Response of Wild Pollinator Assemblages to Management of Restored Wetlands in Central New York, USA

SUNY College of Environmental Science and Forestry Digital Commons @ ESF

Dissertations and Theses

Spring 4-28-2021

Molly Jacobson SUNY College of Environmental Science and Forestry, [email protected]

Follow this and additional works at: https://digitalcommons.esf.edu/etds

Part of the Natural Resources Management and Policy Commons

Recommended Citation Jacobson, Molly, "Response of Wild Pollinator Assemblages to Management of Restored Wetlands in Central New York, USA" (2021). Dissertations and Theses. 224. https://digitalcommons.esf.edu/etds/224

This Open Access Thesis is brought to you for free and open access by Digital Commons @ ESF. It has been accepted for inclusion in Dissertations and Theses by an authorized administrator of Digital Commons @ ESF. For more information, please contact [email protected], [email protected]. RESPONSE OF WILD POLLINATOR ASSEMBLAGES TO MANAGEMENT OF

RESTORED WETLANDS IN CENTRAL NEW YORK, USA

Molly M. Jacobson

A thesis submitted in partial fulfillment of the requirements for the Master of Science Degree State University of New York College of Environmental Science and Forestry Syracuse, New York April 2021

Department of Environmental Biology

Approved by: Melissa K. Fierke, co-Major Professor, Department Chair Michael L. Schummer, co-Major Professor Margaret Bryant, Chair, Examining Committee S. Scott Shannon, Dean, The Graduate School

ACKNOWLEDGEMENTS

I would like to give special thanks to my colleagues and acquaintances who made this research possible. Firstly, to Dr. Melissa Fierke for providing entomological expertise and the lab space and equipment to process and identify my thousands of specimens, and to Dr. Michael

Schummer for his guidance on developing the methodology of the project, the revision process, and acquainting me with the ecology of the MWC. Thank you to Dr. Donald Leopold for imparting some of his vast botanical wisdom, and to Dr. Hyatt Green for his exceptional support and guidance in the statistical analysis process. Thank you to Frank Morlock, Linda Ziemba,

Mark Benjamin, and Kristen Moore for their partnership in coordinating and informing sampling and outreach efforts. A huge thanks to my field techs – Scott Kostka, Nate Carlton, Dylan

Mahalko, Jake Chronister, Gianna Losquadro, Brittnie Fleming, and Shianne Lindsay, without whom data collection would not have been possible. In addition, I would like to thank the NY

Department of Environmental Conservation, Friends of Montezuma, Seneca Meadows Inc., Edna

Bailey Sussman Foundation, Maurice M. & Annette B. Alexander Wetland Research Award,

Leroy C. Stegeman Award in Invertebrate Ecology, Ducks Unlimited, Cargill Inc., Birds Canada,

Eaton Birding Society, and Sigma Xi Scientific Research Honor Society for their support and funding of this research. Thank you to the Danforth Lab at Cornell University for their assistance in validating identifications, and to Sam Droege for his guidance and insights as a master of melittofauna. Thank you to Norma Jean Young, who graciously tolerated me pinning bees on her kitchen table, Anna Brunner for building the study site map, and to Alexander Siddig for his generous moral support. Lastly, but most of all, thank you to my loving parents for being with me every step of the way. Your lifelong encouragement is what got me this far, and I could never repay that kindness enough. You are all the bees’ knees.

iii

TABLE OF CONTENTS

LIST OF TABLES ...... vi LIST OF FIGURES ...... vii LIST OF APPENDICES ...... viii ABSTRACT ...... ix CHAPTER 1: Literature Review and Project Summary...... 1 Importance and Ecology of Pollinating Insects...... 2 Drivers of Pollinator Declines ...... 4 Habitat Conservation for Pollinators ...... 8 Wetland Restoration and Management ...... 10 Potential Importance of Wetlands for Pollinators ...... 13 Project Summary and Objectives ...... 14 CHAPTER 2: Response of Wild Bee Assemblages to Management of Restored Wetlands in Central New York, USA ...... 15 ABSTRACT ...... 16 INTRODUCTION ...... 17 METHODS...... 20 Study Area...... 20 Vegetation Sampling ...... 22 Pollinator Sampling ...... 23 Specimen Preparation ...... 25 Statistical Analysis ...... 26 RESULTS...... 30 Wild Bee Assemblages ...... 30 Plant-pollinator Interactions ...... 31 Vegetative & Environmental Characteristics of Drawdowns ...... 33 Response of Wild Bees and Floral Resources to Wetland Management ...... 34 DISCUSSION ...... 36 Wild Bee Diversity & Ecology at Restored Wetlands...... 36 Species & Patterns of Note ...... 37 Floral Resources...... 41 Influence of Management on Wild Bees and Floral Resources ...... 45

Efficacy of Drawdown Treatments ...... 45 Response of Plant and Bee Assemblages to Treatments and the Wetland Environment .. 47 FIGURES AND TABLES ...... 52 CHAPTER 3: Diversity and Plant-Pollinator Interactions of Flower Flies (Diptera: Syrphidae) in Restored Wetlands in Central New York, USA ...... 87 ABSTRACT ...... 88 INTRODUCTION ...... 89 METHODS...... 91 RESULTS & DISCUSSION ...... 93 TABLES AND FIGURES ...... 98 CHAPTER 4: Evaluation of Research Objectives and Broad Implications ...... 104 Evaluation of Research Objectives ...... 105 Broad Implications ...... 109 Recommendations for Future Study ...... 111 LITERATURE CITED ...... 114 APPENDICES ...... 132 CURRICULUM VITA ...... 135

LIST OF TABLES

Table 2.1. Study sites and treatments ...... 53 Table 2.2. Pearson’s Correlation coefficients for predictor variables used in mixed models...... 55 Table 2.3. List of bee species collected in restored wetlands in central New York in 2019 – 2020 ...... 56 Table 2.4. Bee families collected by pan-trapping and sweep-netting methods...... 60 Table 2.5. Bee genera collected by pan-trapping and sweep-netting methods...... 60 Table 2.6. Most common bee species per treatment ...... 61 Table 2.7. Bee species collected in sweeps of flowers in restored wetlands, with species-level plant-pollinator interaction statistics ...... 65 Table 2.8. Flowering plant species swept in restored wetlands ...... 68 Table 2.9. Most commonly encountered entomophilous plant species per treatment ...... 75 Table 2.10. Model estimates, standard error, and 95% confidence intervals for the top-ranked model(s) for each parameter of interest...... 77 Table 2.11. Model rankings for predictors of plant species richness in central New York wetlands...... 79 Table 2.12. Model rankings for predictors of entomophilous plant species richness in central New York wetands...... 81 Table 2.13. Model rankings for predictors of bee species richness in central New York wetlands...... 83 Table 2.14. Model rankings for predictors entomophilous plant frequency of occurrence in central New York wetlands...... 85 Table 3.1. Syrphid fly species collected in restored wetlands in central New York in 2019 – 2020 ...... 98 Table 3.2. Plant-pollinator interactions between syrphid flies and flowering plant species ...... 101

LIST OF FIGURES

Figure 2.1. GIS map of study sites ...... 52 Figure 2.2. An example of a pan trap setup used in bee surveys ...... 54 Figure 2.3. Most common bee species collected in restored wetlands in central New York ...... 62 Figure 2.4. Venn diagram portraying species unique to treatments and shared among treatments...... 63 Figure 2.5. Sample-based species accumulation curve for each treatment and for the study system as a whole...... 64 Figure 2.6. Flight season of bees collected in restored wetlands in central New York ...... 71 Figure 2.7. Bloom times of flowering plants from study sites, derived from observational data and sweeps from 2019 – 2020...... 72 Figure 2.8. Plant-pollinator interaction network for bee and flowering plant species in central New York wetlands...... 73 Figure 2.9. Most common plant taxa in restored wetlands in central New York...... 74 Figure 2.10. Model predictions for differences in environmental characteristics among treatments ...... 76 Figure 2.11. Regression for fixed effect of invasive graminoid cover on plant species richness with open water held at median value...... 80 Figure 2.12. Regression for fixed effect of open water on plant species richness with invasive cover held at median value ...... 80 Figure 2.13. Regression for fixed effect of monotypic cattail on entomophilous plant species richness with open water held at median value...... 82 Figure 2.14. Regression for fixed effect of open water on entomophilous plant species richness with monotypic cattail cover held at median value...... 82 Figure 2.15. Regression for fixed effect of open water on bee species richness ...... 84 Figure 2.16. Regression for fixed effect of invasive graminoids on entomophilous plant frequency with open water held at median value ...... 86 Figure 2.17. Regression for fixed effect of open water on entomophilous plant frequency with invasive cover held at median value ...... 86 Figure 3.1. Interaction network illustrating plant-pollinator relationships between syrphid flies and flowers in restored wetlands ...... 100 Figure 3.2. Parhelophilus divisus recorded at Howland Island, NMWMA ...... 103 Figure 3.3. Parhelophilus integer captured via pan trap at MNWR ...... 103

vii

LIST OF APPENDICES

Appendix 1. Pollinating wasp species recorded in restored wetlands in central New York, 2019 – 2020...... 132 Appendix 2. Ratio of bee to non-bee pollinators recorded on each flowering plant species swept in restored wetlands...... 133 Appendix 3. Interaction network illustrating plant-pollinator relationships between wasps and flowers...... 134

viii

ABSTRACT

M.M. Jacobson. Response of Wild Pollinator Assemblages to Management of Restored Wetlands in Central New York, USA. 148 pages, 16 tables, 20 figures, 3 appendices, 2021. Journal of Insect Conservation style guide used.

Wetlands are restored and managed primarily to benefit waterfowl with little information about how these activities affect other taxa like pollinators. This study was among the first to survey pollinators in managed wetlands, and characterized the availability of floral resources, as well as bee and syrphid fly assemblages, among wetlands in central New York experiencing hydrological management. Over 80 bee species and 38 syrphid species were collected in 2019 and 2020, totaling 10,170 individuals. Bee and entomophilous plant assemblages did not differ in richness, and differed only weakly in composition, among treatments. Open water, invasive graminoids, and monotypic cattail negatively predicted bee richness and entomophilous plants. Treatments provided the greatest diversity of floral resources in late summer, and were exploited opportunistically by many generalists and some rare or specialized species. Wetland managers should strive for landscape heterogeneity through a mix of drawdowns, and control invasive graminoids to improve habitat quality.

Keywords: Wetland restoration, wetland management, native pollinators, wild bees, Syrphidae, plant-pollinator interactions, New York, Montezuma Wetlands Complex

M.M. Jacobson Candidate for the degree of Master of Science, April 2021 Melissa K. Fierke, Ph.D. Michael L. Schummer, Ph.D. Department of Environmental Biology State University of New York College of Environmental Science and Forestry Syracuse, New York

Chapter 1

LITERATURE REVIEW AND PROJECT SUMMARY

Importance and Ecology of Pollinating Insects

It has been estimated that over 87% of flowering plants depend on animal pollination (Ollerton et al. 2011), including 70% of the world’s most important crop species, which accounts for more than a third of global food production (Klein et al. 2007). The economic value of pollination has been estimated at $215 billion annually worldwide (Gallai et al. 2009), and while a wide range of animal taxa act as pollinators, including birds and small mammals, the primary pollinators worldwide are insects (Winfree et al. 2011). Members of Hymenoptera (bees and wasps), Diptera

(flower and bee flies), Lepidoptera (butterflies and moths), and Coleoptera (longhorn and checkered beetles) all contribute to pollination, however, bees (Apoidea: Anthophila) are most effective, with branched body hairs adapted specifically for pollen collection and a life history that requires visiting flowers to acquire resources for larvae and adults (Winfree et al. 2011;

Danforth et al. 2019). A number of bee species are currently used in agricultural production in the U.S., e.g., Megachile rotundata (Fabricius; alfalfa leafcutter bee, non-native) and Osmia lignaria Say (blue orchard bee, native), but by far the majority of crop pollination is performed by Apis mellifera L., the domestic honeybee (Morse and Calderone 2000; Losey and Vaughan

2006; Pitts-Singer and Cane 2011). Only fairly recently has the importance of wild native pollinators begun to receive attention, despite accounting for over $3 billion in pollination services annually in the U.S. (Losey and Vaughan 2006). Native bees are equivalent to or more effective than honeybees at pollinating several key fruit crops, including apples (Vicens and

Bosch 2000), cherries (Holzschuh et al. 2012), and squash (Artz and Nault 2011), with tomatoes in particular receiving great benefits in fruit set from the sonicating (“buzz-pollinating”) technique by bumble bees of which honeybees are incapable (Greenleaf and Kremen 2006;

Garibaldi et al. 2013). Due to losses of honeybee colonies as a result of Colony Collapse

Disorder in the last decade and a half, more attention has been placed on the potential of native bees to supplement honeybee pollination (Kremen et al. 2002; Garibaldi et al. 2013; Blitzer et al.

2016; Garantonakis et al. 2016). Alongside the critical and irreplaceable services provided to natural ecosystems, the growing importance of native bees for the human food supply dictates that promoting the persistence of healthy, diverse wild pollinator populations must be a priority worldwide.

A majority of the world’s 20,000 bee species, around 77%, are solitary in nature

(Danforth et al. 2019). These bees lack the reproductive division of labor of colonial honeybees and bumble bees, with one female excavating a nest and provisioning brood cells with pollen.

Some bees are strictly solitary, while others may be communal, where aggregates of one species gather in suitable habitat and may share a nest entrance but do not otherwise interact. About 10% of bees exhibit social behavior in some form, which can fall anywhere along a spectrum from semisocial and subsocial, with multiple females cooperating to maintain a nest and reproductive roles shifting over time, to advanced (or obligate) eusociality, where a single queen lays eggs and non-reproductive worker females carry out nesting, foraging, and rearing duties (Schwarz et al.

2007; Lawson et al. 2016; Shell and Rehan 2018; Danforth et al. 2019). Lastly, around 13% of bees are brood parasites (kleptoparasites) or social parasites, which make no nest of their own but instead usurp that of a specific host, laying an egg which will develop into a larva that kills the host larva and eats its provisions, or commandeering the workers of a social species to raise her own young (Danforth et al. 2019). Over 80% of non-parasitic bees are ground-nesting according to recent estimates, while the remainder nest in cavities by chewing the pith from plant stems, mining tunnels in wood, or using preexisting holes above or below ground (Cane 2003;

Harmon-Threatt 2020). The diet of a particular bee species is closely related to its ecological

niche and the pollination services it provides; most are polylectic (generalists), visiting a wide range of flowering plants across unrelated families, whereas others are oligolectic or monolectic specialists which collect pollen from a single family or genus, but may visit other flowers for nectar (Cane 2020). In the northeastern U.S., it is estimated ~ 15% of native bee species are specialists (Fowler 2016).

Syrphid flies (Diptera: Syrphidae) can be effective pollinators of many wild and cultivated plants, and are abundant in a variety of habitats (Rader et al. 2016; Skevington et al.

2019). They often exhibit morphological and/or behavioral mimicry of aculeate Hymenoptera with which they share floral resources. Larvae of syrphids perform a wide array of important services such as nutrient recycling and agricultural biocontrol as predators of aphids. Many species have saprophagous larvae that feed on living or dead aquatic plant matter, sometimes specialized to the roots or stems of certain plants (Skevington et al. 2019). While syrphids and their life histories are well-studied in Europe, significant gaps in information exist for North

American species, thus for many, larval food source, adult floral interactions, and conservation status are still unknown (Skevington et al. 2019). Recent studies suggest their contribution to pollination may in some cases rival or even exceed that of bees due to their abundance and offer benefits to plants by providing redundancy and a diversity of pollination methods (Kearns 2001;

Ssymank et al. 2008; Rader et al. 2016).

Drivers of Pollinator Declines

Declines in wild pollinators have been documented in recent decades around the globe, though particularly thoroughly in Europe and North America (Colla and Packer 2008; Cameron et al.

2011; Senipathi et al. 2015). Many species have experienced severe reductions in geographic range and population size within just the last three decades (Colla and Packer 2008, Cameron et

al. 2011, Colla et al. 2012). Habitat loss and fragmentation mostly associated with agricultural intensification (Grixti et al. 2008; Bommarco et al. 2012; Senapathi et al. 2015; Koh et al. 2016) are considered a primary cause of pollinator declines (Goulson et al. 2008; Potts et al. 2010).

Bumble bees (Bombus spp.) have notably reacted adversely to these anthropogenic changes, with several once-common species undergoing precipitous declines and a few near extinction (Colla and Packer 2008; Cameron et al. 2011; Bartomeus et al. 2013).

Native bees, particularly generalist social species like Bombus, rely on temporally staggered food sources found in plant species-rich systems to meet energy demands for brood- provisioning and colony maintenance throughout their seasonal flight period (Westphal et al.

2009; Rundlöf et al. 2014; Galpern et al. 2017). Modern monocultures offer few resources outside of sporadic mass-flowering events, making most crop fields unsuitable for establishing persistent populations of pollinators unless sufficient adjacent semi-natural habitat exists to supplement their diet (Mandelik et al. 2012). Leguminous cover crops like alfalfa and clover were once widely used to renew soil nitrogen, and served as a critical diet supplement for pollinators, especially long-tongued bumble bees. However, with the advent of nitrogenous fertilizers leguminous cover crops have become much less common, and the loss of this steady resource may have contributed to subsequent bumble bee declines in the midwestern U.S.

(Goulson et al. 2008; Grixti et al. 2009). Additionally, species-poor agricultural landscapes are unlikely to host many of the plants which specialist bees depend on for pollen to provision their young (Morandin and Kremen 2013; Kremen and M’Gonigle 2015). Human-modified systems also likely lack healthy populations of host species for kleptoparasites, potentially leading to further losses in biodiversity at multiple trophic levels.

Besides floral resources, wild bees require nesting sites, either in the ground or cavities

(e.g., plant stems or dead wood). The soil disturbance associated with agricultural activities can destroy below-ground nests, and remove weedy perennial stems and coarse woody debris used for nesting (Fussell and Corbet 1992; Svensson et al. 2000; Shuler et al. 2005).

The potential effect of habitat conversion on pollinating flies is not well understood, due to a scarcity of data on baseline abundances and basic life history information. Unlike bees, syrphid larvae are not homogenous in their dietary and habitat requirements, thus it is possible that species vary widely in their response to anthropogenic change (Kearns 2001; Moquet et al.

2018). Although some studies have revealed compositional or diversity differences in syrphid assemblages with changes in land use, favoring generalist species with broad ranges in modified areas (Bañkowska 1980; Bramquart and Hemptinne 2000; Biesmeijer et al. 2006; Schweiger et al. 2007), others suggest syrphids are overall more tolerant of habitat loss than bees (Rader et al.

2016).

Anthropogenic change has brought with it numerous secondary factors that synergistically exacerbate bee declines. Persistent pesticides used in agriculture, especially neonicotinoids (e.g., imidacloprid), further reduce forage availability, and are known to have a number of sublethal effects on social bees, including reduced colony growth (Whitehorn et al.

2012) and impaired learning ability needed to manipulate flowers (Stanley et al. 2015).

Neonicotinoids and commonly encountered fungicides have also been shown to cause immune suppression in social bees (Pettis et al. 2013; Hladik et al. 2016; Brandt et al. 2016), leaving them vulnerable to high loads of pathogens like Nosema bombi Fantham & Porter, that has in recent decades spilled over to wild bumble bees in large numbers from commercial greenhouse colonies (Cameron et al. 2011; Cameron et al. 2016). This pathogen has been implicated as a

significant contributor to the decline of several North American Bombus species (Cameron et al.

2011; Goulson et al. 2015). Effects of pesticides on solitary bees remain largely unknown, as most studies have focused on measuring the response of managed pollinators like A. mellifera and Bombus impatiens Cresson. However, negative effects on the reproductive success of Osmia mason bees, and lowered abundance and species richness of Lasioglossum sweat bees, have been found in agricultural areas with greater pesticide use (Scott-Dupree et al. 2009; Sandrock et al.

2014; Mallinger et al. 2015). Potential competition for resources with exotic species like A. mellifera adds to existing stress induced by habitat loss, pesticide exposure, and pathogens, and may contribute to diet simplification which has further ramifications for immune health (Goulson

2003; Goulson et al. 2015, Geslin et al. 2017).

Climate change will likely pose additional threats in the near future, and detection of these effects is only possible with long-term datasets currently lacking for most taxa and regions.

Maladaptive range shifts, where species follow elevational gradients into shrinking habitat instead of moving north to track cooler climes, have been observed in some declining Bombus species (Cameron et al. 2011, Ploquin et al. 2013; Tucker and Rehan 2017a; Jacobson et al.

2018). Phenological mismatch and the breakdown of historic plant-pollinator relationships are also possible, as pollinators and their floral hosts might respond at different rates to warmer spring temperatures and an earlier start to the growing season (Burkle et al. 2013; Kerr et al.

2015). The responses of pollinators to these drivers and their interrelated effects vary across body size, sociality, diet, and taxon, making it difficult to assess how species will react given how few have been studied in-depth (Bommarco et al. 2010; Bartomeus et al. 2013; Rader et al.

2014).

Although many bee species have declined, a handful of others have increased in abundance and expanded their range, such as B. impatiens (Colla and Packer 2008; Goulson et al. 2008; Jacobson et al. 2018). Generalists and exotic species that have adapted well to anthropogenic change are able to fill vacancies created by extirpations, taking on a more dominant role in ecosystems than before. Ultimately, pollinator declines and biotic homogenization alter community structure and could lead to reduced diversity and ecosystem services as specialists and functionally redundant species are lost, and plant-pollinator interactions become more simplified (McKinney and Lockwood 1999; Olden et al. 2004).

Habitat Conservation for Pollinators

Concerns for pollinator populations have prompted management action to protect, enhance, or create suitable pollinator habitat. However, difficulties arise when attempting to manage for wild bees, due to a lack of basic information on the natural histories of many species. Even less attention has been paid to non-bee pollinators. Much of the focus of current management for pollinators has been on improving habitat quality and increasing quantity in agriculture- dominated landscapes, especially in Europe, where such practices are known as agri- environment schemes (Kleijn et al. 2006; Pywell et al. 2006). Incorporating strips of wildflowers with a variety of blooming periods between crop fields has been shown to benefit many types of pollinators, by diversifying available floral resources to support generalists and social species like bumble bees (Pywell et al. 2006; Carvell et al. 2007). Permitting crop margins to return to wild forbs can also provide sources of food once crops cease flowering. Hedgerows, margins that have been planted with a diversity of native forbs and shrubs, are particularly useful in offering refuge from disturbance like pesticides and soil upheaval and are more likely to host specialists than other types of herbaceous flower strips (Morandin and Kremen 2013; M’Gonigle et al.

2015). Fallow fields, as well as natural or managed grasslands adjacent to cultivated land, provide additional floral resources for pollinators in homogenized human-modified landscapes

(Mandelik et al. 2012; Toivonen et al. 2015).

Success of pollinator management techniques in agricultural systems can be influenced by larger landscape-level factors. Addition of floral resources may have a greater influence on pollinator diversity in intensively cultivated landscapes than semi-natural areas with diverse crops and native vegetation (Holzschuh et al. 2007; Carvell et al. 2011). This response can be due to the concentration of bees and syrphids in these small areas when few other resources exist, though less so for native hedgerows, which may support independent populations of pollinators (Morandin and Kremen 2013; M’Gonigle et al. 2015). Mixed patterns have been observed when comparing conventional and organic farming techniques, with the diversity and composition of pollinator assemblages influenced not just by farming system but farm size and availability of adjacent uncultivated land, sometimes resulting in few or no measurable benefits of organic management (Holzschuh et al. 2007; Brittain et al. 2010; Tucker and Rehan 2017b).

However, pesticide drift from agriculture can affect pollinators in natural habitats depending on distance, direction of drift, and landscape matrix, adding further considerations to placement of pollinator habitat (Park et al. 2015; Hladik et al. 2016). In such cases where unfarmed habitat was present, it has been suggested that preserving non-crop land offers greater benefits, whereas improving diversity within crop fields is best suited for highly modified, intensively farmed landscapes (Egan and Mortensen 2012).

Composition of the surrounding landscape, and landscape heterogeneity, also influence wild pollinator assemblages, and in recent years other cover types besides agriculture and grassland have begun to receive more attention in surveys. Forests offer coarse woody debris for

nesting and spring-blooming plants for specialists and early-emerging bees (Mandelik et al.

2012), while clearcuts of varying intensities provide ground-nesting sites and a successional gradient of resources over time as reforestation occurs (Romey et al. 2007; Rubene et al. 2015).

Urban and suburban areas have been a growing focus for their potential to host pollinators and to assess the influence of urban sprawl on the landscape (Matteson et al. 2008; Hinners et al. 2012;

Normandin et al. 2017). Complementary habitat use by assemblages of pollinators, with long- flying generalist species using different cover types in a heterogeneous landscape for their seasonal resources, is an essential factor in the persistence and spatiotemporal management of pollinators (Mandelik et al. 2012). However, a paucity of information regarding, in many cases, even baseline data on pollinators in diverse cover types, hinders our ability to prescribe large- scale management for wild pollinators, particularly for those that cannot use fields or crop margins due to diet specialization and are vital for the functioning of native communities.

Wetland Restoration and Management

Wetlands are among the most biologically and economically productive systems in the world

(Costanza et al. 1997), providing numerous services such as nutrient cycling, water purification, flood retention, carbon sequestration, and nursery habitat for species important to fisheries and recreation (Zedler and Kercher 2005). Hundreds of species of terrestrial vertebrate wildlife rely on wetlands for at least some portion of their lifecycle, including over 170 species of birds in

North America, like the millions of waterfowl that depend on the Prairie Pothole Region for breeding habitat (Stewart 1996; NABCI 2016; USFWS 2018). However, less than half of the historic wetland area in the United States remains, lost primarily to draining, filling, and ditching for agriculture, navigation, and residential and industrial development (Dahl 1990; Dahl and

Allord 1996; Jasikoff 2013). New York has experienced around 60% wetland loss, as a result of

agricultural expansion, urban development, and construction of the Erie Canal and other waterways (Barringer et al. 1996; Jasikoff 2013). Remnant wetlands tend to be degraded in quality and often isolated in predominantly agricultural landscapes, with a reduced ability to perform essential ecosystem functions and support wildlife (Zedler et al. 2005). During the last half century, concerns for the persistence of waterfowl populations drove substantial efforts to protect and restore wetlands (Strickland et al. 2009; USFWS 2018). The 1986 North American

Waterfowl Management Plan (NAWMP) established a federal framework to initiate and promote wetland restoration projects, with $1.7 billion in grants issued by the 1989 North American

Wetlands Conservation Act (NAWCA) to undertake these ventures (USFWS 2018, 2020).

Additional support and implementation has come from the United States (US) Federal Migratory

Bird Hunting and Conservation Stamp, US Department of Agriculture Wetland Reserve Program

(WRP), and numerous others, which involve wildlife habitat restoration funding and technical assistance for public and private landowners through mandatory annual hunting fees and incentivized voluntary programs (King et al. 2006; USFWS 2019). As such, one of the primary goals of wetland restoration in North America is to provide sources of food, shelter, and breeding habitat necessary for waterfowl and other wetland-dependent wildlife for hunting, recreational, and ecological purposes (Fredrickson and Taylor 1982; USFWS 2018).

