limited timing and frequency in croplands, but some autumn-seeded of searches relative to other habitats due cereal crops including winter wheat Triticum to crop damage concerns. Hence, our sp. and autumn rye Secale cereale may understanding of cropland use by nesting provide viable nesting habitats (Devries et al. ducks is limited despite the dominance 2008). of cropland as potential nest habitat in In North America, wheat has two distinct many landscapes important to breeding growing seasons. Winter wheat, accounting waterfowl. Where data are available, nest for 70–80% of U.S. wheat production, is survival in cropland is typically low due to planted in the autumn, harvested the predation and destruction of nests by following summer and is generally grown machinery during spring-seeding operations from the Texas Gulf Coast to prairie Canada (Cowardin et al. 1985; Klett et al. 1988; (Acquaah 2005). Spring wheat is planted in Greenwood et al. 1995; Richkus 2002). early spring, harvested in late summer/early Autumn-seeded cereal grains, such as autumn and produced primarily in the winter wheat and autumn rye, however, can prairies of the northern U.S. and southern provide relatively undisturbed nesting cover Canada (Acquaah 2005). Within the PPR, the for birds during the breeding season and majority of wheat grown is the spring- may complement grassland nesting habitats seeded variety. For example, of 75 million that are available to birds. Several recent acres (30.4 million ha) of cropland in prairie studies suggest that autumn cereals may Canada in 2012, ~21 million acres (8.5 provide high value nesting habitat for million ha) were wheat, of which only ~1 breeding waterfowl relative to spring-seeded million acres (0.4 million ha) were winter crops and grasslands. Devries et al. (2008) wheat (Statistics Canada 2012). In North and conducted complete nest searches on 4,247 South Dakota, about 10 million acres (4.1 ha of cropland in southern Saskatchewan, million ha) out of 40 million cropland acres including spring-seeded (wheat and barley) (16.2 million ha) were wheat in 2012, of and autumn-seeded cereals (winter wheat which ~2 million acres (0.8 million ha) were and autumn rye). Autumn rye and spring- winter wheat. Other autumn-seeded cereal seeded crops were used for nesting by five grains like autumn rye and triticale (Triticum × duck species (Mallard, Pintail, Blue-winged Secale hybrid) generally comprise less than a Teal Anas discors, Northern Shoveler A. couple of hundred thousand acres in the clypeata, and Gadwall A. strepera), while PPR. winter wheat was used by all of the Croplands are commonly ignored in aforementioned species, as well as by Green- waterfowl nesting studies despite their winged Teal A. crecca, and Lesser Scaup

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 389

Aythya affinis). Nest densities were 0.39 and cropland stubble (Milonski 1958; Klett et al. 0.25 nests/ha in winter wheat and autumn 1988; Miller & Duncan 1999), initiate nests rye, respectively, compared to 0.03 nests/ha early in the season prior to spring-seeding in spring-seeded cereals. Critically, nest operations (Austin & Miller 1995) and re- survival was consistently very high nest minimally (Austin & Miller 1995; Guyn throughout the nesting season in winter & Clark 2000). Further, Pintail tend to settle wheat and autumn rye (~38% and 18%, in highly cropped landscapes, especially at respectively) whereas survival in spring- high population density (J.H. Devries, seeded crops varied from close to 0% in unpubl. data). Other priority bird species (e.g. early nests to close to winter wheat levels for Long-billed Curlew Numenius americanus) that late nests (Devries et al. 2008). High nest are known to nest early in cropland stubble survival has also been found in winter wheat are also likely to benefit from autumn cereals in comparable studies in North Dakota (Lokemoen & Beiser 1997; Devries et al. (Duebbert & Kantrud 1987; B.R. Skone, 2010). unpubl. data). Further, nest success rates in autumn-seeded crops are generally greater Challenges for management of autumn cereals for than those found in grassland habitats waterfowl throughout much of the PPR (e.g. Klett et al. Given the potential benefits of autumn- 1988; Greenwood et al. 1995). Additional seeded cereals to nesting waterfowl, Ducks research comparing waterfowl nest density Unlimited (DU), an international non-profit and success in winter wheat and adjacent organisation focused on conserving grassland habitat is currently ongoing (B.R. waterfowl habitat, has taken great interest in Skone, unpubl. data). winter wheat. Efforts by DU include Together, the density and success of promoting winter cereals in landscapes that waterfowl nests in autumn-seeded cereals have high wetland densities and attract high suggests that these crops have the potential densities of Pintail and other wetland- to recruit many more waterfowl young to dependent birds. As with most agricultural breeding populations than spring-seeded commodities, the primary drivers of winter cropland, and are comparable to grassland wheat production are agronomic and in this habitats (Devries et al. 2008). Providing high sense, winter wheat has several advantages. nest survival early in the nesting season Winter wheat on average provides a 20% conveys added value, given the importance yield advantage over spring wheat and of early hatched nests to waterfowl generally has lower input costs (Statistics recruitment (e.g. Dzus & Clark 1998). The Canada 2013). Further, indirect benefits value of autumn-seeded cereals may be most include: 1) spreading out the annual evident in landscapes with high breeding workload; 2) earlier seeding of spring- waterfowl populations, many wetlands and seeded crops; 3) decreased exposure to poor extensive croplands. Pintail, especially, could spring seeding weather; 4) winter wheat benefit from expansion of autumn-seeded takes full advantage of spring moisture; cereal crops, as they nest extensively in 5) early growth avoids exposure to

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 390 Private land habitat conservation certain pests; and 6) winter wheat often Cropscience. A focus of the partnership with outcompetes spring grassy weeds. Bayer, “Winter Cereals – Sustainability in Despite the agronomic benefits, barriers Action”, includes additional extension to the growth of winter wheat remain. First, outreach to increase the acreage of winter extremely cold winters, especially in prairie wheat planted in the PPR and expansion of Canada, challenge existing winter wheat winter wheat breeding programmes at several varieties with winter kill. Also in Canada, universities across the U.S. and Canada. Since where hard red spring wheat has been the DU initiated the winter cereals programme in “gold standard” of the grain market, 1999, winter wheat acreage in North Dakota challenges remain with developing markets has increased over 12-fold from 60,000 acres for alternate wheat varieties (Mulik & Koo (24,300 ha) in 1999 to 750,000 acres (303,600 2006). Finally, changing long-held traditional ha) planted in 2012 (USDA NASS 2013). farming practices remains an impediment, Over the same time period in prairie Canada, as seeding in September, shallow seeding, winter wheat has grown from 245,000 acres seeding into standing stubble and earlier (99,100 ha) to 1.1 million acres (459,500 ha, harvests, challenge farmers to make Statistics Canada 2012). substantial changes to their operations. The remaining barriers to expansion To address these challenges, DU of winter wheat can be overcome. New has embraced several non-traditional cold-tolerant varieties are in constant activities for a conservation organisation. development and additional varietal Recognising the limitations of available development is focused on yield, quality and winter wheat varieties, DU provided financial disease resistance. Markets for winter wheat, support for the development of new winter especially in Canada, are beginning to wheat varieties at a time when winter expand, and realised agronomic benefits wheat variety development in Canada was should overcome traditional barriers to concluding. Currently, > 90% of winter autumn-seeding. Finally, further research is wheat varieties grown in prairie Canada are being conducted to determine whether those developed with DU support. Further, winter wheat provides landscape-level DU is investing in collaborative research to impacts on duck and shorebird nest survival improve cold-hardiness of winter wheat in addition to the apparent habitat-specific varieties while providing direct technical increase in nest survival for nests within assistance to farmers regarding best crop winter wheat fields (B. Skone, Montana State management practices. While incentive University, pers. comm.). payments were initially part of the programme, evaluations have shown that The challenges of strategic technical expertise provided by agronomists conservation targeting on was more attractive and sustainable than private land cash incentives (DU, unpubl. data). Ducks Unlimited has recently expanded their winter The need to target conservation strategically wheat programme in partnership with Bayer has long been recognised by policy makers,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 391 ecologists and economists. Faced with landscape-scale conservation delivery. The limited resources (e.g. budgets or labour- original plan and subsequent updates in hours), conservationists cannot select all 1994, 1998 and 2004 established abundant available projects that produce biological waterfowl populations as the plan’s ultimate benefits (e.g. easement locations or goal. A science-based understanding of management activities); thus, they seek to waterfowl habitat requirements throughout select projects that generate the greatest their annual cycle and population responses biological benefits possible given resource to habitat, enabled managers to step-down constraints. What constitutes strategic continental population objectives to targeting and how to achieve it, however, important waterfowl regions; regional can differ substantially across different partnerships between public and private interest groups and academic disciplines. parties (Joint Ventures) were formed to Moreover, the challenges specific to implement management and assess targeting waterfowl conservation on private progress towards the achievement of land depend on the definition of strategic objectives. As in the SHC framework targeting. These challenges become clear if described above, information gathered we begin from the strategic habitat during monitoring efforts was used conservation (SHC) framework developed to improve population-habitat models by the USFWS (USFWS 2008). and provide a sound science-base for The SHC framework describes strategic management actions. targeting as an iterative process involving The NAWMP was substantially revised in biological planning, conservation design 2012 to reflect, “… the rising challenges and delivery, and monitoring and research presented by a changing climate, social that provides feedback to inform the changes, the effects on land-use decisions of process. Essential to the SHC framework global economic pressures, and fiscal are: 1) defining and measuring specific restraint faced by agencies …” (DOI et al. population objectives (i.e. as opposed to 2012:iv). Specifically, the 2012 NAWMP simply focusing on habitat area protected, seeks to formally integrate objectives which are inputs to species-specific for waterfowl populations, habitat objectives); 2) using the best scientific conservation, and societal needs and desires. information, including population-habitat The central thesis of the revised plan is models and decision support tools, to that “… conservation goals can only be inform and update iteratively the SHC achieved with broad public support and by process; and 3) developing partnerships influencing land-use decisions over to design and deliver conservation extensive areas of the continent” (DOI et al. programmes. 2012:12). The 2012 NAWMP recognises For over a quarter of a century the North that most of these areas are privately owned American Waterfowl Management Plan “working lands” noting that, “While (NAWMP; DOI & EC 1986) has served as a some conservation outcomes are achieved transformative model of partnership-based, through regulations and policies, others

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 392 Private land habitat conservation result from collaborations that lead to changes in species-specific populations), can voluntary actions. Support from the lead to conservation plans that are highly public and participation by landowners inefficient (i.e. waste scarce resources; Duke hinges on striking the right balance et al. 2013). between conservation outcomes and the There is a general consensus in the socioeconomic drivers that influence land- economic and ecological literature that three use decisions. That balance is always primary factors affect efficiency of shifting, depending on the relative value conservation delivery: biological benefits, placed on conservation versus other conservation cost and threat of habitat loss drivers.” or conversion (see Newburn et al. 2005 for a Despite general similarities between the review of alternative targeting strategies). NAWMP and SHC frameworks, many Biological benefits, measured in physical details are lacking, especially with respect units or dollars, refer to the outcomes of to strategic targeting of conservation conservation. Although some studies focus on private land. The SHC framework on targeting biological benefits exclusively is “strategic” only in the sense that (e.g. Niemuth et al. 2009), the broader conservation activities are based on specific literature has consistently demonstrated objectives, scientific planning and design, that benefits must be weighed against and regular evaluation. In areas dominated conservation costs to generate efficient by private land, such as the U.S. and conservation plans (Naidoo & Iwamura Canadian prairies, private landowner 2007; Duke et al. 2013). Plans that maximise incentives, land-use change and agricultural benefits only (e.g. by selecting sites for policy significantly complicate the design protection that have the greatest biological and delivery of waterfowl conservation. In value) often lead to inefficient conservation addition, the SHC framework does not outcomes because limited budgets are provide sufficient guidance on how to quickly exhausted on high-benefit, high-cost evaluate conservation success. Metrics to projects (Duke et al. 2013). Incorporating evaluate conservation success are especially costs, using a cost-benefit or return on lacking on private land where conservation investment criterion, tends to increase delivery can be orders of magnitude more conservation efficiency by maximising the expensive than conservation on public land, conservation benefit per dollar expended. spatially targeting conservation is limited by More recently, the literature focused on each landowner’s willingness to accept conservation targeting established the conservation, and many conservation important role that threat plays in designing activities can focus as much on agricultural efficient conservation plans (Merenlender et activities as on ecological activities (e.g. al. 2009). Threat refers to the risk that working land conservation; Lewis et al. biological benefits will be lost in the absence 2011). Measuring success, and thus targeting of conservation. In the case of waterfowl conservation, in terms of biological benefits nesting habitat, for example, threat could as implied by the SHC framework (e.g. refer to the probability that dense grassland

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 393 cover is converted to intensive cropland. require estimates of benefits for various Ignoring threats can also result in conservation alternatives (e.g. changes in conservation inefficiencies because limited recruitment), conservation costs and resources are targeted to areas likely to measures of threat levels. Much of this produce benefits even in the absence of information already exists, particularly for explicit conservation. Incorporating threat, the waterfowl breeding grounds of North along with benefits and costs, therefore America. Population-habitat models exist improves efficiency by targeting limited to predict waterfowl distributions and resources towards projects that generate the response to landscape-level conservation greatest avoided loss per dollar expended (Cowardin et al. 1995; Reynolds et al. 1996; (Newburn et al. 2005; Murdoch et al. 2007; Johnson et al. 2010). Conservation costs can Withey et al. 2012). also be estimated from existing data (e.g. Despite the large and growing strategic published cropland rental rates to proxy conservation targeting knowledge base, for the cost of conservation easements); there has been relatively little research however, significant heterogeneity in costs targeted specifically to the design of across space and information asymmetry efficient waterfowl habitat conservation (i.e. landowner’s private costs are not plans. Several studies reported the costs and observable) imply that estimating benefits of specific management treatments conservation costs at the landscape level (see Williams et al. 1999 for a review) and could be complex. Moreover, cost tends to others explored the cost-effectiveness of be highly correlated with threat. For waterfowl management in hypothetical example, locations that have a high settings (e.g. Rashford & Adams 2007). probability of converting from grassland to Several waterfowl studies have considered intensive cropland will have a high landscape-level conservation targeting but opportunity cost of remaining in grassland, have focused on conservation benefits only and thus high conservation costs. (e.g. Reynolds et al. 2006; Niemuth et al. 2009; Estimating threat or risk of grassland or Johnson et al. 2010), have considered wetland conversion can also be challenging benefits and costs but not threats (Loesch et because conversion in the region is a largely al. 2012), or have considered threats and private decision influenced by economic, benefits but not cost (Stephens et al. 2008). physical and social factors that are not Rashford et al. (2011) demonstrated the completely observable. Agricultural policy cost-benefit-threat tradeoffs associated with reform, global commodity markets and targeting grassland conservation in prairie stochastic weather patterns are difficult to Canada, but their application was for a quantify or predict and can also contribute hypothetical and unrealistic conservation to uncertainty related to threats. Since these scheme (i.e. a fixed payment to all grassland). factors are highly heterogeneous across Given the consensus in the literature, space (e.g. soil quality varies considerably developing truly strategic plans to target across the prairies), conversion risk is likely waterfowl conservation on private land will to be highly heterogeneous (see below).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 394 Private land habitat conservation

Given spatially heterogeneous conversion broad headings of biophysical, economic risk, conservation targeting that considers and policy drivers. only benefits and cost will be inefficient (Newburn et al. 2005). Additionally, Bio-physical drivers effectiveness of voluntary conservation The biophysical attributes of land, such as programmes, such as CRP and WRP, are soil quality, hydrology and slope, directly and also influenced by threat, e.g. land with high indirectly affect private land-use. In some conversion threat and thus high opportunity cases, biophysical attributes may restrict the costs of entering conservation programmes set of land uses that are physically possible is less likely to be enrolled given a fixed (e.g. land too steep to be tilled). Biophysical payment level (Lewis et al. 2011). It is attributes also affect yields that can be therefore crucial to understand factors realised from the land, and thus, the affecting private land-use decisions, and economic returns private landowners can thus habitat conversion risk. derive from alternative land uses. As a result, land with characteristics that are suited to Private land-use decisions and habitat crop production tends to be placed in crop conversion risk production. Studies have found strong Studies in ecology, economics and positive correlations between high soil geography have a long history of modelling quality and grassland conversion (Wright & private land-use change and its drivers (see Wimberly 2013). In addition, PPR grassland e.g. Verburg et al. 2004). Although theories habitats in land capability Class 1 and 2 (i.e. and approaches differ across disciplines (e.g. best soils for agricultural production) were geographers and ecologists tend to focus on found to be 30% to 200% more likely to be social drivers at the macro-scale, whereas converted to cropland than grassland of economists tend to focus on private drivers lower soil capability (Rashford et al. 2010). at the micro-scale), there is a general Likewise, hydric soils, when drained, may consensus that economic, bio-physical and provide productive farmland, and removal social/policy factors drive land-use change. of in-field wetlands can improve the A relatively recent and growing body of efficiency of tillage operations by removing literature which has specifically examined “obstacles” to farm machinery. agricultural land-use change, both in the Climatic conditions have also been found PPR and in the NGP, suggests several key to strongly influence land-use decisions. drivers of land-use change and the risk of Temperature, precipitation and CO2 habitat conversion (Stephens et al. 2008; concentrations affect yield and yield Rashford et al. 2010; Gutzwiller & Flather variability, and thus the economic returns 2011; Rashford et al. 2011; Sohl et al. 2012; and risk associated with alternative land uses Feng et al. 2013; Wright & Wimberly 2013; (Adams et al. 1990). Studies in the PPR and Attavanich et al. 2014). Although the NGP have generally found strong positive identified drivers are not mutually exclusive, correlations between climate change and we categorise and describe them under the grassland habitat conversion (Sohl et al.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 395

2012; Attavanich et al. 2014). Warmer and change impacts on waterfowl in the PPR wetter conditions in the PPR are predicted and reported that ignoring land-use to increase wheat production at the expense response would underestimate the effects of of significant pastureland (Fig. 2; B.S. climate change on waterfowl by as much as Rashford, unpubl. data). Potential effects 10% (300,000 breeding pairs). of climate change, however, are highly heterogeneous across space and are Economic drivers (prices) moderated by other drivers (e.g. soil quality). Economic theory assumes that landowners Research in the NGP, where predicted allocate land to the use that generates the climate changes and soil quality are highly highest discounted stream of returns heterogeneous, indicates that grasslands in (Rashford et al. 2010). Hence, any factors the central Dakotas will be at increasing risk that affect current or future returns will of conversion, while grassland in the drive land-use decisions. Many studies in the western NGP will remain relatively secure NGP and PPR concluded that agricultural or increase (Fig. 3; B.S. Rashford, unpubl. prices or their derivatives (e.g. land rental data). Such climate-induced land-use rates) are important drivers of land-use and changes can exacerbate the effect of climate habitat conversion (Stephens et al. 2008; change on waterfowl and must therefore be Rashford et al. 2010; Rashford et al. 2011; considered when targeting conservation to Feng et al. 2013). For example, recent mitigate climate change. Attavanich et al. research in the NGP indicated that, holding (2014), predicted climate and land-use all else constant, a 10% increase in the

4,000 +2°C +4°C +4°C & +10% Precipitation 3,000

2,000

1,000

0

Hectares (× 1000 ) –1,000

–2,000

–3,000

Corn Hay Wheat Pasture Soybean Sorghum Figure 2. Predicted change in area (1,000s ha) by land use in the North and South Dakota portion of the Prairie Pothole Region for three future climate scenarios (+2°C, +4°C, and +4°C with +10% precipitation) (B.S. Rashford, unpubl. data).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 396 Private land habitat conservation

ND

MT

SD

WY

NE

–14,000 ha Predicted change in grassland 115,000 ha

Figure 3. Predicted change in expected grassland area (ha) in 2030, based on the Intergovernmental Panel on Climate Change’s A2 scenario (B.S. Rashford, unpubl. data; IPCC 2000). ND = North Dakota, SD = South Dakota, NE = Nebraska, WY = Wyoming and MT = Montana. returns to cropland would induce ~22,000 in crop insurance subsidies have been ha of grassland to convert to cropland (B.S. correlated with increases in crop acreage and Rashford, unpubl. data). decreases in CRP enrolment (Feng et al. 2013). Recent research in the NGP suggests Policy drivers (government payments, that the probability of a field being used for crop insurance, conservation crop production would be as much as 30% payments) lower if there were no direct government As discussed above, the U.S. farm bill payments to agricultural producers, which provides a number of programmes that would imply ~5.4 million additional acres provide incentives for private landowners to (2.2 million ha) of grassland (B.S. Rashford, choose certain land uses or production unpubl. data). practices. Subsidised crop insurance can reduce the financial risk of growing crops on Future directions for strategically lower quality soils or in areas with less than targeting conservation on private land suitable climates (i.e. areas that would be Strategically targeting conservation in a more likely to remain in native covers in manner that accounts for economic efficiency, the absence of insurance). Thus, increases private land-use incentives, and threat of loss

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 397 may require a slight reconsideration of the response given the current distribution of current SHC framework. First, conservation land-use; however, if probability of land-use agencies will need to consider alternative change were accounted for, the population measures of efficiency when evaluating response may be substantially different. conservation accomplishments and updating Additionally, targeting conservation on conservation targeting strategies. Simply private land effectively may require wholly evaluating programme outcomes in terms of different decision support tools than the biological benefits will not provide the typical tools that focus on population- comprehensive information necessary to habitat relationships. For instance, the use decide which programmes should be applied of models that complement traditional where. Given limited budgets, the cost of population-habitat models, by informing achieving outcomes across space must be landowners about how different considered to target conservation cost- conservation alternatives (e.g. working-land effectively. Incorporating costs can imply conservation or farm bill programmes) can shifts in conservation focus that are counter be economically compatible with (or to biological targeting, such as focusing beneficial to) their agricultural production. resources in regions where the incremental A site-specific decision support tool, biological benefits are relatively low but the demonstrating the economic and ecological benefits per dollar are high due to low tradeoffs between alternative crop rotations, conservation costs. Similarly, the identification could thus be used as a conservation of priority areas in the SHC framework could delivery mechanism by leading to increased be informed by incorporating costs and adoption of winter wheat. conversion threat. For example, prioritising Lastly, strategically targeting waterfowl habitat protection for breeding waterfowl conservation on private land will require based on pair densities may overlook the fact recognition that many effective conservation that the costs of achieving population “activities” may have little resemblance objectives could be reduced by focusing in to more traditional biological activities. regions with lower pair densities but relatively Forming a partnership, itself an emphasis of less conversion threat (and therefore lower the SHC framework, to lobby politically for conservation costs). “waterfowl friendly” agricultural policies (e.g. The decision-support tools critical to the conservation compliance) or to invest in SHC framework may also need to be agricultural research (e.g. new winter wheat expanded to make it effective for targeting varieties) may prove as effective as more and delivering conservation on private land. traditional direct habitat management. Models of the relationship between habitat Although many non-profit conservation and populations may misinform the SHC groups currently use such activities process if the effects of private land-use as highlighted in previous sections, decisions are not considered. For example, incorporating such activities explicitly within targeting easements in a particular region the SHC framework would focus the may appear to generate a large population process. Additionally, designing conservation

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 398 Private land habitat conservation activities for private land should be D’Abramo, J.B. Davis, S.C. Grado, R.M. incorporated specifically (to improve Kaminski, J.R. Marty, M.D. McConnell, W. leverage) in farm bill conservation Schilling and G. Wang; Point Blue Conservation programmes. Farm bill programmes are not Science (formerly Point Reyes Bird particularly well targeted across space, for Observatory Conservation Science): C. example, because they largely depend on Hickey, M. Reiter, K. Sesser, D. Skalos and voluntary participation and opportunistic K. Strum; USDA NRCS: S.J. Brady and R. enrolment. Consequently, it is difficult to Weber; and USFWS: H.D. Azure, K.E. control the spatial allocation of such Doherty, H. Hoistad, C.R. Loesch, W.A. programmes, because of inherent inability to Meeks, N.D. Niemuth, R.E. Reynolds, A.J. control, or even easily predict, which Ryba, C.L. Stemler, and J. Tirpak. Skillful landowners will choose to participate. A editing of the manuscript was provided by broadened SHC framework that considers L. Webb and E. Rees. existing voluntary enrolment, however, could be used to target other conservation References activities to leverage benefits of farm bill Acquaah, G. 2005. Principles of Crop Production: programmes, for instance by targeting Theory, Techniques, and Technology. Pearson easements near existing CRP to create larger Prentice Hall, Upper Saddle River, New Jersey, blocks of waterfowl nesting cover. USA. Adams, R.M., Rosenzweig, C., Ritchie, J., Peart, Acknowledgements P., Glyer, J.D., McCarl, B.A., Curry, B. & This paper is the product of a special paper Jones, J. 1990. Global Climate Change and Agriculture. Nature 345: 219–224. session “Opportunities and Challenges Alisauskas, R.T. 1998. Winter range expansion Facing Waterfowl Habitat Conservation on and relationships between landscape and Private Lands” (organisers: W.L. Hohman morphometrics of midcontinent Lesser and E. Lindstrom) conducted at the Snow Geese. Auk 115: 851–862. Ecology and Conservation of North Allen, A.W. & Vandever, M.W. 2012. Conservation American Waterfowl Conference held in Reserve Program (CRP) Contributions to Wildlife Memphis, TN, 27–31 January 2013. We are Habitat, Management Issues, Challenges and Policy especially grateful to the Conference Co- Choices – an Annotated Bibliography. U.S. chairs (R.M. Kaminski and J.B Davis, Geological Survey Scientific Investigations Mississippi State University) and Scientific Report 2012–5066. U.S. Geological Survey, Committee for the opportunity to Washington D.C., USA. participate in the meeting. Contributors to Attanavich, W., Rashford, B.S., Adams, R.M. & McCarl, B.A. 2014. Land Use, Climate our 2-day special session were Audubon Change and Ecosystem Services. In J.M. Duke California: M. Iglecia and R. Kelsey; Ducks and J. Wu (eds.), Oxford Handbook of Land Unlimited: M.G. Brasher, J.D. James, M.R. Economics, pp. 255–280. Oxford University Kaminski, M.T. Merendino, M.J. Petrie and Press, UK. A.J. Wiseman; Mississippi State University: A.B. Austin, J.E. & Miller, M.R. 1995. Northern Pintail Alford, J.L. Avery, L.W. Burger, L.R. (Anas acuta). In A. Poole & F. Gill (eds.). The

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 399

Birds of North America, No. 163. The Birds of OECD (ed.), Evaluating Agri-Environmental North America, Inc., Philadelphia, PA, USA. Policies: Design, Practice and Results, pp. Batt, B.D.J., Anderson, M.G., Anderson, C.D. & 293–294. Organization for Economic Co- Caswell, F.D. 1989. The use of Prairie Potholes operation and Development (OECD), Paris, by North American ducks. In A. van der Valk France. (ed.), Northern Prairie Wetlands, pp. 204–227. Claassen, R. 2012. The Future of Environmental Iowa State University Press, Ames, USA. Compliance Incentives in U.S. Agriculture: The Role Beauchamp, W.D., Koford, R.R., Nudds, T.D., of Commodity, Conservation and Crop Insurance Clark, R.G. & Johnson, D.H. 1996. Long- Programs. U.S. Department of Agriculture, term declines in nest success of prairie ducks. Economic Research Service, Economic Journal of Wildlife Management 60: 247–257. Information Bulletin 94. U.S. Department of Blanche, B., Harrell, D. & Saichuk, J. 2009. Weed Agriculture, Economic Research Service, management. In J. Saichuk (ed.), Louisiana Rice Washington D.C., USA. Production Handbook, pp. 3–15. Louisiana State Congressional Budget Office (CBO). 2014. Budget University, College of Agriculture, AgCenter, Estimates for H.R. 2642, the Agricultural Act of Baton Rouge, USA. Accessible at http://www. 2014. Accessible at http://www.cbo.gov/ lsuagcenter.com/NR/rdonlyres/F7E930BF- sites/default/files/cbofiles/attachments/hr 346A-45DA-A989-37FE3D0B0F11/97775/ 2642LucasLtr.pdf (last accessed 14 February pub2331RiceProductionHandbook2014 2014). completebook.pdf (last accessed 29 August Couvillion, B.R., Barras, J.A., Steyer, G.D., Sleavin, 2014). W., Fischer, M., Beck, H., Trahan, N., Griffin, Brady, S.J. 2005. Highly erodible land and B. & Heckman, D. 2011. Land Area Change in Swampbuster provisions of the 2002 Farm Coastal Louisiana from 1932 to 2010. U.S. Act. In J.B. Haufler (ed.), Fish and Wildlife Geological Survey Scientific Investigations Benefits of Farm Bill Conservation Programs: Map 3164, scale 1:265,000. U.S. Geological 2000–2005 Update, pp. 5–15. Technical Survey, Washington D.C., USA. Review 05-2. The Wildlife Society, Bethesda, Cowardin, L.M., Gilmer, D.S. & Shaiffer, C.W. Maryland, USA. 1985. Mallard recruitment in the agricultural Burger, Jr., L.W. 2000. Wildlife responses to the environment of North Dakota. Wildlife Conservation Reserve Program in the Monograph 92: 1–37. Southeast. In W.L. Hohman & D.J. Halloum Cowardin, L.M., Shaffer, T.L. & Arnold, P.M. (eds.), A Comprehensive Review of Farm Bill 1995. Evaluation of Duck Habitat and Estimation Contributions to Wildlife Conservation, 1985–2000, of Duck Population Sizes with a Remote-Sensing pp. 55–73. U.S. Department of Agriculture, Based System. Biological Science Report No. 2. Natural Resources Conservation Service, U.S. Department of the Interior, National Wildlife Habitat Management Institute, Biological Service, Washington D.C., USA. Technical Report USDA/NRCS/WHMI- Cox, Jr., R.R. & Afton, A.D. 1996. Evening 2000. Madison, Mississippi, USA. flights of female Northern Pintails from a Carlyle, W.J. 1997. The decline of summerfallow major roost site. Condor 98: 810–819. on the Canadian Prairies. The Canadian Cox, Jr., R.R. & Afton, A.D. 1997. Use of habitats Geographer/Le Géographe Canadien 41: 267–280. by female northern pintails wintering in Claassen, R. 2005. Has conservation compliance southwestern Louisiana. Journal of Wildlife reduced soil erosion on U.S. cropland? In Management 61: 435−443.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 400 Private land habitat conservation

Ciuzio, E., Hohman, W.L., Martin, B., Smith, M.D., DOI, EC & Secretario de Desarrollo Social Stephens, S., Strong, A.M., and Vercauteren, Mexico. 2012. North American Waterfowl T. 2013. Opportunities and challenges to Management Plan 2012: People Conserving Waterfowl implementing bird conservation on private and Wetlands. U.S. Department of the Interior, lands. Wildlife Society Bulletin 37: 267–277. Fish and Wildlife Service, Washington, D.C., Dahl, T.E. 1990. Wetlands Losses in the United States USA. Accessible at http://www.fws.gov/ 1780s to 1980s. Department of the Interior, birdhabitat/NAWMP/files/NAWMP-Plan- U.S. Fish and Wildlife Service, Washington, EN-may23.pdf (last accessed 27 February D.C., USA. Accessible at http://www.fws.gov/ 2014). wetlands/Documents/Wetlands-Losses-in- Duebbert, H.F. & Kantrud, H.A. 1987. Use of the-United-States-1780s-to-1980s.pdf (last no-till winter wheat by nesting ducks in accessed 19 November 2013). North Dakota. Journal of Soil and Water Dahl, T.E. 2000. Status and Trends of Wetlands in Conservation 42: 50–53. the Conterminous United States 1986 to 1997. Duke, J.M., Dundas, S.J. & Messer, K.M. 2013. U.S. Department of the Interior, Fish Cost-effective conservation planning: Lessons and Wildlife Service, Washington D.C., from economics. Journal of Environmental USA. Accessible at http://www.fws.gov/ Management 125: 126–133. wetlands/Documents/Status-and-Trends- Dzus, E.H. & Clark, R.G. 1998. Brood survival of-Wetlands-in-the-Conterminous-United- and recruitment of Mallards in relation to States-1986-to-1997.pdf (last accessed 19 wetland density and hatching date. Auk 115: November 2013). 311–318. Devries, J.H., Armstrong, L.M., MacFarlane, R.J., Eadie, J.M., Elphick, C.S., Reinecke, K.J. & Miller, Moats, L. & Thoroughgood, P.T. 2008. M.R. 2008. Wildlife values of North American Waterfowl nesting in fall-seeded and spring- ricelands. In S.W. Manley (ed.), Conservation seeded cropland in Saskatchewan. Journal of in Ricelands of North America, pp. 7–90. Wildlife Management 72: 1790–1797. The Rice Foundation, Stuttgart, Arkansas, Devries, J.H., Rimer, S.O. & Walsh, E.M. 2010. USA. Cropland nesting by Long-billed Curlews in Earl, J.P. 1950. Production of Mallards on irrigated southern Alberta. Prairie Naturalist 42: 123– land in the Sacramento Valley, California. 129. Journal of Wildlife Management 14: 332–342. Doherty, K.E., Ryba, A.J., Stemler, C.L., Elphick, C.S. 2000. Functional equivalency Niemuth, N.D. & Meeks, W.A. 2013. between ricefields and seminatural wetland Conservation planning in an era of change: habitats. Conservation Biology 14: 181–191. State of U.S. Prairie Pothole Region. Wildlife Elphick, C.S. & Oring, L.W. 1998. Winter Society Bulletin 37: 546–563. management of Californian rice fields for DOI & EC (U.S. Department of Interior & waterbirds. Journal of Applied Ecology 35: 95– Environment Canada). 1986. North American 108. Waterfowl Management Plan. U.S. Department Feng, H., Hennessy, D.A. & Miao, R. 2013. The of the Interior, Fish and Wildlife Service, effects of government payments on cropland Washington, D.C., USA. Accessible at acreage, conservation reserve program http://www.fws.gov/birdhabitat/NAWMP/ enrollment, and grassland conversion in the files/NAWMP.pdf (last accessed 27 February Dakotas. American Journal of Agricultural 2014). Economics 95: 412–418.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 401

Fleury, B.E. & Sherry, T.W. 1995. Long-term Greenwood, R.J., Sargeant, A.B., Johnson, D.H., population trends of colonial wading birds in Cowardin, L.M. & Shaffer, T.L. 1995. Factors the southern United States: The impact of associated with duck nest success in the crayfish aquaculture on Louisiana populations. Prairie Pothole Region of Canada. Wildlife The Auk 112: 613–632. Monographs 128: 1–57. Foley, J.A., Defries, R., Asner, G.P., Barford, C., Gutzwiller, K.J. & Flather, C.H. 2011. Wetland Bonan, G., Carpenter, S.R., Chapin, F.S., Coe, features and landscape context predict M.T., Daily, G.C., Gibbs, H.K., Helkowski, the risk of wetland habitat loss. Ecological J.H., Holloway, T., Howard, E.A., Kucharik, Applications 21: 968–982. C.J., Monfreda, C., Patz, J.A., Prentice, I.C., Guyn, K.L. & Clark, R.G. 2000. Nesting effort of Ramankutty, N. & Snyder, P.K. 2005. Global Northern Pintails in Alberta. Condor 102: consequences of land use. Science 309: 619–628. 570–574. Heard, L.P., Allen, A.W., Best, L.B., Brady, S.J., Food and Agriculture Organization (FOA). 2010. Burger, W., Esser, A.J., Hackett, E., Johnson, The State of Food Insecurity in the World: Addressing D.H., Pederson, R.L., Reynolds, R.E., Rewa, Food Insecurity in Protracted Crises. Food and C., Ryan, M.R., Molleur, R.T. & Buck, Agriculture Organization of the United P. 2000. In W.L. Hohman & D.J. Halloum Nations, World Summit on Food Security, (eds.), A Comprehensive Review of Farm November 16–18, 2009, Rome, Italy. Bill Contributions to Wildlife Conservation, Accessible at http://www.fao.org/docrep/ 1985–2000. U.S. Department of Agriculture, 013/i1683e/i1683e.pdf (last accessed 19 Natural Resources Conservation Service, November 2013). Wildlife Habitat Management Institute, Frayer, W.E., Monahan, T.J., Bowden, D.C. & Technical Report USDA/NRCS/WHMI- Graybill, F.A. 1983. Status and Trends of Wetlands 2000. Madison, Mississippi, USA. and Deepwater Habitats in the Conterminous Higgins, K.F. 1977. Duck nesting in intensively United States, 1950’s to 1970’s. Colorado State farmed areas of North Dakota. Journal of University, Fort Collins, Colorado, USA. Wildlife Management 41: 232–242. Fredrickson, L.H. & Taylor, T.S. 1982. Management Hoekstra, J.M., Boucher, T.M., Ricketts T.H & of Seasonally Flooded Impoundments for Wildlife. Roberts, C. 2005. Confronting a biome crisis: Resource Publication 148, U.S. Department global disparities of habitat loss and of the Interior, Fish and Wildlife Service, protection. Ecology Letters 8: 23–29. Washington D.C., USA. Hohman, W.L., Moore, J.L., Stark, T.M., Gerard, P. 1995. Agricultural Practices, Farm Policy and Weisbrich, G.A. & Coon, R.M. 1994. the Conservation of Biological Diversity. Biological Breeding waterbird use of Louisiana rice Science Report No. 4. U.S. Department of fields in relation to planting practices. the Interior, National Biological Service, Proceedings of the Southeastern Association of Fish Washington D.C., USA. and Wildlife Agencies 48: 31–37. Goelitz, W.A. 1918. The destruction of nests by Hohman, W.L., Kidd, G., Nelson, R. & Pitre, J. In farming operations in Saskatchewan. Auk 35: press. Opportunities for crane conservation 238–240. through U.S. Department of Agriculture Government of Canada. 1991. The Federal Policy on Conservation Programs. In J.E. Austin & K. Wetland Conservation. Environment Canada. Morrison (eds.), Cranes and Agriculture – Ottawa, Ontario, Canada. a Practical Guide to Conservationists and Land

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 402 Private land habitat conservation

Managers, pp. 000–000. International Union Loesch, C.R., Reynolds, R.E. & Hansen, L.T. for Conservation of Nature, Species Survival 2012. An assessment of re-directing breeding Commission, Crane Specialist Group, Gland, waterfowl conservation relative to predictions Switzerland. of climate change. Journal of Fish and Wildlife Intergovernmental Panel on Climate Change Management 3: 1–22. (IPCC). 2000. IPCC Special Report: Emissions Lokemoen, J.T. & Beiser, J.A. 1997. Bird use and Scenarios, Summary for Policymakers. A Special nesting in conventional, minimum-tillage, Report of IPCC Working Group III, and organic cropland. Journal of Wildlife ISBN: 92-9169-113-5. World Meteorological Management 61: 644–655. Organization, United Nations Environment Lynch-Stewart, P., Rubec, C.D.A., Cox, K.W. & Programme, IPCC. Accessible at http://www. Patterson, J.H. 1993. A Coming of Age: Policy for ipcc.ch/pdf/special-reports/spm/sres-en.pdf Wetland Conservation in Canada. Report No. 93- (last accessed 14 May 2014). 1. North American Wetlands Conservation Johnston, C.A. 2013. Wetland losses due to row Council (Canada), Ottawa, Ontario. crop expansion in the Dakota Prairie Pothole Maizel, M., White, R.D., Gage, S., Osbourne, Region. Wetlands 33: 175–182. L., Root, R., Stitt, S. & Muehlbach, G. Johnson, D.H. 2000. Grassland bird use of 1998. Historical Interrelationships between Conservation Reserve Program fields in the Population Settlement and Farmland in the Northern Great Plains. In W.L. Hohman & Conterminous United States, 1790 to 1992. In D.J. Halloum (eds.), A Comprehensive Review of T.D. Sisk (ed.), Perspectives on the Land-use Farm Bill Contributions to Wildlife Conservation, History of North America: a Context for 1985–2000, pp. 19–33. U.S. Department of Understanding or Changing Environment, pp. 5–13. Agriculture, Natural Resources Conservation Biological Science Report USGS/BRD/BSR- Service, Wildlife Habitat Management 1998-0003. U.S. Geological Survey, Biological Institute, Technical Report USDA/NRCS/ Resources Division, Washington D.C., USA. WHMI-2000. Madison, Mississippi, USA. Marty, J.R. 2013. Seed and waterbird abundances in Johnson, R.R., Granfors, D.A., Niemuth, ricelands in the Gulf Coast Prairies of Louisiana N.D., Estey, M.E. & Reynolds, R.E. 2010. and Texas. M.Sc. thesis, Mississippi State Delineating grassland bird conservation areas University, Starkville, USA. in the U.S. Prairie Pothole Region. Journal of McClain, W.R, Romaire, R.P., Lutz, C.G. & Fish and Wildlife Management 1: 38–42. Shirley, M.G. 2009. Louisiana Crawfish Klett, A.T., Shaffer, T.L. & Johnson, D.H. 1988. Production Manual. Publication No. 2637, Duck nest success in the Prairie Pothole Louisiana State University, College of Region. Journal of Wildlife Management 52: Agriculture, AgCenter, Baton Rouge, USA. 431–440. Accessible at http://www.lsuagcenter.com/ Lewis, D.J., Plantinga, A.J., Nelson, E. & en/crops_livestock/aquaculture/crawfish/ Polasky, S. 2011. The efficiency of voluntary Crawfish+Production+Manual.htm (last incentives policies for preventing biodiversity accessed 5 May 2014). loss. Resource and Energy Economics 33: 192– Merenlender, A.M., Newburn, D., Reed, S.E. & 211. Rissman, A.R. 2009. The importance of Linscombe, R.G. 1972. Crop damage by waterfowl in incorporating threat for efficient targeting southwestern Louisiana. M.Sc. Thesis, Louisiana and evaluation of conservation investments. State University, Baton Rouge, USA. Conservation Letters 2: 240–41.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 403

Millar, J.B. 1989. Perspectives on the Status of Quality, Department of the Interior, Great Canadian Prairie Wetlands. In R.R. Sharitz Lakes Indian Fish and Wildlife Commission, & J.W. Gibbons (eds.), Freshwater Wetlands National Oceanic and Atmospheric and Wildlife, pp. 829−852. U.S. Department Administration. Washington, D.C., USA. of Energy Symposium Series No. 61, U.S. Newburn, D., Reed, S., Berck, P. & Merenlender, Department of Energy, Oak Ridge, Tennessee, A. 2005. Economics of land-use change in USA. prioritizing private lands conservation. Miller, M.R. 1987. Fall and winter foods of Conservation Biology 19: 1411–1420. Northern Pintails in the Sacramento Valley, Nickerson, C., Ebel, R., Borchers, A. & Carriazo, California. Journal of Wildlife Management 51: F. 2011. Major Uses of Land in the United 405−414. States, 2007, EIB-89. U.S. Department of Miller, M.R. & Duncan, D.C. 1999. The Agriculture, Economic Research Service, Northern Pintail in North America: status Washington D.C, USA. Accessible at http:// and conservation needs of a struggling www.ers.usda.gov/media/188404/eib89_ population. Wildlife Society Bulletin 27: 788– 2_.pdf (last accessed November 2013). 800. Niemuth, N.D., Reynolds, R.E., Granfors, Milonski, M. 1958. The significance of farmland D.A., Johnson, R.R., Wangler, B. & Estey, for waterfowl nesting and techniques for M.E. 2009. Landscape-level planning for reducing losses due to agricultural practices. conservation of wetland birds in the U.S. Transactions of the North American Wildlife Prairie Pothole Region. In J.J. Millspaugh Conference 23: 215–227. & F.R. Thompson, III (eds.), Models for Mulik, K. & Koo, W.W. 2006. Substitution Between Planning Wildlife Conservation in Large U.S. and Canadian Wheat by Class. Center for Landscapes, pp. 533–560. Academic Press, Agricultural Policy and Trade Studies Burlington, Massachusetts, USA. Department of Agribusiness and Applied Noss, R.F., LaRoe III, E.T. & Scott, J.M. 1995. Economics, North Dakota State University, Endangered Ecosystems of the United States: a Fargo, North Dakota, USA. Preliminary Assessment of Loss and Degradation. Murdoch, W., Polasky, S., Wilson, K.A., Biological Report 28. U.S. Department of the Possingham, H.P., Kareiva, P. & Shaw, R. Interior, National Biological Service. 2007. Maximizing return on investment in Washington, D.C., USA. conservation. Biological Conservation 139: 375– Peterjohn, B.G. & Sauer, J.R. 1999. Population 388. status of North American grassland birds Naidoo, R. & Iwamura, T. 2007. Global- from the North American Breeding Bird scale mapping of economic benefits Survey, 1966–1996. Studies in Avian Biology 19: from agricultural lands: Implications for 27–44. conservation priorities. Biological Conservation Podruzny, K.M. Devries, J.H., Armstrong, L.M. 140: 40–49. & Rotella, J.J. 2002. Long-term response of National Fish, Wildlife and Plants Climate Northern Pintails to changes in wetlands and Adaptation Strategy Partnership (NFWPCAS). agriculture in the Canadian Prairie Pothole 2012. Draft National Fish, Wildlife and Plants Region. Journal of Wildlife Management 66: Climate Adaptation Strategy: Shared Solutions to 993–1010. Protect Shared Values. Association of Fish and Rashford, B.S. & Adams, R.M. 2007. Improving the Wildlife Agencies, Council on Environmental cost-effectiveness of ecosystem management:

