Aspects of the Ecology and Management of Mottled Ducks in Coastal South Carolina

Automated Template B: Created by James Nail 2011V2.1

Aspects of the ecology and management of mottled ducks in Coastal South Carolina

James Claybourne Shipes

A Thesis Submitted to the Faculty of Mississippi State University in Partial Fulfillment of the Requirements for the Degree of Master of Science in Wildlife and Fisheries Science in the Department of Wildlife, Fisheries, and Aquaculture

Mississippi State, Mississippi

December 2014

James Claybourne Shipes

2014

Aspects of the ecology and management of mottled ducks in Coastal South Carolina

James Claybourne Shipes

Approved:

______J. Brian Davis (Major Professor)

______Richard M. Kaminski (Minor Professor)

______Ernie P. Wiggers (Committee Member)

______Eric D. Dibble (Graduate Coordinator)

______George M. Hopper Dean College of Forest Resources

Name: James Claybourne Shipes

Date of Degree: December 13, 2014

Institution: Mississippi State University

Major Field: Wildlife and Fisheries Science

Major Professor: J. Brian Davis

Title of Study: Aspects of the ecology and management of mottled ducks in Coastal South Carolina

Pages in Study: 124

Candidate for Degree of Master of Science

Mottled ducks (Anas fulvigula) are endemic to Gulf Coastal United States and

Mexico. Birds from Florida, Louisiana, and Texas were released in coastal South

Carolina from 1975-1983, and banding data suggest an expanding South Carolina

population. We radio-marked 116 females in August 2010-2011 in the Ashepoo,

Combahee, and Edisto (ACE) Rivers Basin and used radio telemetry to study habitat

selection, searched for nests of non-radiomarked females, and conducted indicated

breeding pair surveys of mottled ducks at various wetlands. Overall, radiomarked

mottled duck females selected managed wetland impoundments, wetlands containing

planted corn, and brackish wetlands. Overall nest success of 42 nests of unmarked

females was 19%. Modeling results indicated that the area of an island on which a nest

was located was the only variable influencing nest success. Indicated breeding pair

surveys revealed that the size of the wetland was the primary influence of breeding

mottled duck immigration into a wetland.

ACKNOWLEDGEMENTS

I am greatly appreciative to countless individuals and agencies vital to the success of this research. First off, I would like to thank Nemours Wildlife Foundation for providing me with funding, housing, and logistical support. I am very grateful to Ernie

Wiggers, Eddie Mills, Justin Rickenbaker, and Robert Kitler who assisted in project design, data collection, and equipment maintenance. I am grateful for financial assistance from South Carolina Ducks Unlimited and South Carolina Department of

Natural Resources, without it this research would not have been possible. I thank all of the public and private land owners and managers for allowing me access to their properties, mainly Lou Crouch, Ross Caterton, and Mark Purcell. I would sincerely like to thank Dean Harrigal for the airboat rides and help in capturing mottled ducks, you always had a way of making airboat rides an experience.

I am grateful for all of the hard work and enthusiasm of my technician Molly

Kneece. The work was not the most glamorous or enjoyable but you were always enthusiastic and made it interesting. Without your help and friendship, this project would not have been nearly as successful. I wish you the best of luck in your future endeavors and am proud of the research you have conducted with your own project. The mottled ducks in South Carolina are better off because of you and the work you have done.

I am thankful to the late Jack Keener Sr. for all of the willingness to fly and search for radio-marked ducks. Even though I didn’t get to fly as much as I would have

wanted to, he always had a way of making it interesting and scary at times. I would like to express my sincerest gratitude to fellow Mississippi State students Adrian Monroe and

Mark McConnell for all of their assistance on my analysis and writing of this thesis.

Their willingness to help and expertise has helped me beyond words can express.

My genuine appreciation goes to Dr. J. Brian Davis for taking me on as a graduate student at MSU and providing me support throughout the entire process. I would also like to express my gratitude to my committee Dr. Ernie Wiggers and Dr. R. M. Kaminski; your comments and encouragement throughout my graduate career were extremely helpful.

Personally, I need to thank my loving parents and brother who have supported and loved me throughout all of my traveling and work. Since I was young I always said that I wanted to be a wildlife biologist like my father. A lot has changed since that time but neither my aspirations nor my parents support and love for following my dreams has wavered. I am very appreciative and thankful for this support and help throughout the process. I would not be where I am today without them. I would also like to sincerely thank my future wife, Brittany. I know there have been many times over the past seven years where you have supported me in my efforts and watched me drive away. I cannot thank you enough for all of support and patience you have shown me. I am very proud of what you have done and accomplished in your nursing profession and the person you have become. I also could not imagine spending my life with a better person and am excited about what the future holds.

Lastly, I would like to express my gratitude to all of my fellow graduate students and friends I have met here in Mississippi. We have made many great memories hunting,

iii

tailgating, and sharing many other experiences. I look forward to many more memories made with you in the future.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... ii

LIST OF TABLES ...... viii

LIST OF FIGURES ...... x

CHAPTER

I. GENERAL INTRODUCTION ...... 1

Mottled Ducks ...... 4 References ...... 10

II. ANNUAL HABITAT SELECTION BY MOTTLED DUCKS IN COASTAL SOUTH CAROLINA ...... 14

Study Area ...... 16 Methods...... 17 Mottled Duck Capture and Tracking ...... 17 Radio Tracking...... 18 Habitat Categorization ...... 18 Statistical Analysis ...... 19 Home Range Analysis...... 20 Results ...... 21 Radio Instrumentation and Tracking ...... 21 Fall-Winter Habitat Use and Selection ...... 21 Spring-Summer Habitat Use and Selection ...... 22 Home Range Estimates of Female Mottled Ducks ...... 23 Discussion ...... 23 Resource Selection ...... 23 Wetlands with Natural Vegetation and Crops ...... 26 Hunted and Non-Hunted Wetlands ...... 26 Home Range Analysis...... 27 Management Implications ...... 28 References ...... 37

III. SOCIAL INDICES OF BREEDING MOTTLED DUCKS IN COASTAL SOUTH CAROLINA ...... 41

Study Area ...... 44 Ernest F. Hollings ACE Basin NWR ...... 44 Bear Island WMA ...... 45 Nemours Plantation Wildlife Foundation ...... 45 Cheeha Combahee Plantation ...... 46 Methods...... 46 Statistical Analysis ...... 49 Model Selection ...... 52 Results ...... 53 Surveys ...... 53 Detection Probability Modeling Results ...... 53 Immigration Modeling Results ...... 54 Discussion ...... 54 Wetland area ...... 55 Wetland Salinity...... 56 Wetland Water Depth ...... 57 Presence or Absence of Islands ...... 59 Management Implications ...... 59 References ...... 69

IV. NESTING ECOLOGY OF MOTTLED DUCKS IN COASTAL SOUTH CAROLINA ...... 75

Study Area ...... 77 Ernest F. Hollings ACE Basin NWR ...... 77 Bear Island WMA ...... 78 Nemours Plantation Wildlife Foundation ...... 78 Cheeha Combahee Plantation ...... 79 Methods...... 79 Mottled Duck Capture and Marking ...... 79 Radio Tracking...... 80 Nest Sites ...... 81 Vegetation Measurements ...... 82 Ancillary Nest Searches ...... 82 Statistical Analysis ...... 83 Nest Site Selection ...... 83 Nest Survival ...... 85 Modeling ...... 86 Results ...... 86 Radio Instrumentation and Tracking ...... 86 Nests, Initiation Dates, and Clutch Size...... 87 Nest Sites ...... 88 Nest Survival ...... 89 Nest Losses ...... 90 Discussion ...... 90 Nest Site Selection ...... 90 vi

Nest Survival ...... 91 Management Implications ...... 97 References ...... 113

vii

LIST OF TABLES

2.1 Habitat types and proportion of each type within the ACE Basin South Carolina from fall-summer 2010-2012...... 29

2.2 Location and home range data for radio-marked mottled ducks within the ACE Basin, South Carolina, fall 2010-summer 2012...... 30

2.3 Habitat categories, selection ratios, and corresponding 95% Bonferroni confidence intervals for radio-marked mottled ducks (n) in the ACE Basin, South Carolina fall-summer 2010-2012...... 31

3.1 Variables used to model immigration (γ) or detection probability (p) of mottled ducks using wetlands within the ACE Basin, South Carolina, spring-summer 2011-2012...... 61

3.2 Candidate models used for detection probability (ρ) in surveys of breeding mottled ducks (Anas fulvigula), ACE Basin South Carolina, spring-summer 2011-2012...... 62

3.3 Candidate models for mottled duck (Anas fulvigula) immigration used for model averaging ...... 63

3.4 Model-averaged log-scaled coefficient estimates (Dail-Madsen models) and 95% confidence intervals of factors potentially influencing immigration and detection probability of mottled ducks (Anas fulvigula) using wetlands in the ACE Basin, South Carolina, spring-summer 2011-2012...... 64

4.1 Mean (± SE; n) for parameters of mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 99

4.2 Predictor variables used to model daily survival rate of mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 99

4.3 Nest site selection candidate models for mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 100

viii

4.4 Model-averaged parameter estimates, odds ratios, and importance weights of parameters used to describe local scale nest site selection of mottled ducks in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 101

4.5 Descriptive statistics (± SE) used to describe mottled duck nests and matched local pair sites in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 102

4.6 Candidate models used to explain daily survival rates of mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 103

4.7 Model averaged estimates and odds ratio estimates of parameters used to describe daily nest survival of mottled ducks nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 104

4.8 Comparisons of estimated nest success for mottled ducks across the birds’ geographic range (adapted from Durham and Afton 2003)...... 105

LIST OF FIGURES

2.1 Location in the ACE Basin, South Carolina, where radio-marked female mottled ducks were studied, 2010-2012...... 32

2.2 Private, state, and federal lands where female mottled ducks were captured, radio marked, and tracked within the Ashepoo, Combahee, and Edisto River Basin, South Carolina, 2010-2012 ...... 33

2.3 Number of live female mottled ducks with functional transmitters and new mortality signals detected within the Ashepoo, Combahee, and Edisto River Basin, South Carolina during monthly periods from August-April 2010 and 2011...... 34

2.4 Proportions of each habitat type within the Ashepoo, Combahee, and Edisto River Basin from fall-summer, 2010-2012...... 35

2.5 Proportion of radio-marked female mottled duck locations within each habitat category in the Ashepoo, Combahee, and Edisto River Basin from fall-summer 2010-2012...... 36

3.1 Sites of mottled duck breeding surveys, ACE Basin South Carolina, spring-summer, 2011-2012...... 65

3.2 Locations of 11 wetlands (inside yellow) surveyed for breeding mottled ducks in the ACE Basin, South Carolina, spring-summer 2011- 2012...... 66

3.3 Designation of social groups of mottled ducks surveyed in wetlands of the ACE Basin, South Carolina, spring-summer 2011-2012...... 67

3.4 Estimated relationship between mottled ducks (Anas fulvigula) immigration (ducks/wetland) and area of a wetland (ha; solid line), with prediction interval (dashed lines), ACE Basin, South Carolina spring-summer 2011-2012...... 68

4.1 Study area (Ashepoo, Combahee, and Edisto River Basin) within South Carolina, where mottled ducks were captured, radio-marked, and tracked 2010-2012...... 106

4.2 Private, state, and federal lands where female mottled ducks were captured, radio-marked, and tracked in the Ashepoo, Combahee, and Edisto River Basin, South Carolina, 2010-2012...... 107

4.3 Influence of island area (m2; ± SE) on successful and failed mottled duck nests and a matched random site (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 108

4.4 Influence of distance to water (m; ± SE) on successful and failed mottled duck nests and matched random site measurement (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 109

4.5 Influence of vegetation height (cm; ± SE) on successful and failed mottled duck nests and a matched random site measurement (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 110

4.6 Influence of percent bare ground (± SE) on successful and failed mottled duck nests and a matched random site measurement (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 111

4.7 Seasonal trend in clutch size for nesting mottled ducks in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012...... 112

CHAPTER I

GENERAL INTRODUCTION

Wetlands of the Atlantic coast of the United States have historically been some of the most important resources for migrating and wintering waterfowl in the Atlantic

Flyway (Bellrose 1976). The South Atlantic Coastal Zone (SACZ) of the Atlantic

Flyway is recognized as one of the most vital regions to wintering waterfowl in eastern

North America (Gordon et al. 1989). South Carolina wintered an average of 30% of the dabbling ducks within the flyway from 1954-1987 (Gordon et al. 1989, Strange et al.

1989). Wetlands of the Ashepoo, Combahee, and Edisto (ACE) River system, or the

ACE Basin, are especially important to waterfowl (Gordon et al. 1989). In fact, coastal wetlands of the ACE Basin are among some of the most critical wetlands to waterfowl along the entire east coast of the Southeastern United States (Miglarese and Sandifer

1982, Gordon et al. 1989).

The Lowcountry of South Carolina is a specific geographic area encompassing parts of four counties along the southern edge of the state (SCDPRT 2014). The ACE

Basin is a 142,000 ha area within the Lowcountry comprised of freshwater wetlands, brackish and tidal wetlands, bottomland hardwood forests, beaches, and other resources

(Miglarese and Sandifer 1982). Like many coastal environments of the U.S., wetlands of the Lowcountry are threatened by loss or deteriorated quality (Dahl 1999). There was an estimated annual net loss of nearly 3,000 acres of wetlands in South Carolina between 1

1982-1989, primarily from urbanization and rural development (Dahl 1999), and approximately 29% of the state’s wetlands have disappeared since 1780 (Yarrow 2009).

Wetlands of coastal South Carolina are especially vulnerable to loss. Georgia, Florida, and South Carolina experienced the greatest levels of human immigration in the U. S. from 1995-2000, and coastal population density increased by 70 percent in these states from 1980-2003 (Franklin 2003). Nearly 25% of South Carolina’s population lives along the coastline today but is expected to climb to 33% by 2033 (U.S. Census Bureau 2008,

London et al 2009). Thus, anthropogenic change associated with coastal development poses a serious threat to coastal wetlands of South Carolina.

Despite potential threats of quality and loss, coastal wetland impoundments are unique and important resources of the South Carolina coast (Gordon et al. 1989). Coastal

impoundments in this region originated from rice cultivation dating to the pre-

Revolutionary War (Kovacik 1979). Rice was originally grown in small inland

impoundments associated with floodplains of freshwater swamps. Small streams were

used as conduits of irrigation, but unpredictable levels of rainfall and stream flow limited

rice production (Heyward 1937). However, in the early 1700s agricultural producers

figured out tidal cycles of Lowcountry wetlands. Farmers took advantage of diel

variation in water levels from tidal influence, which resulted in consistent access to water

for rice crops (Kovacik 1979). At this time, extensive systems of dikes, canals, and

unique water control structures were built using slave labor (Kovacik 1979). The water

control structures used in rice production were called ‘rice field trunks,’ or trunks, and

connected impoundments with tidal rivers. By adjusting trunk gates either on the inside

or outside of the impoundment, managers could gravity flow water to flood or drain their

fields accordingly. This form of rice culture in the Lowcountry prospered until approximately 1860, prior to the Civil War (Kovacik 1979). Ultimately, the demise of

Lowcountry rice culture resulted from several tropical storms that damaged dikes and crops, saltwater intrusion, and a precipitous decline in slave labor (Kovacik 1979,

Rowland 1987, Gordon et al. 1989).

Former rice plantations were soon purchased by wealthy sportsman and managed as hunting properties (Cuthbert and Hoffius 2009). Impoundments were maintained or restored for waterfowl hunting, and management focused on establishing natural wetland vegetation (Strange et al. 1989, Gordon et al. 1989). Today, managed impoundments

(hereafter managed tidal impoundments) remain viable in areas of the ACE Basin and

contribute significantly to quality migration and winter habitats for waterfowl in the

region (Gordon et al 1989).

An estimated 28,500 ha of South Carolina’s 204,000 ha (14%) coastal tidal

marshes are comprised of functioning managed tidal impoundments (Gordon et al. 1989).

Wetlands in the Lowcountry gradate from natural freshwater to brackish, then to salt

marshes; managed tidal impoundments are interspersed within this gradient. Together,

these wetlands create diverse food and other resources not common to natural intertidal

areas (Gordon et al 1998). These resources provides diverse and vital wetland habitats

for numerous species of migrating and wintering waterfowl in the SACZ (Gray et al.

1987). Gordon et al. (1998) found that dabbling ducks selected managed tidal

impoundments over unmanaged tidal wetlands in South Carolina. Moreover, an

estimated 99% of all mottled ducks have been documented using managed tidal

impoundments during their breeding season in the ACE Basin (SCDNR, unpublished data).

Mottled Ducks

In addition to other migrating and wintering waterfowl, managed tidal impoundments of the SACZ are also critical to locally-breeding mottled ducks (Anas

fulvigula). The Mottled Duck is a non-migratory southern relative of the Mallard (Anas

platyrynchos) and American Black Duck (Anas rubripes) and one of >20 “mallard-like” ducks worldwide (Palmer 1976). The mottled duck is also one of two non-migratory dabbling ducks that inhabit North America (McCracken et al. 2001). The Mottled

Duck’s historic range includes peninsular Florida and coastal areas southwestward from

Mobile, Alabama to Vera Cruz, Mexico (Stutzenbaker 1988, Bielefeld et al.2010,

Bellrose 1980). These geographical locations essentially separated mottled ducks into the

Western Gulf Coast and Florida populations (Bielefeld et al 2010). Banding and genetics

data have both demonstrated that these two populations are distinct with no detectable

gene flow between them (McCracken et al. 2001, Williams et al. 2005). Therefore,

currently there are two recognized and distinct populations of mottled ducks in southern

United States, a Louisiana-Texas Gulf Coast and a Florida population (Bielefeld et al

2010).