Management of restored wetlands generally involves hydrological manipulation, as past agricultural activities disrupted, sometimes irreversibly, seasonal flooding and drying cycles that maintain the health and diversity of wetland ecosystems (Fredrickson and Taylor 1982; Dahl and

Allord 1996). Wetland managers use water control structures to mimic these historic cycles, allowing them to make adaptive decisions regarding water level depths at certain key times of year that serve to meet life history needs of plants and wildlife as outlined in management plans

(Fredrickson and Taylor 1982; Strickland et al. 2009; Schummer et al. 2012; Gray et al. 2013;

Farley 2020; Schummer et al. 2021). In addition, coordinated soil and vegetation disturbance

(e.g., disking and mowing) simulate natural river-scouring and long-term flood events that expose the seed bank, creating conditions necessary for germination of desired annual and perennial plants (Fredrickson and Taylor 1982; Gray et al. 2013). These “moist-soil” management techniques are successful at increasing waterbird habitat use and providing ecosystem services for humans and wildlife (Kaminski et al. 2006; King et al. 2006; O’Neal et al. 2008; Jenkins et al. 2010). However, concern exists that restoration may not be recovering wetland functionality despite meeting management objectives for target wildlife taxa, with some biogeochemical processes not reaching a reference state even after a century (Moreno-Mateos et al. 2012). Vegetative communities of restored wetlands frequently differ from non-impacted sites even decades after restoration, potentially based on hydrological differences, invasion by aggressive exotic species, and whether plantings were conducted during restoration to establish characteristic flora (Gutrich et al. 2009; Matthews and Spyreas 2010; Moreno-Mateos et al.

2012). In addition, wetland conditions produced by management to benefit waterfowl throughout the annual cycle may not be optimally supporting other wetland-dependent species, however assessments are lacking (Rundle and Fredrickson 1981; Nadeau and Conway 2015; Farley 2020).

Few long-term monitoring requirements or metrics of community diversity exist for restored wetlands, thus posing difficulties in assessing restoration progress or responses of lesser-studied non-focal groups (e.g., terrestrial invertebrates) to current management regimes (Brown et al.

1997; Zedler 2000; Matthews and Spyreas 2010).

Potential Importance of Wetlands for Pollinators

Although few studies exist, wetlands restored from agriculture may have great potential to produce resources for pollinators. Several wetland flowering plants are hosts to specialist bees, particularly beggarticks (Bidens spp.), pickerelweed (Pontederia cordata L.), willow (Salix spp.), and native loosestrife species (Lysimachia spp.), and at least two bee species in the northeast –

Hylaeus nelumbonis (Robertson) and Lasioglossum nelumbonis (Robertson) – are believed to be wetland habitat specialists for reasons other than diet (Fowler 2016; Fowler and Droege 2020; S.

Droege, USGS, pers. comm.). Surprisingly, while managed wetlands provide abundant flowering plants that are geographically widespread in North America (e.g., smartweeds, Persicaria spp., milkweeds, Asclepias spp., and beggarticks), diet generalists are also relatively understudied in these ecosystems. In addition, several notable syrphid taxa (e,g., most Eristalinae, many

Syrphinae) are associated with wetlands in larval and/or adult stages (Skevington et al. 2019).

When restored from a degraded state, particularly row crops, wetlands likely contribute substantially to abundance and diversity of pollinators in the landscape. Remnant wetlands in urban or agriculture-dominated landscapes act as refuges for a rich assortment of pollinators, including rare and threatened species, and may offer benefits for farmers through spillover into nearby crop fields (Moroń et al. 2008; Evans et al. 2018; Heneberg et al. 2018; Vickruck et al.

2019). Such refuges could be providing primary or supplemental floral resources for both generalists and specialists, as well as foraging and nesting sites safe from human disturbance, aiding in the persistence of a diverse pollinator assemblage in highly-modified areas.

Additionally, use of wetlands by long-flying generalists underscores their potential importance as complementary habitat at key times during the growing season (O’Neill and O’Neill 2010;

Mandelik et al. 2012; Heneberg et al. 2018; Begosh et al. 2020).

Project Summary and Objectives

Determining how native bees and syrphid flies utilize restored wetlands would be a critical step in elucidating their life histories and habitat requirements, thus increasing our capability to support and protect these wild pollinators. Further, understanding how management actions aimed at providing habitat for migratory birds and other wetland-dependent wildlife affect entomophilous plants and pollinators will help inform decisions to sustain community diversity.

The goal of this study was to evaluate the influence of wetland management techniques on entomophilous plant and native pollinator assemblages and diversity at restored wetlands in central New York. Specific objectives were to 1) quantify presence and frequency of entomophilous plants as resources for native pollinators among wetlands with differing management treatments, 2) describe native bee and fly assemblage diversity in restored wetlands, and their plant-pollinator associations, and 3) determine if native bee assemblages and diversity varied among wetland management treatments.

This thesis consists of four chapters, including a literature review, two data chapters, and a synthesis chapter. Chapter 2 presents the study performed to characterize use of restored wetlands by wild bee pollinators and the influence of drawdown management on availability of floral resources and pollinator assemblages. It is presented as a manuscript for publication in the

Journal of Insect Conservation. Chapter 3 consists of descriptive findings pertaining to syrphid fly diversity and ecology in restored wetlands and is formatted as a short communication for inclusion in the Journal of Insect Conservation. Chapter 4 offers broad implications and concluding remarks.

Chapter 2

RESPONSE OF WILD BEE ASSEMBLAGES TO MANAGEMENT

OF RESTORED WETLANDS IN CENTRAL NEW YORK, USA

ABSTRACT

To effectively protect wild bee pollinators and the services they provide, it is critical to understand their interactions with plants among a diversity of land uses. Wetlands are underrepresented in bee surveys, and although wetland restoration produces quality habitat for breeding and migratory waterbirds, it is still poorly known how hydrological management influences availability of floral resources and bee assemblages. In this study, restored wetlands that were actively or passively managed were surveyed to determine wetland structure and plant and bee assemblages in central New York, June – September 2019 and 2020; over 9,000 bees were collected, representing ≥ 80 species in 25 genera. Nearly 300 unique plant-pollinator associations were recorded, including those previously undocumented for Hylaeus nelumbonis

(Robertson). Pickerelweed, nodding bur-marigold, and swamp smartweed were particularly important for richness and abundance of bees. Bee and entomophilous plant assemblages were similar between treatments, and their diversity and frequency were negatively influenced by percentage of open water, invasive graminoids, and monotypic cattail. Flowers of wetland plant species were 80% more diverse in the later portion of the growing season than the early portion, suggesting wetlands are most important for bees in late summer and autumn. Passively and actively managed wetlands were each used by a rich assortment of generalist and specialist bees; maintaining a complex of wetlands with differing hydrological regimes, and controlling densities of invasive plants like common reed, can diversify the resources available for pollinators while also meeting management goals for waterfowl and other target vertebrate taxa.

Keywords: Wild bees · Wetland restoration · Wetland management · Plant-pollinator interactions · New York

INTRODUCTION

It has been estimated that > 87% of flowering plants depend on animal pollination (Ollerton et al.

2011), including 70% of the world’s most important crop species, which accounts for more than a third of global food production (Klein et al. 2007). The economic value of pollination has been estimated at $215 billion annually worldwide (Gallai et al. 2009), and of this, the contribution by wild pollinators is thought to be > $3 billion in the U.S. (Losey and Vaughan 2006). Concerns over honeybee (Apis mellifera L.) declines have prompted a reevaluation of our current reliance on a single species for a majority of pollination services, with substantial attention now focused on the potential of wild bees to supplement or even eliminate the need for honeybee pollination for some crops (Kremen et al. 2002; Garibaldi et al. 2013; Blitzer et al. 2016; Garantonakis et al.

2016). However, significant declines in many native bee species (particularly Bombus spp.) have been documented in recent decades (Colla and Packer 2008; Cameron et al. 2011; Bartomeus et al. 2013), due primarily to habitat conversion for agriculture (Potts et al. 2010; Senipathi et al.

2015). Current farming practices often fail to provide nesting and floral resources to sustain rare, imperiled, or specialized species (Kleijn et al. 2006; Kremen and M’Gonigle 2015), while also harming pollinators in myriad ways through extensive use of neonicotinoid insecticides (Scott-

Dupree et al. 2009; Whitehorn et al. 2012; Feltham et al. 2014; Mallinger et al. 2015). Spillover of pathogens like Nosema bombi Fantham & Porter from managed bees (Cameron et al. 2011;

Cameron et al. 2016), possible competition with honeybees (Goulson 2003), and climate change effects (Burkle et al. 2013; Jacobson et al. 2018) may also serve to compound and exacerbate drivers of pollinator declines (Goulson et al. 2015).

Extensive efforts have been made to improve habitat quality within intensively farmed landscapes, including use of wildflower strips (Pywell et al. 2006), native plant hedgerows

(Morandin and Kremen 2013; M’Gonigle et al. 2015), and fallow fields (Toivonen et al. 2015), which have shown promising success in increasing abundance and diversity of wild pollinators, including specialists in some cases (Kremen and M’Gonigle 2015). However, there are many landscape-level effects that need to be considered, such as distance to natural habitat and composition of the surrounding habitat matrix which may affect outcomes of restoration or enhancement (Carvell et al. 2011; Egan and Mortensen 2012; Mandelik et al. 2012).

Substantially less attention has been paid to understanding the role other cover types besides agriculture and grassland play in supporting wild bee populations (but see Hanula et al. 2016), as primary or complementary habitats (Mandelik et al. 2012). A heterogeneous landscape will likely further ensure persistence of the full breadth of generalist and specialist bees than modified, homogeneous ecosystems (Holzschuh et al. 2007; Mandelik et al. 2012; Andersson et al. 2013; Moreira et al. 2015).

Wetlands represent a cover type that can be common in agricultural landscapes yet remain undersurveyed for pollinators. In the last 250 years, > 50% of wetlands in the United

States have been drained, filled, or ditched for agriculture, navigation, and urban development, with much of what remains in a degraded state (Dahl 1990; Dahl and Allord 1996). These fragmented remnants are often embedded in homogenized landscapes, yet may serve as refuges for rare species and sites of supplemental floral and nesting resources for bees, which could offer spillover benefits for farmers growing crops requiring insect pollination (Evans et al. 2018;

Heneberg et al. 2018; Vickruck et al. 2019; Begosh et al. 2020). Restoration efforts made possible by the North American Wetlands Conservation Act (NAWCA) and North American

Waterfowl Management Plan (NAWMP) have returned functionality as wildlife habitat,

primarily aimed at waterfowl, to millions of hectares of previously farmed wetlands across North

America (USFWS 2018, 2020).

Wetland restoration often includes installation of water control structures that enable seasonal hydrological manipulation, to mimic historic flooding and drying cycles needed to sustain diverse wetland habitat that meets the life history needs of wetland-dependent plants and animals (Mitsch et al. 2005; Schummer et al. 2012; Gray et al. 2013; Schummer et al. 2021).

Wetlands with water control structures may be passively or actively managed; passively managed wetlands retain water at maximum possible capacity throughout the growing season

(Fleming et al. 2012), whereas active “moist-soil” management involves seasonal dewatering, and additional vegetation and soil disturbance regimes that simulate the effects of natural river scouring, promoting germination of diverse annual and perennial plant communities that support hundreds of species of wildlife (Fredrickson and Taylor 1982; Fleming et al. 2012; Gray et al.

2013). However, a scarcity of information still exists regarding the response of non-target taxa to management performed chiefly to meet the needs of waterfowl. Recent studies suggest wetlands of many kinds (Moroń et al. 2008; Heneberg et al. 2018; Vickruck et al. 2019), including actively managed wetlands (Stephenson et al. 2020), host a rich assemblage of wild bee pollinators. Yet, baseline data on bees inhabiting restored wetlands – their diversity, ecology, and response to management – are largely lacking despite the prevalence and importance of wetlands in the landscape. This lack of information hinders the ability of land managers to make informed decisions to protect and promote wild pollinators using wetlands.

decisions. Specific objectives were to 1) quantify presence and frequency of entomophilous plants as resources for native bee pollinators among wetlands experiencing different hydrological management, 2) describe native bee assemblage diversity and plant-pollinator associations in restored wetlands, and 3) determine if native bee assemblages and diversity varied among wetland management treatments. It was predicted that management treatments would result in distinct plant communities based on seasonal water levels, and distinct bee assemblages due to differences in the abundance and species of flowering plants available.

METHODS

Study Area

This study was conducted in the Montezuma Wetlands Complex (MWC; 43.024079°N, -

76.748412°W) and Seneca Meadows Wetland Preserve (SMWP; 42.937068°N, -76.823104°W) in central New York State, USA (Fig. 2.1). The MWC is > 26,000 ha of wetlands and uplands with a primary focus on preservation, restoration, and management of habitats for migratory birds, particularly waterfowl (Jasikoff 2013; Eckler et al. 2019; Wagner 2020). The MWC includes the United States Fish and Wildlife Service (USFWS) Montezuma National Wildlife

Refuge (MNWR; 3,970 ha), New York State Department of Environmental Conservation

(NYSDEC) Northern Montezuma Wildlife Management Area (NMWMA; 2,860 ha), and surrounding private lands. The SMWP is 230 ha of restored and enhanced wetlands, of which 20 ha are emergent wetlands (McGraw and Larson 2019). Wetlands and uplands at SMWP are subject to chemical and mechanical invasive plant control (B. Zimmerman, Applied Ecological

Services, pers. comm.). The landscape surrounding MWC and SMWP is predominantly row crop agriculture and forest (Jasikoff 2013; Eckler et al. 2019). MWC and SMWP are located in a

globally significant Important Bird Area due to the region’s central position along the Atlantic flyway, with over 320 bird species utilizing the MWC for at least some part of the annual cycle

(Jasikoff 2013; Eckler et al. 2019). Historically, the region contained > 20,000 ha of contiguous wetland habitat and was subject to seasonal flooding of the Seneca River from snow melt and precipitation, but these hydrological functions that once sustained wetland heterogeneity were greatly reduced by river engineering for the NYS Canal System and draining wetlands for agriculture (Jasikoff 2013). To enable wetland managers to mimic historic hydrology, most wetland restoration in the region includes use of water control structures that enable drying and flooding of wetland impoundments. Emergent wetlands at MWC are generally dominated by cattail (Typha spp.), with areas of perennial diversity including other graminoids (e.g.,

Sparganium spp., Carex spp., Cyperus spp.) and entomophilous plants (e.g., Sagittaria,

Nymphaea, many Asteraceae). Mudflats produced by active management results in germination of annual grasses (e.g., Echinocloa spp., Panicum spp.) and annual forbs (e.g., Bidens spp.,

Persicaria spp.) in late summer. Several invasive plant species are established at MWC and undergo regular management, such as common reed (Phragmites australis [Cav.] Trin. ex

Steud), narrow-leaved and hybrid cattail (Typha angustifolia L. and Typha x glauca Godr.), reed- canary grass (Phalaris arundinacea L.), and purple loosestrife (Lythrum salicaria L.).

Wetlands were systematically sampled in 2019 (n = 33) and 2020 (n = 33; total n = 38).

Wetlands were located primarily in the MWC, with 26 at the NMWMA, 9 at the MNWR, and 3 at the SMWP (Fig. 2.1), and ranged in size from 2 – 670 ha (mean = 39.0, median = 10.5). For those > 35 ha, a representative subset was sampled. Wetlands were subjected to one of three hydrologic management treatments, 1) passive, where water was held at or near full-service level

(maximum designed capacity) throughout the growing season with no active drawdown of water

(with water loss primarily from evaporation and evapotranspiration), 2) partial drawdown, where

20 – 50% of the basin was exposed by 1 Aug using active water drawdown, and 3) full drawdown, where nearly 100% of the basin was exposed by 1 Aug with the only remaining water located in the borrow ditch (areas where soil is “borrowed” from to make wetland berms;

Table 2.1). Wetlands experiencing partial or full drawdowns were also often subjected to invasive species control via herbicide application, and mowing and disking which mimic river scouring activities and expose native plant seeds for germination (Fleming et al. 2012; Farley

2020). Initially, sites were chosen by identifying all full drawdown wetlands and pairing them spatially with adjacent partial drawdown and passive wetlands. However, to meet management goals for wildlife habitat (e.g., desired vegetation for food or shelter, heterogeneous cover types) decisions regarding drawdowns were made annually, and influenced by precipitation, landscape context, and invasive species control. Due to these ongoing management actions some sites were not available to help balance the study spatially and numerically in the second year, and thus other suitable sites were substituted. In 2019, 17 passive wetlands, 6 partial drawdowns, and 10 full drawdowns were sampled. In 2020, 22 passive wetlands, 1 partial drawdown, and 10 full drawdowns were sampled. A Mantel test was performed post hoc to identify any spatial autocorrelation between treatments, because a randomized complete block design was not possible. The test confirmed no biases resulting from spatial autocorrelation, so all surveyed wetlands were included in final analyses.

Vegetation Sampling

Surveys were conducted monthly on consecutive fair-weather days (no precipitation, winds < 10 mph, temp. > 60°F) from Jun – Sept, 2019 and 2020. Sampling periods lasted approximately 2

weeks/mo: 2019 = 7 Jun – 24 Jun, 7 Jul – 16 Jul, 1 Aug – 10 Aug, 14 – 15 Sept and 21 – 22 Sept;

2020 = 14 Jun – 4 Jul, 14 Jul – 25 Jul, 10 Aug – 20 Aug, and 5 Sept – 14 Sept. In each wetland,

40 points were located at equidistant intervals along 2 – 4 equally-spaced linear transects.

Transect length and number varied due to the unique dimensions of each wetland, but were configured with points 5 – 20 m apart to capture wetland heterogeneity. Prior studies of these wetlands and those using similar methodology indicated 40 points were sufficient to represent the plant community (Roberts-Pichette and Gillespie 1999; Fleming 2012; Farley 2020).

Vegetation was surveyed using a point-intercept method, where all plant species in contact with a 3 cm diameter PVC pole were identified at each point. Species were identified to lowest possible taxonomic level in the field. Above-water vegetation height (± 5 cm) and water depth (±

5 cm), and if the point occurred in open water (lacking any emergent vegetation) or a mudflat (no measurable vegetation) were also noted for each point.

Pollinator Sampling

Wild pollinator surveys occurred on the same dates as vegetation sampling, with the exception of preliminary sampling 20 May – 25 May 2019 and 5 Jun 2020. Sampling was not performed until flowers of wetland species were present, to avoid biased collection of bees from adjacent cover types. Although bees were the primary focus of collection, other pollinating insects (syrphid flies, sphecid, vespid, and crabronid wasps) were also collected. Surveys used pan trap, or “bee bowl” (passive), and sweep-net (active) methods. Due to logistical limitations, only sweep- netting was performed in September 2019. Use of pan traps followed standard procedures

(Droege et al. 2016). Pan traps were 163 ml Dixie plastic soufflé cups (Georgia-Pacific, Atlanta,

GA) painted with fluorescent yellow, blue, or white paint (Guerra Paint & Pigment Corp., New

York, NY) secured to 1.2 m-height fiberglass stakes (Fig. 2.2). Using a hole-punch and mesh, three drainage holes were made near the rim of each cup to prevent overflow during rain events.

Traps were arranged with colors alternating at even intervals along vegetation sampling transects, totaling 30 pan traps per wetland in 2019 and 24 in 2020 (Droege et al. 2016). Passive sampling effort was reduced in the second field season in order to shift focus to performing more sweeps. Areas of impenetrable woody vegetation and open water > 1 m deep were avoided for trap placement. Traps were filled three-quarters of the way with a 2:1 mixture of soapy water

(Dawn Ultra blue dish soap, Procter & Gamble, Cincinnati, OH) and propylene glycol (Droege et al. 2016). Pan traps were collected after ~ 24 hrs in the field.

Sweep-netting was performed to capture taxa not commonly collected in pan traps (e.g.,

Bombus spp.), on sampling days generally between 900 and 1400 hrs when bees were most active (Droege et al. 2016). However, flowers observed being heavily visited after 1400 hrs were opportunistically swept to maximize documentation of unique plant-pollinator interactions.

Monotypic patches of flowering plants were targeted, but because flower density was highly variable sweeps using standardized timed passes were impractical. Instead, sweeps aimed to capture the diversity of pollinators on flowers by using multiple complete passes totaling ≤ 15 min; for plant species with only a few flowers per site, a pass consisted of stationary observation and netting of any pollinators. All pollinating insects netted were collected. Abundant plant species were swept more than rare ones because of frequency of encounters, thus reflecting the availability of floral resources in the landscape. Opportunistic hand-capture occurred only to collect a lone individual of a previously unrecorded or rare species. An acknowledged exception to the described methodology was pickerelweed (Pontederia cordata L.), which was swept more often due to its status as a host plant for the extremely rare monolege Melissodes apicatus Lovell

and Cockerell, which was known historically from New York and for which contemporary documentation was an aim. This sampling scheme was considered when interpreting pollinator interaction metrics.

Monthly pan trap and sweep data were pooled per wetland for statistical analyses. Any bees collected during pilot studies in May when vegetation or environmental data were not taken, or on upland flowers on wetland edges, were only included in analyses of plant-pollinator interactions.

Specimen Preparation

Specimens were stored in a 70% ethanol solution from time of capture until processing, which occurred throughout the study during non-sampling weeks and into the autumn after the field season. Upon removal from ethanol, specimens were washed in a mixture of Dawn dish soap and warm water to remove pollen and debris. Bees were dried using a hair dryer and mason jar, to prevent hair-matting and ensure even drying necessary for identification. All specimens were pinned and labeled with metadata indicating location, date, collection method, collector, and floral record (if applicable), and given unique ID numbers. Bees were identified to the lowest taxonomic level possible, usually species, using the interactive keys on DiscoverLife

(www.discoverlife.org) and relevant literature (Mitchell 1960, 1962; Rehan and Sheffield 2011;

Sheffield et al. 2011). Most members of the large, morphologically monotonous genus

Lasioglossum (Halictidae) were identified to genus, due to the expert attention required for confident identifications. Two species, L. coeruleum (Robertson) and L. nigroviride (Graenicher) are distinctive and were identified to species when present. Concerted effort was made to identify the few non-metallic Lasioglossum caught in sweeps for inclusion in plant-pollinator

networks, thus the recorded abundances for L. leucozonium (Schrank) and L. truncatum

(Robertson) are not likely representative of the true number collected in the study. Select identifications (primarily Hylaeus spp. and Andrena spp.) were verified by the Danforth Lab at

Cornell University, Ithaca, New York. Identifications were also compared against expertly- identified voucher specimens from the USGS Native Bee Inventory & Monitoring Lab (courtesy of Sam Droege). Information on bees’ diet preferences was obtained from Fowler and Droege

(2020). Voucher photos of select species were posted to BugGuide (www.bugguide.net).

Voucher specimens were deposited in the SUNY-ESF Insect Collection, Cornell University

Insect Collection, and American Museum of Natural History. Novel floral records were submitted to DiscoverLife and occurrence data were uploaded to the Global Biodiversity

Information Facility (GBIF).

Statistical Analysis

Statistical analyses were performed in R Studio 4.0.3 (R Core Team 2020). To determine how completely the bee assemblage was sampled and provide estimates of true species richness, rarefaction tests were conducted in package ‘CommEcol’ (Melo 2019) with Chao-1 (Chao 1987),

Jackknife (Burnham and Overton 1978), and Bootstrap estimates (Efron 1979), and sample- based species accumulation curves were developed for each wetland treatment. A Venn diagram was made using the ‘VennDiagram’ package (Chen and Boutros 2011) to visualize unique and shared species among treatments. Multinomial Species Classification Method (CLAM) tests

(Chazdon et al. 2011) in package ‘vegan’ (Oksanen et al. 2020) were used on bee species data to identify possible habitat generalists and specialists. This test can only compare two treatments at a time, so one was performed for each combination of the three treatments. Although all

combinations of treatments were tested, due to the low number of species unique to partial drawdowns and smaller sample size of wetlands in this treatment, it was determined that a comparison of full to passive drawdowns yielded the most accurate estimates.

Data from floral sweeps were used to generate plant-pollinator visual interaction networks using the function plotweb in the ‘bipartite’ package (Dormann 2008). Networklevel and specieslevel functions provided statistics relevant to community function. At the network level, weighted nestedness (Galeano et al. 2009) and connectance (Dunne et al. 2002) measure stability and resiliency of plant-pollinator systems; weighted nestedness quantifies overlap between generalist and specialist interactions, providing a metric of community complexity, and connectance measures what proportion of the possible interactions between bee and flower that can occur were present (or documented). A small number indicates a low functional redundancy for pollination services that may leave the system vulnerable if species losses occur. At the species level, degree, normalized degree, and Pollinator Service Index (PSI) quantified unique interactions among bee and flower species and the contribution of each pollinator species to the functioning of the system (Dormann et al. 2008; Delmas et al. 2018). PSI attempts to quantify the importance of each bee species to the persistence of the plant community by utilizing the proportion of the foraging visits by a bee species that go to each plant species, and the proportion of visits that a plant species receives by each bee species (Dormann et al. 2008). A score of a maximum 1 indicates a monolectic specialization where both bee and plant are dependent on each other, whereas 0 indicates the pollinator is essentially redundant enough, or contributes so little, that its loss would not matter to the system. Thus, bees that either have a broad host range or are exclusive to a single host can potentially rank high, if they account for a large proportion of visitors to a flower species.

Phenology plots for both flowering plants and bees were created using the ‘ggplot2’ package (Wickham 2016), utilizing collection, sweep, and observational data on bloom times at each site across years to aid in visualization of bee flight seasons and turnover of floral resources available to pollinators across the growing season. The Wetland Indicator Status reported for each plant species was obtained through the New York Flora Atlas (newyork.plantatlas.usf.edu).

Linear and generalized linear mixed models were constructed to test for differences in environmental variables among treatments, using packages ‘lmerTest’ (Kuznetsova et al. 2017) and ‘MASS’ (Venables and Ripley 2002). A Poisson distribution was used for dependent variables that did not meet assumptions of normality, and a negative binomial distribution was used for proportional dependent variables. Mean water depth, mean vegetation height, and percentage open water, monotypic cattail, invasive graminoids, and mudflat were dependent variables and treatment, month, and treatment × month were fixed predictor variables. Year was included as a covariate to control for annual variation in climate not included in models, and wetland as a repeated measure to account for sampling the same sites across months and years.

Predictor variables were considered significant at α = 0.05.

To determine factors influencing composition of plant and pollinator assemblages, Non- metric Multi-dimensional Scaling (NMDS) ordination was investigated using the ‘vegan’ package, with a Bray-Curtis distance matrix (999 permutations) and applying a Wisconsin double standardization to count data (Bray and Curtis 1957; Noy-Meir et al. 1975). The number of dimensions was limited to three, to ensure results could be interpreted and visualized for management use. Environmental characteristics were fit and overlain onto the ordination to illustrate potential directional relationships, and permutational ANOVAs (PERMANOVA) were used to test for multivariate differences in vegetation and bee assemblage composition among

treatments, month, treatment × month, and environmental conditions (e.g., percentage open water). Homogeneity of dispersion among treatments for vegetation and bee count data was tested using the betadisper function in ‘vegan’ to assess bias introduced by unequal sample sizes

(Anderson 2006). A significant p-value (p > 0.05) returned from an ANOVA indicates heterogeneity in group variance and may result in an overestimation of differences in community structure by the PERMANOVA when caused by the group with the smallest sample size

(Anderson and Walsh 2013). Datasets showed heterogeneous dispersion, and a Tukey’s HSD test revealed that for bee and plant assemblages, partial drawdowns were the source of heterogeneity.