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 404 Private land habitat conservation

An application to waterfowl production. Riemer, G. 2005. Land-use policy change and American Journal of Agricultural Economics 89: the ramifications for stewardship and 755–768. waterfowl conservation in Saskatchewan. In Rashford, B.S., Walker, J.A. & Bastian, C.T. T.A. Radenbaugh & G.C. Sutter (eds.), 2010. Economics of grassland conversion to Managing Changing Prairie Landscapes, pp. cropland in the Prairie Pothole Region. 11−22. University of Regina Press, Regina, Conservation Biology 25: 276–284. Saskatchewan, Canada. Rashford, B.S., Bastian, C.T. & Cole, J.G. 2011. Ringelman, J.K., Forman, K.J., Granfors, D.A., Agricultural land-use change in prairie Johnson, R.R., Lively, C.A., Naugle, D.E., Canada: Implications for wetland and Niemuth, N.D. & Reynolds, R.E. 2005. Prairie waterfowl conservation. Canadian Journal of Pothole Joint Venture 2005 Implementation Agricultural Economics 59: 185–205. Plan. U.S. Fish and Wildlife Service, Prairie Rave, D.P. & Cordes, C.L. 1993. Time-activity Pothole Joint Venture, Denver, Colorado, budget of northern pintails using nonhunted USA. ricefields in southwest Louisiana. Journal of Ryan, M.R. 2000. Impact of the Conservation Field Ornithology 64: 211–218. Reserve Program on wildlife conservation in Reynolds, R.E. 2000. Waterfowl responses to the Midwest. In W.L. Hohman & D.J. the Conservation Reserve Program in the Halloum (eds.), A Comprehensive Review of Northern Great Plains. In W.L. Hohman & Farm Bill Contributions to Wildlife Conservation, D.J. Halloum (eds.), A Comprehensive Review of 1985–2000. pp. 45–54. U.S. Department of Farm Bill Contributions to Wildlife Conservation, Agriculture, Natural Resources Conservation 1985–2000. pp. 35–43. U.S. Department of Service, Wildlife Habitat Management Agriculture, Natural Resources Conservation Institute, Technical Report USDA/NRCS/ Service, Wildlife Habitat Management WHMI-2000. Madison, Mississippi, USA. Institute, Technical Report USDA/NRCS/ Samson, F. & Knopf, F. 1994. Prairie WHMI-2000. Madison, Mississippi, USA. conservation in North America. BioScience 44: Reynolds, R.E., Cohan, D.R. & Johnson, 418–421. M.A. 1996. Using landscape information Sargeant, A.B., Greenwood, R.J., Sovada, M.A. & approaches to increase duck recruitment in Shaffer, T.L. 1993. Distribution and Abundance the Prairie Pothole Region. Transaction of the of Predators that Affect Duck Production: the North American Wildlife and Natural Resource Prairie Pothole Region. Resource Publication Conference 61: 86–93. 194. U.S. Department of Interior, Fish and Reynolds, R.E., Shaffer, T.L., Loesch, C.R. & Wildlife Service, Washington D.C., USA. Cox, Jr., R.R. 2006. The Farm Bill and duck Skagen, S.K. & Thompson, G. 2007. Northern production in the Prairie Pothole Region. Plains/Prairie Potholes Regional Shorebird Wildlife Society Bulletin 34: 963–974. Conservation Plan (Revised). U.S. Department of Richkus, K.D. 2002. Northern Pintail nest site Interior, Fish and Wildlife Service, Washington selection, nest success, renesting ecology and D.C., USA. survival in the intensively farmed prairies of Smith, R. J. & Sullivan, J.D. 1980. Reducing red southern Saskatchewan: An evaluation of the rice grain in rice fields by winter feeding of ecological trap hypothesis. Ph.D. thesis, ducks. Arkansas Farm Research 29: 3. Louisiana State University, Baton Rouge, Sohl, T.L., Sleeter, B.M., Saylera, K.L., Bouchard, Louisiana, USA. M.A., Reker, R.R., Bennett, S.L., Sleeter, R.R.,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 Private land habitat conservation 405

Kanengieter, R.L. & Zhue, Z. 2012. Spatially USDA NASS. 2014. 2012 Census of Agriculture, explicit land-use and land-cover scenarios for United States Summary and State Data, Volume the Great Plains of the United States. 1, Part 51 (AC-12-A-51). United States Agriculture, Ecosystems and Environment 153: Department of Agriculture, Washington D.C., 1–15. USA. Accessible at http://www.agcensus.usda. Statistics Canada. 2011. 2011 Census: Census of gov/Publications/2012/Full_Report/Volume Agriculture. Statistics Canada, Ottawa, Ontario, _1,_Chapter_1_US/usv1.pdf (last accessed Canada. 5 May 2014). Statistics Canada. 2012. 2012 Census: Census of USDA NRCS (U.S. Department of Agriculture, Agriculture. Statistics Canada, Ottawa, Natural Resources Conservation Service). Ontario, Canada. 2000. 1997 National Resources Inventory. Statistics Canada. 2013. Estimated Areas, Yield, U.S. Department of Agriculture, Natural Production, Average Farm Price and Total Farm Resources Conservation Service, Resources Value of Principal Field Crops, in Imperial Units, Inventory Division, Washington, D.C., Annual. Table 001-0017-CANSIM (database). USA. Last updated December 4, 2013. Statistics USDA NRCS. 2013. 2007 National Resources Canada, Ottawa, Ontario, Canada. Inventory. U.S. Department of Agriculture, Stephens, S.E., Rotella, J.J., Lindberg, M.S., Taper, Natural Resources Conservation Service, M.L. & Ringleman, J.K. 2005. Duck nest Resources Inventory Division, Washington survival in the Missouri Coteau of North D.C., USA. Accessible at http://www.nrcs. Dakota: landscape effects at multiple usda.gov/wps/portal/nrcs/detail/national/ spatial scales. Ecological Applications 15: 2137– technical/nra/nri/?cid=stelprdb1117258 (last 2149. accessed 12 February 12 2014). Stephens, S.E., Walker, J.A., Blunck, D.R., USFWS (U.S. Fish and Wildlife Service). 2008. Jayaraman, A., Naugle, D.E., Ringleman, J.K. Strategic Habitat Conservation Handbook: A & Smith, A.J. 2008. Predicting risk of habitat guide to Implementing the Technical Elements of conversion in native temperate grasslands. Strategic Habitat Conservation, Version 1.0. U.S. Conservation Biology 22: 1320–1330. Department of Interior, Fish and Wildlife Taft, O.W. & Elphick, C.S. 2007. Waterbirds on Service, Washington D.C., USA. Working Lands: Literature Review and Verburg, P.H., Schot, P.P. Dijst, M.J. & Veldkamp, Bibliography Development. Technical Report. A. 2004. Land use change modeling: Current National Audubon Society, New York, USA. practice and research priorities. GeoJournal 61: USDA FSA (U.S. Department of Agriculture, 309–324. Farm Service Agency). 2013. U.S. Non-Cropland Watmough, M.D. & Schmoll, M.J. 2007. to Cropland Report. Washington, D.C., USA. Environment Canada’s Prairie & Northern Habitat Accessible at http://www.fsa.usda.gov/FSA/ Monitoring Program Phase II: Recent habitat trends webapp?area=newsroom&subject=landing& in the Prairie Habitat Joint Venture. Technical topic=foi-er-fri-dtc (last accessed 15 January Report Series No. 493. Environment Canada, 2014). Canadian Wildlife Service, Edmonton, USDA NASS (U.S. Department of Agriculture, Alberta, Canada. National Agricultural Statistics Service). 2013. Webster, E. & Levy, R. 2009. Weed management. In R. Acreage report. United States Department of Saichuk (ed.), Louisiana Rice Production Handbook, Agriculture, Washington D.C., USA. pp. 46–71. Louisiana State University, College

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 406 Private land habitat conservation

of Agriculture, AgCenter, Baton Rouge, Withey, J.C., Lawler, J.J., Polasky, S., Plantinga, A.J., USA. http://www.lsuagcenter.com/en/ Nelson, E.J., Kareiva, P., Wilsey, C.B., Schloss, communications/publications/Publications C.A., Nogeire, T.M., Ruesch, A., Ramos Jr., J. +Catalog/Crops+and+Livestock/Rice/Rice & Reid, W. 2012. Maximising return on +Production+Handbook.htm.m (last accessed conservation investments in the conterminous 5 May 2014). USA. Ecology Letters 15: 1249–1256. Williams, B.K., Koneff, M.D. & Smith, D.A. 1999. Wright, C.K. & Wimberly, M.C. 2013. Recent Evaluation of waterfowl conservation under land use change in the Western Corn Belt the North American Waterfowl Management threatens grasslands and wetlands. Proceedings Plan. Journal of Wildlife Management 63: of the National Academy of Sciences 110: 4134– 417–440. 4139.

Photograph: Aerial view of southwest Louisiana rice fields, by John K. Saichuk/Louisiana State University AgCenter.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 368–406 407

Estimating habitat carrying capacity for migrating and wintering waterfowl: considerations, pitfalls and improvements

CHRISTOPHER K. WILLIAMS1*, BRUCE D. DUGGER2, MICHAEL G. BRASHER3, JOHN M. COLUCCY4, DANE M. CRAMER4, JOHN M. EADIE5, MATTHEW J. GRAY6, HEATH M. HAGY7, MARK LIVOLSI1, SCOTT R. MCWILLIAMS8, MARK PETRIE9, GREGORY J. SOULLIERE10, JOHN M. TIRPAK11 & ELISABETH B. WEBB12

1Department of Entomology and Wildlife Ecology, University of Delaware, 250 Townsend Hall, Newark, Delaware 19716, USA. 2Department of Fisheries and Wildlife, Oregon State University, Corvallis, Oregon 97331, USA. 3Ducks Unlimited, Inc., Gulf Coast Joint Venture, 700 Cajundome Blvd, Lafayette, Louisiana 70506, USA. 4Ducks Unlimited, Inc., Great Lakes/Atlantic Region, 1220 Eisenhower Place, Ann Arbor, Michigan 48108, USA. 5Department of Wildlife, Fish and Conservation Biology, University of California, 1 Shields Avenue, Davis, California 95616, USA. 6Department of Forestry, Wildlife and Fisheries, University of Tennessee, 274 Ellington Plant Sciences Building, Knoxville, Tennessee 37996, USA. 7Forbes Biological Station – Bellrose Waterfowl Research Center, Illinois Natural History Survey, University of Illinois at Urbana-Champaign, P.O. Box 590, Havana, Illinois 62644, USA. 8Department of Natural Resources Science, University of Rhode Island, 105 Coastal Institute, Kingston, Rhode Island 02881, USA. 9Ducks Unlimited Inc. Western Regional Office, 17800 SE Mill Plain Blvd. Suite 120, Vancouver, Washington 98683, USA. 10U.S. Fish and Wildlife Service, Upper Mississippi River and Great Lakes Region Joint Venture Science Office, 2651 Coolidge Road, Suite 101, East Lansing, Michigan 48823, USA. 11U.S. Fish and Wildlife Service, Gulf Coastal Plains & Ozarks Landscape Conservation Cooperative, 700 Cajundome Blvd, Lafayette, Louisiana 70506, USA. 12U.S. Geological Survey, Missouri Cooperative Fish and Wildlife Research Unit, Department of Fisheries and Wildlife Sciences, University of Missouri, Columbia, Missouri 65211, USA. *Correspondence author: E-mail: [email protected]

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 408 Waterfowl carrying capacity models

Abstract Population-based habitat conservation planning for migrating and wintering waterfowl in North America is carried out by habitat Joint Venture (JV) initiatives and is based on the premise that food can limit demography (i.e. food limitation hypothesis). Consequently, planners use bioenergetic models to estimate food (energy) availability and population-level energy demands at appropriate spatial and temporal scales, and translate these values into regional habitat objectives. While simple in principle, there are both empirical and theoretical challenges associated with calculating energy supply and demand including: 1) estimating food availability, 2) estimating the energy content of specific foods, 3) extrapolating site-specific estimates of food availability to landscapes for focal species, 4) applicability of estimates from a single species to other species, 5) estimating resting metabolic rate, 6) estimating cost of daily behaviours, and 7) estimating costs of thermoregulation or tissue synthesis. Most models being used are daily ration models (DRMs) whose set of simplifying assumptions are well established and whose use is widely accepted and feasible given the empirical data available to populate such models. However, DRMs do not link habitat objectives to metrics of ultimate ecological importance such as individual body condition or survival, and largely only consider food-producing habitats. Agent-based models (ABMs) provide a possible alternative for creating more biologically realistic models under some conditions; however, ABMs require different types of empirical inputs, many of which have yet to be estimated for key North American waterfowl. Decisions about how JVs can best proceed with habitat conservation would benefit from the use of sensitivity analyses that could identify the empirical and theoretical uncertainties that have the greatest influence on efforts to estimate habitat carrying capacity. Development of ABMs at restricted, yet biologically relevant spatial scales, followed by comparisons of their outputs to those generated from more simplistic, deterministic models can provide a means of assessing degrees of dissimilarity in how alternative models describe desired landscape conditions for migrating and wintering waterfowl.

Key words: agent-based models, bioenergetics, carrying capacity, daily ration models, energy demand, energy supply, waterfowl.

Population-based, habitat conservation of habitat objectives and deliver programmes in the North American Waterfowl Management regions that support the majority of ducks Plan (NAWMP 2012) is planned and wintering in the United States. Conservation implemented by regional, collaborative plans developed by these JVs are based on partnerships named Joint Ventures (JVs). the premise that food during the non- Since 1987, JVs have spent approximately breeding period can limit demographics and US$5 billion to conserve or manage 7.8 thus population trends for waterfowl (i.e. the million ha of habitat. Nine JVs set waterfowl food limitation hypothesis). This hypothesis

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 409 is supported by research indicating that duck 1999; Goss-Custard et al. 2002; Goss- body condition correlates with winter habitat Custard et al. 2003). In its simplest form, conditions (Delnicki & Reinecke 1986; carrying capacity may be expressed in terms Lovvorn 1994; Thomas 2004; Heitmeyer of duck energy-days (DED): 2006; Moon et al. 2007), which influences diet Food available () g dry weight  quality (Loesch & Kaminski 1989), and True metabolisable energy moreover that body condition influences ()kcal/g dry weight survival (e.g. Moon & Haukos 2006; Bergan & DED = (1) Daily energy expenditure Smith 1993) and the timing of migration () phenology (Heitmeyer 1988, 2006). At the kcal/day population level, winter habitat conditions Thus, under the assumption that DRMs can influence the distribution of ducks within reasonably reflect foraging dynamics of and across winters (Nichols et al. 1983; Hepp free-ranging waterfowl, useful calculations & Hines 1991; Lovvorn & Baldwin 1996; of carrying capacity require estimates of: Pearse et al. 2012). Finally, there is evidence 1) habitat-specific food production (g dry for cross-seasonal influences, with winter and weight per unit area), 2) functional migration habitat conditions influencing availability of waterfowl foods (g dry weight subsequent productivity (Heitmeyer & per unit area; e.g. Greer et al. 2009), 3) true Fredickson 1981; Kaminski & Gluesing 1987; metabolisable energy of available foods Raveling & Heitmeyer 1989; Guillemain et al. (kcal per g dry weight; Miller & Reinecke 2008; Devries et al. 2008; Anteau & Afton 1984), 4) daily energy requirements of target 2009). species (kcal), and 5) region- and species- Most JVs use a bioenergetics model to specific population targets (Petrie et al. estimate habitat carrying capacity and 2011). The actual forms of models being project habitat needs to support waterfowl used by JVs are more sophisticated than the populations at target levels during the non- simple equation depicted above. For breeding season (e.g. Prince 1979; Reinecke example, most JVs model energy supply and et al. 1989; Petrie et al. 2011). Bioenergetics demand in time and space (e.g. Central Valley models represent a class of resource JV 2006; Pacific Coast JV 2004) with the depletion models and those used by winter understanding that energy supplies may be habitat JVs often take the form of daily influenced by natural or intentional flooding ration models (DRMs; Goss-Custard et al. of habitats and that demand may vary 2003). While DRMs can take different temporally based on population size, forms, they generally aggregate food energy migration chronology, changes in species density across multiple habitat patches composition, physiological needs, weather (using either habitat-specific values or and other endogenous or exogenous factors. average values across habitats) and divide by Regardless of model sophistication, all daily energy demands of a target duck DRMs require some estimate of energy species to estimate the theoretical carrying supply and demand. capacity of a given area (Miller & Newton While simple in principle, there are

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 410 Waterfowl carrying capacity models empirical and theoretical challenges model structures and the implications of associated with estimating energy supply and new model advances and pitfalls, for future demand. When estimating energy supply, bias conservation planning directed towards can occur from inaccurate estimates of: 1) ducks at staging and wintering sites. food availability, 2) energy content of specific foods, 3) extrapolating site-specific estimates Energy supply of food availability to landscapes for focal Calculating energy supply for waterfowl species, or 4) assuming estimates from a requires an empirical measure of habitat- single species of waterfowl apply to other specific food production and availability to species. Likewise, quantitative challenges exist waterfowl. Here we define food availability when estimating energy demand during the as the production of food minus an amount non-breeding period. Beyond the challenge of not exploitable by waterfowl (i.e. the giving- estimating regional population size (Soulliere up density, or food availability threshold; et al. 2013), those planning waterfowl Greer et al. 2009; Hagy & Kaminski 2012a). management programmes may also face When food availability is known or can be biased estimates of: 1) resting metabolic rate, reasonably estimated, energy supply can be 2) cost of daily behaviours, and 3) costs of calculated using energetic values of each thermoregulation or tissue synthesis. Finally, food to individual species and extrapolated bioenergetic models can take various forms across the area of interest. Below, we ranging from simple DRMs to spatially- explore each of these components in explicit, agent-based models that incorporate increased detail. additional mechanistic details of the systems being modelled. However, the conditions Estimating food production dictating when more complicated models Direct and indirect methods have been are required is not thoroughly understood developed for estimating food availability in (Goss-Custard et al. 2003). the environment. For dabbling ducks, To address some of these challenges, a clipping of inflorescences, extracting soil special session was convened at the 6th cores and sweep nets are the primary tools North American Duck Symposium to used to estimate the biomass of seeds, consider fundamental aspects of the DRMs nektonic and benthic forage produced most commonly used in conservation (Dugger et al. 2008; Kross et al. 2008a; planning undertaken by winter habitat Evans-Peters et al. 2010; Hagy & Kaminski JVs in North America. A comprehensive 2012a). assessment and comparison of the strengths, Vegetative food production is often weaknesses and utility of the full suite of estimated using floristic measurements (e.g. conservation planning models for wintering Gray et al. 1999a, 2009; Naylor et al. 2005) or waterfowl were beyond the scope of this by assessing seed, tuber and plant part paper. Rather, talks addressed several key biomass at the end of the growing season elements associated with estimating energy (Kross et al. 2008a). Seeds can be threshed supply and demand, along with alternative from inflorescences and collected using core

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 411 samples (Kross et al. 2008a), predicted using Traditional core samplers used for vegetation morphology indices (Gray et al. sampling duck foods range from 5–10 cm in 1999a) using visual assessments (Naylor et al. diameter (Swanson 1983; Stafford et al. 2005), or measured directly using other 2006; Greer et al. 2007; Hagy et al. 2012b; means (Gray et al. 2013). Initially, Laubhan Smith et al. 2012) and 5–10 cm in depth & Fredrickson (1992) developed equations (Greer et al. 2007; Kross et al. 2008a; that predicted seed production using Olmstead 2010; Hagy & Kaminski phytomorphological measurements. These 2012b), with deeper samplers used for models were not widely used because they larger, longer-necked taxa (e.g. 30 cm for required extensive field measurements swans; Santamaria & Rodriguez-Girones and predictions outside of the region of 2002). In relatively shallow water, core development were unreliable (Gray et al. samplers may yield simultaneous density 1999a). Gray et al. (2009) determined that estimates for submersed aquatic vegetation the area of a seed head was a reliable (e.g. Myriophalum sp., Ceratophyllum sp., Elodia predictor of seed mass and developed a sp.), nektonic and benthic invertebrates, simplified process of using desktop seeds and tubers (Swanson 1978). In deeper scanners to predict seed production. Naylor water or in areas with high densities of et al. (2005) described a process for ranking aquatic vegetation, core samplers may be moist-soil habitat quality for waterfowl in used in combination with sweep nets California, USA, based on visual estimates (Murkin et al. 1996; Tidwell et al. 2013), of plant composition and forage quality. exclusion devices (Straub et al. 2012) and Stafford et al. (2011) replicated this box samplers (Synchra & Adamek 2010) to technique in Illinois and found that the provide better measures of food availability. index explained 65% of the variation in Regardless of the direct sampling method moist-soil plant seed biomass collected used, samples require extensive time to from core samples. These rapid assessment process in the laboratory. Food items techniques allow wetland managers to typically are sorted from the plant, soil and obtain efficiently general estimates of detritus by hand using a series of graduated seed biomass for waterfowl in moist-soil sieves, dried to constant mass in a forced air wetlands without extensive and costly oven, identified and weighed by species or laboratory or field work. However, appropriate biological classification. This Evans-Peters (2010) suggested these visual process is tedious and costly (Stafford assessments omit ≤ 30% of seed biomass in et al. 2011). Sub-sampling is a well-vetted the seed bank. Moreover, accuracy of visual approach for reducing processing time (e.g. assessments of standing vegetation during Proctor & Marks 1974; Schroth & Kolbe the growing season or prior to vegetation 1993; Murkin et al. 1996; Reinecke & senescence and inundation by water may not Hartke 2005; Smith et al. 2012). Waterfowl accurately reflect food densities at later researchers have recently applied soil core points in time useful to managers (Greer et sub-sampling and verified that overall means al. 2007; Fleming et al. 2012). are similar between sub-sampled and whole-

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 412 Waterfowl carrying capacity models processed samples for moist-soil seeds dabbling ducks (Hagy & Kaminski 2012b; (Hagy et al. 2011; Stafford et al. 2011), rice Olmstead et al. 2013). Additionally, biomass Oryza sp. grains (Stafford et al. 2006) estimates may be biased due to incomplete and macroinvertebrates (M. Livolsi et al. recovery of seeds during sample processing, University of Delaware, unpubl. data), through non-detection, loss or destruction. although variance associated with the Hagy et al. (2011) reported that c. 14% of estimates may increase (Hagy et al. 2011). known seeds were not recovered from core Currently, there is little information on samples during sorting, and that recovery the optimal sample size for core samples in rates depended on seed size. Thus, energy waterfowl habitats, and some information information based on incomplete recovery suggests optimal sample size may be could underestimate food availability by difficult to predict (Reinecke & Hartke 2005; 10–20% (Hagy et al. 2011), whereas analyses Marty 2013). While current studies often that do not account for actual diet and food base sample size on financial and temporal use bias could overestimate food availability constraints (Sherfy et al. 2000; Evans-Peters by as much as 47% (Hagy et al. 2011; et al. 2012), 20–30 samples per patch have Olmstead et al. 2013). been shown to result in coefficient of Finally, biomass estimates may be variable variation (C.V.) values of < 10% (Dugger among locations and geographical regions et al. 2008; Greer et al. 2009; Evans-Peters (e.g. Stafford et al. 2006a; 2011; Kross et al. et al. 2012). Preliminary power analysis 2008a; Evans-Peters et al. 2012; Hagy & of saltmarsh systems indicates that c. 40 Kaminski 2012b); thus, use of local samples per habitat type/location/time productivity estimates may be biased when period would be reasonable for some scaled to regional levels. A recent simulation habitats, but other habitats such as “high- of core sampling indicated that detection marsh” and rice fields show greater probabilities for food items varied by food variability and may require significantly densities, corer size and the underlying more samples (K. Ringelman, University of pattern of food distribution (A. Behney, Delaware, unpubl. data; Marty 2013). More Southern Illinois University, unpubl. data). comprehensive research is needed to While biologists can use methodological provide appropriate sample sizes for improvements to reduce local variance, they quantifying food availability within different should acknowledge the possibility of habitat types. geographic variation, random versus clumped Food biomass estimates may also be food distributions within habitats and the influenced by improper inclusion of prey possibility that birds may not follow an ideal seldom consumed or not energetically free distribution or forage optimally. profitable to waterfowl. For example, Therefore, biologists may wish to consider 30–70% of seeds sorted from core samples sampling in multiple locations to provide a collected in the Mississippi Alluvial Valley, better representation of values for regional- USA, in autumn had little nutritional value scale habitat management and conservation or were likely not consumed by most (e.g. Stafford et al. 2006; Kross et al. 2008a).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 413

Estimating food availability month. Stafford et al. (2006) reported that 58% of waste rice in the MAV fields Unbiased, precise estimates of food decomposed post-harvest in autumn, production and its energy density are not compared to 14% and 8% loss from sufficient for understanding the energy granivory and germination, respectively. supply available to migrating and wintering Foster et al. (2010b) found that monthly rate waterfowl, because all foods produced of loss of seeds for corn Zea sp., sorghum may not be available when waterfowl access Sorghum sp. and soybean Glycine sp. ranged habitats, for instance due to the depth or from 64 –84% post-harvest. Decomposition extent (i.e. surface area) of flooding (Kross rates likely vary with latitude as warmer et al. 2008b; Foster et al. 2010). Assessment temperatures would contribute to higher of availability includes accounting for rates. Thus, production estimates are useful physical accessibility and energy acquired to evaluate management actions, but given costs of foraging (Hagy et al. 2012b). sampling should either be timed to coincide Use of an area by migrating or wintering with waterfowl arrival or appropriate waterfowl may lag considerably from the adjustments are needed to initial estimates, time of seed maturation, and reliance on to account for seed loss not attributable to food production estimates from the end waterfowl foraging. of a growing season may overestimate The physical availability of foods for food available to the birds. Seed abundance waterfowl depends on the birds’ ability to in agricultural habitats can decrease extract foods from wetlands (e.g. Nolet et al. substantially between seed maturation and 2001). Studies of seed and invertebrate the arrival of waterfowl, especially in mid to biomass have traditionally assumed that southern latitudes of the United States every seed or invertebrate captured in the (Manley et al. 2004; Stafford et al. 2006; 5–10 cm deep core samples was available Greer et al. 2009; Foster et al. 2010b). to foraging waterfowl. However, species However, Marty (2013) reported an increase with different foraging behaviours and in abundance of waste rice in the Gulf morphologies (e.g. diving ducks Aytha sp. Coastal Prairie rice fields of Louisiana and versus dabbling ducks Anas sp.) can affect Texas during autumn, due to the production how much of the production is actually of a second unharvested rice crop (ratoon) available (Nudds & Kaminski 1984; Murkin in late autumn, following an initial late et al. 1996; Sherfy et al. 2000; Evans- summer harvest of rice. Seed loss rates vary Peters 2010; Olmstead 2010). Additionally, with hydrology and are attributable to foraging efficiency varies with sediment germination, consumption by non-target depths and seed type, and it is likely that wildlife and decomposition (Stafford seeds buried at greater depths may be less et al. 2006; Greer et al. 2007; Foster et al. profitable energetically (Nolet et al. 2001; 2010a). Hagy et al. (2012a) noted that Smith et al. 2011). However, if biologists decomposition of moist-soil seeds in sample systems with large amounts of flooded emergent wetlands was c. 18% per macroinvertebrates moving within the soil

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 414 Waterfowl carrying capacity models column, especially in a tidal system, deeper predictions at the patch level (van Gils et al. sampling depths may be more appropriate 2004), but are compatible with daily ration to estimate food availability (although this models currently used to predict carrying has yet to be tested). capacity for large regions (e.g. Soulliere et al. From behavioural and energetic 2007) and may have some value if patch- perspectives, estimates of food availability specific data are unavailable or impractical to must be adjusted when food densities are too obtain. Therefore, incorporating critical food low for energetically profitable foraging (van densities at the patch level will likely increase Gils et al. 2004; Nolet et al. 2006; Hagy et al. the accuracy of food availability estimates. 2012b). There is evidence for a critical food Ancillary modelling has indicated that failure density that remains after individuals either to apply foraging thresholds accurately at the give up foraging (Greer et al. 2009) or no patch level could affect food availability longer remove food from patches despite estimates by as much as 60%, and this bias continued foraging effort (Hagy 2010; Hagy varies with values of foraging thresholds and et al. 2012b). However, the critical food seed density (Pearse & Stafford 2014; H. density varies among habitats, regions and Hagy, Mississippi State University, unpubl. potentially even between foraging patches data). that differ in food composition (Baldassarre & Bolen 1984; Naylor 2002; Greer et al. 2009; True metabolisable energy of foods Hagy 2010). For example, Hagy (2010) noted True metabolisable energy (TME, kcal/g) that foraging thresholds varied widely for represents the amount of energy an moist-soil wetlands, but they were likely at individual bird receives from a food least 200 kg/ha in natural moist-soil item, after accounting for metabolic wetlands with a wide variety of seed taxa faecal losses and also endogenous urinary present. Hagy et al (2012b) reported residual losses as metabolised energy (Miller & seed densities exceeding 250kg/ha in moist- Reinecke 1984). The TME provides a soil wetlands in the MAV after waterfowl more accurate estimate of metabolised ceased to remove additional foods, whereas energy than apparent metabolisable energy Naylor (2002) reported residual densities (AME), because TME accounts for faecal of 30–160 kg/ha in these wetlands in and urinary losses (Miller & Reinecke California. Greer et al. (2009) and Baldassarre 1984; Karasov 1990). The TME of & Bolen (1984) reported residual densities waterfowl foods are important components of 50 kg/ha or less in flooded rice and dry for accurate assessments of waterfowl corn fields, respectively. Gray et al. (2013) bioenergetics and energetic carrying provided updated estimates of available food capacity. It may be calculated: a) indirectly, for waterfowl in agricultural fields, moist-soil using a regression model of total excretory wetlands and bottomland hardwoods, and energy on total food intake, or b) incorporated critical food densities into experimentally, determined by feeding birds these estimates. Application of fixed critical a controlled diet and measuring excretory food densities probably results in inaccurate energy (Sibbald 1975; Sibbald 1979;

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 415

Kaminski & Essig 1992). The TME values seed TME value of 2.5 kcals/g for their can then be used to calculate available projections of moist-soil seeds; however, energy by multiplying the mass of a food data from the Lower Klamath National item by its TME value and extrapolating the Wildlife Refuge, in northern California, resulting energy value across an area of indicate that mean seed TME values can be interest. less than 50% of this estimate (Dugger et al. Nonetheless, there is a lack of TME 2008). Most existing TME data are for seed values for common waterfowl foods and species from the Midwestern United States, species. Current studies have focused on and employing TME averages may therefore TME values for Mallard Anas platyrynchos, be inappropriate in regions where seed American Black Duck A. rubripes, Northern composition is significantly different. Pintail A. acuta, Canada Goose Branta Calculating potential available energy across canadensis, Blue-winged Teal A. discors and large areas using variable or biased TME Carolina Wood Duck Aix sponsa (Hoffman & values may result in meaningless estimates Bookhout 1985; Jorde & Owne 1988; Petrie of carrying capacity. Thus, additional 1994; Reinecke et al. 1989; Petrie et al. 1998; research efforts focused on deriving TME Sherfy 1999; Sherfy et al. 2001; Checkett et al. values experimentally for a wide variety of 2002; Kaminski et al. 2003; Ballard et al. common food items for waterfowl species 2004; Dugger et al. 2007; J. Coluccy et al., will increase precision and accuracy Ducks Unlimited, unpubl. data). There is of energetic carrying capacity estimates. uncertainty associated with applying a TME Alternatively, because such comprehensive value to a species other than the one from analyses may be impractical, future tests that which it was derived. However, closely extrapolate partial knowledge may be of use. related bird species will likely have similar For example, researchers could derive TME TME values for a given food item due to values for a single plant seed species for a similarities in gut morphology. While this range of duck species that differ in body hypothesis needs to be tested, substituting size and diet (Green-winged Teal A. crecca, TME values among similar bird species may American Wigeon A. americana, Northern suffice when evaluating energy content of Shoveler A. clypeata, Mallard) to yield insight food items until species-specific TME values into the extrapolation and applicability of become available. TME values for species that have not been Another problem with the lack of included in TME experiments. Alternatively, information on TME values is that existing researchers could relate TME to nutritional studies have typically determined TME composition of the seeds because values for only a few seed or invertebrate digestibility of multi-species forage is species across a select few families. related to the amount of indigestible fibre. Therefore, researchers are forced to “fill in Additionally, researchers may evaluate if the gaps” by assigning TME values based on known TME values of common waterfowl educated guesses with information from few foods can be predicted from gross energy studies. For example, many JVs use a mean (GE) estimates of these foods (e.g.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 416 Waterfowl carrying capacity models

Kaminski et al. 2003). If GE would explain technology, the availability of higher significant variation in TME (e.g. ≥ 70%) resolution imagery, the integration of NWI such models may be used cost-effectively data with other geospatial data sources for habitat conservation planning and (e.g. LiDAR, soil maps, etc.) and the implementation. development of standardised techniques for wetland identification and delineation have Extrapolating energy supply to the substantially improved the NWI (Dahl et al. landscape 2009; Knight et al. 2013). However, NWI Because of the significant potential for data are currently only available for 89% of biological and sampling error discussed continental United States, and the average above, the notion of extrapolating estimated date of the NWI for most of the U.S. is useable energy to a landscape level should from the 1980s. Currently, NWI data are be approached with caution. Inherently, the being updated at a rate of ≤ 2% per year due potential for multiplication of errors may to funding reductions (J. Coluccy, Ducks cause landscape-level variance to be too Unlimited, pers. comm.). Similar limitations large to make meaningful management exist for quantifying non-wetland habitat recommendations. However, an equally types (e.g. county crop data or seasonally- difficult problem is to quantify correctly flooded cropland). These limitations have what habitat is actually available to hampered efforts to estimate habitat waterfowl species. availability accurately for wintering waterfowl, Various geospatial data have been used to especially at frequencies desired for quantify characteristics of important maintaining a contemporary understanding waterfowl habitats (e.g. the National of the landscape carrying capacity. Wetlands Inventory (NWI) and National In addition to error associated with Land Cover Data). However, there are estimating wetland habitats correctly, inherent limitations associated with the biologists need to consider potential indirect accuracy of these data. For example, the impacts of human developments and NWI established by the U.S. Fish and disturbance that make available habitat Wildlife Service to conduct a nationwide avoided and thus reduce carrying capacity inventory of wetlands by type is widely (Korschgen & Dahlgren 1992; St. James et used for quantifying the availability of al. 2013). Often, estimates of available wetlands on staging and wintering areas. habitat are quickly calculated with the Unfortunately, the NWI does not capture assumption that all wetlands are created and classify all wetlands accurately because equal and waterfowl have unimpeded access NWI maps are derived from aerial photo- to them. However, avoidance behaviour may interpretation with varying limitations due occur at various temporal and spatial scales, to scale, photograph quality, inventory ranging between not settling in an area techniques and other factors (Federal and/or not utilising a space to its maximum Geographic Data Committee 2008; Dahl potential (Korschgen & Dahlgren 1992; et al. 2009). Recent advances in GIS Laundré et al. 2010; Hine et al. 2013). While

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 417 some waterfowl species such as Canada capacity requires an estimation of the Geese and Mallard may have adapted well to energy needs of individual waterfowl on any an altered landscape, other species of special given location and day. Energy requirements concern appear to be particularly sensitive of a wild vertebrate (or Daily Energy to disturbance (Korschgren & Dahlgren Expenditure, DEE) are usually estimated as 1992) including Atlantic Brant Branta the sum of the energy costs of maintenance bernicla, American Black Ducks, Canvasback (or Resting Metabolic Rate, RMR), activity Aythya valisineria and Lesser Scaup Aythya and thermoregulation (King 1973; Servello affinis. Additionally, ducks with a smaller et al. 2005; Fig. 1). If the animal is growing body size and shorter longevity (e.g. Green- or reproducing, then an additional energy winged Teal) than larger species (e.g. cost associated with this production must be Mallard) may take more risks to forage or added (Fig. 1). If the animal is not in a otherwise use wetlands that are hunted, and steady state and storing energy, then this thus exhibit reduced avoidance behaviour cost also must be added to the daily energy (St. James 2011). costs for the animal (Fig. 1). Most For planning purposes, it is important for bioenergetics approaches currently used by resource managers to understand how JVs for estimating DEE of waterfowl do waterfowl separate themselves from not account for the energy costs of anthropogenic development and respond to thermoregulation, production and storage, disturbance, and how these factors influence thereby reducing DEE to an estimate of their ability to extract critical food resources energy expenditure through daily activities from habitats. For example, on Lake St. Clair (Fig. 1). We will summarise current views on between Michigan and Ontario, autumn- estimating resting metabolic rate, the staging diving ducks shifted their use of energetic costs of daily activity and the costs traditional feeding and loafing areas on the of thermoregulation. U.S. side of the lake to new areas on the Canadian side (Shirkey 2012). Warmer Resting metabolic rate weather and associated increased angling The RMR is usually defined as energy costs and hunting activity on the U.S. side of maintaining an animal’s physiological throughout autumn and early winter is systems under a certain, restricted set of considered to have resulted in a significant conditions including: 1) no activity, 2) no shift in habitats important to Canvasback cost of staying warm or cool when and Lesser Scaup (Shirkey 2012). Diving measured at ambient temperatures that are ducks had options at this location, but within the thermoneutral zone of the obviously disturbance is an important animal, 3) no growth or reproduction, and management consideration at staging sites. 4) the bird is at steady state and not accumulating any fat or protein reserves or Energy demand depleting them. Because of these restrictive In addition to knowing energy availability in conditions, there have been few replicated the landscape, calculating habitat carrying RMR studies for duck species. However,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 418 Waterfowl carrying capacity models

A simple model of bioenergecs Daily Energy Acvity Expenditure DEE = 3 x RMR Resng Metabolic Rate RMR = 457 x (body mass)0.77

+ Thermoregulaon

Producon Storage (reproducon, growth)

Figure 1. A conceptual bioenergetics model of the key components that comprise Daily Energy Expenditure (DEE) of a wild vertebrate. DEE is usually estimated as the sum of the energy costs of maintenance (or Resting Metabolic Rate, RMR, calculated from body mass; Miller & Eadie 2006), times 3 to account for activity. Most bioenergetics approaches do not account for the energy costs of thermoregulation, production and storage, thereby reducing DEE to an estimate of energy expenditure through daily activities. there is established literature on the considered. Miller & Eadie (2006) used all allometric scaling of RMR of a wide variety available data from waterfowl RMR to of animals (Kleiber 1932, 1961; Prince 1979; update estimates of the slope (b) and Schmidt-Nielsen 1984; Peters 1983; Calder intercept (a) parameters for waterfowl. For 1984; Brown & West 2000) following: JV carrying capacity modelling, this provides a relatively straightforward method RMR = a(Mass)b (2) to estimate RMR for this component of where a = a mass proportionality bioenergetics modelling (but see caveats in coefficient, Mass = body mass (kg), and b = Miller & Eadie 2006). slope of the regression line on a log scale. In general, the accuracy of RMR predictions Activity expenditures from body size improves substantially when Daily Energy Expenditure is based on the subsets of species such as waterfowl are previously estimated RMR times the cost of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 419 activity. Most bioenergetics models used by factorial increases in energy expenditure JVs calculate DEE as over RMR, a time budget can be converted into estimates of energy expenditure or an DEE = RMR × 3 (3) energy budget (Albright et al. 1983; Paulus with the multiplier “3” accounting for the 1988). Therefore, equation (3) can be average amount of energy expended on expanded to account for multiple activities: activity in any given day (King 1974; Prince n 1979; Miller & Eadie 2006). However, if DEE = [](RMR  a i )Ti (4) ducks are storing energy, then this adds i=1 further energy costs (e.g. Heitmeyer 2006 where ai = the activity-specific factorial used a 3.4 multiplier for Mallard in winter). increase in RMR for the ith behavioural

While this methodology provides an activity and Ti = the proportion of time estimate of DEE, there are questions as to engaged in the activity within the 24-hr whether this simple approach accurately cycle. For example, Wooley (1976) added accounts for variability in behaviour due to heart rate monitors to five Black Ducks and external variations (e.g. temperature, tide, estimated multiplier values of 1.7 for time of day, month, latitude, harvest feeding, 1.2 for resting, 2.1 for comfort, 2.2 pressure, disturbance, etc.) that are known to for swimming, 2.2 for alert, 12.5 for flying, influence both daily activities and DEE 1.7 for walking, 2.4 for agonistic and 2.4 for (Weathers 1979; Albright et al. 1983; courtship. Finally, the activity-specific

Brodsky & Weatherhead 1985; Morton et al. factorial increase in RMR, ai, was estimated 1989). Miller & Eadie (2006) demonstrated by Wooley (1976). However, due to the that estimates of carrying capacity were crude nature of this study as compared to a highly sensitive to the multiple of RMR more controlled respirometry study, the used, as well as the mass proportionality validity of the estimates has been queried. coefficient (a) from the allometric equation. For example, how could comfort behaviour Thus, the use of a single multiplier be energetically more taxing than feeding? If for adjusting RMR is likely an these estimates are biased, the implications oversimplification. Depending on the for scaling up to the population level duck- sophistication of the planning process, JVs use days could be flawed. Additional studies could use more refined estimates of activity that measure BMR accurately for many costs derived from measured time-activity waterfowl species (see McKechnie & Wolf budgets of waterfowl in a given area for 2004), and consider how environmental which DEE is to be estimated (see Weathers factors including cold ambient temperatures et al. 1984; Miller & Eadie 2006). Time- affect BMR and DEE (e.g. McKechnie 2008; activity budgets rely on extensive McKinney & McWilliams 2005), can behavioural observations to determine a provide more precise estimates of DEE for time budget or the percentage of time free- waterfowl. living individuals spend in different Estimating DEE through time-energy behavioural states. Using behaviour-specific budgets is labour intensive, time consuming,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 420 Waterfowl carrying capacity models assumes random observability, does not ducks, closer to the estimated 75,000 Black account for energetically taxing flight Ducks estimated by mid-winter surveys behaviour and is historically limited (USFWS MBDC 2014). to diurnal observations (Jorde & Owen 1988). Recently, Jones (2012) conducted Cost of thermoregulation a comprehensive analysis of different The assumption that waterfowl are not measures of daily activity energy incurring energy costs of thermoregulation, expenditures using American Black Ducks production and storage requires some in coastal New Jersey. Black Duck behaviour scrutiny. Waterfowl and all endotherms (including flying) was quantified during must increase energy expenditure when morning crepuscular, diurnal, evening ambient temperature is below the lower crepuscular and nocturnal periods to create critical temperature (LCT) or above the a 24-hr time-energy budget. Behaviours upper critical temperature (UCT) of the and energy expenditure differed between animal (Fig. 2): periods and months, with greatest hourly m *ΔT (5) energy expenditure during the morning c LCT-Ta crepuscular period and lowest during the where mc is the slope of increasing metabolic nocturnal period. Additionally, precipitation, energy above the lowest critical temperature Δ temperature and tide influenced variation (LCT) and TLCT-Ta is the difference in in Black Duck behaviours over the ambient temperature from the lowest critical 24-hr period. Moreover, anthropogenic temperature. This allows for an expansion disturbance factors influenced behaviour of equation (4): including increased feeding during diurnal n and nocturnal periods on areas open to DEE = ()(RMR  a i )+CT Ti  (6) hunting when the hunting season was i=1 closed, and increased resting on areas closed where CT = the cost of thermoregulation at to hunting regardless of whether the a specified temperature (kJ/bird/h) in hunting season was open. As a result of the addition to activity-specific increases to detailed time energy budget, DEE was RMR. The UCT is rarely considered, as estimated as being 1,218 ± s.e. 19.36 most temperate waterfowl avoid regions kJ/bird/day, or 2.4 times RMR. The same with temperatures above the UCT. estimate calculated using RMR times 3 was However, waterfowl during winter may 21% greater at 1,545 kJ/bird/day, or 21% often encounter periods when the ambient higher. If one were to apply both estimates temperature is below their LCT. For to 101,017 ha of New Jersey coastal habitat example, the LCT of brant is 7.5°C for which energy supplies were estimated (Morehouse 1974); below this ambient (Cramer et al. 2012), autumn carrying temperature, brant would expend additional capacity would be ~55,000 ducks while the energy to stay warm. Generally, because 24-hr time energy budget estimates would thermal conductivity of water is 23 times predict a carrying capacity of ~70,000 greater than that of air, waterfowl sitting on