There has been considerable debate about classification of these two populations

(Davis 2007). Bellrose (1980) considered mottled and Florida ducks as two subspecies

(Anas fulvigula maculosa and Anas fulvigula fulvigula, respectively). Additionally, the

Florida Fish and Wildlife Conservation Commission considers the Florida duck a unique

subspecies in its conservation planning and population management. McCraken et al. 4

(2001) used genetic analysis and differentiated these two populations of mottled ducks.

Apparently there is no detectible gene flow between the populations or at least none in the past several generations (McCraken et al. 2001). These findings are consistent with data from banding records; between 1922 and 1998, 42,369 mottled ducks were banded in the United States, and 4,240 were recovered (U.S. Geological Survey, Biological

Resources Division, Patuxent Wildlife Research Center, Laurel, Maryland). Of these,

2,811 were banded in the Gulf coast region and 1,273 were banded in Florida. No mottled ducks banded in the Gulf coast region were discovered in Florida or vice versa

(McCraken et al. 2001). Despite this information, Moorman and Gray (1994) and the

Birds of North America (Bielefeld et al. 2010) recognize only one subspecies, Anas

fulvigula, highlighting the importance of further research on the genetic variation and

subsequent classification of mottled ducks in these two regions.

Wetland managers for the South Carolina Department of Natural Resources

(SCDNR) banded and released approximately 1,285 Mottled Ducks at the Santee River

Delta in Charleston and Georgetown Counties, and the lower ACE Basin in Beaufort,

Charleston, and Colleton Counties, South Carolina from 1975-1983 (South Carolina

Department of Natural Resources [DNR], unpublished data). Most released birds

originated from Louisiana and Texas but 26 individuals were imported from Florida

(South Carolina DNR, unpublished data). Of the original mottled ducks banded and

released in South Carolina, 107 (8%) were later identified either through direct or indirect

band recoveries from 1975-1986 (South Carolina DNR, unpublished data). Only 7 (6%)

of these ducks were recovered outside of South Carolina. Hence, mottled ducks

originally released in South Carolina have demonstrated strong fidelity to that state and neighboring ones, such as Georgia (South Carolina DNR, unpublished data).

Breeding populations of Florida mottled ducks exhibit considerable variation over brief temporal periods, such as 5-7 years, which likely occurs because of drought

(Bielefeld et al 2010). However, recent analysis of historical population survey data demonstrates slight population growth of Florida mottled ducks from 1985-2006

(Bielefeld 2006). In Louisiana, Midwinter survey data indicate a stable breeding population of mottled ducks (0.7% annual increase; 1971-2009) but a slightly decreasing trend in Texas (2.8% annual decline; Bielefeld et al 2010). Nonetheless, the Louisiana and Texas populations of mottled ducks, and Florida populations, both demonstrate either stability or a slight increase across these historic ranges (Bielefeld et al 2010).

Beginning in 2008, the South Carolina DNR prioritized banding of mottled ducks

in coastal South Carolina. Current banding data suggest that the South Carolina mottled

duck population may be as great as 15,000-20,000 individuals (South Carolina DNR,

unpublished data). Moreover, mottled ducks currently comprise 4% of total duck harvest

on state WMAs; South Carolina DNR, unpublished data). Although banding data have

not been rigorously analyzed at this time, the data suggest an expanding population of

mottled ducks in South Carolina. This trend also suggests that wetland resources must be

sufficient to support some level of locally-breeding and wintering mottled ducks in South

Carolina.

Despite apparent growth in this population, virtually nothing is known about the

annual ecology of the species in South Carolina. Weng (2006) studied genetics and

habitat use of mottled ducks on Bear Island Wildlife Management Area and determined

that mottled ducks showed a preference for shallow wetlands (3-45 cm, average = 14.24 cm) likely because of the availability of Chironimidae invertebrates in them. In northern prairie environments of the northern United States and Canada, wetlands with a near equal interspersion of water and wetland vegetation, deemed a hemi-marsh, supported more waterfowl than wetlands containing different ratios of vegetation and water

(Murkin et al. 1992). In contrast, mottled ducks at Bear Island WMA seemed to be linked more to vegetation edges of wetlands regardless of the ratio of interspersed open water and vegetation (Weng 2006).

Despite the dearth of science-based knowledge on mottled duck ecology and management in South Carolina, I am personally aware that the bird is popular among wetland land managers, hunters, and bird watchers. Several studies of habitat ecology during the breeding season and winter have been conducted elsewhere in the species range (Stutzenbaker 1988, Zwank et al 1989, Johnson 1991, Davis 2012, Moon et al

2012). Although many of these studies will be useful to me in my research on mottled ducks in South Carolina, wetlands occupied by the bird in the ACE Basin are unique relative to other regions. For example, rice agriculture is still prominent in the Louisiana

Coastal Plain and Texas Mid-Coast regions, where producers planted an average of

129,240 ha and 45,292 ha of rice per year, respectively, from 2000-2010 (USDA 2010).

Previous research in these two regions investigated selection of agricultural habitats by mottled ducks (Zwank et al 1989, Durham and Afton 2006, Haukos et al 2010). Mottled ducks reportedly used flooded agricultural fields considerably during the breeding season, but general habitat use was quite variable the rest of the year (Stutzenbaker 1988,

Holbrook et al 2000, Durham and Afton 2006, Bielefeld et al 2010). Other habitats

considered important to mottled ducks in the western gulf coast include coastal marsh, cattle pastures, shallow bays, and grassland prairies (Stutzenbaker 1988, Durham and

Afton 2006, Bielefeld et al. 2010, Davis 2012). Amid coastal wetland habitats, mottled ducks apparently select brackish and intermediate marshes over other coastal habitats

(Davis 2012, Moon et al 2012).

Historically, mottled ducks inhabited fresh emergent and brackish wetlands inland and along both coastlines of Florida, but today these birds use a variety of habitats because of wetland loss and degradation (Bielefeld et al. 2010). It is estimated that >50% of mottled ducks in Florida use urban/suburban wetlands such as retention ponds and drainage ditches (Bielefeld 2008, Bielefeld et al. 2010). In more rural regions of Florida, such as the Everglades Agricultural Area, mottled ducks use seasonally flooded ponds and marshes associated with major rivers and lakes, storm-water treatment areas, the

Everglades, and the Everglades Agricultural Area (Johnson et al. 1991, Bielefeld 2008,

Bielefeld et al. 2010). Lotter (1969) speculated that mottled ducks in Florida only moved to coastal areas in response to dry conditions in the prairie regions of the state. In South

Carolina, Weng (2006) found that mottled ducks preferred 3-45 cm water depths during summer at Bear Island WMA. This behavior may have been related to acquisition of aquatic invertebrates, and possibly as a mechanism to avoid depredation by the American

Alligator (Alligator mississippiensis; Weng 2006).

From 1983-1986, the South Carolina DNR systematically conducted aerial

surveys during the presumed breeding season to determine habitat use by these birds.

Approximately 99% of mottled ducks were observed in brackish (5-20 ppt.) and saline

(>20 ppt.) managed tidal impoundments (South Carolina DNR). Despite the sometimes

high salinities, these results suggest the importance of remnant rice field impoundments in the annual cycle of mottled ducks in South Carolina (SCDNR unpublished data).

The primary objectives of my study were at least four-fold: (1) identify movements and habitats used (coarse-scale, i.e., 2nd Order Selection; Johnson 1980) by

radio-marked mottled ducks during the non-breeding period (Chapter 1), (2) estimate

indicated breeding pairs (IBPs) of local mottled ducks (Chapter 2), (3) determine habitat

characteristics (3rd and 4th Order selection; Johnson 1980) selected and study subsequent

nesting ecology of Mottled Ducks (Chapter 3), and (4) identify and recommend habitat

management practices that benefit nesting Mottled Ducks in the ACE Basin.

References

Bellrose, F. C. 1976. Ducks, geese, and swans of North America. Stackpole, Harrisburg, USA.

Bellrose, F. C. 1980. Ducks, geese, and swans of North America. Third Edition. Stackpole Books, Harrisburg, Pennsylvania, USA.

Bielefeld, R. R. 2008. Mottled duck survey report. Fish and Wildlife Research Institute.

Bielefeld, Ronald R., Michael G. Brasher, T. E. Moorman and P. N. Gray. 2010. Mottled Duck (Anas fulvigula), The Birds of North America (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology; http://bna.birds.cornell.edu/bna/species/081 (Accessed December 12, 2013).

Bielefeld, R.R. & Cox, R.R., Jr. 2006. Survival and cause-specific mortality of adult female mottled ducks in east-central Florida. Wildlife Society Bulletin 34: 388– 394.

Cuthbert, Robert B. and Stephen G. Hoffius, Eds. (2009). Northern Money, Southern Land: The Lowcountry Plantation Sketches of Chlotilde R. Martin. Columbia: University of South Carolina Press.

Dahl, T. 1999. South Carolina’s wetlands: status and trends, 1982-1989. http://www.fws.gov/wetlands/_documents/gSandT/StateRegionalReports/SouthC arolinaWetlandsStatusTrends1982to1989.pdf (Accessed October 7, 2013).

Davis, B. E. 2007. Habitat use, movements, and survival of radiomarked female mallards in the lower Mississippi Alluvial Valley. Thesis, Louisiana State University, Baton Rouge, USA.

Davis, B. E. 2012. Habitat use, movements, and ecology of female mottled ducks on the gulf coast of Louisiana and Texas. Dissertation, Louisiana State University, Baton Rouge, USA.

Durham, R. S. and A. D. Afton. 2006. Breeding biology of mottled ducks on agricultural lands in southwestern Louisiana. Southeastern Naturalist 5:311-316.

Franklin, R. S., U.S. Census Bureau. 2003. DomesticMigration Across Regions, Divisions, and States: 1995-2000. Washington, DC: DOC, U.S. Census Bureau. Available from:(accessed December 12, 2003).

Gordon, D. H., B. T. Gray, R. D. Perry, M.B. Prevost, T. H. Strange, and R. K. Williams. 1989. South Atlantic Coastal Wetlands. Pages 57–92 in L. M. Smith, R. L. Pederson, and R. M. Kaminski, editors. Habitat management for migrating and wintering waterfowl in North America. Texas Tech University Press, Lubbock, Texas.

Gordon, D. H., B. T. Gray, and R.M. Kaminski 1998. Dabbling duck habitat associations during winter in coastal South Carolina. Journal of Wildlife Management 62: 569- 580.

Haukos, D., S. Martinez, and J. Heltzel. 2010. Characteristics of ponds used by breeding mottled ducks on the chenier plain of the Texas gulf coast. Journal of Fish and Wildlife Management 1:93-101.

Heyward, Duncan C. 1937. Seed from Madagascar. Chapel Hill: University of North Carolina Press.

Hoekman, S. T., T. S. Gabor, M. J. Petrie, R. Maher, H. R. Murkin, and M. S. Lindberg. 2006. Population dynamics of mallards breeding in agricultural environments in eastern Canada. Journal of Wildlife Management 70:121–128.

Holbrook, R. S., F. C. Rohwer, and W. P. Johnson. 2000. Habitat use and productivity of mottled ducks on the Atchafalaya River Delta, Louisiana. Proceedings of Annual Conference of Southeastern Association of Fish and Wildlife Agencies 54:292- 303.

Holbrook, R. S. 1997. Ecology of nesting mottled ducks at the Atchafalaya River delta, Louisiana. M.Sc. Thesis, Louisiana State University, Baton Rouge, USA.

Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65–71.

Johnson, D. H., G. L. Krapu, K. J. Reinecke, and D. G. Jorde. 1985. An evaluation of condition indices for birds. Journal of Wildlife Management 49:569-575.

Johnson, F. A., F. Montalbano III, and T. C. Hines. 1984. Population dynamics and status of the mottled duck in Florida. Journal of Wildlife Management 48:1137–1143.

Johnson, F. A., F. Montalbano III, J. D. Truitt, and D. R. Eggeman. 1991. Distribution, abundance, and habitat use by mottled ducks in Florida. Journal of Wildlife Management 55:476–482.

Johnson, W.P., F. C. Rohwer, and M. Carloss. 1996. Evidence of nest parasitism in mottled ducks. Wilson Bulletin 108:187-189.

Kovacik, C. F. 1979. South Carolina rice coast landscape changes. Proceedings Tall Timbers Ecology Management Conference 16:47-65.

LaHart, D. E., and G. W. Cornwell. 1970. Habitat preference and survival of Florida duck broods. 24th Annual Conference of the Southeastern Association of Game and Fish Commissioners 117-120.

London, J. B., C. S. Dyckman, J. S. Allen, C. C. St. John, I. L. Wood. 2009. An assessment of shoreline management options along the South Carolina coast. Report Submitted to: Office of Ocean and Coastal Resource Management, South Carolina Department of Health and Environmental Control.

Lotter, C. F. 1969. Habitat requirements and procedures for measuring various population parameters of the Florida duck, Anas platyrhynchos fulvigula. Ridgeway, M.S. Thesis, University of Florida, Gainesville, Florida, USA.

McCracken, Kevin G., William P. Johnson, and Frederick H. Sheldon. 2001. Molecular population genetics, phylogeography, and conservation biology of the mottled duck (Anas fulvigula). Conservation Genetics 2: 87-102.

Miglarese, J. V., and P. A. Sandifer, editors. 1982. An ecological characterization of South Carolina wetland impoundments. South Carolina Wildlife and Marine Resources Department, Marine Re-sources Research Institute Technical Report Number 51.

Moon, J. A. 2012. Movements, Habitat Use, Survival, and potential implications of Climate Change on Mottled Ducks (Anas fulvigula) in the Texas Chenier Plain Region. Dissertation, Stephen F. Austin State University, Nacogdoches, USA.

Moorman, T. E. and P. N. Gray. 1994. Mottled Duck. In Poole, A. and F. Gill, editors. (Eds.). The Birds of North America. Academy of Natural Science. Philadelphia, PA. 19 pp.

Moorman, A. M., T. E. Moorman, G. A. Baldassarre, and D. M. Richard. 1991. Effects of saline water on growth and survival of mottled duck ducklings in Louisiana. Journal of Wildlife Management 55: 471-476.

Murkin, E. J., Murkin, H. R., & Titman, R. D. (1992). Nectonic invertebrate abundance and distribution at the emergent vegetation-open water interface in the Delta Marsh, Manitoba, Canada. Wetlands 12: 45-52.

Palmer, Ralph S. 1976. Handbook of North American Birds. Vol. 2, Pt. 1. Yale Univ. Press, New Haven.

Paulus, S. L., and M. W. Weller. 1988. Social behavior and pairing chronology of mottled ducks during autumn and winter in Louisiana USA. Waterfowl in winter; symposium. University of Minnesota Press: Minneapolis, MN. 59-70.

Rowland, Lawrence S. 1987. “Alone on the River:” The Rise and Fall of the Savannah River Rice Plantations of St. Peter’s Parish, South Carolina. The South Carolina Historical Magazine 88: 121-150.

South Carolina Department of Parks Recreation and Tourism. 2014. Resort islands and lowcountry of South Carolina. < http://www.discoversouthcarolina.com/see- do/explore-the-state/islands-coasts/lowcntry-resort-isles/default.aspx> (Accessed November 2 2013).

Strange, T. H., D. H. Gordon, and J. A. Timmerman, Jr. 1989. Atlantic Flyway perspectives and expectations. Transactions of the North American Wildlife and Natural Resources Conference 54: 496-511.

Stutzenbaker, C. D. 1988. The Mottled Duck: Its Life History, Ecology, and Management. Texas Parks and Wildlife Department. Austin, TX. 209 pp.

U.S. Bureau of the Census (1). (2008). Projections of the Total Population of States: 1995 to 2025.

U. S. Department of Agriculture [USDA]. 2010. National Agricultural Statistics Service [NASS]. Quick Stats. < http://quickstats.nass.usda.gov/>. Accessed 25 July 2010.

Weng, G. J. 2006. Ecology and population genetics of mottled ducks within the south Atlantic coastal zone. M.Sc. Dissertation. University of Georgia. Athens, GA. 100.

Williams C. L., Fedynich A. M., Pence D. B., Rhodes O. E. Jr. 2005. Evaluation of allozyme and microsatellite variation in Texas and Florida Mottled Ducks. Condor, 107, 155–161.

Williams, C. L., R. C. Brust, T. T. Fendley, G. R. Tiller Jr., and O. E. Rhodes Jr. 2004. A comparison of hybridization between mottled ducks (Anas fulvigula) and mallards (A. platyrhyncos) in Florida and South Carolina using microsatellite DNA analysis. Conservation Genetics 6:445-453.

Yarrow, G. 2009. Wetland Management and Protection. Clemson Cooperative Extension, Extension Forestry and Natural Resources.

Zwank, P. J., P. M. McKenzie, and E. B. Moser. 1989. Mottled duck habitat use and density indices in agricultural lands. Journal of Wildlife Management 53:110-114.