Shannon-Wiener indices were calculated to estimate alpha diversity of bee assemblages for each treatment, and Sørensen’s Dissimilarity indices to compare beta diversity of bee assemblages through overlap of species among treatments. The Sørensen’s Dissimilarity index ranges from 0-

1, where 0 indicates the assemblages are exactly the same, and 1 indicates they are completely different.

Competing hypotheses for models explaining predictors of variation in plant, entomophilous plant, and bee species richness, and frequency of occurrence of entomophilous plants per wetland were tested using mixed models with the same covariate and random effects as noted above. Prior to model-building, independent variables were tested for correlation using

Pearson correlation coefficients and variables at -0.50 > r > 0.50 were not included together in models as fixed effects (Mukaka 2012; Schober et al. 2018; Table 2.2). An information theoretic approach was used for model selection and to calculate Akaike’s information criterion (AIC) for each competing model (Burnham and Anderson 2002). For each dependent variable, ΔAIC and

AIC weights (wi) were used to select the most parsimonious model(s). Tukey’s Honest

Significant Difference (HSD) pairwise comparison tests were used at α = 0.05 among treatments and months of top models.

RESULTS

Wild Bee Assemblages

A total of 9,046 bee specimens of ≥ 80 species were collected in 2019 and 2020 (Table 3), representing five families: Halictidae (75.2%), Apidae (21.2%), Colletidae (2.8%), Andrenidae

(0.44%), and Megachilidae (0.29%). Most specimens of genus Lasioglossum, as well as some individuals in other genera, could not be identified to the species level, thus the true species richness can be presumed to be substantially greater. Differences in passive sampling efforts between 2019 and 2020 had a negligible effect on bee abundance or diversity captured at study sites (2019: 4,379 bees of ≥ 48 spp.; 2020: 4,501 bees of ≥ 71 spp.). Pan traps captured 7,599 bees (84%) of ≥ 55 spp., and 1,447 (16%) of ≥ 61 spp. were collected in sweeps. There were ≥

36 bee species collected in pan traps and sweeps, while ≥ 25 species were only collected in sweeps (Tables 4 and 5). In June 2,641 specimens of ≥ 51 spp. were collected, 3,941 of ≥ 37 spp. in July, 1,938 of ≥ 49 spp. in August, and 463 of ≥ 31 spp. in September. An additional 63 individuals of ≥ 6 spp. were collected in pilot sampling in May 2019, including a single Colletes inaequalis Say, a species not captured during other sampling periods. Excluding unidentified

Lasioglossum, 17 species were common to all sampling periods. The five most abundant species were Agapostemon virescens (Fabricius; n = 654), Apis mellifera (n = 579), Melissodes bimaculatus (Lepeletier; n = 516), Agapostemon splendens (Lepeletier; n = 360), and Bombus impatiens Cresson (n = 222), comprising 64% of all non-Lasioglossum bees caught. These were, for the most part, the most abundant species for each drawdown as well (Fig. 2.3, Table 2.6).

There were 17 species (21%) represented as singletons. In full drawdowns ≥ 55 species were recorded (10 unique), ≥ 31 spp. in partial drawdowns (1 unique), and ≥ 66 spp. in passive wetlands (22 unique). There were 25 species shared among all treatments (Fig. 2.4). CLAM tests identified only two potential habitat specialists, with 100% (n = 93) of Dufourea novaeangliae

(Robertson) collected in passive wetlands, and 84% (n = 51) of Halictus rubicundus (Christ) collected in full drawdowns (Fig. 2.3). Other species were unique to treatments, but were not common enough for the test to draw conclusions about specialization. Of the species documented, 60 (75%) were polylectic (pollen generalists), while 14 (17.5%) were oligolectic and 6 (7.5%) were monolectic. Comparing nesting habits, 51 species (64.5%) were ground- nesting, 24 (30.4%, excluding A. mellifera) were cavity-nesting, and 4 (5.1%) were kleptoparasites (Table 2.3). Rarefaction using the Bootstrap test estimated a minimum true richness of 92 bee species for the study area (Fig. 2.5); other tests offered greater estimates (first- order Jackknife = 109, Chao-1 = 118).

Plant-pollinator Interactions

Bees were collected from 60 flowering plant species in 30 families; several additional species were swept unsuccessfully, with no insect visitors captured or observed (Tables 2.7 and 2.8). By month, ≥ 33 spp. were netted in June from 23 plant species, ≥ 21 spp. from 12 plant species in

July, ≥ 38 spp. from 26 plant species in August, and ≥ 23 spp. from 13 plant species in

September.

A phenology diagram of bees collected shows that 42 species had a flight season of at least one month, and 30 species with a flight season of two months or more (Fig. 2.6). Genera such as Andrena spp. and Colletes spp. had shorter documented flight seasons. There were 30

species documented from only a single week over the duration of sampling. However, bee phenology may not be accurate for all species due to scarcity or low detection. For the purposes of this study, flowers of wetland species are defined as those having a Wetland Indicator Status of FAC, FACW, or OBL; while FACU and a few UPL plants were recorded at study sites, these were above the full-service level line (FSL; the greatest amount of water an impoundment is designed to hold) and cannot be considered regular or likely inhabitants of wetlands. In the early portion of the growing season, defined here as late-May to late-June, flowers of 26 wetland species were observed blooming (Fig. 2.7). During the mid-season (July to early-August) flowers of 29 wetland species were blooming. In the late season (mid-August to September) flowers of

47 wetland species were in bloom. The only species recorded blooming throughout the entire study period was wild mustard (Brassica rapa L.), however this is an upland species only detected outside of FSL and had poor visitation by bees.

A plant-pollinator network was constructed using pooled sweep data from all treatments and months to characterize interactions recorded among ≥ 64 bee species and 60 flower species in wetlands (Fig. 2.8). A total of 296 unique species interactions were represented. The weighted nestedness of the system was 0.58 out of 1, and the connectance was 0.08 out of 1. More A. mellifera were caught on flowers (n = 429) than any other bee species. It also had the greatest degree, or number of unique floral interactions, as it was recorded on 36 flowering plant species, and had the greatest normalized degree (0.60), which is scaled to the number of possible floral hosts (Table 2.7). Among native bees (excluding the aggregate Lasioglossum), Bombus impatiens had the highest degree (26) and normalized degree (0.43). When examining PSI,

Colletes latitarsis Robertson had the greatest score with the maximum 1.0, as its only documented host was ground-cherry (Physalis longifolia Nutt.) and it was the plant’s only

pollinator (Table 2.7). The next greatest species was A. mellifera at 0.54. When examining flowers, the species with the greatest degree (19) and normalized degree (0.29) was pickerelweed, followed by swamp smartweed (Persicaria hydropiperoides [Michx.]; 17, 0.28) and giant goldenrod (Solidago gigantea Aiton; 16, 0.26; Table 2.8).

Vegetative & Environmental Characteristics of Drawdowns

During surveys, 161 species of plants were recorded, including 128 in passive wetlands, 125 in full drawdowns, and 71 in partial drawdowns. The five most common species were rice cutgrass

(Leersia oryzoides [L.], n = 1,291, present at 12.2% of survey points), white waterlily

(Nymphaea odorata Aiton, n = 1,141, 10.8%), narrow-leaved cattail (n = 1,008, 9.5%), purple loosestrife (n = 927, 8.8%), and reed-canary grass (n = 868, 8.2%). Of the plants recorded, 96 species (59.6%) are considered entomophilous, when including plants that are both wind and insect pollinated. Overlap existed between the most frequently encountered entomophilous plants in each drawdown (Fig. 2.9, Table 2.9).

Proportion of open water (F2,251 = 50.7, p < 0.001), mean water depth (F2,251 = 56.8, p <

0.001), proportion of invasive graminoids (F2,251 = 7.1, p = 0.001), proportion of mudflat (F2,251 =

22.4, p < 0.001), mean vegetation height (F2,251 = 7.4, p < 0.001), and proportion of monotypic cattail (F2,251 = 20.6, p < 0.001) varied among treatments (Fig. 2.10). However, results of Tukey’s

HSD determined partial drawdowns were only significantly different from the other two treatments for mean water depth. A significant overall effect of month was detected for open water (F3,251 = 5.1 p=0.018), mean water depth (F3,251=8.0, p < 0.001), and vegetation height

(F3,251 = 29.2, p < 0.001), but partial drawdowns did not differ significantly by month for any predictor variable.

Response of Wild Bees and Floral Resources to Wetland Management

At three dimensions, stress levels of the NMDS ordination fits for plants and bees were ≥ 0.2, indicating a complexity of these data beyond what could be easily applicable from a management perspective or visualized in a useful way by the NMDS. PERMANOVA tests revealed that the composition of plant assemblages was most strongly influenced by mean water depth (F1,258 =

2 22.4, p = 0.001, R = 0.08), followed closely by proportion of open water (F1,258 = 20.1, p =

2 2 0.001, R = 0.07), and drawdown treatment (F2,257 = 8.8, p = 0.001, R = 0.06). Other factors were significant at p = 0.001, but with low R2 values (R2 = 0.02). For bee assemblages, the only factor

2 2 to have an R > 0.02 was sampling month (F3,225 = 6.6, p = 0.001, R = 0.08). Treatment was significant but explained little of the observed variation in bee assemblages (F2,225 = 2.3, p =

0.001, R2 = 0.02).

The Shannon-Wiener Diversity index was calculated for bee assemblages at 1.0 ± 0.14 in full drawdowns, 0.63 ± 0.22 in partial drawdowns, and 0.84 ± 0.10 in passive wetlands.

Shannon-Wiener indices differed between full and partial drawdowns (p = 0.01), but not between partial and passive (p = 0.22) or full and passive (p = 0.13). The Sørensen Dissimilarity Index was calculated for bee assemblages to be 0.38 between full and partial drawdowns, 0.45 between partial and passive wetlands, and 0.29 between full and passive wetlands.

The top model for plant species richness contained a negative effect of proportion of open water, but a positive effect of proportion of points with invasive graminoids (Tables 2.10 and

2.11). At median proportion open water (median = 20%), greatest plant species richness was predicted as 20.8 (95% CI: 18.4 – 23.2) species per wetland in August with invasives at 90% frequency and least as 11.6 (95% CI: 10.0 – 13.2) species in June when invasives were at 0% frequency (Fig. 2.11). At median frequency of occurrence of invasive graminoids (median =

21.3%), greatest plant richness was predicted as 16.3 (95% CI: 14.9 – 17.7) in August with 0% open water and least at 10.6 (95% CI: 8.7 – 12.4) in June with 95% open water (Fig. 2.12). For entomophilous plant richness, the top model indicated a negative effect of percentage of open water and monotypic cattail cover (Tables 2.10 and 2.12). At median percentage open water

(median = 20%), greatest richness was predicted as 8.3 (95% CI: 7.4 – 9.2) in September with

0% monotypic cattail cover, and least at 4.6 (95% CI: 2.7 – 6.4) in June at 55% cover (Fig. 2.13).

At median percentage of monotypic cattail cover (median = 2.5%), the greatest entomophilous plant richness was predicted as 8.7 (95% CI: 7.8 – 9.6) species in September with 0% open water, least as 4.7 (95% CI: 3.5 – 5.9) in June at 95% open water (Fig. 2.14). The next best competing model for entomophilous plant richness (∆AIC = 0.7, Tables 2.10 and 2.12) included percentage of invasive graminoids and open water. At median percentage of invasive graminoids

(median = 21.3%), greatest entomophilous plant richness was predicted as 8.2 (95% CI: 7.3 –

9.2) in September with 0% open water and least as 4.7 (95% CI: 3.5 – 5.8) in June with 95% open water. At median percentage open water (median = 20%), greatest entomophilous plant richness was predicted as 9.3 (95% CI: 7.8 – 10.7) species per wetland in September with invasives at 90% frequency and least as 5.8 (95% CI: 4.8 – 6.9) species in June when invasives were at 0% frequency. For bee species richness, the top model contained a negative effect of percentage of open water (Tables 2.10 and 2.13). Greatest predicted richness was 7.1 (95% CI:

6.2 – 8.2) species per wetland in August at 0% open water, and lowest as 2.1 (95% CI: 1.5 – 2.9) in September at 95% open water (Fig. 2.15). The top model for frequency of occurrence of entomophilous plants detected that the response varied negatively with percentage of open water and percentage of invasive graminoids (Tables 2.10 and 2.14). At median proportion open water

(median = 20%), greatest frequency was 74.8% (95% CI: 66.5% – 84.0%) in August with 0%

invasives and least was 27.4% (95% CI: 23.0% – 32.7%) in June at 90% invasive plant occurrence (Fig. 2.16). At median percentage of invasive graminoids (median = 21.3%), the greatest predicted frequency was 69.3% (95% CI: 62.0% –77.5%) occurrence in August at 0% open water, the least was 33.3% (95% CI: 28.4% – 39.1%) in June at 95% open water (Fig.

2.17).

DISCUSSION

Wild Bee Diversity & Ecology at Restored Wetlands

In this study, 9,046 specimens in 25 genera were collected, comprising 80 named bee species and an undetermined number of Lasioglossum spp. (Table 2.3). This richness is considerably less than that of most geographically similar surveys of distinct upland habitat types (blueberry fields

– 124, Bushmann and Drummond 2015; agro-ecosystems – 112, Tucker and Rehan 2017b; apple orchards – 104, Russo et al. 2015; forest canopy and understory – 90, Urban-Mead et al. 2021) but rank closely with other wetland studies (≥ 80, Zarrillo and Stoner 2019; 83, Stephenson et al.

2020; but see Vickruck et al. 2019 with 132). However, species richness in the study would increase with identification of Lasioglossum specimens, making such comparisons limited in utility at this time. Although abundances of the most common species detected, including

Agapostemon virescens (7.2%), Apis mellifera (6.4%), and Bombus impatiens (2.4%), are expected for this region, two others – Melissodes bimaculatus (5.7%) and Agapostemon splendens (4.0%) – were collected at unusually high frequencies. Surveys in coastal (Ascher et al. 2014) and urban (Matteson et al. 2008) New York, the White Mountains of New Hampshire

(Tucker and Rehan 2017a), Connecticut (Zarrillo and Stoner 2019), and Maine (Bushmann and

Drummond 2015) collected these two species rarely or not at all. Conversely, common

generalists like Augochlora pura (Say; 0.80%), Halictus ligatus Say (0.77%), and especially

Augochlorella aurata (Smith; 0.13%) were infrequent to rare in this study. This difference is pronounced when compared to Stephenson et al. (2020), where A. aurata made up 47% of individuals in their survey of managed wetlands in the lower Mississippi Alluvial Valley

(LMAV). Whether this assemblage is due to a selection by some species for wetlands or simply emblematic of conditions in this region of New York would require similar surveys in other parts of the state.

Species & Patterns of Note

A survey of trap-nesting Hymenoptera in upland portions of MWC performed by O’Neill and

O’Neill (2010) documented 12 species of bees, 7 of which were megachilids not detected in this study. Almost all megachilids captured in this survey were caught at the wetland edge rather than the basin, and Megachile were regularly observed on upland plants; along with the findings of

O’Neill and O’Neill, these patterns suggest megachilids are more species-rich at MWC than documented here, but no conclusions can be drawn as to whether or why these bees avoid wet areas. Few parasitic species were recorded (0.12%); notably missing were Coelioxys (parasitic on

Megachile) and Bombus (Psithyrus). While representation of megachilids was poor, suitable hosts for cuckoo bumble bees were abundant, thus these species may have escaped detection. A lengthier pollinator survey at MWC would likely continue to document new and potentially rare or scarce species, as is usually the case in bee surveys. Hymenoptera populations tend to fluctuate spatially and temporally, and detection contains an element of chance which can be biased by sampling methodology, resulting in singletons often representing a significant portion of species lists (Williams et al. 2001; Tucker and Rehan 2016; Rhoades et al. 2018; Portman et al. 2020).

The study detected Bombus fervidus (Fabricius; n = 25) and B. borealis Kirby (n = 1), two species of greatest conservation need in New York. B. fervidus has declined substantially in the northeastern U.S. (Colla and Packer 2008; Colla et al. 2012; Jacobson et al. 2018), but appears, according to state surveys, to be persisting in New York in greater numbers than the northern New England states (Bushmann and Drummond 2015; Goldstein and Ascher 2016;

Tucker and Rehan 2017a; NYNHP 2019). This species was collected from all three properties and in seven of the eight sampling periods (Fig. 2.6). B. fervidus was an uncommon, but regular component of the bee fauna in the study wetlands, making particular use of pickerelweed as a floral resource (Fig. 2.8). Given its trend of severe decline in this region, restored wetlands could be valuable supplemental habitat for this species. B. borealis was represented in this study by a single record on monkeyflower (Mimulus ringens L.). Central New York is at the southernmost edge of the range for this species, with most historic records from the Adirondacks, although recent statewide surveys have documented this bee in nearly half of all New York counties

(NYSDEC 2015; NYNHP 2019; E. White, NYSDEC, pers. comm.). Previous studies suggest this species is declining (Colla et al. 2012) or stable (Colla and Packer 2008) but generally low in abundance, and further efforts are needed to assess its status in the northeastern U.S. Given its small sample size, and that our sites were not within its core range, no conclusion can be drawn as to the usefulness of wetlands to this species.

Hylaeus nelumbonis (Robertson) is associated with wetlands but its life history and ecology are still largely unknown. A total of 137 specimens were collected, from wetlands representing all treatments, properties, and sampling periods (Fig. 2.6), and in 31 of 38 sites

(though slightly more common in passive wetlands; Fig. 2.3). Mitchell (1960) lists this species as a specialist on lotus (Nelumbo) and waterlily (Nymphaea), with few published floral records

since (Zarrillo et al. 2016; Gibbs et al. 2017). In this survey H. nelumbonis was documented (n =

38) on 14 plant species (Fig. 2.8, Table 2.7), with most new to literature: pickerelweed (7), hemp dogbane (Apocynum cannabinum L.; 5), blue flag iris (Iris versicolor L.; 5), purple loosestrife

(4), arrowhead (Sagittaria latifolia Willd.; 3), cow/hairy vetch (Vicia cracca/villosa; 3), flowering rush (Butomus umbellatus L.; 2), buttonbush (Cephalanthus occidentalis L.; 2), sulphur cinquefoil (Potentilla recta L.; 2), white waterlily (1), swamp smartweed (1), hedge bedstraw (Galium album Mill.; 1), water horehound (Lycopus americanus Muhl. ex W.Bartram;

1), and monkeyflower (1). Of these, monkeyflower and bedstraw were represented only by males. These novel floral records suggest H. nelumbonis has a broader diet than previously thought. Also of note were two males which resembled those described in Ascher et al. (2014) and Zarrillo et al. (2016), having a black first tergum but lacking basal elevations on the third and fourth sterna which would point to H. schwarzii (Cockerell; not documented in this study).

These specimens were not included in the species total and left as Hylaeus sp.

Honeybees were extremely common, especially in the latter half of the season. NMWMA rents space to local farmers who have multiple hives established in three locations < 5 km apart, within foraging distance of most study sites (Frank Morlock, NYSDEC, pers. comm.).

Honeybees had the greatest number of recorded floral interactions (n = 36) and abundance caught in sweeps (n = 429), and were ranked by PSI (Pollinator Service Index) as the second most important species in the system (PSI = 0.54; Table 2.7). They made up the majority of bee visitors for 9 plant species, notably hemp dogbane (81.4%), boneset (Eupatorium perfoliatum L.;

75.0%), Joe-Pye weed (Eutrochium maculatum [L.]; 62.7%), and nodding bur-marigold (Bidens cernua L.; 61.5%), with the latter having the greatest honeybee abundance on any one plant species (n = 120; Fig. 2.8). It has been proposed that honeybees contribute to wild bee declines

through resource competition which could lead to added stress for social bees, resulting in lower fitness and compromised immune systems (Thomson 2004; Goulson and Sparrow 2009; Goulson et al. 2015). Honeybees may also be able to transmit pathogens and viruses to bumble bees through indirect contact via flowers (Graystock et al. 2015; Alger et al. 2019a, b). Resources consumed by a single honeybee colony in a summer are equivalent to those needed to produce

100,000 solitary bee progeny, and when translated to an average 40-hive apiary, this number becomes 4 million solitary bees not produced in the landscape (Cane and Tepedino 2016). It is unclear if honeybees in this system are excluding native bees from resources, as it was not possible to have hives experimentally removed to measure native bee response. Honeybees can additionally alter dynamics of plant-pollinator interactions, including increased pollination of purple loosestrife, which besides aiding reproduction and spread of this invasive species, can threaten the persistence of other wetland species through pollen contamination (Grabas and

Laverty 1999; Geslin et al. 2017). Indeed, honeybees accounted for nearly 60% of bee visits to purple loosestrife in this survey (Fig. 2.8). Geslin et al. (2017) cite the need for studies comparing plant-pollinator networks before and after the removal of honeybees from a system; given that honeybees were such a large component of plant-pollinator interactions at MWC, which may influence the resiliency and nestedness of the system, such research would be valuable for future management considerations. Studies in other protected areas have found negative effects of honeybee colonies on native bees with increasing colony densities (Shavit et al. 2009; Torné-Noguera et al. 2016); if a goal of land managers is to promote robust and diverse pollinator assemblages, caution should be exercised when choosing to host non-native species on conservation land, as honeybees may deter native species from preferred foodplants and reduce

accessible resources (Walther-Hellwig et al. 2006; Roubik and Villanueva-Gutiérrez 2009;

Hatfield et al. 2018).

Floral Resources

In this survey, 20 diet specialist bee species were documented, including Dufourea novaeangliae

(pickerelweed shortface bee) which is wetland-dependent based on floral host (Table 2.3). Most others were specialized to multiple genera in the Asteraceae, including wetland beggarticks

(Bidens) but upland goldenrods (Solidago) and asters (Symphyotrichum) as well (Fowler and

Droege 2020). Availability of floral resources in restored wetlands increased substantially in the latter half of the growing season, with 80% more wetland flower species in bloom in the late season (Aug-Sep) than the early season (May-June; Fig. 2.7). Interestingly, the flowers shown to support the greatest diversity of bees were a mix of species with long bloom periods (e.g., pickerelweed, swamp smartweed, swamp vervain, Verbena hastata L.) and short ones (e.g., giant goldenrod, nodding bur-marigold), as well as those occurring at high and low frequencies in the landscape (Figs. 2.7 and 2.9, Table 2.8), and in differing hydrological conditions. Some of these may serve to provide season-long, predictable resources for generalist bees, while others offer a pulse of abundant resources for social species and specialists in late summer and autumn.

Species richness of bees was greater on pickerelweed than any other plant (Fig. 2.9,

Table 2.8), with 19 species of bee visitors, despite being detected at only 6 sites, all of which were passive wetlands. Over half (52.9%) of its bee visitors, and nearly half (48.5%) of all its pollinators, were Bombus spp., primarily B. impatiens (n = 77) and B. griseocollis (DeGeer; n =

76; Fig. 2.8). The long blooming period of pickerelweed offered continuous resources to sustain bumble bees through most of the colony cycle, which could be particularly valuable where passively managed (or unmanaged remnant) wetlands exist in a species-poor agricultural matrix

with otherwise sporadic flower availability (Westphal et al. 2009; Rundlöf et al. 2014; Vickruck et al. 2019). Despite the prevalence of honeybees in the study area, they were almost never

(0.88%) collected on pickerelweed, suggesting it is a beneficial plant to manage for to avoid potential honeybee resource competition. The pickerelweed specialist D. novaeangliae was a more frequent visitor to this plant than any other bee (n = 91) and was found at all three properties, though mostly (92.5%) at SMWP. The thriving presence of this wetland-dependent monolege suggests that wetland restoration and management at MWC and SMWP has produced sufficient habitat needed to support rare or specialized pollinators. Although the extremely rare monolege Melissodes apicatus was not recorded, it may have gone undetected, as historic records exist from Seneca and Oswego counties and MWC represents a large swath of remaining and reclaimed habitat for this bee in the state (CUIC 2019, GBIF 2019a). Further targeted surveys may be successful in documenting this species.

Nodding bur-marigold was the most commonly encountered entomophilous plant in full drawdowns, and when present often formed expansive near-monotypic patches reaching up to

95% frequency of occurence (Fig. 2.9, Table 2.9). Although used by 12 species of bees, over

85% of bee visits (70% of all pollinator visits) were A. mellifera and B. impatiens, indicating it is an important food source for late-flying social bees but not attracting the abundance or diversity of solitary bees expected for its presence in the landscape – sympatric Asteraceae such as sneezeweed (Helenium autumnale L.) and giant goldenrod had nearly as many, or more, recorded bee species despite far less frequency of occurrence (Fig. 2.8, Table 2.8). Nodding bur- marigold is likely contributing substantially to honeybee populations as a profuse autumn pollen and nectar source. Removal of honeybee colonies may alter pollination services to this plant, and

could make space for use by smaller bees or even result in changes to the frequency of bur- marigold in the landscape; however, this would need to be tested empirically.

While several species of smartweed (Persicaria spp.) were encountered, their attractiveness to bees differed vastly, with swamp smartweed hosting the greatest richness (17 spp.) and abundance of pollinators (Table 2.8). Unlike other smartweeds in the wetlands, it had a low (< 0.5 m), dense mat-forming growth habit, and when present at a site tended to be abundant

(Schummer et al. 2012). While it was preferred by small bees (Hylaeus, Ceratina,

Lasioglossum), it was also visited by honeybees and many medium to large wasps (e.g.,

Isodontia mexicana [de Saussure] and Dolichovespula arenaria [Fabricius]; Appendices 1 and

3). In contrast, annual smartweed species (P. pensylvanica L., P. lapathifolia L., Polygonum persicaria Gray) received far fewer recorded pollinator visits during this study (Table 2.8).

Annual smartweeds are promoted in moist-soil units by land managers for seed production as food for migratory waterfowl (Fredrickson and Taylor 1982; Strickland et al. 2009; Jasikoff

2013; Eckler et al. 2019), however, the importance of perennial aquatic and semi-aquatic smartweeds as long-blooming nectar sources should not be overlooked.

A notable plant-pollinator association was that of bur-reed (Sparganium spp.), traditionally considered a wind-pollinated graminoid, however in this study it was documented and observed being infrequently visited by Lasioglossum (Dialictus), augochlorine bees, and syrphid flies. Abundant at many sites, it had staggered flowering periods such that while a fraction of plants was in bloom others were setting seed, resulting in flowers present almost all season (Fig. 2.7). Native bees have been known rarely in the literature to visit anemophilous wetland plants, such as Bombus and Andrena on sedges (Carex), and Bombus on rushes (Juncus;

(Pojar 1973; Moisan-Deserres et al. 2014; Saunders 2018). Saunders (2018) found records of

syrphids on bur-reed but none of bees. Facultative entomophily in plants like bur-reed can be easily overlooked in sweep surveys and could be providing supplemental pollen sources to small generalist bees in wetlands when few other flowers are available.