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 421

9 Costs of thermoregulation

8 Active Thermoneutral zone heating 7 (TNZ): Range of temperatures with

) 6

–1 minimum metabolism h

–1 5 LCT UCT 4 (ml g 2 O

V 3 ENDOTHERM 2 TNZ Passive cooling 1 ECTOTHERM 0 0 10 20 30 40 Ambient temperature (°C)

Figure 2. Any endothermic animal incurs additional energy costs when ambient temperatures drop below their lower critical temperature (LCT). These additional costs are linearly related to ambient temperatures below the LCT and are directly a function of the insulative properties of the animal. From Hiebert & Noveral (2007). water would need to expend much more there is a need for empirical studies of the energy to stay warm compared to birds energy costs of thermoregulation in other sitting on shore at the same ambient air waterfowl species to provide a strong temperature. While Richman & Lovvorn foundation for such estimates. (2011) did not find noticeable differences in energy costs for Common Eiders Somateria Alternative modelling mollissima, in cold water or air, McKinney & frameworks McWilliams (2005) estimated that the energy costs of thermoregulation in water While addressing sources of variation could contribute as much as 13–23% associated with energy supply and demand is of DEE for Bufflehead Bucephala albeola fundamental to reliable bioenergetics during winter in southern New England. models, there is also value in considering if Therefore, if DEE is estimated for alternative modelling frameworks could waterfowl during winter when temperatures improve conservation planning for wintering are typically below the LCT, including the waterfowl. The DRMs have provided a explicit energy cost of thermoregulation will useful approach for estimating bioenergetic reduce bias in DEE estimates. However, needs and landscape carrying capacity for

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 422 Waterfowl carrying capacity models waterfowl in winter for most non-breeding acquire energy reserves in preparation for JVs in North America for several reasons: migration), and interaction with other 1) they are well-established and widely fitness-maximising strategies (e.g. risk accepted, 2) they provide a tool for aversion, courtship and mate defence), translating waterfowl population objectives generally are not included in DRMs. A into habitat-based objectives, which is further limitation of current DRMs is that essential to accomplish NAWMP continental they consider only energy and not other goals, 3) they are based on data that can be nutrient resource needs from foraging obtained and validated from field surveys/ habitat. Yet other habitats, such as roosting research (e.g. food abundance and daily or refuge sites with better thermal energy demand), and 4) they allow managers characteristics, reduced disturbance or fewer and planners to evaluate the effect of large- predators, may also be important for reliable scale habitat changes on the availability of conservation planning. Finally, the suite of food resources and so provide an ability to response variables in DRMs is limited, undertake scenario planning (Central Valley resulting in a conglomerate energy supply JV Implementation Plan 2006). and summed population energy demand However, there are a number of that yields a surplus or deficit determined on limitations of the DRMs. The DRMs are the habitat base available. not spatially explicit, because they assume One of the challenges for JVs concerned no cost of travelling between food patches, with science-based habitat conservation for and food availability is considered to be non-breeding waterfowl is to develop a relatively uniform across the landscape. The measure of how achieving habitat objectives set of simplifying assumptions incorporated ultimately affects waterfowl demographic into most DRMs (e.g. ideal free foragers) parameters. Current DRMs do not make precludes consideration of how habitat this link, which limits the ability of habitat heterogeneity or bird distribution patterns managers to integrate planning models with (spatially or temporally) influence carrying demographic models that predict effects capacity. Likewise, in DRM, energy demand of regional actions on waterfowl dynamics is summed over all individuals regardless of at the continental scale. An alternative sex and age, usually over extended time approach to DRMs is use of agent-based periods (bi-weekly), and with DEEs that are models (hereafter referred to as ABMs). assumed to be invariant over time (fixed Unlike the top-down population-based energy costs). In some cases, JVs will sum approach of DRMs (summed energy DEE across all species based on an average demand and supply functions across species or representative body size. and habitats), ABMs instead represent a Changes in energy expenditures “bottom-up” approach where systems are throughout the non-breeding season, due modelled as collections of unique to changing food availability, temporal individuals or “agents”. Unlike more differences in thermoregulatory and formal mathematical population models, in individual state-strategies (e.g. the need to ABMs the system dynamics emerge from

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 423 interactions of individuals and their ABM referred to as SWAMP (Spatially- environment. Models are solved by explicit Waterbird Agent-based Modeling simulation instead of analytical solutions. Program) to model overwintering waterfowl Given plausible and realistic “rules” of bioenergetics (Miller et al. 2014; K. interaction and behaviour, the outcome is Ringelman et al., unpubl. data; J. Eadie et al., determined by simulation of how the agents unpubl. data). This model is intended to in the model respond according to the rules provide similar functionality as DRMs, and parameters defined. although it is one of the few ABMs The ABMs are now beginning to see specifically designed for use in landscape- broad application in conservation fields, based conservation management. The first although they have been used in ecological prototype has been tested, validated and research for well over two decades peer-reviewed (Miller et al. 2014). (DeAngelis & Gross 1992; Sutherland & Allport 1994; Goss-Custard et al. 2003; What are the advantages of an agent- Grimm & Railsback 2005; Stillman 2008). based approach? McLane et al. (2011) recently provided a Why might we need a potentially more comprehensive review of use and utility of complex approach to plan for habitat needs ABMs in wildlife ecology and management. of migrating and wintering waterfowl? Several ABMs have been developed that Conceptually and pragmatically, exploring could provide a platform for modification and developing an ABM approach includes and use by the NAWMP community to nine potential advantages. First, ABMs model and plan for waterfowl and waterbird link the behaviour of individuals with use of managed wetlands. Stillman’s population- or community-level processes – MORPH programme (Stillman 2008; West “scaling upwards”. This provides an et al. 2011) provides a foraging model that opportunity to incorporate individual has been applied to coastal birds in estuarine variation among birds and different environments and has broad potential as a populations in different locations or over platform for extension to other waterbirds time. It reveals local effects that could have and waterfowl. Pettifor et al. (2000) larger scale impacts (local sinks) that might developed a spatially explicit, individual- allow planners to target conservation efforts based, behavioural model to examine the more effectively. Second, ABMs provide the annual cycle of migratory geese. Mathevet ability to model individual and population et al. (2003) developed an ABM model as performance metrics of ultimate interest to a management tool for waterfowl JVs, such as body condition and survival. conservation incorporating farming and Thus, an ABM approach may enable hunting practices in France. However, they functional integration of habitat changes on did not model duck energetics explicitly and wintering and staging areas with regional and instead relied on a spatially-located DRM. continental demographic impacts. Third, Most recently, the Eadie & Shank research ABMs provide a more mechanistic structure group at UC Davis developed a prototype for foraging behaviour and dynamics that

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 424 Waterfowl carrying capacity models allows for inclusion of more biologically advantages and potential new insights, we realistic behaviour (e.g. foraging rules, patch have identified at least three challenges of choice criteria, flocking, etc.) than a generic implementing an ABM approach. First, the DRM that subsumes and potentially richness of possible combinations of confounds a large amount of important deterministic and stochastic processes can biology. Fourth, ABMs permit spatially make it daunting to simulate through a explicit analysis of the effects of alternative sufficient range of scenarios to be confident management regimes on the area and of generating results with a high level of geographic distribution of wetland habitats. repeatability and generality. Sensitivity Fifth, ABMs provide a useful tool for cross- analyses are essential, and careful thought boundary JV planning. The condition of and description (and some simplification) birds in one JV region might influence their are necessary, especially in the early decision about when to migrate to another development of a model. Second, ABMs region, and the condition of birds when they can be complex with a large number of arrive in a JV region might influence how parameters and functions to be estimated. long they stay. The ABMs provide a method This complexity deterred progress with by which to track and link body condition ABM, because models were specific to a and movement of birds across larger spatial particular situation, not transparent, and not scales. Sixth, the ABMs allow planners to easily communicated or vetted by the model large numbers (millions) of birds in scientific and management communities real time, and on large and small spatial (Grimm & Railsback 2005). However, the scales, using GIS layers as input. Seventh, field has become sufficiently advanced ABMs can incorporate other important and well-defined protocols have been determinants of habitat use and carrying established (e.g. Grimm et al. 2006, 2010). capacity such as disturbance and dispersion Yet, many parameters are unknown or based of non-foraging (refuge) habitat. Eighth, on expert opinion. The ABMs are not alone ABMs offer the potential to expand the in this regard; traditional DRMs also include capacity to generalise across taxa, including a large number of variables and can be waterfowl, shorebirds and other wetland- extremely complex (e.g. the TRUEMET dependent wildlife. Ninth, we can use bioenergetic modelling application includes these models to integrate more directly up to 77 time-dependent and 41 time- and completely with existing models of independent variables). Lastly, ABMs, like water management, in-stream fish habitat, DRMs, require a strong foundation of urban growth, and other spatially-based empirical data upon which to base the conservation issues (J. Fleskes, U.S. model and validation. Hence, the sampling Geological Survey, pers. comm.). and calculation challenges discussed above for DRMs also apply to ABMs. The ABMs Challenges of an agent-based approach allow for heterogeneity in many of these Agent-based models are not a panacea for parameters (e.g. DEE as a function of daily conservation planning. While they offer mass). Although integrating heterogenity is

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 425 more realistic, it can also generate increased conditions to key individual (e.g. body variation in predicted outcomes. As with any condition) and population performance (e.g. modelling exercise, the quality of output is survival) metrics. dictated by the quality and validity of the Admittedly, there is a trade-off between parameters and functions used as input. complexity and utility, and we need to be Field and laboratory research to refine vigilant against creating increasingly parameters, estimate unknown parameters complex models simply because it is and validate outputs are essential. possible. However, the ability to address Given these challenges, do we really need uncertainty in a more formal and explicit more complex models? Goss-Custard et al. manner, the value of a mechanistic (2003) tested carrying capacity predictions approach that potentially can link habitat to between a DRM and a spatial depletion demography, and the ability to consider model (SDM) and concluded that smaller-scale spatially-explicit planning for predictions were similar under certain waterfowl conservation (e.g. by providing circumstances. Comparisons of the non-foraging habitat and refuge areas), output from SWAMP with TRUEMET provide a strong argument to continue corroborate this similarity (K. Ringelman developing and learning from ABM et al., unpubl. data). Hence, DRMs may be approaches. sufficient for basic determination of foraging habitat needs, with the caveat that Future research and DRMs may not reflect energy balance management considerations perfectly and uncertainty may be high. So, when do we need spatially explicit Research has made substantial advances to behaviour-based models? We envision address empirical uncertainties and test several conditions when an ABM approach assumptions of JV biological planning models would be valuable, when: 1) time and energy over the past several decades. Such research costs of foraging differ between patches, has greatly increased our confidence in the 2) cost of movement between patches is utility of DRMs as a conservation planning great and varies over time (e.g. with patch tool under certain applications. However, depletion close to refuges), 3) sequence of it has also revealed a need to embrace use of patches varies but is important, more formal means of identifying model 4) distance from roosting sites or refuges to components and parameters and refining foraging patches varies and changes over them, while also considering alternative time, 5) juxtaposition and location of modelling frameworks that may produce habitats are important, 6) non-foraging more reliable conservation planning strategies. habitat is important and managers need to Perturbation Analyses or Structural Decision assess consequences and interactions of Making are useful tools for considering such disturbance, sanctuary, predation and public uncertainty (Caswell 2001; Hoekman et al. access, among other non-foraging needs, 2002, 2006; Coluccy et al. 2008), and managers and 7) there is a need to link habitat planning the conservation of wintering

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 426 Waterfowl carrying capacity models waterfowl may want to consider expanding uncertainties and requirements for scientific their use of such tools. investigations to parameterise, evaluate and During our special session and this refine them. Thus, deliberate considerations review, we identified a subset of key must be made to determine if the advantages simplifying assumptions common to DRMs, of more complex models outweigh their some of which deserve additional scrutiny greater financial and logistical costs. An and refinement (e.g. constant BMR initial approach could include developing an multipliers). With few exceptions, potential ABM at a restricted yet biologically relevant implications of these assumptions and spatial scale, followed by a comparison model uncertainties for habitat conservation of their outputs to those generated from objectives and priorities have not been more simplistic, deterministic models as a quantified (cf. Miller & Eadie 2006). means of assessing degrees of dissimilarity Variability, whether originating from natural in how alternative models describe processes or sampling strategies, is an desired landscape conditions. Logically, integral part of any biological system being sophisticated models with outputs only modelled (e.g. Saether et al. 2008), yet marginally different from simpler models most DRMs currently being used are would likely not be worthy of adoption (e.g. deterministic. Even in the few cases where Goss-Custard et al. 2003). Moreover, because quantitative analyses of variability have conservation plans are implemented by the occurred, the results have not yet been broader JV community, understanding of widely incorporated into conservation key models and model-based conservation recommendations or used to prioritise recommendations is essential for maximum future investment in science. Thus, a partner engagement and support. fundamental question is whether habitat Finally, the challenges associated with objectives arrived at using models based developing and refining better biological solely on measures of central tendency models are likely to increase as JVs explore (e.g. mean population abundance, mean strategies to achieve the integrated goals of TME values) can produce landscapes the current version of NAWMP (NAWMP necessary to achieve NAWMP goals (Straub 2012), which will likely require investments et al. 2012). in human dimensions that have until now Recent progress in our understanding of been at the periphery of JV activities. waterfowl foraging ecology and the Skillful evaluation of the costs and benefits successful application of ABMs to inform to include financial, ecological and habitat conservation for coastal waterbirds evaluation of alternative carrying capacity (e.g. Stillman & Goss-Custard 2010) are models and additional model refinements compelling JV conservation planners to should invoke prudent decisions about their consider the development and use of necessity. Just as the waterfowl management alternative frameworks for establishing community has applied rigorous science to habitat objectives. However, ABMs ensure efficient and effective expenditure of may be accompanied by even greater limited conservation dollars, so too should it

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 427 be the foundation for expenditure of Brodsky, L.M. & Weatherhead, P.J. 1985. Time and science resources, which are often far more energy constraints on courtship in wintering limited. American black ducks. Condor 87: 33–36. Brown, J.H. & West, G.B. (eds.) 2000. Scaling in Acknowledgements Biology. Oxford University Press, Oxford, UK. We wish to thank Joe Fleskes, Matt Miller, Calder, W.A., III. 1984. Size, Function, and Life Samantha Richman, Kevin Ringelman History. Harvard University Press, Cambridge, Massachusetts, USA. and Jeff Schank for reviews and advice Callicutt, J.T., Hagy, H.M. & Schummer, M.L. regarding the content of this paper. The use 2011. The food preference paradigm: a of trade, product, industry or firm names or review of autumn-winter food use by North products is for informative purposes only American dabbling ducks (1900–2009). Journal and does not constitute an endorsement by of Fish and Wildlife Management 2: 29–40. the U.S. Government or other sponsoring or Caswell, H. 2001. Matrix Population Models. Second participating agencies. edition. Sinauer, Sunderland, Massachusetts, USA. References Central Valley Joint Venture. 2006. Central Valley Albright, J.J., Owen, Jr., R.B. & Corr, P.O. 1983. Joint Venture Implementation Plan: Conserving The effects of winter weather on the Bird Habitat. U.S. Fish and Wildlife Service, behavior and energy reserves of black ducks Sacramento, California, USA. in Maine. Transactions of the Northeast Section of Checkett, J.M., Drobney, R.D., Petrie, M.J. & the Wildlife Society 40: 118–128. Graber, D.A. 2002. True metabolizable Anteau, M.J. & Afton, A.D. 2009. Lipid reserves energy of moist-soil seeds. Wildlife Society of lesser scaup (Aythya affinis) migrating Bulletin 30: 1113–1119. across a large landscape are consistent with Coluccy, J.M., Yerkes, T., Simpson, R., Simpson, the “spring condition” hypothesis. The Auk J.W., Armstrong, L. & Davis, J.I. 2008. 126: 873–883. Population dynamics of breeding mallards in Baldassarre, G.A. & Bolen, E.G. 1984. Field- the Great Lakes states. Journal of Wildlife feeding ecology of waterfowl wintering in Management 72: 1181–1187. the Southern High Plains of Texas. Journal of Cramer, D.M., Castelli, P.M., Yerkes, T. & Williams, Wildlife Management 48: 63–71. C.K. 2012. Food resource availability for Ballard, B.M., Thompson, J.E., Petrie, M.J., American black ducks wintering in southern Checkett, M. & Hewitt, D.G. 2004. Diet and New Jersey. The Journal of Wildlife Management nutrition of northern pintails wintering along 76: 214–219. the southern coast of Texas. Journal of Dahl, T.E., Dick, J., Swords, J. & Wilen, Wildlife Management 68: 371–382. B.O. 2009. Data Collection Requirements and Bergan, J.F. & Smith, L.M. 1993. Survival rates of Procedures for Mapping Wetland, Deepwater and female mallards wintering in the Playa Lakes Related Habitats of the United States. Division Region. The Journal of Wildlife Management 57: of Habitat Resource Conservation, National 570–577. Standards and Support Team, Madison, Bousquet, F. & Le Page, C. 2004. Multi-agent Wisconsin, USA. simulations and ecosystem management: a DeAngelis D.L. & Gross, L.J. 1992. Individual- review. Ecological Modelling 176: 313–332. based models and approaches in ecology:

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 428 Waterfowl carrying capacity models

populations, communities and ecosystems. In Evans-Peters G.R., Dugger, B.D., Petrie, M.J. D.L. DeAngelis & L.J. Gross (eds.), Individual- 2012. Plant community composition and based Models and Approaches in Ecology. waterfowl food production on Wetland Chapman & Hall, New York, USA. Reserve Program easements compared to DeAngelis, D.L. & Mooij, W.M. 2005. Individual- those on managed public lands in western based modeling of ecological and Oregon and Washington. Wetlands 32: 391– evolutionary processes. Annual Review of 399. Ecology, Evolution, and Systematics 36: 147–168. Federal Geographic Data Committee. 2008. de Kroon, H., van Groenendael, J. & Ehrlen, J. Federal Geographic Data Committee Wetlands 2000. Elasticities: A review of methods and Inventory Mapping Standard. FGDC Wetland model limitations. Ecology 81: 607–618. Subcommittee and Wetland Mapping Delnicki, D. & Reinecke, K.J. 1986. Mid-winter Standard Workgroup. Reston, Virginia, food use and body weights of mallards and USA. Accessible at http://www.fgdc.gov/ wood ducks in Mississippi. The Journal of participation/working-groups-subcommittees/ Wildlife Management 50: 43–51. wsc/ (last accessed 18 August 2014). Devries, J.H., Armstrong, L.M., MacFarlane, R.J., Fleming, K.S., Kaminski, R.M., Tietjen, T.E., Moats, L. & Thoroughgood, P.T. 2008. Schummer, M.L., Ervin, G.N. & Nelms, K.D. Waterfowl nesting in fall-seeded and spring- 2012. Vegetative forage quality and moist-soil seeded cropland in Saskatchewan. Journal of management on Wetlands Reserve Program Wildlife Management 72: 1790–1797. lands in Mississippi. Wetlands 32: 919–929. Dolman, P.M. & Sutherland, W.J. 1995. The Foster, M.A., Gray, M.J., Harper, C.A. & Walls, response of bird populations to habitat loss. J.G. 2010a. Comparison of agricultural seed Ibis 137: S38–S46. loss in flooded and unflooded fields on the Dolman, P.M. & Sutherland, W.J. 1997. Spatial Tennessee National Wildlife Refuge. Journal patterns of depletion imposed by foraging of Fish and Wildlife Management 1: 43–46. vertebrates: Theory, review and meta-analysis. Foster, M.A., Gray, M.J. & Kaminski, R.M. Journal of Animal Ecology 66: 481–494. 2010b. Agricultural seed biomass for Dugger, B.D., Moore, M.L., Finger, R.S. & Petrie, migrating and wintering waterfowl in the M.J. 2007. True metabolizable energy for Southeastern United States. Journal of Wildlife seeds of common moist-soil plant species. Management 74: 489–495. Journal of Wildlife Management 71: 1964– Fredrickson, L.H. & Reid, F.A. 1988. 1967. Nutritional values of waterfowl foods. Dugger, B.D., Petrie, M.J. & Mauser, D. 2008. A Waterfowl Management Handbook – Fish and Bioenergetic Approach to Conservation Planning for Wildlife Leaflet 13.1.1. U. S. Fish and Wildlife Waterfowl at Lower Klamath and Tule Lake Service, Washington D.C., USA. National Wildlife Refuge. Technical Report Gill, J.A., Sutherland, W.J. & Norris K. 2001. submitted to United States Fish and Wildlife Depletion models can predict shorebird Service. US Fish & Wildlife Service, Tulelake, distribution at different spatial scales. California, USA. Proceedings of the Royal Society Biological Sciences Evans-Peters, G.R. 2010. Assessing biological Series B 268: 369–376. values of Wetlands Reserve Program wetlands Goss-Custard, J.D., Clarke, R.T., Briggs, K.B., for wintering waterfowl. M.Sc. thesis, Oregon Ens, B.J., Exo, K.M., Smit, C., Beintema, A.J., State University, Corvallis, USA. Caldow, R.W.G., Catt, D.C., Clark, N.A., Dit

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 429

Durell, S.E.A.L.V., Harris, M.P., Hulscher, waterfowl. The Journal of Wildlife Management J.B., Meininger, P.L., Picozzi, N., Prys-Jones, 71: 1561–1566. R., Safriel, U.N. & West, A.D. 1995. Greer, D.M., Dugger, B.D., Reinecke, K.J. & Population consequences of winter habitat Petrie, M.J. 2009. Depletion of rice as food loss in a migratory shorebird. I. Estimating of waterfowl wintering in the Mississippi model parameters. Journal of Applied Ecology Alluvial Valley. Journal of Wildlife Management 32: 320–336. 73: 1125–1133. Goss-Custard, J.D., West, A.D., Clarke, R.T., Grimm, V. 1999. Ten years of individual-based Caldow, R.W.G & Dit Durell, S.E.A.L.V. modeling in ecology: what have we learned 1996. The carrying capacity of coastal and what could we learn in the future? habitats for oystercatchers. In J. D. Goss- Ecological Modeling 115: 129–148. Custard (ed.), The Oystercatcher: from Individuals Grimm, V. & Railsback, S.F. 2005. Individual-Based to Populations. Oxford Ornithology Series, Modeling and Ecology. Princeton University Oxford University Press, Oxford, UK. Press, New Jersey, USA. Goss-Custard, J.D., Stillman, R.A., West, A.D., Grimm, V., Revilla, E., Berger, U., Jeltsch, F., Caldow, R.W.G. & McGrorty, S. 2002. Mooij, W.M., Railsback, S.F., Thulke, H.H., Carrying capacity in overwintering migratory Weiner, J., Wiegand, T. & DeAngelis, D.L. birds. Biological Conservation 105: 27–41. 2005. Pattern-oriented modeling of agent Goss-Custard, J.D., Stillman, R.A., Caldow, based complex systems: lessons from ecology. R.W.G. & West, A.D. 2003. Carrying capacity Science 310: 987–991. in overwintering birds: when are spatial Grimm,V., Berger, U., Bastiansen, F., Eliassen, S., models needed? Journal of Applied Ecology 40: Ginot, V., Giske, J., Goss-Custard, J., Grand, 176–187. T., Heinz, S.K., Huse, G., Huth, A., Jepsen, Gray, M.J., Kaminski, R.M. & Brasher, M.G. J.U., Jørgensen, C., Mooij, W.M., Müller, B., 1999a. A new method to predict seed yield of Pe’er, G., Piou, C., Railsback, S.F., Robbins, moist-soil plants. Journal of Wildlife Management A.M., Robbins, M.M., Rossmanith, E., Rüger, 63: 1269–1272. N., Strand, E., Souissi, E., Stillman, R.A., Gray, M.J., Kaminski, R.M., & Weerakkody, G. Vabø, R., Visser, U. & DeAngelis, D.L. 2006. 1999b. Predicting seed yield of moist-soil A standard protocol for describing plants. Journal of Wildlife Management 63: individual-based and agent-based models. 1261–1268. Ecological Modelling 198: 115–126. Gray, M.J., Foster, M.A. & Peña Peniche, Grimm, V., Berger, U., DeAngelis, D.L., Polhill, L.A. 2009. New technology for estimating J.G., Giske, J. & Railsback, J.S.F. 2010. The seed production of moist-soil plants. Journal ODD protocol: A review and first update. of Wildlife Management 73: 1229–1232. Ecological Modeling 221: 2760–2768. Gray, M.J., Hagy, H.A., Nyman, J.A. & Stafford, Guillemain, M., Elmberg, J., Arzel, C., Johnson, J.D. 2013. Management of wetlands for A.R. & Simon, G. 2008. The income- wildlife. In C.A. Davis & J.T. Anderson capital breeding dichotomy revisited: late (eds.), Wetland Techniques, Volume 3. Springer, winter body condition is related to breeding Secaucus, New Jersey, USA. success in an income breeder. Ibis 150: 172– Greer, A.K., Dugger, B.D., Graber, D.A. & Petrie, 176. M.J. 2007. The effects of seasonal flooding Hagy, H.M. 2010. Winter food and waterfowl on seed availability for spring migrating dynamics in managed moist-soil wetlands in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 430 Waterfowl carrying capacity models

the Mississippi Alluvial Valley. Ph.D. thesis. Hoekman, S.T., Mills, L.S., Howerter, D.W., Mississippi State University, Mississippi State, Devries, J.H. & Ball, I.J. 2002. Sensitivity USA. analyses of the life cycle of midcontinent Hagy, H.M. & Kaminski, R.M. 2012a. Winter mallards. Journal of Wildlife Management 66: waterbird and food dynamics in autumn- 883–900. managed moist-soil wetlands of the Hoekman, S.T., Gabor, T.S., Petrie, M.J., Maher, Mississippi Alluvial Valley. Wildlife Society R., Murkin, H.R. & Lindberg, M.S. 2006. Bulletin 36: 512–523. Population dynamics of Mallards breeding in Hagy, H.M. & Kaminski, R.M. 2012b. Apparent agricultural environments in eastern Canada. seed use by ducks in the Mississippi Alluvial Journal of Wildlife Management 70: 121 –128. Valley. Journal of Wildlife Management 76: Hoffman, R.D. & Bookhout, T.A. 1985. 1053–1061. Metabolizable energy of seeds consumed by Hagy, H.M., Straub, J.N. & Kaminski, R.M. 2011. ducks in Lake Erie marshes. Transactions of the Estimation and correction of seed recovery North American Wildlife and Natural Resources bias from moist-soil cores. Journal of Wildlife Conference 50: 557–565. Management 75: 959–966. Jones III, O.E. 2012. Constructing a 24 hour Heitmeyer, M.E. 1988. Body composition of time-energy budget for American black female mallards in winter in relation to ducks wintering in coastal New Jersey. M.Sc. annual cycle events. Condor 90: 669–680. thesis, University of Delaware, Newark, USA. Heitmeyer, M.E. 2006. The importance of winter Jorde, D.G. & Owen, Jr., R.B. 1988. Efficiency floods to mallards in the Mississippi Alluvial of nutrient use by American black ducks Valley. Journal of Wildlife Management 70: wintering in Maine. Journal of Wildlife 101–110. Management 59: 209–214. Heitmeyer, M.E. & Fredrickson, L.H. 1981. Do Kaminski, R.M. & Essig, H.W. 1992. True wetland conditions in the Mississippi Delta metabolizable energy estimates for wild and hardwoods influence mallard recruitment? game-farm mallards. Journal of Wildlife Transactions of the North American Wildlife and Management 56: 321–324. Natural Resources Conferences 46: 44–57. Kaminski, R.M. & Gluesing, E.A. 1987. Density- Hepp, G.R. & Hines, J.E. 1991. Factors affecting and habitat-related recruitment in mallards. winter distribution and migration distance of The Journal of Wildlife Management 51: 141–148. Wood Ducks from southern breeding Kaminski, R.M., Davis, J.B., Essig, H.W., populations. Condor 93: 884–891. Gerard, P.D. & Reinecke, K.J. 2003. True Hiebert, S.M. & Noveral, J. 2007. Are chicken metabolizable energy for wood ducks from embryos endotherms or ectotherms? A acorns compared to other waterfowl foods. laboratory exercise integrating concepts in Journal of Wildlife Management 67: 542–550. thermoregulation and metabolism. Advances Karasov, W.H. 1990. Digestion in birds: in Physiology Education. 31:97–109. chemical and physiological determinants and Hine, C.S., Hagy, H.M., Yetter, A.P., Horath, ecological implications. Studies in Avian Biology M.M., Smith, R.V. & Stafford, J.D. 2013. 13: 391–415. Waterbird and Wetland Monitoring at the Emiquon King, J.R. 1973. Energetics of reproduction in Preserve: Final Report 2007–2013. Illinois birds. In D.S. Farner (ed.), Breeding Biology Natural History Survey Technical Report of Birds. National Academy of Science, 2013 (20), Champaign, Illinois, USA. Washington, D.C., USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 431

King, J.R. 1974. Seasonal allocation of time Lovvorn, J.R. & Baldwin, J.R. 1996. Intertidal and energy resources in birds. Nuttall and farmland habitats of ducks in the Puget Ornithological Club Publication 15: 4–70. Sound region: a landscape perspective. Kleiber, M. 1932. Body size and metabolism. Biological Conservation 77: 97–114. Hilgardia 6: 315–353. Luke S., Cioffi-Revilla, C., Panait, L., Sullivan, K. Kleiber, M. 1961. The Fire of Life. An Introduction & Balan, G. 2005. MASON: A Multi-agent to Animal Energetics. Wiley, New York, USA. Simulation Environment. Simulation 81: Knight, J.F., Tolcser, B.T., Corcoran, J.M. & 517–527. Rampi, L.P. 2013. The effects of data Marty, J.R. 2013. Seed and waterbird abundances selection and thematic detail on the in ricelands in the Gulf Coast Prairies accuracy of high spatial resolution wetland of Louisiana and Texas. M.Sc. Thesis. classifications. Photogrammetric Engineering and Mississippi State University, USA. Remote Sensing 79: 613–623. Mathevet R, Bousquet, F.C., Le Page, C. & Korschgen, C.E. & Dahlgren, R.B. 1992. Human Antona, M. 2003. Agent-based simulations disturbances of waterfowl: causes, effects, of interactions between duck population, and management. Waterfowl Management farming decisions and leasing of hunting Handbook No. 13. U.S. Department of the rights in the Camargue (Southern France). Interior, Washington D.C., USA. Ecological Modelling 165: 107–126. Kross, J., Kaminski, R.M., Reinecke, K.J., Penny, McKechnie, A.E. 2004. The allometry of avian E.J. & Pearse, A.T. 2008a. Moist-soil seed basal metabolic rate: good predictions need abundance in managed wetlands in the good data. Physiological and Biochemical Zoology Mississippi Alluvial Valley. Journal of Wildlife 77: 502–521. Management 72: 707–714. McKechnie, A.E. 2008. Phenotypic flexibility in Kross, J., Kaminski, R.M., Reinecke, K.J. & basal metabolic rate and the changing view Pearse, A.T. 2008b. Conserving waste rice for of avian physiological diversity: a review. wintering waterfowl in the Mississippi Journal of Comparative Physiology B 178: 235– Alluvial Valley. Journal of Wildlife Management 247. 72: 1383–1387. McKinney, R.R.A. & McWilliams, S.T.R. 2005. Laubhan, M.K. & Fredrickson, L.H. 1992. A new model to estimate daily energy Estimating seed production of common expenditure for wintering waterfowl. The plants in seasonally flooded wetlands. Journal Wilson Bulletin 117: 44–55. of Wildlife Management 56: 329–337. McLane A.J., Semeniuk, C., McDermid, G.J. & Laundré, J.W., Hernandez, L. & Ripple, W.J. 2010. Marceau, D.J. 2011. The role of agent-based Landscape of fear: ecological implication of models in wildlife ecology and management. being afraid. The Open Ecology Journal 3: 1–7. Ecological Modelling 222: 1544–1556. Loesch, C.R. & Kaminski, K.M. 1989. Winter Miller, M.R. & Eadie, J.M. 2006. The allometric body-weight patterns of female mallards fed relationship between resting metabolic rate agricultural seeds. The Journal of Wildlife and body mass in wild wintering waterfowl Management 53: 1081–1087. (Anatidae) and an application to estimation Lovvorn, J.R. 1994. Nutrient reserves, probability of winter habitat requirements. The Condor of cold spells and the question of reserve 108:166–177. regulation in wintering canvasbacks. Journal of Miller, M.R. & Newton, W.E. 1999. Population Animal Ecology 63: 11–23. energetics of northern pintails wintering in

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 432 Waterfowl carrying capacity models

the Sacramento Valley, California. Journal of Naylor, L.W., Eadie, J.M., Smith, W.D., Eichholz, Wildlife Management 63: 1222–1238. M. & Gray, M.J. 2005. A simple method to Miller, M.R. & Reinecke, K.J. 1984. Proper predict seed yield in moist-soil habitats. expression of metabolizable energy in avian Wildlife Society Bulletin 33: 1335–1341. energetics. The Condor 86: 396–400. Nichols, J.D., Reinecke, K.J. & Hines, J.E. 1983. Miller M.L., Ringelman, K.M., Schank, J.C. Factors affecting the distribution of mallards & Eadie, J.M. 2014. SWAMP: An agent- wintering in the Mississippi Alluvial Valley. based model for wetland and waterfowl Auk 100: 932–946. conservation and management. Simulation 90: Nolet, B.A., Langevoord, O., Bevan, R.M., 52–68. Engelaar, K.R., Klaassen, M., Mulder, R.J.W. Moon, J.A. & Haukos, D.A. 2006. Survival of & Van Dijks, S. 2001. Spatial variation in female northern pintails wintering in the tuber depletion by swans explained by Playa Lakes Region of northwestern Texas. difference in net intake rates. Ecology 82: Journal of Wildlife Management 70: 777–783. 1655–1667. Moon, J.A., Haukos, D.A. & Smith, L.M. 2007. Nolet, B.A., Gyimesi, A. & Klaassen, R.H. 2006. Declining body condition of northern Prediction of bird-day carrying capacity on a pintails wintering in the Playa Lakes staging site: a test of depletion models. region. Journal of Wildlife Management 71: Journal of Animal Ecology 75: 1285–1282. 218–221. North American Waterfowl Management Plan Morehouse, K.A. 1974. Development, energetics Committee. 2012. North American Waterfowl and nutrition of captive Pacific Brant (Branta Management Plan: People Conserving Waterfowl bernicla orientalis, Tougarinov). Ph.D. thesis, and Wetlands. U.S. Fish and Wildlife Service, University of Alaska, Fairbanks, Alaska, USA. Washington D.C., USA. Morton, J.M., Fowler, A.C. & Kirkpatrick, R.L. Nudds, T.D. & Kaminski, R.M. 1984. Sexual 1989. Time and energy budgets of American size dimorphism in relation to resource black ducks in winter. Journal of Wildlife partitioning in North American dabbling Management 53: 401–410. ducks. Canadian Journal of Zoology 62: 2009– Murkin, H.R., Wrubleski, D.A. & Reid, F.A. 1996. 2012. Sampling invertebrates in aquatic and Olmstead, V.G. 2010. Evaluation of management terrestrial habitats. In T.A. Bookout (ed.), strategies on moist-soil seed availability Research and Management Techniques for Wildlife and depletion on Wetland Reserve Program and Habitats. Fifth edition. The Wildlife Society, sites in the Mississippi Alluvial Valley. Bethesda, Maryland, USA. M.Sc. thesis, Arkansas Tech University, National Ecological Assessment Team. 2006. Russellville, USA. Strategic Habitat Conservation – Final Report of Olmstead, V.G., Webb, E.B. & Johnson, the National Ecological Assessment Team. U.S. R.W. 2013. Moist-soil seed biomass and Geological Survey and U.S. Fish and Wildlife species richness on Wetland Reserve Program Service, Washington D.C., USA. Easements in the Mississippi Alluvial Valley. Naylor, L.W. 2002. Evaluating moist-soil seed Wetlands 33: 197–206. production and management in Central Pacific Coast Joint Venture. 2004. Pacific Coast Joint Valley Wetlands to determine habitat needs Venture: Coastal Northern California Component: for waterfowl. M.Sc. thesis, University of Strategic Plan. U.S. Fish and Wildlife Service, California, Davis, California, USA. Sacramento, California, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 433

Paulus, S.L. 1988. Time-activity budgets of Society. The Wildlife Society, Madison, nonbreeding Anatidae: a review. In M. W. Wisconsin, USA. Weller (ed.), Waterfowl in Winter, pp. 135–165 Proctor, J.R. & Marks, C.F. 1974. The University of Minnesota Press. Minneapolis, determination of normalizing transformations USA. for nematode count data from soil Pearse, A.T. & Stafford, J.D. 2014. Error samples and of efficient sampling schemes. propagation in energetic carrying capacity Nematologica 20: 395–406. models. Journal of Conservation Planning 10: Railsback, S.F & Grimm, V. 2012. Agent-based and 17–24. Individual-based Modeling: A Practical Introduction. Pearse, A.T., Kaminski, R.M., Reinecke, K.J. & Princeton University Press, New Jersey, USA. Dinsmore, S.J. 2012. Local and landscape Raveling, D.G. & Heitmeyer, M.E. 1989. associations between wintering dabbling Relationships of population size and ducks and wetland complexes in Mississippi. recruitment of pintails to habitat conditions Wetlands 32: 859–869. and harvest. Journal of Wildlife Management 53: Peters, R.H. 1983. The Ecological Implications 1088–1103. of Body Size. Cambridge University Press, Reinecke, K.J. & Hartke, K.M. 2005. Estimating Cambridge, UK. moist-soil seeds available to waterfowl with Petrie, M.J. 1994. True metabolizable energy of double sampling for stratification. Journal of waterfowl foods. M.Sc. thesis, University of Wildlife Management 69: 794–799. Missouri, Columbia, USA. Reinecke, K.J, Kaminski, R.M., Moorhead, D.J., Petrie, M.J., Drobney, R.D. & Graber, D.A. 1998. Hodges, J.D. & Nassar, J.R. 1989. Mississippi True metabolizable energy estimates of Alluvial Valley. In L.M. Smith, R.L. Pederson Canada goose foods. Journal of Wildlife & R.M. Kaminski (eds.), Habitat Management Management 62: 1147–1152. for Migrating and Wntering Waterfowl in North Petrie, M. J., Brasher, M.G., Soulliere, G.J., Tirpak, America. Texas Tech University Press, J.M., Pool, D.B. & Reker, R.R. 2011. Guidelines Lubbock, USA. for Establishing Joint Venture Waterfowl Population Richman, S.E. & Lovvorn, J.R. 2011. Effects of Abundance Objectives. North American air and water temperatures on resting Waterfowl Management Plan Science Support metabolism of auklets and other diving birds. Team Technical Report No. 2011-1, U.S. Fish Physiological and Biochemical Zoology 84: and Wildlife Service, Washington D.C., USA. 316–332. Pettifor, R.A., Caldow, R.W.G., Rowcliffe, J.M., Ringelman, K.M. 2014. Predator foraging Goss-Custard, J.D., Black, J.M., Hodder, behavior and patterns of avian nest success: K.H., Houston, A.I., Lang, A. & Webb, J. What can we learn from an agent-based 2000. Spatially explicit, individual-based, model? Ecological Modelling 272: 141–149. behavioural models of the annual cycle of Sæther, B.E., Lillegård, M., Grøtan, V., Drever, two migratory goose populations. Journal of M.C., Engen, S., Nudds, T. & Podruzny, Applied Ecology 37: 103–135. K.M. 2008. Geographical gradients in the Prince, H.H. 1979. Bioenergetics of postbreeding population dynamics of North American dabbling ducks. In T. A. Bookhout (ed.), prairie ducks. Journal of Animal Ecology 77: Waterfowl and Wetlands – An Integrated Review. 869–882. Proceedings of the 1977 Symposium of Santamaría, L. & Rodríguez-Gironés, M.A. 2002. the North Central Section of The Wildlife Hiding from swans: optimal burial depth of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 434 Waterfowl carrying capacity models

sago pondweed tubers foraged by Bewick’s Smith, R.V., Stafford, J.D., Yetter, A.P., Whelan, swans. Journal of Ecology 90: 303–315. C.J., Hine, C.S. & Horath, M.M. 2011. Schmidt-Nielsen, K. 1984. Scaling: Why is Animal Foraging thresholds of spring migrating Size so Important? Cambridge University Press, dabbling ducks in central Illinois. Illinois Cambridge, UK. Natural History Survey Technical Report Schroth, G. & Kolbe, D. 1994. A method of 2011(No. 37), Champaign, Illinois, USA. processing soil core samples for root studies Smith, R.V., Stafford, J.D., Yetter, A.P., Horath, by subsampling. Biology and Fertility of Soils 18: M.M., Hine, C.S. & Hoover, J.P. 2012. 60–62. Foraging ecology of fall-migrating Servello, F.A., Hellgren, E.C. & McWilliams, S.R. shorebirds in the Illinois River valley. PLoS 2005. Techniques for wildlife nutritional ONE 7(9):e45121. ecology. In C.E. Braun (ed.), Research and Soulliere, G.J., Potter, B.A., Coluccy, J.M., Gatti, Management Techniques for Wildlife and Habitats. R.C., Roy, C.L., Luukkonen, D.R., Brown, The Wildlife Society, Washington D.C., P.W. & Eichholz, M.W. 2007. Upper Mississippi USA. River and Great Lakes Region Joint Venture Sherfy, M.H. 1999. Nutritional value and Waterfowl Habitat Conservation Strategy. U.S. management of waterfowl and shorebird Fish and Wildlife Service, Fort Snelling, foods in Atlantic coastal moist-soil Minnesota, USA. impoundments. Ph.D. thesis, Virginia Soulliere, G.J., Loges, B.W., Dunton, E.M., Polytechnic Institute and State University, Luukkonen, D.R., Eichholz, M.E. & Koch, Blacksburg, Virginia, USA. K.E. 2013. Monitoring Midwest waterfowl Sherfy, M.H., Kirkpatrick, R.L. & Richkus, K.D. during the non-breeding period: priorities, 2000. Benthos core sampling and Chironomid challenges and recommendations. Journal of vertical distribution: implications for assessing Fish and Wildlife Management 4: 395–405. shorebird food availability. Wildlife Society St. James, E.A. 2011. Effect of hunting Bulletin 28: 124–130. frequency on duck abundance, harvest and Sherfy, M.H., Kirkpatrick, R.L. & Webb, Jr., K.E. hunt quality in Mississippi. M.Sc. thesis, 2001. Nutritional consequences of gastolith Mississippi State University, USA. ingestion in blue-winged teal: a test of the St. James, E.A., Schummer, M.L., Kaminski, hard-seed-for-grit hypothesis. Journal of R.M., Penny, E.J., Burger, L.W. 2013. Wildlife Management 65: 406–414. Effect of weekly hunting frequency on Shirkey, B.T. 2012. Diving duck abundance and duck abundances in Mississippi Wildlife distribution on Lake St. Clair and western Management Areas. Journal of Fish and Lake Erie. M.Sc. thesis, Michigan State Wildlife Management 4: 144–150. University, East Lansing, USA. Stafford, J.D., Kaminski, R.M., Reinecke, K.J. & Sibbald, I.R. 1975. The effect of level of feed Manley, S.W. 2006. Waste rice for waterfowl intake on metabolizable energy values in the Mississippi Alluvial Valley. Journal of measured with adult roosters. Poultry Science Wildlife Management 70: 61–69. 54: 1990–1997. Stafford, J.D., Yetter, A.P., Hine, C.S., Smith, R.V. Sibbald, I.R. 1979. Effects of level of feed input, & Horath, M.M. 2011. Seed abundance for dilution of test material, and duration of waterfowl in wetlands managed by the excreta collection on true metabolizable Illinois Department of Natural Resources. energy values. Poultry Science 58: 1325–1329. Journal of Fish and Wildlife Management 2: 3–11.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 Waterfowl carrying capacity models 435