CHAPTER II

ANNUAL HABITAT SELECTION BY MOTTLED DUCKS IN COASTAL SOUTH

CAROLINA

The annual cycle of waterfowl consists of several physiological and behavioral

distinct but interconnected events and periods, including breeding, summer molting, fall

and spring migration, and winter (Bellrose 1980, Smith et al. 1989, Batt et al. 1992,

Baldassarre and Bolen 2006). Most Nearctic species of dabbling ducks nest at northern

latitudes and winter in central to southern United States or farther south in North, Central,

or South America (Bellrose 1980, Baldassarre and Bolen 2006). However, some species

of dabbling ducks fulfill annual-cycle events in comparatively restricted geographic

areas. For example, individuals of western (e.g., California) breeding mallards (Yarris et

al. 1994, McLandress et al. 1996), southern populations of wood ducks (Aix sponsa;

Bellrose and Holm 1994) and hooded mergansers (Baldassarre 2014), and Gulf coastal and Florida mottled ducks live year-round in local to regional areas (Stutzenbaker 1988,

Bielefeld and Brasher 2010). Indeed, knowledge of annual autecology is critical to meet annual-cycle needs of all waterfowl, including non-migratory species such as mottled ducks (Anas fulvigula) (Wilson 2007, Bielefeld and Brasher 2010).

Mottled ducks are endemic to peninsular Florida and coastal regions of Alabama,

Mississippi, Louisiana, and Texas (Bielefeld and Brasher 2010). Wildlife managers and private conservationists from South Carolina and elsewhere translocated 1,285 mottled 14

ducks from Louisiana, Texas, and Florida to two coastal regions of South Carolina from

1975-1983, because the species did not occur there in the 1970s, yet available habitat was believed suitable for the species (South Carolina DNR, unpublished data). Mottled duck populations in South Carolina have grown and the birds have dispersed across South

Carolina and Georgia (Weng 2006). An estimated 15,000-20,000 mottled ducks are believed to occur in South Carolina since 2010 (South Carolina DNR, unpublished data).

Mottled ducks have been studied throughout their endemic range (Stutzenbaker

1988, Zwank et al. 1989, Johnson et al. 1991, Davis 2012, Moon 2012, Varner 2013).

Unlike South Carolina, rice agriculture remains in the Louisiana Coastal Plain and Texas

Mid-Coast, and mottled ducks readily use flooded ricelands (Davis 2012). Although mottled ducks use flooded ricelands during the breeding season, a complex of agricultural and other habitats and resources, including coastal marsh, pastures, and prairies are important for non-breeding birds (Stutzenbaker 1988, Holbrook et al. 2000, Durham and

Afton 2006, Bielefeld et al. 2010). Mottled ducks apparently select brackish and fresh marshes over other coastal habitats within the Western Gulf Coast (Davis 2012, Moon

2012).

Mottled ducks seem most adapted to coastal, wet prairie, and largely tree-less

wetlands or agricultural lands. Wintering dabbling ducks use managed tidal

impoundments and unmanaged natural tidal marsh habitats in coastal South Carolina, but

greatest densities of dabblers occurred in managed wetlands (Gordon et al. 1989, Gordon

et al. 1998). Natural marshes of coastal South Carolina exhibit a semi-diurnal tidal cycle

where water depths fluctuate ≈ 2.29 m (Ricketts et al. 2011). In contrast, Gulf coastal

tidal marshes have an amplitude of <1m (Dardeu et al 1992). Gulf coastal mottled ducks

use flooded pastures, associated wetlands, and flooded croplands, neither of which occur in the coastal region of South Carolina. Apparently, only one study of breeding mottled ducks has been conducted in coastal South Carolina, which reported that birds preferred water depths of 3-45 cm in managed tidal impoundments (Weng 2006). However, Weng

(2006) did not study other potentially important habitat covariates such as vegetation communities, wetland size, and wetland juxtaposition within habitats used in the landscape.

Given the apparent growing populations of mottled ducks in South Carolina, and, as part of a larger study of annual autecology of mottled ducks, we initiated exploratory research to assess habitat use of these birds during fall-spring. Wetland managers also desire information on specific habitat needs of mottled ducks during the annual cycle in this region to guide landscape habitat conservation planning and management for mottled ducks and other waterbirds. Herein, we examined coarse-scale habitat use (i.e., 2nd Order

Selection; specific habitat availability versus use by mottled ducks within the region;

Johnson 1980). Following Gordon et al. (1989, 1998), we predicted a priori that mottled

ducks would be associated with managed tidal impoundments in the Ashepoo,

Combahee, and Edisto (ACE) Rivers Basin.

Study Area

We studied in the 182,115 ha estuarine complex of the ACE Basin located in the

southern half of South Carolina’s coastline in Beaufort, Charleston, and Colleton

Counties (Figure 2.1). The ACE Basin is one of 27 national research reserves bio-

monitored by the National Estuarine Research Reserve System and contains 128,000 ha

of land protected by state, federal, private, and nonprofit organizations (NOAA 2012). 16

The ACE Basin contains pine and hardwood uplands, forested wetlands, fresh, brackish and salt water tidal marshes, barrier islands, and beaches (SCDNR 2013). Our study included wetlands and associated habitats on private, state, and federal lands including

Ernest F. Hollings ACE Basin National Wildlife Refuge (NWR), Bear Island Wildlife

Management Area (WMA), Nemours Plantation Wildlife Foundation, and Cheeha

Combahee Plantation (Figure 2.2).

Methods

Mottled Duck Capture and Tracking

We captured mottled ducks on our study area using night-lighting techniques

during June-September 2010-2011 when most birds were wing-molting (Merendino et al.

2005, Mills et al. 2011). We transported all captured females to the laboratory at

Nemours Plantation and prepared them for radio-marking. We outfitted hatch year and

after hatch year females either with a harness style transmitter (21 g; Advanced

Telemetry Systems model A2300, Isanti MN) or an intra-abdominal transmitter (18 g;

Holohil Systems Model RI-2D, Carp, Ontario, Canada) in 2011, but only used intra-

abdominal transmitters in 2012, because of apparent mass failure of transmitters in 2011.

We used a team of skilled veterinarians to implant transmitters, details of which were

provided by Shipes (unpublished data). We followed Dwyer’s (1972) method for

attaching harness transmitters and Korschgen et al. (1996) for intra-abdominal

transmitters. Both transmitters were < 3% of the bird’s body mass at capture and had an

estimated 30-month lifespan. Overall 115 females were transmitted, with a mean bird

mass of 752.7 ± 6.3 g. Each transmitter was equipped with a mortality sensor that

doubled the pulse rate when the transmitter was motionless > 8 hours and likely 17

indicative of the birds’ death. We banded each mottled duck with a standard USGS aluminum leg band.

Radio Tracking

We began monitoring radio-marked females’ 5-days post release. Weekly from

August-December 2010 and 2011, we located radio-marked females using an aircraft

equipped with strut-mounted 4-element Yagi antennas (Gilmer et al. 1981). Beginning in

January 2011 and 2012, we located radio-marked birds >3 times per week from a truck

using an ATS receiver (Model R4000, Advanced Telemetry Systems, Inc. Isanti,

Minnesota) and a handheld 3-element Yagi antenna. Upon detecting a mortality signal,

we attempted to retrieve the transmitter and determine cause of bird’s death (Sargeant et

al. 1998).

Habitat Categorization

We used ArcMap 10 and overlaid bird locations onto digital orthophoto images of

the study area and created a 100% minimum convex polygon (MCP) to determine area

used by radio-marked mottled ducks (Legagneux et al. 2008). We created polygons

within the 100% MCP to differentiate managed tidal impoundments from natural tidal

marsh. We specified several different habitat and management categories for the

polygons: 1) Management regime; i.e., whether or not managed tidal impoundments

received active hydrological management by managers, 2) salinity regime; i.e.,

intermediate (1-5 ppt.), brackish (5-15 ppt.), and brackish/salt (15-25 ppt.), 3) hunting

regime; whether waterfowl hunting occurred or did not within a wetland, and 4)

Vegetation coverage; whether wetlands contained natural wetland vegetation only or

natural vegetation and agricultural crops, such as corn. In GIS, we assigned each of these four categorical variables to individual polygons and used the join function in ArcMap 10 to assign or link categories and individual female locations. Lastly, we calculated the total area of each category within the 100% MCP so proportions of each could be determined for subsequent analyses of habitat selection.

Statistical Analysis

We assessed habitat selection by female mottled ducks using the Manly et al.

(2002) selection ratio approach, where: wj = uj / aj and uj was the proportion of use of the

habitat class j and aj is the proportion of this habitat assumed available to individual

mottled ducks within their MCP. Habitat selection ratios are a simplified case of

resource selection functions where each resource unit is classified into distinct categories

(Manly et al. 2002). As described, our distinct resource units included hydrological

management, salinity category, vegetation types within wetlands, and presence or

absence of hunting. We examined selection of these four resource units for two different

time periods: 1) August- 31 January, which encompassed post-breeding through mid-

winter periods, including the waterfowl hunting season; and 2) 1 February- 1 June (i.e.,

post-waterfowl hunting through breeding season).

We used chi–square analysis to determine whether female mottled ducks used

resource units proportionately similar and whether habitat use occurred in proportion to

availability (Manly et al. 1993, Rogers and White 2007). We then calculated selection

ratios and accompanying 95% Bonferroni confidence intervals to determine if specific

resource units were used disproportionally to their estimated and assumed availability,

implying units were selected or alternatively avoided by mottled ducks (Thomas and 19

Taylor 1990, Manly et al. 2002, Rogers and White 2007). Selection ratio values greater than one imply selection for a category while ratio values less than one suggest avoidance. Any 95% Bonferroni confidence intervals that include one demonstrate no apparent selection. We used program Fishtel (v 1.4) to calculate chi-square statistics and selection ratios (Rogers and White 2007). This aspect of the Manly et al. (1993) analysis does consider individual variation associated with resource selection, but it allows pooled observations among all individuals in a sample population. Therefore, resulting selection ratios are provided at a population and not individual level (Manly et al. 1993).

For the 1 August- 31 January and the 1 February- 1 June periods, we evaluated habitat selection by mottled ducks relative to hydrological management and salinity. We evaluated habitat selection relative to presence or absence of hunting only during the regular South Carolina waterfowl hunting season (i.e., 20 November - 30 January). We evaluated habitat selection relative to wetlands only with natural vegetation versus those with agricultural crops and natural vegetation for the period 15 November-15 February, when we were certain crops were available to mottled ducks.

Home Range Analysis

In addition to MCP, I used the fixed kernel method only to get an estimate of

home range sizes for 10 radio-marked females for basic knowledge (Worton 1989). I

used Hawth’s Tools (Beyer 2004) for ArcGIS 9.3 to calculate the 95% fixed kernel

estimator. Kernel methods free the utilization distribution estimate from parametric

assumptions and provide a means of smoothing location data (Giles 2010, Worton 1989).

They also have well-understood consistent statistical properties and are used in both

univariate and multivariate probability density estimation (Worton 1989). I calculated 20

likelihood cross validation (CVh) smoothing parameters for fixed kernel home ranges. I

selected likelihood cross validation because it tends to produce home range estimates

with a better fit and less variability then least squares cross validation (LSCVh) (Worton

1989).

Results

Radio Instrumentation and Tracking

We radio-marked 80 and 36 female mottled ducks from August-September 2010

and 2011, respectively. Because of failed transmitters (primarily backpacks), non-

detection of radio-marked females in the study area after their release, or their death, we

obtained 1,241 locations from 67 females (58%) from fall-spring 2010-2012 and used

these data in analyses for habitat selection. All subsequent results herein are from

females with implanted transmitters.

Fall-Winter Habitat Use and Selection

We obtained data from 67 females (n = 693 locations) to investigate habitat selection during fall-winter relative to hydrologic management and salinity category.

Females used hydrologically managed wetlands >9 times more than unmanaged wetlands

2 (χ 66 = 92.15, P = 0.01; Table 1). Additionally, females’ use of managed wetlands

2 exceeded availability of these wetlands (χ 67 = 606.40, P < 0.001), implying females

selected these habitats (wi = 2.02 [1.95, 2.09]) but avoided unmanaged wetlands (wi =

0.21 [0.15, 0.27]); Table 1).

2 Females also differentially used wetlands relative to salinity designation (χ 132 =

2 321.90, P < 0.001), and use was disproportional to availability of the wetlands (χ 133 =

612.97, P < 0.001). Females were 12 times more likely to select brackish over

intermediate wetlands and 5 times more likely to select brackish over brackish/saline

wetlands (Table 1).

We obtained data from 55 females (n = 354 locations) to investigate habitat selection in relation to presence or absence of waterfowl hunting in wetlands used by females during South Carolina’s waterfowl hunting season. Females differentially used

2 hunted and non-hunted wetlands (χ 54 = 110.0, P < 0.001), and use was disproportional

2 to availability (χ 55 = 112.2, df = 55, P < 0.001). However, the Bonferroni 95% confidence interval included 1; therefore, we could not infer females avoided hunted wetlands (Table 1).

We evaluated selection by 55 radio-marked females (n = 364 locations) for

wetlands with natural vegetation versus those with a combination of crops and native

2 vegetation. Females used both types dissimilarly (χ 54 = 78.9, P = 0.015) and

2 disproportionally to their availability (χ 55 = 93.5, P < 0.001). Selection ratios indicated

that females were 2 times more likely to select wetlands containing crops than only

natural vegetation (Table 1).

Spring-Summer Habitat Use and Selection

We used data from 36 radio-marked females (n = 548 locations) from 1 February-

1 June to investigate selection of habitats relative to hydrological management and

salinity. Crops in wetlands generally were toppled, decomposed, or eliminated by

managers soon after the waterfowl hunting season, thus only leaving wetlands with native

vegetation and none others for comparison of duck use. Females differentially used

2 managed and unmanaged wetlands (χ 35 = 755.3, P < 0.001), and use was disproportional 22

2 to availability of these wetlands (χ 36 = 773.7, P < 0.001). Female mottled ducks were

>30 times more likely to select managed wetlands (wi = 2.204 [2.155, 2.252]) over

unmanaged wetlands (wi = 0.065 [0.027, 0.103]; Table 1).

2 Female mottled ducks differentially used wetlands of varying salinities (χ 68 =

277.1, P < 0.001) throughout the spring and summer. Females use of brackish wetlands

2 exceeded availability of these wetlands (χ 69 = 722.5, P < 0.001), implying females

selected brackish wetlands (wi = 2.038 [1.81, 2.27]) but avoided intermediate (wi = 0.16

[0.03, 0.29]) and brackish/salt wetlands (wi = 0.39 [0.07, 0.73]; Table 1).

Home Range Estimates of Female Mottled Ducks

I used 591 locations obtained from 10 different female mottled ducks to estimate

home range size. These 10 female mottled ducks were the only radio-marked birds with

>30 locations. Estimates of 95% fixed kernel home range size ranged from 1,004 ha to

6,644 for these 10 females, with 50% core areas ranging in size from 204 ha to 1,045 ha.

95% fixed kernel home range estimates averaged 3,217 (± 608) ha, while 50% core areas

averaged 628 ha. (Table 1.2).

Discussion

Resource Selection

Our study was the first to demonstrate empirically habitat selection by radio- marked female mottled ducks in South Carolina. The importance of managed, brackish wetlands to mottled ducks in the ACE Basin supported our predictions and reinforced

Gordon et al. (1989) recognition of the importance of managed brackish wetlands to non- breeding waterfowl in the region. Our findings also corroborate habitat associations of

breeding season mottled ducks observed during aerial surveys in the Santee River Basin

(South Carolina Department of Natural Resources [SCDNR], unpublished data). Pilot

biologists estimated that approximately 99% of all mottled ducks observed during 1983-

1986 occurred in managed brackish (5-20 ppt.) and saline (>20 ppt) wetlands (SCDNR

unpublished data). Gordon et al. (1998) also reported more migrating and wintering

dabbling ducks in managed tidal impoundments than unmanaged tidal wetlands. Indeed,

selection of managed brackish wetlands by radio-marked mottled ducks in our study

seemed consistent across fall-winter and spring-summer seasons.

Managed tidal impoundments and unmanaged tidal wetlands in South Carolina

differ considerably, especially in vegetation communities and surface-water area (Gordon

et al. 1989). Managed tidal impoundments are typically shallowly flooded (< 1 m).

Water levels are managed in summer to promote growth of wetland emergent and

submerged aquatic plants, such as dwarf spikerush (Eleocharis parvula) and wigeongrass

(Ruppia maritima; Gordon et al. 1989). Foliage and seeds of these species provide

quality food for wintering dabbling ducks in this region (Prevost 1978, Gordon et al.

1989). Moreover, vegetation in managed tidal impoundments is often actively

manipulated to create interspersed herbaceous vegetation and open water (i.e., hemi-

marsh) that enhances access and other activities by wintering waterfowl (Gordon et al.

1989, Gordon et al. 1998, Smith et al. 2004). In contrast, unmanaged tidal areas in our

study area were dominated by smooth cordgrass (Spartina alterniflora), with black

needlerush (Juncus roemarianus) forming largely homogenous patches in higher

elevations. Also, saltmarsh bulrush (Schoenoplectus robustus), giant cordgrass (Spartina

cynosuroides), and soft-stemmed bulrush (Schoenoplectus tabernaemontani) are

interspersed with Spartina alterniflora (Ricketts 2011). Unmanaged tidal wetlands undergo a semidiurnal tide cycle (NOAA 2012) where daily frequency, duration, and depth of water vary spatially and temporally. These natural dynamics often promote tall, dense wetland vegetation with limited open water that is difficult to access by waterfowl

(Kaminski and Prince 1981, Gordon et al. 1998). Consequently, most open water in these

tidal systems is associated with river and creek channels, which are relatively deeply

flooded (>1 m; Gordon et al 1998).