Some entomophilous flowers were surprisingly lacking in pollinators despite high frequencies and substantial search effort. Arrowhead and white waterlily represented two abundant, long-blooming flowers in passive and partial drawdown wetlands (Figs. 2.7 and 2.9,

Table 2.9). Yet, neither the abundance nor diversity of their bee pollinators relative to other flower species reflected this presence in the landscape (Table 2.8). While central New York is outside of the current northernmost range limit of the hibiscus specialist Ptilothrix bombiformis

(Cresson; Sam Droege, USGS, pers. comm.) the near-complete lack of pollinators on swamp rose-mallow (Hibiscus moscheutos L.) was also unexpected given its abundance at MNWR within and outside of study sites (pers. observation). European frogbit (Hydrocharis morsus- ranae L.) and bladderwort (Utricularia vulgaris ssp. machroriza LeConte) are two types of submerged aquatic vegetation (SAV) which have showy flowers and are known to be visited by syrphid flies and sweat bees (Scribailo and Posluszny 1983; Płachno et al. 2018), yet in spite of their often high bloom densities and long-lasting flowering periods in passive wetlands no pollinators were observed or collected. The parasitic vine dodder (Cuscuta sp.) was a target of sweep-netting due to its status as host for the rare Colletes ciliatus Patton, but also had no recorded pollinator visits. Many abundant entomophilous plants in our wetlands that could provide steady resources were, in this survey, underutilized by bees (Table 2.8).

While plant-pollinator interactions were fairly complex (weighted nestedness = 0.58) and spanned 30 plant families, floral visitation and usage of available resources by bees were unevenly distributed given the patterns documented, reflecting the low connectance value (0.08)

for the system. Flower species varied widely in their apparent attractiveness to bees across taxonomy, bloom phenology, and frequency in the landscape. However, this survey indicated a diversity of floral resources exist in restored wetlands for bees, in all management treatments, and particularly in the latter half of the growing season (Fig. 2.7, Table 2.9).

As has been investigated by others (reviewed in Portman et al. 2020), pan traps favored halictid bees (e.g., Lasioglossum and Agapostemon) as well as Melissodes, while a greater diversity of species were collected in sweeps despite much lower frequency of occurrence

(Tables 2.4 and 2.5). Several species common in traps (e.g., A. splendens, M. bimaculatus, M. trinodis Robertson) were rarely or never caught in sweeps of flowers, making it difficult to ascertain their role as pollinators in wetland systems, or confirm that they use wetland habitat at all. This phenomenon highlights a potential shortcoming of this form of sampling methodology; study wetlands were embedded in a diverse matrix of open, forested, and farmed land through which it is presumed bees can move freely, thus bees captured in traps could represent assemblages from multiple habitat types incidentally attracted to traps while foraging or in transit. However, it does suggest that bees from the surrounding landscape are generally willing to enter areas with standing water or dense vegetation to investigate potential food sources.

Influence of Management on Wild Bees and Floral Resources

Efficacy of Drawdown Treatments

Important in the assessment of differences in pollinator assemblages among treatments was to initially evaluate if treatments resulted in different environmental and vegetative characteristics.

Differences between full drawdowns and passive wetlands were detected for mean water depth, mean vegetation height, and percentage of open water, mudflat, and monotypic cattail cover

(Fig. 2.10). Partial drawdowns had intermediate characteristics that overlapped with other treatments. Treatments generally progressed in expected ways; decreases in water levels were detected in passive wetlands due to evaporation, and decreases in percentage of open water in full drawdowns occurred as drawdowns took effect. Vegetation height increased over the course of the growing season, although this difference was only significant in June (Fig. 2.10). Wetlands in the northeast tends to start producing new herbaceous growth later than surrounding upland

(Gleason and Cronquist 1991), with some sites still completely bare in June due to the need for drawdown conditions to stimulate germination of moist-soil annuals (Fredrickson and Taylor

1982). Vegetation rapidly gained height following drawdowns, accounting for the observed differences. These findings are mostly consistent with those of Farley (2020), who performed similar surveys at MWC for vegetation and wetland structure in 2016 – 2018, the years immediately prior to this study. Annual climate (e.g., drought) and site-specific variation (e.g., topography), resulted in variability in management outcomes and overlap among treatments for some environmental characteristics. High summer temperatures often caused partial drawdowns to dry completely, and many passive wetlands naturally assumed a partial drawdown appearance from evaporation, which was also noted by Farley (2020). Differences among wetlands in the timing and speed of prescribed drawdowns (and soil disturbance) can have additional effects on resulting habitat structure (Fredrickson and Taylor 1982, Strickland et al. 2009, Fleming et al.

2012, Gray et al. 2013). These sources of natural and introduced variation can lead to greater habitat diversity, but can also make it more difficult to characterize treatments on a coarse scale.

Management decisions are made annually based on climate, landscape context (missing habitat types), and needs of target taxa (Fredrickson and Taylor 1982, Gray et al. 2013, Eckler et al.

2019). It is important that treatments be monitored throughout the growing season and water

levels altered appropriately, as an adaptive management strategy to ensure expected trajectories are followed and habitat goals are met (Williams and Brown 2016, Farley 2020).

Response of Plant and Bee Assemblages to Treatments and the Wetland Environment

Few to no differences in plant and bee assemblages among treatments were detected by NMDS ordination or linear modeling. The complexities of plant and bee assemblages at three dimensions of the NMDS resulted in high stress levels that precluded straightforward interpretation for management. Sørensen Dissimilarity indices and Shannon-Wiener indices also indicated that bee assemblages and diversity were fairly similar among treatments. These findings corroborate Stephenson et al. (2020) who detected no appreciable difference in bee assemblages of restored wetlands that were actively and passively managed in the LMAV. The phenological turnover of bee species in the landscape accounted for more of the variation in bee assemblage composition than any aspect of the wetland environment. This result is not surprising given observed patterns; most Andrena were documented only in June, followed by abundant

Agapostemon collected June to July and Melissodes in mid-July through late August, with A. mellifera and B. impatiens comprising the majority of September collections (Fig. 2.6). Each treatment had unique bee species (Fig. 2.4); however, almost all were collected in numbers too low to draw conclusions about habitat selection. Only two species were found to statistically select one drawdown type over another; D. novaeangliae was only captured at passively managed sites where pickerelweed was present, and Halictus rubicundus selected for full drawdowns (Fig. 2.3). The majority of frequently encountered plants (e.g., rice cutgrass, narrow- leaved cattail, bur-reed) and entomophilous plants (e.g., purple loosestrife, smartweed spp., white waterlily) were found in varying densities across all treatments (Fig. 2.9, Table 2.9).

Additionally, within-site microhabitat diversity and the persistence of some emergent plants after

drawdowns contributed to vegetative similarity among treatments. Still, many species were more common in some treatments than others, such as arrowhead and nodding bur-marigold (Fig. 2.9), which may change the habitat structure and resources for pollinators and wildlife.

Gross scale metrics that are more easily measurable and applicable to wetland managers, including plant and bee diversity and frequency of entomophilous plant occurrence, were influenced most strongly by open water, invasive graminoid cover, and/or monotypic cattail cover. Assemblages showed negative relationships with predictors except in the case of plant species richness and the second (ΔAIC = 0.7) top model for entomophilous plant richness, where richness actually increased with increasing invasive graminoid cover while decreasing with greater percentage of open water (Tables 2.10-2.12, Figs. 2.11 and 2.12). This outcome was unexpected, given that entomophilous plant richness responded negatively to monotypic cattail in the top-ranking model (Figs. 2.13 and 2.14), and frequency of entomophilous plants responded negatively to invasive graminoid cover (Fig. 2.17). Narrow-leaved cattail, common reed, and reed-canary grass can produce dense monotypic stands that prevent growth of desirable native flora, which can degrade habitat quality for wildlife through alterations to marsh structure and loss of resources (Benoit and Askins 1999; Able and Hagan 2003; Zedler and Kercher 2004;

Chambers et al. 2012, Schummer et al. 2021). However, cattail presence did not necessarily inhibit flowering plants in this study; despite 95% of recorded cattail being narrow-leaved or hybrid cattail, only 41% of survey points with cattail were monotypic. Sites frequently had entomophilous plants (e.g., swamp milkweed, Asclepias incarnata L., bulb-bearing water hemlock, Cicuta bulbifera [L.], water horehound) intermixed with narrow-leaved cattail. One possibility for the positive relationships is that the hydrological conditions that promote reed and reed-canary grass, which at MWC tended to be moist soil or shallow standing water, are also

favored by a variety of other wetland plants, whereas monotypic cattail was primarily found in deep water habitats shared by few other species (and by definition excluded other plants from survey points). Open water was a negative predictor for entomophilous plant richness and frequency (Figs. 2.13 and 2.16), and for bee richness (Fig. 2.15), despite some abundant entomophilous SAV like frogbit and bladderwort being included in this metric. Traps in open water isolated from edge habitat captured few bees, but traps in standing water with vegetation – even dense cattail – were visited (pers. observation). Although the foraging distances of bees differ greatly based on body size (Greenleaf et al. 2007; Zurbuchen et al. 2010), it is reasonable to speculate that most bees may be averse to exploring expansive open water with little emergent vegetation to safely rest or forage on.

This study provides evidence that restored wetlands with greater frequencies of invasive graminoids have a lower frequency of floral resources for bees, particularly when combined with open water, such as a half-vegetated marsh dominated by cattail and common reed. Half- vegetated marshes can be highly beneficial to breeding marsh birds and mammals (Bishop et al.

1979; Kaminski and Prince 1981; Murkin et al. 1982), thus are a valuable component of protected wetland areas. However, locating them in proximity to other drawdown types, or managing for flower-rich wet edges could improve suitability of the landscape for bees.

Reducing densities of invasive plant species opens more space for native entomophilous plants, which often double as sources of seeds, nectar, nesting material, or shelter for birds and other wildlife (Fredrickson and Taylor 1982, Strickland et al. 2009). While few studies have assessed the response of wild bees to removal of invasive plants, the work of Fiedler et al. (2011) showed gains in bee diversity and changes in assemblage composition following removal of glossy buckthorn (Frangula alnus Mill.) from fens, making invasive species control a promising means

for improving habitat quality for pollinators in wetlands. Schummer et al. (2021) also found lower aquatic macroinvertebrate densities with greater open water and invasive plant cover at

MWC, suggesting that multiple invertebrate guilds could benefit from efforts to control invasive graminoids and maintain a diverse wetland habitat matrix.

Given that few differences were found in bee assemblages among treatments, and that most species were habitat and diet generalists (Table 2.3), results suggest that wetlands hold value as complementary habitat for pollinators that opportunistically use wetland floral resources. Restored wetlands also host scarce or specialized species that rely on this habitat type for parts of their life histories, making them indispensable for bees not targeted by traditional pollinator habitat creation methods (i.e., gardens, wildflower strips, meadows). Bees and entomophilous plants responded to specific environmental conditions rather than broad treatment categorizations, and wetlands with greater open water, monotypic cattail cover, and invasive graminoids can expect fewer floral resources for pollinators. Continued control of invasive plant species will likely improve habitat quality for target and non-target species. Locating wetlands undergoing different types of management together on the landscape will offer a diverse mix of plant families and bloom turnover to support pollinators. In particular, actively managed wetlands likely fill a phenological gap in the later part of the growing season (Fig. 2.7), offering a large pulse of pollen and nectar sources after farm crops, forest ephemerals, and many field and roadside plants have ceased blooming. However, floral resources were present in lower densities throughout the growing season in all treatments, with several key long-blooming species (e.g., white waterlily, pickerelweed, bur-reed) in wetlands with standing water. The frequency of some species can be increased without compromising waterfowl production goals if incorporated into the planning for vegetation and soil disturbance in actively managed wetlands (Strickland et al.

2009; Stephenson et al. 2020). When wetlands, especially restored wetlands, are located in an agriculture-dominated landscape, they can offer substantial benefits as refuges and supplemental sources of pollen and nectar for local pollinators (Evans et al. 2018, Heneberg et al. 2018,

Vickruck et al. 2019, Begosh et al. 2020). Restoration of wetlands from agriculture at MWC and

SMWP has added floral resources back into the landscape that are being utilized by over 80 species of bees, lending support to the idea that wetland restoration can be an important tool for pollinator conservation in modified landscapes. Current management at MWC creates a matrix of wetland habitats that produce diverse floral resources and sustain a rich bee fauna.

Management should seek to maintain this landscape heterogeneity, or increase it when possible through the further reclamation of agricultural land, so the needs of pollinators and vertebrate wildlife are met simultaneously.

FIGURES AND TABLES

Figure 2.1. GIS map of study sites. Property boundaries of state and federal protected lands are in color. Water bodies are blue.

Table 2.1. Study sites and treatments. Those labeled ‘n/a’ were not sampled in that year.

Latitude Longitude Treatment Wetland Property (°N) (°W) 2019 2020 Arum 43.084809 -76.670052 Howland Island (NMWMA) Full Full Breeder 43.086738 -76.67568 Howland Island (NMWMA) Full Full Brooder 43.079874 -76.676026 Howland Island (NMWMA) Partial Passive CCC 43.078453 -76.671453 Howland Island (NMWMA) Passive Passive Cook Pond 43.093032 -76.674077 Howland Island (NMWMA) Partial Partial Coot Pond 43.069119 -76.672581 Howland Island (NMWMA) Passive Passive Gander 43.078854 -76.667678 Howland Island (NMWMA) Passive Passive Goose 43.081846 -76.670174 Howland Island (NMWMA) Passive Passive Headquarters Pond 43.07637 -76.677913 Howland Island (NMWMA) Passive Passive Loosestrife 43.069864 -76.68096 Howland Island (NMWMA) Partial Full Lost Pond 43.074113 -76.667126 Howland Island (NMWMA) Partial Passive Carncross 43.07709 -76.710335 NMWMA n/a Full Colvin Marsh 43.087823 -76.759967 NMWMA Passive Passive Deep Muck 43.07369 -76.723074 NMWMA Passive Passive Guy's Marsh 43.068753 -76.74105 NMWMA Passive Passive Loop Rd. Flats 43.025869 -76.709511 NMWMA n/a Full M&M 43.058076 -76.726409 NMWMA Full Full Malone Marsh 43.075124 -76.744851 NMWMA Passive Passive Martens Tract 43.086218 -76.702692 NMWMA Full Passive Mitigation Marsh 43.073667 -76.717763 NMWMA Passive Passive Muckrace Flats 43.072632 -76.739588 NMWMA Partial Passive Mulligan 43.057614 -76.720628 NMWMA Passive n/a Teal Pond 43.086039 -76.708101 NMWMA n/a Passive Torrey Marsh 43.062369 -76.734227 NMWMA Passive Full Unit 1 43.069981 -76.715483 NMWMA Full Full Unit 2 43.065776 -76.718004 NMWMA Full Passive Eaton Marsh 42.989611 -76.73505 MNWR Full n/a Knox-Marsellus Marsh 43.011276 -76.752952 MNWR Partial Full LaRues Lagoon 42.972346 -76.737876 MNWR Full n/a Main Pool 42.973583 -76.74426 MNWR Full Passive Millennium Marsh 42.981653 -76.76864 MNWR Passive n/a Puddler Marsh 43.007974 -76.742951 MNWR n/a Passive Sandhill Crane Unit 43.044989 -76.732591 MNWR n/a Full Seneca Flats 42.975202 -76.736989 MNWR Passive Passive Stoll Marsh 43.012165 -76.746221 MNWR Full n/a "BFP" 42.929153 -76.820746 SMWP Passive Passive "Pickerel" 42.942279 -76.818986 SMWP Passive Passive "Snapper" 42.932579 -76.818885 SMWP Passive Passive

Figure 2.2. An example of a pan trap setup used in bee surveys. Bowls are adhered to shelf and stake using Velcro and rubber bands, to prevent tipping from wind or animal interference.

Table 2.2. Pearson’s Correlation coefficients for predictor variables used in mixed models.

PointsEntoPlants

EntoRich 0.397

BeeRich 0.178 0.006

AvgPointRich 0.176 0.658 0.511

VegRich 0.688 0.229 0.878 0.185

Invasives 0.448 0.251 0.079 0.269 -0.320

0.579 0.011 -0.171 0.036 -0.083 -0.423

MonoCat PointsCat 0.778 0.796 0.282 0.163 0.096 0.117 -0.262

PointsMud -0.130 -0.106 -0.021 -0.034 -0.257 0.116 -0.091 -0.324

PointsOpen -0.289 -0.208 -0.016 -0.399 -0.504 -0.591 -0.176 -0.444 -0.153

MedianWaterDepth 0.791 -0.338 -0.130 0.061 -0.331 -0.516 -0.479 -0.177 -0.424 0.045

AvgWaterDepth 0.975 0.811 -0.371 -0.107 0.073 -0.312 -0.503 -0.479 -0.169 -0.414 0.052

AvgVegHeight -0.565 -0.580 -0.696 -0.190 0.618 0.325 0.699 0.584 0.658 0.078 0.491 0.126

Year 0.114 -0.186 -0.176 -0.051 0.155 -0.004 -0.167 -0.049 0.274 0.265 0.205 0.263 0.161

Year AvgVegHeight AvgWaterDepth MedianWaterDepth PointsOpen PointsMud PointsCat MonoCat Invasives VegRich AvgPointRich BeeRich EntoRich PointsEntoPlants

Year AVH AWD MWD PO PM PC MC Inv VR BR ER PEP Year 0.114 -0.186 -0.176 -0.051 0.155 -0.004 -0.167 -0.049 0.274 0.205 0.263 0.161 AVH -0.565 -0.580 -0.696 -0.190 0.618 0.325 0.699 0.584 0.078 0.491 0.126 AWD 0.975 0.811 -0.371 -0.107 0.073 -0.312 -0.503 -0.169 -0.414 0.052 MWD 0.791 -0.338 -0.130 0.061 -0.331 -0.516 -0.177 -0.424 0.045 PO -0.289 -0.208 -0.016 -0.399 -0.504 -0.176 -0.444 -0.153 PM -0.130 -0.106 -0.021 -0.034 0.116 -0.091 -0.324 PC 0.778 0.796 0.282 0.096 0.117 -0.262 MC 0.579 0.011 0.036 -0.083 -0.423 Inv 0.448 0.079 0.269 -0.320 VR 0.229 0.878 0.185 BR 0.178 0.006 ER 0.397 PEP AVH = average vegetation height MC = points with monotypic cattail AWD = average water depth Inv = points with invasive graminoids MWD = median water depth VR = plant species richness PO = points with open water BR = bee species richness PM = points with mudflat ER = entomophilous plant species richness PC = points with cattail PEP = points with entomophilous plants

Table 2.3. List of bee species collected in restored wetlands in central New York in 2019 – 2020, with relevant life history information. Species denoted with † are non-native. Pollen specificity: P = polylectic, O = oligolectic, M = monolectic. Nesting habit: G = ground, C = cavity, K = kleptoparasite.

Species No. Caught Pollen Specificity Nesting Habit Andrenidae Andrena (Callandrena s.l.) helianthi (Robertson 1891) 1 M G Andrena (Callandrena s.l.) placata Mitchell 1960 4 O G Andrena (Callandrena s.l.) simplex Smith 1853 4 O G Andrena (Cnemidandrena) hirticincta Provancher 1888 1 O G Andrena (Gonandrena) fragilis Smith 1853 2 M G Andrena (Holandrena) cressonii Robertson 1891 1 P G Andrena (Melandrena) commoda Smith 1879 2 P G Andrena (Melandrena) pruni Robertson 1891 1 P G Andrena (Melandrena) vicina Smith 1853 4 P G Andrena (Plastandrena) crataegi Robertson 1893 2 P G Andrena (Scrapteropsis) alleghaniensis Viereck 1907 1 P G Andrena (Scrapteropsis) ilicis Mitchell 1960 1 P G Andrena (Taeniandrena) wilkella (Kirby 1802)† 3 O G Andrena (Trachandrena) mariae (Robertson 1891) 2 M G Andrena (Trachandrena) miranda Smith 1879 1 P G Andrena (Trachandrena) nuda Robertson 1891 3 P G Perdita (Perdita) octomaculata (Say 1824) 1 O G Pseudopanurgus andrenoides (Smith 1853) 6 O G Apidae Anthophora (Clisodon) terminalis Cresson 1869 20 P C Anthophora (Melea) bomboides Kirby 1837 2 P G Apis (Apis) mellifera Linnaeus 1758† 579 P C Bombus (Cullumanobombus) griseocollis (DeGeer 1773) 166 P G Bombus (Cullumanobombus) rufocinctus Cresson 1863 3 P G

Bombus (Pyrobombus) bimaculatus Cresson 1863 37 P G Bombus (Pyrobombus) impatiens Cresson 1863 222 P G Bombus (Pyrobombus) sandersoni Franklin 1913 4 P G Bombus (Pyrobombus) vagans Smith 1854 4 P G Bombus vagans/sandersoni 2 - G Bombus (Subterraneobombus) borealis Kirby 1837 1 P G Bombus (Thoracobombus) fervidus (Fabricius 1798) 25 P G Ceratina (Zadontomerus) calcarata Robertson 1900 29 P C Ceratina (Zadontomerus) dupla Say 1837 66 P C Ceratina (Zadontomerus) mikmaqi Rehan & Sheffield 2011 16 P C Ceratina spp. 3 - C Eucera (Peponapis) pruinosa (Say 1837) 5 M G Melissodes (Eumelissodes) agilis Cresson 1878 23 O G Melissodes (Eumelissodes) dentiventris Smith 1854 5 O G Melissodes (Eumelissodes) druriellus (Kirby 1802) 18 O G Melissodes (Eumelissodes) trinodis Robertson 1901 91 O G Melissodes (Heliomelissodes) desponsus Smith 1854 21 O G Melissodes (Melissodes) bimaculatus (Lepeletier 1825) 516 P G Melissodes spp. 30 - G Nomada spp. 8 - K Triepeolus pectoralis (Robertson 1897) 1 P K Xylocopa (Xylocopoides) virginica (Linnaeus 1771) 24 P C Colletidae Colletes compactus Cresson 1868 1 O G Colletes inequalis Say 1837 1 P G Colletes latitarsis Robertson 1891 5 M G Colletes simulans Cresson 1868 4 O G Hylaeus (Hylaeus) annulatus (Linnaeus 1758) 4 P C Hylaeus (Hylaeus) mesillae (Cockerell 1896) 9 P C

Hylaeus (Prosopis) affinis (Smith 1853) 26 P C Hylaeus (Prosopis) illinoisensis (Robertson 1896) 21 P C Hylaeus (Prosopis) modestus Say 1837 15 P C Hylaeus (Prosopis) nelumbonis (Robertson 1890) 137 P C Hylaeus spp. 34 - C Halictidae Agapostemon (Agapostemon) sericeus (Förster 1771) 43 P G Agapostemon (Agapostemon) splendens (Lepeletier 1841) 360 P G Agapostemon (Agapostemon) virescens (Fabricius 1775) 654 P G Augochlora (Augochlora) pura (Say 1837) 72 P C Augochlorella aurata (Smith 1853) 12 P G Dufourea novaeangliae (Robertson 1897) 93 M G Halictus (Nealictus) parallelus Say 1837 5 P G Halictus (Odontalictus) ligatus Say 1837 70 P G Halictus (Protohalictus) rubicundus (Christ 1791) 51 P G Halictus (Seladonia) confusus Smith 1853 16 P G Lasioglossum (Dialictus) coeruleum (Robertson 1893) 6 P C Lasioglossum (Dialictus) nigroviride (Graenicher 1911) 3 P C Lasioglossum (Leuchalictus) leucozonium (Schrank 1781)† 1 P G Lasioglossum (Sphecodogastra) truncatum (Robertson 1901) 2 P G Lasioglossum spp. 5412 - - Sphecodes coronus Mitchell 1956 1 P K Sphecodes sp. 1 - K Megachilidae Heriades (Neotrypetes) carinata Cresson 1864 2 P C Hoplitis (Alcidamea) pilosifrons (Cresson 1864) 3 P C Megachile (Callomegachile) sculpturalis Smith 1853 1 P C Megachile (Litomegachile) mendica Cresson 1878 7 P C Megachile (Megachile) montivaga Cresson 1878 3 P G Megachile (Megachile) relativa Cresson 1878 2 P C

Megachile (Sayapis) pugnata Say 1837 1 O C Megachile (Xanthosarus) latimanus Say 1823 3 P C Osmia (Diceratosmia) conjuncta Cresson 1864 2 P C Osmia (Melanosmia) atriventris Cresson 1864 1 P C Osmia (Melanosmia) bucephala Cresson 1864 1 P C

Table 2.4. Bee families collected by pan-trapping and sweep-netting methods.

Family Pan Trap Sweep Total Andrenidae 7 (17.5%) 33 (82.5%) 40 Apidae 957 (49.8%) 964 (50.2%) 1921 Colletidae 150 (58.4%) 107 (41.6%) 257

Halictidae 6476 (95.2%) 326 (4.8%) 6802 Megachilidae 9 (34.6%) 17 (65.4%) 26

Table 2.5. Bee genera collected by pan-trapping and sweep-netting methods.

Genus Pan Trap Sweep Total

Agapostemon 1038 (98%) 19 (2%) 1057 Andrena 7 (21%) 26 (79%) 33 Anthophora 19 (86%) 3 (14%) 22 Apis 150 (26%) 429 (74%) 579 Augochlora 58 (81%) 14 (19%) 72 Augochlorella 8 (67%) 4 (33%) 12 Bombus 49 (11%) 415 (89%) 464 Ceratina 50 (44%) 64 (56%) 114 Colletes 2 (18%) 9 (82%) 11 Dufourea 2 (2%) 91 (98%) 93

Eucera 5 (100%) 0 5 Halictus 131 (92%) 11 (8%) 142 Heriades 0 2 (100%) 2

Hoplitis 2 (67%) 1 (33%) 3 Hylaeus 148 (60%) 98 (40%) 246 Lasioglossum 5237 (97%) 187 (3%) 5424

Megachile 3 (18%) 14 (82%) 17 Melissodes 679 (96%) 25 (4%) 704 Nomada 2 (25%) 6 (75%) 8 Osmia 4 (100%) 0 4 Perdita 0 1 (100%) 1 Pseudopanurgus 0 6 (100%) 6 Sphecodes 2 (100%) 0 2 Triepeolus 0 1 (100%) 1 Xylocopa 3 (12%) 21 (88%) 24

Table 2.6. Most common bee species per treatment, with raw and relative abundances. Genus Lasioglossum not shown.

Full Drawdown # Collected % Total Catch per Treatment Agapostemon virescens 301 10.7

Apis mellifera 224 8.0 Melissodes bimaculatus 163 5.8 Agapostemon splendens 66 2.3

Bombus griseocollis 61 2.2

Partial Drawdown A. mellifera 77 7.2

A. virescens 70 6.5 M. bimaculatus 47 4.4 Ceratina dupla 23 2.1 Bombus impatiens 19 1.8

Passive Management M. bimaculatus 306 6.0

A. splendens 287 5.6 A. virescens 281 5.5 A. mellifera 263 5.1 B. impatiens 156 3.1

Figure 2.3. Most common (n > 10) bee species (excluding Lasioglossum) collected in restored wetlands in central New York and the percentage they represent of each treatment’s total catch, arranged by family.

Figure 2.4. Venn diagram portraying species unique to treatments and shared among treatments.

Figure 2.5. Sample-based species accumulation curve for each treatment and for the study system as a whole.