Stillman R.A. 2008. MORPH – an individual- beds – a pilot study. Turkish Journal of Fisheries based model to predict the effect of and Aquatic Sciences 10: 161–167. environmental change on foraging animal Tidwell, P.R., Webb, E.B., Vrtiska, M.P., Bishop, populations. Ecological Modelling 216: 265–276. A.A. 2013. Diets and food selection of Stillman R.A. & Goss-Custard, J.D. 2010. female mallards and blue-winged teal during Individual-based ecology of coastal birds. spring migration. Journal of Fish and Wildlife Biological Reviews 85: 413–434. Management 4: 63–74. Stillman, R.A., Goss-Custard, J.D., West, A.D., Thomas, D.R. 2004. Assessment of Waterfowl Durell, S.E.A.L.V., McGrorty, D.S., Caldow, Body Condition to Evaluate the Effectiveness R.W.G., Norris, K.J., Johnstone, I.G., Ens, of the Central Valley Joint Venture. M.Sc. B.J., van der Meer, J. & Triplet, P. 2001. Thesis. University of California, Davis, Predicting shorebird mortality and California, USA. population size under different regimes of U.S. Fish and Wildlife Service’s Migratory shellfishery management. Journal of Applied Bird Data Center (USFWS MBDC). 2014. Ecology 38: 857–868. Midwinter Waterfowl Index Database. U.S. Fish Straub, J.N., Gates, J.L., Schultheis, R.D., Yerkes, and Wildlife Service, Patuxent, Maryland, T., Coluccy, J.M. & Stafford, J.D. 2012. USA. Accessible at http://mbdcapps.fws.gov/ Wetland food resources for spring-migrating (last accessed 14 March 2014). ducks in the upper-Mississippi River and van Gils, J.A., Edelaar, P., Escudero, G. & Great Lakes Region. Journal of Wildlife Piersma, T. 2004. Carrying capacity models Management 76: 768 –777. should not use fixed prey density thresholds: Sutherland, W.J. 1998. The effect of local change a plea for using more tools of behavioural in habitat quality on populations of ecology. Oikos 104: 197–204. migratory species. Journal of Applied Ecology Weathers, W.W. 1979. Climate adaptations in 35: 418–421. avian standard metabolic rate. Oecologia 42: Sutherland, W.J. & Allport, G.A. 1994. A spatial 81–89. depletion model of the interaction between Weathers, W.W., Buttemer, W.A., Hayworth, A.M. bean geese and wigeon with the consequences & Nagy, K.A. 1984. An evaluation of time- for habitat management. Journal of Animal budget estimates of daily energy expenditure Ecology 63: 51–59. in birds. Auk 101: 459–472. Swanson, G.A. 1978. A water column sampler for West A.D., Stillman, R.A., Drewitt, A., Frost, N.J., invertebrates in shallow wetlands. Journal of Mander M. & Miles, C. 2011. MORPH – a Wildlife Management 42: 670–672. user-friendly individual-based model to Swanson, G.A. 1983. Benthic sampling for advise shorebird policy and management. waterfowl foods in emergent vegetation. Methods in Ecology and Evolution 2: 95–98. Journal of Wildlife Management 47: 821–823. Wooley, Jr., J.B. 1976. Energy expenditure of the Synchra, J. & Adamek, Z. 2010. Sampling black duck under controlled and free-living efficiency of the Gerking sampler and sweep conditions. M.Sc. thesis, University of Maine, net in pond emergent littoral macrophyte Orono, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 407–435 436

Annual variation in food densities and factors affecting wetland use by waterfowl in the Mississippi Alluvial Valley

HEATH M. HAGY1*, JACOB N. STRAUB2, MICHAEL L. SCHUMMER3,5 & RICHARD M. KAMINSKI4

1*Illinois Natural History Survey, Forbes Biological Station – Bellrose Waterfowl Research Center, University of Illinois at Urbana-Champaign, Havana, Illinois 62644, USA. 2State University of New York at Plattsburgh, Center for Earth and Environmental Science, Plattsburgh, New York 12901, USA. 3Long Point Waterfowl, PO Box 160, Port Rowan, Ontario N0E 1M0, Canada. 4Mississippi State University, Department of Wildlife, Fisheries and Aquaculture, Box 9690, Mississippi State, Mississippi 39762, USA. 5Current address: State University of New York at Oswego, Department of Biological Sciences, 30 Centennial Drive, Oswego, New York 13126, USA. *Correspondence author. E-mail: [email protected]

Abstract Spatial and temporal heterogeneity in habitat quantity and quality, weather and other variables influence the production of food and the distribution of waterfowl, making it difficult to predict carrying capacity accurately. Food densities for waterfowl, which are key parameters of energetic carrying capacity models, were examined in managed moist-soil wetlands and bottomland hardwood forests in or near the Mississippi Alluvial Valley (MAV) of the southern United States of America, to determine variation in those densities across wetlands and years. Secondly, the relationship between migratory waterfowl density in managed wetlands and local and mid-latitude factors north of the study area was examined to identify mechanisms influencing waterfowl density at latitudes used during winter. At individual wetlands and within years, food densities were highly variable, but coefficients of variation (CV) at the scale of the MAV and nearby areas across years were relatively low (moist-soil CV = 21%, bottomland hardwood forest CV = 11%). Local precipitation was inversely related to waterfowl density in managed moist-soil wetlands, and this relationship was stronger than other local and mid-latitude factors including weather severity and temperature. Our data suggest that simplistic daily ration models may reasonably incorporate fixed estimates of food density for managed moist-soil wetlands and bottomland hardwood forests to predict energetic carrying capacity of waterfowl habitat at the scale of the MAV across multiple years. However, substantial variation in food densities among

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 Variation in waterfowl food densities 437

locations and time periods likely limits the utility and accuracy of these models when scaled down temporally or spatially. Therefore, the challenge in predicting annual carrying capacity for waterfowl in the MAV likely depends less on precisely estimating food densities at the scale of individual wetlands and more on determining spatial and temporal availability of habitats that contain food resources for waterfowl.

Key words: bottomland hardwood forest, conservation planning, dabbling duck, daily ration model, migration, moist-soil, weather severity.

Waterfowl ecologists use predictive models Newton 1998; Schummer et al. 2010; Hagy to develop habitat conservation objectives & Kaminski 2012a,b). Moreover, even if sufficient to meet the energy demands available in some parts of the migratory of migrating and wintering waterfowl range, foods may not be encountered by populations in North America and elsewhere waterfowl because of variation within and (Soulliere et al. 2007; Reinecke et al. 1989). between years in the timing of migration These models require estimates of food and regional movements by waterfowl density and foraging demand, with the latter (Bellrose et al. 1979; Schummer et al. 2010; reflecting the number and duration of stay of Krementz et al. 2011, 2012; O’Neal et al. individuals foraging in a given area (Williams 2012; Hagy et al. 2014). Currently, energetic et al. 2014). Many previous studies have carrying capacity models used by some Joint aimed to measure the density and availability Ventures of the North American Waterfowl of food in habitats used by waterfowl, but Management Plan and other conservation few researchers have examined variability in partners include fixed parameters (i.e. food densities at multiple temporal and constants) which may not account for spatial scales (cf. Lovvorn & Gillingham spatio-temporal variation in food density or 1996). Moreover, variation or changes in other factors that result in a mismatch of regional and habitat use by waterfowl can foods becoming available and waterfowl have significant effects on carrying capacity being present to access those foods model predictions of habitat requirements (Soulliere et al. 2007; Williams et al. 2014). In (Hagy et al. 2014). order to develop habitat conservation Food availability for waterfowl can objectives effectively, conservation planners be influenced by a range of factors require an understanding of variation in including annual production and seasonal carrying capacity estimates resulting from decomposition of plant and animal foods, variable parameter estimates and the depletion of food resources by wildlife mechanisms underlying waterfowl habitat other than waterfowl, diet selectivity by use to determine priority habitats for foragers, ice and snow cover, duration and conservation (Schummer et al. 2010; Hagy & depth of flooding, disturbance by humans Kaminski 2012b; Beatty et al. 2014b; Hagy et and natural predators and photoperiodic al. 2014; Williams et al. 2014). cues triggering migration (Rees 1981; Recently, information has become

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 438 Variation in waterfowl food densities available to help explain waterfowl is a need for studies that simultaneously movements and habitat use during winter in consider factors within wintering areas relation to landscape composition (Pearse et and those occurring at latitudes north al. 2012; Beatty et al. 2014b). In addition to of wintering areas, which may cause landscape-scale factors measured close to movements of birds into wintering areas wintering and stopover sites used by large and influence their access to foods, numbers of waterfowl, factors north of subsequent fitness and conservation wintering areas could affect the southward planning (Lovvorn & Baldwin 1996; Haig et movement of individuals and these may be al. 1998; Pearse et al. 2012). The annual useful in further explaining and predicting variation in food density in managed moist- wetland use (Schummer et al. 2010). For soil wetlands and bottomland hardwood migratory species such as waterfowl, habitat forests in and near the Mississippi Alluvial selection is likely a hierarchical process Valley (MAV), an important wintering area and factors affecting selection may vary for North American waterfowl at the temporally and interact spatially (Beatty et al. continental level (Reinecke et al. 1989), was 2014a,b). For example, the cumulative examined to determine variation in effects of decreasing temperatures, parameter estimates used in energetic freezing of wetlands and snow cover can carrying capacity models at two spatial cause regional decreases in abundance scales (within and across study wetlands in of waterfowl at autumn staging areas the MAV) and across years. Secondly, the (Schummer et al. 2010), but once birds reach relative influence of factors measured not their southern wintering grounds where only locally but also at a mid-latitude harsh weather conditions are less common, location on migratory waterfowl densities in other factors such as precipitation (which managed wetlands was investigated. Our influences wetland availability), food objectives were to: 1) describe variation in availability and intrinsic wetland factors may food densities across wintering areas used influence habitat use (Davis et al. 2009; Hagy by migrating waterfowl during autumn and & Kaminski 2012b; Dalby et al. 2013). winter in the MAV, and 2) determine factors Factors related to migratory movements that influence waterfowl densities on from mid-latitude areas to more southerly managed wetlands to better inform the wintering grounds may influence the conservation planning process. abundance of birds within southern areas and allow comparison of the relative Methods influence of local and mid-latitude factors on site use by the birds. Variation in food abundance Although a number of studies have Waterfowl density and associated food examined waterfowl movements and densities were estimated in moist-soil abundance in relation to food, habitat wetlands and food density was also juxtaposition and other factors influencing estimated in bottomland hardwood forests movements along the migration route, there in or near the MAV (Reinecke et al. 1989;

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 Variation in waterfowl food densities 439

Hagy & Kaminski 2012a,b). Data presented lost, missed or destroyed during processing here and related sample collection methods (Hagy et al. 2011), seeds and tubers of plant have previously been described in detail taxa thought to be avoided or infrequently by Hagy and Kaminski (2012b) and consumed by waterfowl were removed from Straub (2012), but different analyses were density estimates (Hagy & Kaminski 2012a) conducted to address our novel objectives. and coefficients of variation (CV) were Moist-soil and bottomland forest wetlands estimated for each wetland (CV = s.d./mean are used extensively by many species of density for all foods combined in non- waterfowl, especially dabbling ducks (Anas manipulated wetland plots), across wetlands sp.), for provision of food resources and for each year, and across years for both other life-history needs (Reinecke et al. moist-soil wetlands and bottomland 1989). Moist-soil wetlands provide hardwood forests. abundant natural seeds and tubers after they are flooded, which typically occurs in late Local and mid-latitude factors autumn or early winter, and bottomland affecting waterfowl density hardwood wetlands provide hard mast Waterfowl were enumerated by species from (e.g. acorns) throughout winter which elevated hides during crepuscular periods typically become available during periodic 2–3 times weekly at each wetland plot from bottomland flooding events (Reinecke et al. first flooding (i.e. November–December) 1989; Hagy & Kaminski 2012b; Straub through to waterfowl leaving the wetlands 2012). To estimate annual densities of and surrounding area (i.e. late February), sound red oak Quercus sp. acorns, we during winters 2006–2009 (Hagy & installed and checked 1-m2 seed traps Kaminski 2012b). We combined densities of monthly from November through February all dabbling duck species observed and 2009–2013 at five study wetlands across the analysed only this variable as dabbling ducks MAV (Straub 2012). Additionally, to comprised >90% of observations and their estimate moist-soil seed and tuber densities, densities were positively correlated with we collected 10 benthic core subsamples densities of all waterbirds combined (Hagy during November or December (i.e. before & Kaminski 2012b). most wintering waterfowl accessed wetlands To accomplish our goal of investigating and depleted foods) in 2006–2008 at each of factors affecting waterfowl densities in three wetland plots immediately following wetlands in the study area, we examined the flooding that had been either: 1) mown, influence of weather and habitat variables 2) disced, or 3) not manipulated during the measured within or near wetlands in the autumn prior to flooding, for 22 wetlands MAV (i.e. local) and weather variables in or near the MAV (2006 = 6 wetlands, measured c. 200 km north of our study area 2007 = 9 wetlands, 2008 = 7 wetlands). (i.e. mid-latitude) where large concentrations Laboratory processing followed Hagy and of waterfowl may winter if they are not Kaminski (2012b) and Straub (2012); seed encouraged to migrate further south by and tuber densities were adjusted for seeds weather or other factors (Schummer et al.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 440 Variation in waterfowl food densities

2010). Concurrent assessment of local (nlme; Pinheiro et al. 2014) to assess wintering area and mid-latitude variables variation in dabbling duck density across addresses the hypothesis that waterfowl managed moist-soil wetlands in relation densities might be influenced by extrinsic to various explanatory variables. A set factors (e.g. a weather severity index (WSI); of candidate models (Table 1), each Schummer et al. 2010), and these events representing a unique biologically-plausible may exert a greater influence than local scenario, was built and models were conditions in southerly areas (i.e. factors compared for explanatory support using near and within the southerly wetlands) to Akaike’s information criterion adjusted for which the birds migrate. Weather data from small sample size (AICc; Burnham & central Missouri, which is in the northern Anderson 2002). Because study wetlands part of the typical wintering range for were repeatedly sampled within a season, Mallard Anas platyrhynchos following the observation date was included as a repeated Mississippi Flyway, were assumed to provide effect in the model. Additionally, the a reasonable representation of the effects of management category for each wetland weather on duck movements from northern plot (i.e. mow, disc or no manipulation) staging and wintering areas into the MAV, nested within wetland was designated as a including our study wetlands. Weather data random effect because evidence (i.e. lowest

(i.e. WSI, cumulative precipitation during AICc) suggested that this increased the winter (October 1 to observation date), and explanatory power of our global model mean daily temperature) from the closest (Zuur et al. 2009). Models were developed to weather station and water depth gauge assess support for mid-latitude factors, local readings from within each wetland were factors, a combination of both local and used to evaluate local influences on mid-latitude factors, the effects of year and waterfowl densities. Water depths and date of surveys, and finally a null model weather data were recorded on the same day containing only the intercept. Inspection of as waterfowl densities. For more northern residual plots and histograms indicated that (mid-latitude) parts of the wintering dabbling duck density (i.e. the response range, the same weather variables were variable) was not normally distributed and acquired. Data were acquired from the had heterogeneous variance when plotted Historical Climatology Network National against independent variables. Dabbling Oceanographic Atmospheric Service duck density therefore was natural log weather stations at Farmington, Missouri transformed prior to analysis. Parameter (mid-latitude) and also from weather estimates from the most parsimonious stations closest to (≤ 50 km from) study model were back-transformed to describe wetlands (at Batesville, Corinth, Greenwood, the size of each effect. We provide marginal Starkville and Yazoo City in Mississippi; and conditional R2 statistics as a means to Covington, Jackson and Union City in assess the fit of each candidate model Tennessee). (Nakagawa & Schielzeth 2013). Marginal R2 Linear mixed models were used in R describes the proportion of variance

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 Variation in waterfowl food densities 441

Table 1. Results of linear mixed models predicting dabbling duck density in managed moist- soil wetlands in or near the Mississippi Alluvial Valley during late autumn and winter 2006–2009, with the difference between each model-specific Akaike Information Criteria adjusted for small sample size (ΔAICc) and that of the top model. Model variables include local (LO) and mid-latitude (MO) estimates of cumulative winter precipitation (PrecipW), water depth (Depth), a weather severity index (WSI), temperature (Temp), year, management practice (autumn mowing, discing, or no management; Treat) and Julian day.

Δ R2 R2 Model AICc AICc marg cond

LOPrecipW+LODepth 2248.3 0 0.04 0.34 LOPrecipW + LODepth + MOPrecipW 2254.6 6.3 0.05 0.34 LOTemp+LODepth+LOPrecipW 2257.4 9.1 0.04 0.35 LOPrecipW + LODepth + MOWSI 2257.5 9.2 0.04 0.35 LODepth + MOPrecipW 2262.4 14.1 0.00 0.44 LOPrecipW + LOTemp + LODepth + MOWSI 2263.5 15.2 0.04 0.36 LOPrecipW + LOTemp + LODepth + MOPrecipW 2263.6 15.3 0.05 0.34 LOPrecipW + LODepth + MOWSI + MOPrecipW 2263.7 15.4 0.05 0.34 LOTemp + LOPrecipW + LODepth + MOWSI + 2269.3 21.0 0.05 0.35 MOPrecipW LODepth + MOWSI + MOPrecipW 2271.8 23.5 0.00 0.45 LOTemp + LODepth + MOWSI + MOPrecipW 2276.1 27.8 0.00 0.44 LODepth 2286.9 38.6 0.00 0.44 LOTemp+LODepth 2294.9 46.6 0.00 0.41 LODepth + MOWSI 2297.2 48.9 0.00 0.45 LOPrecipW 2407.0 158.7 0.03 0.36 LOPrecipW+LOTemp 2415.7 167.4 0.03 0.36 LOPrecipW + MOWSI 2416.6 168.3 0.03 0.36 MOPrecipW 2417.9 169.6 0.00 0.44 LOPrecipW + MOWSI + MOPrecipW 2420.8 172.5 0.04 0.35 MOPrecipW + MOTemp 2427.3 179.0 0.00 0.45 MOPrecipW + MOWSI 2427.5 179.2 0.00 0.45 Treat 2449.7 201.4 0.07 0.29 Null (Intercept) 2462.0 213.7 0.00 0.43 Year 2468.9 220.6 0.01 0.39 Julian day 2469.9 221.6 0.02 0.36 LOTemp 2470.5 222.2 0.00 0.42 MOTemp 2472.7 224.4 0.00 0.44 MOWSI 2472.8 224.5 0.00 0.45

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 442 Variation in waterfowl food densities explained by the fixed factor(s) while density) was natural log transformed to conditional R2 describes the proportion of normalize residuals and homogeneity of variance explained by both the fixed and variances among years and wetlands. Results random factors (Nakagawa & Schielzeth were considered significant at P < 0.05. 2013). Additionally, a general linear mixed model Results was used in SAS 9.3 to evaluate the effects For managed moist-soil, CVs for seed of year and wetland on food density in and tuber densities ranged from 9–77% managed moist-soil wetlands in late autumn within wetlands (x– = 31%, n = 22) and from (i.e. approximately early November, before 32–115% across years (x– = 69%, n = 3). waterfowl used wetlands) by performing a Overall, the CV of the annual mean seed different analysis of data presented by Hagy density across years and wetlands was less and Kaminski (2012b). Year and wetland (CV = 21%) than within years or most were included as fixed effects and wetland wetlands (Fig. 1). For bottomland hardwood management practice (i.e. discing, mowing forests, CVs for red Oak acorn densities or no manipulation of robust moist-soil ranged from 16–60% within wetlands vegetation; see Hagy & Kaminski 2012b) (x– = 33% n = 5) and from 11–29% across within each wetland plot was included as a wetlands (x– = 18%, n = 4). Overall, the CV random effect. The response variable (food of the annual mean acorn density across

120%

100%

80%

60%

40% Coeff i c ent ofi at on var 20%

0% 2006 2007 200 8

2006 2007 2008 Overall 2006–200 8

Figure 1. Coefficients of variation (%) for means of seed and tuber density for individual wetland plots, across wetland plots within years, and across wetland plots and years (overall) during late autumn 2006–2008 in managed moist-soil wetlands (n = 22 unmanipulated moist-soil wetland plots) in or near the Mississippi Alluvial Valley.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 Variation in waterfowl food densities 443

70% 60% 50% 40% 30% 20% 10% Coeff i c ent ofi at on var 0% 2009 2010 2011 2012 2009–2012

2009 2010 2011 2012 Overall

Figure 2. Coefficients of variation (%) for means of red Oak acorn production density for individual wetlands, across wetlands within years, and across wetlands and years (overall) during late autumn and winter 2009–2012 in bottomland hardwood forests (n = 5 wetlands surveyed repeatedly) in or near the Mississippi Alluvial Valley. years and wetlands was less (CV = 11%) precipitation during winter decreased than individual years or most wetlands predicted numbers of dabbling ducks in (Fig. 2). Food densities in managed moist- managed moist-soil wetlands by 1 duck/ha. soil wetlands in late autumn varied by Models containing only mid-latitude or wetland (F16,32 = 6.56, P < 0.001) but mid-latitude plus local factors were not Δ did not differ among years (F2,13 = 2.01, competitive ( AICc > 6.3). Although we P = 0.174) (Fig. 3). had low model uncertainty, the proportion On evaluating the explanatory variables of the variance explained by depth and local thought to influence dabbling duck density, winter precipitation (R2 = 0.04) was less the top model contained factors measured than the variance explained by the only at the local scale (Table 1). Dabbling combination of fixed effects and random duck densities in managed moist-soil variables (R2 = 0.34). wetlands were negatively associated with local winter precipitation (i.e. an assumed Discussion correlate of local wetland availability) and Small-scale spatial or temporal estimates positively associated with mean water depth of food density (e.g. among wetlands or of wetland plots, although confidence years) in managed moist-soil wetlands and intervals associated with the beta estimate bottomland hardwood forests were highly overlapped zero for water depth and thus we variable and means from these estimates did not explore that relationship further were relatively imprecise, which is consistent (Fig. 4). A 9.6 cm increase in local with other studies in similar habitats

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 444 Variation in waterfowl food densities

1,000 900 800 700 600 500 400 300 200 Seed and tu b er dens i ty (kg/ha) 100 0 2006 2007 2008

Figure 3. Seed and tuber density (dry weight, in kg/ha ± s.e.) during late autumn 2006–2008 in managed moist-soil wetlands (n = 64 wetland plots) in or near the Mississippi Alluvial Valley (data from Hagy & Kaminski 2012b).

(Stafford et al. 2006; Kross et al. 2008; direct relationships between food resources Evans-Peters et al. 2012; Straub 2012; and waterfowl distribution seem to be Olmstead et al. 2013). However, coefficients difficult to detect without incorporating of variation were relatively low and means additional local environmental conditions were similar when estimated across wetlands into habitat models (Fleming 2010; Tapp and years. Thus, use of fixed food densities 2013; Weegman 2013). in simplistic daily ration models for Hagy and Kaminski (2012b) presented conservation planning purposes at a large data indicating that early winter food spatial scale appears to be a reasonable densities varied with management practice practice. At the MAV scale and across in moist-soil wetlands; they considered year the years of the study, food densities and wetland as random effects relative to in managed moist-soil wetlands and their research questions, but did not test bottomland hardwood forests were them explicitly. Herein, re-examination of relatively constant; however, the spatial their data across years indicated that while distribution of that food within the region waterfowl food densities varied by wetland, varied annually. Because waterfowl are there was not an apparent annual difference highly mobile and respond to changing in their study area (Fig. 3). The variation habitat conditions, they are likely able to between individual wetlands was much move within the landscape and respond to greater than across years. Similarly, Straub changing distributions of food and habitat (2012) reported that acorn densities in availability. In fact, many factors other than bottomland hardwood forests fluctuated food density likely influence habitat use, and greatly across years for individual wetlands;

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 Variation in waterfowl food densities 445

Water depth

Local precipitation

–0.06 –0.05 –0.04 –0.03 –0.02 –0.01 0.00 0.01 0.02 Partial Regression Coefficient Figure 4. Direction and relative effect size (partial regression coefficient with 95% confidence intervals) for variables in the top model predicting dabbling duck densities during late autumn 2006–2008 in managed moist-soil wetlands (n = 64 wetland plots) in or near the Mississippi Alluvial Valley. however, at the scale of the MAV, annual al. 2008), bottomland hardwood forests and estimates were similar. Although some agricultural rice fields (Stafford et al. 2006), bottomland hardwood forests produced few food production for waterfowl has been acorns in some years, low yield never shown to be highly variable between sites occurred at all wetlands in the same year. but much less variable across years and large Across years, MAV-wide estimates of red regions, such as the MAV. Variation at Oak acorn abundance were precise (CV = individual wetlands may be influenced by 11%), but variability across wetlands was management practices (Hagy & Kaminski great (Fig. 2). Stafford et al. (2006) and Kross 2012b), management frequency and et al. (2008) both reached similar conclusions intensity (Brasher et al. 2007; Olmstead et al. for seeds in rice fields and moist-soil 2013), and other environmental factors that wetlands, respectively, in the MAV. Thus, are difficult to predict accurately. Therefore, while daily ration models incorporating the challenge in predicting annual carrying fixed food densities may reasonably predict capacity in the managed wetlands and carrying capacity at large spatial scales, with bottomland hardwood forests of the MAV all other parameters being equal, substantial likely depends less on accounting for annual variation among locations and time periods differences in site-specific food densities likely limits the predictive accuracy of these and more on availability of those wetlands models when scaled down spatially. as habitats suitable for waterfowl. However, In managed moist-soil wetlands (Kross et flooding of foods at the appropriate time

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 446 Variation in waterfowl food densities

(Greer et al. 2007) and to the appropriate precipitation in more northerly portions of depth (Hagy & Kaminski 2012b) may be the Mallard’s wintering range (Krementz et a challenge in some years. To date, we al. 2012; Beatty et al. 2014b). Distributions are aware of few attempts to quantify and duration of stay of waterfowl at autumn functionally available habitats at a scale such staging areas can vary with wetland area, as a joint venture region (but see Soulliere et disturbance (Stafford et al. 2010), wetland al. 2007), despite the clear need to determine forage quality (O’Neal et al. 2012), and quantify food availability. precipitation (Krementz et al. 2011, 2012) We examined the relative influences of and vegetation characteristics (Moon mid-latitude and local factors on waterfowl & Haukos 2008), but others have failed densities in managed wetlands in and near to show a relationship between food the MAV and determined that local abundance and waterfowl use of wetlands precipitation, an assumed surrogate of (Percival et al. 1998; Straub 2008; Brasher wetland availability, was most influential in 2010; Fleming 2010). Weather has been explaining variation in dabbling duck shown to influence regional abundance of density. Managed wetlands with extensive ducks (Schummer et al. 2010), but Krementz water control capabilities, such as the et al. (2012) reported that temperature and wetlands included in our study (see Hagy & the onset of freezing conditions were not Kaminski 2012b), are often flooded before significant determinants of departure date many passively or non-managed wetlands during autumn migration. Cumulative and are available when waterfowl first arrive research suggests that there is a significant in the late autumn and early winter. Thus, degree of plasticity in the autumn migration waterfowl are likely attracted to these of duck species to the wintering grounds managed wetlands initially but may later (Bellrose et al. 1979; Krementz et al. 2012). colonise passively filled or temporary We posit that a suite of factors including wetlands following periods of sufficient photoperiod, the location and timing of rainfall (Beatty et al. 2014a). Given the rapid severe weather events and habitat availability declines in waterfowl food densities in and quality interact with considerable inter- managed wetlands documented by Hagy and intraspecific variation to determine and Kaminski (2012b), our results suggest migratory patterns, but continued research that waterfowl may move to alternative is necessary to differentiate their relative locations, such as agricultural fields (Pearse contribution to the timing of waterfowl et al. 2012), following precipitation events to migration and habitat use. acquire newly-available food on flooded At a continental or flyway scale, a suite of farmland. factors influences waterfowl habitat Interestingly, evidence is accumulating selection, including wetland availability that suggests local habitat availability may be (Beatty et al. 2014b). Krementz et al. (2011) a better predictor of duck density in and Hagy et al. (2014) anecdotally noted managed wetlands than food density that, during spring- and autumn-migration, (Hagy 2010; Tapp 2013) or weather and Mallard stopover use and duration of stay

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 Variation in waterfowl food densities 447 may have been related to precipitation food acquisition), temporal availability of and the availability of local wetlands, patches within and among years and respectively. Interestingly, we identified a individual patch value rather than aggregate relationship between local precipitation and food availability to improve accuracy and duck densities in the wintering region of the utility at smaller scales. In reality, extensive MAV, which might also suggest that once inter- and intraspecific variation in ducks migrate to latitudes where influences migration timing and life-history strategies of photoperiod, weather severity and other add considerable uncertainty to energetic factors decrease (i.e. wintering areas; models (Hagy et al. 2014) and efforts Schummer et al. 2010), wetland habitat aimed at reducing these uncertainties or availability is also an important driver of quantifying the relative effects on energetic habitat selection (Webb et al. 2010; Pearse carrying capacity models at annual and et al. 2012). Thus, conservation planning longer-term timescales would be beneficial models can benefit from considering (see Notaro et al. 2014; Schummer et al. timing and extent of flooding (i.e. wetland 2014). inundation) when determining habitat In summary, our data indicate that objectives at large spatial scales (e.g. Joint annual food densities in managed wetlands Ventures). and bottomland hardwood forests were Although challenging to build and generally stable across the 3-year study at a parameterise, spatially explicit models that regional scale used by a Joint Venture for incorporate variables (e.g. precipitation) that conservation planning. It may be beneficial account for spatial and temporal variability for conservation planners to quantify in habitat availability may be needed longer-term variability in food resources and for more accurate predictions of food examine factors influencing the availability availability and site use by ducks. However, of these resources to waterfowl in other relatively simplistic daily ration models that regions used by wintering waterfowl. incorporate fixed estimates of food density Furthermore, if the results are extended are likely adequate for predicting energetic to other wintering areas and habitat carrying capacity at the MAV scale for types, quantifying and facilitating habitat waterfowl that are highly mobile and availability within the landscape for can respond rapidly to changing habitat waterfowl, in sufficient quantities to meet conditions and availability. A critical next their energetic demands at the appropriate step in improving the accuracy of energetic time and location, is a challenge worthy of carrying capacity models is estimating the additional exploration. spatial and temporal extent and variability of habitats by modelling the wetland areas Acknowledgements suitable for exploitation of food resources We thank the many individuals who assisted by waterfowl (Williams et al. 2014). Future during collection of the data used in this modelling attempts could incorporate the manuscript. Many thanks are due to two spatial arrangement of patches (i.e. costs of anonymous reviewers, L. Webb and E. Rees

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 448 Variation in waterfowl food densities for their helpful suggestions which greatly Dalby, L., Fox, A.D., Petersen, I.K., Delany, S. improved this manuscript. We thank the & Svenning, J.C. 2013. Temperature does not Illinois Natural History Survey, the Forest dictate the wintering distributions of European and Wildlife Research Center (FWRC) of dabbling duck species. Ibis 155: 80–88. the Mississippi State University (MSU) Davis, B.E., Afton, A.D. & Cox, R.R., Jr. 2009. Habitat use by female mallards in the lower Department of Wildlife, Fisheries and Mississippi Alluvial Valley. The Journal of Aquaculture, the United States Forest Wildlife Management 73: 701–709. Service, Ducks Unlimited and other Evans-Peters, G.R., Dugger, B.D. & Petrie, M.J. organisations for funding and support. 2013. Plant community composition and waterfowl food production on Wetland References Reserve Program Easements compared to Beatty, W.S., Kesler, D.C., Webb, E.B., Raedeke, those on managed public lands in western A.H., Naylor, L.W. & Humburg, D.D. 2014a. Oregon. Wetlands 32: 391–399. The role of protected area wetlands in Fleming, K.S. 2010. Effects of management and waterfowl habitat conservation: implications hydrology on vegetation, winter waterbird for protected are network design. Biological use, and water quality on Wetlands Reserve Conservation 176: 144–152. Program lands, Mississippi. M.Sc. thesis, Beatty, W.S., Webb, E.B., Kesler, D.C., Raedeke, Mississippi State University, Mississippi State, A.H., Naylor, L.W. & Humburg, D.D. 2014b. USA. Landscape effects on mallard habitat Greer, A.K., Dugger, B.D., Graber, D.A. & Petrie, selection at multiple spatial scales during the M.J. 2007. The effects of seasonal flooding non-breeding period. Landscape Ecology 29: on seed availability for spring migrating 989–1000. waterfowl. The Journal of Wildlife Management Bellrose, F.C., Paveglio, Jr. F.L. & Steffeck, 71: 1561–1566. D.W. 1979. Waterfowl populations and the Haig, S.M., Mehlman, D.W. & Oring, L.W. 1998. changing environment of the Illinois River Avian movements and wetland connectivity valley. Illinois Natural History Survey Bulletin 32: in landscape conservation. Conservation Biology 1–54. 12: 749–758. Brasher, M.G. 2010. Duck use and energetic Hagy, H.M. 2010. Winter food and waterfowl carrying capacity of actively and passively dynamics in managed moist-soil wetlands managed wetlands in Ohio during autumn in the Mississippi Alluvial Valley. Ph.D. and spring migration. Ph.D. thesis, The Ohio Dissertation, Mississippi State University, State University, Columbus, USA. Mississippi State, Mississippi, USA. Brasher, M.G., Steckel, J.D. & Gates, R.J. 2007. Hagy, H.M. & Kaminski, R.M. 2012a. Apparent Energetic carrying capacity of actively and seed use by ducks in the Mississippi Alluvial passively managed wetlands for migrating Valley. Journal of Wildlife Management 76: ducks in Ohio. Journal of Wildlife Management 1053–1061. 71: 2532–2541. Hagy, H.M. & Kaminski, R.M. 2012b. Winter Burnham, K.P. & Anderson, D.R. 2002. Model waterbird and food dynamics in autumn- Selection and Multimodel Inference: a Practical managed moist-soil wetlands of the Information-Theoretic Approach. Second Edition. Mississippi Alluvial Valley. Wildlife Society Springer-Verlag, New York, USA. Bulletin 36: 512–523.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 Variation in waterfowl food densities 449

Hagy, H.M., Straub, J.N. & Kaminski, R.M. 2011. Notaro, M., Lorenz, D., Hoving, C. & Schummer, Estimation and correction of seed recovery M. 2014. 21st Century projections of bias from moist-soil cores. Journal of Wildlife snowfall and winter severity across central- Management 75: 959–966. eastern North America. Journal of Climate Hagy, H.M., Yetter, A.P., Stodola, K.W., Horath, doi: http://dx.doi.org/10.1175/JCLI-D-13- M.M., Hine, C.S., Ward, M.P., Benson, T.J., 00520.1. Smith, R.V. & Stafford, J.D. 2014. Stopover Newton, I. 1998. Population Limitation in Birds. duration of mallards during autumn in the Academic Press, San Diego, California, USA. Illinois River Valley. Journal of Wildlife O’Neal, B.J., Stafford, J.D. & Larkin, R.P. 2012. Management 78: 747–752. Stopover duration of fall-migrating dabbling Krementz, D.G., Asanti, K. & Naylor, L.W. ducks. Journal of Wildlife Management 76: 2011. Spring migration of mallards from 285–293. Arkansas as determined by satellite telemetry. Olmstead, V.G., Webb, E.B. & Johnson, R.W. Journal of Fish and Wildlife Management 2: 2013. Moist-soil seed biomass and species 156–168. richness on Wetland Reserve Program Krementz, D.G., Asanti, K. & Naylor, L.W. 2012. Easements in the Mississippi Alluvial Valley. Autumn migration of Mississippi Flyway Wetlands 33: 197–206. mallards as determined by satellite telemetry. Pearse, A.T., Kaminski, R.M., Reinecke, K.J. & Journal of Fish and Wildlife Management 3: Dinsmore, S.J. 2012. Local and landscape 238–251. associations between wintering dabbling Kross, J., Kaminski, R.M., Reinecke, K.J., Penny, ducks and wetland complexes in Missisppi. E.J. & Pearse, A.T. 2008. Moist-soil seed Wetlands 32: 859–869. abundance in managed wetlands in the Percival, S.M., Sutherland, W.J. & Evans, P.R. Mississippi Alluvial Valley. Journal of Wildlife 1998. Intertidal habitat loss and wildfowl Management 72: 707–714. numbers, applications of a spatial depletion Lovvorn, J.R. & Baldwin, J.R. 1996. Intertidal model. Journal of Applied Ecology 35: 57–63. and farmland habitats of ducks in the Puget Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & R Sound region: a landscape perspective. Core Team. 2014. Nlme: Linear and Nonlinear Biological Conservation 77: 97–114. Mixed Effects Models. R package version 3.1-117. Lovvorn, J.R. & Gillingham, M.P. 1996. Accessible at http://CRAN.R-project.org/ Food dispersion and foraging energetics: a package=nlme (last accessed 08.07.14). mechanistic synthesis for field studies of Rees, E.C. 1982. The effect of photoperiod avian benthivores. Ecology 77: 435–451. on the timing of spring migration in the Moon, J.A. & Haukos, D.A. 2008. Habitat use of Bewick’s Swan. Wildfowl 33: 119–132. female northern pintails in the playa lakes Reinecke, K.J., Kaminski, R.M., Moorhead, D.J., region of Texas. Proceedings of the Annual Hodges, J.D. & Nassar, J.R. 1989. Mississippi Conference of the Southeastern Association of Fish Alluvial Valley. In L.M. Smith, R.L. Pederson and Wildlife Agencies 62: 32–87. & R.M. Kaminski (eds.), Habitat Management Nakagawa, S. & Schielzeth, H. 2013. A general for Migrating and Wintering Waterfowl in North and simple method for obtaining R2 from America, pp. 203–247. Texas Tech University generalized linear-mixed effects models. Press, Lubbock, USA. Methods in Ecology and Evolution 4: 133– Schummer, M.L., Kaminski, R.M., Raedeke, A.H. 142. & Graber, D.A. 2010. Weather-related indices

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 450 Variation in waterfowl food densities

of autumn-winter dabbling duck abundance Mississippi State University, Mississippi State, in middle North America. Journal of Wildlife USA. Management 74: 94–101. Straub, J.N., Gates, R.J., Schultheis, R.D., Yerkes, Schummer, M.L., Cohen, J., Kaminski, R.M., T., Coluccy, J.M. & Stafford, J.D. 2012. Brown, M.E. & Wax, C.L. 2014. Atmospheric Wetland food resources for spring-migrating teleconnections and Eurasian snow cover as ducks in the upper Mississippi River and Great predictors of a weather severity index in Lakes region. Journal of Wildlife Management 76: relation to Mallard Anas platyrhynchos autumn– 768–777. winter migration. Wildfowl (Special Issue No. Tapp, J.L. 2013. Waterbird use and food 4): 451–469. availability on Wetland Reserve Program Soulliere, G.J., Potter, B.A., Coluccy, J.M., Gatti, Easements enrolled in the Migratory Bird R.C., Roy, C.L., Luukkonen, D.R., Brown, Habitat Initiative. M.Sc. thesis, University of P.W. & Eichholz, M.W. 2007. Upper Mississippi Missouri, Columbia, USA. River and Great Lakes Region Joint Venture Webb, E.B., Smith, L.M., Vrtiska, M.P. & Lagrange, Waterfowl Habitat Conservation Strategy. U.S. T.G. 2010. Effects of local and landscape Fish and Wildlife Service, Fort Snelling, variables on wetland bird habitat use during Minnesota, USA. migration through the rainwater basin. Journal Stafford, J.D., Kaminski, R.M., Reinecke, K.J. & of Wildlife Management 74: 109–111. Manley, S.W. 2006. Waste rice for waterfowl Weegman, M.M. 2013. Waterbird and seed in the Mississippi Alluvial Valley. Journal of abundances in Migratory Bird Habitat Initiative Wildlife Management 70: 61–69. and non-managed wetlands in Mississippi Stafford, J.D., Horath, M.M., Smith, R.V., Yetter, and Louisiana. M.Sc. thesis, Mississippi State A.P. & Hine, C.S. 2010. Historical and University, Mississippi State, USA. contemporary characteristics and waterfowl Williams, C.K., Dugger, B., Brasher, M., Collucy, use of Illinois River valley wetlands. Wetlands J., Cramer, D.M., Eadie, J.M., Gray, M., Hagy, 30: 565–576. H.M., Livolsi, M., McWilliams, S.R., Petrie, Straub, J.N. 2008. Energetic carrying capacity of M., Soulliere, G.J., Tirpak, J. & Webb, L. 2014. habitats used by spring migrating waterfowl Estimating habitat carrying capacity for in the Upper Mississippi and Great Lakes migrating and wintering waterfowl: Region during spring migration. M.Sc. considerations, pitfalls and improvements. thesis, Ohio State University, Columbus, Wildfowl (Special Issue No. 4): 407–435. USA. Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A. Straub, J.N. 2012. Estimating and modelling red & Smith, G.M. 2009. Mixed Effects Models and oak acorn production and abundance in the Extensions in Ecology with R. Springer, New Mississippi Alluvial Valley. Ph.D. thesis, York, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 436–450 451

Atmospheric teleconnections and Eurasian snow cover as predictors of a weather severity index in relation to Mallard Anas platyrhynchos autumn–winter migration

MICHAEL L. SCHUMMER1,5*, JUDAH COHEN2, RICHARD M. KAMINSKI3, MICHAEL E. BROWN4 & CHARLES L. WAX4

1Long Point Waterfowl, PO Box 160, Port Rowan, Ontario N0E 1M0, Canada. 2Atmospheric and Environmental Research, Lexington, Massachusetts USA. 3Mississippi State University, Department of Wildlife, Fisheries, and Aquaculture, Mississippi State, Mississippi 39762 USA. 4Department of Geosciences, Mississippi State University, Mississippi State, Mississippi 39762, USA. 5Current address. SUNY-Oswego, Department of Biological Sciences, 30 Centennial Drive, Oswego, New York 13126, USA. *Correspondence author. E-mail: [email protected] Abstract Research on long-term trends in annual weather severity known to influence migration and winter distributions of Mallard Anas platyrhynchos and other migratory birds is needed to predict effects of changing climate on: 1) annual distributions and vital rates of these birds, 2) timing of habitat use by migratory birds, and 3) demographics of the hunters of these species. Weather severity thresholds developed previously for Mallard were used to calculate weather severity and spatially- depicted Weather Severity Index Anomalies (± km2, WSIA), in comparison with normal conditions, for Mallard in eastern North America from November–January 1950–2008. We determined whether WSIA differed among decades and analysed the effects of atmospheric teleconnections and Eurasian snow cover on annual variation in WSIA. Weather severity was mildest (+ WSIA) during the 2000s compared to other decades and differed substantially from the 1960s and 1970s (– WSIA). The Arctic Oscillation Index explained substantial variation in WSIA during El Niño and La Niña episodes, but not when the Oceanic Niño Index was neutral. Eurasian snow cover models accurately predicted if the WSIA would be greater or less than normal for 75% of the years studied. Our results may provide a partial explanation for recent observations of interrupted or reduced migration to southern latitudes by Mallard and other migratory birds during autumn–winter. Our models also provide ecologists with teleconnection models to help predict future distributions of Mallard and potentially

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 452 Predictors of Mallard migration

other migratory birds in eastern North America. Future investigations could include testing the influence of WSIA on Mallard survival and on annual movements and distributions for other migratory birds, to provide a better understanding of the influences of climate and changes in climate on population dynamics and the need to conserve particular habitats.