Mottled ducks in Louisiana and Texas may shift habitat use from marsh and wet

prairie habitats during winter and pre-breeding periods to agricultural lands (e.g., rice)

during the breeding seasons (Stutzenbaker 1988, Zwank et al. 1989, Davis 2012). We did

not observe a similar shift in habitat selection from non-breeding to breeding periods in

our study. Wetlands with planted crops (e.g., corn) were most important to mottled ducks

during winter, and these resources were unavailable during breeding because wetlands

are drained post-hunting season in preparation for subsequent management. We suspect

that relatively stable water conditions in managed brackish and saline wetlands unlike

that of natural tidal marshes contributed to great use of brackish-saline wetlands by

mottled ducks.

Across much of their range, mottled ducks seemingly select fresh and

intermediate wetlands over brackish and brackish/salt wetlands during winter and

breeding seasons (Johnson et al. 1991, Davis 2012, Wehland 2012). However, an

apparent affinity by mottled ducks in the ACE Basin for brackish wetlands may be

related to vegetation composition and water management. Intermediate wetlands of the

ACE Basin are dominated by early successional and herbaceous emergent plants such as

rice cutgrass (Leersia oryzoides) and panic grasses (Panicum spp.). Management strategies in these wetlands prioritize quality moist-soil plant communities for fall migrating and wintering waterfowl (Kross et al. 2008, Schummer et al. 2012). However, when mottled ducks use these wetlands in summer, likely while brooding, vegetation is dense and uniform, seeds have not yet matured and shed, and water levels are very shallow (< 0.3 m).

Wetlands with Natural Vegetation and Crops

Female mottled ducks in our study appeared to select wetlands with planted crops, primarily corn, over those with only natural vegetation during fall-winter (15 November-

15 February). Davis (2012) observed female mottled ducks shifting from natural wetlands to ricelands from winter to the breeding period. We likely did not observe this trend because wetlands containing crops generally were available during winter and the waterfowl hunting season and typically were drained shortly thereafter. We cannot fully explain use of agricultural wetlands by mottled ducks but corn contains high energy and is among the most metabolizable of waterfowl foods (Kaminski et al. 2003). One hypothesis is that, like congeneric mallards, mottled ducks may consume corn when it is flooded and available. However, we did not study food habits of mottled ducks in this study so we can only speculate at this time.

Hunted and Non-Hunted Wetlands

Our selection ratio results suggest that hunting had little influence on wetland

selection, however our 95% Bonferroni confidence intervals included 1 therefore we

could not infer whether females selected or avoided hunted wetlands. State and private

managed lands and part of ACE Basin NWR (4,781 ha) comprised much of our study area. State-owned and private lands are hunted during the legal waterfowl hunting season but hunt frequency is variable. Bear Island WMA is a 4,864 ha wetland complex that allows approximately 20 hunting parties once per week (D. Harrigal, SCDNR, personal communication). This level of hunting disturbance may not preclude these areas from serving as refugia for mottled ducks and other waterfowl on non-hunted days. We had no information on the frequency and intensity of hunting on private lands. Hunting pressure likely varied considerably per individual landowner prerogatives. The ACE Basin NWR was the only property in our study area prohibiting waterfowl hunting. Radio-marked mottled ducks did use wetlands there but not in proportion to their availability, perhaps because of abundant intermediate wetlands. The secretive nature and early pairing chronology of mottled ducks may acclimate or adapt them to exploit habitats regardless of hunting (Stutzenbaker 1988, Grand 1992). During winter most Anatinae are gregarious and typically form flocks (Bellrose 1980, Baldassarre and Bolen 2006), but mottled ducks most often occur in pairs amid small wetlands or ponds away from other ducks (Paulus and Weller 1988, Stutzenbaker 1988, Zwank et al. 1989). The basic life history and behavior of mottled ducks may make them less risk averse to hunting than other species. Generally, managed brackish wetlands seem vital to mottled ducks in the

ACE Basin, but more information is needed on ecology and habitat use, especially how it relates to breeding and brood rearing.

Home Range Analysis

Home range estimates showed that mottled ducks within the ACE Basin South

Carolina utilized fairly small home ranges as compared to migratory waterfowl 27

throughout the fall-spring time period. Our home range estimates were also on average smaller than home ranges found for mottled ducks in Florida (Varner et al 2013).

Average home range size for mottled ducks within our study was 3,217 ha, whereas average home range size for mottled ducks in Florida averaged 53,800 ha (Varner et al.

2013). 95% home range estimates and 50% core use areas of the 12 females within our study also varied greatly in size in relation to one another (Table 1.2).

Management Implications

An important pattern observed in my study was the near exclusive use of managed tidal impoundments. This seems to contrast habitat use patterns in birds’ historic range of Florida and the Gulf coast region, where natural marsh habitats seem to be used frequently. Future efforts should focus on sustaining current levels of management of managed tidal impoundments, especially brackish and brackish/salt wetlands, and sustaining current stable water conditions during spring-summer period.

Priority should also be placed on reclaiming areas where management may have lapsed

either because of damages incurred to water control structures or insufficient funding to

manage the impoundments.

Female mottled ducks appear to select flooded crop fields when available during

winter. Increased production of crops within existing wetlands may provide beneficial

habitats for mottled ducks during winter, but it also might come at a cost. Areas

designated for agricultural crop production would be unavailable to mottled ducks during

spring-summer. A second possible negative consequence of increasing acres of flooded

corn is possible attraction of wintering mallards, such as feral birds known to have been

released along the Eastern Shore for decades (Weng 2006). To maintain genetic integrity 28

of mottled ducks, corn could attract mallards and promote pair-binding and subsequent swamping of mottled duck genes (Williams et al. 2005). Perhaps future research should focus on providing alternative agricultural crops to corn. Focus should be placed on crops that can be grown and made available throughout the winter and subsequent breeding season. Overall, a more rigorous assessment of mottled duck habitat requirements is needed in the Lowcountry. Balancing annual cycle needs of mottled ducks with wintering waterfowl provides a fruitful and interesting challenge, much like that in California and resident western mallards (Yarris et al. 1994, McLandress et al.

1996, Ackerman et al. 2004).

Table 2.1 Habitat types and proportion of each type within the ACE Basin South Carolina from fall-summer 2010-2012.

Habitat type Total Hectares Proportion of total Managed wetlands 7,781 0.4372 Unmanaged wetlands 10,017 0.5628 Intermediate wetlands 6,047 0.3398 Brackish wetlands 7,658 0.4303 Brackish/salt wetlands 4,093 0.2300 Agricultural vegetationa 992 0.0557 Natural vegetationb 16,806 0.9443 Hunted wetlands 16,093 0.9042 Non-hunted wetlands 1,706 0.0958 aWetlands where agricultural vegetation is present in combination with natural vegetation. bWetlands where only natural vegetation is present.

Table 2.2 Location and home range data for radio-marked mottled ducks within the ACE Basin, South Carolina, fall 2010-summer 2012.

95% fixed kernel Mottled Duck ID Number of locations 50% core use area (ha) home range (ha) 151.279 61 1,506 346 151.399 54 5,330 1,008 151.199 63 4,342 1,045 151.359 61 1,004 275 151.860 42 1,602 386 151.800 57 3,221 819 151.780 44 4,082 982 151.539 50 6,644 833 150.532 43 3,388 386 150.632 40 1,054 204 Mean ± SE 51.5 ± 2.8 3217.3 ± 106.6 628.4 ± 22.3

Table 2.3 Habitat categories, selection ratios, and corresponding 95% Bonferroni confidence intervals for radio-marked mottled ducks (n) in the ACE Basin, South Carolina fall-summer 2010-2012.

Fall-Winter Spring-Summer Probability Probability Wetland Resource use Selection ratio wi Resource use Selection ratio wi n of selection n of selection characteristic (P-value)a (CI)c (P-value)a (CI)c (P-value)b (P-value)b Managed 2.02 (1.95, 2.09) 2.20 (2.16, 2.25) 67 <0.001 <0.001 36 <0.001 <0.001 Unmanaged 0.21 (0.15, 0.26) 0.06 (0.03, 0.10)

Agricultural 1.92 (1.15, 2.69) Vegetation d 55 0.02 <0.001 Natural 0.95 (0.90, 0.99)

31 Vegetation

Hunted areas 1.03 (0.97, 1.08) Non-hunted 55 <0.001 <0.001 0.77 (0.22, 1.31) areas

Intermediate 0.16 (0.02, 0.29) 0.16 (0.03, 0.30) Brackish 67 <0.001 <0.001 2.04 (1.81, 2.27) 36 <0.001 <0.001 2.04 (1.81, 2.27) Brackish/Salt 0.39 (0.07, 0.72) 0.39 (0.07, 0.73) aChi-square results to detect differential use of habitat regimes. bChi-square results to detect differential selection of habitat covariates. cSelection ratios and corresponding 95% Bonferroni confidence intervals for habitats encountered by radio-marked mottled ducks. dBlanks denote that wetland characteristic was not applicable

Figure 2.1 Location in the ACE Basin, South Carolina, where radio-marked female mottled ducks were studied, 2010-2012.

Figure 2.2 Private, state, and federal lands where female mottled ducks were captured, radio marked, and tracked within the Ashepoo, Combahee, and Edisto River Basin, South Carolina, 2010-2012

Figure 2.3 Number of live female mottled ducks with functional transmitters and new mortality signals detected within the Ashepoo, Combahee, and Edisto River Basin, South Carolina during monthly periods from August-April 2010 and 2011.

Figure 2.4 Proportions of each habitat type within the Ashepoo, Combahee, and Edisto River Basin from fall-summer, 2010- 2012.

Figure 2.5 Proportion of radio-marked female mottled duck locations within each habitat category in the Ashepoo, Combahee, and Edisto River Basin from fall-summer 2010-2012.

References

Baldassarre, G. A. 2014. Ducks, Geese, and Swans of North America. Fourth edition. Johns Hopkins University Press. Baltimore, USA.

Baldassarre, G. A., and E. G. Bolen. 2006. Waterfowl Ecology and Management. Second edition. John Wiley & Sons, Inc. New York, USA.

Batt, B. D. J., Afton A. D., Anderson M.G., Ankney C. D., Johnson D. H., Kadlec J. A. & Krapu G. L. 1992 Ecology and Management of Breeding Waterfowl. University of Minnesota Press, Minnesota, USA.

Bellrose, F. C. 1980. Ducks, geese, and swans of North America. Third Edition. Stackpole Books, Harrisburg, Pennsylvania, USA.

Bellrose, F. C. and D. J. Holm. 1994. Ecology and Management of the Wood Duck. Stackpole Books, Harrisburg, Pennsylvania, USA.

Bielefeld, R. R., M. G. Brasher, T. E. Moorman and P. N. Gray. 2010. Mottled Duck (Anas fulvigula), The Birds of North America (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology;

Dardeu, M. R., R. F. Modlin, W. W. Schroeder, and J. P. Stout. 1992. Estuaries. In Biodiversity of the southeastern United States: Aquatic communities, ed. C. T. Hackney, S.M. Adams, and W.M. Martin, 614–744. New York: Wiley.

Davis, B. E. 2012. Habitat use, movements, and ecology of female mottled ducks on the Gulf Coast of Louisiana and Texas. Dissertation, Louisiana State University, Baton Rouge, LA, USA.

Durham, R. S. and A. D. Afton. 2006. Breeding biology of mottled ducks on agricultural lands in southwestern Louisiana. Southeastern Naturalist 5:311-316.

Dwyer, T. J. 1972. An adjustable radio-package for ducks. Bird Banding 43:282-284.

Gilmer, D. S., L. M. Cowardin, R. L. Duval, L. M. Mechlin, C. W. Shaiffer, and V. B. Kuechle. 1981. Procedures for the use of aircraft in wildlife biotelemetry studies. U.S. Fish and Wildlife Service, Resource Publication 140, Washington, D.C., USA.

Gordon, D. H., B. T. Gray, and R.M. Kaminski 1998. Dabbling duck habitat associations during winter in coastal South Carolina. Journal of Wildlife Management 62: 569- 580.

Grand, J. B. 1992. Breeding chronology of Mottled Ducks in a Texas coastal marsh. Journal of Field Ornithology 63:195–202.

Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61:65-71.

Johnson, F. A., F. Montalbano III, J. D. Truitt, and D. R. Eggeman. 1991. Distribution, abundance, and habitat use by mottled ducks in Florida. Journal of Wildlife Management 55:476–482.

Kaminski, R. M., and H. H. Prince. 1981. Dabbling duck and aquatic macroinvertebrate responses to manipulated wetland habitat. Journal of Wildlife Management 45:1- 15.

Kaminski, R. M., J. B. Davis, H. W. Essig, P. D. Gerard, and K. J. Reinecke. 2003. True metabolizable energy for wood ducks from acorns compared to other waterfowl foods. Journal of Wildlife Management 67:542–550

Korschgen, C. E., K. P. Kenow, A. Gendron-Fitzpatrick, W. L. Green, and F. J. Dein. 1996. Implanting intra-abdominal radio-transmitters with external whip antennas in ducks. Journal of Wildlife Management 60:132–137.

Kross, J., R. M. Kaminski, K. J. Reinecke, E. J. Penny, and A. T. Pearse. 2008. Moist-soil seed abundance in managed wetlands in the Mississippi Alluvial Valley. Journal of Wildlife Management 72:707-714.

Legagneux, P., C. Blaize, F. Latraube, J. Gautier, and V. Bretagnolle. 2008. Variation in home-range size and movements of wintering dabbling ducks. Journal of Ornithology 150:183-193.

Manly, B. F. J., L. L. McDonald, D. L. Thomas. 1993. Resource Selection by Animals: Statistical Design and Analysis for Field Studies. Chapman and Hall, London, United Kingdom, p. 192.

Manly, B. F. J., L. L. McDonald, and D. L. Thomas. 2002. Resource selection by animals: statistical design and analysis for field studies. 2nd edition. Kluwer Academic Publishers.

McLandress, M. R., G. S. Yarris, A. E. H. Perkins, D. P. Connelly, and D. G. Raveling. 1996. Nesting biology of mallards in California. Journal of Wildlife Management 60:94-107.

Merendino M. T., D. S. Lobpries, J. E. Neaville, J. D. Ortego, W. P. Johnson. 2005. Regional differences and long-term trends in lead exposure in mottled ducks. Wildlife Society Bulletin 33:1002–1008

Mills, W. E., D. E. Harrigal, S. F. Owen, W. F. Dukes, D. A. Barrineau, and E. P. Wiggers. 2011. Capturing clapper rails using thermal imaging technology. The Journal of Wildlife Management 75(5):1218-1221.

National Oceanic and Atmospheric Administration. 2012. Tide tables 2012, east coast of North America, Charleston, SC. < http://tidesandcurrents.noaa.gov/map/> accessed April 14, 2014.

Prevost, M. B. 1978. Production and utilization of waterfowl foods in brackish impoundments in South Carolina. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 32:60-70.

Ricketts, C. E. 2011. Multi-scale analyses of habitat selection and demographic consequences for South Carolina clapper rails (Rallus longirostris), Georgia. M.Sc. Thesis, University of Georgia, Athens, USA.

Rogers, K. B., and G. C. White. 2007. Analysis of movement and habitat use from telemetry data. In Analysis and interpretation of freshwater fisheries data, eds. C.S. Guy, and M.L. Brown, 625–676. Bethesda, Maryland: American Fisheries Society.

Sargeant, A. B., M. A. Sovada, and R. J. Greenwood. 1998. Interpreting evidence of depredation of duck nests in the Prairie Pothole Region. Northern Prairie Publication 1030. 72pp 1998.

Schummer, M.L., H. M. Hagy, K. S. Fleming, J. C. Cheshier, and J. T. Callicutt. 2012. A guide to moist-soil wetland plants of the Mississippi Alluvial Valley. University Press of Mississippi, Jackson, USA.

Smith, L. M., D. A. Haukos and R. M. Prather. 2004. Avian response to vegetative pattern in playa wetlands during winter. Wildlife Society Bulletin 32:474-480. 39

Smith, L. M., R. L. Pederson, and R. M. Kaminski. 1989. Habitat management for migrating and wintering waterfowl in North America. Texas Tech University Press, Lubbock, USA.

South Carolina Department of Natural Resources, South Carolina Code of Laws. 2013. Marine Resources Act, Section 50-5-80.

Stutzenbaker, C. D. 1988. The Mottled Duck: Its Life History, Ecology, and Management. Texas Parks and Wildlife Department. Austin, TX. 209 pp.

Thomas, D. L., and E. J. Taylor. 1990. Study designs and tests for comparing resource use and availability. Journal of Wildlife Management 54:322-330.

Varner, D. M. 2013. Habitat use, survival, and reproductive success of female Florida mottled ducks (Anas fulvigula fulvigula) using the Everglades Agricultural Area and urban habitats of south-eastern Florida, Auburn. M.Sc. Thesis, Auburn University, Auburn, USA.

Wehland, E. M. 2012. Survival and post-breeding habitat use of mottled ducks in the western gulf coast. Dissertation, Texas A&M University, Kingsville, TX, USA.