Table 2.7. Bee species collected in sweeps of flowers in restored wetlands in central New York, with species-level plant-pollinator interaction statistics. PSI = Pollinator Service Index.

Species No. Caught Floral Interactions (Degree) Normalized Degree PSI Andrenidae Andrena alleghaniensis 1 1 0.02 0.33 Andrena crataegi 2 1 0.02 0.11 Andrena fragilis 1 1 0.02 0.07 Andrena helianthi 1 1 0.02 0.17 Andrena hirticincta 1 1 0.02 0.02 Andrena ilicis 1 1 0.02 0.06 Andrena mariae 2 1 0.02 0.11 Andrena miranda 1 1 0.02 0.01 Andrena nuda 3 1 0.02 0.17 Andrena placata 3 1 0.02 0.07 Andrena simplex 4 1 0.02 0.10 Andrena vicina 4 2 0.03 0.21 Andrena wilkella 2 2 0.03 0.13 Perdita octomaculata 1 1 0.02 0.02 Pseudopanurgus andrenoides 6 1 0.02 0.15 Apidae Anthophora terminalis 3 3 0.05 0.35 Apis mellifera 429 36 0.60 0.54 Bombus bimaculatus 32 6 0.10 0.10 Bombus borealis 1 1 0.02 0.17 Bombus fervidus 9 4 0.07 0.03 Bombus griseocollis 154 14 0.23 0.30 Bombus impatiens 206 26 0.43 0.24 Bombus rufocinctus 3 3 0.05 0.03 Bombus sandersoni 4 4 0.07 0.38

Bombus vagans 4 3 0.05 0.10 Bombus vagans/sandersoni 2 2 0.03 0.04 Ceratina calcarata 19 7 0.12 0.12 Ceratina dupla 32 11 0.18 0.13 Ceratina mikmaqi 11 7 0.12 0.13 Ceratina sp. 2 1 0.02 0.05 Melissodes agilis 1 1 0.02 0.00 Melissodes bimaculatus 7 3 0.05 0.30 Melissodes druriellus 12 3 0.05 0.11 Melissodes spp. 5 3 0.05 0.04 Nomada spp. 6 1 0.02 0.19 Triepeolus pectoralis 1 1 0.02 0.02 Xylocopa virginica 21 8 0.13 0.13 Colletidae Colletes latitarsis 5 1 0.02 1.00 Colletes simulans 4 1 0.02 0.10 Hylaeus affinis 10 5 0.08 0.30 Hylaeus annulatus 1 1 0.02 0.01 Hylaeus illinoisensis 14 8 0.13 0.14 Hylaeus mesillae 9 4 0.07 0.15 Hylaeus modestus 6 3 0.05 0.11 Hylaeus nelumbonis 38 14 0.23 0.21 Hylaeus spp. 20 11 0.18 0.18 Halictidae Agapostemon sericeus 10 7 0.12 0.13 Agapostemon splendens 4 3 0.05 0.11 Agapostemon virescens 5 4 0.07 0.17 Augochlora pura 14 8 0.13 0.06 Augochlorella aurata 4 3 0.05 0.07 Dufourea novaeangliae 91 1 0.02 0.27

Halictus confusus 4 4 0.07 0.07 Halictus ligatus 5 3 0.05 0.13 Halictus rubicundus 2 2 0.03 0.02 Lasioglossum coeruleum 1 1 0.02 0.01 Lasioglossum truncatum 2 2 0.03 0.17 Lasioglossum spp. 184 32 0.53 0.36 Megachilidae Heriades carinata 2 2 0.03 0.07 Hoplitis pilosifrons 1 1 0.02 0.10 Megachile latimanus 2 1 0.02 0.20 Megachile mendica 7 5 0.08 0.07 Megachile montivaga 2 1 0.02 0.22 Megachile relativa 2 2 0.03 0.02 Megachile sculpturalis 1 1 0.02 0.02

Table 2.8. Flowering plant species swept in restored wetlands in central New York, with Wetland Indicator Status and species-level plant-pollinator interaction statistics. Species denoted with † are non-native.

Successful Bee species Bee sampling Normalized Family Flowering plant species Common name WIS associations abundance events degree Adoxaceae Sambucus canadensis black elderberry FACW 3 3 1 0.05 Alismataceae Alisma subcordatum water plantain OBL 2 3 2 0.03 Sagittaria latifolia arrowhead OBL 5 17 5 0.08 Apocynaceae Apocynum cannabinum hemp dogbane FAC 5 43 3 0.08 Asclepias incarnata swamp milkweed OBL 7 60 6 0.11 Asclepias syriaca common milkweed UPL 6 37 3 0.09 Asteraceae Bidens cernua nodding bur-marigold OBL 12 195 12 0.18 Bidens frondosa devil's beggarticks FACW 4 6 3 0.06 Bidens tripartita ssp. comosa three-lobed beggarticks FACW 3 3 1 0.05 Cirsium arvense† creeping thistle FACU 1 1 1 0.02 Erigeron philadelphicus daisy fleabane FAC 10 31 4 0.15 Eupatorium perfoliatum common boneset FACW 6 36 2 0.09 Eutrochium maculatum spotted Joe-pye weed OBL 6 51 3 0.09 Helenium autumnale sneezeweed FACW 10 41 3 0.15 Pilosella caespitosa† yellow hawkweed NI 1 3 1 0.02 Solidago gigantea giant goldenrod FACW 16 41 2 0.26 Balsaminaceae Impatiens capensis spotted jewelweed FACW 8 31 4 0.12 Brassicaceae Brassica rapa† wild mustard UPL 1 1 1 0.02 Butomaceae Butomus umbellatus† flowering rush OBL 3 25 2 0.05 Caprifoliaceae Dipsacus fullonum† teasel FACU 1 1 1 0.02 Caryophyllaceae Silene vulgaris† bladder campion NI 2 3 1 0.03 Convolvulaceae Convolvulus arvensis† field bindweed NI 3 6 1 0.05 Cuscuta sp. dodder - 0 0 0 - Cornaceae Cornus amomum silky dogwood FACW 1 1 1 0.02

Cornus racemosa gray dogwood FAC 7 14 2 0.11 Cucurbitaceae Echinocystis lobata wild cucumber FACW 0 0 0 - Fabaceae Desmodium canadense Canada tick-trefoil FAC 4 10 1 0.06 Trifolium repens† white clover FACU 7 19 2 0.11 Vicia cracca/villosa† cow/hairy vetch NI 5 10 1 0.08 Hydrocharitaceae Hydrocharis morsus-ranae† European frogbit OBL 0 0 0 - Iridaceae Iris pseudacorus† yellow flag iris OBL 3 4 2 0.05 Iris versicolor blue flag iris OBL 4 9 3 0.06 Lamiaceae Lycopus americanus water horehound OBL 4 8 2 0.06 Mentha arvensis wild mint FACW 2 15 1 0.03 Monarda fistulosa wild bergamot FACU 4 8 1 0.06 Stachys palustris† marsh woundwort OBL 6 13 1 0.09 Lentibulariaceae Utricularia vulgaris ssp. macrorhiza bladderwort OBL 0 0 0 - Lythraceae Decodon verticillatus swamp loosestrife OBL 5 7 1 0.08 Lythrum salicaria† purple loosestrife OBL 11 68 6 0.15 Malvaceae Abutilon theophrasti† velvetleaf FACU 5 9 1 0.08 Hibiscus moscheutos swamp rose mallow OBL 2 2 2 0.03 Nymphaeaceae Nymphaea odorata white waterlily OBL 3 28 5 0.05 Orobanchaceae Agalinis tenuifolia common gerardia FACW 2 8 2 0.05 Penthoraceae Penthorum sedoides ditch stonecrop OBL 3 4 1 0.05 Phrymaceae Mimulus ringens Allegheny monkeyflower OBL 3 6 2 0.05 Plantaginaceae Chelone glabra white turtlehead OBL 3 3 1 0.05 Penstemon digitalis foxglove beardtongue FAC 1 1 1 0.02 Polygonaceae Persicaria amphibia water smartweed OBL 3 4 1 0.05 Persicaria hydropiperoides swamp smartweed OBL 17 80 7 0.28 Persicaria lapathifolia pale smartweed FACW 1 1 1 0.02 Persicaria pensylvanica Pennsylvania smartweed FACW 3 10 2 0.06 Persicaria sagittata arrow-leaved tearthumb OBL 0 0 0 - Pontederiaceae Pontederia cordata pickerelweed OBL 19 342 13 0.29

Primulaceae Lysimachia nummularia† moneywort FACW 0 0 0 -

Lysimachia terrestris swamp candles OBL 2 3 1 0.03 Ranunculaceae Anemone canadensis Canada anemone FACW 3 5 3 0.05 Clematis virginiana Virgin's bower FAC 3 3 1 0.05 Ranunculus sp. buttercup - 3 3 3 0.05 Rosaceae Potentilla recta† sulphur cinquefoil NI 1 2 1 0.02 Rosa multiflora† multiflora rose FACU 10 18 3 0.15 Rubiaceae Cephalanthus occidentalis buttonbush OBL 7 38 2 0.11 Galium album† hedge bedstraw FACU 4 6 2 0.08 Galium sp. bedstraw - 1 1 1 0.02 Solanaceae Physalis longifolia longleaf ground cherry NI 1 5 1 0.02 Typhaceae Sparganium sp. bur-reed OBL 1 2 1 0.02 Verbenaceae Verbena hastata swamp vervain FACW 13 39 7 0.20

Figure 2.6. Flight season of bees collected in restored wetlands in central New York, derived from specimen data and delimited by week, 20-May to 22-Sep.

Figure 2.7. Bloom times of flowering plants from study sites, derived from observational data and sweeps from 2019 – 2020. Time periods are relative and not directly comparable to bee phenology.

Figure 2.8. Plant-pollinator interaction network for ≥ 64 bee and 60 flowering plant species in central New York wetlands, derived from sweep data. Width of bars are relative to the number of individuals captured on a flower species (number of interactions).

Figure 2.9. Most common (n > 100) plant taxa in restored wetlands in central New York and the percentage of points per treatment each was detected at, arranged by functional group (e.g., graminoids, submerged aquatic vegetation, emergent).

Table 2.9. Most commonly encountered entomophilous plant species per treatment, with raw and relative abundances.

Full Drawdown # of Survey Points Present % Survey Points Present Bidens cernua 388 12.1 Persicaria pensylvanica 317 9.9 Lythrum salicaria 289 9.0 Bidens frondosa 186 5.8 Sparganium spp. 183 5.7

Partial Drawdown Nymphaea odorata 220 19.7 Sparganium spp. 110 9.8 Sagittaria latifolia 109 9.7 L. salicaria 95 8.5 P. pensylvanica 64 5.7

Passive Management N. odorata 854 13.7 L. salicaria 543 8.7 S. latifolia 475 7.6 Sparganium spp. 350 5.6 Persicaria hydropiperoides 286 4.6

Figure 2.10. Model predictions for differences in environmental characteristics among treatments (bars indicate 95% confidence intervals). Different letters denote significance with Tukey HSD ⍺ = 0.05. Those in color had a significant overall effect of month and are separated by both treatment and month; significance of overall month effects are given, color-coded, in the top left of each panel. Within-treatment month effects are shown with the predicted values, and are not applicable between treatments.

Table 2.10. Model estimates, standard error, and 95% confidence intervals for the top-ranked model(s) for each parameter of interest.

Response Parameter β SE 95% CI Bee Richness Intercept -345.5 0.982 -519.966 to -171.011 Points with Open Water -0.016 0.004 -0.024 to -0.008 Year 0.172 0.000 0.086 to 0.258 Month (June) -0.076 0.077 -0.229 to 0.084 (July) -0.075 0.073 -0.222 to 0.074 (September) -0.610 0.105 -0.827 to -0.394 Plant Richness Intercept -6741.0 747.0 -8210.698 to -5269.025 Points with Invasive Graminoids 0.189 0.042 0.106 to 0.271 Points with Open Water -0.087 0.027 -0.142 to -0.032 Year 3.345 0.370 2.616 to 4.073 Month (June) -2.425 0.507 -3.421 to -1.425 (July) -0.038 0.488 -0.999 to 0.923 (September) -0.149 0.487 -1.108 to 0.810 Entomophilous Plant Richness Intercept -3581.0 503.7 -4571.858 to -2589.584 Points with Monotypic Cattail Cover -0.103 0.046 -0.194 to -0.011 Points with Open Water -0.069 0.017 -0.102 to -0.035 Year 1.778 0.249 1.286 to 2.268 Month (June) -1.247 0.327 -1.890 to -0.603 (July) -0.140 0.310 -0.749 to 0.469 (September) 0.185 0.308 -0.420 to 0.791 Intercept -3939.0 476.4 -4876.62 to -2999.02 Points with Invasive Graminoids 0.055 0.027 0.002 to 0.108 Points with Open Water -0.055 0.018 -0.090 to 0.020 Year 1.954 0.236 1.489 to 2.419 Month

(June) -1.338 0.323 -1.974 to -0.701 (July) -0.172 0.311 -0.785 to 0.441 (September) 0.095 0.311 -0.517 to 0.706 Survey Points with Entomophilous Plants Intercept -205.800 102.100 -405.755 to -5.880 Points with Invasive Graminoids -0.021 0.003 -0.026 to -0.015 Points with Open Water -0.012 0.003 -0.018 to -0.007 Year 0.104 0.051 0.005 to 0.203 Month (June) -0.264 0.073 -0.407 to -0.122 (July) -0.039 0.070 -0.177 to 0.099 (September) -0.030 0.071 -0.168 to 0.108

Table 2.11. Model rankings for predictors of plant species richness in central New York wetlands.

Rank Model AICc ΔAICc wa 1 INVASIVE+POINTS_OPEN+YEAR+MONTH 1405.2 0.0 0.98 2 INVASIVE+YEAR+MONTH 1412.7 7.5 0.02 3 MONO_CAT+POINTS_OPEN+YEAR+MONTH 1419.0 13.8 0.00 4 POINTS_OPEN+YEAR+MONTH 1422.0 16.8 0.00 5 AVG_WATER_DEPTH_CM+YEAR+MONTH 1426.3 21.1 0.00 6 POINTS_CAT+YEAR+MONTH 1431.1 25.9 0.00 7 MONO_CAT+YEAR+MONTH 1437.0 31.8 0.00 8 TRMT+YEAR+MONTH 1441.0 35.8 0.00 9 INVASIVE+YEAR 1445.9 40.7 0.00 10 POINTS_OPEN+YEAR 1449.8 44.6 0.00 11 AVG_WATER_DEPTH_CM+YEAR 1458.9 53.7 0.00 12 MONO_CAT+YEAR 1478.8 73.6 0.00 13 TRMT+YEAR 1487.8 82.6 0.00

Figure 2.11. Regression for fixed effect of invasive graminoid cover on plant species richness with open water held at median value, presented with 95% confidence intervals.

Figure 2.12. Regression for fixed effect of open water on plant species richness with invasive cover held at median value, presented with 95% confidence intervals. 80

Table 2.12. Model rankings for predictors of entomophilous plant species richness in central New York wetands.

Rank Model AICc ΔAICc wa 1 MONO_CAT+POINTS_OPEN+YEAR+MONTH 1167.6 0.0 0.44 2 INVASIVE+POINTS_OPEN+YEAR+MONTH 1168.3 0.7 0.31 3 POINTS_OPEN+YEAR+MONTH 1170.3 2.7 0.11 4 MONO_CAT+AVG_WATER_DEPTH_CM+YEAR+MONTH 1170.7 3.1 0.09 5 AVG_WATER_DEPTH_CM+YEAR+MONTH 1172.4 4.8 0.04 6 INVASIVE+YEAR+MONTH 1175.8 8.2 0.01 7 MONO_CAT+YEAR+MONTH 1181.3 13.7 0.00 8 POINTS_CAT+YEAR+MONTH 1183.9 16.3 0.00 9 TRMT+YEAR+MONTH 1185.2 17.6 0.00 10 POINTS_OPEN+YEAR 1189.8 22.2 0.00 11 AVG_WATER_DEPTH_CM+YEAR 1194.9 27.3 0.00 12 INVASIVE+YEAR 1201.5 33.9 0.00 13 MONO_CAT+YEAR 1211.9 44.3 0.00 14 POINTS_CAT+YEAR 1215.6 48.0 0.00 15 TRMT+YEAR 1220.2 52.6 0.00

Figure 2.13. Regression for fixed effect of monotypic cattail on entomophilous plant species richness with open water held at median value, presented with 95% confidence intervals.

Figure 2.14. Regression for fixed effect of open water on entomophilous plant species richness with monotypic cattail cover held at median value, presented with 95% confidence intervals.

Table 2.13. Model rankings for predictors of bee species richness in central New York wetlands.

Rank Model AICc ΔAICc wa 1 POINTS_OPEN+YEAR+MONTH 1138.3 0.0 0.84 2 MEDIAN_WATER_DEPTH_CM+YEAR+MONTH 1143.2 4.9 0.07 3 AVG_WATER_DEPTH_CM+YEAR+MONTH 1144.2 5.9 0.04 4 RICHNESS+YEAR+MONTH 1146.8 8.5 0.01 5 AVG_VEG_HT_CM+YEAR+MONTH 1147.1 8.8 0.01 6 POINTS_CAT+YEAR+MONTH 1147.5 9.2 0.01 7 INVASIVE+YEAR+MONTH 1148.6 10.3 0.00 8 TRMT+YEAR+MONTH 1149.2 10.9 0.00 9 ENTO_RICH+YEAR+MONTH 1149.4 11.1 0.00 10 MONO_CAT+YEAR+MONTH 1150.9 12.6 0.00 11 MONTH+YEAR 1151.0 12.7 0.00 12 POINTS_ENTO_PLANTS+YEAR+MONTH 1153.1 14.8 0.00 13 MONTH 1158.8 20.5 0.00 14 POINTS_OPEN+MONO_CAT+YEAR 1167.5 29.2 0.00 15 POINTS_OPEN+YEAR 1168.1 29.8 0.00 16 POINTS_OPEN+INVASIVE+YEAR 1168.3 30.0 0.00 17 MEDIAN_WATER_DEPTH_CM+YEAR 1173.8 35.5 0.00 18 RICHNESS+YEAR 1174.6 36.3 0.00 19 AVG_VEG_HT_CM+YEAR 1175.0 36.7 0.00 20 INVASIVE+YEAR 1175.3 37.0 0.00 21 AVG_WATER_DEPTH_CM+YEAR 1175.5 37.2 0.00 22 ENTO_RICH+YEAR 1176.5 38.2 0.00 23 TRMT+YEAR 1176.8 38.5 0.00 24 TRMT 1176.8 38.5 0.00 25 MONO_CAT+YEAR 1178.6 40.3 0.00 26 POINTS_ENTO_PLANTS+YEAR 1180.7 42.4 0.00

Figure 2.15. Regression for fixed effect of open water on bee species richness, presented with 95% confidence intervals.

Table 2.14. Model rankings for predictors entomophilous plant frequency of occurrence in central New York wetlands.

Rank Model AICc ΔAICc wa 1 INVASIVE+POINTS_OPEN+YEAR+MONTH 1898.3 0.0 1.00 2 INVASIVE+YEAR+MONTH 1915.1 16.8 0.00 3 POINTS_CAT+POINTS_OPEN+YEAR+MONTH 1917.8 19.5 0.00 4 POINTS_CAT+YEAR+MONTH 1923.3 25.0 0.00 5 INVASIVE+YEAR 1929.4 31.1 0.00 6 TRMT+YEAR+MONTH 1939.5 41.2 0.00 7 AVG_WATER_DEPTH_CM+YEAR+MONTH 1941.2 42.9 0.00 8 MEDIAN_WATER_DEPTH_CM+YEAR+MONTH 1941.7 43.4 0.00 9 POINTS_OPEN+YEAR+MONTH 1942.0 43.7 0.00 10 YEAR+MONTH 1942.5 44.2 0.00 11 POINTS_OPEN+YEAR 1948.0 49.7 0.00 12 TRMT+YEAR 1948.4 50.1 0.00 13 AVG_WATER_DEPTH_CM+YEAR 1951.6 53.3 0.00 14 MEDIAN_WATER_DEPTH_CM+YEAR 1951.9 53.6 0.00

Figure 2.16. Regression for fixed effect of invasive graminoids on entomophilous plant frequency with open water held at median value, presented with 95% confidence intervals.

Figure 2.17. Regression for fixed effect of open water on entomophilous plant frequency with invasive cover held at median value, presented with 95% confidence intervals.

Chapter 3

DIVERSITY AND PLANT-POLLINATOR INTERACTIONS OF FLOWER FLIES

(DIPTERA: SYRPHIDAE) IN RESTORED WETLANDS IN CENTRAL NEW YORK, USA

ABSTRACT

Syrphid flies (Diptera: Syrphidae) are important pollinators often overlooked in surveys targeting more charismatic taxa like wild bees. Information about distribution, life history, and ecology are lacking for many species, impeding efforts to conserve native pollinators. Here, passive and active syrphid collection accompanying a wild bee survey in a restored wetland complex in central New York revealed over three dozen species, including a sizeable population of the rare

Parhelophilus divisus (Loew), along with plant-pollinator interactions for 27 syrphid species and

35 flowering plant species. Saprophagous syrphids are associated with decaying organic matter in wetland habitats, and visit numerous wetland flowering plants, making them important to consider when attempting to understand wetland pollination systems. Syrphid pollinators serve to benefit from wetland restoration and, in turn, be affected by management actions aimed at mimicking seasonal hydrology to produce diverse habitat primarily for waterfowl. These findings illustrate the need for more dedicated surveys for non-bee pollinators, because rare or unusual species may be present on protected land but cannot be factored into management until documented.

Keywords: Wild pollinators · Wetland restoration · Wetland management · Syrphidae ·

Parhelophilus divisus · New York

INTRODUCTION

Syrphid or flower flies (Diptera: Syrphidae), are an abundant and diverse component of the insect fauna in a wide range of habitats. Their larvae can be aphidophagous, making them important biocontrol agents in agricultural systems, or aquatic detritivores, providing essential nutrient-cycling services, among other lifestyles. Adult syrphid flies are for the most part generalist flower visitors, and contribute to pollination of hundreds of wild plant species and numerous crops (Larson et al. 2001; Tooker et al. 2006; Ssymank et al. 2008; Rader et al. 2016), though their role as pollinators has been understudied compared to bees (Hymenoptera: Apoidea:

Anthophila). It has been proposed that under some circumstances (e.g., low temperatures, arctic or alpine environments) syrphids are the predominant pollinators of many flowers, and although in general they are considered less effective pollinators as they lack the specialized hairs of bees, their presence can benefit plants by providing redundancy and filling gaps when bee pollinators are infrequent or absent (McCall and Primack 1992; Jauker and Wolters 2008; Rader et al. 2013;

Rader et al. 2016).

As agricultural intensification and land-use changes threaten persistence of wild bees

(Potts et al. 2010; Cameron et al. 2011; Senipathi et al. 2015), the effect of anthropogenic modification on syrphid flies is less clear. Some studies have found syrphids to be more resilient to loss of habitat than bees (Rader et al. 2016), while others have documented a compositional shift in syrphid assemblages to widespread, broad-ranging generalists in more highly impacted areas (Bañkowska 1980; Bramquart and Hemptinne 2000; Biesmeijer et al. 2006). Although non- bee pollinators are rarely the focus of conservation efforts in agricultural areas, agri-environment schemes and native plantings within crop fields designed for bees and butterflies have in many cases shown to benefit syrphids (Morandin and Kremen 2013; Holland et al. 2015; M’Gonigle et

al. 2015; Warzecha et al. 2017). However, heterogeneity among syrphids for larval habitat and diet requirements precludes a broad assessment of their response to land-use change and pollinator habitat creation, with effects varying at the local and landscape scales in relation to both size and quality of habitat (Schweiger et al. 2007; Moquet et al. 2017).

Less than half of the historic wetland area in the United States, and in the state of New

York, remains since the time of the American Revolution, due primarily to draining, filling, and ditching for agriculture, navigation, industry, and residential development (Dahl 1990; Dahl and

Allord 1996). Through the 1986 North American Waterfowl Management Plan (NAWMP), the

1989 North American Wetlands Conservation Act (NAWCA), and other protections, millions of acres of converted and degraded wetlands have been restored to provide quality habitat for waterfowl and other wetland-dependent wildlife (USFWS 2018, 2020). Management of restored wetlands often involves water level manipulation through the use of water control structures that allow seasonal flooding and dewatering of impoundments to mimic historic hydrological cycles that sustained wetland processes and diversity (Fredrickson and Taylor 1982; Strickland et al.

2009). Soil and vegetation disturbance additionally simulate river-scouring events that promote germination of desired plants, to provide sources of food and shelter chiefly for waterfowl

(Fredrickson and Taylor 1982; Gray et al. 2013). Wetland management can be passive, where water levels are held at maximum capacity throughout the growing season, or active, where water drawdowns timed to migration of waterfowl and shorebirds create moist-soil conditions conducive to annual plant production (Fredrickson and Taylor 1982; Fleming et al. 2012; Gray et al. 2013). While the effects of wetland restoration and management are relatively well-studied for waterbirds (Kaminski et al. 2006; King et al. 2006; O’Neal et al. 2008; Farley 2020), the

response of other taxa, such as invertebrates, is still largely unknown (but see De Szalay and

Resh 2000; Schummer et al. 2021).

Many syrphid flies have aquatic saprophagous larvae, making them dependent on wetland habitats, and as adults are likely abundant and important pollinators of wetland flora.

Wetlands embedded in agricultural landscapes have shown to help support local syrphid diversity and potentially provide supplemental resources, which may also benefit nearby crops through improved pollination (Stewart et al. 2016; Begosh et al. 2020). Wetland restoration may be an important tool to increase pollinator populations in the landscape by offering habitat for larval syrphids and floral resources for adults. Yet, few surveys have been performed to provide even basic information about syrphid diversity and ecology in restored wetlands, which is the precursor to more in-depth quantitative studies to measure the utility of wetland restoration and management to syrphids and the pollination services they provide. Here, descriptive results are given of syrphids caught in passive pan trap and sweep-netting surveys at restored wetland sites in central New York, with the goal of adding useful information about pollinator use of managed wetlands to the literature.

METHODS

Surveys were conducted in the Montezuma Wetlands Complex (MWC; 43.024079°N, -

76.748412°W) and Seneca Meadows Wetland Preserve (SMWP; 42.937068°N, -76.823104°W) in central New York State, USA, June to September, 2019 and 2020 (Fig 2.1). The MWC is >

26,000 ha of wetlands and uplands, and includes the United States Fish and Wildlife Service

(USFWS) Montezuma National Wildlife Refuge (MNWR; 3,970 ha), New York State

Department of Environmental Conservation (NYSDEC) Northern Montezuma Wildlife

Management Area (NMWMA; 2,860 ha), and surrounding private lands (Jasikoff 2013; Eckler et al. 2019; Wagner 2020). The SMWP is 230 ha of restored and enhanced wetlands, of which 20 ha are emergent wetlands (McGraw and Larson 2019). The landscape surrounding MWC and

SMWP is predominantly row crop agriculture and forest (Jasikoff 2013; Eckler et al. 2019).

Wetlands were systematically sampled in 2019 (n = 33) and 2020 (n = 33; total n = 38), and were subjected to one of three hydrologic management treatments – passive, partial drawdown, or full drawdown – as described in Farley (2020), Jacobson (2021), and Schummer et al. (2021).