Key words: Arctic Oscillation, climate, El Niño, migration, waterfowl, weather, winter.

The winter distribution of migratory birds is Northward shifts in winter range and influenced by a range of variables including changes in autumn migration phenology can food and habitat availability, weather, arise from changes in weather severity, food evolutionary and ecological mechanisms, and habitat availability, or a combination body condition and anthropogenic of these (Gordo 2007; Newton 2008; factors (Miller et al. 2005; Newton 2007, Schummer et al. 2010). Concurrent with 2008; Schummer et al. 2010). Evidence milder weather during winter, habitat has suggests that global climate change has become increasingly available to waterfowl lengthened growing seasons, increased through wetland conservation programmes winter temperatures and decreased snow in northern and mid-latitude regions of accumulation at many locations worldwide North America (NAWMP 2004; Dahl 2006). (Field et al. 2007). Concurrently, delays in This has led to considerable debate regarding autumn migration and changes in the birds’ the mechanism(s) determining current winter distributions have been documented winter distributions of Nearctic waterfowl in North America and Eurasia (Sokolov et al. (National Flyway Council and Wildlife 1999; Cotton 2003; La Sorte & Thompson Management Institute 2006, Greene and 2007; Brook et al. 2009; Sauter et al. 2010). Krementz 2008; Brook et al. 2009). Research Many migratory birds distribute annually into the long-term changes in weather along a latitudinal gradient as a function of conditions known to elicit southerly physiological tolerances to the severity of migrations by waterfowl therefore is needed winter weather (Root & Schneider 1995). to understand influences of weather severity Although food and habitat availability, and habitat availability on the birds’ winter refugia, flooding and other exogenous factors distribution and migration phenology (e.g. can mitigate influences of weather severity Schummer et al. 2010). Further, success of on bird energy budgets, climate remains a wetland restoration is often measured by primary determinant of winter distributions monitoring the recolonisation and use of of many species (Root 1988). Thus, climate habitats by wildlife (Jordan et al. 1987; warming is often cited as the mechanism Scodari 1997), though confirming the underlying a shift in range towards the pole outcome of restoration efforts is difficult (Crick 2004; Gordo 2007; La Sorte & without concurrent quantification of the Thompson 2007; Nevin et al. 2010). weather factors that influence distribution.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 453

Increasingly, studies are identifying eastern North America (i.e. the Snow changes in the frequency and amplitude of Advance Index; SAI). Thus, we reasoned atmospheric teleconnections as influencing that the SAI may predict severe winter temporal and spatial distributions of weather months in advance of traditional populations and species (Stenseth et al. atmospheric teleconnections and provide 2003; Wang & Schimel 2003). Atmospheric seasonal forecasts of waterfowl migration teleconnnections are defined as recurring, timing and intensity. persistent, large-scale patterns of Using temperature and snow data, atmospheric pressure, circulation and Schummer et al. (2010) developed several temperature anomalies occurring over competing models to explain rates of thousands of square kilometres (Bridgman change in relative abundance of Mallard & Oliver 2006). Prominent teleconnections Anas platyrhynchos at mid-latitude staging dominant in the North American sector areas in North America (i.e. Missouri) during include the El Niño Southern Oscillation autumn–winter. The model which best (ENSO), the Pacific North America explained annual variation in the rate of teleconnection pattern (PNA), the Arctic change in Mallard abundance was calculated Oscillation (AO) and the North Atlantic daily for November–January inclusive as the Oscillation (NAO; Bridgman & Oliver mean daily temperature –(°C)+ the number 2006). A substantial portion of the warming of consecutive days where the mean trends in North America over the past 20 temperature was ≤ 0°C + snow depth + the years have been attributed to sustained number of consecutive days with snow positive phases in atmospheric oscillations cover (i.e. the Weather Severity Index, WSI; (Serreze & Barry 2005; Hurrell & Deser Schummer et al. 2010). Temperatures of 2009). Further, some models indicate < 0°C were given a positive algebraic sign increased frequency and amplitude of (i.e. indicating more severe weather positive AO and NAO phases as a result of and accumulating positively when added global climate change (Corti et al. 1999; Gillet to other variables in the model), and et al. 2003). Positive phase teleconnections temperatures > 0°C were given a negative may result in increased frequency of mild sign. Here we use the WSI to rank severity winters and changes in winter distributions of winter weather for nearly six decades and and migration phenology of birds (Cotton determine differences in mean decadal rank 2003). The AO and NAO can only be for the winters 1950/51–2008/09 (hereafter forecast accurately c. 7 days in advance; thus 1950–2008) in eastern North America we sought additional indices of winter (see Fig. 1). Eastern North America was weather used to produce long-term seasonal selected because it corresponded with the forecasts for eastern North America (see Fig. geographic area in which the WSI was 1). Cohen and Jones (2011) detected a strong developed (Schummer et al. 2010) and correlation between the advance of Eurasian ecoregions used in conservation planning snow cover during October and the winter for Nearctic waterfowl (i.e. Mississippi and AO, which influences winter temperatures in Atlantic Flyways: Bellrose 1980; NAWMP

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 454 Predictors of Mallard migration

NOVEMBER DECE MBER

JANUARY

Legend HCN weather station Banding region 13 Banding region 14 Banding region 15 Banding region 16

Figure 1. Long-term mean location (solid line) ± 95% C.I. (dashed line) of the threshold weather severity index (WSI = 8) for Mallard in the Mississippi Flyway (USFWS banding regions 13–14) and the Atlantic Flyway (USFWS banding regions 15–16), for winters 1950–2008, calculated using Historical Climatology Network (HCN) data (black dots).

2012). The period 1950–2008 was used Methods because monitoring long-term metrics of Mallard breeding population and habitat Weather data dynamics (i.e. the Breeding Waterfowl and Sixty two weather stations in the United Habitat Survey: Baldassarre & Bolen 2006) States Historical Climatology Network commenced during the 1950s. We used data (HCN) at c. ≥ 35°N were selected to provide through to 2008, which was the end of the long-term state-wide weather data for North initial funding period for data compilation Dakota, South Dakota, Nebraska, Kansas, and analysis (Zimmerman 2009). As the Minnesota, Iowa, Missouri, Wisconsin, overall aim of the study was to investigate if Illinois, Michigan, Indiana, Kentucky, indices used to forecast winter weather (sensu Ohio, New York, Pennsylvania, West Cohen & Jones 2011) could be used to Virginia, Maryland, Virginia, Vermont, predict WSI and potentially provide longer- Delaware, New Jersey, New Hampshire, term projections of Mallard distributions Massachusetts, Connecticut, Rhode Island associated with climate change, we conclude and Maine (Quinlan et al. 1987; Williams et al. by relating our findings to global climate and 2006; Fig. 1). The median latitude HCN climate change models. station was selected within each state and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 455 inspected for missing temperature and snow Schummer et al. 2010); however, months data. If > 10% of daily temperature or snow with ≥ 7 consecutive days of missing WSI data were missing, we alternately selected values were omitted from the analyses. stations immediately north and south until a Annual monthly mean WSI values were weather station that contained < 10% calculated from the daily WSI values for missing data was identified. For states each HCN station, for November–January ≥ 28,500 km2 in area (i.e. larger than 1950–2008 (n = 10,788; PROC MEANS in Maryland), two additional weather stations SAS Institute 2009). For each month, we were also included (one in the north and also calculated a long-term (1950–2008) the other in the south of the state), starting monthly mean (± 95% C. I.s) from annual with the 25% and 75% quartile latitude monthly mean WSI values for each HCN weather stations and using the HCN station. Monthly PNA, AO and aforementioned criterion for missing data. NAO (November–January) and three- We also selected a station in Michigan’s month mean ENSO indices (September– upper peninsula which met this criterion. November, SON; October–December, Only two HCN weather stations in Nebraska OND; and November–January, NDJ) were were found that met our quality criteria. obtained from the National Oceanic Daily mean temperature and snow data and Atmospheric Administration/Climate were obtained for November–January Prediction Center at College Park, 1950–2008 from selected HCN weather Maryland, USA (NOAA 2010). stations to calculate a daily WSI for each station (n = 330,832 values), as follows: Spatial analyses WSI = Temp + Temp days + Snow depth + Interpolation analyses were conducted using Snow days (following Schummer et al. 2010), the natural neighbour method in ArcView where Temp = mean daily temperature (°C; Spatial Analyst, with a cell output size with temperatures < 0°C given a positive of 0.07 degrees (c. 7.7 km; ESRI 2009), to algebraic sign to indicate more severe depict the long-term average WSI spatially, weather, and those > 0°C a negative sign, as and to calculate ± 95% C.I. for monthly described above); Temp days = consecutive mean WSI values (for November– days with mean temperature ≤ 0°C; Snow January,1950–2008) from the 62 HCN depth = (snow depth, in cm) × 0.394; and weather stations (Fig. 1). Schummer et al. Snow days = consecutive days ≥ 2.54 cm of (2010) reported that daily WSI values of snow. Values for snow depth and snow days ≥ 7.2 coincided with an increased likelihood convert to 1 inch of snow, a measurement of Mallard leaving mid-latitude locations in used in weather forecasting and reporting North America. We assumed that Mallard in the United States. For days when reacted similarly to weather severity temperature and snow depth data were throughout the range of our present study missing, we interpolated missing values as as in our earlier study area (at c. 33˚–49˚N), being the mean of data recorded on days and a conservative threshold WSI value of 8 next before and after these date(s) (see therefore was used to demarcate by month

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 456 Predictors of Mallard migration and year the geographic extent of the WSI Decadal analysis of winter severity threshold value. Three polylines were index anomalies (WSIA) digitised as layers representing the long-term We summed WSIA from banding regions 13 average (LTA) WSI and the upper and lower and 14 (hereafter, Mississippi Flyway) for 95% C.I.s for each month of the study (Fig. each month and year, likewise summed data 1). We conducted the same interpolation for regions 15 and 16 (Atlantic Flyway) to analyses to produce a monthly WSI follow regional scales at which Mallard threshold demarcation line for November– migrate, and thus conservation planning January 1950–2008 inclusive (n = 232 decisions are made for Nearctic waterfowl maps). The areas above and below the 95% (Baldassarre & Bolen 2006), and because C.I.s were digitised as polygon layers to data from individual banding regions were 2 determine the area (km ) assumed to be not normally distributed. We also calculated available or unavailable to Mallard, based on mean WSIA within the Mississippi Flyway the LTA and 95% C.I.s (Fig. 1). The “extract (MSF), the Atlantic Flyway (AF) and the by mask” function in ArcView Spatial combined MSF and AF means for the three- Analyst (ESRI 2009) was then used to month period (November–January; NDJ). 2 develop polygon areas (i.e. km ) above and Data passed normality and equal variance below the 95% C.I. within the U.S. Fish and tests (P > 0.05). A one-way Analysis of Wildlife Service waterfowl banding regions Variance (ANOVA) therefore was used to 13–16 in eastern North America (hereafter, test the null hypothesis that there is no banding regions; Fig. 1). Banding regions difference between the six decadal periods were modified by extending the northern (1950s–2000s inclusive) in the WSIA values border to capture areas of southern Canada recorded (α = 0.05; Sokal & Rohlf 1981; where data interpolation of WSI occurred. Systat Software Inc. 2008). We predicted Within the banding regions, we subtracted that WSIA during the 2000s would be 2 polygon areas (i.e. km ) below the lower greater in comparison with other decades 95% C.I. from polygon areas above the because of observations of interrupted or upper 95% C.I. to determine changes to net delayed (i.e. “short-stopping”) migration by 2 potential habitat for Mallard (± km ; Mallard to southern latitudes in eastern hereafter the weather severity index North America in the last decade (Greene anomaly, WSIA) for each region. Thus, the & Krementz 2008; Brook et al. 2009; Nevin WSIA is the area assumed available or et al. 2010). unavailable for use by Mallard compared to the LTA by banding region on a monthly Effects of atmospheric basis (November–January) for each year teleconnections (1950–2008). Data were summarised by We used an information-theoretic approach banding regions because they represent (Burnham & Anderson 2002) to investigate different Mallard populations and migratory climatic factors potentially influencing flyways (i.e. the Mississippi and Atlantic variation in WSIA. Five candidate Flyways). teleconnection-based models, associated

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 457 with weather patterns in eastern North long-term regional weather patterns America, were developed a priori for this becomes difficult (Bridgman & Oliver 2006). analysis as listed below. North Atlantic Oscillation (NAO) El Niño Southern Oscillation (ENSO) The NAO is often associated with Climate variability resulting from ENSO interdecadal and interannual shifts in can produce wide-ranging ecological ecological processes in marine and terrestrial consequences (Stenseth et al. 2003). ENSO systems of Europe and North America is most often characterised as El Niño, La (Hurrell et al. 2003). A positive NAO index Niña or Neutral. El Niño occurs when increases the likelihood of mild, wet winters the equatorial Pacific Ocean sea surface in the eastern United States while increased temperatures (SSTs) are unusually warm, cold accompanies a negative NAO index over whereas La Niña occurs when the same the same area (Hurrell 1995). Because the region is dominated by unusually cold SSTs, NAO is also correlated with the Arctic and Neutral occurs when SSTs are near Oscillation (AO), there is considerable debate normal. The Oceanic Niño Index (ONI) is as to whether the NAO is merely a regional the principal tool for monitoring and expression of the AO, which is a hemispheric assessing ENSO. To be considered a full El mode of climatic variability (Ambaum et al. Niño or La Niña episode, the ONI must 2001; Hurrell et al. 2003). Moreover, scientists exceed ± 0.5 (˚C) for ≥ 5 consecutive are uncertain if the NAO or the AO has the overlapping 3-month seasons (Tozuka et al. greatest influence on winter climates and 2005; NOAA 2010). During winter months, related ecological processes in North America a positive ONI (El Niño, ≥ +0.5) is (Aanes et al. 2002; Stenseth et al. 2003). associated with increased precipitation and colder than normal temperatures in the Arctic Oscillation (AO) southeast United States, but warmer than Effects of the AO on North American normal temperatures and reduced ice climate are most pronounced from coverage in the Great Lakes region and December through March (Serreze & Barry eastern Canada (Ropelewski & Halpert 1987; 2005). A positive AO index reduces the Halpert & Ropelewski 1992; Assel et al. number of cold air intrusions east of the 2000). A negative ONI (La Niña, ≤ –0.5) Rocky Mountains, causing much of the commonly results in warmer and drier eastern United States to experience warmer conditions over the southeast United States, winters than normal. A negative AO index but colder than normal air penetrating into produces the opposite effect over the same the northern Great Plains of Canada and the area (Serreze & Barry 2005; Bridgman & United States (Ropelewski & Halpert 1987; Oliver 2006). Halpert & Ropelewski 1992). During ONI neutral phases (–0.5 to 0.5), equatorial Pacific Pacific North American Oscillation (PNA) Ocean SSTs appear to have much less Like the AO, the PNA tends to be most influence on global climates and forecasting pronounced during the winter months. The

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 458 Predictors of Mallard migration positive phase of the PNA is associated with and PNA among the ONI categories of El above average temperatures in western Niño, La Niña and Neutral (P < 0.05). Data Canada and the Pacific coast of the United passed normality and equal variance tests States, but below average temperatures (P > 0.05, Sokal & Rohlf 1981). If an across the south-central and southeast ANOVA approached significance (P < 0.10), United States. The PNA has also been we did not include that atmospheric associated with moisture variability in the teleconnection in models containing ONI same region (Coleman & Rogers 1995; categories. Positive associations were Rogers & Coleman 2003). Anomalies in detected between the AO and NAO (0.64 ≤ ≤ regional temperatures, ice cover and snow r58 0.76, P < 0.001) and the ONI and ≤ ≤ cover have been related to the PNA phase PNA (0.37 r58 0.40, P < 0.01). The PNA ≤ (Assel 1992; Serreze et al. 1998). Also, also varied with ONICAT (3.49 F2,55 positive PNA index values are more ≤ 8.16, P < 0.05) but not with the AO and ≤ ≤ frequent during El Niño episodes (Wang & NAO (0.64 F2,55 0.09, P > 0.05). Because Fu 2000). of the latter results, we did not include the AO and PNA within models containing the Plausible interactions and combined influences NAO and ONICAT, respectively. Thus, we Several studies have investigated modulation evaluated 18 candidate models for explaining of the ONI by other atmospheric variation in WSIA: 1) AO, 2) NAO, 3) PNA, teleconnections (Gershunov & Barnett 1998, 4) ONI (categorical, ONICAT), 5) ONI Bridgman & Oliver 2006) and the combined (continuous, ONI), 6) ONICAT AO, influences of atmospheric teleconnections 7) ONICAT NAO, 8) ONICAT × AO, on winter temperature, precipitation 9) ONICAT × NAO, 10) ONI AO, 11) ONI and snowfall (Serreze et al. 1998). For NAO, 12) ONI × AO, 13) ONI × NAO, example, during El Niño or La Niña 14) PNA AO, 15) PNA NAO, 16) PNA × episodes, inclusion of other atmospheric AO, 17) PNA × NAO, and 18) the NULL. In teleconnections as covariates improves addition to appropriate multivariate models, explained variation in precipitation and we modelled interactions of continuous temperature throughout much of North variables and ONI category to determine if America (Higgins et al. 2000). However, including the relationship among slopes several atmospheric teleconnections are improved the model fit. often correlated and their combination in An Akaike’s Information Criterion (AICc) models could create statistical bias. We was calculated for each model (PROC calculated Pearson’s correlation coefficients MIXED; SAS Institute 2009). Competing for relationships between continuous models were ranked according to Δ AICc variables (NAO, AO, PNA) and did not values and selection was based on the lowest include combinations of model predictors Δ AICc value (Burnham & Anderson 2002; that were correlated (r ≥ 0.70; Dormann et al. Littell et al. 2007). We considered models 2013). We also used ANOVA to test the null competitive when Δ AICc values were ≤ 2 hypothesis of no difference in NAO, AO units of the best model (i.e. with Δ AICc =

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 459

0; Burnham & Anderson 2002). Year was 2009 to determine whether SAI was better included as a random variable and variance than random chance of it predicting WSIA. components (VC) were used from a suite of SAI is a strong predictor for AO (Cohen tested covariance structures (i.e. CS, UN, & Jones 2011) and NAO and AO are TOEP, AR(1), ARH(1), VC), because VC correlated (Hurrell et al. 2003) so we only produced the best fit models (lowest AICc included SAI and the null in our analysis. values: Littell et al. 2007). We calculated Analyses were conducted on 1972–2008

Akaike weights (wi) to assess relative support data because the SAI values were available for each atmospheric teleconnection- only for 1972 onwards. Whether removing based model in explaining variation in ONI-neutral years improved the utility of WSIA. When top and competing models the model for predicting WSIA was also contained an interaction effect, results investigated. were interpreted from the slopes and intercepts of the relationship between the Results dependent and interacting explanatory Decadal analysis of WSIA variables (Gutzwiller & Riffell 2007; SAS Institute 2009). The WSIA rank recorded for NDJ each year did not differ significantly across the decades Effects of the snow advance index (SAI) (F5,52 = 1.31, P = 0.26, n.s.; Table 1). The To evaluate whether the SAI was better 2000s had the highest mean WSIA but also than random at predicting WSIA during greatest range in values (Table 1; Appendix November–January, AICc values for the SAI 1). Inspection of data indicated that winter model were compared with those for the 2000/01 was a statistical outlier (with WSIA null models (PROC MIXED; SAS Institute > 200 units in comparison with the next

Table 1. Descriptive statistics for Weather Severity Index Anomalies (WSIA , ± thousands of km2 ), by decade, for November–January in eastern North America.

Decade n Mean (± s.e.) Median 25% Quartile 75% Quartile WSIA WSIA WSIA WSIA (× 1,000 km2) (× 1,000 km2) (× 1,000 km2) (× 1,000 km2)

1950s 9 69.0 (50.5) 72.5 –77.4 210.0 1960s 10 –4.9 (25.1) –15.8 –57.6 52.9 1970s 10 –33.5 (32.9) –16.2 –123.5 74.0 1980s 10 68.7 (57.9) 60.2 –51.9 252.5 1990s 10 14.3 (43.9) 8.5 –113.3 110.3 2000s 9 129.2 (85.4) 129.4 49.6 272.1

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 460 Predictors of Mallard migration coldest winters, whereas WSIA values for the portion of the variation in WSIA was two warmest winters differed by c. 10 units) explained by the AO during El Niño and La and ranked as the most negative WSIA in Niña episodes but not during Neutral the 58 years of the study (Appendix 1). conditions (Table 2, Fig. 2). On removing 2000/01 data, the mean WSIAs differed significantly among decades Effects of SAI

(ANOVA: F5,51 = 2.99, P = 0.02). Post-hoc The SAI (AICc = 477.3) was a better Tukey pair-wise comparisons with 2000/01 predictor of WSIA than the null model data removed detected that WSIA was (AICc = 481.9), but utility of the model for greater in the 2000s (x– = 197.7 ± 57.9) than predicting WSIA was relatively poor (Fig. 3). the 1970s (P = 0.01) and 1960s (P = 0.04) but Including the ONI category (El Nino, La no other decadal differences were detected Nina or Neutral) did not improve the for WSIA (P ≥ 0.30, n.s.; Table 1). predictive ability of our model (AICc = 480.7). Despite poor model fit, the SAI did Effects of atmospheric predict correctly whether WSIA would be teleconnections positive (mild winter compared to normal) Models containing interactions of AO or or negative (severe winter compared to NAO with ONICAT were the only models normal) in 27 (75%) of 36 years (Fig 3). that were ≤ 2 Δ AICc units from the best model for all locations and time periods Discussion (Table 2). The AO and NAO explained a So far as we are aware, this study is the first greater portion of variation (R2) in WSIA to assess the influence of long-term trends during El Niño and La Niña episodes than in annual weather severity (i.e. ± WSIA) on during Neutral conditions for all locations the autumn–winter distribution of Mallard and periods. Models derived using 3-month in the Nearctic (Schummer et al. 2010). We means generally were better fit models (i.e. detected a trend toward less severe weather 2 by wi and R ) than those using monthly data. (+ve WSIA) during recent decades (1990s In the Mississippi Flyway, AO had a greater and 2000s) but also detected that severe influence on WSIA during NDJ, whereas events (–ve WSIA) occurred during this the NAO was more influential in the period (e.g. winter 2000/2001). The results Atlantic Flyway (Table 2). Greatest weight of identified that WSIA was positively related to evidence explaining variation in WSIA for the Arctic and North Atlantic Oscillation Mississippi and Atlantic Flyways combined Indices during El Niño and La Niña was associated with the interaction of AO episodes but not during ONI Neutral with ONICAT (Table 2). However, AICc conditions. Similar to other studies, Arctic values for the interaction of NAO and and North Atlantic Oscillation Indices ONICAT were < 2 units from the top were found to be strongly correlated model (ONICAT × AO) and received (Ambaum et al. 2001; Hurrell et al. 2003). substantial weight of evidence (Table 2). For Our models indicate that weather severity, eastern North America, a substantial which is known to influence autumn–winter

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 461 d by Category 2 R i AIC cw AIC c Δ ) are included. i c Models ONICAT × AOONICAT 1409.3 × NAOONICAT 4.80 1449.4 0.08 0.01 × NAOONICAT EL (0.00), LA (0.30), NA (0.00) 1417.7 0.49 ONICAT × AO EL (0.31), LA (0.21), NA (0.08) 2.30 ONICAT × NAO 1374.9 0.24 1376.3 0.00 EL (0.01), LA (0.17), NA (0.10) × NAO 1.40ONICAT 0.67 1378.4 0.33 EL (0.14), LA (0.55), NA (<0.00) EL (0.18), LA (0.32), NA (<0.00) 0.01 × NAOONICAT 1321.6 0.49 × NAOONICAT EL (0.47), LA (0.49), NA (0.01) 1292.3 5.20 × AOONICAT 0.07 0.00 × AOONICAT 1293.6 EL (0.20), LA (0.25), NA (0.01) 0.66 × NAOONICAT 1401.0 1.30 EL (0.66), LA (0.48), NA (0.00) 1402.3 0.34 0.00 1.30 EL (0.60), LA (0.55), NA (0.04) 0.66 0.34 EL (0.32), LA (0.64), NA (0.01) EL (0.28), LA (0.27), NA (0.01) El Niño, La Niña and Neutral), AO = Arctic Oscillation Index, NAO = North = Arctic Oscillation Index, NAO La Niña and Neutral), AO El Niño, b Period November (N)November × NAO ONICAT 1404.5 0.00 0.92 EL (0.02), LA (0.23), NA (0.06) December (D) × AO ONICAT (J)January 1449.3 0.00NDJ × AO ONICAT 0.51 EL (0.16), LA (0.51), NA (0.12) 1415.4December (D) 0.00 × AO ONICAT 0.76 (J)January 1378.3 EL (0.25), LA (0.46), NA (0.03) 0.00NDJ × AO ONICAT 0.51 EL (0.41), LA (0.60), NA (0.04) 1316.4 0.00 0.93 EL (0.34), LA (0.43), NA (0.06) by category:by LA = La Niña and NA Neutral. EL = El Niño, 2 a R ) for linear relationships between Weather Severity Index Anomalies (WSIA, ± thousands Severity Information Weather Criteria (AIC c ) for linear relationships between Akaike’s ) and atmospheric teleconnection models in the Mississippi and Atlantic Flyways, in November–January 1950–2008. Eighteen in November–January ) and atmospheric teleconnection models in the Mississippi Atlantic Flyways, 2 November is not included in the Atlantic Flyway because most observationsof outside were is not included in the Atlantic Flyway November 1). the study area (see Fig. ONICAT = Oceanic Niño Index categoryONICAT ( e.g. derived derived Mississippi Flyway = USFWS banding regions 13 and 14; Atlantic Flyway = USFWS banding regions 15 and 16 (see Fig. 1). = USFWS banding regions 15 and 16 (see Fig. USFWS banding regions 13 and 14; Atlantic Flyway = Mississippi Flyway candidate models for explaining variation in WSIA were considered for each time period (including combinations of AO, NAO, PNA, time period (including combinations of NAO, for each considered in WSIA were candidate models for explaining variation AO, models in each further and year; Methods section provides ONI (ONI), categorical details); the top two continuous ONI (ONICAT) case are presented here. km Table 2. Table Location Mississippi Flyway a b c d Atlantic Oscillation Index. Only models with weight ofAtlantic Oscillation Index. Only models with weight evidence ( w Atlantic Flyway Mississippi and Atlantic Flyways NDJ

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 462 Predictors of Mallard migration

El Niño La Niña Neutral

600 600 600 R2=0.32, n = 18 R2=0.64, n = 19 R2=0.01, n = 21 WSIA = 75.98 + (85.31 × AO) WSIA = 45.60 + (136.30 × AO) 400 400 400

200 200 200 × 1,000) 2

0 0 0

–200 –200 –200

–400 –400 –400 WSIA area (km

–600 –600 –600 –3 –2 –1 0 1 2 3 –3 –2 –1 0 1 2 3 –3 –2 –1 0 1 2 3 Arctic Oscillation Index

Figure 2. Relationships between the Arctic Oscillation Index (AO) and Weather Severity Index Anomalies (WSIA, ± thousands of km2) in eastern North America by Oceanic Niño Index category (El Niño, La Niña and Neutral) for November–January, 1950–2008.

Mallard migration (Bellrose 1980; Schummer potential explanation for recent observations et al. 2010), is reduced during El Niño and La of delayed autumn–winter migration in Niña episodes when the Arctic and North Nearctic Mallard and other migratory birds Atlantic Oscillation Indices are in a positive (National Flyway Council and Wildlife phase. Thus, we think Mallard migration Management Institute 2006; Nevin et al. 2010) may be interrupted or delayed (i.e. and may be helpful in modelling future “short-stopping”) during these conditions. autumn–winter distributions of Mallard (and Investigation of the usefulness of the SAI, possibly other migratory birds) in eastern as a long-term seasonal predictor of weather North America and more widely (see Notaro influencing Mallard migration, provided et al. 2014). Colleagues are encouraged to test mixed results. The SAI predicted accurately for the influence of WSIA on the movements the coming winter (NDJ) as being either less and annual distribution of Mallard and other or more severe than normal 75% of the waterfowl, its influence on waterfowl hunter time, but its capacity to predict WSIA was demographics and behaviour, and whether limited. We suggest continued investigation WSIA affects survival rates and trends in of the SAI with additional weather indices. population size. To facilitate these analyses we Long-term forecasting of weather known to suggest development of weather indices for influence Mallard and other waterfowl other waterfowl and web-based tools for migrations would be helpful for managers distribution of these data to scientists and charged with providing waterfowl habitat at managers. Overall, the results suggest that key times of migration throughout the non- recent observations of delayed waterfowl breeding season and for managing timing of migration (National Flyway Council and hunting seasons (Schummer et al. 2010). Wildlife Management Institute 2006; Brook et The results of the study provides a al. 2009) may be related to reduced weather

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 463

600 R2 = 0.17, n = 36 WSIA = 37.96 + (–75.88 × SAI)

400

200 × 1,000 ) 2

0

–200 WIS A area (k m

–400

–600 –4 –2 0 2 4 Snow advance index Figure 3. Relationship between the Snow Advance Index (SAI) and Weather Severity Index Anomalies (WSIA, ± thousands of km2) in eastern North America for November–January 1972–2008. Data within blue areas were correctly designated as WSIA greater or lesser than normal by SAI (27 of 36 years, 75%), whereas red areas were incorrectly designated. severity at northern and mid-latitudes in waterfowl (Greene & Krementz 2008). In eastern North America. addition, such information could potentially The results can also be used for aid conservationists in predicting future determining the relative contributions of winter distributions of Nearctic waterfowl weather severity and habitat availability to and habitat protection, management and the winter distribution and migration restoration needs under various climate phenology of waterfowl and other migratory change models (e.g. Ruosteenoja et al. 2003; wildlife. WSIA was quantified here for La Sorte & Jetz 2010). Further, WSIA could Mallard in eastern North America, but be included in models used to determine the availability and quality of wetland habitat, benefits of wetland restoration because human-related disturbance, waterfowl habitat use by birds (i.e. a metric of population sizes, and other factors restoration success) can be highly dynamic potentially influencing habitat use, migration and may be related to the severity of winter and population dynamics should also be weather in addition to habitat quality and taken in account (Bellrose 1980; Kaminski & other metrics (Newton 1998). Gluesing 1987; Newton 2008). Such an A sustained positive Arctic Oscillation evaluation would add clarity to the debate Index could result in increased frequency of regarding the mechanism(s) that influence mild winters (+ve WSIA), during which the current winter distributions of Nearctic decreased ice cover, snowfall and increased

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 464 Predictors of Mallard migration temperatures may allow Mallard to remain at 2010). We encourage including a broader suite more northern latitudes during autumn– of influences such as potential species winter (Schummer et al. 2010). Overall, interactions, physiology and energy- reduced weather severity may result in dependent “bottle-necks” at different stages delayed migration or reduced numbers of of migration, concurrent habitat changes, and Mallard and possibly other birds migrating other biotic interactions, to increase biological to southern latitudes in North America. realism in future analyses (Seavy et al. 2008; La Nearctic waterfowl (millions of birds), use a Sorte & Jetz 2010). Further examination of diversity of aquatic and terrestrial foraging Mallard distribution using satellite telemetry niches, and can feed at rates capable of and volunteer observation programmes (e.g. causing strong trophic influences (Newton Christmas Bird Count, Mallard Migration 1998; Abraham et al. 2005; Baldassaree & Network) in relation to the WSI and WSIA Bolen 2006). Sustained northern shifts in would provide further validation of temporal autumn–winter distributions of these and spatial distributions of these birds. Other abundant species could increase foraging Nearctic waterfowl and migratory species may pressure at northern latitudes while reducing react differently to annual variation in winter such effects at southerly locations and severity and changing climate (Crick 2004; cause changes in trophic relationships (i.e. Sauter et al. 2010), for instance because habitat trophic cascades) throughout eastern North generalists (e.g. Mallard) often respond to America (Crick 2004; Inkley et al. 2004). An climate change more readily than habitat increase in foraging intensity at more specialists (La Sorte & Jetz 2010). Thus, we northern latitudes during autumn and winter also encourage continued research aimed at may also deplete the food available for understanding the effects and threats of a waterfowl during spring migration in some changing climate for a variety of migratory locations (Straub et al. 2012; Greer et al. species (Thomas et al. 2001; Walther et al. 2007; Long Point Waterfowl, unpubl. data). 2002). Predicting future distributions of waterfowl using forecasts of WSIA may help Acknowledgements conservationists develop adaptive plans to Finances were provided by the Forest meet the habitat needs of waterfowl and and Wildlife Research Center (FWRC), other migratory birds in a changing climate Mississippi State University (MSU); the James (Lehikoinen et al. 2006; Seavy et al. 2008; La C. Kennedy Endowed Chair in Waterfowl Sorte & Jetz 2010, Notaro et al. 2014). and Wetlands Conservation; the United We used a simplistic WSI developed for States Forest Service, Center for Bottomland Mallard and other dabbling ducks (Schummer Hardwoods Research, Stoneville, Mississippi; et al. 2010) to examine their potential past and and the Department of Geosciences, MSU. future distributions. Results from our study Statistical advice was provided by G. Wang. corroborate those for European Mallard We thank D. Krementz and B. Cooke for which showed reduced winter migration providing helpful comments during the distance with long-term warming (Sauter et al. development of the manuscript. We also

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 465 thank C. Zimmerman for completing a pilot Burnham, K.P. & Anderson, D.R. 2002. Model project (Zimmerman 2009) and for assisting Selection and Multimodel Inference: a Practical with compilation of weather data. L. Tucker Information-Theoretic Approach. Second Edition. assisted with weather data collection and Springer-Verlag, New York, New York, USA. error checking. J. Cohen is supported by the Bridgman, H.A. & Oliver, J.E. 2006. The global climate system: patterns, processes and teleconnections. National Science Foundation grant BCS- Cambridge University Press, Cambridge, 1060323. Our manuscript has been approved UK. for publication as MSU-FWRC journal article Cohen, J. & Jones, J. 2011. A new index for WFA-310. more accurate winter predictions. Geophysical Research Letters 38: L21701. References Coleman, J.S.M. & Rogers, J.C. 1995. Ohio Aanes, R., Saether, B.E., Smith, F.M. Cooper, E.J., River Valley winter moisture condition Wookey, P.A. & Øritsland, N.A. 2002. The associated with the Pacific-North American Arctic Oscillation predicts effects of climate teleconnection pattern. Journal of Climate 16: change in two trophic levels in a high-arctic 969–981. ecosystem. Ecology Letters 5: 445–453. Corti, S., Molteni, F. & Palmer, T.N. 1999. Abraham, K.F., Jefferies, R.L. & Rockwell, R.F. Signature of recent climate change in 2005. Goose-induced changes in vegetation frequencies of natural atmospheric circulation and land cover between 1976 and 1997 in an regimes. Nature 398: 799–802. Arctic coastal marsh. Arctic, Antarctic, and Cotton, P.A. 2003. Avian migration phenology Alpine Research 37: 269–275. and global climate change. Proceedings of Assel, R.A. 1992. Great Lakes winter-weather 700- Nature and Science 100: 12219–12222. hPa PNA teleconnections. Monthly Weather Crick, H.Q.P. 2004. The impact of climate Review 120: 2156–2163. change on birds. Ibis 146:48–56. Assel R.A., Janowiak, J., Boyce, D., O’Connors, C., Dahl, T.E. 2006. Status and Trends of Wetlands in Quinn, F.H. & Norton, D.C. 2000. Laurentian the Conterminous United States 1998 to 2004. Great Lakes ice and weather conditions for U.S. Department of the Interior, Fish and the 1998 El Niño winter. Bulletin of the Wildlife Service, Washington D.C., USA. American Meteorological Society 81: 703–718. Dormann, C.F., Elith, J., Bacher, S., Buchmann, Ambaum, M.H.P., Hoskins, B.J. & Stephenson, C., Carl, G., Carré, G., Diekötter, T., García D.B. 2001. Arctic Oscillation or North Atlantic Marquéz, J., Gruber, B., Lafourcade, B., Oscillation? Journal of Climate 14: 3495–3507. Leitão, P.J., Münkemüller, T., McClean, C., Baldassarre, G.A. & Bolen, E.G. 2006. Waterfowl Osborne, P., Reineking, B., Schröder, B., Ecology and Management. Krieger, Malabar, Skidmore, A.K., Zurell, D., Lautenbach, S. Florida, USA. 2013. Collinearity: a review of methods to Bellrose, F.C. 1980. Ducks, Geese and Swans of North deal with it and a simulation study evaluating America. Stackpole Books, Mechanicsburg, their performance. Ecography 36: 27–46. Pennsylvania, USA. Environmental Systems Research Institute (ESRI). Brook, R.W., Ross, R.K., Abraham, K.F., 2009. ArcGIS 9.2 Desktop Help. Accessible at Fronczak, D.L. & Davies, J.C. 2009. Evidence www.webhelp.esri.com/arcgisdesktop/9.2/ for black duck winter distribution change. index.cfm? TopicName=welcome (last Journal of Wildlife Management 73: 98–103. accessed 15 June 2014).

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 466 Predictors of Mallard migration

Field, C. B., Mortsch, L.D., Brklacich, M., Forbes, Halpert, M.S. & Ropelewski, C.F. 1992. Surface D.L., Kovacs, P., Patz, J.A., Running, S.W., temperature patterns associated with the Scott, M.J., Andrey, J., Cayan, D., Demuth, M., Southern Oscillation. Journal of Climate 5: Hamlet, A., Jones, G., Mills, E., Mills, S., 577–593. Minns, M.K., Sailor, D., Saunders, M., Scott, Higgins, R.W., Leetmaa, A., Xue, Y. & Barston, A. D., Solecki, W., and MacCracken, M. 2000. Dominant factors influencing the 2007. North America: Climate change seasonal predictability of U.S. precipitation 2007: impacts, adaptation and vulnerability and surface air temperature. Journal of Climate – contribution of working group II to 13: 3994–4017. the fourth assessment report of the Hurrell, J.R. 1995. Decadal trends in the North Intergovernmental Panel on Climate Change. Atlantic Oscillation and relationships to In M.L. Parry, F. Canziani, J.P. Palutikof, P.J. regional temperature and precipitation. Science van der Linden & C.E. Hanson (eds.), North 269: 676– 679. America: Climate Change 2007, pp. 617– 620. Hurrell, J.R. & Deser, C. 2009. North Atlantic Cambridge University Press, Cambridge, UK. climate variability: The role of the North Gershunov, A. & Barnett, T.P. 1998. Interdecadal Atlantic Oscillation. Journal of Marine Systems modulation of ENSO teleconnections. Bulletin 78: 28–41. of the American Meteorological Society 80: 2715– Hurrell, J.R., Kushnir, Y., Visbeck, M. & Ottersen, 2725. G. 2003. An overview of the North Atlantic Gillet, N.P., Baldwin, M.P. & Allen, M.R. 2003. Oscillation. The North Atlantic Oscillation: Climate change and the North Atlantic Climate Significance and Environmental Oscillation. The North Atlantic Oscillation: Impact. Geophysical Monograph Series 134: 1–35. Climate Significance and Environmental Inkley, D.B., Anderson, M.G., Blaustein, A.R., Impact. Geophysical Monograph 134: 193–199. Burkett, V.R., Felzer, B., Griffith, B., Price, J. Gordo, O. 2007. Why are bird migration dates & Root, T.L. 2004. Global climate change and shifting? A review of weather and climate wildlife in North America, Wildlife Society effects on avian migration phenology. Climate Technical Review 04-2. The Wildlife Society, Research 35: 37–58. Bethesda, Maryland, USA. Greene, A.W. & Krementz, D.G. 2008. Mallard Jordan, W.R., Gilpin, M.E., & Aber, J.D. 1987. harvest distributions in the Mississippi and Restoration ecology: A Synthetic Approach to Central Flyways. Journal of Wildlife Management Ecological Research. Cambridge University 72: 1328–1334. Press, Cambridge, UK. Greer, A.K., Dugger, B.D., Graber, D.A. & Petrie, Kaminski, R.M. & Gluesing, E.A. 1987. M.J. 2007. The effects of seasonal flooding on Density-and habitat-related recruitment in seed availability for spring migrating waterfowl. mallards. Journal of Wildlife Management 51: Journal of Wildlife Management 71: 1561–1566. 141–148. Gutzwiller, K.J. & Riffell, S.K. 2007. La Sorte, F.A. & Jetz, W. 2010. Avian Using statistical models to study temporal distributions under climate change: towards dynamics of animal-landscape relations. In improving projections. Journal of Experimental J.A. Bissonette & I. Storch (eds.), Temporal Biology 213: 862–869. Dimensions of Landscape Ecology: Wildlife La Sorte, F.A. & Thompson, F.R. 2007. Poleward Responses to Variable Resources, pp. 93–118. shifts in winter ranges of North American Springer Inc., New York, New York, USA. birds. Ecology 88: 1803–1812.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 467

Lehikoinen, A., Kilpi, M. & Öst, M. 2006. Winter change: northward shifts in early winter climate affects subsequent breeding success abundance. American Birds 109: 10–15. of common eiders. Global Change Biology 12: Notaro, M., Lorenz, D., Hoving, C., & Schummer, 1355–1365. M.L. 2014. 21st Century projections of Littell, R.C., Milliken, G.A., Stroup, W.W., snowfall and winter severity across central- Wolfinger, R.D. & Schabenberger, O. 2007. eastern North America. Journal of Climate SAS for Mixed Models. Second edition. SAS online early: doi: http://dx.doi.org/10.1175/ Institute Inc., Cary, North Carolina, USA. JCLI-D-13-00520.1. Miller, R.M., Takekawa, J.Y., Fleskes, J.P., Quinlan, F.T., Karl, T.R. & Williams, C.N., Jr. 1987. Orthmeyer, D.L., Casazza, M.L. & Perry, W.M. United States Historical Climatology Network 2005. Spring migration of Northern Pintails (HCN) Serial Temperature and Precipitation Data. from California’s Central Valley wintering area Carbon Dioxide Information Analysis tracked with satellite telemetry: routes, timing Center, Oak Ridge National Laboratory and destinations. Canadian Journal of Zoology 83: NDP-019, Oak Ridge, Tennessee, USA. 1314–1332. Rogers, J.C. & Coleman, J.S.M. 2003. National Flyway Council and Wildlife Management Interactions between the Atlantic/ Institute (NFC and WMI). 2006. National Multidecadal Osciallation, El Niño/La Niña, Duck Hunter Survey 2005: National Report. and the PNA in winter Mississippi Valley Wildlife Management Institute, Washington stream flow. Geophysical Research Letters 30: D.C., USA. 1518–1522. National Oceanic and Atmospheric Ropelewski, C F. &. Halpert, M.S. 1987. Global and Administration (NOAA). 2010. Climate regional scale precipitation patterns associated Prediction Center: El Niño-Southern Oscillation with the El Niño/Southern Oscillation. Index. Available at www.cpc.noaa.gov/ Monthly Weather Review 115: 1606– 1626. products/precip/CWlink/MJO/enso.shtml Root, T. 1988. Energy constraints on avian #history (last accessed 6 May 2012). distributions and abundances. Ecology 69: North American Waterfowl Management Plan 330–339. (NAWMP) Committee. 2004. North American Root, T. & Schneider, S.H. 1995. Ecology and Waterfowl Management Plan. 2004. Implementation climate: research strategies and implications. Framework: Strengthening the Biological Foundation. Science 269: 334–341. Canadian Wildlife Service, U.S. Fish & Ruosteenoja, K., Carter, T.R., Jylha, K., & Wildlife Service and Secretaria de Medio Tuomenvirta, H. 2003. Future climate in World Ambiente y Recursos Naturales. Accessible at Regions: an Intercomparison of Model-based http://www.fws.gov/birdhabitat/NAWMP/ Projections for the New IPCC Emissions Scenarios. Planstrategy.shtm (last accessed 15 June 2014). Finnish Environment Institute, Helsinki, Newton, I. 1998. Population Limitation in Birds. Finland. Academic Press, San Diego, USA. SAS Institute. 2009. SAS/STAT User’s Guide. SAS Newton, I. 2007. Weather-related mass mortality Institute, Cary, North Carolina, USA. events in migrants. Ibis 149: 453–467. Sauter, A., Korner-Nievergelt, F. & Jenni, L. Newton, I. 2008. The Migration Ecology of Birds. 2010. Evidence of climate change effects Academic Press, San Diego, USA. on within-winter movements of European Nevin, D.K., Butcher, G.S. & Bancroft, G.T. Mallards Anas platyrhynchos. Ibis 152: 600– 2010. Christmas bird counts and climate 609.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 468 Predictors of Mallard migration

Schummer, M.L., Kaminski, R.M., Raedeke, A.H. Systat Software, Inc. 2008. Systat 11.0. Richmond, & Graber, D.A. 2010. Weather-related indices California, USA. of autumn–winter dabbling duck abundance Thomas, D.E., Blondel, J., Perret, R., Lambrechts, in middle North America. Journal of Wildlife M.M. & Speakman, J.R. 2001. Energetic and Management 74: 94–101. fitness costs of mismatching resource supply Scodari, P.F. 1997. Measuring the Benefits of Federal and demand in seasonally breeding birds. Wetland Programs. Environmental Law Institute, Science 291: 2598–2600. Washington D.C., USA. Tozuka, T, Luo, J.-J., Masson, S., Behera, S.K. & Seavy, N.E.,. Dybala, K.E. & Snyder, M.A. 2008. Yamagata, T. 2005. Annual ENSO simulated Perspectives in ornithology: climate models in a coupled ocean atmosphere model. and ornithology. Auk 125: 1–10. Dynamics of Atmospheres and Oceans 39: 41–60. Serreze M.C. & Barry, R.G. 2005. The Arctic US Fish and Wildlife Service (USFWS). 2010. Climate System. Cambridge University Press, Upper Mississippi River National Wildlife Refuge Cambridge, UK. Fall Flight Surveys: How are Fall Flights Conducted? Serreze, M.C., Clark, M.P., McGinnis, D.L., & Accessible at http://www.fws.gov/midwest/ Robinson, D.A. 1998. Characteristics of UpperMississippiRiver/Documents/InfoFF. snowfall over the eastern half of the United pdf (last accessed 20 October 2010). States and relationships with principal modes Walther, G., Post, E., Convey, P., Menzel, A., of low-frequency atmospheric variability. Parmesan, C., Beebee, T.C.J., Fromentin, J., Journal of Climate 11: 234–250. Hugh-Guldberg, O. & Bairlein, F. 2002. Sokal, R.R., & Rohlf, F.J. 1981. Biometry. W.H. Ecological response to recent climate Freeman and Company, San Francisco, USA. change. Nature 416: 389–395. Sokolov, L.V., Markovets, M.Y. & Morozov, Y.G. Wang, G. & Schimel, D. 2003. Climate change, 1999. Long-term dynamics of the mean date climate modes, and climate impacts. Annual of autumn migration in passerines on the Review of Environmental Resources 28: 1–28. Courish Spit of the Baltic Sea. Avian Ecology Wang, H. & Fu, R. 2000. Winter monthly mean and Behaviour 2: 1–18. atmospheric anomalies over the North Pacific Stenseth, N.C., Ottersen, G., Hurrell, J.W., and North America associated with El Niño Mysterud, A., Lima, M., Chan, K.-S., Yoccoz, SSTs. Journal of Climate 13: 3435–3447. N.G., & Ådlandsvik, B. 2003. Studying climate Williams, C.N., Jr., Menne, M.J., Vose, R.S. & effects on ecology through the use of climate Easterling, D.R. 2006. United States Historical indices, the North Atlantic Oscillation, El Climatology Network Daily Temperature, Niño Southern Oscillation and beyond. Precipitation, and Snow Data. Oak Ridge National Proceeding of the Royal Society of London, B Laboratory/Carbon Dioxide Information Biological Science 270: 2087–2096. Analysis Center-118, NDP-070, Oak Ridge, Straub, J.N., Gates, R.J., Schultheir, R.D., Yerkes, Tennessee, USA. T., Coluccy, J.M. & Stafford, J.D. 2012. Zimmerman, C.E. 2009. Long-term trend analysis Wetland food resources for spring-migrating of climatic factors influencing autumn-winter ducks in the Upper Mississippi River and migration of Mallards in the Mississippi Great Lakes region. Journal of Wildlife Flyway. M.Sc. thesis, Mississippi State Management 76: 768–777. University, Mississippi State, Mississippi, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 Predictors of Mallard migration 469

Appendix 1. Rank of Weather Severity Index Anomalies (WSIA, ± thousands of km2) for eastern North America (Atlantic and Mississippi Flyway combined) in November–January 1950–2008.