Weng, G. J. 2006. Ecology and population genetics of mottled ducks within the south atlantic coastal zone. M.Sc. Dissertation. University of Georgia. Athens, GA. 100.

Wilson, B. C. 2007. North American Waterfowl Management Plan, Gulf Coast Joint Venture: Mottled duck conservation plan. North American waterfowl Management Plan, Albuquerque, New Mexico, USA.

Yarris, G. S., M. R. McLandress, and A. E. H. Perkins. 1994. Molt migration of postbreeding female mallards from Suisun Marsh, California. Condor 96:36-45.

Zwank, P. J., P. M. McKenzie, and E. B. Moser. 1989. Mottled duck habitat use and density indices in agricultural lands. Journal of Wildlife Management 53:110-114.

CHAPTER III

SOCIAL INDICES OF BREEDING MOTTLED DUCKS IN COASTAL SOUTH

CAROLINA

Techniques to estimate population sizes of breeding ducks in principal nesting areas of North America, such as the annual May Pond Breeding Population Survey, are long-established and rigorous (Cowardin and Blohm 1992). In addition to annual breeding ground surveys, another technique known as indicated breeding pairs (IBPs) has been used for decades to study sociability in breeding ducks (Heinroth 1911, Dzubin

1957). Heinroth (1911) was among the first to describe behaviors of breeding ducks, such as aerial chases in relation to territoriality in mallards (Anas platyrhynchos). Later,

Hochbaum (1944), Dzubin (1957, 1969), McKinney (1969), and others established protocol for inventorying behavior and breeding efforts of canvasback (Aythya valisineria) and other prairie ducks using IBPs (Serie and Cowardin 1990, Oring and

Sayler 1992). Researchers continue to use and refine IBPs to better understand linkages between social indices and ecological parameters such as nest success, female success, and duckling fledging rates (Brasher et al. 2002, Arnold et al. 2008, Pagano and Arnold

2009).

The IBP surveys were originally designed for prairie nesting waterfowl but can be used for ducks in different geographic landscapes (Cowardin et al. 1995, Lemelin et al.

2010, Haukos et al. 2010, Zimmerman et al. 2011). Inherent in social indices is that 41

changes in abundance of different social groups (e.g., pairs, lone males, flocked males) reveal important information about nesting behavior and timing of breeding activities

(Dzubin 1969, Hochbaum et al. 1987, Serie and Cowardin 1990, Arnold et al. 2008).

Basic waterfowl life histories, such as birds’ temporal nesting patterns, and environmental and landscape configuration (e.g., grassland, overwater substrates) may influence intraspecific variation in breeding behavior (Sowls 1955, Raveling 1978,

Langford and Driver 1979, Hammond and Johnson 1984, Baldassarre and Bolen 2006).

Social indices have been used to study mottled ducks and other sexually monomorphic species in parts of the species endemic range (Grand 1992, Haukos et al.

2010, Zimmerman 2011). Developing unique survey techniques for breeding coastal mottled ducks is necessary because traditional or annual range-wide breeding surveys for the species do not exist like that for prairie nesting ducks (Haukos et al. 2010). Grand

(1992) and Durham and Afton (2006) hypothesized that nesting chronology of mottled ducks in Louisiana and Texas was directly influenced by wetland availability, which is influenced by precipitation patterns and unpredictable tropical storms and hurricanes. In the ACE Basin, unmanaged natural tidal wetlands and managed tidal impoundments are the primary resources available to waterfowl during the year (Prevost 1978, Gordon et al.

1989, Gordon et al. 1998). However, important breeding habitats for ACE Basin mottled ducks are virtually unknown.

A population of mottled ducks was established in southeastern South Carolina following release of 1,285 birds beginning in the mid 1970’s (South Carolina Department of Natural Resource [SCDNR] unpublished data, Weng 2006). Mottled ducks released to

South Carolina originated from south Louisiana, Texas, and Florida. Nearly 40 years

have elapsed post-release, and mottled duck populations have apparently grown in South

Carolina based on banding data and observations of wetland managers (SCDNR unpublished data). However, only one previous investigator has studied ecology of mottled ducks in South Carolina and they witnessed frequently breeding pairs in several locations in coastal South Carolina and Georgia (Weng 2006). Surveys of brooding, molting, and roosting habitats for mottled ducks were conducted in this region of South

Carolina in the early-mid 1980s (SCDNR, unpublished data). Populations of the birds were thought to be increasing then, and nearly all mottled ducks observed used managed brackish and saline impoundments in the ACE Basin (SCDNR, unpublished data).

To help sustain or expand breeding mottled ducks in South Carolina, baseline information on the ecology of the species is needed. The ACE Basin, formed by the

Ashepoo, Combahee, and Edisto Rivers, is an 182,115 ha region of historical waterfowl importance (Gordon et al. 1989, 1998) and largely comprised of state, federal, and private lands with considerable wetlands and associated habitats (Gordon et al. 1989). Private landowners in the ACE Basin are dedicated waterfowl and wetlands conservationists, and many private lands contain perpetual conservation easements demonstrating collective interest in sustaining wetlands and all waterfowl populations. The ACE Basin lacks active ricefields or extensive upland habitats that might otherwise serve as nesting habitats for mottled ducks or other dabbling ducks (Cowardin et al. 1985, McLandress et al. 1996, Durham and Afton 2003), and preferred nesting substrates are basically unknown in the ACE Basin.

As part of a greater effort to study mottled duck nesting ecology, I initiated IBP surveys to locate potential areas of nesting activity throughout my study area. My goal

was not to derive a population estimate of breeding mottled ducks but merely to use IBPs as an indicator of nesting activity in the ACE Basin. My goal was 1) to develop an index to occupancy of wetlands and associated habitats possibly used by breeding mottled ducks and 2) use survey data to identify site-habitat characteristics that seemed important to nesting mottled ducks and hence focus nest searching efforts in those areas.

Study Area

The ACE Basin is a 182,115 ha estuarine complex formed by the Ashepoo,

Combahee, and Edisto rivers along the southern half of South Carolina’s coastline

between the cities of Beaufort and Charleston (UTM: 530202, 3614843). The ACE

Basin is one of 27 national research reserves monitored by the National Estuarine

Research Reserve System, and contains 128,000 ha of land protected by state, federal,

private, and nonprofit organizations (NOAA 2012). The ACE Basin contains pine and

hardwood uplands, forested wetlands, fresh, brackish and salt water tidal marshes, barrier

islands, and beaches (SCDNR 2013). Included in this study were numerous private,

state, and federal lands including Ernest F. Hollings ACE Basin National Wildlife Refuge

(NWR), Bear Island Wildlife Management Area (WMA), Nemours Plantation Wildlife

Foundation, and Cheeha Combahee Plantation.

Ernest F. Hollings ACE Basin NWR

Ernest F. Hollings ACE Basin National Wildlife Refuge (hereafter Ace Basin

NWR) is a 4,781 ha complex located within portions of Beaufort, Colleton, and

Charleston County, South Carolina. The refuge contains 2 separate units, the Edisto

River unit (2,913 ha), and the Combahee River unit (1,846 ha). These areas combined

contain 1,598 ha of tidal marsh, 1,214 ha of managed tidal impoundments, 485 ha of bottomland hardwood forest, and 1,133 ha of upland forest (Nareff 2009). This refuge mostly contains freshwater to slightly brackish (0-5 ppt.) tidal marsh and managed tidal impoundments (Ricketts 2011). I conducted my surveys of social indices in one wetland within the ACE Basin NWR. This wetland was located on the Combahee River unit, and comprised a wetland with intermediate (1-5 ppt.) salinity.

Bear Island WMA

Bear Island WMA is a 4,864 ha complex located between the Edisto and Ashepoo

Rivers in southern Colleton County, and is owned and managed by the SCDNR.

Management objectives at Bear Island are numerous and include providing quality habitat for waterfowl, other migratory game birds, resident game species, and non-game and endangered species; public hunting and fishing opportunities; and wildlife observation (NOAA 2012). Bear Island WMA contains 2,179 ha of managed wetlands (n

= 30 impoundments), 2,025 ha of natural tidal marsh, 496 ha of woodlands, and 163 ha of agricultural lands (NOAA 2012). Salinities within managed tidal impoundments and natural tidal marsh range from slightly brackish to brackish (5-20ppt). I conducted surveys of social indices within 5 wetlands on Bear Island WMA. Salinities within these wetlands fall under the category of brackish (5-20 ppt.).

Nemours Plantation Wildlife Foundation

The Nemours Plantation Wildlife Foundation (hereafter, Nemours) is a privately

operated non-profit foundation with a focus on education, outreach, research, and land

stewardship (Ricketts 2011). Nemours is 3,986 ha and located along the Combahee

River in northeastern Beaufort County. Nemours contains 1,575 ha of pine forest, 946 ha of pine/hardwood forest, 206 ha of bottomland hardwood forest, 195 ha of fields or open

habitats, 25 ha of upland wetlands, 805 ha of managed tidal impoundments, and 99 ha of

salt marsh. I conducted surveys of social indices within 4 wetlands on Nemours.

Salinities within these wetlands fall under the category of brackish (5-20 ppt.).

Cheeha Combahee Plantation

Cheeha Combahee Plantation (hereafter, Cheeha) is a privately owned 4,856 ha

recreational property located between the Cheeha and Combahee Rivers in southwestern

Colleton County. Cheeha is privately-owned and prioritizes conservation land

stewardship, research, and hunting and fishing. Cheeha contains 500 ha of managed

wetlands (n = 2 impoundments), 3,631 ha upland forest, and 725 ha of bottomland

hardwood forests. I studied mottled ducks in Cheeha’s managed tidal impoundments. I

conducted surveys of social indices within one wetland on Cheeha. Salinities within this

wetland fell under the category of brackish/salt (20-30 ppt.).

Methods

I used IBPs to survey pre-breeding and breeding mottled ducks in the study area

during spring and summer 2011 and 2012 (Dzubin 1969, Sauder et al. 1971, Cowardin et

al. 1995, Pagano and Arnold 2009). I used my personal knowledge of the ACE Basin

and maps to identify a population of habitats, mostly wetlands, that mottled ducks might

select for nesting. For study site selection I used a priori criteria that encompassed

representative characteristics of the region such as variable water management regimes

among wetlands, average salinities, and geographic location within the ACE Basin

(Figure 3.2). Because virtually nothing is known about wetland selection and the breeding ecology of mottled ducks in South Carolina, I was judicious in selecting what I hypothesized to be representative wetland and associated environments available to nesting birds. Furthermore, wetland types in my study area were diverse so I wanted to

ensure I had the representative resources in my sampling scheme. I selected one wetland

in the northern ACE Basin (x=526322, y=3614814) where 1) salinities were least (0-5

ppt.) compared to other regions and 2) the wetland was typically drained in spring and

summer and water levels were shallow (0-15 cm) to promote early succession plants that

subsequently benefit migrating and wintering waterfowl. My sampling was limited to

one wetland in northern ACE Basin because of restricted landowner access and lack of

representative wetlands in the area. I selected four wetlands in the central ACE Basin

(x=530164, y=3614188) because there was greater wetland diversity relative to fresh and saltwater systems. These four wetlands were selected based on permitted access granted by private landowners. Wetlands in the central region are designated in context of

“legal” saltwater-freshwater interface; ‘legal’ interface signifies the divide between salt and freshwater resources, differentiating commercial from sport fishing (SCDNR Section

50-5-80). I had greater access to private lands in this region. Wetlands in the central

ACE Basin ranged from low (e.g., 2 ppt.) to high (e.g., 20 ppt.) salinities and water levels

fluctuated, from 5 cm to 500 cm to enhance growth of desirable emergent plants such as

dwarf spikerush (Eleocharis spp.) and salt-marsh bulrush (Scirpus robustis). I used a

random number generator to randomly select for sampling five wetlands among a pool of

29 wetlands at Bear Island Wildlife Management Area (x=548804 y=3608205) in the southwestern ACE Basin. These wetlands usually contained mid to high salinities (e.g.,

5-22 ppt.) and water levels that fluctuated from 5-500 cm to enhance growth of dwarf

spikerush, salt-marsh bulrush, and widgeon grass (Ruppia maritima). Lastly I located

and sampled one wetland in southeastern ACE Basin, a region with few wetlands and

limited access. This wetland contained high salinities (15-35 ppt.) and flooding was

normally maintained through spring and summer to promote spikerush, widgeon grass,

and other desirable species.

I conducted IBPs on selected wetlands once weekly from sunrise until 1.5 hours

after sunrise (McLandress et al. 1996) from approximately the first week of March

through the last week of June in 2011 and 2012. I conducted 15 survey-days in each

wetland each of the two breeding seasons. I followed established IBP protocol as closely

as possible (Ballard et al 2001, Brasher et al. 2002, Arnold et al. 2008) but manpower limitations precluded me from conducting a double observer approach. Given that there is no previous history of basic nesting ecology of mottled ducks in the ACE Basin, I was mostly interested first in discovering possible nesting site-habitats of these birds and less concerned with devising a rigorous breeding population estimate. Given this rationale, I

did not include a second observer to conduct IBPs alongside me (Sensu Pagano and

Arnold 2009). I did, however, have a second observer that conducted IBPs at the same

time as me but on wetlands independent of mine. The goal was to maximize numbers of

wetlands surveyed for potentially breeding mottled ducks given our lack of basic

knowledge about the species nesting ecology in South Carolina.

Prior to initiating IBP surveys independently, I trained the second observer who

assisted me in both years of the study. During training and prior to conducting

independent surveys, each of us observed mottled duck activity at the same time and site

but did not share information until after completion of a survey (1.5 hours post-sunrise).

After several observation-days and concluding enough parity existed in detecting birds between us, we then conducted surveys independently in separate wetlands (Cowardin et

al 1995). I and the second observer counted every mottled duck that could be seen within

either the entire wetland or a specified portion of the wetland where we could witness all

mottled ducks that came into the area of the wetland we could reliably detect

(McLandress et al. 1996). I classified birds by their status and gender (Haukos et al.

2010, Grand 1992). I used IBP criterion described by Cowardin et al. (1995) that was

slightly adjusted from Hammond (1969) and Dzubin (1969). I deemed that pairs, lone

males, and groups of two males and one female represented mottled duck breeding pairs

(Arnold et al. 2008). I used these designations because this grouping comprised 87% of

the total number of mottled ducks surveyed during both years (Figure 3.3). I only

conducted surveys when weather did not impede clear visibility of birds (Arnold et al.

2008). At the end of each survey, I inserted a salinity meter (YSI model 63; YSI

Incorporated, Yellow Springs, Ohio) to the bottom of the wetland where we had just

observed birds (Haukos et al. 2010), or by consulting with wetland managers who

maintained careful records on water delivery among impoundments and subsequent

changes in salinity. I estimated wetland water depth by permanently affixing a meter

stick in a wetland or consulting records of wetland managers who archived the data.

Statistical Analysis

To estimate abundance of breeding mottled ducks within wetlands in the ACE

Basin, I applied a generalized open-population hierarchical N-mixture model (hereafter

Dail-Madsen Model). I chose Dail-Madsen for two primary reasons: 1) I did not want to 49

assume population closure in mottled ducks because of current uncertainty about how birds use wetlands in the ACE Basin and 2) because I was primarily interested in changes in mottled duck abundance with respect to potentially important covariates I accounted

for in the 11 sampled wetlands. The Dail-Madsen model allows for estimating organism

abundance from repeated counts within or across seasons while loosening the closure

assumption made by previous N-mixture models (e.g. Royle 2004). A population is

deemed closed if there are no births, deaths, immigration, or emigration occurring.

Therefore the Dail-Madsen model allows for estimating abundance from a population

where births, deaths, immigration, and emigration may realistically occur during survey

periods. Another advantage is that Dail-Madsen estimates several demographic

parameters including abundance (λ), apparent survival rate (including mortality and

emigration; ω), recruitment rate (hereafter immigration; γ), and detection probability (p).

This model limits site-level covariates only on λ, and site-level and observation-level

covariates on recruitment rate γ, apparent survival rate ω, and detection probability p

(Fisk and Chandler 2011, Peterson 2014). The Dail-Madsen model also allows for

determining the effects of variables on demographic parameters using several different

types of survey methods. This method allowed us to estimate the relationship of wetland

variables on recruitment rate given that our surveys were conducted on wetlands or

portions of wetlands of different sizes. To test the effect of variables on immigration of

mottled ducks the Dail-Madsen model does not calculate estimates on a per area basis.

Given that we are not attempting to estimate a population size, the Dail-Madsen model

functions adequately to estimate the effects of variables on mottled duck immigration

given that our surveys were conducted on wetlands of varying sizes. It is also realistic to

recognize that detection probabilities of observers can vary in relation to their ability to count and detect birds. Detection of birds typically is not 100%, even for the more experienced observers and more detectable species (Pagano and Arnold 2009). Failing to

account for incomplete detectability can lead to substantial underestimation of abundance

(Pagano and Arnold 2010). Modeling surveys using the Dail-Madsen approach allows

the observer to model detection probability based on varying parameters and account for

imperfect detection, and avoid biases in abundance. Moreover, I was not interested in a

rigorous population estimate but an index to mottled duck breeding activities.

To apply the Dail-Madsen model I used the package Unmarked in program R (v.