Syrphid flies were collected using survey methods of Jacobson (2021). While the specific research aim was to document diversity of bees (Hymenoptera: Apoidea: Anthophila) in restored wetlands and their response to hydrological management, opportunistic collection of syrphids and other non-bee pollinators was an a priori decision. Standardized pan trap and sweep-netting techniques were used to capture pollinating insects and record plant-pollinator interactions. Pan traps consisted of 163 ml Dixie plastic soufflé cups (Georgia-Pacific, Atlanta, GA) painted fluorescent yellow, blue, or white (Guerra Paint & Pigment Corp., New York, NY) and affixed to

1.2 m-high fiberglass stakes, arranged with colors alternating at 30 (in 2019) or 24 (in 2020) equidistant points along 2 – 4 equally spaced linear transects at each site. Transect length and number varied due to unique dimensions of each wetland, but were configured with points 5 – 20 m apart to capture wetland heterogeneity. Traps were filled with a soapy water and propylene glycol mixture and deployed for 24 hrs. Sweeps targeted monotypic patches of individual flowering plant species. Although syrphids were a focal taxon, methodology was designed to capture bees, thus it is likely that the syrphids collected are not representative of their diversity in the study area; cryptic species not easily recognizable as Syrphidae in the field during sweeps, or those that are more readily captured in sweeps of vegetation rather than flowers, are probably

underrepresented here. All samples were carefully inspected for syrphids in the lab. Point- intercept vegetation surveys were also conducted alongside pollinator collection on established transects to characterize wetland management treatments and availability of floral resources.

Specimens were pinned and labeled with pertinent metadata, and were identified to lowest taxonomic level possible, using the taxonomy and keys in Skevington et al. (2019). Females of some genera were not possible to identify to species. Voucher specimens were deposited in the

SUNY-ESF Insect Collection, and voucher photos of select species were posted to BugGuide

(bugguide.net). Occurrence records were uploaded to the Global Biodiversity Information

Facility (GBIF). Capture rates of syrphid species were too infrequent to detect patterns in diversity or assemblage composition by treatment or environmental variables using Non-Metric

Multidimensional Scaling ordination or linear regression, thus descriptive findings are presented.

A plant-pollinator interaction network was constructed in R (R Core Team 2020) using package

‘bipartite’ (Dormann 2008).

RESULTS & DISCUSSION

In this survey 1,124 syrphids were collected, representing ≥ 38 species in 22 genera (Table 3.1).

The most abundant species was Toxomerus marginatus (Say, n = 441), comprising 39.2% of the catch. A majority of syrphids (n = 928, 82.6%) were collected in pan traps versus sweeps (n =

196, 17.4%). A total of 61 entomophilous plant species were swept for pollinating insects, however syrphids were only documented on 35 species. There were 8 syrphid species collected only in sweeps, and 11 species only in pan traps. Of the most abundant genera (n > 20), some were strongly biased towards collection in traps: Parhelophilus (94.0%), Chalcosyrphus

(91.7%), Toxomerus (87.4%), and Eurimyia (79.6%). Only two abundant genera were collected at equal or greater frequencies in sweeps: Platycheirus (52.4%) and Tropidia (50.0%).

Tropidia quadrata (Say) had the greatest number of unique floral interactions (n = 13), while Toxomerus marginatus was the most abundant on flowers (n = 50; Fig. 3.1, Table 3.2).

Flowers supporting the highest abundance of syrphids were flowering rush (Butomus umbellatus

L., n = 47), nodding bur-marigold (Bidens cernua L., n = 28), pickerelweed (Pontederia cordata

L., n = 26), and purple loosestrife (Lythrum salicaria L., n = 16; Fig. 3.1). Two of these, flowering rush and purple loosestrife, are invasive in the Great Lakes region. Flowering rush also hosted the greatest species richness of syrphid flies (n = 11) followed by nodding bur-marigold

(n = 10; Fig. 3.1). In context, flowering rush was the only flower whose dipteran pollinators outnumbered bee pollinators in abundance and richness, though crabronid wasps in subtribe

Crabronina were also more frequent visitors than bees (n = 41; Appendices 2 and 3). Syrphids

(Parhelophilus laetus [Loew] and Toxomerus sp.) also were recorded and observed visiting flowers of bur-reed (Sparganium), a typically wind-pollinated graminoid, though scattered records in the literature exist of syrphids feeding on this plant (Leereveld 1984) and other anemophilous wetland species (Saunders 2018). Bur-reed was an abundant resource at many study sites, and infrequent visitation by bees was also noted (Jacobson 2021).

A few species of note were collected in this survey. Parhelophilus divisus (Loew) was a frequently encountered syrphid (n = 65), though considered rare in North America with likely <

50 locality records (GBIF 2019b; Skevington et al. 2019). Specimens were collected primarily in

June (85.0%) but were documented from 17-Jun to 10-Sep (Fig. 3.2). All but two individuals were collected from the Howland Island property of NMWMA, with those remaining two from

SMWP. Parhelophilus divisus was documented from nine sites in 2019, representing all three

hydrological management treatments (full drawdown, partial drawdown, and passive management), though all sites had some amount of standing water present throughout the growing season. Parhelophilus divisus was not collected again in 2020 in sites that underwent a complete drying to a mud or moist-soil state (through drawdown or natural evaporation) that year, and was not collected from any new sites in 2020. This genus is associated with cattail

(Typha) as larvae, and require relatively high quality habitat (Martin Hauser, CDFA, pers. comm.). Sites at which this species was recorded shared commonalities in vegetation; all but one site had relatively high densities of white waterlily (Nymphaea odorata Aiton) and the majority had abundant bur-reed, while few other sites at MWC where P. divisus was absent had comparable densities of these plants. Additionally, although some sites with P. divisus had abundant cattail, if this species was solely associated with cattail it would have been expected in many more locations, as cattail marshes were common throughout the MWC outside of Howland

Island. This being said, many sites on the Island were located close to each other, making interpretations potentially spatially biased. Adults of P. divisus have been recorded on lotus

(Nuphar; Skevington et al. 2019); unfortunately in this survey no individuals were caught in sweeps or observed on flowers, but one female was observed resting on white waterlily vegetation (Fig. 3.2).

Howland Island is surrounded by the Seneca River with extensive tracts of floodplain forest throughout the lowland areas, and is known already to host other state rare or imperiled species, including Kentucky coffee tree (Gymnocladus dioicus [L.]), Indiana bat (Myotis sodalis

Miller and Allen), and breeding populations of cerulean (Setophaga cerulea [Wilson]) and prothonotary warblers (Protonotaria citrea [Boddaert]; Jasikoff 2013; Edinger et al. 2014;

Eckler 2019). The presence of P. divisus in great numbers suggests this parcel of land may be

supporting unusual invertebrate species as well; many examples of exemplary or endangered natural communities exist on protected land in New York (Edinger et al. 2014), and inventories focused on invertebrate taxa may reveal hitherto unknown species of concern or interest.

Another scarce Parhelophilus, P. integer (Loew), was documented (n = 5) in this survey, from 11-Jun to 1-Aug (Fig. 3.3). Unlike P. divisus, which were centered at one location, P. integer seemed to be widespread at low abundances across MWC, with each individual captured at a different site ≥ 1.5 km apart. The majority of sites had ≥ 40% cattail cover, with a tendency towards half-vegetated marshes (Schummer et al. 2021), though two were completely dry for part of the growing season, suggesting the individuals may have originated in a nearby wetland.

One floral interaction – buttonbush (Cephalanthus occidentalis L.) – was recorded (Fig. 3.1,

Table 3.2), which matches previous records (Skevington et al. 2019).

One more notable species was Polydontomyia curvipes (Wiedemann), which was detected (n = 3) at NMWMA. This species favors, and may specialize in, brackish wetlands, with few records known outside of this habitat type (Skevington et al. 2019; J. Skevington, CNC, pers. comm.). NMWMA, along with an adjacent parcel managed by The Nature Conservancy, is recognized as one of few locations in the eastern U.S. with remaining examples of inland salt marsh; this rare and globally endangered natural community is characterized by salt-tolerant flora growing on seasonally inundated saline mudflats, the result of groundwater affected by inland salt springs (Eallonardo and Leopold 2013; Eckler et al. 2019; NYNHP 2021).

Polydontomyia curvipes was recorded at sites ranging 0.7 – 4.0 km from the known inland salt marsh locations. While this species is capable of moving some distance, as males exhibit hilltopping behavior (congregating at higher elevations to seek females), they typically do not stray far from natal sites (J. Skevington, pers. comm.). Interestingly, halophytic vegetation

characteristic of inland salt marshes, including American golden dock (Rumex fueginus Phil.), salt marsh bulrush (Bolboschoenus maritimus [L.]), salt-meadow grass (Diplachne fusca ssp. fascicularis [Lam.]), and annual salt marsh aster (Symphyotrichum subulatum var. subulatum

[Michx.]; Eallonardo and Leopold 2013; NYNHP 2021), was documented during vegetation surveys in several managed wetlands at NMWMA not adjacent to the known inland salt marshes

(notably 43.062369°N, -76.734227°W, and 43.025869°N, -76.709511°W; M. Jacobson, unpublished data). Although P. curvipes was not found at these particular sites, it may indicate localized saline influence in more locations than previously documented, providing a wider area of suitable larval habitat for this species from which it may forage short distances.

Information about the life history and plant-pollinator interactions for many syrphid pollinators is still lacking, preventing researchers from assessing their response to land use change and habitat restoration efforts. This survey provides data and documents rare or unusual species that bring to attention the potential for wetland restoration and management to create high quality habitat for wetland-dependent syrphid pollinators. Unlike most bees, saprophagous syrphids are closely linked to wetland habitats during larval and adult stages, making them more likely to benefit from wetland restoration. Pollinators are often overlooked when evaluating effects of wetland management on flora and fauna; it has been shown here that syrphids are frequent visitors to flowers of many wetland plants, likely playing a significant role in the pollination of some species, making them an important taxon to consider in future studies and management plans. Pollinator surveys in all habitat types, especially wetlands, should endeavor a priori to target syrphids and other pollinator taxa alongside bees, to gain a more complete understanding of pollination systems and expand our current knowledge of the ranges, life history, and ecology of all pollinators.

TABLES AND FIGURES

Table 3.1. Syrphid fly species collected in restored wetlands in central New York in 2019 – 2020. Species denoted with † are non- native.

Species No. Caught Wetland Dependent Floral Interactions Allograpta obliqua (Say 1823) 7 N 1 Anasimyia chrysostoma (Wiedemann 1830) 1 Y 1 Anasimyia sp. 5 Y 1 Chalcosyrphus (Xylotomima) nemorum (Fabricius 1805) 24 Y 1 Chrysotoxum pubescens Loew 1864 1 N - Chrysotoxum sp. 1 N - Epistrophe sp. 1 N - Eristalis anthophorina (Fallén 1817) 48 Y 4 Eristalis arbustorum (Linnaeus 1758)† 2 N 1 Eristalis dimidiata Wiedemann 1830 16 N 3 Eristalis flavipes Walker 1849 1 N 1 Eristalis tenax (Linnaeus 1758)† 1 N - Eristalis transversa Wiedemann 1830 5 N 1 Eupeodes americanus (Wiedemann 1830) 5 N - Eupeodes americanus/pomus 4 N - Eurimyia stipata (Walker 1849) 147 Y 7 Helophilus latifrons Loew 1863 5 Y - Mallota bautias (Walker 1849) 2 N 1 Neoascia (Neoascia) metallica (Williston 1882) 1 Y 1 Neoascia metallica/tenur 1 Y 1 Neoascia (Neoascia) tenur (Harris 1780) 3 Y 1 Neoascia sp. 3 Y 1

Orthonevra nitida (Wiedemann 1830) 8 Y 1 Parhelophilus divisus (Loew 1863) 65 Y - Parhelophilus integer (Loew 1863) 5 Y 1 Parhelophilus laetus (Loew 1863) 91 Y 5 Parhelophilus obsoletus (Loew 1863) 1 Y 1 Parhelophilus rex Curran & Fluke 1926 2 Y - Parhelophilus sp. 3 Y - Platycheirus granditarsis (Forster 1771) 1 Y - Platycheirus immarginatus (Zetterstedt 1849) 3 Y 2 Platycheirus quadratus (Say 1823) 7 Y 3 Platycheirus rosarum (Fabricius 1787) 1 Y 1 Platycheirus spp. 9 N 2 Polydontomyia curvipes (Wiedemann 1830) 3 Y 1 Sphaerophoria contigua Macquart 1847 2 N 1 Sphaerophoria philanthus (Meigen 1822) 3 N 1 Sphaerophoria spp. 11 N 3 Syritta pipiens (Linnaeus 1758)† 3 N 2 Temnostoma balyras (Walker 1849) 1 N - Toxomerus geminatus (Say 1823) 125 N 10 Toxomerus marginatus (Say 1823) 441 N 9 Toxomerus politus (Say 1823) 5 N 3 Tropidia quadrata (Say 1824) 44 Y 13 Xanthogramma flavipes (Loew 1863) 1 N - Xylota sp. 1 N - Syrphidae sp. 4 - -

Figure 3.1. Interaction network illustrating plant-pollinator relationships between syrphid flies and flowers as documented by sweep-netting in central New York wetlands. Width of bars are indicative of the number of recorded interactions.

100

Table 3.2. Plant-pollinator interactions between syrphid flies and flowering plant species, arranged by syrphid species. Only records at the species level are included.

Species Floral Associations Allograpta obliqua Brassicaceae: Brassica rapa Anasimyia chrysostoma Ranunculaceae: Anemone canadensis Chalcosyrphus nemorum Lythraceae: Lythrum salicaria Eristalis anthophorina Alismataceae: Sagittaria latifolia; Asteraceae: Bidens cernua, Solidago gigantea; Ranunculaceae: Clematis virginiana Eristalis arbustorum Butomaceae: Butomus umbellatus Eristalis dimidiata Asteraceae: Bidens cernua; Butomaceae: Butomus umbellatus; Rosaceae: Rosa multiflora Eristalis flavipes Asteraceae: Cirsium arvense Eristalis transversa Asteraceae: Bidens cernua Eurimyia stipata Alismataceae: Alisma subcordatum; Butomaceae: Butomus umbellatus; Iridaceae: Iris pseudacorus, Iris versicolor; Orobanchaceae: Agalinis tenuifolia; Pontederiaceae: Pontederia cordata; Verbenaceae: Verbena hastata Mallota bautias Rosaceae: Rosa multiflora Neoascia metallica Alismataceae: Sagittaria latifolia Neoascia tenur Asteraceae: Bidens cernua Orthonevra nitida Rubiaceae: Cephalanthus occidentalis Parhelophilus integer Rubiaceae: Cephalanthus occidentalis Parhelophilus laetus Butomaceae: Butomus umbellatus, Iridaceae: Iris versicolor; Penthoraceae: Penthorum sedoides; Pontederiaceae: Pontederia cordata; Typhaceae: Sparganium sp. Parhelophilus obsoletus Ranunculaceae: Anemone canadensis Platycheirus immarginatus Butomaceae: Butomus umbellatus; Fabaceae: Trifolium repens Platycheirus quadratus Asteraceae: Bidens cernua, Eupatorium perfoliatum; Butomaceae: Butomus umbellatus Platycheirus rosarum Balsaminaceae: Impatiens capensis Polydontomyia curvipes Asteraceae: Bidens cernua Sphaerophoria contigua Asteraceae: Erigeron philadelphicus Sphaerophoria philanthus Asteraceae: Erigeron philadelphicus Syritta pipiens Asteraceae: Bidens cernua; Butomaceae: Butomus umbellatus

101

Toxomerus geminatus Asteraceae: Bidens cernua, Erigeron philadelphicus; Butomaceae: Butomus umbellatus; Lamiaceae: Monarda fistulosa; Lythraceae: Lythrum salicaria; Polygonaceae: Persicaria hydropiperoides; Pontederiaceae: Pontederia cordata; Ranunculaceae: Anemone canadensis; Rubiaceae: Cephalanthus occidentalis, Galium album Toxomerus marginatus Apocynaceae: Asclepias incarnata; Asteraceae: Erigeron philadelphicus, Solidago gigantea; Butomaceae: Butomus umbellatus; Lythraceae: Lythrum salicaria; Polygonaceae: Persicaria hydropiperoides; Pontederiaceae: Pontederia cordata; Ranunculaceae: Ranunculus sceleratus; Verbenaceae: Verbena hastata Toxomerus politus Asteraceae: Bidens cernua, Eupatorium perfoliatum; Pontederiaceae: Pontederia cordata Tropidia quadrata Adoxaceae: Sambucus canadensis; Apocynaceae: Apocynum cannabinum, Asclepias syriaca; Asteraceae: Bidens frondosa; Brassicaceae: Brassica rapa; Lamiaceae: Lycopus americanus, Mentha arvensis, Stachys palustris; Lythraceae: Lythrum salicaria; Polygonaceae: Persicaria hydropiperoides; Pontederiaceae: Pontederia cordata; Ranunculaceae: Anemone canadensis; Rosaceae: Rosa multiflora

102

Figure 3.2. Parhelophilus divisus (Loew) female (left) and male (right), recorded at Howland Island, NMWMA. Photographs: Molly Jacobson. Living specimen photographed at Headquarters Pond.

Figure 3.3. Parhelophilus integer (Loew) female captured via pan trap at MNWR. Photograph: Molly Jacobson.

103

Chapter 4

EVALUATION OF RESEARCH OBJECTIVES AND BROAD IMPLICATIONS

Evaluation of Research Objectives

As outlined in Chapter 1, objectives of my research were as follows:

1) Quantify presence and frequency of entomophilous plants as resources for native pollinators among wetlands with differing management treatments

During vegetation surveys, 96 species of entomophilous flowering plants were recorded.

Assessment of the entomophilous plant assemblage using non-metric methods (NMDS) suggested difference among treatments, but substantial variation remained. Species richness of entomophilous plants did not differ among management treatments, but varied negatively with percentage open water and monotypic cattail. The frequency of occurrence of entomophilous plants, represented as the number of survey points with ≥ one entomophilous plant species, also did not differ among treatments, but did vary negatively with percentage open water and invasive graminoid cover. Floral resources were most diverse in the latter half of the growing season, when annuals such as beggarticks and smartweeds that had germinated as a result of drawdowns, and moist-soil perennials in the Asteraceae (e.g., goldenrod, boneset, sneezeweed), were in bloom, most prominently in full drawdown wetlands. Nodding bur-marigold in particular reached up to 95% frequency on transects in September when in full bloom. However, there were many long-blooming perennials in wetlands with standing water, including pickerelweed, white waterlily, and arrowhead, that offered predictable floral resources, although some of these did not receive the level of pollinator visitation expected based on their abundance. It is possible that during the early and mid-season, uplands provided enough resources to divert pollinators from heavy usage of certain available emergent wetland flowers that may, for reasons not quantified here such as nectar production, be less attractive than alternatives. While an immense amount of data were collected in the field, including standardized vegetation surveys, per-site

105

observations on flower phenology, and hundreds of photographs of field sites and flowering plants, a quick method (e.g., quadrats or visual categorical assessment) of quantifying density of actively blooming floral resources in the field would have been beneficial (e.g., Holzschuh et al

2007; Romey et al. 2007; Stephenson et al. 2020). With a larger field team, more such data could be collected, so availability of floral resources could be more thoroughly analyzed. However, information about floral resources present at pollinator survey sites, regardless of the form it takes, is highly valuable if not crucial for understanding how pollinators distribute themselves within habitats and across the landscape. All pollinator studies must endeavor to collect these data in some way, to provide context for patterns of species diversity and abundance. Further refinement of this study for publication should include compiling of phenological data with vegetation surveys to estimate frequency of blooming entomophilous plants in each site, which can then be incorporated into linear models.

2) Describe native bee and fly assemblage diversity in restored wetlands, and their plant- pollinator associations

This study is among few to survey pollinators in wetlands, especially restored wetlands that are rotationally managed as habitat for migratory birds, primarily waterfowl (see Stephenson et al.

2020), providing a valuable opportunity to understand the ecology of bee and fly pollinators in an undersurveyed cover type. Over 80 species of bees, 38 species of syrphid flies, and 30 species of wasp were recorded in this study, each of which contribute to wetland pollination with varying efficiency and degrees of specialization. Some taxa were underrepresented, including kleptoparasites, megachilids, early-flying bees, and some guilds of syrphids. These gaps may be remedied by performing long-term monitoring, as well as targeted surveys earlier in the season when wet woody plants like willows (Salix spp.) are flowering. It may also be worthwhile to

106

investigate the influence of trap height on documented species diversity, as parasitic bees spend much of their time at ground level, searching for host nest entrances. Traps in this study were often elevated above surrounding vegetation to be consistent across sites, but this placement may have failed to capture some taxa. Expanded methodologies to more directly target dipterans would also increase the number of syrphid species documented, as many are best collected in vegetation rather than at flowers. A majority of species detected using wetland resources are common or fairly common generalists found in a variety of habitats, although some species occurred at frequencies higher (e.g., Agapostemon splendens) or lower (e.g., Augochlorella aurata) than expected. Without sampling of adjacent habitats in the MWC, or replicated surveys in wetlands in other parts of the northeastern U.S. and Canada, conclusions cannot be drawn as to whether observed abundances are due to habitat selection or local variation in pollinator populations. A few specialists and scarce or declining species were collected in relatively substantial numbers, such as Bombus fervidus, Hylaeus nelumbonis, and Parhelophilus divisus, providing insights into previously unknown ecological interactions and the importance of wetlands for supporting species that may have more specific needs not easily met by agricultural, suburban, or open upland habitats. Bees, pollinating flies, and pollinating wasps were collected from 61 entomophilous plant species in 30 families, representing over 430 unique interactions.

Information collected through monotypic sweep-netting is invaluable for understanding not just the distribution of pollinator species, but the functions they perform in ecosystems across multiple spatiotemporal scales and the resources they require to persist in a rapidly changing world. It is pertinent that pollinator surveys endeavor whenever possible to collect these data, instead of sweeps performed indiscriminately, so such studies can be of maximum use in advancing pollinator conservation efforts.

107

3) Determine if native bee assemblages and diversity varied among wetland management treatments

Native bee species richness was best predicted by a negative relationship with percentage of open water in wetlands rather than treatment. The composition of bee assemblages differed between treatments, however this effect explained little of the variation present in the system.

Time of year (sampling period) better explained differences in bee species and abundances than any other environmental factor measured, indicating that factors outside of the studied wetlands are likely driving bee assemblages, and phenological turnover was the best predictor of which species would be detected in wetlands during the growing season. The majority use of wetland flowers by habitat and diet generalists in this survey suggests wetlands act mainly as an opportunistic resource alongside naturally co-occurring cover types like upland or bottomland forest and open areas. Although MWC and SMWP have the benefit of this diverse habitat matrix, many wetlands are remnants embedded in agriculture or residential development.

Nevertheless, the presence of non-managed wetlands in modified landscapes have been shown to benefit pollinators (Stewart et al. 2016; Heneberg et al. 2018; Vickruck et al. 2019), and complementary habitat that offers critically-timed floral resources, such as wetlands in late summer and fall, can greatly influence colony success of social bees like Bombus (Westphal et al. 2009; Mandelik et al. 2012; Rundlöf et al. 2014; Galpern et al. 2017). Maintaining a matrix of wetlands managed with a diversity of drawdown intensities and timings can provide pollinators with abundant foraging options, and restoration even on a limited scale has the potential to increase resources in the landscape, given the productivity of wetlands and the willingness of pollinators to use them. Preventing ongoing degradation of existing and restored wetlands by invasive plants like common reed or reed-canary grass is a challenge that affects many wetland-

108

dependent wildlife (Benoit and Askins 1999; Able and Hagan 2003; Zedler and Kercher 2004;

Chambers et al. 2012), and control of these species is crucial if wetlands are to support the densities and diversity of floral resources needed to maximize benefit to pollinators.

Broad Implications

The goal of this research was to improve our understanding about how wetland management aimed at providing habitat for breeding and migratory birds affects pollinators and the floral resources they require. While the surveys performed here are regionally restricted, they provide novel information which will aid future efforts in pollinator conservation and wetland management, and are hopefully the precursor to more studies of similar kind. Continued encroachment of agricultural, industrial, and urban development into remaining natural areas demands practices that maximize functional and biodiversity benefits of protected lands.

Wetland restoration has been driven in part by the goal to provide the necessary resources to support historical waterfowl populations on more limited hectares than before, through management that favors profuse seed-producing plants, aquatic invertebrates, and submerged aquatic vegetation when most needed during the annual cycle (Fredrickson and Taylor 1982;

King et al. 2006; Fleming et al. 2012). Similarly, the expansion of agriculture places an increasing demand on pollination services while continuing to remove essential habitat (Koh et al. 2016), necessitating that available space for pollinator habitat is used efficiently to meet the dietary and nesting needs of as many native species as possible, often targeting imperiled genera like Bombus. While much focus has been placed on preservation and restoration of open land

(e.g., prairies, meadows), wetlands host both generalist and specialist pollinators (Heneberg et al.

2018; Vickruck et al. 2019; Stephenson et al. 2020), some of which are not well-supported by traditional means (wildflower strips, forb margins, suburban gardens). Thus, wetland restoration,

109

especially in agriculture-dominated landscapes, is poised to offer unique benefits to at-risk vertebrate and invertebrate wildlife taxa.

Great potential exists in partnerships among agencies and organizations working in wetland management and pollinator conservation, to optimize planning, efforts, and outcomes to meet the needs of more species on protected lands. Additionally, recognition of the importance of wetlands to pollinators may provide incentive for landowners who have unmanaged or degraded wetlands on their property, particularly farmers growing insect-pollinated crops, to enroll in voluntary federal easement programs like the Wetland Reserve Program (WRP) and

Conservation Reserve Enhancement Program (CREP), with wide-reaching ecological benefits

(Kaminski et al. 2006, King et al. 2006, O’Neal et al. 2008). These steps forward are not possible without a foundational framework of knowledge around pollinator habitat use, plant-pollinator interactions, and effects of management on floral and nesting resources in wetland environments.

In 2016, New York published a comprehensive Pollinator Protection Plan which addressed key issues surrounding threats to wild and managed pollinators, with plans to increase habitat acreage and surveying efforts within the state. It calls for state natural areas to manage

“…for ecologically diverse and structurally complex habitats to benefit a wide variety of wildlife, including pollinators” (NYSDEC 2016, p. 23), citing that the state owns and manages land in a variety of uses and cover types, including wetlands. However, although more detailed outlines for enhancement and creation of open habitat is provided, no further elaboration is given on how wetlands may also be used to promote pollinator populations. To have robust and resilient pollinator populations, and to support the full breadth of native pollinator diversity required for native flora to persist, a heterogeneous landscape is vital (Holzschuh et al. 2007;

Hinners et al. 2012; Mandelik et al. 2012), which aligns with the needs of other wildlife as well.

110

Well-informed decisions regarding selection and acquisition of land for protection, and management for biodiversity and ecosystem function, can only be made when scientific research reveals the habitat requirements of native species and quantifies wildlife response to management. The Pollinator Plan highlights this necessity for further research into pollinator distribution and conservation status: “The need for baseline data on many species of native pollinators is essential to understanding how to conserve their populations” (NYSDEC 2016, p.