Winter WSIA (× 1,000 km2) Rank* Winter WSIA (× 1,000 km2) Rank*

1950/51 –26.8 21 1979/80 292.0 55 1951/52 –93.5 13 1980/81 93.8 39 1952/53 197.0 49 1981/82 –51.9 18 1953/54 283.4 54 1982/83 339.2 56 1954/55 72.5 35 1983/84 –161.5 5 1955/56 –106.9 11 1984/85 33.1 29 1956/57 117.9 45 1985/86 –197.1 4 1957/58 249.0 52 1986/87 –0.4 27 1958/59 –72.0 15 1987/88 87.3 38 1959/60 96.8 41 1988/89 252.5 53 1960/61 –137.5 6 1989/90 –114.4 9 1961/62 –27.8 19 1990/91 86.1 37 1962/63 –72.5 14 1991/92 110.3 44 1963/64 –57.6 16 1992/93 –27.1 20 1964/65 52.9 32 1993/94 –0.8 26 1965/66 106.5 43 1994/95 173.8 47 1966/67 47.0 31 1995/96 –217.4 2 1967/68 –3.7 25 1996/97 –113.3 10 1968/69 –53.0 17 1997/98 –17.8 28 1969/70 –103.3 12 1998/99 227.6 51 1970/71 –17.5 22 1999/00 217.1 50 1971/72 94.8 40 2000/01 –418.5 1 1972/73 –14.8 23 2001/02 447.7 58 1973/74 –9.0 24 2002/03 54.8 33 1974/75 99.2 42 2003/04 71.7 34 1975/76 74.0 36 2004/05 129.4 46 1976/77 –198.9 3 2005/06 189.8 48 1977/78 –135.6 7 2006/07 437.1 57 1978/79 –123.5 8 2007/08 33.9 30

* Rank 1 = most severe winter; 58 = least severe.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 451–469 470

Waterfowl populations of conservation concern: learning from diverse challenges, models and conservation strategies

JANE AUSTIN1*, STUART SLATTERY2 & ROBERT G. CLARK3

1US Geological Survey, Northern Prairie Wildlife Research Center, Jamestown, North Dakota 58401, USA. 2Ducks Unlimited Canada, Institute for Waterfowl and Wetland Research, Oak Hammock, Manitoba MB R0C 2Z0, Canada. 3Environment Canada, Prairie and Northern Wildlife Research Center, Saskatoon, Saskatchewan S7N 0X4, Canada. *Correspondence author. E-mail [email protected]

Abstract There are 30 threatened or endangered species of waterfowl worldwide, and several sub-populations are also threatened. Some of these species occur in North America, and others there are also of conservation concern due to declining population trends and their importance to hunters. Here we review conservation initiatives being undertaken for several of these latter species, along with conservation measures in place in Europe, to seek common themes and approaches that could be useful in developing broad conservation guidelines. While focal species may vary in their life- histories, population threats and geopolitical context, most conservation efforts have used a systematic approach to understand factors limiting populations and to identify possible management or policy actions. This approach generally includes a priori identification of plausible hypotheses about population declines or status, incorporation of hypotheses into conceptual or quantitative planning models, and the use of some form of structured decision making and adaptive management to develop and implement conservation actions in the face of many uncertainties. A climate of collaboration among jurisdictions sharing these birds is important to the success of a conservation or management programme. The structured conservation approach exemplified herein provides an opportunity to involve stakeholders at all planning stages, allows for all views to be examined and incorporated into model structures, and yields a format for improved communication, cooperation and learning, which may ultimately be one of the greatest benefits of this strategy.

Key words: Anatidae, conservation strategy, decision framework, population model, status and trends.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 471

More than 20 species or populations of Nature (IUCN; IUCN 2013), although high waterfowl in North America, with diverse degrees of uncertainty were noted regarding life-histories, have experienced substantial limiting factors. Problems associated with declines over the past 25 years, or their habitat loss, fragmentation and degradation numbers remain well below conservation (e.g. reduced water quality) have continued goals (Table 1). Duck species of unabated since Green’s (1996) work (An et conservation concern range from the non- al. 2007; Dahl & Stedman 2013; Junk et al. migratory Mottled Duck Anas fulvigula, 2013). Hence, many challenges faced by which has small populations of limited North American waterfowl have relevance distribution, to migratory scaup (Greater globally, even if those populations are not Scaup Aythya marila and Lesser Scaup A. considered threatened by global standards. affinis, combined hereafter as scaup) and sea In this paper we examine approaches ducks (Tribe: Mergini) with continental to addressing contemporary challenges distributions. While some species share faced by several duck species of special traits, such as geographic overlap of scaup management concern in North America: and scoter Melanitta sp. breeding ranges in Mottled Duck, American Black Duck A. the boreal forests of North America, others rubripes (hereafter Black Duck), Northern seem to have little in common (e.g. Northern Pintail, scaup and sea ducks. We also Pintail Anas acuta and sea ducks). These examine conservation challenges facing declines and persistent low populations have the Common Eider Sometaria mollissima in concerned biologists, managers and hunters western Europe, where collaborative alike (Miller & Duncan 1999; Austin et al. research has developed but eider monitoring 2000). One aspect shared across species is and management depends in large part on considerable uncertainty about the factors agreement among many countries. Although that may be limiting populations, which all of these species are designated as being creates substantial challenges for developing of “least concern” by international nature effective conservation strategies. conservation agencies, they have become A wide range of environmental factors focal species for several reasons. First, in the pose threats to the persistence of many case of the North American dabbling duck, sea duck and goose populations and diving ducks, all are numerically globally. Of 228 waterfowl taxa (sub-species important harvested species valued by level) investigated by Green (1996), 48 hunters (Raftovich & Wilkins 2013). For vulnerable or endangered taxa (37 ducks instance, Northern Pintail, Lesser Scaup and and sea ducks; 11 geese) were threatened Black Duck are highly prized by hunters mainly by habitat loss, hunting and in the Pacific, Mississippi and Atlantic predation by invasive species. These same Flyways, respectively, for a variety of threats were also the most common among cultural reasons. Second, all of these species the 29 threatened or endangered duck and have experienced substantial population goose species recently assessed by the declines at some point in the past 30 years, International Union for Conservation of with no evidence of strong recoveries

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 472 Waterfowl populations of conservation concern Polysticta stelleri Polysticta Sometaria fischeri Anas wyvilliana b Anas laysanensis Laysan Duck Laysan Steller’s Eider (Alaska) Steller’s Spectacled Eider Duck Hawaiian At-risk species Bucephala islandica a,b Branta bernicla Histrionicus histrionicus Melanitta fusca Clangula hyemalis Branta sandvicensis Chen canagica Branta hutchinsonii Anas fulvigula Melanitta americana Anser albifrons elgasi Sometaria spectabilis Hawaiian Goose Hawaiian Emperor Goose Black Scoter Black White-winged Scoter High Arctic Brant Western Barrow’s Goldeneye (eastern) Barrow’s King Eider Goose Tule Goose Cackling Mottled Duck Long-tailed Duck Species of special concern a Anas rubripes Branta c. occidentalis Anas americana Anas acuta Aythya marila Aythya Aythya affinis Aythya Summary of of conservation below objectives species that were North American waterfowl the North American Waterfowl Branta canadensis Canadian Wildlife Service Waterfowl Committee 2012, USFWS 2008, IUCN 2013. Canadian Wildlife Service Waterfowl NAWMP 2012, Appendix A (Tables 1, 2, 3). 1, 2, A (Tables 2012, Appendix NAWMP American Wigeon Greater Scaup Atlantic Flyway Canada Goose (resident)Atlantic Flyway (eastern) Harlequin Duck Management Plan (NAWMP) in 2011, and species listed by federal agencies in 2011, and species listed by or International Union for the Conservation ofManagement Plan (NAWMP) Nature (IUCN) as being “at-risk” in 2013. Duck American Black Lesser Scaup Dusky Canada Goose Table 1. Table objective NAWMP Below Northern Pintail a b

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 473

(Zimpfer et al. 2013; Ekroos et al. 2012). highlight the value of planning models for Population sizes of Northern Pintail, scaup making decisions when much uncertainty is and Black Duck remain below goals involved. We believe this approach is established by the North American applicable whether the population of Waterfowl Management Plan (NAWMP; concern is the Lesser Scaup, which is still NAWMP 2012). Third, population common in North America, or a globally management objectives achieved through threatened species. harvest regulations should be guided by science, and the manifold reasons for Conceptual framework persistently low populations have not been A conceptual framework is here defined as adequately resolved. This situation can an organisation of ideas into a set of logical create debates between advocates of steps to solve a problem and develop conservative harvest regulations or season strategies to achieve desired goals. For closures and proponents of liberal harvest North American waterfowl, those goals quotas who may question the lack of are population levels sufficient to meet evidence for adverse harvest effects on conservation and societal demands and populations. And, fourth, these species are implicit in subsequent discussions. provide unique opportunities to learn about Conservation efforts generally follow a the application of formal decision analysis framework that begins with a broad (e.g. Clemen 1996; Conroy & Peterson approach to the formulation of hypotheses 2012; Gregory et al. 2012) to address about why populations either decline or concerns surrounding the management remain below conservation goals. The and conservation of harvested duck strength of this approach lies in proposing populations, while these species remain plausible hypotheses to explain low relatively common. These taxa represent a populations, a process that typically involves range of conservation goals, geographic a thorough evaluation of existing evidence scope, confidence in survey results, and debate about defensible and sometimes availability of data to inform hypotheses speculative explanations for population and models, modelling approaches and patterns. One way of visualising this is with organisational history. Additionally, a decision tree (adapted from Platt’s (1964) planning efforts within each taxon generally logical tree), as was used to summarise and follow a robust conservation framework illustrate explanations for low populations (sensu strategic habitat conservation; of Northern Pintail (J. Eadie, University of Johnson et al. 2009) that facilitates California-Davis, cited in Miller et al. 2003) systematic and collaborative planning, and scaup (Fig. 1). Mechanisms that could typically under the auspices of NAWMP produce observed population changes infrastructure (NAWMP 2012). The goals of allow for explicit predictions about the this paper are to review the conceptual expected demographic responses to specific framework, demonstrate how the management or policy actions. This framework was applied in case studies and approach provides a stronger conceptual

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 474 Waterfowl populations of conservation concern

Population No decline

Yes

Recruitment Survival

Boreal Priarie Pothole Hunting Cross-seasonal ecosystem Region

Habitat loss and Habitat loss and Nutrition fragmentation fragmentation General

Climate change Disease effects on Disease wetlands

Figure 1. Graphical representation of a decision tree (modified from a logical tree; Platt 1964), designed to represent main working hypotheses proposed for scaup population status or declines (e.g. recruitment, survival) and putative mechanisms responsible for such changes (see Table 2). Management or policy alternatives would be implemented to improve demographic rates (survival or reproductive success), and then evaluated for effectiveness with targeted monitoring, research or adaptive management programmes. Factors affecting recruitment in the boreal ecosystem differ from those affecting recruitment in the Prairie Pothole Region; other factors affecting recruitment cross seasons for both breeding regions. Hypotheses were generated during waterfowl community workshops (Austin et al. 2000), as well as via research, monitoring and modelling studies. framework for integrating critical steps by of models have greatly expanded over the pinpointing: 1) likely bottlenecks to positive last few decades, enabling biologists to population growth rates; 2) suites of integrate and simultaneously to model management or policy actions with the potential drivers of demographic variation potential to alleviate these bottlenecks; encountered on breeding, staging and 3) predicted demographic and population wintering areas, in order to predict responses to these actions; and 4) population change (Mattsson et al. 2012; monitoring required to evaluate the Osnas et al. 2014). Model objectives and effectiveness of management interventions. structure reflect existing hypotheses or Population models, whether qualitative or primary issues of concern, the availability of quantitative, are at the core of implementing data and potential management actions. In this conceptual framework. The use of addition, integrated models are increasingly population models to inform conservation able to leverage multiple sources of limited actions has a long history (Shaffer 1981; data (Schaub & Abadi 2011). Models Caswell 2000). The role and sophistication therefore are fundamental in our case

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 475 studies because they codify the decision tree coastal Texas during 1994–2005 (Johnson and management actions, and thus allow 2009) and relatively stable trends elsewhere measurable predictions about expected (Bielefeld et al. 2010). It is also one of the demographic outcomes under different least studied Anatini in North America. The management and conservation scenarios. species is considered to be of conservation Additionally, this structure can be applied to concern because of its restricted distribution, other management goals (e.g. hunter relatively small population sizes, loss and recruitment and retention; NAWMP 2012), degradation of key coastal habitats in the and used to identify key uncertainties. WGC (Wilson 2007) and introgressive There is considerable optimism, indeed hybridization with Mallard Anas platyrhynchos expectation, that these model-based in the Florida sub-population (Table 2). approaches will be pivotal in setting new, In the WGC, the primary conservation integrated objectives for NAWMP in the concerns are the degradation and loss of next several years (NAWMP 2012; Osnas et critical habitats, notably coastal and inland al. 2014). In the following case studies, we palustrine marshes, rice fields and native examine how application and outcomes of prairie and pastureland important for the framework evolved under the unique life nesting (Wilson 2007). Highly variable history, data availability and socio-political breeding propensity, which affects settings for each duck species. population growth rates, may be tied to wetland conditions (Rigby & Haukos 2012). Case studies In Florida, similar concerns about the loss and degradation of wetland habitats have Mottled Duck raised questions about the duck’s nutritional Background – The Mottled Duck is a non- status, which can affect reproductive success migratory species with two genetically (Florida Fish and Wildlife Conservation distinct sub-populations, one in Florida, the Commission 2011). Because of small home other occurring along the Western Gulf ranges, the species can be sensitive to local Coast (WGC) portions of Alabama, habitat changes and harvest pressure. Mississippi, Louisiana, Texas and northeast Hybridization with feral Mallard is however Mexico. The combined population estimate the main threat to Mottled Duck in Florida. is c. 172,000 birds (M. Brasher, U.S. Fish and Both regions share concerns about harvest Wildlife Service, and R. Bielefeld, Florida rates and potential impacts of climate Fish and Wildlife Commission, pers. change on habitat conditions, with likely comm.). Acquiring reliable status and trends increased frequency of severe weather and information has been hampered by the lack further habitat loss (Florida Fish and of long-term, range-wide surveys corrected Wildlife Conservation Commission 2011). for visibility bias. Spring surveys in Florida Approach – Conservation plans were suggest that the sub-population there has developed based on expert opinion and been stable (at c. 53,300) since 1984, but local limited existing data, and implemented for surveys and indices suggest declines in both Florida (Florida Fish and Wildlife

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 476 Waterfowl populations of conservation concern b R, S R, S, P R, S, b b SS SS S S S, P North Baltic R, S, P R, S, b b b b S, P R, P R, P R, S, P R, S, b b SS S R, S R, S R, S, P R, S, P R, S, b b S SS S R, S, P R, S, R, S, P R, S, b b b b SS SS SS S S XX R, P Duck Duck Pintail Scaup sea ducks Eider R, S, P R, S, P R, S, Mottled Black Northern American Common = probability of Occurrence of breeding. is noted with an X. hybridization b Hypothesised factors causing declines or low populations ofHypothesised factors causing declines or low species of selected waterfowl conservation concern. Vital rates: shipping, wind-power development shipping, wind-power migration and wintering areas grounds Bioaccumulation ofBioaccumulation contaminants from foods on Presumed cause(s) Disturbance on migration and wintering areas from Commercial exploitation of food resources on wintering R = recruitment, S = survival, P Table 2. Table Habitat loss and degradation on breeding groundsHabitat loss and degradation on wintering grounds P R, S, Habitat changes due to climate change Habitat changesagriculture related to Mallard with Interspecific hybridization Lead poisoning from spent shotgun pellets Altered predator-prey relations on breeding groundsAltered predator-prey Human disturbance during breeding Diseases and parasites Over-harvest Contaminants (lead poisoning from spent shotgun pellets) Habitat changes on spring migration areas

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 477

Conservation Commission 2011) and WGC In the WGC, a sex-specific matrix model sub-populations (Wilson 2007). Experts identified female annual survival as an identified factors most likely to limit sub- important factor in Mottled Duck population growth, which stimulated population dynamics and potential target for research to elucidate how those factors were management actions (Johnson 2009). A affecting vital rates, such as breeding pattern of high breeding season survival and distribution, nesting effort and survival. low breeding incidence suggested a trade- Florida’s plan focused on addressing off between nesting effort and female uncertainties related to hybridization with survival (Rigby & Haukos 2012). Combined, Mallard by developing tools to identify these models indicate that improved habitat species and hybrids more accurately, and on quality will be critical for conserving this assessing the impact of wetland quality on species in the WGC region. Two main productivity and the energy demands of this conservation actions identified by the model species. Results from studies on habitat use are the enhancement and restoration of patterns for urban and rural Mottled Ducks coastal marshes (primarily for creating in Florida improved predictions of Mottled suitable (i.e. low salinity) brood habitat), and Duck distribution and habitat use during the restoration of coastal prairie and multiple periods of the annual cycle associated wetlands to enhance nesting and under contrasting water conditions propensity, nest success and brood survival. (Bielefeld & Cox 2006; Varner 2013, 2014). Partners have developed a spatially-explicit The findings should improve the targeting decision support tool to aid delivery of of habitat conservation actions, and also Mottled Duck habitat conservation in improve the efficiency and effectiveness of locations where demographic responses the annual population surveys. Recent are likely to be more favourable. Finally, information on temporal and spatial implementation of an annual range-wide, patterns of survival in Florida (Bielefeld & visibility-corrected survey of Mottled Cox 2006) have improved predictions of Ducks in the WGC will likely reduce how future habitat loss and alteration uncertainties about population sizes and (including continued urbanisation and trends. wetland creation associated with urban development and the Comprehensive American Black Duck Everglades Restoration Plan; Anonymous Background – The American Black Duck is 1999), will affect the Mottled Duck sub- distributed in eastern North America from population. New techniques based on Ontario to the Maritime Provinces and plumage characteristics (R. Bielefeld, pers. south through states of the Mississippi comm.) will be valuable for assessing the Flyway and the Atlantic Flyway. Historically, extent and distributional aspects of Mottled it was the most abundant dabbling duck in Duck x Mallard hybridization, and thus eastern North America and also the most for identifying the most appropriate heavily harvested (Rusch et al. 1989). conservation actions. Estimates from the Mid-Winter Waterfowl

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 478 Waterfowl populations of conservation concern

Inventory, conducted annually across most 1) loss in the quantity or quality of breeding key wintering areas in the U.S., indicated the habitats; 2) loss in the quantity or quality of population experienced a rapid and wintering habitats; 3) harvest; and 4) sustained decline of > 50% between 1955 competitive interactions or hybridization and the 1990s (Conroy et al. 2002). with Mallards. They concluded that no Christmas Bird Count data, a citizen-science single factor could explain the Black Duck survey programme conducted annually in decline. A common theme across these selected areas, suggest that Black Duck issues and trends is uncertainty about the numbers declined in the southern and role of density dependence on reproduction central portion of wintering range during and survival, and potential cross-seasonal 1966–2003 but that populations in the influences of putative density-dependent northeast were stable (Link et al. 2006). effects. Also unclear is the degree to which While these wintering surveys provide the competition and hybridization with Mallards longest time period to assess trends, they may have interacted with other factors such both suffer from substantial shortcomings, as harvest and habitat changes (Nudds et al. such as temporal and spatial variation in 1996; Petrie et al. 2000). Although numerous survey effort and methodology. Breeding investigations have addressed these issues, ground surveys conducted with more there is a lack of consensus about the role rigorous methods since 1990 indicate stable these factors play in limiting the population. or slightly increasing trends (Zimpfer et al. Approach – The Black Duck Joint Venture 2013). Contrasting population trends among (BDJV) was established in 1989 as the first these three surveys could be related to “species joint venture” (JV) to implement counting methods or temporal shifts in and coordinate a cooperative population winter distributions. However, some have monitoring, research and communications raised questions about possible regional programme to provide information required differences in population demographics to manage Black Duck populations and (Conroy et al. 2002; Black Duck Joint restore numbers to the NAWMP goal of Venture 2008). 640,000 breeding birds (NAWMP 2012). Researchers and managers have proposed Initial priorities included development and several hypotheses to explain the historic implementation of improved surveys to decline of Black Duck populations (Table monitor breeding populations and harvest, 2), including over-harvest, competition and directed projects to provide estimates of hybridization with Mallard, decrease in vital rates and habitat requirements, research quality and quantity of wintering and to incorporate spatial information into the breeding habitat, parasites and disease (e.g. breeding ground survey to identify habitat duck viral enteritis) and environmental features affecting Black Duck abundance, contaminants (e.g. lead shot, mercury, and development of a life-cycle model DDT). Conroy et al. (2002) found support (Conroy et al. 2002) and a model to estimate for four major, continental-scope factors autumn age-ratios. that may influence Black Duck populations: The annual life-cycle model provides a

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 479 mechanistic description of population Ducks corresponding to the 1990–2012 growth, assesses hypotheses concerning period; 3) maintaining habitat to support factors potentially limiting Black Duck desired abundance, distribution and population growth, and links hypotheses to harvest opportunity; and 4) increasing parameters that could be estimated from understanding of the density-dependence available data. Sensitivity analyses were used mechanism and of factors limiting the to explore effects of statistical uncertainty in species to make increasingly more informed parameter values on population growth decisions. Underlying the framework is the rates. Results indicated that reproductive Conroy et al. (2002) model of Black Duck rates were positively influenced by breeding population dynamics and habitat, with habitat quantity and negatively influenced by competing hypotheses on the role of Black Duck and Mallard densities, and that density dependence on reproduction on the the proportion of Black Ducks harvested breeding grounds, survival on wintering also declined with increasing densities of grounds (post-hunting season), carry-over both species (Conroy et al. 2002). Conroy et effects of wintering habitat conditions and al.’s (2002) modelling work formalised changes in movement patterns from uncertainties about factors that influence breeding to wintering areas. Drivers of the Black Duck population and provided the vital rates include weather and carrying foundation for an adaptive management capacity as affected by habitat loss and framework. habitat management. Strength of density The BDJV, in partnership with the dependence in winter is assumed to be Eastern Habitat and Atlantic Coast JVs, related to energy intake and expenditure (e.g. is developing a decision framework that per capita food supply and weather integrates habitat and population conditions). Harvest is included in the management that will enable the JV to decision framework but is not a focus of produce an objective, science-based management actions. Research is underway estimate of carrying capacity and make to address key uncertainties related to recommendations for revising the NAWMP energetic capacities on the wintering population goal (Black Duck Joint Venture grounds, return on investment for winter 2008; Devers et al. 2011). The framework habitat restoration and to improve focuses on area of habitat restored or parameter estimates such as post-season protected at the Bird Conservation Region survival rates in relation to variation in (BCR; North American Bird Conservation weather and food (Osnas et al. 2014). Initiative 2000) level, framed as a resource allocation issue. Decision framework Northern Pintail objectives include: 1) achieving the Background – The Northern Pintail is one of NAWMP population goal under a harvest the most abundant dabbling ducks in North strategy of 98% maximum sustainable America. The main breeding habitats are in yield; 2) maintaining current distribution Alaska and the Prairie Pothole Region of of breeding and wintering Black southern Canada and the northern U.S.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 480 Waterfowl populations of conservation concern

Great Plains, and winter habitats are along expressed uncertainty about the ability of the coasts and throughout the southern the traditional waterfowl survey (survey U.S.A. The species is closely associated with strata 1–50; Zimpfer et al. 2013) to count temporary and seasonal wetlands, and breeding Northern Pintail reliably during historically pintail numbers have tracked dry years on the prairies, when the pintails wetland conditions on the prairies (Miller et may overfly the region to settle in al. 2003). Population levels were high during unsurveyed areas further north. the 1950s and 1970s (5.5–9.9 million), Empirical research since the 2001 Pintail periodically fell below 4 million birds during Workshop (Miller et al. 2003), undertaken short-term droughts on the prairies during both at breeding sites and on the wintering the 1960s–1980s, then fell to record lows grounds, has helped to fill many information during an extensive prairie drought in gaps and reduced some uncertainties, such 1988–1991 (1.8–2.3 million). Despite greatly as those relating agricultural practices to improved wetland conditions in the prairies nest survival (Podruzny et al. 2002; since the mid-1990s, pintail numbers over Kowalchuck 2012; J. Devries, Ducks the last decade have averaged 3.2 million, Unlimited Canada, unpubl. data) and 43% below the NAWMP goal of 5.6 million migration chronology relative to timing of (Zimpfer et al. 2013). Most of the recent surveys within traditional survey areas decline occurred in Prairie Canada, and the (Miller et al. 2005). For the migration and once-close relationship between numbers of winter periods, research findings have breeding pintail and number of prairie generally downplayed the importance of wetlands has weakened substantially since low survival rates (Miller et al. 2005; Haukos the 1990s (Podruzny et al. 2002). et al. 2006; Fleskes et al. 2007; Rice et al. Three main biological hypotheses have 2010). However, studies of Northern been suggested to account for the pintail Pintail wintering on the Texas Gulf decline. The most plausible is that Coast identified new concerns about low conversion of prairie to cropland and overwinter survival associated with the loss changing cropping practices on the breeding of wetlands and rice agriculture (Moon & grounds, especially in prairie Canada, has Haukos 2006; Anderson 2008). These reduced nest success (Table 2). But it was unexpected results and the implementation also speculated that fewer females nested of a national harvest strategy for Northern (persistently) due to cross-seasonal effects Pintail in 1997 (USFWS 2010) were among from reduced habitat quality during winter the factors elevating the importance of and spring migration. Finally, over-harvest harvest rates in population dynamics and higher mortality of adult females during models. the breeding season due to diseases Approach – The Pintail Action Group (primarily Avian Botulism Clostridium (PAG) was created in 2003, operating as a botulinum and Avian Cholera Pasturella working group under NAWMP, with a multocida) and predation have also been mission to advocate and support the suggested. Meanwhile, biologists have also coordination and evaluation of Northern

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 481

Pintail management and research among and internal (i.e. density-dependent) JVs, North American Flyways, government mechanisms, and that those mechanisms agencies, and organisations (Duncan et al. may interact. These assumptions in turn lead 2003). JVs have since pursued large-scale to the dual assumptions that habitat habitat programmes on key breeding areas management by JVs (or JVs encompassing and maintenance of key migration and main pintail regions of North America) has wintering areas. For example, the Prairie a direct influence, and that harvest Habitat JV, which encompasses the Prairie management has an indirect influence on Pothole Region in Canada, has developed population-specific vital rates through programmes to encourage conversion of density-dependent mechanisms. This spring-seeded cropland to more pintail- linkage allows simultaneous prediction friendly uses (Devries et al. 2008), such as of the effects of harvest and habitat Winter Wheat Triticum aestivum and forage management on continental pintail crops. Other conservation efforts include population dynamics (Table 3). direct land protection and enhancement, agricultural partnerships, and policy Greater and Lesser Scaup initiatives. The PAG coordinated work to Background – Greater and Lesser Scaup construct an empirically based meta- cannot be distinguished in aerial surveys so population model that integrates the effects the species are usually combined and of habitat and harvest on vital rates, and identified as ‘scaup’ in population estimates. provides a platform to link habitat change Their combined range is the most and regional management actions to widespread of North American diving key demographic rates and population ducks. Greater Scaup breed primarily in responses (Mattson et al. 2012). The model tundra regions from western Alaska to approach and structure is described below eastern Canada, with some also breeding in and by Osnas et al. (2014). the boreal forest, whereas Lesser Scaup The predictive life-cycle model (Mattson breed largely in boreal and prairie regions. In et al. 2012) enables evaluation of how winter most scaup are found along the alternative habitat and harvest management coasts of the Pacific and Atlantic Oceans, strategies simultaneously influence the Gulf of Mexico and the Great Lakes, continental-scale pintail population although Lesser Scaup also winter on inland dynamics. This was the first model to waters. The combined breeding populations integrate habitat and harvest explicitly into a of scaup declined from 5.7–7.6 million birds modelling framework, the goal of the in the 1970s to a record low of 3.25 million NAWMP Joint Task Group (Anderson et al. birds in 2006 before showing signs of partial 2007). Mattson et al. (2012) discuss the recovery; 4.2 million scaup were reported in general assumptions, common to most 2013 (Zimpfer et al. 2013). The current other species models, that population population estimate remains 33% below dynamics are regulated by external the NAWMP goal of 6.3 million. The (i.e. habitat and harvest management) prolonged decline and uncertainties about

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 482 Waterfowl populations of conservation concern different habitat conditions. conservation goals. continental-scale population dynamics. continental-scale scaup population dynamics and hunter recruitmentcontinental-scale and retention. Conservation applications female survival and nesting effortClarify relationship between under Guide habitat actions (protection and restoration of critical wetlands). Framework for cost-effective allocation of for cost-effective Framework conservation resources. estimated costs and benefits ofEvaluate management actions. effects of evaluate Simultaneously harvest and habitat management on or actions to manageIdentify best areas species. Mottled Duck ageMottled Duck and sex, and to identify Mottled Duck- population and harvest estimates; research. Improve based on Mallard hybrids are most prevalent. hybrids plumage characteristics life-cycle modelannual dynamics. model model hunter model Model : Model to differentiate Florida Accurate identification of in surveys species and hybrids and where GulfWestern Coast : Sex-specific survival) in population vital rates (female annual Pinpoint critical Integrated life-cycle and habitat best areas or actions to achieve Guide habitat actions; identify Integrated life-cycle and habitat effects of evaluate Simultaneously harvest and habitat management on Integrated habitat and life-cycle, and management Identify research priorities. a a Examples of for species of models developed conceptual and predictive conservation concern in North America, and their uses a in guiding research and conservation programmes. References that describe details of and conservation References in guiding research programmes. case- studies. in respective these models are given Table 3. Table Species Mottled Duck Duck Black Northern Pintail Scaup

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 483 dynamics. population dynamics. Conservation applications Sustainability of harvestfor harvest and recommendations levels. Impact of on population cholera periodic mortality from avian events life-cycle model model for wintering in the Bering Sea of and design availability marine protected areas. Model ( continued ). Eider Eider Atlantic Common Demographic model Common Pacific stage-based, matrix Stochastic, and management Identify research priorities. Spectacled Eider Energy balance simulation Assess implications of food resource changing climate, changing Long-tailed Duck Matrix population model Determine impacts of lead poisoning and subsistence harvest on See Osnas et al . (2014) for details. Table 3 Table Species Sea ducks a

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 484 Waterfowl populations of conservation concern factors contributing to the low numbers The Climate Change-Habitat Hypothesis resulted in both species being listed as “focal suggests that warming climate in northern species of concern” by the U.S. Fish and breeding regions has reduced the abundance Wildlife Service (USFWS 2011). The largest or quality of wetland habitats for scaup at decline occurred in the boreal forest, the large scales, potentially reducing food core breeding region for Lesser Scaup, but resources, availability of nesting or brood- numbers also declined in prairie Canada. rearing habitat, and breeding propensity; Numbers of scaup in the tundra survey altered habitat conditions may also have strata, presumed to be Greater Scaup, have altered scaup’s exposure to predators, been stable or slightly increasing since the reducing nest success or adult female 1970s. Hence, the main focus of concern is survival. This hypothesis is founded on data on Lesser Scaup. indicating that the greatest change in annual Specific hypotheses explaining the mean temperatures coincides with the population decline were first put forward by location of core Lesser Scaup breeding Austin et al. (2000) and Afton & Anderson habitats in the western boreal forest, and (2001) and with further debate evolved evidence for substantial long-term declines into six key hypotheses (Table 2, Fig. 1). in wetland areas in Alaska’s boreal region The Disease Hypothesis (i.e. contaminants) (Riordan et al. 2006). The Climate Change- proposes that environmental contaminants Mismatch Hypothesis asserts that earlier spring have had a negative effect on scaup survival phenology and warmer water temperatures and productivity; this was based on known in northern breeding wetlands has caused environmental contamination of wintering invertebrates (the scaup’s main food and staging areas (primarily selenium and resource) to advance their reproductive PCBs) and on high levels of contaminants cycles, possibly reducing their availability to recorded in some preferred scaup foods scaup later in the season (see Drever et al. such as the exotic Zebra Mussels Dreissena 2012). The Predation Hypothesis postulates polymorpha (Custer and Custer 2000; Petrie et that fluctuations in predators and alternate al. 2007). The Spring Condition Hypothesis prey indirectly affect waterfowl productivity posits that body condition during migration (Brook et al. 2005). A Harvest Impact and pre-breeding has declined compared to Hypothesis was put forward to acknowledge historic levels due to reduced food possible links between harvest management abundance or quality on spring migration and scaup population size, but was not areas, and subsequently reduced body considered a strong contributor to the scaup condition has negatively affected scaup decline (Afton & Anderson 2001). survival and productivity (e.g. through lower Approach – A Scaup Action Team (SAT) breeding propensity, smaller clutch sizes, was created in 2008, also as a special and later nest initiation dates). Original working group under the auspices of the concerns about widespread habitat changes NAWMP, to help strengthen the biological on the breeding grounds have been foundations of conservation programmes. refocused to two inter-related hypotheses. The interests of the SAT and the listing of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 485 scaup as a “Migratory Birds of Management tundra (Greater Scaup). In workshops (e.g. Concern” (USFWS 2011) led to Austin et al. 2010), experts formulated development of a conservation action plan. competing hypotheses about causes of The SAT is using a structured decision- population changes (Fig. 1) and identified making process, starting by framing the measurable features (attributes), such as problem (resource allocation among wetland density or percent of the landscape alternative management actions) and in cropland, that likely influenced vital rates identifying fundamental objectives of scaup (Table 2) and that could be influenced conservation planning not only in terms of through management (or policy) actions. objectives for scaup populations and their Functional relationships were then habitats but also for scaup hunter developed for each vital rate and measurable populations. The foundation of the decision attribute and incorporated into the scaup framework is based on predictive models for model (Austin et al. 2010). Density both scaup and hunter populations, linked dependence is incorporated in two parts of via harvest rate, the former building on the life-cycle model. During breeding, the work of Flint et al. (2006) and Koons et al. mechanism of density dependence is via (2006). probability of breeding (habitat and/or The prototype predictive model for scaup food limitation in boreal and tundra is designed as a top-down decision regions). For birds in late winter (i.e. after the framework to address three objectives: 1) hunting season), density dependence may achieve landscape conditions (continental operate through survival, with food carrying capacity, i.e. habitat) capable of limitation as the primary mechanism. The supporting target populations; 2) ensure model relates the number of ducks in the desired levels of sustainable harvest; and 3) post-hunting population to survival during sustain the diving duck hunting tradition (i.e. the following season (here, late winter–early diving duck hunter population). The spring) with either compensatory or additive framework provides a means to identify the harvest mortality. The process allowed many best areas or actions to be targeted for issues of uncertainty to be identified (e.g. managing scaup and for learning (reducing interactions among alternative actions, lag uncertainty). The framework explicitly links effects of environmental change and two life-cycle models, one for scaup reliability of vital rate estimates). populations and a second for diving The scaup life-cycle model is explicitly duck hunters, and identifies alternative linked to a simple model of diving duck management actions and their (inter-) hunters, which identifies putative factors relationships to scaup and/or hunter vital and their functional relationships affecting rates. hunter recruitment and retention. The two The scaup life-cycle model incorporates models are linked via an empirically based separate population estimates and respective harvest rate parameter, and are both vital rates for three breeding regions: prairie projected forward through time to estimate and boreal regions (Lesser Scaup) and numbers of scaup, number of scaup

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 486 Waterfowl populations of conservation concern harvested, and numbers of diving duck Maine (Table 1). Spectacled Eider Sometaria hunters under different habitat and harvest fisheri and Alaskan-breeding population of regulations scenarios. The model also can Steller’s Eiders Polysticta stelleri are listed as be used to explore potential impacts of threatened in the U.S. Where population large-scale ecosystem change on scaup data do exist, trends of several sea duck reproduction and carrying capacity. species were correlated with large-scale Ultimately, the model will provide the oceanic regime shifts, although the direction necessary framework to perform decision of relationships varied within and among analyses and evaluate estimated costs and species, and these populations appear to benefits of specific management actions as have been stable or increasing for the last 20 well as to make transparent, informed trade- years (Flint 2013). offs among multiple objectives. For many species, ecological knowledge in the early 1990s was insufficient to identify North American sea ducks priority threats or factors contributing to Background – Among the least studied of apparent declines. Threats related to loss North American waterfowl are 15 species of and degradation of breeding and wintering sea ducks (Mergini). Their distributions fall habitats, and the implications to long-term largely in remote arctic or marine areas, health and security of populations, are outside of traditional survey areas, so shared by multiple sea duck populations reliable indices of their populations and (Table 2). Habitat-related threats include oil, productivity have been lacking. Moreover, gas and wind power development, shellfish some groups of sea ducks have not been aquaculture on staging and wintering differentiated to species during surveys areas, and effects of changing climate on (three species of scoters Melanitta sp.; critical habitat. Harvest threatens several Common Goldeneye Bucephala clangula populations (SDJV Management Board and Barrow’s Goldeneye B. islandica; and 2008). Other issues of concern include Red-breasted Merganser Mergus serrator bioaccumulation of contaminants, effects of and Common Merganser M. merganser). disease and parasites (e.g. Avian Cholera Consequently, abundance, relative densities die-offs affecting Common Eiders), and population trends cannot be accurately consumption of spent lead shot and estimated for most sea duck populations. disturbances from shipping lanes and Eight of 22 species or populations are offshore wind power development. thought to be below historic levels and 5 are Approach – Evolving awareness and thought to be at or above historic levels; the concerns surrounding habitat, contaminants status of remaining species remains and harvest for all sea duck species led to the unknown (Bowman et al. 2015). Since establishment of the Sea Duck JV (SDJV) in 1986, Barrow’s Goldeneye and the eastern 1998 as a multi-species JV to advance sea population of Harlequin Ducks Histrionicus duck conservation. The focus of the SDJV histrionicus have been listed as species of to date has primarily been to fill key concern in Canada and as threatened in information gaps on population trends, vital