3.0.3) using the pcountOpen module (Fiske and Chandler 2011). I tested covariates only

on immigration (γ) and detection probability (p) because I was only interested in what

variables attracted (immigration) pre-breeding and breeding mottled ducks. I examined

possible effects of date, year, area, and observer on detection probability (p), and I a

priori selected ecologically reasonable covariates to test against immigration (γ)

including; date, salinity, area, water depth, and the presence or absence of islands within

the wetland. The pcountOpen module allows you to specify the data distribution best

suited for the data one collects, including negative binomial, poisson, or zero-inflated

poisson (Fiske et al. 2014). To determine the most representative distribution, I ran

global models of the 3 possible distributions and chose the one with the lowest AIC value

and that which generated realistic estimates (no extraneous CIs). Based on modeling

results I selected a negative binomial distribution. The Dail-Madsen model also allows

partitioning of data into primary and secondary survey periods. Primary periods are

deemed those periods in which the population is closed (i.e. no births, deaths,

immigration, or emigration) and can be estimated as different years, months, and survey periods (Dail and Madsen 2011, Fiske et al. 2014). Populations can be assumed to be open between, but not within, separate primary periods. Within each primary period, it is

feasible to have multiple survey periods, or secondary surveys, which can improve model

estimates. I only selected primary survey periods because I assumed populations were

open between survey periods.

Model Selection

Using an information theoretic approach, I developed ecologically reasonable

models potentially important to detection of mottled ducks (p) and wetland use

(immigration) by pre-breeding and breeding mottled ducks (γ) (Table 2.1). I modeled 5

and 4 parameters on immigration (γ) and detection (p), respectively. I used a two-step

process to conceive models that could influence detection and abundance. I first

developed 13 models for p keeping γ fully parameterized (Lebreton et al. 1992). I

evaluated and ranked these models based on Akaike’s Information Criterion (AIC).

When I determined the model with the lowest AIC value for p, I included the model for p

with the most support for all models of γ. I then created 15 models for γ and evaluated

and ranked these models as previously stated. I acknowledge that variable and model

selection may be imprecise because run times for the Dail-Madsen model prevented me

from running all possible model combinations. Given that there were 120 possible

models, and the average run time for each model was around 4 hours, I did not have time

to run all possible combinations. Thus, I further inspected candidate variables of

supported covariates and models by examining model weights and p-values and

associated 95% confidence intervals (CIs) of covariates. If covariates had coefficients 52

with 95% CIs that overlapped zero, I assumed the variable effect was inadequate and eliminated it from further consideration. Lastly, I model-averaged parameter estimates across all models within 5 AIC units of the most supported model (Burnham and

Anderson 2002).

Results

Surveys

I conducted 15 surveys on 11 different wetlands from March-June in both 2011

and 2012 (n = 330 surveys) and counted 4,472 mottled ducks in ACE Basin wetlands.

While holding all other variables at their mean value, the estimated abundance of mottled

ducks was 13.6 ducks per wetland per survey across years. Relative to social groups of

mottled ducks, I most often observed pairs (1,497 pairs), lone males (n = 444), and two

males and one female (n = 215 groups) (Figure 3.3).

Detection Probability Modeling Results

Of 13 candidate models I used to explain imperfect detection, there was one top

model and no competing models <2∆AIC from the top model. I model averaged

variables present in the top model that accounted for imperfect detection to ensure that

95% CIs did not overlap zero and variables were indeed biologically relevant (Table 2.2).

Covariates including year, survey date, and area (ha) of the wetland surveyed were supported in this approach (Table 2.2). Levels of influence of these factors included: date

(β=−1.29, 95% CI=−1.40-−1.02), year of survey (β=−0.52, 95% CI=−0.94-−0.08), and

area of the wetland (β=0.73, 95% CI=0.58-0.81).

Immigration Modeling Results

I used 15 candidate models to explain immigration of mottled ducks into

wetlands. This resulted in one top model and two competing models <2∆AIC from that

most supported model (Table 2.3). I model averaged covariates in the competing models

and found support for area of an entire wetland that was surveyed, wetland salinity, and

presence or absence of islands within the surveyed wetland. Area of an entire wetland

surveyed was positively associated (β=0.31, 95% CI=0.25-0.39) with breeding mottled duck immigration and was the only variable that had a biologically meaningful effect, having 95% confidence intervals that did not overlap zero (Table 2.4, Fig. 2.4). Salinity of the wetland (β=0.03, 95% CI=-0.08-0.13) and presence or absence of islands (β=-0.03,

95% CI=-0.14-0.08) did not have biologically meaningful effects but should not be

ignored. That is, wetland salinity was positively correlated with mottled duck

immigration, whereas presence of islands was negatively correlated with immigration.

Discussion

Application of social indices to study breeding activity of ducks has been used for decades in birds’ principal breeding ranges in North America, but limited research has been conducted on mottled ducks in their endemic range, and my study is the first of its kind in South Carolina (Stutzenbaker 1988, Anderson and Titman 1992, Grand 1992,

Arnold et al. 2008, Haukos et al. 2010). Pairs of mottled ducks in Louisiana will use and

defend ponds formed from water temporarily pooling in freshwater prairie habitats that

range from 10 to 130 ha (Stutzenbaker 1988, Haukos et al. 2010). Flooded rice fields

also provide important feeding and loafing sites for mottled ducks breeding on

agricultural lands in Louisiana (Durham and Afton 2006). Habitats available to mottled 54

ducks in South Carolina differ in structure and availability from habitats elsewhere, such as in coastal Louisiana and Florida. Thus, my study was important to begin assessment of relations between breeding mottled ducks and ACE Basin habitats.

Wetland area

Wetland area often positively influences dabbling duck abundance among species

and during different periods of the annual cycle (e.g. [wintering, Pearse et al. 2012;

migration, Webb et al. 2010; and breeding, Kaminski and Weller 1992). Breeding

mottled ducks also seem to respond to wetland size as Haukos et al. (2010) found positive

correlation between this variable and abundance of mottled duck pairs. Similarly, I found

that immigration of mottled ducks into wetlands of the ACE Basin was positively

correlated with wetland area, but caution is warranted when comparing these dynamics

between mine and Haukos et al. (2010) study areas. The average size of wetlands I

monitored was 120 ha, while wetlands studied by Haukos et al. (2010) averaged 0.12 ha.

The relationship between mottled duck immigration and wetland area that I documented

could be the result of several different factors. First and as I discuss in chapter 3, all nests

that I found during this study occurred within managed tidal impoundments and I never

detected nests in upland habitats. I found that wetlands, particularly those containing

islands, seem critical to breeding mottled ducks in the ACE Basin. Concomitantly,

mottled ducks are territorial during breeding seasons and commonly defend nest sites and loafing ponds from other mottled duck pairs (Stutzenbaker 1988). Therefore, given the importance of wetlands to these birds during breeding season and their territorial behavior, I rationalize that larger wetlands accommodate more pairs, which likely influences the number of available nesting islands in ponds. 55

Larger wetlands may also provide a greater breadth of habitat benefits, such as greater shallow water edge which might sustain a greater number of individual birds

(Kaminski and Prince 1981, Brown and Dinsmore 1986, Smith et al. 2004, Webb 2010).

Second, larger wetlands may provide mottled ducks with more resources in one site, such as food and cover, and hence accommodate more breeding pairs (Nudds 1983, Webb

2010). Wetlands within the ACE Basin are typically a mix of emergent and submerged wetland vegetation and open water habitats (Gordon et al 1989, Weng 2006). Larger wetlands may result in a greater level of interspersion between vegetation and open water

(e.g., hemi-marsh) and therefore more habitat complexity could attract breeding mottled ducks (Kaminski and Prince 1981, Johnson and Grier 1988).

Wetland Salinity

At the outset of this study I hypothesized that salinity would influence

immigration of breeding mottled ducks into wetlands in the ACE Basin. That is, I

predicted that mottled duck immigration would be negatively correlated with wetland

salinity. Haukos et al. (2010) found mottled duck breeding pairs in Louisiana to select

freshwater marsh in greater proportion than intermediate, brackish, and saline marshes.

Other studies have reported that mottled ducks select fresh and intermediate wetlands

over brackish and brackish/salt wetlands during both winter and breeding (Johnson et al.

1991, Davis 2012, Wehland 2012). I hypothesized that mottled ducks would avoid

wetlands with higher salinities (e.g., > 20 ppt.) considering the potentially lethal aspects

of high salinity and duckling mortality (e.g. Moorman et al 1991). However, results from

my models suggested that salinity had no biologically meaningful effect on mottled duck

immigration. I did observe wide-ranging salinities (0-35 ppt) throughout my study, but 56

only 10% of surveys were conducted when wetland salinities ranged from fresh- intermediate. The fact that so few wetlands were fresh to intermediate resulted from drought and abnormally high salinity levels within wetlands during the two years of surveying. Wetlands remained flooded given that managers had access to constant water supplies from rivers, but droughts caused water within the rivers to contain abnormally high salinity levels as well. Wetland salinities in areas I studied tended to remain fairly high in both years. I question if enough variation occurred across salinity gradients to provide a clear understanding of the relation between immigration of breeding mottled ducks and salinity levels.

In Louisiana, breeding pairs of mottled ducks were frequently observed feeding in saline wetlands, especially habitats dominated by widgeongrass (Ruppia maritima)

(Haukos et al. 2010, Stutzenbaker 1988). Management emphasis on most of the wetlands that I studied in the ACE Basin prioritize brackish, brackish/salt, and saline conditions to benefit dwarf spikerush and widgeongrass (Ruppia maritime), aquatic species that wetland managers strive to make available to wetland birds throughout the year. I could not determine if mottled ducks I observed on surveys were breeding birds that were there foraging, non-breeding birds, as mottled ducks tend to have low nesting propensity

(Dugger et al.2010, Varner et al.2014), or more likely some combinations of different social groups of birds.

Wetland Water Depth

I also hypothesized when I initiated my study that wetland depth would drive wetland use of breeding mottled ducks in the ACE Basin. Several studies have reported that mottled ducks use shallowly flooded wetlands (4-20 cm) throughout the entire annual 57

cycle in greater proportion than those flooded more deeply (Haukos et al. 2010, Bielefeld and Cox 2006, Grand 1992, Stutzenbaker 1988). Landers et al (1976) also reported that wintering waterfowl in coastal South Carolina mainly consumed plant species present within shallowly flooded wetlands (<1 m). I hypothesized that immigration of mottled ducks into wetlands in my study would be influenced by wetland depths (Baldassarre and

Bolen 2006, Colwell and Taft 2000, Murkin et al. 1997, Fredrickson and Reid 1988,

Fredrickson and Taylor 1982, Kaminski and Prince 1981), but modeling results showed

no biological influence on mottled duck immigration. During the two years that I

conducted surveys water levels within wetlands remained relatively constant; average

depth of wetlands in surveyed wetlands was 10 cm, and ranged from 0 cm to 27 cm.

Wetlands in which I conducted surveys endure considerable active management annually to promote growth of desired waterfowl foods (Gordon et al. 1989, Chapter 1).

Maintenance of rather constant and shallow levels (i.e. 5-30 cm) is required for these desired plants (Gordon et al. 1998, Gordon et al.1989, Landers 1976). Active management in wetlands and access to relatively predictable wetland conditions could have resulted in a lack of a dynamic relationship between mottled duck immigration and

changes in water depth that occur elsewhere in the birds’ natural range (Varner 2013,

Haukos et al. 2010, Bielefeld and Cox 2006, Grand 1992, Zwank 1989, Stutzenbaker

1988). Mottled ducks frequently respond to environmental conditions and water level

fluctuations by abandoning more deeply flooded wetlands for shallowly-flooded wetland

prairie environments in Louisiana, Texas, and Florida (Varner 2013, Haukos et al. 2010,

Bielefeld and Cox 2006, Grand 1992, Zwank 1989, Stutzenbaker 1988).

Presence or Absence of Islands

Perhaps the most important factors affecting nest success of mottled ducks in my study was the presence of nests on islands and size of islands on which nests occurred

(Chapter 3). Duebbert et al. (1983) surmised that island nesting by waterfowl in certain circumstances is a deliberate choice to achieve greater reproductive success than might be

attainable otherwise. For this reason I reasoned that wetlands containing islands would

attract more breeding mottled ducks than wetlands without islands. Despite the

importance of islands on nest success, my social indices models provided little support

that islands affected use of wetlands by mottled ducks. This pattern is consistent with

results of our nest site selection in chapter 3. Nest site selection results show that female

mottled ducks do not select nest sites on islands over other nest sites within wetlands

such as levees encompassing impoundments. It appears to me that female mottled ducks

do not necessarily select wetlands for breeding solely on the presence or absence of

island habitats, but based on a more complex combination of wetland variables. In this

way mottled ducks may select wetlands for breeding based on this combination of

variables but may benefit from the presence and availability of island habitats within

wetlands that are ultimately chosen.

Management Implications

My study of social indices of breeding mottled ducks suggested that wetland size

is critical to immigration of birds into wetlands of the ACE Basin. Breeding mottled

ducks appear to be attracted to larger wetlands. Wetland managers could consider

providing larger wetlands and wetland complexes amid high quality breeding habitats

(e.g., islands). However, I temper my recommendation with a caveat that we also need to 59

understand mottled brood ecology. If for example, smaller and more numerous versus fewer but larger wetland habitats benefitted brood survival (sensu MacArthur and Wilson

1967), waterfowl conservation should strive to protect or restore some combination of

wetland types. Nonetheless, current availability of wetlands for mottled ducks seems to

be contributing to some levels of breeding activity in the species. In closing, I believe

there are at least three important recommendations resulting from my work: 1) future

research should ensure that other nesting substrates (e.g., upland habitats) truly are not

available in this region to verify that I did not miss potentially important nesting habitats;

2) wetland managers will need to strive to determine how to balance habitat needs of

breeding mottled ducks with those of wintering waterfowl, and 3) determine the carrying

capacity of ACE Basin habitats and the relationship between habitat availability and

desired abundances or densities of breeding mottled ducks. These questions, especially

number 3, may be difficult to answer but seem fruitful for understanding mottled duck

breeding and other annual cycle ecology in the ACE Basin.

Table 3.1 Variables used to model immigration (γ) or detection probability (p) of mottled ducks using wetlands within the ACE Basin, South Carolina, spring-summer 2011-2012.

Covariate Name Description Modeled On Date Date that the survey was conducted γ & p Area Area of the entire wetland surveyed γ & p Depth Average depth of the wetland when surveyed γ Salinity Salinity of the wetland when surveyed γ Island Presence or absence of islands within the wetland γ Observer Observer that conducted the survey p Year Year that the survey was conducted p

Table 3.2 Candidate models used for detection probability (ρ) in surveys of breeding mottled ducks (Anas fulvigula), ACE Basin South Carolina, spring-summer 2011-2012.

a b c d e γ λ ω ρ AIC ΔAIC wi cumltvWt . . . Area + Year + Date 1626.89 0.00 1.00 1.00 . . . Observer + Date + Area 1656.85 29.96 <0.0001 1.00 . . . Area + Date 1659.81 32.91 <0.0001 1.00 . . . Observer + Date + Year 1672.84 45.95 <0.0001 1.00 . . . Observer + Date 1677.72 50.83 <0.0001 1.00 . . . Year + Date 1705.97 79.08 <0.0001 1.00 . . . Date 1706.93 80.04 <0.0001 1.00 . . . Area + Year 1782.07 155.18 <0.0001 1.00 . . . Area 1797.65 170.75 <0.0001 1.00 . . . Observer + Area 1798.91 172.01 <0.0001 1.00 . . . Observer + Year 1806.22 179.32 <0.0001 1.00 62 . . . Observer 1810.18 183.29 <0.0001 1.00

. . . Year 1811.73 184.84 <0.0001 1.00 . . . Null 1812.49 185.60 <0.0001 1.00 a +denotes additive effect. b Akaike’s Information Criterion. c Difference between current model and the model with lowest AIC. d Relative likelihood of the current model (i) based on AIC value. e Cumulative weight of the current model plus all previous models.

Table 3.3 Candidate models for mottled duck (Anas fulvigula) immigration used for model averaging

a b c d e γ λ ω ρ AIC ΔAIC wi cumltvwi Area . . Date+Year+Area 1600.70 0 0.44 0.44 Area + Island . . Date+Year+Area 1602.39 1.69 0.19 0.62 Salinity + Area . . Date+Year+Area 1602.46 1.76 0.18 0.80 Salinity + Depth + Area . . Date+Year+Area 1604.03 3.33 0.08 0.89 Salinity + Area + Island . . Date+Year+Area 1604.31 3.61 0.07 0.96 Date + Salinity + Depth + Area . . Date+Year+Area 1605.39 4.69 0.04 1.00 Salinity + Island . . Date+Year+Area 1638.78 38.08 <0.0001 1.00 Salinity + Depth + Island . . Date+Year+Area 1640.67 39.97 <0.0001 1.00 Salinity . . Date+Year+Area 1649.97 49.27 <0.0001 1.00 Date + Salinity . . Date+Year+Area 1651.60 50.90 <0.0001 1.00 Salinity + Depth . . Date+Year+Area 1651.86 51.16 <0.0001 1.00 Island . . Date+Year+Area 1665.41 64.71 <0.0001 1.00 63 Date + Island . . Date+Year+Area 1667.20 66.50 <0.0001 1.00

Depth . . Date+Year+Area 1673.74 73.04 <0.0001 1.00 Date . . Date+Year+Area 1674.81 74.11 <0.0001 1.00 Null . . Date+Year+Area 1812.49 211.79 <0.0001 1.00 λ is initial abundance, γ is recruitment (immigration), ω is apparent survival (emigration and mortality), and ρ is detection probability. A dot signifies no covariate effect. Data are from 350 surveys on 11 wetlands in the ACE Basin, South Carolina, spring-summer 2011-2012.a +denotes additive effect. b Akaike’s Information Criterion. c Difference between current model and the model with lowest AIC. d Relative likelihood of the current model (i) based on AIC value. e Cumulative weight of the current model plus all previous models.