24). Large-scale standardized surveys like the New York Natural Heritage Program’s Empire

State Native Pollinator Survey, which targets bee and non-bee pollinators, is a step in the right direction for acquiring the necessary data to make effective coarse-scale land management decisions. However, localized surveys on state and federal lands can be of great value in informing fine-scale, tailored management plans, identifying populations of rare or imperiled species, and providing opportunities to engage the public through outreach, education, and participation in community science initiatives (e.g., Bioblitzes). As diverse cover types are studied for their unique contributions to supporting pollinators, it is hoped that future installments of the Pollinator Plan, and similar plans elsewhere, will reflect new knowledge with recommendations for protection of underrepresented natural communities, and resources allocated to interdisciplinary partnerships to strengthen and unify conservation efforts on local, state, and federal levels.

Recommendations for Future Study

Certain aspects of moist-soil management may be affecting native bees in ways not measured in this study, which would be a valuable topic for future research. Soil disturbance regimes and invasive species control – which can include mowing, disking, herbicide application, and prescribed burns (Strickland et al. 2009; Hazelton et al. 2014; Eckler et al. 2019) – may reduce

111

suitability of full drawdowns for bees by eliminating mid-season floral resources (e.g., swamp milkweed) and destroying nesting sites in areas above full-service level. Disking after drawdown often leaves sites as bare, overturned soil until late season annuals like Pennsylvania smartweed germinate, and in this study these sites were at times colonized by exotic weed species (e.g., velvetleaf, Abutilon theophrasti Medik., goosefoot spp., bull thistle, Cirsium vulgare [Savi]) in the interim. In addition, woody vegetation was scarce in actively managed wetlands, as also noted by Stephenson et al. (2020). While some shrub species (e.g., buttonbush, dogwoods) were present, primarily in passive wetlands, woody species like willows (which host 9 specialist

Andrena in the northeast, though only one was detected in this survey; Fowler 2016) or red maple (Acer rubrum L.; Batra 1985; Urban-Mead et al. 2021) that offer valuable spring resources to bees were rarely recorded within study sites. Woody vegetation is often managed to maintain early successional marsh habitat and promote seed-producing annuals for waterfowl (Fredrickson and Taylor 1982; Strickland et al. 2009; Eckler et al. 2019). If wetlands exist with adjacent natural habitat where moist-soil woody species are present this may be a non-issue, but in the case of fragmented remnant wetlands, or restored sites in an agricultural landscape, leaving space for these unique resources could help support a diverse bee fauna. Testing the effect these moist- soil management actions may be having on wild bees was beyond the scope of this study but worthy of future consideration.

To be of greatest use to wetland management and pollinator conservation efforts, this study should be repeated at restored wetland complexes in other parts of North America, similar to Stephenson et al. (2020) in the LMAV. This survey was focused in scope to MWC and

SMWP in central New York, thus untangling observed patterns in bee assemblages from artifacts of the study location was difficult. Future studies should compare actively and passively

112

managed wetlands at multiple protected areas across a wider geographic area to account for such possible bias. While not possible given the time and logistical limitations of this study, sampling in surrounding cover types would provide more insight into complementary habitat use.

Sampling in adjacent agricultural areas would allow quantitative analysis into the potential for restored wetlands to act as refuges for pollinators in fragmented, modified landscapes (Evans et al. 2018; Vickruck et al. 2019; Begosh et al. 2020). There are few studies on pristine unimpacted wetlands or degraded wetlands available for comparison to understand if restoration is effective at supporting pre-disturbance levels of pollinator diversity.

Although bees are generally considered the most effective and efficient pollinators in temperate environments, a complete understanding of pollination systems and plant-pollinator interactions cannot be achieved without examining all major pollinator taxa together. The preceding chapters organized results separately by bees and flies, for the most part excluding wasps, and as such some broader conclusions about wetland pollination services may not have been possible. For example, some flowers of wetland plants provided resources primarily to non- bee pollinators in this study (Appendix 2), and if these interactions are overlooked, the value of certain plants may be underestimated in management decisions. However, collecting that data in the first place is the crucial step to making these connections, and it is imperative that pollinator surveys pay more attention to non-bee pollinators. This study provides novel descriptive and quantitative information about wild pollinators in managed wetlands. It is my hope that other wetland surveys and management evaluations will be undertaken to fill remaining knowledge gaps so resources may be directed most efficiently to meet the needs of North America’s native pollinators.

113

LITERATURE CITED

Able K, Hagan S (2003) Impact of common reed, Phragmites australis, on essential fish habitat: influence on reproduction, embryological development, and larval abundance of mummichog (Fundulus heteroclitus). Estuaries 26(1):40–50.

Alger S, Burnham P, Boncristiani H, Brody A (2019) RNA virus spillover from managed honeybees (Apis mellifera) to wild bumblebees (Bombus spp.). PLoS ONE 14(6): e0217822. doi:10.1371/journal.pone.0217822

Alger S, Burnham P, Brody A (2019) Flowers as viral hot spots: honey bees (Apis mellifera) unevenly deposit viruses across plant species. PLoS ONE 14(9): e0221800. doi:10.1371/journal.pone.0221800

Anderson M (2006) Distance-based tests for homogeneity of multivariate dispersions. Biometrics 62:245–253.

Anderson M, Walsh D (2013) PERMANOVA, ANOSIM, and the Mantel test in the face of heterogeneous dispersions: What null hypothesis are you testing? Ecol Monogr 83(4):557– 574. doi:10.1890/12-2010.1

Andersson G, Birkhofer K, Rundlöf, Smith H (2013) Landscape heterogeneity and farming practice alter the species composition and taxonomic breadth of pollinator communities. Basic Appl Ecol 14:540–546. doi:10.1016/j.baae.2013.08.003

Artz D, Nault B (2011) Performance of Apis mellifera, Bombus impatiens, and Peponapis pruinosa (Hymenoptera: Apidae) as pollinators of pumpkin. J Econ Entomol 104(4):1153– 1161. doi:10.1603/EC10431

Ascher J, Kornbluth S, Goelet R (2014) Bees (Hymenoptera: Apoidea: Anthophila) of Gardiners Island, Suffolk County, New York. Northeast Nat 21(1):47–71. doi:10.1656/045.021.0105

Bañkowska R (1980) Fly communities of the family Syrphidae in natural and anthropogenic habitats of Poland. Mem Zoologi 33:3–93.

Barringer T, Williams J, Lumia D (1996) New York wetland resources. In: Fretwell J, Williams J, Redman P (compilers), National Water Survey on Wetland Resources. United States Geological Survey Water Supply Paper 2425, pp 291–296.

Bartomeus I, Ascher J, Gibbs J, Danforth B, Wagner D, Hedtke S, Winfree R (2013) Historical changes in northeastern US bee pollinators related to shared ecological traits. P Natl Acad Sci USA 110(12):4656–4660. doi:10.1073/pnas.1218503110

Batra S (1985) Red maple (Acer rubrum L.), an important early spring food resource for honey bees and other insects. J Kansas Entomol Soc 58(1):169–172.

114

Begosh A, Smith L, Park C, Mcmurry S, Lagrange T (2020) Effects of wetland presence and upland land use on wild hymenopteran and dipteran pollinators in the rainwater basin of Nebraska, USA. Wetlands 40:1017–1031. doi:10.1007/s13157-019-01244-w

Benoit L, Askins R (1999) Impact of the spread of Phragmites on the distribution of birds in Connecticut tidal marshes. Wetlands 19(1):194–208.

Biesmeijer J, Roberts S, Reemer M, Ohlemüller R, Edwards M, Peeters T, Schaffers A, Potts S, Kleukers R, Thomas C, Settele J, Kunin W (2006) Parallel declines in pollinators and insect- pollinated plants in Britain and the Netherlands. Science 313(5785):351–354.

Bishop R, Andrews R, Bridges R (1979) Marsh management and its relationship to vegetation, waterfowl, and muskrats. Iowa Acad Sci 86(2):50–56.

Blitzer E, Gibbs J, Park M, Danforth B (2016) Pollination services for apple are dependent on diverse wild bee communities. Agric Ecosyst Environ 221:1–7. doi:10.1016/j.agee.2016.01.004

Bommarco R, Biesmeijer J, Meyer B, Potts S, Pöyry J, Roberts S, Steffan-Dewenter I, Öckinger E (2010) Dispersal capacity and diet breadth modify the response of wild bees to habitat loss. Proc R Soc B 277:2075–2082. doi:10.1098/rspb.2009.2221

Bommarco R, Lundin O, Smith H, Rundlöf M (2012) Drastic historic shifts in bumble-bee community composition in Sweden. Proc R Soc B 279:309–315. doi:10.1098/rspb.2011.0647

Brandt A, Gorenflo A, Siede R, Meixner M, Büchler R (2016) The neonicotinoids thiacloprid, imidacloprid, and clothianidin affect the immunocompetence of honey bees (Apis mellifera L.). J Insect Physiol 86:40–47. doi:10.1016/j.jinsphys.2016.01.001

Branquart E, Heptinne J (2000) Selectivity in the exploitation of floral resources by hoverflies (Diptera: Syrphinae). Ecography 23(6):732–742.

Bray J, Curtis J (1957) An ordination of the upland forest communities of southern Wisconsin. Ecol Monogr 27(4):325–349. doi:10.2307/1942268

Brittain C, Bommarco R, Vighi M, Settele J, Potts S (2010) Organic farming in isolated landscapes does not benefit flower-visiting insects and pollination. Biol Conserv 143(8):1860–1867. doi:10.1016/j.biocon.2010.04.029

Brown S, Smith K, Batzer D (1997) Macroinvertebrate responses to wetland restoration in northern New York. Environ Entomol 26(5):1016–1024.

Burkle L, Marlin J, Knight T (2013) Plant-pollinator interactions over 120 years: loss of species, co-occurrence, and function. Science 339(6127):1611–1615. doi:10.1126/science.1232728

115

Burnham K, Anderson D (2002) Model selection and multimodel inference: a practical information-theoretic approach, 2nd ed. Springer-Verlag, New York.

Bushmann S, Drummond F (2015) Abundance and diversity of wild bees (Hymenoptera: Apoidea) found in lowbush blueberry growing regions of downeast Maine. Environ Entomol 44(4):975–989. doi:10.1093/ee/nvv082

Cameron S, Lim H, Lozier J, Duennes M, Thorp R (2016) Test of the invasive pathogen hypothesis of bumble bee decline in North America. PNAS 113(16):4386–4391. doi:10.1073/pnas.1525266113

Cameron S, Lozier J, Strange J, Koch J, Cordes N, Solter L, Griswold T, Robinson G (2011) Patterns of widespread decline in North American bumble bees. P Natl Acad Sci USA 108(2):662–667. doi:10.1073/pnas.1014743108

Cane J (2003) Exotic nonsocial bees (Hymenoptera: Apiformes) in North America: ecological implications. In: Strickler K, Cane J (eds), For nonnative crops, whence pollinators of the future? Entomological Society of America, Lanham, Maryland

Cane J (2020) A brief review of monolecty in bees and benefits of a broadened definition. Apidologie 52:17–22. doi:10.1007/s13592-020-00785-y

Cane J, Tepedino V (2017) Gauging the effect of honey bee pollen collection on native bee communities. Conserv Lett 10(2):205–210. doi:10.1111/conl.12263

Carvell C, Meek W, Pywell R, Goulson D, Nowakowski M (2007) Comparing the efficacy of agri-environment schemes to enhance bumble bee abundance and diversity on arable field margins. J Appl Ecol 44:29–40. doi:10.1111/j.1365-2664.2006.01249.x

Carvell C, Osborne J, Bourke A, Freeman S, Pywell R, Heard M (2011) Bumble bee species’ responses to a targeted conservation measure depend on landscape context and habitat quality. Ecol Appl 21(5):1760–1771. doi:10.1890/10-0677.1

Chambers R, Meyerson L, Dibble K (2012) Ecology of Phragmites australis and responses to tidal restoration. In: Roman C, Burdick D (eds) Tidal marsh restoration: a synthesis of science and management. Island Press, Washington, DC, pp 81–96. doi:10.5822/978-1- 61091-229-7_5

Chao A (1987) Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43(4):783–791.

Chazdon R, Chao A, Colwell R, Lin S, Norden N, Letcher S, Clark D, Finegan B, Arroyo J (2011) A novel statistical method for classifying habitat generalists and specialists. Ecology 92:1332–1343. doi:10.1890/10-1345.1

116

Chen H, Boutros P (2011) VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R. BMC Bioinformatics 12(35):1–7.

Colla S, Gadallah F, Richardson L, Wagner D, Gall L (2012) Assessing declines of North American bumble bees (Bombus spp.) using museum specimens. Biodivers Conserv 21:3585–3595. doi:10.1007/s10531-012-0383-2

Colla S, Packer L (2008) Evidence for decline in eastern North American bumblebees (Hymenoptera: Apidae), with special focus on Bombus affinis Cresson. Biodivers Conserv 17:1379–1391.

Costanza R, d’Arge R, de Groot R, Farber S, Grasso M, Hannon B, Limburg K, Naeem S, O’Neill R, Paruelo J, Raskin R, Sutton P, van den Belt M (1997) The value of the world’s ecosystem services and natural capital. Nature 387:253–260.

[CUIC] Cornell University Insect Collection (2019) CUIC_ENT00029054. https://www.discoverlife.org/mp/20l?id=CUIC_ENT00029054 Accessed 07 Mar 2021.

Dahl T (1990) Wetland loss since the revolution. National Wetlands Newsletter 12(6):16–17.

Dahl T, Allord G (1996) History of wetlands in the conterminous United States. In: Fretwell J, Williams J, Redman P (compilers), National Water Survey on Wetland Resources. United States Geological Survey Water Supply Paper 2425, pp 19–26.

Danforth B, Minckley R, Neff J (2019) The solitary bees: biology, evolution, conservation. Princeton University Press, Princeton, New Jersey

Delmas E, Besson M, Brice M, Burkle L, Dalla Riva G, Fortin M, Gravel D, Guimarães P, Hembry D, Newman E, Olesen J, Pires M, Yeakel J, Poisot T (2018) Analysing ecological networks of species interactions. Biol Rev 94(1):16–36. doi:10.1111/brv.12433

De Szalay F, Resh V (2000) Factors influencing macroinvertebrate colonization of seasonal wetlands: responses to emergent plant cover. Freshwater Biol 45:295–308.

Dormann C, Gruber B, Fruend J (2008) Introducing the bipartite package: analysing ecological networks. R news 8:8–11.

Droege S, Engler J, Sellers E, O’Brien L (2016) National protocol framework for the inventory and monitoring of bees. National Wildlife Refuge System, U.S. Fish and Wildlife Service, Fort Collins, Colorado

Dunne J, Williams R, Martinez N (2002) Food-web structure and network theory: the role of connectance and size. PNAS 99(20):12917–12922. doi:10.1073/pnas.192407699

117

Eallonardo A, Leopold D (2014) Inland salt marshes of the Northeastern United States: stress, disturbance, and compositional stability. Wetlands 34:155–166. doi:10.1007/s13157-013- 0493-y

Eckler J, Morlock F, Palermo M, Bonk E, Mahoney J, Kennedy H (2019) Habitat management plan for Northern Montezuma Wildlife Management Area, 2020–2029. New York Department of Environmental Conservation.

Edinger G, Evans D, Gebauer S, Howard T, Hunt D, Olivero A, eds (2014) Ecological communities of New York State, second edition. A revised and expanded edition of Carol Reschke’s Ecological Communities of New York State. New York Natural Heritage Program, New York State Department of Environmental Conservation, Albany, NY.

Efron B (1979) Bootstrap methods: another look at the Jackknife. Ann Stat 7(1):1–26.

Egan J, Mortensen D (2012) A comparison of land-sharing and land-sparing strategies for plant richness conservation in agricultural landscapes. Ecol Appl 22(2):459–471. doi:10.1890/11- 0206.1

Evans E, Smart M, Cariveau D, Spivak M (2018) Wild, native bees and managed honey bees benefit from similar agricultural land uses. Agric Ecosyst Environ 268:162–170. doi:10.1016/j.agee.2018.09.014

Farley E (2020) Influence of water level management on vegetation and bird use of restored wetlands in the Montezuma Wetlands Complex. Thesis, SUNY College of Environmental Science and Forestry, 113 pp.

Feltham H, Park K, Goulson D (2014) Field realistic doses of pesticide imidacloprid reduce bumblebee pollen foraging efficiency. Ecotoxicology 23:317–323. doi:10.1007/s10646-014- 1189-7

Fiedler A, Landis D, Arduser M (2011) Rapid shift in pollinator communities following invasive species removal. Restor Ecol 20(5):593–602. doi:10.1111/j.1526-100X.2011.00820.x

Fleming K, Kaminski R, Tietjen T, Schummer M, Ervin G, Nelms, K (2012) Vegetative forage quality and moist-soil management on Wetlands Reserve Program lands in Mississippi. Wetlands 32(5): 919–929.

Fowler J (2016) Specialist bees of the Northeast: host plants and habitat conservation. Northeast Nat 23(2):305–320. doi:10.1656/045.023.0210

Fowler J, Droege S (2020) Pollen specialist bees of the eastern United States. https://jarrodfowler.com/specialist_bees.html. Accessed 20 Feb 2021.

118

Fredrickson L, Taylor T (1982) Management of seasonally flooded impoundments for wildlife. U.S. Fish and Wildlife Service, Resource Publication 148, Washington, D.C., USA.

Fussell M, Corbet S (1992) The nesting places of some British bumble bees. J Apicult Res 31(1):32–41. doi:10.1080/00218839.1992.11101258

Galeano J, Pastor J, Iriondo J (2009) Weighted-interaction nestedness estimator (WINE): a new estimator to calculate over frequency matrices. Environ Model Softw 24:1342–1346.

Gallai N, Salles J, Settele J, Vaissière B (2009) Economic valuation of the vulnerability of world agriculture confronted with pollinator decline. Ecol Econ 68:810–821. doi:10.1016/j.ecolecon.2008.06.014

Galpern P, Johnson S, Retzlaff J, Chang D, Swann J (2017) Reduced abundance and earlier collection of bumble bee workers under intensive cultivation of a mass-flowering prairie crop. Ecol Evol 7:2414–2422. doi:10.1002/ece3.2856

Garantonakis N, Varikou K, Birouraki A, Edwards M, Kalliakaki V, Andrinopoulos F (2016) Comparing the pollination services of honey bees and wild bees in a watermelon field. Sci Hortic-Amsterdam 204:138–144. doi:10.1016/j.scienta.2016.04.006

Garibaldi L, Steffan-Dewenter I, Winfree R, Aizen M, Bommarco R, Cunningham S, Kremen C, Carvalheiro L, Harder L, Afik O, Bartomeus I, Benjamin F, Boreux V, Cariveau F, Chacoff N, Dudenhöffer J, Freitas B, Ghazoul J, Greenleaf S, Hipólito J, Holzschuh A, Howlett B, Isaacs R, Javorek S, Kennedy C, Krewenka K, Krishnan S, Mandelik Y, Mayfield M, Motzke I, Munyuli T, Nault B, Otieno M, Petersen J, Pisanty G, Potts S, Rader R, Ricketts T, Rundlöf M, Seymour C, Schüepp C, Szentgyörgyi H, Taki H, Tscharntke T, Vergara C, Viana B, Wanger T, Westphal C, Williams N, Klein A (2013) Wild pollinators enhance fruit set of crops regardless of honey bee abundance. Science 339(6127):1608–1611. doi:10.1126/science.1230200

[GBIF] GBIF Backbone Taxonomy (2019) Parhelophilus divisus (Loew, 1863) in GBIF Secretariat. Checklist dataset https://doi.org/10.15468/39omei Accessed via GBIF.org on 14 Mar 2021.

[GBIF] GBIF Backbone Taxonomy (2019) Melissodes apicatus Lovell & Cockerell, 1906 in GBIF Secretariat (2019). Checklist dataset https://doi.org/10.15468/39omei Accessed via GBIF.org on 07 Mar 2021.

Geslin B, Gauzens B, Baude M, Dajoz I, Fontaine C, Henry M, Ropars L, Rollin O, Thébault E, Vereecken N (2017) Massively introduced managed species and their consequences for plant–pollinator interactions. Adv Ecol Res 57:147–199. doi:10.1016/bs.aecr.2016.10.007

Gibbs J, Ascher J, Rightmyer M, Isaacs R (2017) The bees of Michigan (Hymenoptera: Apoidea: Anthophila), with notes on distribution, taxonomy, pollination, and natural history. Zootaxa 4352(1):001–160. doi:10.11646/zootaxa.4352.1.1

119

Gleason H, Cronquist A (1991) Manual of vascular plants of northeastern United States and adjacent Canada, 2nd ed. New York Botanical Garden Press, Bronx, New York

Goldstein P, Ascher J (2016) Taxonomic and behavioral composition of an island fauna: a survey of bees (Hymenoptera: Apoidea: Anthophila) on Martha’s Vineyard, Massachusetts. Proc Entomol Soc Wash 118(1):37–92. doi:10.4289/0013-8797.118.1.37

Goulson D (2003) Effects of introduced bees on native ecosystems. Annu Rev Ecol Evol S 34:1– 26.

Goulson D, Lye G, Darvill B (2008) Decline and conservation of bumble bees. Annu Rev Entomol 53:191–208. doi:10.1146/annurev.ento.53.103106.093454

Goulson D, Nicholls E, Botias C, Rotheray E (2015) Bee declines driven by combined stress from parasites, pesticides, and lack of flowers. Science 347(6229): 1255957-1–1255957-9. doi:10.1126/science.1255957

Goulson D, Sparrow K (2009) Evidence for competition between honeybees and bumblebees; effects on bumblebee worker size. J Insect Conserv 13:177–181. doi:10.1007/s10841-008- 9140-y

Grabas G, Laverty T (1999) The effect of purple loosestrife (Lythrum salicaria L.; Lythraceae) on the pollination and reproductive success of sympatric co-flowering wetland plants. Écoscience, 6(2):230–242. doi:10.1080/11956860.1999.11682524

Gray M, Hagy H, Nyman J, Stafford J (2013) Management of wetlands for wildlife. In: Anderson J, Davis C (eds) Wetland techniques volume 3: applications and management. Springer, Netherlands, pp 121–180

Graystock P, Goulson D, Hughes W (2015) Parasites in bloom: flowers aid dispersal and transmission of pollinator parasites within and between bee species. Proc R Soc B 282: 20151371. doi:10.1098/rspb.2015.1371

Greenleaf S, Kremen C (2006) Wild bee species increase tomato production and respond differently to surrounding land use in Northern California. Biol Conserv 133:81–87. doi:10.1016/j.biocon.2006.05.025

Greenleaf S, Williams N, Winfree R, Kremen C (2007) Bee foraging ranges and their relationship to body size. Oecologia 153:589–596. doi:10.1007/s00442-007-0752-9

Grixti J, Wong L, Cameron S, Favret C (2009) Decline of bumble bees (Bombus) in the American Midwest. Biol Conserv 142:75–84. doi:10.1016/j.biocon.2008.09.027

120

Gutrich J, Taylor K, Fennessy M (2009) Restoration of vegetation communities of created depressional marshes in Ohio and Colorado (USA): the importance of initial effort for mitigation success. Ecol Eng 35:351–368. doi:10.1016/j.ecoleng.2008.09.018

Hanula J, Ulyshen M, Horn S (2016) Conserving pollinators in North American forests: a review. Nat Area J 36(4):427–439. doi:10.3375/043.036.0409 Harmon-Threatt (2020) Influence of nesting characteristics on health of wild bee communities. Annu Rev Entomol 65:39–56. doi:10.1146/annurev-ento-011019-024955

Hatfield R, Jepsen S, Vaughan M, Black S, Lee-Mäder E (2018) An overview of the potential impacts of honey bees to native bees, plant communities, and ecosystems in wild landscapes: recommendations for land managers. The Xerces Society for Invertebrate Conservation, Portland, Oregon, 12 pp.

Hazelton E, Mozdzer T, Burdick D, Kettenring K, Whigham D (2014) Phragmites australis management in the United States: 40 years of methods and outcomes. AoB PLANTS 6: plu001. doi:10.1093/aobpla/plu001

Heneberg P, Bogusch P, Řezác M (2018) Numerous drift sand “specialists” among bees and wasps (Hymenoptera: Aculeata) nest in wetlands that spontaneously form de novo in arable fields. Ecol Eng 117:133–139. doi:10.1016/j.ecoleng.2018.04.007

Hinners S, Kearns C, Wessman C (2012) Role of scale, matrix, and native habitat in supporting a diverse suburban pollinator assemblage. Ecol Appl 22(7):1923–1935. doi:10.1890/11-1590.1

Hladik M, Vandever M, Smalling K (2016) Exposure of native bees foraging in an agricultural landscape to current-use pesticides. Sci Total Environ 542:469–477. doi:10.1016/j.scitotenv.2015.10.077

Holland J, Smith B, Storkey J, Lutman P, Aebischer N (2015) Managing habitats on English farmland for insect pollinator conservation. Biol Conserv 182:215–222. doi:10.1016/j.biocon.2014.12.009

Holzschuh A, Steffan-Dewenter I, Kleijn D, Tscharntke T (2007) Diversity of flower-visiting bees in cereal fields: effects of farming system, landscape composition and regional context. J Appl Ecol 44:41–49. doi:10.1111/j.1365-2664.2006.01259.x

Jacobson M, Tucker E, Mathiasson M, Rehan S (2018) Decline of bumble bees in northeastern North America, with special focus on Bombus terricola. Biol Conserv 217:437–445. doi:10.1016/j.biocon.2017.11.026

Jasikoff T (2013) Montezuma National Wildlife Refuge comprehensive conservation plan. United States Fish and Wildlife Service, Montezuma National Wildlife Refuge, Seneca Falls, New York

121

Jauker F, Wolters V (2008) Hover flies are efficient pollinators of oilseed rape. Oecologia 156:819–823. doi:10.1007/s00442-008-1034-x

Jenkins W, Murray B, Kramer R, Faulkner S (2010) Valuing ecosystem services from wetland restoration in the Mississippi Alluvial Valley. Ecol Econ 69:1051–1061. doi:10.1016/j.ecolecon.2009.11.022

Kaminski M, Baldassarre G, Pearse A (2006) Waterbird responses to hydrological management of Wetlands Reserve Program habitats in New York. Wildlife Soc B 34(4):921–926. doi:10.2193/0091-7648(2006)34[921:WRTHMO]2.0.CO;2

Kaminski R, Prince H (1981) Dabbling duck and aquatic macroinvertebrate responses to manipulated wetland habitat. J Wildl Manage 45(1):1–15.