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 487 rates, habitat use, and delineation of regulations towards sustainable harvest functional populations (SDJV Management levels, identify vital rates most responsive to Board 2008). Conservation efforts involve a management action or requiring further coordinated international approach (mainly research, quantify population-level risk to U.S. and Canada, but also Russia and the Exxon Valdez oil spill, and assess size Greenland). Partners used existing data and requirements for marine protected areas. expert opinion to rank species and research Most models generally demonstrated high priorities for species known or believed and stable annual female survival, the vital to be facing significant threats. The SDJV rate to which population changes were most has supported programmes to develop sensitive. However, given these patterns in and improve population monitoring and adult survival, fecundity parameters (nest delineation, such as winter sea duck surveys success and especially duckling survival) off the Atlantic Coast and on the Great were more often indicated as potential Lakes, counting Black Scoters Melanitta targets for management actions. americana molting in James Bay, and delineating functional populations using Common Eider in the Baltic satellite telemetry and genetic markers. Background – The Common Eiders of the Because of the diversity of species and Baltic/Wadden Sea flyway breed in Sweden, issues, biologists have pursued targeted Finland, Denmark, Norway, Estonia, the research projects rather than broad Netherlands and Germany and winter conceptual models more generally mainly in Denmark, Germany, the applicable to seaducks. Netherlands, Sweden, Norway and Poland. The targeted sea duck projects have led to This population has been well-studied and development of at least eight different has been the subject of long-term population models that have or can aid international monitoring programmes decision-makers (see Table 3 for examples). because of its status in the European Model types included stage-based matrix harvest. Recent evidence from mid-winter projections (for Common Eider: Gilliland et surveys suggests the population may al. 2009; Iles 2012; Wilson et al. 2012; for have experienced a substantial decline. King Eider Sometaria spectabilis: Bentzen & Coordinated aerial surveys in the Dutch, Powell 2012; for Long-tailed Duck Clangula German and Danish Wadden Sea show hyemalis: Schamber et al. 2009), reverse-time numbers halved from c. 320,000 in 1993 to c. capture-recapture (White-winged Scoter 160,000 in 2007; coordination of counts in Melanitta fusca: Alisauskas et al. 2004), other winter regions is weaker, leading to individual-based models (Harlequin Duck: substantial uncertainties in overall trends Harwell et al. 2012), and spatially-explicit (Ekroos et al. 2012). Ability to assess simulations of energy balance (Spectacled population trends across the entire winter Eider: Lovvorn et al. 2009). While these range is further compromised by changes in models are too numerous to review here, count methodology from “total counts” to they have been used to assess and guide aerial survey methods that rely on distance

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 488 Waterfowl populations of conservation concern sampling and spatial modelling. The “best” some nesting colonies have been attributed estimates of winter totals for the years 1991, to varying causes, usually associated with 2000, and 2009 were of 1,181,000, 760,000 changes in predation, including greater and 976,000 Common Eiders, respectively. predation of incubating females and eggs by Ekroos et al. (2012) questioned whether the White-tailed Sea Eagles Haliaeetus albicilla in apparent increase between 2000 and 2009 Finland and invasive American Mink was real or due to: 1) changes in survey Neovison vison that have reduced reproductive methods; 2) the generation of mid-winter success and female survival in Sweden counts from some states using data collected (Desholm et al. 2002). Declines in other over several winters during 2006–2010; or 3) colonies have been linked to lower duckling birds short-stopping further east in survival (related to density dependent response to milder winters, where they may regulation and viral disease), competition be less well counted. These negative with other waterbirds, and poor pre-nesting population trends contrast with breeding- body condition in spring (Desholm et al. ground surveys that show no consistent 2002). Other issues of concern include trends in breeding abundance (Desholm et disease and parasite infestations affecting al. 2002; BirdLife International 2004). survival and reproduction, pollutants Desholm et al. (2002) suggested the decline generally, lead poisoning in Finland, avian may represent a decline in numbers of non- cholera in Denmark, bycatch in gill nets, and breeding “floaters”, which would not be offshore collisions with high-speed boats represented in breeding ground counts but and offshore structures such as wind would be included in winter counts. There is turbines and bridges. The impact of harvest also evidence of a decline in the adult sex on the population is also unclear (Gilliland et ratio among Common Eiders harvested in al. 2009). Denmark (Ekroos et al. 2012). Hence, there Approach – Because of its global are substantial underlying uncertainties importance to many waterbirds in the about winter survey data and population flyway, the Danish Wadden Sea has been demographics within different breeding recognised as a Ramsar site, a Natura-2000 regions. site, an Important Bird Area, a Man and The most immediate threat to the Baltic/ Biosphere Reserve, and a World Heritage Wadden Sea population is commercial Site. It is encompassed under the Western exploitation of shellfish, which has likely Palearctic Anatidae Agreement (WPAA), reduced food availability to eiders and is Trilateral Governmental Conference (The linked to mass starvation of Common Netherlands, Germany and Denmark), and Eiders in some years (Table 2) and regional the 1995 Agreement on the Conservation of reductions in other years (Camphuysen et al. African-Eurasian Migratory Waterbirds 2002). Furthermore, declines may be related (AEWA; Boere & Piersma 2012). The latter to unknown factors causing delays among provides the best legal, intergovernmental females in first year of breeding combined instrument for collaborative management with reduced breeding propensity. Decline in of the Wadden Sea. However, these

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 489 international conventions and collaborative (1) Population Assessments. While many partnerships have not been entirely effective species are of concern because of low in protecting waterbirds and their habitats in numbers or perceptions of substantial the Wadden Sea (Boere & Piersma 2012). population decline, robust data for Conflicting economic and political interests assessing population trajectories are of the multiple nations continue to often lacking. As well, spatial variation challenge conservation planning and in demographic rates, where such implementation in the Baltic/Wadden Sea data exist, suggests that population region. Conservation and management of trajectories might be driven by sub- the Common Eider, and other sea ducks in populations. However, demographically Europe, would be greatly enhanced by the distinct sub-populations are not clearly development of a conservation plan to identified for many species. help prioritise, coordinate and implement (2) Demographic trends and relationships. We research, monitoring and management have little information on spatial and actions (also see Elmberg et al. 2006). temporal patterns in survival and reproduction for many species of Decision making in the face of conservation concern. Furthermore, uncertainty functional relationships between demographic parameters and habitat We have outlined a basic framework that quality or other limiting factors, plus integrates critical steps for defining actions underlying biological mechanisms to conserve populations of concern, driving those patterns, are often ranging from identifying plausible unknown. These include harvest, cross- hypotheses about factors influencing seasonal and density-dependent effects. demographic parameters and population status to determining suites of potentially (3) Status and trends of key limiting factors. effective management or policy actions to Models assume linkages between monitoring the outcomes of those actions. demography and habitat quantity and We also provided examples of how this quality or the presence of other framework has been used for several stressors, e.g. “invasive” species (genetic waterfowl taxa, including the development competitors), contaminants, or of sophisticated planning models that predators. Key to targeting conservation quantify key relationships between stressors, action and evaluating the outcome of demography and desired management management actions is an understanding outcomes (Table 3). These are essential of how environmental conditions steps towards addressing population change due to and in spite of concerns and revealing critical research management actions. However, such needs. Uncertainties exist at each stage of information is often absent at spatial or the framework, which generally can be temporal scales consistent with the scale grouped into the following categories: of conservation concerns.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 490 Waterfowl populations of conservation concern

(4) Predicting future relationships in a changing Despite these many uncertainties, many world. Modelled relationships between conservation decisions must be made now. limiting factors and waterfowl Often these decisions are time sensitive and demography are built on expert opinion cannot wait for perfect information. As and/or existing data. However, managers Nichols et al. (2011) highlight, such decisions cannot assume that systems they are are regular occurrences for population trying to manage are static (i.e. constant managers. There is a large field of adaptive through time and space), and therefore management and structured decision- that known current values are useful for making that describes an active, transparent predicting future patterns. For example, and defensible process for arriving at climate change may induce changes in the decisions and reducing uncertainty to ecological processes that drive patterns of inform future decisions (e.g. Williams et al. waterfowl distribution and demography, 2002; Conroy et al. 2012). It is not our intent which may alter those patterns (Nichols et to repeat this information here, but rather to al 2011). This potential change in system focus on the use of population models for dynamics through time may be difficult advancing adaptive decision-making. to predict, but is valuable to explore We recognise that models are an over- (Drever et al. 2012). simplification of complex relationships, with inherent errors, uncertainties and (5) Predicting outcomes and cost effectiveness of assumptions. However, models are key management or policy actions. Uncertainty in components to adaptive management the above categories can hinder the because they provide a defensible structure identification and implementation of from which to communicate, make appropriate conservation actions. For decisions, and learn about population example, competing hypotheses about dynamics and the impacts of our decisions. relationships between limiting factors Both conceptual and quantitative models and demography may lead to different articulate contrasting views about how the management strategies. Moreover, limited systems we are trying to influence operate, ability to control how management allowing us to predict potential outcomes of actions are deployed (e.g. due to alternate conservation actions. Further, by unplanned financial constraints), and also specifying key relationships, parameterizing the effectiveness of actions given equations and conducting sensitivity environmental variation, make predicting analyses, we bring key information gaps and and also realising desired outcomes debate into greater focus. This focus can challenging. Uncertainty regarding the inform research agendas, strengthen success of conservation outcomes fundraising efforts, and guide development confounds estimating return on of conservation programmes. Finally, investment, an essential component in competing hypotheses about relationships determining how best to allocate the between limiting factors and demography limited finances allocated to conservation. can be weighted based upon the degree of

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 491 confidence we have in the probability that (Green 1996). While the approach we they are correct (Nichols et al. 2011). These described herein has been used for several weights can be modified as we learn through data-rich species, we argue that it is equally directed research or implementation applicable for data-poor species. This of conservation programmes arising benefit is due, in part, to the ability from strategic decision-making, thereby of conceptual models to help shape improving future decisions. Thus, while not and communicate biological reasoning. a panacea for every situation, models However, we fully recognise the challenges provide a mechanism for structured, long- associated with conserving species that term learning; the most crucial research migrate across multiple countries that questions and monitoring needs typically potentially have different levels of resources emerge during this process and can be (people and financial resources) and integrated with conservation action plans. perspectives on goals and collaboration for conservation. The approach we outlined can Conclusions be useful for rapidly assessing risks and The structure of conservation efforts has guiding conservation efforts for diverse evolved somewhat differently for each of waterfowl species of conservation concern. the waterfowl species of concern, reflecting different issues, histories, geopolitical Acknowledgements context and associated uncertainties about This paper is based on a special session at current and future system dynamics and the 2013 Ecology and Conservation of management effectiveness. However, the North American Waterfowl Symposium examples we present share common (ECNAW). We would like to thank the components: formulation of hypotheses at ECNAW scientific committee for accepting initial stages; application of conceptual our session proposal and guiding and quantitative models that integrate development of content. We are also greatly hypotheses with conservation actions; indebted to the following presenters and development of formal conservation their coauthors; without your hard work frameworks and plans based on adaptive there would have been no session: R.R. management; and use of collaborations and Bielefeld, T. Bowman, P. Devers, J.-M. partnerships, largely through the NAWMP’s DeVink, K. Guyn, A.D. Fox and J. Lovvorn. JVs. We believe this approach provides the Finally, we thank A. Green, B. Mattsson, E. most defensible, and perhaps repeatable, Rees and L. Webb for their valuable method for allocating limited resources and comments on earlier versions of this paper. advancing learning. Review of IUCN threats for threatened References and endangered waterfowl worldwide Afton, A.D. & Anderson, M.G. 2001. Declining indicated great uncertainty in fully scaup populations: a retrospective analysis of understanding the limiting factors and long-term population and harvest survey data. actions required to alleviate their impacts Journal of Wildlife Management 65: 781–796.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 492 Waterfowl populations of conservation concern

Alisauskas, R.T., Traylor, J.J., Swoboda, C.J. 2010. Development of Scaup Conservation Plan. & Kehoe, F.P. 2004. Components of Unpublished report from the Third population growth rate for white-winged Scaup Workshop – Migration, Winter and scoters in Saskatchewan, Canada. Animal Human Dimensions, 21–25 September Biodiversity and Conservation 27: 1–12. 2009, Memphis Tennessee, USA. U.S. Anderson, J.T. 2008. Survival, habitat use and Geological Survey, Jamestown, North movements of female Northern Pintails Dakota, USA. wintering along the Texas coast. M.Sc. thesis, Bentzen, R.L. & Powell, A.N. 2012. Population Texas A & M University, Kingsville, USA. dynamics of King Eiders breeding in Anderson, M.G., Caswell, D., Eadie, J.M., northern Alaska. Journal of Wildlife Management Herbert, J.T., Huang, M., Humburg, D.D., 76: 1011–1020. Johnson, F.A., Koneff, M.D., Mott, S.E., Bielefeld, R.R. &. Cox, R.R., Jr. 2006. Survival Nudds, T.D., Reed, E.T., Ringelman, J.K., and cause-specific mortality of adult female Runge, M.C. & Wilson, B.C. 2007. Report from mottled ducks in east-central Florida. Wildlife the Joint Task Group for Clarifying North Society Bulletin 34: 388–394. American Waterfowl Management Plan Population Bielefeld, R.R., Brasher, M.G., Moorman, T.E. & Objectives and Their Use in Harvest Management. Gray, P.N. 2010. Mottled Duck (Anas NAWMP Joint Task Group unpublished fulvigula). In A. Poole (ed.), The Birds of North report, U.S. Fish & Wildlife Service and U.S. America Online. Cornell Lab of Ornithology, Geological Survey, Washington D.C., USA. Ithaca, New York, USA. Accessible at Accessible at http://nawmprevision.org/ http://bna.birds.cornell.edu/bna/. sites/default/files/jtg_final_report.pdf. BirdLife International. 2004. Birds in Europe: An, S., Li, H., Guan, B., Zhou, C., Wang, Z., Population Estimates, Trends and Conservation Deng, Z., Zhik, Y., Lui, Y., Xu, C., Fang, S., Status. BirdLife Conservation Series No. 12. Jiang, J. & Li, H. 2007. China’s natural BirdLife International, Wageningen, the wetlands: past problems, current status and Netherlands. future challenges. Ambio 36: 355–342. Black Duck Joint Venture. 2008. Triennial report to Anonymous. 1999. Final Feasibility Report and the North American Waterfowl Management Plan Program Environmental Impact Statement. Central Committee, July, 2008. Unpublished report, and Southern Florida Project Comprehensive Review U.S. Fish and Wildlife Service, Laurel, Study. U.S. Army Corps of Engineers, Maryland, USA. Southern Florida Water Management District. Boere, G.C. & Piersma, T. 2012. Flyway Accessible at http://www.evergladesplan.org. protection and the predicament of our Austin, J.E., Afton, A.D., Anderson, M.G., Clark, migratory birds: a look at international R.G., Custer, C.M., Lawrence, J.S., Pollard, conservation policies and the Dutch Wadden J.B. & Ringelman, J.K. 2000. Declining scaup Sea. Oceans and Coastal Management 68: populations: issues, hypotheses and research 157–168. needs. Wildlife Society Bulletin 28: 254–263. Bowman, T.D, Silverman, E.D., Gilliland, S.G. & Austin, J.E., Boomer, G.S., Brasher, M., Clark, Leirness, J.B. 2015. Status and trends of R.G., Cordts, S.D., Eggeman, D., Eichholz, North American sea ducks: reinforcing the M.W., Hindman, L., Humburg, D.D., Kelley, need for better monitoring. In J.-P.L. Savard, J.R., Jr., Kraege, D., Lyons, J.E., Raedeke, D.V. Derksen, D. Esler & J.M. Eadie (eds.), A.H., Soulliere, G.J. & Slattery, S.M. Ecology and Conservation of North American Sea

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 493

Ducks, pp. 1–27. Studies in Avian Biology (in Dahl, T.E. & Stedman, S.M. 2013. Status and press), CRP Press, New York, USA. trends of wetlands in the coastal watersheds of the Brook, R.W., Duncan, D.C., Hines, J.E., Carrière, Conterminous United States 2004 to 2009. S. & Clark, R.G. 2005. Effects of small U.S. Department of the Interior, Fish and mammal cycles on productivity of boreal Wildlife Service and National Oceanic ducks. Wildlife Biology 11: 3–11. and Atmospheric Administration, National Camphuysen, C.J., Berrevoets, C.M., Cremers, Marine Fisheries Service, Washington D.C., H.J.W.M., Dekinga, A., Dekker, R., Ens, USA. B.J., van der Have, T.M., Kats, R.K.H., Desholm, M., Christensen, T.K., Scheiffarth, Kuiken, T., Leopold, M.F., van der Meer, J. G., Hario, M., Andersson, Å., Ens, B., & Piersma, T. 2002. Mass mortality of Camphuysen, C.J., Nilsson, L., Waltho, C.M., Common Eiders (Somateria mollissima) in Lorentsen, S.-H., Kuresoo, A., Kats, R.K.H., the Dutch Wadden Sea, winter 1999/ Fleet, D.M. & Fox, A.D. 2002. Status of 2000: starvation in a commercially exploited the Baltic/Wadden Sea population of the wetland of international importance. Biological Common Eider Somateria m. mollissima. Conservation 106: 303–317. Wildfowl 53: 167–204. Canadian Wildlife Service Waterfowl Committee. Devers, P.K., McGowan, C., Mattson, B., Brook, 2012. Population Status of Migratory Game Birds R., Huang, M., Jones, T., McAuley, D. & in Canada: November 2011. Canadian Wildlife Zimmerman, G. 2011. American Black Duck Service Migratory Birds Regulatory Report Adaptive Management – Preliminary Integrated Number 37. Canadian Wildlife Service, Habitat and Population Dynamics Framework. Gatineau, Quebec, Canada. A Case Study from the Structured Decision Caswell, H. 2000. Prospective and retrospective Making Workshop, 13–17 September 2010. perturbation analyses: Their roles in National Conservation Training Center, conservation biology. Ecology 81: 619–627. Shepherdstown, West Virginia, USA. Clemen, R.T. 1996. Making Hard Decisions: an Devries, J.H., Armstrong, L.M., MacFarlane, R.J., Introduction to Decision Analysis. Second Edition. Moats, L. & Thoroughgood, P.T. 2008. Duxbury Press, Belmont, California, USA. Waterfowl nesting in fall seeded and Conroy, M.J. & Peterson, J.T. 2012. Decision spring seeded cropland in Saskatchewan. Making in Natural Resource Management: A Journal of Wildlife Management 72: 1790–1797. Structured, Adaptive Approach. John Wiley & Drever, M.C., Clark, R.G., Derksen, C., Slattery, Sons, New York, USA. S.M., Toose, P. & Nudds, T.D. 2012. Conroy, M.J., Miller, M.W. & Hines, J. E. 2002. Population vulnerability to climate change Identification and synthetic modelling of linked to timing of breeding in boreal ducks. factors affecting American Black Duck Global Change Biology 18: 480–492. populations. Wildlife Monographs 150: 1– Duncan, D.C., Miller, M.R., Guyn, K.L., Flint, 64. P.L. & Austin, J.E. 2003. Part 2. A Custer, C.M. & Custer, T.W. 2000. Organochlorine management and research plan for Northern and trace element contamination in wintering Pintails. In M.R. Miller, D.C. Duncan, K.L. and migrating diving ducks in the southern Guyn, P.L. Flint, & J.E. Austin (eds.), The Great Lakes, USA, since the zebra mussel Northern Pintail in North America: The Problem invasion. Environmental Toxicology and Chemistry and a Prescription for Recovery, pp. 28–38. Ducks 19: 2821–2829. Unlimited, Sacramento, California, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 494 Waterfowl populations of conservation concern

Ekroos, J., Fox, A.D., Christensen, T.K., Petersen, Greenland and Canada. Wildlife Biology 15: I.K., Kilpi, M., Jónsson, J.E., Green, M., 24–36. Laursen, K., Cervencl, A., de Boer, P., Green, A.J. 1996. Analyses of globally threatened Nilsson, L., Meissner, W., Garthe, S. & Öst, Anatidae in relation to threats distribution, M. 2012. Declines amongst breeding Eider migration patterns, and habitat use. Somateria mollissima numbers in the Baltic/ Conservation Biology 10: 1435–1445. Wadden Sea flyway. Ornis Fennica 89: 81–90. Gregory, R., Failing, L., Harstone, M., Long, G., Elmberg, J., Nummi, P., Pöysä, H., Sjöberg, K., McDaniels, T. & Ohlson, D. 2012. Structured Gunnarsson, G., Clausen, P., Guillemain, Decision Making: A Practical Guide to M., Rodrigues, D. & Väänänen, V.M. 2006. Environmental Management Choices. John Wiley The scientific basis for new and sustainable & Sons, New York, USA. management of European migratory ducks. Harwell, M.A., Gentile, J.H. & Parker, K.R. 2012. Wildlife Biology 12: 121–127. Quantifying population-level risks using and Fleskes, J.P., Yee, J.L., Yarris, G.S., Miller, M.R. & individual based model: Sea otters, Harlequin Casazza, M.L. 2007. Northern Pintail and Ducks, and Exxon Valdez oil spill. Integrated Mallard survival in California relative to Environmental Assessment and Management 8: habitat, abundance and hunting. Journal of 503–522. Wildlife Management 71: 2238–2248. Haukos, D.A., Miller, M.R., Orthmeyer, D.L., Flint, P.L. 2013. Changes in size and trends of Takekawa, J.Y., Fleskes, J.P., Casazza, M.L., North American sea duck populations Perry, W.M. & Moon, J.A. 2006. Spring associated with North Pacific oceanic regime migration of Northern Pintails from Texas shifts. Marine Biology 160: 59–65. and New Mexico, USA. Waterbirds 29: Flint, P.L., Grand, J.B., Fondell, J.A. & Morse, J.A. 127–136. 2006. Population dynamics of Greater Scaup Iles, D.T. 2012. Drivers of nest success and breeding on the Yukon-Kuskokwim Delta, stochastic population dynamics of the Alaska. Wildlife Monographs 162: 1–22. common eider (Somateria mollissima). Florida Fish and Wildlife Commission. 2011. A Unpublished M.Sc. thesis, Utah State Conservation Plan for the Florida Mottled Duck. University, Logan, Utah, USA. Florida Fish & Wildlife Commission, International Union for Conservation of Nature Tallahassee, Florida, USA. (IUCN). 2013. IUCN Red List of Threatened Gilchrist, H.G., Gilliland, S., Rockwell, R., Species. Version 2013.2. IUCN, Gland, Savard, J.-P., Robertson, G.J. & Merkel, F.R. Switzerland. Accessible at www.iucnredlist.org 2001. Population Dynamics of the Northern (last accessed 11 December 2013). Common Eider in Canada and Greenland: Results Johnson, F.A. 2009. Variation in population of a Computer Simulation Model. Canadian growth rates of Mottled Ducks in Texas and Wildlife Service Unpublished Report. Louisiana. U.S. Geological Survey National Wildlife Research Centre, Carleton Administrative Report to the U.S. Fish and University, Ottawa, Ontario, Canada. Wildlife Service. U.S. Geological Survey, Gilliland, S.G., Gilchrist, H.G., Rockwell, R.F., Gainesville, Florida, USA. Robertson, G.J., Savard, J-P.L., Merkel, F. Junk, W.J., An, S., Finlayson, C.M., Gopal, B., & Mosbech, A. 2009. Evaluating the Kvêt, J., Mitchell, S.A., Mitsch, W.J. & sustainability of harvest among northern Robarts, R.D. 2013. Current state of Common Eiders Sometaria mollissima borealis in knowledge regarding the world’s wetlands

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 495

and their future under global climate change: The problem and a Prescription for Recovery, a synthesis. Aquatic Science 75: 151–167. pp. 6–25. Ducks Unlimited, Sacramento, Koons, D.N., Rotella, J.J., Willey, D.W., Taper, M., California, USA. Clark, R.G., Slattery, S., Brook, R.W, Miller, M.R., Takekawa, J.Y., Fleskes, J.P., Corcoran, R.M. & Lovvorn, J.R. 2006. Lesser Orthmeyer, D.L., Casazza, M.L. & Perry, Scaup population dynamics: what can be W.M. 2005. Spring migration of Northern learned from available data? Avian Pintails from California’s Central Valley Conservation and Ecology 1(3): 6. wintering area tracked with satellite Kowalchuk, T.A. 2012. Breeding ecology of Northern telemetry: routes, timing and destinations. Pintails in prairie landscapes: tests of habitat Canadian Journal of Zoology 83: 1314–1332. selection and reproductive trade-off models. Moon, J.A. & Haukos, D.A. 2006. Survival of Unpublished Ph.D. thesis, University of female Northern Pintails wintering in the Saskatchewan, Saskatoon, Canada. Playa Lakes Region of northwestern Link, W.A., Sauer, J.R. & Niven, D.K. 2006. A Texas. Journal of Wildlife Management 70: hierarchical model for regional analysis of 777–783. population change using Christmas Bird Nichols, J.D., Koneff, M.D., Heglund, P.J., Count Data, with application to the Knutson, M.G., Seamans, M.E., Lyons, American Black Duck. Condor 108: 13–24. J.E., Morton, J.M., Jones, M.T., Boomer, Lovvorn, J.R., Grebmeier, J.M., Cooper, L.W., G.S. & Williams, B.K. 2011. Climate Bump, J.K. & Richman, S.E. 2009. Modeling change, uncertainty and natural resource marine protected areas for threatened eiders management. Journal of Wildlife Management in a climatically changing Bering Sea. 75: 6–18. Ecological Applications 19: 1596–1613. North American Bird Conservation Initiative. Mattsson, B.J., Runge, M.C., Devries, J.H., Boomer, 2000. Bird Conservation Region Descriptions. U.S. G.S., Eadie, J.M., Haukos, D.A., Fleskes, J.P., Fish and Wildlife Service, Division of Bird Koons, D.N., Thogmartin, W.E., & Clark, R.G. Habitat Conservation, Arlington, Virginia, 2012. A modeling framework for integrated USA. Accessible at http://www.nabci-us.org/ harvest and habitat management of North aboutnabci/bcrdescrip.pdf. American waterfowl: case-study of Northern North American Waterfowl Management Plan Pintail metapopulation dynamics. Ecological (NAWMP) Committee. 2012. North Modelling 225: 146–158. American Waterfowl Management Plan Miller, M.R. & Duncan, D.C. 1999. The 2012: People conserving waterfowl and Northern Pintail in North America: status wetlands. North American Waterfowl and conservation needs of a struggling Management Plan Committee Report to the population. Wildlife Society Bulletin 27: Canadian Wildlife Service and the U.S. Fish 788–800. and Wildlife Service. U.S. Department of the Miller, M.R., Duncan, D.C., Guyn, K.L., Flint, Interior, Washington D.C., USA. P.L. & Austin, J.E. (eds.). 2003. Part 1. Nudds, T.D., Miller, M.W. & Ankney, C.D. 1996. Proceedings of the Northern Pintail Black ducks: harvest, mallards or habitat? Workshop, 25 March 2001, Sacramento, In J. T. Ratti (ed.), Seventh International California, USA. In M.R. Miller, D.C. Waterfowl Symposium, pp. 50–60. Institute for Duncan, K.L. Guyn, P.L. Flint & J.E. Austin Wetland and Waterfowl Research, Memphis, (eds.), The Northern Pintail in North America: Tennessee, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 496 Waterfowl populations of conservation concern

Osnas, E.E., Runge, M.C., Mattsson, B.J., Austin, Stotts, V.D. 1989. Population ecology J.E., Boomer, G.S., Clark, R.G., Devers, P., and harvest of the American black duck: Eadie, J.M., Lonsdorf, E.V. & Tavernia, B.G. a review. Wildlife Society Bulletin 17: 379– 2014. Managing harvest and habitat as 406. integrated components. Wildfowl (Special Schamber, J.L., Flint, P.L., Grand, J.B., Wilson, Issue No. 4): 305–328. H.M. & Morse, J.A. 2009. Population Petrie, M.J., Drobney, R.D. & Sears, D.T. 2000. dynamics of Long-tailed Ducks breeding on Mallard and Black Duck breeding parameters the Yukon-Kuskokwim Delta, Alaska. Arctic in New Brunswick: a test of the reproductive 62: 190–200. rate hypothesis. Journal of Wildlife Management Schaub, M. & Abadi, F. 2011. Integrated 64: 832–838. population models: a novel analysis Petrie, S.A, Badzinski, S.S. & Drouillard, K.G. framework for deeper insight into population 2007. Contaminants in lesser and greater dynamics. Journal of Ornithology 152: S227– scaup staging on the lower Great Lakes. S237. Archives for Environmental Contaminants and Sea Duck Joint Venture (SDJV) Management Toxicology 52: 580–589. Board. 2008. Sea Duck Joint Venture Platt, J. 1964. Strong inference. Science 146: Strategic Plan 2008–2012. U.S. Fish and 347–353. Wildlife Service, Anchorage, Alaska, USA, Podruzny, K.M., Devries, J.H., Armstrong, L.M. and Canadian Wildlife Service, Sackville, & Rotella, J.J. 2002. Long-term response of New Brunswick, Canada. Accessible at Northern Pintails to changes in wetlands and http://seaduckjv.org. agriculture in the Canadian Prairie Pothole Shaffer, M.L. 1981. Minimum population sizes Region. Journal of Wildlife Management 66: for species conservation. BioScience 31: 993–1010. 131–134. Raftovich, R.V. & Wilkins, K.A. 2013. Migratory U.S. Fish and Wildlife Service. 2010. Northern bird hunting activity and harvest during the Pintail Harvest Strategy 2010. U.S. Department 2011–12 and 2012–13 hunting seasons. U.S. Fish of the Interior, Fish and Wildlife Service, and Wildlife Service, Laurel, Maryland, USA. Division of Migratory Bird Management, Rice, M.B., Haukos, D.A., Dubovsky, J.A. & Washington D.C., USA. Runge, M.C. 2010. Continental survival and U.S. Fish and Wildlife Service. 2011. Migratory recovery rates of Northern Pintails using Bird Program, Focal Species Strategy. U.S. band-recovery data. Journal of Wildlife Department of Interior, Fish and Wildlife Management 74: 778–787. Service, Division of Migratory Bird Rigby, E.A. & Haukos, D.A. 2012. Breeding Management, Arlington, Virginia, USA. season survival and breeding incidence of Accessible at www.fws.gov/migratorybirds/ female Mottled Ducks on the Upper Texas currentbirdissues/management/FocalSpecies. Gulf Coast. Waterbirds 35: 260–269. html. Riordan, B., Verbyla, D. & McGuire, A.D. 2006. Varner, D.M., Bielefeld, R.R. & Hepp, G.R. 2013. Shrinking ponds in subarctic Alaska based on Nesting ecology of Florida Mottled Ducks 1950–2002 remotely sensed images. Journal of using altered habitats. Journal of Wildlife Geophysical Research: (2005–2012)111.G4. Management 77: 1002–1009. Rusch, D.H., Ankney, C.D., Boyd, H., Longcore, Varner, D.M., Hepp, G.R. & Bielefeld, R.R. 2014. J.R., Montalbano, F., Ringelman, J.K. & Annual and seasonal survival of adult female

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 Waterfowl populations of conservation concern 497

Mottled Ducks in southern Florida, USA. Wilson, H.M., Flint, P.L., Powell, A.N., Grand, The Condor 116: 134–143. J.B. & Moran, C.L. 2012. Population ecology Williams, B., Nichols, J.D. & Conroy, M.J. 2002. of breeding Pacific Common Eiders on the Analysis and Management of Animal Populations: Yukon-Kuskokwim Delta, Alaska. Wildlife Modeling, Estimation and Decision Making. Monographs 182: 1–28. Academic Press, San Diego, California, USA. Zimpfer, N.L., Rhodes, W.E., Silverman, E.D., Wilson, B.C. 2007. North American Waterfowl Zimmerman, G.S. & Richkus, K.D. 2013. Management Plan, Gulf Coast Joint Venture: Trends in waterfowl breeding populations, Mottled Duck Conservation Plan. North 1955–2013. U.S. Fish and Wildlife Service American Waterfowl Management Plan, Administrative Report, 12 July 2013. U.S. Fish Albuquerque, New Mexico, USA. and Wildlife Service, Laurel, Maryland, USA.

Photograph: Common Eider males and females at Joekulsarlon, Iceland, by Imagebroker/FLPA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 470–497 498

Waterfowl in Cuba: current status and distribution

PEDRO BLANCO RODRÍGUEZ1, FRANCISCO J. VILELLA2* & BÁRBARA SÁNCHEZ ORIA3

1Instituto de Ecología y Sistemática, Ministerio de Ciencia, Tecnología y Medio Ambiente, Apartado Postal 8010, La Habana 10800, Cuba. 2U.S. Geological Survey, Mississippi Cooperative Fish and Wildlife Research Unit, Department of Wildlife, Fisheries and Aquaculture, Mississippi State University, Mississippi 39762, USA. 3Instituto de Ecología y Sistemática, Ministerio de Ciencia, Tecnología y Medio Ambiente, Apartado Postal 8010, La Habana 10800, Cuba. *Correspondence author. E-mail: [email protected]

Abstract Cuba and its satellite islands represent the largest landmass in the Caribbean archipelago and a major repository of the region’s biodiversity. Approximately 13.4% of the Cuban territory is covered by wetlands, encompassing approximately 1.48 million ha which includes mangroves, flooded savannas, peatlands, freshwater swamp forests and various types of managed wetlands. Here, we synthesise information on the distribution and abundance of waterfowl on the main island of Cuba, excluding the numerous surrounding cays and the Isla de la Juventud (Isle of Youth), and report on band recoveries from wintering waterfowl harvested in Cuba by species and location. Twenty-nine species of waterfowl occur in Cuba, 24 of which are North American migrants. Of the five resident Anatid species, three are of conservation concern: the West Indian Whistling-duck Dendrocygna arborea (globally vulnerable), White-cheeked Pintail Anas bahamensis (regional concern) and Masked Duck Nomonyx dominicus (regional concern). The most abundant species of waterfowl wintering in Cuba include Blue-winged Teal A. discors, Northern Pintail A. acuta, and Northern Shoveler A. clypeata. Waterfowl banded in Canada and the United States and recovered in Cuba included predominantly Blue-winged Teal, American Wigeon and Northern Pintail. Banding sites of recovered birds suggest that most of the waterfowl moving through and wintering in Cuba are from the Atlantic and Mississippi flyways. Threats to wetlands and waterfowl in Cuba include: 1) egg poaching of resident species, 2) illegal hunting of migratory and protected resident species, 3) mangrove deforestation, 4) reservoirs for irrigation, 5) periods of pronounced droughts, and 6) hurricanes. Wetland and waterfowl conservation efforts continue across Cuba’s extensive system of protected areas. Expanding

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 Waterfowl in Cuba 499

collaborations with international conservation organisations, researchers and governments in North America will enhance protection of waterfowl and wetlands in Cuba.

Key words: Anatidae, Caribbean, Cuba, conservation, habitat, management.

The Caribbean islands are a priority area island group in the West Indies (Fig. 1). It globally for biodiversity conservation harbours the greatest biological diversity and because of the high rate of habitat loss in the degree of endemism in the West Indies; over region (Brooks et al. 2006; Shi et al. 2005). 50% of its flowering plants and 32% of The archipelago straddles the boundary of its vertebrates are unique to the country the Neotropical and Nearctic regions with (ACC-ICGC 1978; Woods 1989; González tropical and subtropical climates. Rainfall 2002; Rodríguez-Schettino 2003; Borroto patterns in the insular Caribbean are highly & Mancina 2011). Despite its regional variable, and on many islands precipitation importance, very little published information exceeds potential evapotranspiration, a on waterfowl in Cuba, including on their condition that provides ample water to distribution and general ecology, has become sustain wetland environments (Lugo 2002). available to scientists working on these With a total land mass of 110,860 km2 species in other parts of their range (Scott & and over 1,600 offshore islands and cays, Carbonell 1986; Santana 1991). Given the Cuba represents the largest and most diverse scarcity of publications on Cuban waterfowl

Figure 1. Map of the West Indies indicating major island groups.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 500 Waterfowl in Cuba and threats to wetland conservation in the Approximately 30% of the 1.48 million region, it is important to summarise ha of Cuban wetlands are included in the available information for the benefit of the national system of protected areas. Some of broader scientific community. Furthermore, the most important wetlands include: the condensing the available literature on complex of lagoons south of Pinar del Río waterfowl and wetlands of Cuba into a single province, the Birama marshes in Granma document may also be useful for researchers province, the Río Máximo wildlife refuge in and managers interested in the region. Camagüey province and the Lanier Swamp Wetland types in Cuba include mangrove in the Isla de la Juventud. With a total area of forest (riparian and estuarine), freshwater 450 km2, the Zapata Swamp (22°01’–22°40’ marsh, seasonally flooded savanna, swamp N, 80°33’–82°09’ W) is the largest and most forest, riverine wetlands and managed complex drainage system in the Caribbean wetlands such as salt pans and rice fields (Kirkconnell et al. 2005). Some 625,354 ha (Borhidi et al. 1993). Coastal regions of of this large wetland complex are protected Cuba feature large expanses of mangrove as a Biosphere Reserve. Here we present forest characterised by the four tree species information on the distribution and common in the Caribbean (Red Mangrove abundance of waterfowl on the main Rhyzophora mangle, Black Mangrove Avicennia island of Cuba, excluding the numerous germinans, White Mangrove Laguncularia surrounding cays and the Isle of Youth. We racemosa and Buttonwood Mangrove also report on band recoveries, by species Conocarpus erectus). The interior regions of and location, for waterfowl caught and the main island of Cuba are characterised by ringed in Canada and the United States that flat topography where seasonally flooded were harvested in mainland Cuba. There savannas are found. These savannas have been no formal waterfowl banding represent the most floristically diverse programmes to date within Cuba. wetlands of the Caribbean and include a great number of endemic palm species Methods (Armenteros et al. 2007). Dominant species We reviewed and summarised information of savanna wetlands include Eleocharris on abundance patterns and geographic interstincta, Claudium jamaicense, Paspalum distribution from a large number of giganteum, Cyperus sp., Isoetes palustris, unpublished reports and publications for Erianthus giganteus, Thalia geniculata, Nymphaea the period 1975–2010 (most notably from odorata and Brasenia scheberi. Swamp forests Garrido & Schwartz 1968; Garrido 1980; are characterised by arboreal elements and Llanes et al. 1987; Sánchez et al. 1991; Torres epiphytes with canopy heights of up to 20 & Solana 1994; Acosta & Mugica 1994; m. Here the dominant species include Goossen et al. 1994; Morales & Garrido Tabebuia angustata, Fraxinus cubensis, Annona 1996; Melián 2000, Rodríguez 2000; Peña glabra, Gueltarda combiri, Bucida palustris, et al. 2000; Wiley et al. 2002; Barrios et al. Hibiscus tiliaceus and Chrysobalanus icaco 2003). Presence of duck species and (Armenteros et al. 2007). location coordinates were transferred to

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 Waterfowl in Cuba 501

118 cartographic quadrangles (37 × 18.5 km, (Raffaele et al. 1998), highlighting the scale 1:100,000) covering the entire main regional importance of Cuba for waterfowl island. Residence categories for each species in the Caribbean. North American in Cuba followed Garrido and Kirkconnell migratory waterfowl contribute greatly to (2000). For instance, bimodal resident (BR) the widespread distribution of ducks in refers to species (e.g. Wood Duck) that Cuba and were registered in 98 (83%) of the include both permanent breeding residents 118 topographic quadrangles (Fig. 2). The and also a small number of transient migratory species are almost exclusively migrants; winter resident and transient from North America, with the possible (WR-T) refers to species that mostly winter exception of the White-faced Whistling in Cuba but with some individuals that Duck Dendrocygna viduata which comes from occur as transients as they move through Central and/or South America and is Cuba to and from wintering sites on the considered an “accidental” species in mainland (Central and South America); Cuba. Migrant waterfowl most frequently mostly transient winter residents (T-WR) are recorded in topographic quadrangles species that only occur as transients during included Blue-winged Teal Anas discors, migration peaks; introduced breeding American Wigeon A. americana, Northern residents (I-BR) are introduced species (e.g. Shoveler A. clypeata, Northern Pintail the Muscovy Duck) known to breed in A. acuta and Fulvous Whistling-duck Cuba; and accidental (Ac) species occur only Dendrocygna bicolor. A total of 1,842 bands occasionally in Cuba. from 11 waterfowl species were recovered in Band recovery information of waterfowl Cuba during 1930–2010 (Table 2). Of these, harvested in Cuba from 1930–2010 was 91.5% were recovered from Blue-winged obtained from the U.S. Geological Survey’s Teal, 2.2% from American Wigeon and Bird Banding Laboratory and the 2.1% from Northern Pintail. A small Canadian Bird Banding Office (Blanco & number of Wood Duck banded in Florida Sánchez 2005). Location information from and Georgia were recovered in Cuba, field surveys and band recoveries were suggesting that, in addition to the georeferenced and incorporated in a permanent breeding residents, occasional Geographic Information System using transient individuals arrive from North ArcView 3.1 (ESRI 2001). America; the species is therefore classed as a bimodal resident (Blanco & Sánchez 2005; Results Garrido & Kirkconnell 2011). Waterfowl in Cuba are represented by 29 Arrival dates and presence of migratory species in 14 genera (Garrido & Kirkconnell waterfowl have been reported by various 2011; Raffaele et al. 1998), of which 24 Cuban researchers working in natural are migratory species with varying degrees wetlands and fields with rice Oryza sp. of residence (Table 1). Species recorded cultivation. Unfortunately, much of this in Cuba represent 93.5% (29 of 31) of information is only available in scientific waterfowl reported for the West Indies journals published in Cuba or in regional

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 502 Waterfowl in Cuba

Table 1. Waterfowl species in Cuba by residence type according to Garrido and Kirkconnell (2011). Residence categories include: accidental (Ac), permanent resident (PR), bimodal resident (BR), bimodal resident and transient (BR-T) winter resident and transient (WR-T), mostly transient winter resident (T-WR), introduced breeding resident (I-BR).