Table 3.4 Model-averaged log-scaled coefficient estimates (Dail-Madsen models) and 95% confidence intervals of factors potentially influencing immigration and detection probability of mottled ducks (Anas fulvigula) using wetlands in the ACE Basin, South Carolina, spring-summer 2011-2012.

Estimate SE Lower 95% CI Upper 95% CI Abundance: Intercept 1.76 0.09 1.63 1.88

Immigration: Intercept 1.05 0.08 0.97 1.20 Area 0.31 0.03 0.27 0.36 Island -0.03 0.06 -0.11 0.05 Salinity 0.03 0.05 -0.05 0.10 Depth 0.03 0.04 -0.03 0.10 Date 0.07 0.09 -0.06 0.21 Apparent Survival: Intercept 0.47 0.16 0.21 0.68

Detection: Intercept 3.13 0.58 2.67 3.86 Area 0.73 0.29 0.58 0.81 Date -1.29 0.13 -1.40 1.02 Year -0.52 0.35 -0.94 0.08

Figure 3.1 Sites of mottled duck breeding surveys, ACE Basin South Carolina, spring- summer, 2011-2012.

Figure 3.2 Locations of 11 wetlands (inside yellow) surveyed for breeding mottled ducks in the ACE Basin, South Carolina, spring-summer 2011-2012.

Figure 3.3 Designation of social groups of mottled ducks surveyed in wetlands of the ACE Basin, South Carolina, spring- summer 2011-2012.

Figure 3.4 Estimated relationship between mottled ducks (Anas fulvigula) immigration (ducks/wetland) and area of a wetland (ha; solid line), with prediction interval (dashed lines), ACE Basin, South Carolina spring-summer 2011- 2012.

References

Anderson, M. G., and R. D. Titman. 1992. Spacing patterns. Pages 251-289 in Batt B. D. J., A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, editors. Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, USA.

Arnold, T. W., A. M. Pagano, J. H. Devries, R. B. Emery, D. W. Howerter, and B. L. Joynt. 2008. Social indices of breeding productivity in parkland mallards. Journal of Wildlife Management 72: 224-230.

Baldassarre, G. A., and E. G. Bolen. 2006. Waterfowl Ecology and Management. Second edition. John Wiley & Sons, Inc. New York, USA.

Ballard, B. M., M. T. Merendino, R. H. Terry, T. C. Tacha. 2001. Estimating abundance of breeding mottled ducks in Texas. Wildlife Society Bulletin 29: 1186-1192.

Bielefeld, R. R. & Cox, R. R. Jr. 2006. Survival and cause-specific mortality of adult female mottled ducks in east-central Florida. Wildlife Society Bulletin 34: 388– 394.

Brasher, M. G., R. M. Kaminski, and L. W. Burger. 2002. Evaluation of indicated breeding pair criteria to estimate mallard breeding populations. Journal of Wildlife Management 66: 985-992.

Brown, M. and J. J. Dinsmore. 1986. Implications of marsh size and isolation for marsh bird management. Journal of Wildlife Management 50:392-397.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference; A practical information-theoretic approach. Second edition. Springer, New York, USA.

Colwell, M.A. & Taft, O.W. 2000. Waterbird communities in managed wetlands of varying water depth. Waterbirds, 23, 45-55.

Cowardin, L. M. 1969. Use of flooded tim ber by waterfowl at the Montezuma National Wildlife Refuge. Journal of Wildlife Management 33: 829-42.

Cowardin, L. M., D. S. Gilmer, and C. W. Shaiffer. 1985. Mallard recruitment in the agricultural environment of North Dakota. Wildlife Monographs, no. 92.

Cowardin, L. M., and R. J. Blohm. Breeding population inventories and measures of recruitment. Pages 423-445 in B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, editors. Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, Minnesota.

Cowardin, L. M., T. L. Shaffer, and P. M. Arnold. 1995. Evaluation of duck habitat and estimation of duck population sizes with a remote-sensing based system. U.S. National Biological Service, Biological Science Report 2, Washington, D.C., USA.

Dail, D. and L. Madsen. 2011. Models for Estimating Abundance from Repeated Counts of an Open Metapopulation. Biometrics. 67, pp 577-587.

Davis, B. E. 2012. Habitat use, movements, and ecology of female mottled ducks on the gulf coast of Louisiana and Texas. Dissertation, Louisiana State University, Baton Rouge, USA.

Duebbert, H. F., J. T. Lokemoen, and D. E. Sharp. 1983. Concentrated nesting of mallards and gadwalls on Miller Lake Island, North Dakota. Journal of Wildlife Management 47: 729-740.

Dugger, B. D., R. Finger, and S. L. Melvin. 2010. Nesting ecology of mottled ducks Anas fulvigula in interior Florida, USA. Wildfowl 60: 95-105.

Durham, R. S. and A. D. Afton. 2006. Breeding biology of mottled ducks on agricultural lands in southwestern Louisiana. Southeastern Naturalist 5: 311-316.

Dzubin, A. 1957. Pairing display and spring and summer flights of the mallard. Blue Jay 15:10-13.

Dzubin, A. 1969. Assessing breeding populations of ducks by ground counts. In: Saskatoon wetlands seminar, pp. 178-230. Canadian Wildlife Service Report Series No. 6, Ottawa, Ontario, Canada.

Fiske, I. and R.B. Chandler. 2011. Unmarked: An R package for fitting hierarchical models of wildlife occurrence and abundance. Journal of Statistical Software, 43(10), 1-23.

Fredrickson, L. H., and F. A. Reid. 1988. Waterfowl use of wetland complexes. Waterfowl Management Handbook- Fish and Wildlife Leaflet 13.1.1. U. S. Fish and Wildlife Service, Washington, D.C., USA

Fredrickson, L. H., and T. S. Taylor. 1982. Management of seasonally flooded impoundments for wildlife. U.S. Fish and Wildlife Service Resource Publication 148.

Gordon, D. H., B. T. Gray, and R. M. Kaminski. 1998. Dabbling duck habitat associations during winter in coastal South Carolina. Journal of Wildlife Management 62: 569-580.

Grand, J. B. 1992. Breeding chronology of Mottled Ducks in a Texas coastal marsh. Journal of Field Ornithology 63: 195–202.

Hammond, M. C. 1969. Notes on conducting waterfowl breeding population surveys. Pages 238-258 in Saskatoon Wetlands Seminar. Canadian Wildlife Service Report Series No. 6, Ottawa, Ontario, Canada.

Hammond, M. C., & D. H. Johnson. 1984. Effects of weather on breeding ducks in North Dakota. US. Fish Wildlife Service, Fish Wildlife Technical Report 1.

Haukos, D., S. Martinez, and J. Heltzel. 2010. Characteristics of ponds used by breeding mottled ducks on the chenier plain of the Texas gulf coast. Journal of Fish and Wildlife Management 1: 93-101.

Heinroth, O. 1911. Beiträge zur Biologie, nämentlich Ethologie und Physiologie der Anatiden. Proceedings of the International Ornithological Congress 5:589-709.

Heyward, D. C. 1937. Seed from Madagascar. University of North Carolina Press, Chapel Hill, USA.

Hochbaum, H. A. The canvasback on a prairie marsh. American Wildlife Institute, Washington D.C.

Hochbaum, G. S., F. D. Caswell, B. C. Turner, and D. J. Nieman. 1987. Relationships among social components of duck breeding populations, production and habitat conditions in prairie Canada. Transactions of the North American Wildlife and Natural Resources Conference 52:310-319.

Johnson, F. A., F. Montalbano III, J. D. Truitt, and D. R. Eggeman. 1991. Distribution, abundance, and habitat use by mottled ducks in Florida. Journal of Wildlife Management 55: 476–482.

Kaminski, R. M., and M. W. Weller. 1992. Breeding habitats of Nearctic waterfowl. Pages 568-589 in B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D., Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, editors. Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, USA.

Kaminski, R. M., and H. H. Prince. 1981. Dabbling duck and aquatic macroinvertebrate responses to manipulated wetland habitat. Journal of Wildlife Management 45:1- 15.

Kovacik, C. F. 1979. South Carolina rice coast landscape changes. Proceedings Tall Timbers Ecology Management Conference 16: 47-65. 71

LaGrange, T. G., and J. J. Dinsmore. 1989. Plant and Animal Community Responses to Restored Iowa Wetlands. Prairie Naturalist 21(1): 29-48.

Landers, J. L., A. S. Johnson, P. H. Morgan, and W. P. Baldwin. 1976. Duck foods in managed tidal impoundments in South Carolina. Journal of Wildlife Management 40: 721-728.

Langford, W. A. & Driver, E.A. 1979. Quantification of the relationship between Mallard nest initiation and temperature. Wildfowl, 30, 31–34.

Lebreton, J. D., Burnham, K.P., Clobert, J. and Anderson, D.R. 1992. Modeling survival and testing biological hypotheses using marked animals: A unified approach with case studies. Ecol. Monogr. 62, 67–118.

Lemelin, L. V., M. Darveau, L. Imbeau, D. Bordage. 2010. Wetland use and selection by breeding waterbirds in the boreal forest of Quebec, Canada. Wetlands 30:321- 332.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography, monographs in population biology. Princeton University Press, Princeton, New Jersey, USA. 244 pages.

McKinney, F. 1965. Spacing and chasing in breeding ducks. The Wildfowl Trust. Annual Report 76:92-106.

McLandress, M. R., G. S. Yarris, A. E. H. Perkins, D. P. Connelly, and D. G. Raveling. 1996. Nesting Biology of Mallards in California. Journal of Wildlife Management 60: 94-107.

McCarty J. P. 2001. Ecological consequences of recent climate change. Conservation Biology 15: 320–331.

Murkin, H. R., E. J. Murkin, and J. P. Ball. 1997. Avian habitat selection and prairie wetland dynamics: a 10-year experiment. Ecological Applications 7: 1144-1159.

Nareff, G. E. 2009. Ecological value and bird use of managed impoundments and tidal marshes of coastal South Carolina. Thesis, University of Georgia, Athens, USA.

Nudds, T. D. 1983. Niche dynamics and organization of waterfowl guilds in variable environments. Ecology 64:319-330.

Oring, L. W., and R. D. Sayler. 1992. The mating systems of waterfowl. Pages 190-213 in B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, editors. Ecology and management of breeding waterfowl. University of Minnesota Press, Minneapolis, Minnesota, USA.

Pagano, A. M., and T. W. Arnold. 2009. Detection probabilities for ground-based breeding waterfowl surveys. Journal of Wildlife Management 73:392-398.

Pearse, A. T., R. M. Kaminski, K. J. Reinecke, and S. J. Dinsmore. 2012. Local and landscape associations between wintering dabbling ducks and wetland complexes in Mississippi. Wetlands 32:859-869

Raveling, D. G. 1978. The timing of egg laying by northern geese. Auk 95:294-303.

Rehm, E.M. & G. A. Baldassarre. 2007. The influence of interspersion on marsh bird abundance in New York. Wilson Journal of Ornithology 119, 648–654.

Ricketts, C. E. 2011. Multi-scale analyses of habitat selection and demographic consequences for South Carolina clapper rails (Rallus longirostris), Georgia. Thesis, University of Georgia, Athens, USA.

Royle, J. A. 2004. N-mixture models for estimating population size from spatially replicated counts. Biometrics 60, 108–115.

Sargeant, A. B., M. A. Sovada, and R. J. Greenwood. 1998. Interpreting evidence of depredation of duck nests in the prairie pothole region. U.S. Geological Survey, Northern Prairie Wildlife Research Center, Jamestown, ND and Ducks Unlimited, Inc., Memphis, TN. Jamestown, ND: Northern Prairie Wildlife Research Center Online. Accessed 20 November, 2013.

Sauder, D. W., R. L. Linder, R. B. Dahlgren, and W. L. Tucker. 1971. An evaluation of the roadside technique for censusing breeding waterfowl. Journal of Wildlife Management 35:538-543.

Serie, J. R. and L. M. Cowardin. 1990. Use of social indices to predict reproductive success in canvasbacks. Journal of Wildlife Management 54:66-72.

South Carolina Department of Natural Resources, South Carolina Code of Laws. 2013. Marine Resources Act, Section 50-5-80.

Sowls, L. K. 1955. Prairie ducks. Stackpole Co. Harrisburg, Pa., and Wildlife Management Institute, Washington, D.C. 193pp.

Stutzenbaker, C. D. 1988. The Mottled Duck: Its Life History, Ecology, and Management. Texas Parks and Wildlife Department. Austin, TX. 209 pp.

Varner, D. M, G. R. Hepp, R. R. Bielefeld. 2014. Movements and seasonal use of habitats by rural and urban female mottled ducks in southeast Florida. Journal of Wildlife Management

Webb, E. B., Smith L. M., Vrtiska M. P., Lagrange T. G. 2010. Effects of local and landscape variables on wetland bird habitat use during migration through the rainwater basin. Journal Wildlife Management 74:109–119

Weng, G. J. 2006. Ecology and population genetics of mottled ducks within the south atlantic coastal zone. Dissertation. University of Georgia, Athens, GA. 100.

Zimmerman, S. G., J. R. Sauer, W. A. Link, and M. Otto. 2011. Composite analysis of black duck breeding population surveys in eastern North America. Journal of Wildlife Management 76:1165-1176.

Zwank, P. J., P. M. McKenzie, and E. B. Moser. 1989. Mottled duck habitat use and density indices in agricultural lands. Journal of Wildlife Management 53:110-114.

CHAPTER IV

NESTING ECOLOGY OF MOTTLED DUCKS IN COASTAL SOUTH CAROLINA

The breeding season of North American ducks (i.e., nesting and brood rearing) relative to other periods of the annual cycle is regarded as having the greatest influence on population dynamics (Cowardin et al. 1985, Johnson et al. 1982, Hoekman et al.

2002). Several biotic and abiotic factors may affect success of ground nesting ducks at local levels (Dzubin 1969, Cowardin et al. 1985, Greenwood et al. 1995, Sargeant et al.

1998, Hoekman et al. 2002). Nest success may be influenced by habitat such as uplands or islands in which birds select for nesting (Greenwood et al. 1995, Hoekman et al. 2006,

Kaminski et al. 2013), field or wetland size (Stephens et al. 2005), proximity of the nesting area to other lands such as wetlands and croplands and other landscape features including patch size and distribution of nesting environments (Kirsch 1969, Gloutney and

Clark 1997, Clark and Shutler 1999, Hoekman et al. 2006).

Several species of Nearctic ducks are non-migratory and use subtropical coastal and inland marshlands year around, including the mottled duck (Anas fulvigula), which is endemic to the Gulf coast and Florida and has been well studied in those regions

(Stieglitz and Wilson 1968, Baldassarre 2014, Baker 1983, Stutzenbaker 1988, Durham and Afton 2003). A population of mottled ducks was established in southeastern South

Carolina, following release of 1,285 birds beginning in the mid 1970’s (South Carolina

Department of Natural Resource [SCDNR] unpublished data, Weng 2006). Mottled 75

ducks released to South Carolina originated from south Louisiana, Texas, and Florida.

Nearly 40 years have elapsed post-release, and mottled duck populations have apparently grown in South Carolina based on banding data and observations of wetland managers

(SCDNR unpublished data). However, only one previous investigator has studied ecology of mottled ducks in South Carolina (Weng 2006).

Breeding pairs of mottled ducks are frequently observed in several locations in coastal South Carolina and Georgia (Weng 2006). Surveys of brooding, molting, roosting habitats in this region were conducted in South Carolina in the early-mid 1980s

(SCDNR, unpublished data). Mottled duck populations were thought to be increasing then, and nearly all mottled ducks observed used managed brackish and saline impoundments in the ACE Basin (SCDNR, unpublished data).

To help sustain or expand breeding mottled ducks in South Carolina, baseline information is needed. The ACE Basin, formed by the Ashepoo, Combahee, and Edisto

Rivers, is an 182,115 ha region of historical waterfowl importance (Gordon et al. 1989,

1998) and largely comprised of state, federal, and private lands with considerable wetlands and associated habitats (Gordon et al. 1989). Private landowners in the ACE

Basin are dedicated waterfowl and wetlands conservationists, and many private lands contain perpetual conservation easements demonstrating collective interest in sustaining wetlands and all waterfowl populations. The ACE Basin lacks active ricefields or extensive upland habitats that might otherwise serve as nesting habitats for mottled ducks or other dabbling ducks (Cowardin et al. 1985, Durham and Afton 2003, McLandress et al. 1996). Therefore, I initiated a study of nesting mottled ducks in coastal areas of the

ACE Basin in South Carolina. My objectives were to discover the habitats in which

mottled ducks nest, estimate nest success, and begin to acquire baseline information of mottled nesting ecology in coastal South Carolina. Based on preliminary observations of pairs of mottled ducks in the ACE Basin, I hypothesized that island hummocks located amid managed wetlands would provide important nesting substrates.