Kearns C (2001) North American dipteran pollinators: assessing their value and conservation status. Conserv Ecol 5(1):5. http://www.consecol.org/vol5/iss1/art5/

Kerr J, Pindar A, Galpern P, Packer L, Potts S, Roberts S, Rasmont P, Schweiger O, Colla S, Richardson L, Wagner D, Gall L, Sikes D, Pantoja A (2015) Climate change impacts on bumblebees converge across continents. Science 349(6244):177–180. doi:10.1126/science.aaa7031

King S, Twedt D, Wilson R (2006) The role of the Wetland Reserve Program in conservation efforts in the Mississippi Alluvial Valley. Wildlife Soc B 34(4):914–920. doi:10.2193/0091- 7648(2006)34[914:TROTWR]2.0.CO;2

Kleijn D, Baquero R, Clough Y, Díaz M, De Esteban J, Fernández F, Gabriel D, Herzog F, Holzschuh A, Jöhl R, Knop E, Kruess A, Marshall E, Steffan-Dewenter I, Tscharntke T, Verhulst J, West T, Yela J (2006) Mixed biodiversity benefits of agri-environment schemes in five European countries. Ecol Lett 9:243–254. doi:10.1111/j.1461-0248.2005.00869.x

Klein A, Vaissière B, Cane J, Steffan-Dewenter I, Cunningham S, Kremen C, Tscharntke T (2007) Importance of pollinators in changing landscapes for world crops. Proc Roy Soc B 274:303–313. doi:10.1098/rspb.2006.3721

Koh I, Lonsdorf E, Williams N, Brittain C, Isaacs R, Gibbs J, Ricketts T (2016) Modeling the status, trends, and impacts of wild bee abundance in the United States. PNAS 113(1):140– 145. doi:10.1073/pnas.1517685113

Kremen C, M’Gonigle L (2015) Small-scale restoration in intensive agricultural landscapes supports more specialized and less mobile pollinator species. J Appl Ecol 52:602–610. doi:10.1111/1365-2664.12418

Kremen C, Williams N, Thorp R (2002) Crop pollination from native bees at risk from agricultural intensification. PNAS 99(26):16812–16816. doi:10.1073pnas.262413599

122

Kuznetsova A, Brockhoff P, Christensen R (2017) LmerTest package: tests in linear mixed effects models. J Stat Softw 82(13):1–26. doi:10.18637/JSS.V082.I13

Larson B, Kevan P, Inouye D (2001) Flies and flowers: taxonomic diversity of anthophiles and pollinators. Can Entomol 133:439–465.

Lawson S, Ciaccio K, Rehan S (2016) Maternal manipulation of pollen provisions affects worker production in a small carpenter bee. Behav Ecol Sociobiol 70(11):1891–1900. doi:10.1007/s00265-016-2194-z

Leereveld H (1984) Anthecological relations between reputedly anemophilous flowers and syrphid flies. VI. Aspects of the anthecology of Cyperaceae and Sparganium erectum L. Acta Bot Neerl 33(4):475–482.

Losey J, Vaughan M (2006) The economic value of ecological services provided by insects. Bioscience 56(4):311–323. doi:10.1641/0006-3568(2006)56[311:TEVOES]2.0.CO;2

Matthews J, Spyreas G (2010) Convergence and divergence in plant community trajectories as a framework for monitoring wetland restoration progress. J Appl Ecol 47:1128–1136. doi:10.1111/j.1365-2664.2010.01862.x

McCall C, Primack R (1992) Influence of flower characteristics, weather, time of day, and season on insect visitation rates in three plant communities. Am J Bot 79(4):434–442.

McKinney M, Lockwood J (1999) Biotic homogenization: a few winners replacing many losers in the next mass extinction. Trends Ecol Evol 14(11):450–453.

M’Gonigle L, Ponisio L, Cutler K, Kremen C (2015) Habitat restoration promotes pollinator persistence and colonization in intensively managed agriculture. Ecol Appl 25(6):1557–1565. doi:10.1890/14-1863.1

Mallinger R, Werts P, Gratton C (2015) Pesticide use within a pollinator-dependent crop has negative effects on the abundance and species richness of sweat bees, Lasioglossum spp., and on bumble bee colony growth. J Insect Conserv 19:999–1010. doi:10.1007/s10841-015- 9816-z

Mandelik Y, Winfree R, Neeson T, Kremen C (2012) Complementary habitat use by wild bees in agro-natural landscapes. Ecol Appl 22(5):1535–1546. doi:10.1890/11-1299.1

Matteson K, Ascher J, Langellotto G (2008) Bee richness and abundance in New York City urban gardens. Ann Entomol Soc Am 101(1):140–150. doi:10.1603/0013- 8746(2008)101[140:BRAAIN]2.0.CO;2

McGraw M, Larson J (2019) Progressive Waste – Seneca Meadows: Wetland mitigation status annual report 9, volume 2: fauna status report 2019. Applied Ecological Services. Prepared for Seneca Meadows, Inc.

123

Melo A (2019) CommEcol: community ecology analyses. R package version 1.7.0. https://CRAN.R-project.org/package=CommEcol

Mitchell T (1960) Bees of the eastern United States, vol 1. Technical Bulletin No 141, North Carolina Agricultural Experiment Station. Raleigh, North Carolina

Mitchell T (1962) Bees of the eastern United States, vol 2. Technical Bulletin No 152, North Carolina Agricultural Experiment Station. Raleigh, North Carolina

Mitsch W, Zhang L, Anderson C, Altor A, Hernández M (2005) Creating riverine wetlands: ecological succession, nutrient retention, and pulsing effects. Ecol Eng, 25(5): 510–527. doi:10.1016/j.ecoleng.2005.04.014

Moisan-Deserres J, Girard M, Chagnon M, Fournier V (2014) Pollen loads and specificity of native pollinators of lowbush blueberry. J Econ Entomol 107(3):1156–1162. doi: 10.1603/EC13229

Moquet L, Laurent E, Bacchetta R, Jacquemart A (2018) Conservation of hoverflies (Diptera, Syrphidae) requires complementary resources at the landscape and local scales. Insect Conserv Diver 11:72–87. doi:10.1111/icad.12245

Morandin L, Kremen C (2013) Hedgerow restoration promotes pollinator populations and exports native bees to adjacent fields. Ecol Appl 23(4):829–839. doi:10.1890/12-1051.1

Moreira E, Boscolo D, Viana B (2015) Spatial heterogeneity regulates plant-pollinator networks across multiple landscape scales. PLoS ONE 10(4): e0123628. doi:10.1371/journal.pone.0123628

Moreno-Mateos D, Power M, Comín F, Yockteng R (2012) Structural and functional loss in restored wetland ecosystems. PLoS Biol 10(1):e1001247. doi:10.1371/journal.pbio.1001247

Morón D, Szentgyörgyi H, Wantuch M, Celary W, Westphal C, Settele J, Woyciechowski M (2008) Diversity of wild bees in wet meadows: implications for conservation. Wetlands 28(4):975–983.

Morse R, Calderone N (2000) The value of honey bees as pollinators of U.S. Crops in 2000. Bee Culture 128:1–15.

Mukaka M (2012) A guide to appropriate use of correlation coefficient in medical research. Malawi Med J 24:69–71.

Murkin H, Kaminski R, Titman R (1982) Responses by dabbling ducks and aquatic invertebrates to an experimentally manipulated cattail marsh. Can J Zool 60:2324–2332.

124

Nadeau C, Conway C (2015) Optimizing water depth for wetland-dependent wildlife could increase wetland restoration success, water efficiency, and water security. Restor Ecol 23(3):292–300. doi:10.1111/rec.12180

[NYNHP] New York Natural Heritage Program (2020) 2019 Summary for the Empire State Native Pollinator Survey January – December 2019. Albany, New York

[NYNHP] New York Natural Heritage Program (2021) Online conservation guide for inland salt marsh. https://guides.nynhp.org/inland-salt-marsh/. Accessed 13 Mar 2021.

[NYSDEC] New York State Department of Environmental Conservation (2015) State wildlife action plan species of greatest conservation need: high priority bees. State Wildlife Action Plan Species Assessments. Albany, New York

[NYSDEC] New York State Department of Environmental Conservation, Department of Agriculture and Markets (2016) New York State pollinator protection plan. Albany, New York

Normandin E, Vereecken N, Buddle C, Fournier V (2017) Taxonomic and functional trait diversity of wild bees in different urban settings. PeerJ 5:e3051. doi:10.7717/peerj.3051

Noy-Meir I, Walker D, Williams W (1975) Data transformations in ecological ordination. J Ecol 63(3):779–800.

Olden J, Poff N, Douglas M, Douglas M, Fausch K (2004) Ecological and evolutionary consequences of biotic homogenization. Trends Ecol Evol 19(1):18–24. doi:10.1016/j.tree.2003.09.010

O’Neal B.J., Heske E.J., and Stafford J.D. 2008. Waterbird response to wetlands restored through the Conservation Reserve Enhancement Program. J Wildlife Manage 72(3):654–664. doi:10.2193/2007-165

O’Neill K, O’Neill J (2010) Cavity-nesting wasps and bees of central New York State: the Montezuma Wetlands Complex. Northeast Nat 17(3):455–472.

Oksanen J, Blanchet F, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin P, O'Hara R, Simpson G, Solymos P, Stevens M, Szoecs E, Wagner H (2020) Vegan: community ecology package. R package version 2.5-7. https://CRAN.R-project.org/package=vegan.

Ollerton J, Winfree R, Tarrant S (2011) How many flowering plants are pollinated by animals? Oikos 120:321–326. doi:10.1111/j.1600-0706.2010.18644.x

Park M, Blitzer E, Gibbs J, Losey J, Danforth B (2015) Negative effects of pesticides on wild bee communities can be buffered by landscape context. Proc R Soc B 282(1809):20150299. doi:10.1098/rspb.2015.0299

125

Pettis J, Lichtenberg E, Andree M, Stitzinger J, Rose R, vanEngelsdorp D (2013) Crop pollination exposes honey bees to pesticides which alters their susceptibility to the gut pathogen Nosema ceranae. PLoS One 8(7):e70182. doi:10.1371/journal.pone.0070182

Pitts-Singer T, Cane J (2011) The alfalfa leafcutting bee, Megachile rotundata: the world’s most intensively managed solitary bee. Annu Rev Entomol 56:221–237. doi:10.1146/annurev- ento-120709-144836

Płachno B, Stpiczyńska M, Adamec L, Miranda V, Świątek P (2018) Nectar trichome structure of aquatic bladderworts from the section Utricularia (Lentibulariaceae) with observation of flower visitors and pollinators. Protoplasma 255:1053–1064. doi:10.1007/s00709-018-1216- 2

Ploquin E, Herrera J, Obeso J (2013) Bumblebee community homogenization after uphill shifts in montane areas of northern Spain. Oecologia 173(4):1649–1660.

Pojar J (1973) Pollination of typically anemophilous salt marsh plants by bumble bees, Bombus terricola occidentalis Grne. Am Midl Nat 89(2):448–451.

Portman Z, Bruninga-Socolar B, Cariveau D (2020) The state of bee monitoring in the United States: a call to refocus away from bowl traps and towards more effective methods. Ann Entomol Soc Am 113(5):337–342. doi:10.1093/aesa/saaa010

Potts S, Biesmeijer J, Kremen C, Neumann P, Schweiger O, Kunin W (2010) Global pollinator declines: trends, impacts, and drivers. Trends in Ecol Evol 25(6):345–353. doi:10.1016/j.tree.2010.01.007

Pywell R, Warman E, Hulmes L, Nuttall P, Sparks T, Critchley C, Sherwood A (2006) Effectiveness of new agri-environment schemes in providing foraging resources for bumblebees in intensively farmed landscapes. Biol Conserv 129:192–206. doi:10.1016/j.biocon.2005.10.034

Rader R, Bartomeus I, Garibaldi L, Garratt M, Howlett B, Winfree R, Cunningham S, Mayfield M, Arthur A, Andersson G, Bommarco R, Brittain C, Carvalheiro L, Chacoff N, Entling M, Foully B, Freitas B, Gemmill-Herren B, Ghazoul J, Griffin S, Gross C, Herbertsson L, Herzog F, Hipólito J, Jaggar S, Jauker F, Klein A, Kleijn D, Krishnan S, Lemos C, Lindström S, Mandelik Y, Monteiro V, Nelson W, Nilsson L, Pattemore D, Pereira N, Pisanty G, Potts S, Reemer M, Rundlöf M, Sheffield C, Scheper J, Schüepp C, Smith H, Stanley D, Stout J, Szentgyörgyi H, Taki H, Vergara C, Viana B, Woyciechowski M (2016) Non-bee insects are important contributors to global crop pollination. PNAS 113(1):146–151. doi:10.1073/pnas.1517092112

Rader R, Bartomeus I, Tylianakis J, Laliberté E (2014) The winners and losers of land use intensification: pollinator community disassembly is non-random and alters functional diversity. Diversity Distrib 20:908–917. doi:10.1111/ddi.12221

126

Rader R, Edwards W, Westcott D, Cunningham S, Howlett B (2013) Diurnal effectiveness of pollination by bees and flies in agricultural Brassica rapa: implications for ecosystem resilience. Basic Appl Ecol 14:20–27. doi:10.1016/j.baae.2012.10.011

R Core Team (2020) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/

Rehan S, Sheffield C (2011) Morphological and molecular delineation of a new species in the Ceratina dupla species group (Hymenoptera: Apidae) of eastern North America. Zootaxa 2873:35–50.

Rhoades P, Griswold T, Waits L, Bosque-Pérez N, Kennedy C, Eigenbrode S (2017) Sampling technique affects detection of habitat factors influencing wild bee communities. J Insect Conserv 21:703–714. doi: 10.1007/s10841-017-0013-0

Roberts-Pichette P, Gillespie L (1999) Terrestrial vegetation biodiversity monitoring protocols. Ecological Monitoring and Assessment Network Publication 9. Burlington, Ontario.

Romey W, Ascher J, Powell D, Yanek M (2007) Impacts of logging on midsummer diversity of native bees (Apoidea) in a northern hardwood forest. J Kansas Entomol Soc 80(4):327–338.

Roubik D, Villanueva-Gutiérrez R (2009) Invaisve Africanized honey bee impact on native solitary bees: a pollen resource and trap nest analysis. Biol J Linn Soc 98:152–160. doi:10.1111/j.1095-8312.2009.01275.x

Rubene D, Schroeder M, Ranius T (2015) Diversity patterns of wild bees and wasps in managed boreal forests: effects of spatial structure, local habitat and surrounding landscape. Biol Conserv 184:201–208. doi:10.1016/j.biocon.2015.01.029

Rundle W, Fredrickson L (1981) Managing seasonally flooded impoundments for migrant rails and shorebirds. Wildlife Soc B 9(2):80–87.

Rundlöf M, Persson A, Smith H, Bommarco R (2014) Late-season mass-flowering red clover increases bumble bee queen and male densities. Biol Conserv 172:138–145. doi:10.1016/j.biocon.2014.02.027

Russo L, Park M, Gibbs J, Danforth B (2015) The challenge of accurately documenting bee species richness in agroecosystems: bee diversity in eastern apple orchards. Ecol Evol 5(17):3531–3540. doi:10.1002/ece3.1582

Sandrock C, Tanadini L, Pettis J, Biesmeijer J, Potts S, Neumann P (2014) Sublethal neonicotinoid insecticide exposure reduces solitary bee reproductive success. Agr Forest Entomol 16:119–128. doi:10.1111/afe.12041

127

Saunders M (2018) Insect pollinators collect pollen from wind-pollinated plants: implications for pollination ecology and sustainable agriculture. Insect Conserv Diver 11:13–31. doi:10.1111/icad.12243

Schober P, Boer C, Schwarte L (2018) Correlation coefficients: appropriate use and interpretation. Anesth Analg 126:1763. doi:10.1213/ANE.0000000000002864

Schummer M, Eason K, Hodges T, Farley E, Sime K, Coluccy J, Tozer D (2021) Response of aquatic macroinvertebrate density and diversity to wetland management and structure in the Montezuma Wetlands Complex, New York. J Great Lakes Res, in press.

Schummer M, Hagy H, Fleming K, Cheshier J, Callicutt J (2012) A guide to moist-soil wetland plants of the Mississippi Alluvial Valley. University Press of Mississippi, Jackson

Schwarz M, Richards M, Danforth B (2007) Changing paradigms in insect social evolution: insights from halictines and allodapines bees. Annu Rev Entomol 52:127–150. doi:10.1146/annurev.ento.51.110104.150950

Schweiger O, Musche M, Bailey D, Billeter R, Diekötter T, Hendrickx F, Herzog F, Liira J, Maelfait J, Speelmans M, Dziock F (2007) Functional richness of local hoverfly communities (Diptera, Syrphidae) in response to land use across temperate Europe. Oikos 116:461–472. doi:10.1111/j.2007.0030-1299.15372.x

Scott-Dupree C, Conroy L, Harris C (2009) Impact of currently used or potentially useful insecticides for canola agroecosystems on Bombus impatiens (Hymenoptera: Apidae), Megachile rotundata (Hymenoptera: Megachilidae), and Osmia lignaria (Hymenoptera: Megachilidae). J Econ Entomol 102(1):177–182.

Scribailo R, Posluszny U (1984) The reproductive biology of Hydrocharis morsus-ranae. I. Floral biology. Can J Bot 62:2779–2787. doi:10.1139/b84-370

Senapathi D, Carvalheiro L, Biesmeijer J, Dodson C, Evans R, McKerchar M, Morton R, Moss E, Roberts S, Kunin W, Potts S (2015) The impact of over 80 years of land cover changes on bee and wasp pollinator communities in England. Proc R Soc B 282(1806):1–8. doi:10.1098/rspb.2015.0294

Shavit O, Dafni A, Ne’eman G (2009) Competition between honeybees (Apis mellifera) and native solitary bees in the Mediterranean region of Israel – implications for conservation. Isr J Plant Sci 57:171–183. doi:10.1560/IJPS.57.3.171

Sheffield C, Ratti C, Packer L, Griswold T (2011) Leafcutter and mason bees of the genus Megachile Latreille (Hymenoptera: Megachilidae) in Canada and Alaska. Can J Arthropod Identif 18:1–107. doi:10.3752/cjai.2011.18

128

Shell W, Rehan S (2018) Behavioral and genetic mechanisms of social evolution: insights from incipiently and facultatively social bees. Apidologie 49:13–30. doi:10.1007/s13592-017- 0527-1

Shuler R, Roulston T, Farris G (2005) Farming practices influence wild pollinator populations on squash and pumpkin. J Econ Entomol 98(3):790–795. doi:10.1603/0022-0493-98.3.790

Skevington J, Locke M, Young A, Moran K, Crins W, Marshall S (2019) Field Guide to the Flower Flies of Northeastern North America. Princeton University Press, Princeton, NJ. 512 p.

Ssymank A, Kearns C, Pape T, Thompson F (2008) Pollinating flies (Diptera): a major contribution to plant diversity and agricultural production. Biodiversity 9(1&2):86–89.

Stanley D, Smith K, Raine N (2015) Bumblebee learning and memory is impaired by chronic exposure to a neonicotinoid pesticide. Sci Rep-UK 5(16508):1–10.

Stephenson P, Dowling A, Krementz D (2020) Bee communities of emergent wetlands under restoration in the Lower Mississippi Alluvial Valley of Arkansas. Southeast Nat 19(3):472– 490. doi:10.1656/058.019.0303

Stewart R (1996) Wetlands as bird habitat. In: Fretwell J, Williams J, Redman P (compilers), National Water Survey on Wetland Resources. United States Geological Survey Water Supply Paper 2425, pp 49–56.

Stewart R, Andersson G, Brönmark C, Klatt B, Hansson L, Zülsdorff V, Smith H (2016) Ecosystem services across the aquatic – terrestrial boundary: linking ponds to pollination. Basic Appl Ecol 18:13–20. doi:10.1016/j.baae.2016.09.006

Strickland B, Kaminski R, Nelms K, Tullos A, eds (2009) Waterfowl habitat management handbook for the Lower Mississippi River Valley. Mississippi State University Extension Services Publication 1864, Mississippi State, USA

Svensson B, Lagerlöf J, Svensson B (2000) Habitat preferences of nest-seeking bumble bees (Hymentoptera: Apidae) in an agricultural landscape. Agric Ecosyst Environ 77:247–255.

Thomson D (2004) Competitive interactions between the invasive European honey bee and native bumble bees. Ecology 85(2):458–470.

Toivonen M, Herzon I, Kuussaari M (2015) Differing effects of fallow type and landscape structure on the occurrence of plants, pollinators and birds on environmental fallows in Finland. Biol Conserv 181:36–43. doi:10.1016/j.biocon.2014.10.034

Tooker J, Hauser M, Hanks L (2006) Floral host plants of Syrphidae and Tachinidae (Diptera) of central Illinois. Ann Entomol Soc Am 99(1):96–112.

129

Torné-Noguera A, Rodrigo A, Osorio S, Bosch J (2015) Collateral effects of beekeeping: impacts on pollen-nectar resources and wild bee communities. Basic Appl Ecol 17(3):199– 209. doi:10.1016/j.baae.2015.11.004

Tucker E, Rehan S (2016) Wild bee pollination networks in northern New England. J Insect Conserv 20(2):325–337. doi:10.1007/s10841-016-9870-1

Tucker E, Rehan S (2017a) High elevation refugia for Bombus terricola (Hymenoptera: Apidae) conservation and wild bees of the White Mountain National Forest. J Insect Sci 17(1):1–10. doi:10.1093/jisesa/iew093

Tucker E, Rehan S (2017b) Wild bee community assemblages across agricultural landscapes. J Agr Urban Entomol, 33(1):77-–104. doi:10.3954/1523-5475-33.1.77

Urban-Mead K, Muñiz P, Gillung J, Espinoza A, Fordyce R, van Dyke M, McArt S, Danforth B (2021) Bees in the trees: diverse spring fauna in temperate forest edge canopies. Forest Ecol Manag 482: 118903. doi:10.1016/j.foreco.2020.118903

[USFWS] United States Fish and Wildlife Service (2018) North American Waterfowl Management Plan (NAWMP) update – connecting people, waterfowl, and wetlands. Washington D.C., USA

[USFWS] United States Fish and Wildlife Service (2019) Federal Duck Stamp Program fact sheet. Washington D.C., USA

[USFWS] United States Fish and Wildlife Service (2020) North American Wetlands Conservation Act progress report 2018–2019. Washington D.C., USA

Venables W, Ripley B (2002) Modern applied statistics with S, fourth ed. Springer, New York

Vicens N, Bosch J (2000) Pollinating efficacy of Osmia cornuta and Apis mellifera (Hymenoptera: Megachilidae, Apidae) on ‘Red Delicious’ apple. Environ Entomol 29(2):235–240. doi:10.1603/0046-225X(2000)029[0235:PEOOCA]2.0.CO;2

Vickruck J, Best L, Gavin M, Devries J, Galpem P (2019) Pothole wetlands provide reservoir habitat for native bees in prairie croplands. Biol Conserv 232:43–50. doi:10.1016/j.biocon.2019.01.015

Wagner M (2020) Montezuma Wetlands Complex land protection strategy. Ducks Unlimited, Inc. Syracuse, New York

Walther-Hellwig K, Fokul G, Frankl R, Büchler R, Ekschmitt K, Wolters V (2006) Increased density of honeybee colonies affects foraging bumblebees. Apidologie 37:517–532. doi:10.1051/apido:2006035

130

Warzecha D, Diekötter T, Wolters V, Jauker F (2017) Attractiveness of wildflower mixtures for wild bees and hoverflies depends on some key plant species. Insect Conserv Diver 11(1):32– 41. doi:10.1111/icad.12264

Westphal C, Steffan-Dewenter I, Tscharntke T (2009) Mass flowering oilseed rape improves early colony growth but not sexual reproduction of bumblebees. J Appl Ecol 46:187–193. doi:10.1111/j.1365-2664.2008.01580.x

Whitehorn P, O’Connor S, Wackers F, Goulson D (2012) Neonicotinoid pesticide reduces bumble bee colony growth and queen production. Science 336(6079):351–352. doi:10.1126/science.1215025

Wickham H (2016) ggplot2: elegant graphics for data analysis. Springer-Verlag New York.

Williams B, Brown E (2016) Technical challenges in the application of adaptive management. Biol Conserv 195:255–263. Doi:10.1016/j.biocon.2016.01.012

Williams N, Minckley R, Silveira F (2001) Variation in native bee faunas and its implications for detecting community changes. Conserv Ecol 5(1):1–21.

Winfree R, Bartomeus I, Cariveau D (2011) Native pollinators in anthropogenic habitats. Annu Rev Ecol Evol Syst 42:1–22. doi:10.1146/annurev-ecolsys-102710-145042

Zarrillo T, Ascher J, Gibbs J, Stoner K (2016) New and noteworthy records of bees (Hymenoptera: Apoidea: Anthophila) for Connecticut. J Kansas Entomol Soc 89(2):138–157. doi:10.2317/0022-8567-89.2.138

Zarrillo T, Stoner K (2019) The bee fauna of an Atlantic coastal plain tidal marsh community in southern New England, USA. J Melittology 86:1–34. doi:10.17161/jom.v0i86.7334

Zedler J (2000) Progress in wetland restoration. Tree 15(10):402–407.

Zedler J, Kercher S (2004) Causes and consequences of invasive plants in wetlands: opportunities, opportunists, and outcomes. Crit Rev Plant Sci 23(5):431–452. doi:10.1080/07352680490514673

Zedler J, Kercher S (2005) Wetland resources: status, trends, services, and restorability. Annu Rev Env Resour 30:39–74. doi:10.1146/annurev.energy.30.050504.144248

Zurbuchen A, Landert L, Klaiber J, Müller A, Hein S, Dorn S (2010) Maximum foraging ranges in solitary bees: only few individuals have the capability to cover long foraging distances. Biol Conserv 143:669–676. doi:10.1016/j.biocon.2009.12.003

131

APPENDICES

Appendix 1. Pollinating wasp species recorded in restored wetlands in central New York, 2019 – 2020. Species denoted with † are non-native.

Species No. Caught Floral Interactions Crabronidae Astata sp. 6 - Bembix americana Fabricius 1793 26 4 Bicyrtes quadrifasciatus (Say 1824) 1 1 Cerceris fumipennis Say 1837 1 1 Mimesa sp. 2 1 Mimumesa sp. 1 1 Oxybelus sp. 1 1 Pemphredon sp. 1 - Philanthus gibbosus (Fabricius 1775) 3 2 Saygorytes phaleratus (Say 1837) 1 1 Sphecius speciosus (Drury 1773) 1 1 Tachytes sp. 5 3 Crabronina spp. 220 6 Sphecidae Isodontia mexicana (de Saussure 1867) 6 2 Sphex (Sphex) pensylvanicus Linnaeus 1763 2 2 Thynnidae Myzinum quinquecinctum (Fabricius 1775) 1 - Vespidae Ancistrocerus sp. 1 1 Ancistrocerus adiabatus (de Saussure 1852) 2 - Dolichovespula arenaria (Fabricius 1775) 31 7 Dolichovespula maculata (Linnaeus 1763) 4 1 Eumenes crucifera Provancher 1888 2 2 Parancistrocerus perennis (de Saussure 1857) 2 1 Polistes dominula (Christ 1791)† 2 1 Polistes fuscatus (Fabricius 1793) 3 1 Stenodynerus ammonia (de Saussure 1852) 1 - Symmorphus canadensis (de Saussure 1855) 1 - Vespa crabro Linnaeus 1758† 1 1 Vespula germanica (Fabricius 1793)† 3 1 Vespula maculifrons (du Buysson 1905) 51 7 Vespula vidua (de Saussure 1854) 1 1

132

Appendix 2. Ratio of bee to non-bee pollinators recorded on each flowering plant species swept in restored wetlands in central New

York.

133

Appendix 3. Interaction network illustrating plant-pollinator relationships between wasps and flowers as documented by sweep-netting in central New York wetlands. Width of bars are indicative of the number of recorded interactions.

134

CURRICULUM VITA

135

136

137

138

139