Common name Scientific name Category of residence

White-faced Whistling-Duck Dendrocygna viduata Ac Black-bellied Whistling-Duck Dendrocygna autumnalis PR West Indian Whistling-Duck Dendrocygna arborea PR Fulvous Whistling-Duck Dendrocygna bicolor I-BR Greater White-fronted Goose Anser albifrons Ac Snow Goose Chen caerulescens Ac Canada Goose Branta canadensis Ac Tundra Swan Cygnus columbianus Ac Muscovy Duck Cairina moschata I-BR Wood Duck Aix sponsa BR-T Gadwall Anas strepera Ac American Wigeon Anas americana WR-T Mallard Anas platyrhynchos T-WR Blue- winged Teal Anas discors WR-T Cinnamon Teal Anas cyanoptera Ac Northern Shoveler Anas clypeata WR-T White-cheeked Pintail Anas bahamensis PR Northern Pintail Anas acuta WR-T Green-winged Teal Anas crecca WR-T Canvasback Aythya valisineria Ac Redhead Aythya americana Ac Ring-necked Duck Aythya collaris WR-T Greater Scaup Aythya marila Ac Lesser Scaup Aythya affinis WR-T Bufflehead Bucephala albeola Ac Hooded Merganser Lophodytes cucullatus Ac Red-breasted Merganser Mergus serrator WR-T Masked Duck Nomonyx dominicus PR Ruddy Duck Oxyura jamaicensis BR-T

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 Waterfowl in Cuba 503

Figure 2. Number of waterfowl species recorded per topographic quadrangle in Cuba (2007–2013). publications, which may be difficult to access of 15 Cuban provinces originated in by researchers outside the Caribbean (Acosta three Canadian provinces (Manitoba, et al. 1992; Blanco 1996; Blanco et al. 1996; Saskatchewan and Ontario) as well as the Sánchez & Rodríguez 2000; Sánchez et al. states of Illinois, Iowa, Louisiana, Michigan, 2008). Published reports indicate the major Minnesota, Missouri, North Dakota and entry points for migrant waterfowl include South Dakota (Fig. 3). Bands were recovered the Zapata Swamp lagoons, the Río Máximo mostly from September–April, coinciding refuge and mangrove wetlands in cays of with the months when migratory waterfowl the Sabana-Camagüey archipelago. Coastal are most commonly found in the Caribbean lagoons in the south-central part of the (Raffaele et al. 1998). These results suggest a island adjacent to the rice producing regions prominent role for both the Mississippi and (e.g. Sur del Jíbaro) are also important arrival Atlantic Flyways in the stopover and sites for the birds (Mugica 2000; Mugica et al. migration patterns of waterfowl in Cuba. 2001; Acosta and Mugica 2006). Migratory waterfowl begin to arrive in Band recovery data indicated that Cuba during August and into early waterfowl recovered in Cuba originated in September. Small flocks comprised of Blue- 10 Canadian provinces and 34 states of the winged Teal, Northern Pintail and Northern USA. Overall, c. 84.6% of the 1,842 banded Shoveler containing 20–100 birds are waterfowl recovered in Cuba originated common in coastal wetlands during this from 24 U.S. states and Canadian provinces period. Numbers of these and other species in the eastern and central regions of North gradually increase until they reach a peak in America. The remaining recoveries include a November, then decrease in coastal areas as small number of birds originating in states wintering waterfowl move to interior of the western United States (e.g. California wetlands on the island (Fong et al. 2005; and Idaho). The greatest number of duck Kirkconnell et al. 2005). During February species and individuals recovered in 10 and March, numbers of Blue-winged Teal

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 504 Waterfowl in Cuba 1 Villa 1 21 12 1 12 1 53 10 2 333 2 4 53651 1 113 1 31 1 1 9 21 4 3 1 121 1 1 3 11 1 1 14 31442 7 4 12122 12 1 12 403 290 28 96 52 51 219 128 111 63 83 91 33 26 Number of waterfowl in North America (U.S.A. and Canada) recovered in Cuba the years 1930–2010. Cuban provinces recovered Canada) and Number of in North America (U.S.A. waterfowl Clara (VC), Sancti Spíritus (SSp), Ciego de Ávila (CAv), Camagüey (Cam), Las Tunas (Tun), Holguín (Hol), Granma (Gra), Santiago (Tun), Camagüey (Cam), Las Tunas Sancti Spíritus (SSp), Ciego Clara (VC), de Ávila (CAv), de Cuba (Stgo), Guantánamo (Gmo). Table 2. Table (IJ), Pinar del Río (PR), Habana (H), Ciudad de la (CH), Matanzas (Mtz), Cienfuegos (Cfg), include: Isla de la Juventud SpeciesA. sponsa IJ PR H CH Mtz Cfg VC SSp CAv Cam Tun Hol Gra Stgo Gmo A. crecca A. clypeata A. Americana A. acuta A. collaris D. bicolour D. A. affinis A. cyanoptera A. discors A. strepera

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 Waterfowl in Cuba 505

States and Provinces of origin: 1–5 Cuban provinces 6–10 Cuban provinces 11–15 Cuban provinces

Number of species recovered: 1–2 species 3–4 species 5–7 species

Figure 3. Origin of bands recovered in Cuba (1930–2010) indicating number of Cuban provinces represented and duck species recovered in US states and Canadian provinces. Information provided by the U.S. Geological Survey’s Bird Banding Laboratory and the Canadian Bird Banding Office. and Northern Pintail gradually increase in waterfowl resident in Cuba (e.g. for White- coastal wetlands as birds prepare for cheeked Pintail Anas bahamensis and West northern migration back to their breeding Indian Whistling-duck Dendrocygna arborea; grounds (Blanco & Sánchez 2005). Fig. 4), less is known regarding their nesting Cuba harbours six species of resident ecology and productivity. Yet such data waterfowl, including introduced breeding are important for species conservation, species. Nesting normally occurs from particularly as three resident species are April–October though many species have classified as being of global and regional extended nesting seasons (Garrido & conservation concern (González et al. Kirkconnell 2011). The Fulvous Whistling- 2012; BirdLife International 2013): the duck was introduced to Cuba in 1931 and West Indian Whistling-duck (globally occurs at a limited number of sites on the vulnerable), White-cheeked Pintail (regional island (Garrido & García 1975). The species concern) and Masked Duck Nomonyx nests in flooded forests surrounding some dominicus (regional concern). Moreover, of Cuba’s major reservoirs such as while Cuba harbours the largest numbers Mampostón (Mayabeque province) and of these resident species in the Caribbean Leonero (Granma province). Nests have (e.g. around 10,000 West Indian Whistling- been reported in cavities of various tree ducks), information suggests that Cuban species including the Cuban Royal Palm populations of resident ducks are declining Roystonea regia. Although information is (Acosta & Mugica 2006). available on the presence and distribution of Cuba boasts a vast network of protected

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 506 Waterfowl in Cuba

Figure 4. Location records for White-cheeked Pintail (top, photo: Alberto Puente) and West Indian Whistling-Duck (bottom, photo: Mike Morel) in Cuba and on the Isle of Youth. areas including marine and terrestrial is largely due to the contribution of ecosystems. Approximately 16.9% of the migratory species from North America. terrestrial surface of Cuba (10.98 million ha) While Cuba is much larger in size than is protected by 253 different conservation other Caribbean islands, the diversity of units (CNAP 2009). Within this network of waterfowl present likely reflects the conservation sites, 27 of the protected areas relatively undisturbed condition of most harbour some of the most important wetland ecosystems, including interior as locations for Cuba’s threatened resident well as coastal regions containing extensive waterfowl (Fig. 5). These include: four areas of mangrove forest (Giri et al. 2011). biosphere reserves, five national parks, four Moreover, the proximity and interspersion ecological reserves, seven wildlife refuges of many of the principal wetlands of Cuba and all six Ramsar sites designated within to rice production areas likely benefits not Cuba (CNAP 2009; Aguilar 2010). only resident but also migratory waterfowl. Rice cultivation has long been a Discussion component of Cuban agriculture, and it The broad geographic distribution of currently represents the second most waterfowl across the main island of Cuba important crop (in terms both of the area and their diverse taxonomic representation planted and in yield) after sugarcane

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 Waterfowl in Cuba 507

Figure 5. Important waterfowl areas and the protected areas network in Cuba (CNAP 2009).

Saccharum sp. Areas in rice production total c. weed seeds rather than rice seeds, and their 150,000 ha and are concentrated along the waste adds nutrients to the soil, promoting southern part of the island bordering coastal an energy flow between the rice paddies and wetlands in the provinces of Pinar del Río, the nearby wetlands (Elphick 2000; Mugica et Matanzas, Sancti Spiritus, Camaguey and al. 2006). Ongoing research and outreach by Granma. Following the dissolution of the the avian ecology group of the University of Soviet Union in 1991, Cuba was faced with Havana has helped greatly in changing major economic challenges. The island attitudes of farmers, and has encouraged nation responded with a large-scale food them to manage rice fields to support production programme and experimented biodiversity at these sites (Mugica et al. 2006). with alternatives to industrialised farming Band recoveries suggest that the due to lack of chemical fertilisers and Mississippi and Atlantic Flyways contribute pesticides (FAO 2002). At present, the greatly to the diversity of waterfowl species most extensive rice production regions are in wintering in and migrating through Cuba close proximity to natural wetlands, (Fig. 3). Recent advances in the study of facilitating the waterbirds’ use of the rice migratory strategies for terrestrial birds fields as feeding areas. Further, the general suggest that North American warblers lack of pesticide and herbicide use promotes Parulidae sp. exhibit similar overwintering high levels of vertebrate and invertebrate patterns. For instance, stable isotope analysis biodiversity (Mugica et al. 2006). indicated that warblers wintering in western Consequently, waterbird populations thrive Cuba originate from New England states, in the rice-producing areas of Cuba. while some warblers which winter further Waterbirds are an important biotic east in Cuba and in the rest of the Caribbean component of the rice agro-ecosystem. are derived from southern Appalachians Most waterbirds feed on invertebrates and populations (Faaborg et al. 2010). These new

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 508 Waterfowl in Cuba techniques may be useful to provide further (Erwin et al. 1984). Illegal harvest of insights into the biogeography of waterfowl mangrove for charcoal production continues in Cuba and the rest of the Caribbean, and in remote coastal regions of Cuba. Further, would greatly enhance information derived illegal logging is ongoing in areas of from the banding data. palm forest and swamp forest where Efforts to increase communication resident species such as the West Indian between Cuban waterfowl biologists and Whistling-duck nest. Fires also degrade these banding laboratories in the United States savannas and seasonally flooded forests. and Canada should be expanded (Blanco & In recent years Cuba has experienced Sánchez 2005), not least because waterfowl periods of pronounced droughts resulting in species resident in the southeastern United lower water levels and, consequently, States (e.g. Wood Duck and Fulvous reduced productivity of wetlands (Sims Whistling-duck) may move regularly & Vogelmann 2002). Finally, hurricanes between Cuba and the continent (Turnbull et can impact coastal wetlands due to al. 1989). The links between mainland and storm timing, frequency and intensity, insular populations of these species and the which in turn can alter coastal wetland functional role of Cuban wetlands in their hydrology, geomorphology, biotic structure, annual cycle are still unknown. Quantitative productivity and nutrient cycling (Michener studies on population ecology and habitat et al. 1997). Hurricane impact on waterbirds relationships of breeding resident species highlights the importance of establishing are also considered a priority by Cuban long-term studies for identifying complex biologists, for informing the conservation environmental and ecological interactions and management of waterfowl and wetlands that may otherwise be dismissed as in the region (Acosta & Mugica 2006). stochastic processes (Green et al. 2011). Approximately 1.19 million ha of Cuba is considered a high priority country wetlands are currently protected under for biodiversity conservation within the various conservation categories in Cuba Caribbean basin region, yet it remains largely (ACC-ICGC 1993; Aguilar 2010). Despite ignored by most conservation organisations the extensive network of protected areas in North America, and few long-term and environmental legislation aimed at conservation programmes have been expanding protection of mangrove forest, established by international NGOs. The Cuban wetlands and waterfowl face state of U.S.-Cuba relations should not numerous threats. Although Cuban exclude the island-nation from regional legislation prohibits the harvest of duck conservation programmes (e.g. the Atlantic species classified as threatened, subsistence Coast Joint Venture), given the prominent hunting of resident waterfowl persists across role of Cuba’s wetland resources compared several regions of Cuba. Similarly, illegal to the rest of the Caribbean. International harvest of eggs from threatened waterfowl cooperation with Cuban scientists, species and other waterbirds occurs in Cuba, universities and environmental organisations as it does in other islands of the Caribbean should be expanded if an integrated and

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 Waterfowl in Cuba 509 effective conservation strategy for wetlands structure of meio- and macrofauna in and waterfowl in the Caribbean region is to seagrass meadows and mangroves from NW be achieved (Margulis & Kunz 1984; Santana shelf of Cuba (Gulf of Mexico). Revista de 1991). Investigaciones Marinas 28: 139–150. American Ornithologists’ Union. 2009. 49th Acknowledgements Supplement to the American Ornithologists’ Union’s Check-list of North American birds. B.S.O and F.J.V. dedicate this paper to the Accessible at http://www.aou.org/checklist memory of our dear friend and collaborator, (last accessed on 3 March 2014). Pedro Blanco Rodríguez, who passed away Barrios, O., Blanco, P. & Soriano, R. 2003. before publication of this manuscript. We Nuevos registros de aves en Cayo Sabinal, are grateful to our colleagues at the Institute Camagüey, Cuba. Journal of Caribbean of Ecology and Systematics for support and Ornithology 16: 22–23. assistance during field expeditions. L. Webb Bellrose, F.C. 1976. Duck, Geese and Swans of and J.M. Wunderle provided helpful and North America. Stackpole Books, Harrisburg, constructive comments to an earlier version Pennsylvania, USA. BirdLife International. 2013. Threatened Birds of of the manuscript. Any use of trade, firm or the World. Accessible at http:www.birdlife.org product names is for descriptive purposes (last accessed on 3 March 2014). only and does not imply endorsement by the Blanco, P. 1996. Censos de aves acuáticas en un U.S. Government. humedal costero de la Ciénaga de Zapata, Cuba. Avicennia 4–5: 51–55. References Blanco, P. & Sánchez, B. 2005. Recuperación de ACC-ICGC 1993. Estudio Geográfico Integral. aves migratorias neárticas del orden Ciénaga de Zapata. Publicaciones del Servicio Anseriformes en Cuba. Journal of Caribbean de Información y Traducciones, La Habana, Ornithology 18: 1–6. Cuba. Blanco, P., Zuñiga D., Gómez, R., Socarras, E., Acosta, M. & Mugica, L. 1994. Notas sobre la Suárez M. & Morera, F. 1996. Aves del comunidad de aves del Embalse Leonero, sistema insular Los Cayos de Piedra, Sancti provincia Granma. Ciencias Biológicas 27: Spíritus, Cuba. Oceanides 11: 49–56. 169–171. Borroto, R. & Mancina, C. (eds.). 2011. Mamíferos Acosta, M., Morales, J., González, M. & Mugica, en Cuba. UPC Print, Vaasa, Finland. L. 1992. Dinámica de la comunidad de aves Borhidi, A., Muñoz, O. & Del Risco, E. 1993. de la playa La Tinaja, Ciego de Ávila, Cuba. Plant communities of Cuba. I. Fresh and salt Ciencias Biológicas 24: 44–58. water, swamp and coastal vegetation. Acta Acosta, M. & Mugica, L. 2006. Reporte final de Botanica Hungarica 29: 337–376. aves acuáticas en Cuba. Facultad de Biología, Brooks, T.M., Mittermeier, R.A., da Fonseca, Universidad de La Habana, La Habana, Cuba. G.A.B., Gerlach, J., Hoffman, M., Lamoreux, Aguilar, S. (ed.). 2010. Áreas Importantes para la J.F., Mittermeier, C.G., Pilgrim, J.D. & Conservación de las Aves en Cuba. Editorial Rodrigues, A.S.L. 2006. Global biodiversity Academia, La Habana, Cuba. conservation priorities. Science 313: 58–61. Armenteros, M., Williams, J.P., Hidalgo, G. Centro Nacional de Áreas Protegidas (CNAP). & González-Sansón, G. 2007. Community 2009. Plan del Sistema Nacional de Áreas

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 510 Waterfowl in Cuba

Protegidas 2009–2013. Centro Nacional de Garrido, O.H. & Schwartz, A. 1968. Anfibios, Áreas Protegidas, Havana, Cuba. reptiles y aves de la Península de Elphick, C.S. 2000. Functional equivalency Guanahacabibes, Cuba. Poeyana 53: 1–68. between rice fields and the semi-natural Giri, C., Ochieng, E., Tieszen, L.L., Zhu, Z., wetland habitat. Conservation Biology 14: 181– Singh, A., Loveland, T., Masek, J. & Duke, N. 191. 2011. Status and distribution of mangrove Erwin, R.M., Kushlan, J.A., Luthin, C., Price, I.M. forests of the world using earth observation & Sprunt, A. 1984. Conservation of colonial satellite data. Global Ecology and Biogeography waterbirds in the Caribbean basin: summary 20: 154–159. of a panel discussion. Colonial Waterbirds 7: González, H. (ed.). 2002. Aves de Cuba. UPC 139–142. Print, Vaasa, Finland. Environmental Systems Research Institute González, H., Rodríguez Shettino, L., Rodríguez, (ESRI). 2001. Getting to know ArcGIS A., Mancina, C.A. & Ramos, I. 2012. Libro desktop. Environmental Systems Research Rojo de los Vertebrados de Cuba. Editorial Institute, Inc., Redlands, California, USA. Academia, La Habana, Cuba. Faaborg, J., Holmes, R.T., Anders, A.D., Goossen, J. P., Blanco, P., Sirois, J. & Gonzalez, Bildstein, K.L., Dugger, K.M., Gauthreaux, H. 1994. Waterbird and shorebird count in S.A., Heglund, P., Hobson, K.A., Jahn, A.E., the province of Matanzas, Cuba. Canadian Johnson, D.H., Latta, S.C., Levey, D.T., Wildlife Service Technical Report Series 170: 1–18. Marra, P.P., Merkord, C.L., Nol, E., Green, M.C., Hill, A., Troy, J.R., Holderby, Z. & Rothstein, S.I., Sherry, T.W., Sillet, T.S., Geary B. 2011. Status of breeding reddish Thompson, F. & Warnock, N. 2010. egrets on Great Inagua, Bahamas with Recent advances in understanding migration comments on breeding territoriality and the systems of New World land birds. Ecological effects of hurricanes. Waterbirds 34: 213– Monographs 80: 3–48. 217. Food and Agriculture Organization (FOA). 2002. Kirkconnell, A., Stotz, D.F. & Shoplaand, M. Cuba. FAO Rice Information Bulletin, FAO, (eds.). 2005. Cuba: Peninsula de Zapata. Rome, Italy. Rapid Biological Inventories Report No. 7. The Fong, A., Maceira, D., Alverson, W.S. & Wachter, Field Museum, Chicago, Illinois, USA. T. (eds.). 2005. Cuba: Parque Nacional Llanes, A., Kirkconnell, A., Posada, R.M. & “Alejandro de Humboldt”. Rapid Biological Cubillas, S. 1987. Aves de Cayo Saetía, Inventories Report No. 14. The Field Archipiélago de Camagüey, Cuba. Miscelaneas Museum, Chicago, Illinois, USA. Zoológicas, Instituto Zoologia, Academia de Ciencias Garrido, O.H. 1980. Los vertebrados terrestres de Cuba 35: 3–4. de la Península de Zapata. Poeyana 203: 1–49. Lugo, A.E. 2002. Conserving Latin American and Garrido, O.H. & García, F. 1975. Catálogo de las Aves Caribbean mangroves: issues and challenges. de Cuba. Editorial Academia, La Habana, Cuba. Madera y Bosques 8 (Supplement): 5–25. Garrido, O.H. & Kirkconnell, A. 2000. Birds of Margulis, L. & Kunz, T.H. 1984. Glimpses of Cuba (Helm Field Guides). Christopher Helm biological research and education in Cuba. Publishers, London, UK. BioScience 34: 634–639. Garrido, O.H. & Kirkconnell, A. 2011. Aves de Melián, L.O. 2000. Inventario de las aves en Cuba. Comstock Publishing Associates, zonas húmedas de San Miguel de Parada. Ithaca, New York, USA. Biodiversidad de Cuba Oriental IV: 90–93.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 Waterfowl in Cuba 511

Michener, W.K., Blood, E.R., Bildstein, K.L., Colorados, Pinar del Río, Cuba. Inv. Mar. Brinson, M.M. & Gardner, L.R. 1997. Cicimar 6: 247–249. Climate change, hurricanes and tropical Sánchez, B., Blanco, P., Oviedo, R., Hernández, storms, and rising sea level in coastal A., del Pozo, P., Lamela, W., Torres, M. & wetlands. Ecological Applications 7: 770–801. Rodríguez, R. 2008. Composición y Morales, J. & Garrido, O.H. 1996. Aves y reptiles abundancia de las comunidades de aves en de cayo Sabinal, Archipiélago Sabana localidades de la provincia de Cienfuegos, Camagüey, Cuba. El Pitirre 9: 9–11. Cuba. Poeyana 496: 10–19. Mugica, L. 2000. Estructura espacio temporal y Santana, E.C. 1991. Nature conservation and relaciones energéticas en la comunidad de sustainable development in Cuba. Conservation aves de la arrocera Sur del Jíbaro, Sancti Biology 5: 13–17. Spíritus, Cuba. Ph.D. thesis, Universidad de Scott, D.A. & Carbonell M. 1986. A Directory La Habana, La Habana, Cuba. of Neotropical Wetlands. International Union Mugica, L., Acosta, M. & Dennis, D. 2001. for Conservation of Nature and Natural Dinámica temporal de la comunidad de aves Resources (IUCN), Conservation Monitoring asociada a la arrocera Sur del Jíbaro. Biología Centre. Cambridge, U.K. 15: 86–97. Shi, H., Singh, A., Kant, S., Zhu, Z. & Waller, E. Mugica, L., Acosta, M., Denis, D., Jiménez, A., 2005. Integrating habitat status, human Rodríguez, A. & Ruiz, X. 2006. Rice culture population pressure, and protection status in Cuba as an important wintering site for into biodiversity conservation priority migrant waterbirds from North America. In setting. Conservation Biology 19: 1273–1285. G.C. Boere, C.A. Galbraith & D.A. Stroud Sims, H. & Vogelmann, K. 2002. Popular (eds.), Waterbirds Around the World, pp. 172– mobilization and disaster management in 176. The Stationery Office, Edinburgh, U.K. Cuba. Public Administration and Development, Peña, C., Fernández, A., Navarro, N., Reyes, E. & 22: 389–400. Sigarreta, S. 2000. Avifauna asociada al sector Torres, A. & Solana, E. 1994. Listado de las aves costero de Playa Corinthia, Holguín, Cuba. observadas dentro del corredor migratorio El Pitirre 13: 31–34. de Gibara, provincia Holguín; Cuba. Garciana Raffaele, H., Wiley, J., Garrido, O.H., Keith, A. & 22: 1–4. Raffaele, J. 1998. Birds of the West Indies. Turnbull, R.E., Johnson, F.A. & Brakhage, D.H. Princeton University Press, Princeton, New 1989. Status, distribution and foods of Jersey, USA. Fulvous Whistling-Ducks in south Florida. Rodríguez, F. 2000. Lista de los vertebrados del Journal of Wildlife Management 53: 1046– Parque Natural bahía de Naranjo, Cuba. 1051. Biodiversidad de Cuba Oriental V: 143–146. Wiley, J.W., Ruiz, A., Pérez, E., Faife, M., Díaz, L., Rodríguez Shettino, L. (ed.). 2003. Anfibios y González, M., Rivero, Y., Chirino, G., Soto, Reptiles de Cuba. UPC Print, Vaasa, Finland. O., Morejón, R., Vales, A. & Ibarra, M.E. Sánchez, B. & Rodríguez, D. 2000. Avifauna 2002. Bird survey in the mogote vegetational associated with the aquatic and coastal complex in the Sierra del Infierno, Pinar del ecosystems of Cayo Coco, Cuba. El Pitirre Río, Cuba. El Pitirre 15: 7–15. 13: 68–75. Woods, C.A. 1989. Biogeography of the West Sánchez, B., García, M.E. & Rodríguez, D. 1991. Indies past, present and future. Sandhill Aves de Cayo Levisa, Archipiélago de los Crane Press, Gainesville, Florida, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 498–511 512

Assessment of the 6th North American Duck Symposium (2013): “Ecology and Conservation of North American Waterfowl” LUCIEN P. LABORDE, JR.1*, RICHARD M. KAMINSKI2 & J. BRIAN DAVIS2

1School of Renewable Natural Resources, Louisiana State University, Baton Rouge, Louisiana 70803, USA. 2Department of Wildlife, Fisheries and Aquaculture, Mississippi State University, Mississippi State, Mississippi 39762, USA. *Correspondence author. E-mail: [email protected] Abstract Using web-based technology, we e-mailed a survey link to all 450 conferees of the 6th North American Duck Symposium (NADS 6), “Ecology and Conservation of North American Waterfowl” (ECNAW), seeking feedback from attendees in order to guide the organisation of future waterfowl and other wildlife symposia. Twelve questions were posed to evaluate the 2013 all-waterfowl symposium and a further 18 questions to assist planning future similar meetings. A total of 284 responses (63%) were received; the feedback suggested that NADS 6 was well organised, that it presented relevant information and that it was valuable to conferees. Perceptions of respondents on the structure of NADS 6 (i.e. whether presentation on all waterfowl should be included, or only on ducks) may not be representative of attendees of previous NADS meetings, as these focused on duck species. Nevertheless, respondents suggested that future symposia should continue on a 3-year rotation and retain its 4-day format with four morning plenary sessions, concurrent afternoon oral presentations and evening poster and mentee-mentor sessions. Respondents also recommended a maximum of three concurrent afternoon sessions and indicated that future symposia might continue to embrace geese and swans as well as ducks. The results suggest a need for officials of NADS to determine future meeting frameworks and venues. Web-based surveys provide a useful tool for conference evaluation and can promote effective design and relevance of future meetings and related events.

Key words: conference evaluation, ducks, NADS, survey, symposium, waterfowl.

Waterfowl (Anatidae: ducks, geese and (Baldassarre & Bolen 2006; Grado et al. swans) are important birds ecologically, 2011; Green & Elmberg 2014). They environmentally and economically have been foci of continual research,

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 Assessment of the 6th North American Duck Symposium 513 conservation and recreational endeavours in ultimately through the creation of NADS, North America since the early 20th century Inc., a non-profit organisation established to (Bellrose 1976; Baldassarre & Bolen 2006). facilitate future symposia and workshops. To help sustain waterfowl resources in Six NADS have been convened to date, in: North America, scientists and managers 1) Baton Rouge, Louisiana, USA (1997); have convened conferences and symposia 2) Saskatoon, Saskatchewan, Canada (2000); periodically to communicate contemporary 3) Sacramento, California, USA (2003); knowledge about these birds and their 4) Bismarck, North Dakota, USA (2006); habitats, particularly for species and 5) Toronto, Ontario, Canada (2009); and populations of conservation concern (e.g. 6) Memphis, Tennessee, USA (2013). Canadian Wildlife Service 1969; Bookhout Locations generally have rotated between 1979; Boyd 1983; Whitman & Meredith the United States and Canada, among 1987; Weller 1988; Smith et al. 1989; administrative waterfowl flyways, and Fredrickson et al. 1990; Batt et al. 1992; generally in northern and southern locations Rusch et al. 1998). The inaugural North of North America. Each NADS was American Duck Symposium and Workshop organised under the direction of a scientific (NADS 1) was convened in Baton Rouge, committee, which had discretion regarding Louisiana, USA in 1997. This seminal event the theme, content and venue. The Science attracted professionals and students from Programme Committee for NADS 6 agreed North America and Europe to address the the symposium would be expanded to include ecology and management of wild ducks, to all taxa of waterfowl. Thus, NADS 6 was sub- synthesise acquired knowledge and to titled “Ecology and Conservation of North convey future needs and directions for American Waterfowl” (ECNAW), with the research, management and conservation. North American Arctic Goose Conference An important objective of NADS from and Workshop and the International inception has been to attract students to Sea Duck Conference included as joint present their research and promote their partners in NADS 6/ECNAW (http:// professionalism among colleagues. The www.northamericanducksymposium.org). founders of NADS believed that addressing Feedback and evaluation are crucial for research questions and management issues improving wildlife science and conservation related to sustaining duck populations, programmes, and also for stakeholder maintaining the wildfowling tradition and engagement (Sholtes 1988; Jacobson 2012; advancing ecological studies, as well as the Lauber et al. 2012). Although five NADS involvement of the next generation of have been held previously, none were students, were of paramount importance. evaluated by surveying the conferees. Additionally, the founders considered that Following NADS 6/ECNAW, its lead waterfowl ecologists had led major advances organisers (R.M. Kaminski and J.B. Davis) in avian ecology, analytical procedures and decided to conduct a post-symposium conservation, and they therefore sought to assessment of the meeting and asked the perpetuate and develop this legacy, senior author (L.P. Laborde, Jr.), with

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 514 Assessment of the 6th North American Duck Symposium human-dimensions and survey-sampling limited to one per Internet Protocol (IP) skills, to develop a questionnaire. The address to minimise poll crashing (i.e. primary objective was to poll participants on multiple responses per attendee; Dillman et al. evaluative and planning criteria for future 2009). Survey protocols ensured anonymity NADS and similar meetings. We believe and confidentiality and were approved by the results from this survey may also benefit Louisiana State University Agriculture Center wildlife and natural resources professionals Institutional Review Board (Protocol in planning and implementing other large Number HE 13-7). We collected responses conferences and symposia. through to 21 March 2013. We calculated the margin of error as the 95% confidence Methods interval for the true population value of The NADS 6/ECNAW post-symposium responses, following Dillman et al. (2009). survey was developed to address 12 Chi-square tests (α = 0.05) were used to evaluative, 18 planning and three assess non-response bias and to test demographic-related questions. We evaluated frequencies of response among demographic conference sessions using an ordinal scale classes. Simple descriptive statistics are of 1 = “not valuable”, 2 = “marginally presented to analyse evaluative and planning valuable”, 3 = “moderately valuable”, and questions. 4 = “highly valuable”, and likewise used an ordinal scale of 1 = “strongly disagree”, Results 2 = “disagree”, 3 = “neither agree nor We received 284 (63%) responses to the disagree”, 4 = “agree”, and 5 = “strongly survey. Based on this response rate and the agree” to rank agreement with statements population of 450 conferees, we report addressing the symposium venue, scheduling, results within a margin of error of ± 4%, programme and costs (Dillman et al. 2009). indicating that 19 out of 20 times (i.e. 95% Additionally, we invited open comments of occasions) the true population value will from conferees. Confirmed e-mail addresses be within 4 percentage points of our were obtained from all 450 participants and reported sample estimate. We used three the survey was distributed on 21 February demographic variables to evaluate non- 2013, three weeks after the symposium. Each response bias. Respondents were 85% male conferee was asked to complete the survey, (15% female), but gender proportions of and each e-mail contained an embedded link respondents did not differ significantly from χ2 to the survey using Qualtrics™ v. 12000 non-respondents ( 1 = 2.82, P = 0.093, (Qualtrics Labs, Inc., Provo, Utah, USA; n.s.). Age distribution of respondents was Vaske 2008). We contacted conferees up to ≤ 25 (11%), 26–35 (30%), 36–45 (24%), three times at 5-day intervals to elicit their 46–55 (19%), 56–65 (13%), and > 65 years response and then thanked all respondents. (3%), but ages of non-respondents were not An alternative response system – via available for comparison. By occupation, electronic document, post or e-mail – was 23% of respondents were students, 15% also provided (Vaske 2008). Responses were were academicians and 62% were grouped

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 Assessment of the 6th North American Duck Symposium 515 as professionals, including biologists, presented at the symposium, logistics, the managers, administrators and retirees. A host hotel and nearby venues. Mean ratings higher proportion of the academicians who ranged from 3.3–4.4 (s.d. = 0.7–1.1), attended the meeting responded to the indicating their assessment ranged from survey (89%) than did students (62%) or moderate to strong agreement (Table 1). χ2 professionals (57%; 2 = 11.96, P = 0.002). Eighteen questions addressed respondents’ Respondents evaluated each of five preferences for future NADS, of which five conference sessions separately. The four daily specifically addressed the format of future plenary sessions of the symposium were symposia. For NADS 7 (scheduled to be attended by ≥ 88% of respondents, and their held in Annapolis, Maryland, USA; February ratings averaged 3.2–3.3 (s.d. = 1.1–1.3), 2016), morning plenary and afternoon oral indicating that their assessment of plenaries presentations, of the same length as NADS ranged from moderate to high value. One day 6, were favoured by 63% and 77% of before the grand opening of the symposium, respondents respectively. During NADS 6, there was a special session on the 2012 6–7 concurrent sessions were held during revision of the North American Waterfowl three afternoons, but feedback indicated that Management Plan; it was attended by 28% of this was too many, with 58% of respondents the respondents who arrived early to the preferring only 2–3 concurrent sessions, and meeting and rated it, on average, moderately 32% suggesting 4–5 concurrent sessions. valuable (mean = 3.1, s.d. = 1.5). The Only 2% wanted to continue the NADS 6 remaining seven evaluative statements format of 6–7 sessions being held at the considered the relevancy of information same time. Given options for convening

Table 1. Level of agreement with statements evaluating the North American Duck Symposium and Workshop 6, Ecology and Conservation of North American Waterfowl (2013), held at Memphis, Tennessee, USA.

Statement Meana s.d. n

Information presented was directly relevant to my work 4.4 0.7 281 The registration cost was a fair value 4.1 0.8 283 The symposium was well organised 4.2 0.9 280 Adequate time was allowed for breaks 4.2 0.8 280 Adequate time was devoted to issues of waterfowl management 4.0 0.8 281 The hotel cost was fair relative to amenities received 3.3 1.1 279 Service from the hotel staff was excellent 4.3 0.8 280 aRated on a scale of 1 = strongly disagree, to 5 = strongly agree.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 516 Assessment of the 6th North American Duck Symposium more frequent symposia, 66% of (Table 2). Responses to eight statements respondents preferred the current format of ranged from neutral to agreeable (means = a 4-day symposium every three years. Given 3.1–4.0, s.d. = 0.7–1.1; Table 2), indicating options of integrated plenary topics in a 4- that evenings were preferred for poster day or two consecutive 2-day formats with sessions, that the student mentee-mentor different registration options and costs, 73% session should be continued, door prizes preferred the current format of 4 days with should be given to students and four different daily plenary sessions and a professionals during breaks between single registration fee. Cross-tabulation of sessions, and that the cost of public the above five questions about the format of transport between airports and hotels future symposia confirmed that ≥ 50% of should be considered on choosing the hotel. the three major occupational groups (i.e. Respondents neither agreed nor disagreed students, academicians and professionals) that breakfasts should be provided as part of ranked alternatives identically as described registration costs, and there was no above. consensus that speakers and entertainment Eleven statements addressing the during lunch enhanced the symposium. symposium venue, scheduling, programme Although nearly all respondents did not and costs were rated as described previously wish to continue the NADS 6 format of

Table 2. Level of agreement with statements for planning the North American Duck Symposium and Workshop 7 (scheduled to be held at Annapolis, Maryland, USA in February, 2016).

Statement Meana s.d. n

Professional and student presentations should be intermingled 4.0 0.9 280 Evening receptions were a good time for poster sessions 3.9 0.8 279 The Student Mentor-Mentee session should be continued 3.9 0.9 273 Door prizes for students should be continued 4.0 0.8 278 Door prizes should be availed to all conferees 3.1 1.0 279 Speakers and entertainment during lunch enhanced the programme 2.8 1.1 279 Breakfast should not be included to reduce registration costs 2.3 1.1 280 Alcoholic beverages should be included during evening receptions 3.4 1.1 281 Restaurants and amenities should be available within walking distance 4.3 0.7 281 The conference should be within a $30 cab ride of an airport 3.7 0.9 279 Future symposiums should address “ducks only” and not all waterfowl 2.1 1.1 282 aRated on a scale of 1 = strongly disagree, to 5 = strongly agree.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 Assessment of the 6th North American Duck Symposium 517

6–7 concurrent afternoon sessions, which respondents representing age classes from was necessary to accommodate many ≤ 25 to > 65 years, and there being no presenters at the all-waterfowl symposium, significant difference in the gender of 73% of respondents “strongly disagree” responding and non-responding conferees, or “disagree” that future NADS should we believe that the responses were address ducks only. reasonably representative of those attending Two questions addressed participation in the conference. A representative sample NADS 7. A total of 172 respondents (65%) from a majority of the surveyed population indicated they were “likely” or “very likely” is considered more relevant than a high to attend NADS 7, and 35 respondents response rate for generalisations from (12%) volunteered to serve on an organising survey results (Vaske 2008). committee for NADS 7. One hundred and Survey results suggested that NADS two respondents (36%) offered comments, 6/ECNAW was well organised, that it of which 47 were congratulatory in nature. presented relevant content and was valuable Twenty-nine comments indicated there were to conferees. Responses and comments too many concurrent afternoon sessions, suggested that the host hotel rates should and 26 comments stated that the daily cost fall within federal and other expense of the host hotel exceeded federal and some guidelines, and that the location of the state expenses limits. For additional details, host hotel should be within walking the complete survey and its summarised distance of restaurants and other amenities. results and comments are available on Respondents also suggested that future the NADS 6/ECNAW website (http:// NADS should continue on a 3-year rotation www.northamericanducksymposium.org/ and retain its single registration, 4-day index.cfm?page=survey) or from the senior format with 4 morning plenary sessions. author. There was an overwhelming preference to reduce the number of concurrent sessions Discussion in future NADS, likely because previous We surveyed conferees of NADS 6/ECNAW, NADS featured non-concurrent sessions a symposium that embraced all waterfowl enabling possible holistic attendance of (i.e. ducks, geese and swans), unlike previous sessions by conferees. Nonetheless, > 70% NADS which focused on ducks alone. The of respondents expressed the preference results and interpretations therefore reflect that future symposia embrace all taxa of data from respondents attending this all- waterfowl. Because this opinion reflected waterfowl conference and may not be the perceptions only of respondents to the perception of attendees of NADS 1–5. NADS 6/ECNAW survey, we conclude Nonetheless, the results likely will be useful there is a need for the scientific committee for planning future NADS and similar large of future NADS to work with NADS, Inc. meetings. Because the overall response rate to determine if subsequent NADS should was > 60%, with > 55% of conferees in focus on ducks or be inclusive of all each of the occupation classes responding, waterfowl taxa. Indeed, there are major

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 518 Assessment of the 6th North American Duck Symposium trade-offs between holding an all-waterfowl comparable and not confounded by survey symposium with the number of concurrent methodology. Additionally, we recommend sessions in large meetings, during which that other wildlife and natural resources conferees would be unable to attend all conferences and symposia conduct similar sessions, presenters (notably students) post-meeting surveys. When combined with would not have an opportunity to address early programme and hotel planning, and most conferees (if speaking in one of with the involvement of experienced several concurrent sessions), and there committee volunteers and fundraisers from may also be competition for attendance previous symposia, post-conference surveys and fund-raising between NADS and can promote the effective design and other waterfowl, ornithological or wildlife relevance of future events. conferences. Multiple concurrent sessions at all-waterfowl symposia may thus lessen Acknowledgements opportunities for students to gain knowledge We are indebted to the public- and private- and receive expert feedback on their work, sector sponsors of NADS 6/ECNAW. We which has been identified as an important also thank the scientists and managers objective for NADS by NADS, Inc. who composed the Scientific Program Access to web-based survey tools and Committee and the Graduate Student to the conferees’ e-mail addresses make Scholarship and Awards Committee, the electronic post-symposia surveys an graduate students at Mississippi State inexpensive and relatively efficient method University, especially Justyn Foth who helped for evaluating meetings and planning future to ensure success of the symposium, the events. While we did not incur any direct session chairs and moderators, and the staff costs to administer the survey, future of the Peabody Hotel, Memphis, Tennessee, surveyors may experience charges for USA. We thank all conferees of NADS/ development and use of a survey ECNAW, especially those who completed the instrument. The NADS 6 post-symposia post-symposia survey and enabled this survey implied the relevance and value of evaluation. We thank A. Afton, D. Haukos, symposia presentations, the frequency A.D. Fox and E.C. Rees for reviewing an and possible format of future symposia, earlier draft of this paper. Finally, we thank and general guidelines for the location and the School of Renewable Natural Resources cost of host hotels. The survey was also at Louisiana State University for supporting able to identify volunteers for organising design, implementation and evaluation of the committees of the next symposium. survey, and the Forest and Wildlife Research As far as we are aware, the survey of the Center (FWRC), Mississippi State University, NADS 6/ECNAW conferees was the first for supporting J.B. Davis and R.M. Kaminski formal evaluation of an international as symposia co-chairs. This manuscript has waterfowl conference. We recommend using been approved for publication by the FWRC similar electronic survey methods for and the Department of Wildlife, Fisheries, evaluating future NADS so that the data are and Aquaculture (WFA) at Mississippi

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 Assessment of the 6th North American Duck Symposium 519

State University as FWRC/WFA publication Green, A.J. & Elmberg, J. 2014. Ecosystem WF379. services provided by waterbirds. Biological Reviews 89: 105–122. References Jacobson, S.K. 2012. Communications and Baldassarre, G.A. & Bolen, E.G. 2006. Waterfowl Outreach. In N. J. Silva (ed.), The Wildlife Ecology and Management, 2nd Edition. Krieger Techniques Manual, Volume 2, 7th Edition, pp. Publishing Company, Malabar, Florida, USA. 2–42. The Johns Hopkins University Press, Batt, B.D.J., Afton, A.D., Anderson, M.G., Ankney, Baltimore, Maryland, USA. C.D., Johnson, D.H., Kadlec, J.A. & Krapu, Lauber, T.B., Decker, D.J., Leong, K.M., Chase, G.L. (eds.) 1992. Ecology and Management of L.C. & Schusler, T.M. 2012. Stakeholder Breeding Waterfowl. University of Minnesota engagement in wildlife management. In Press, Minneapolis, Minnesota, USA. D.J. Decker, S.J. Riley & W.F. Siemer (eds.), Bellrose, F.C. 1976. Ducks, Geese and Swans of Human Dimensions of Wildlife Management, pp. North America, 2nd Edition. Stackpole Books, 139–156. The Johns Hopkins University Harrisburg, Pennsylvania, USA. Press, Baltimore, Maryland, USA. Bookhout, T.A. (ed.) 1979. Waterfowl and Wetlands Rusch, D.H., Samuel, M.D., Humburg, D.D. & – an Integrated Review. Proceedings of the 39th Sullivan, B.D. 1998. Biology and Management Mid-west Fish and Wildlife Conference, of Canada Geese. Proceedings of the Madison, Wisconsin. LaCrosse Printing International Canada Goose Symposium, Company, LaCrosse, Wisconsin, USA. Milwaukee, Wisconsin, USA. Boyd, H. (ed.) 1983. First western hemisphere Scholtes, P.R. 1988. The Team Handbook: How to use waterfowl and waterbird symposium. Teams to Improve Quality. Joiner Associates Inc. Canadian Wildlife Service, Catalogue No. Madison, Wisconsin, USA. CW66-63/1983E. Canadian Wildlife Service, Smith, L.M., Pederson, R.L. & Kaminski, R.M. Ottawa, Ontario, Canada. 1989. Habitat Management for Migrating and Canadian Wildlife Service. 1969. Saskatoon Wintering Waterfowl in North America. Texas wetlands seminar. Canadian Wildlife Service Tech University Press, Lubbock, Texas, USA. Report Series, No. 6. Canadian Wildlife Vaske, J.J. 2008. Survey Research and Analysis: Service, Ottawa, Ontario, Canada. Applications in Parks, Recreation and Human Dillman, D.A., Smyth, J.D. & Christian, L.M. Dimensions. Venture Publishing, Inc., State 2009. Internet, Mail, and Mixed-mode Surveys: The College, Pennsylvania, USA. Tailored-design Method, 3rd Edition. John Wiley Weller, M.W. (ed.) 1988. Waterfowl in Winter. & Sons, Inc., Hoboken, New Jersey, USA. University of Minnesota Press, Minneapolis, Fredrickson, L.H., Burger, G.V., Havera, S.P., Minnesota, USA. Graber, D.A., Kirby, R.E. & Taylor, T.S. (eds.) Whitman, W.R. & Meredith, W.H. (eds.) 1987. 1990. Proceedings of the 1988 North American Waterfowl and Wetlands Symposium: Proceedings Wood Duck Symposium. St. Louis, Missouri, of a Symposium on Waterfowl and Wetlands USA. Management in the Coastal Zone of the Atlantic Grado, S.C., Hunt, K.M., Hutt, C.P., Santos, X. & Flyway. Delaware Department of Natural Kaminski, R.M. 2011. Economic impacts of Resources and Environmental Control waterfowl hunting in Mississippi derived Document No. 40-05/87/07/01. Delaware from a state-based mail survey. Human Coastal Management Program, Dover, Dimensions of Wildlife 16: 100–113. Delaware, USA.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 512–519 520 Flushing Mallard (and a likely male Mallard x American Black Duck hybrid) in an interspersed bottomland-hardwood and moist-soil in an interspersed bottomland-hardwood hybrid) Duck male Mallard x American Black Flushing Mallard (and a likely Photograph: Photograph: of in the Mississippi Alluvial Valley Kennedy. wetland C. James Mississippi, by

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 520 521

Photograph: Prairie Pothole Region of South Dakota depicting the modification of wetland basins for agricultural production, by Tim McCabe/USDA Natural Resources Conservation Service.

Photograph: Greater White-fronted and Lesser Snow Geese in a Louisiana rice field, by John K. Saichuk/Louisiana State University AgCenter.

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 521 522

WWT wetland centres WWT centres offer a unique visitor experience where everyone can enjoy wetland habitats and their wildlife all year round in accessible and comfortable surroundings.

WWT Slimbridge (Includes WWT Head Office) Slimbridge Gloucestershire GL2 7BT United Kingdom T +44 (0)1453 891900 F +44 (0)1453 890827 E [email protected] [email protected] http : // www.wwt.org.uk

WWT Martin Mere WWT Arundel Burscough, Ormskirk, Mill Road, Arundel, Lancashire L40 0TA West Sussex BN18 9PB T +44 (0)1704 895181 T +44 (0)1903 883355 WWT National WWT Caerlaverock Wetland Centre Wales Eastpark Farm, Caerlaverock, Llwynhendy, Llanelli, Dumfriesshire DG1 4RS Carmarthenshire SA14 9SH T +44 (0)1387 770200 T +44 (0)1554 741087 WWT Castle Espie WWT Washington Ballydrain Road, Comber, Pattinson, Washington, County Down BT23 6EA Tyne & Wear NE38 8LE T +44 (0)28 9187 4146 T +44 (0)191 416 5454 WWT London Wetland Centre WWT Welney Queen Elizabeth’s Walk, Barnes, Hundred Foot Bank, Welney, London SW13 9WT Nr Wisbech, Norfolk PE14 9TN T +44 (0)20 8409 4400 T +44 (0)1353 860711

©Wildfowl & Wetlands Trust Wildfowl (2014) Special Issue 4: 522