Study Area

The ACE Basin is a 182,115 ha estuarine system formed by the Ashepoo,

Combahee, and Edisto Rivers along the southern half of South Carolina’s coastline

between the cities of Beaufort and Charleston (UTM: 530202, 3614843). The ACE

Basin is one of 27 national research reserves monitored by the National Estuarine

Research Reserve System and contains 128,000 ha of land protected by state, federal,

private, and nonprofit organizations (NOAA 2012). The ACE Basin contains pine (Pinus

spp.) and hardwood uplands, forested wetlands, fresh, brackish and marine tidal marshes,

barrier islands, and beaches (SCDNR 2013). My study area also included numerous

private, state, and federal lands including Ernest F. Hollings ACE Basin National

Wildlife Refuge (NWR), Bear Island Wildlife Management Area (WMA), Nemours

Plantation Wildlife Foundation, and Cheeha Combahee Plantation (Figure 4.2).

Ernest F. Hollings ACE Basin NWR

Ernest F. Hollings ACE Basin NWR (hereafter Ace Basin NWR) is a 4,781 ha

complex located within portions of Beaufort, Colleton, and Charleston county, South

Carolina. The refuge contains two separate units, the Edisto River unit (2,913 ha), and

the Combahee River unit (1,846 ha). These areas combined contain 1,598 ha of tidal

marsh, 1,214 ha of managed tidal impoundments, 485 ha of bottomland hardwood forest,

and 1,133 ha of upland forest (Nareff 2009). This refuge mostly contains freshwater to slightly brackish (0-5 ppt.) tidal marsh, and managed tidal impoundments (Ricketts

2011). I studied only in the managed tidal impoundments at ACE Basin NWR.

Bear Island WMA

Bear Island WMA is a 4,864 ha complex located between the Edisto and Ashepoo

Rivers in southern Colleton County, and is owned and managed by the SCDNR.

Management objectives at Bear Island are numerous and include providing quality

habitat for waterfowl, other migratory game birds, resident game species, and non-game

and endangered species; public hunting and fishing opportunities; and wildlife

observation (NOAA 2012). Bear Island WMA contains 2,179 ha of managed wetlands (n

= 30 impoundments), 2,025 ha of natural tidal marsh, 496 ha of woodlands, and 163 ha of agricultural lands (NOAA 2012). Salinities within managed tidal impoundments and natural tidal marsh range from slightly brackish to brackish (5-20ppt). I studied in the managed tidal impoundments and natural tidal marsh contained by Bear Island WMA.

Nemours Plantation Wildlife Foundation

The Nemours Plantation Wildlife Foundation (hereafter, Nemours) is a privately operated non-profit foundation with a focus on education, outreach, research, and land stewardship (Ricketts 2011). Nemours is 3,986 ha and located along the Combahee

River in northeastern Beaufort County. Nemours contains 1,575 ha of pine forest (Pinus spp.), 946 ha of pine/hardwood forest, 206 ha of bottomland hardwood forest, 195 ha of fields or open habitats, 25 ha of upland wetlands, 805 ha of managed tidal

impoundments, and 99 ha of salt marsh. I studied mottled ducks in the managed tidal impoundments at Nemours.

Cheeha Combahee Plantation

Cheeha Combahee Plantation (hereafter, Cheeha) is a privately owned 4,856 ha

recreational property located between the Cheeha and Combahee Rivers in southwestern

Colleton County. Cheeha is privately-owned and prioritizes conservation land

stewardship, research, and hunting and fishing. Cheeha contains 500 ha of managed

wetlands (n = 2 impoundments), 3,631 ha upland forest, and 725 ha of bottomland

hardwood forests. I studied mottled ducks in managed tidal impoundments of Cheeha.

Methods

Mottled Duck Capture and Marking

I captured mottled ducks on our study area using night-lighting techniques during

June-September 2010-2011 when birds were in remigial molt (Merendino et al. 2005,

Mills et al. 2011). I transported all captured females to the laboratory at Nemours

Plantation and prepared them for radio-marking. I outfitted hatch-year and after-hatch-

year females either with a harness style transmitter (21 g; Advanced Telemetry Systems

model A2300, Isanti MN) or an intra-abdominal transmitter (18 g; Holohil Systems

Model RI-2D, Carp, Ontario, Canada) in 2011. I only used intra-abdominal transmitters

in 2012, because of apparent failure of transmitters in 2011.

A team of skilled veterinarians implanted all transmitters, and procedures were

approved by Mississippi State University Institutional Animal Care and Use Committee

(IACUC) protocol number 12-005. At the time of capture and radio-marking, the doctors

convened in our laboratory at Nemours. Birds chosen to receive an intra-abdominally implanted radio were first anesthetized with isoflurane and then underwent surgery

(Korschgen et al. 1984, Olsen et al. 1992). After birds recovered from surgery (i.e., 30-

45 min.), they were placed in individual poultry crates and provided water. Birds were monitored for > 2 hours post-surgery to ensure full recovery (Olsen et al, 1992).

Eventually, I transported all birds back to their capture wetland and released transmitted birds together.

Veterinarians followed Dwyer’s (1972) method for attaching harness transmitters and Korschgen et al. (1996) for intra-abdominal transmitters. Both transmitters were <

3% of the bird’s body mass at capture and had an estimated 30-month lifespan. Each transmitter was equipped with a mortality sensor that doubled the pulse rate when the transmitter was motionless > 8 hours and likely indicative of the birds’ death. I banded each mottled duck with a standard USGS aluminum leg band.

Radio Tracking

I began monitoring radio-marked females’ 5-days post release. While I completed coursework at Mississippi State University, a Nemours Plantation wildlife biologist located radio-marked females using an aircraft equipped with strut-mounted 4- element Yagi antennas weekly from August-December 2010 and 2011 (Gilmer et al.

1981). Beginning in January 2011 and 2012, I located radio-marked birds >3 times per week from a truck using an ATS receiver (Model R4000, Advanced Telemetry Systems,

Inc. Isanti, Minnesota) and a handheld 3-element Yagi antenna. Upon detecting a mortality signal, I attempted to retrieve the transmitter and determine cause of each bird’s death from predation or other agents (e.g., Sargeant et al. 1998). 80

I located radio-marked females at least three times a week from early February through early July winter into spring and summer, attempting to detect any pre-nesting or egg-laying behaviors. I located females with a pickup truck, airboat, a jon boat equipped with a surface drive mud motor, by walking levees or wading through marshes, or scanning for birds from atop a large observation tower on one of my study areas. Given

the expansive tidal marsh and diverse inland wetlands of my study areas, use of standard

all-terrain vehicles and the cable-chain drag methods used in northern prairies to nest

search was impossible (Klett et al. 1986, Grant and Shaffer 2012). When I detected a

radio-marked female in the same location for three consecutive days and multiple times

during the day, I assumed the bird was nesting, and I triangulated the female’s location

(Dugger et al. 2010).

Nest Sites

When I was confident of locating a nest, I either flushed the female from the nest

or revisited the location later as determined by triangulation (Klett et al. 1986,

McPherson et al. 2003). Upon finding the nest, I determined clutch size and candled eggs

to estimate initiation and hatch dates (Weller 1956, Bellrose 1980). I used 25 days as the

incubation length plus clutch size to back calculate initiation date, assuming laying rate of

one egg per day (Stutzenbaker 1988). I monitored females suspected of nesting at least

once every three days to confirm continued nesting activity (Dugger et al. 2010). I

revisited nest sites as soon as I presumed a nest had hatched, was depredated before

hatch, or if the female had abandoned the nest to determine its fate. For unsuccessful

nests, I attempted to discern whether nests were abandoned, depredated, or otherwise

destroyed by predators (Klett et al. 1986, 1988, Lariviere and Messier 1998, Dugger et al. 81

2010). For any nest that I deemed as abandoned one day following my initial visit, I categorized it as researcher-induced and excluded each from analysis (McPherson 2003,

Dugger et al. 2010).

Vegetation Measurements

During my final nest visit, I recorded relevant metrics from vegetation bound by a

0.25 m² (0.5m x 0.5m) frame centered on the nest. I identified or measured: 1) vegetation species present within the frame, 2) dominant vegetation height (mean height of vegetation recorded in each of the four cardinal directions inside of the frame (cm), 3) percent cover of each species of vegetation within the frame; I counted number of stems of each species inside the frame and determined proportions accordingly, 4) percent cover over the nest, 5) stem density (number of stems of each species inside the frame), and 6) distance to nearest water (Dugger et al. 2010, Rush et al 2010, Ricketts 2011). I also measured the same set of habitat characteristics inside the 0.25m² sampling frame at a local site that was randomly paired with each nest. I selected local sites by using a random number generator to determine a random distance (≤100 M) and bearing from each nest (Ricketts 2011).

Ancillary Nest Searches

In addition to my sample of radio-marked birds, I also searched and discovered nests of non-instrumented females. I searched for nests in upland and wetland habitats.

I searched for nests in wetland and upland environments by walking through vegetation, using an airboat, or using a jon boat with a surface drive motor to flush females from nests (Stutzenbaker 1988, Huseby 2001). I marked the location of discovered nests with

a handheld GPS unit and defined active nests as those containing at least 1 egg (Klett et

al. 1981). When I detected nests that were being incubated, I candled eggs to estimate

approximate hatch date and backdate the clutches to nest initiation date as previously

described above (Weller 1956, Bellrose 1980). I revisited nests approximately every 4

days (range 3-7 days) to determine if the nest was active or if it had failed, usually when

females were away from nests during recesses. When I determined the fate of a nest

(successful if > 1 egg hatched), I obtained the same vegetation measurements that I

acquired from nests of radio-marked females.

Statistical Analysis

Nest Site Selection

To examine factors influencing local nest site selection by mottled ducks, I used conditional logistic regression in program R (version 3.0.3). This method compared case vs. control response variables in relation to measured covariates (Hosmer and Lemeshow

2000, Ricketts 2011). Each nest site was separated and entered into its own strata. Cases

(nest sites; given a value of 1) were matched with controls (paired random sites; given a value of 0) and each pair represented a single stratum. Emphasis is placed on the overall probability of covariates to adequately describe a case versus a control, and not on

estimating the parameters for each stratum (Ricketts 2011). I used this approach to

analyze selection of each nest site compared to its paired local site. I also used t-tests to

compare vegetation characteristics between years (Kaminski et al. 2013).

I used an information theoretic approach to create 28 a priori candidate models

that might explain nest site selection by mottled ducks (Burnham and Anderson 2002). I

evaluated and ranked candidate models based on a second order Akaike Information 83

Criterion corrected for small sample size (AICc; Akaike 1973, Hurvich and Tsai 1989).

Generally the use of AICc is advocated if the ratio of the sample size (n) to the number of parameters within a model (K) is < 40 (e. g. n/K < 40). The model explaining most

variation in probability of site selection would have the lowest AICc value and the

greatest Akaike weight of evidence, wi. I created a confidence set of models which

included all models with values of ∆AICc ≤ 2 (Burnham and Anderson 2002). A

composite model was created by including all of the parameter estimates present in the

confidence set of models. Furthermore, I model-averaged parameter estimates across

models within 2 AIC units of the most supported model (Burnham and Anderson

2002:170). Model averaging reduces bias and increases precision on parameter

estimates, compared to using a single top model (Burnham and Anderson 2002).

I calculated confidence intervals and scaled odds ratios of the composite model to

address the effect of individual covariates on the dependent variable of site selection

versus the random site (Hosmer and Lemeshow 2000, Ricketts 2011). I calculated scaled

odds ratios because unscaled odds ratios represent the effect of a one unit change on nest

site selection (i.e. a change in vegetation height from 1 cm to 2 cm). However, this

change may not be a true reflection of something biologically meaningful, so I designated

the scaled effect of a 10 cm change in vegetation height versus a 1 cm change on the

probability of a nest being present (Ricketts 2011). I assessed variable importance based

on Akaike weights by summing the Akaike weights (wi) across all candidate models

containing a specific parameter.

Nest Survival

I used the logistic exposure method to estimate daily nest survival (Shaffer 2004).

A primary advantage of this approach is that it permits comparison of daily survival rate estimates to results obtained with other methods (i.e., Mayfield 1975) that assumed constant daily survival rates (Shaffer 2004). The logistic exposure method assumes the probability of survival of each nest is independent of other nests, a characteristic common to other analytical techniques (e.g., Mayfield 2002). However, the logistic exposure approach is advantageous because it does not assume daily nest survival probabilities are constant across nest days, recognizing that they could vary with observation interval length and with respect to differing values of covariates (Shaffer 2004). Moreover, the logistic method allows for analysis of nests found at different times within-seasonally.

The difference between nest survival computation approaches lies in the use of a logit link function, where the logistic exposure method contains a “nuisance” variable recognizing that nests vary in their exposure time (Shaffer 2004). This variable is exponent 1/t (eqn 1), where the denominator, t, represents the number of days in an observation interval and Θ represents the probability of success (survival) during that period. This exponent is necessary to account for the fact that the probability of surviving an interval is dependent on the interval length (Shaffer 2004). The equation is as follows:

1 훩푡 g (Θ) = ln ( 1) (4.1) 1−훩푡

Modeling

Using an information theoretic approach, I developed 17 candidate models potentially important to daily survival of mottled duck nests. I evaluated and ranked candidate models based on a second order Akaike Information Criterion corrected for small sample size (AICc; Akaike 1973, Hurvich and Tsai 1989). I also included a global model that contained all covariates and a null model that assumed constant survival that contained no explanatory variables (Burnham and Anderson 2002, Ricketts 2011). The most probable model had the lowest AICc value and the highest Akaike weight of evidence, wi. I created a composite model by including all parameters present in models where ∆AIC≤ 2 (Burnham and Anderson 2002). I weighted each parameter estimate by the Akaike weights (wi) from each candidate model in which it appeared, and summed these weighted values resulting in the model averaged estimate (Burnham and Anderson

2002). I calculated 95% confidence intervals and scaled odds ratios of the composite model to assist in interpreting modeling results (Ricketts 2011). I also used t-tests to

compare nest initiation dates and clutch sizes between years (Kaminski et al. 2013).

Results

Radio Instrumentation and Tracking

My research team and I outfitted 80 and 36 female mottled ducks ( mass =

752.7 ± 6.3 g [SE], n = 116) with radio transmitters during August-September 2010 and

2011, respectively. We instrumented 40 females with back-pack telemeters and 40

females with intra-abdominally implanted transmitters in 2010; all 36 females received

intra-abdominally implanted transmitters in 2011. Eight days after the adjustment period,

I detected a mortality signal from an implant telemetered bird in 2010, and none was 86

detected within 14 days of deployment in 2011. As fall and winter 2010-2011 progressed, I encountered extreme transmitter malfunction, unexplained disappearance of transmitted birds, or unexplained mortality of females outfitted with back-pack and intra- abdominally implanted transmitters. Near beginning of the breeding season in February

2011, only 4 female mottled ducks with functional implanted transmitters remained on the study area, and none was detected by early March 2011. I never detected any nesting by radio-marked mottled ducks in spring 2011. By the beginning of the 2012 breeding season, I located 12 telemetered females in the study area, and 8 females survived the entire breeding season in 2012 and remained on the study area. I found three nests initiated by radio-marked females in 2012.

Nests, Initiation Dates, and Clutch Size

I found and monitored a total of 45 mottled duck nests; 42 nests were initiated by non-instrumented females (2011, n = 25; 2012, n = 20). I estimated nest initiation dates

for all discovered nests. Earliest nest initiation dates were 25 March and 24 February

2011 and 2012, respectively, while latest initiation dates were 5 June and 8 May 2011

and 2012, respectively. I could not determine if any nests were first or subsequent

nesting attempts. Mean nest initiation date was earlier in 2012 (24 March) than 2011 (27

April; t41 = 1.98, P = 0.0001). Clutch size for nests surviving to incubation was greater in

2012 than 2011 (t26 = 3.09, P = 0.005). Mean clutch size was 9.4 eggs in 2012 (SE = 0.5;

n = 14; range = 6-12) and 7.6 eggs in 2011 (SE = 0.33; n = 12; range = 6-10). I detected

a seasonal decline in clutch size in 2011 (R2 = 0.37, P = 0.02, n = 14) but not in 2012 (R2

= 0.02, P = 0.63, n = 12; Figure 4.7).

Nest Sites

I searched approximately 40 ha of upland and 1,530 ha of wetland areas for nests in both years combined. Uplands were mostly fallow fields containing broomsedge

(Andropogon virginicus), bluestem grasses (Andropogon spp.), and ragweeds (Ambrosia

spp.) contiguous with managed or non-managed wetlands. I never found a mottled duck

nest in uplands, and all 45 nests were found in managed tidal impoundments or on

impoundment levees. I located nests either in impoundments that were not flooded at the

time of searching (n = 11), on levees of these impoundments (n = 6), or on islands of

emergent vegetation in flooded wetlands (n = 28). An average of 63% of all mottled duck nests were found on islands of emergent vegetation within wetlands (n = 14 of 25

nests in 2011; n = 14 of 20 nests in 2012). I never detected more than three species of

plants at nest sites. Plant species at nest sites or in the immediate vicinity included bunch

cordgrass (Spartina bakeri; 74% of 45 nests), salt marsh bulrush (Scirpus robustis; 40%), and giant cordgrass (Spartina cynosuroides; 10%).

I evaluated models formulated a priori to explain variation in nest site selection.

The percentage of vegetation coverage surrounding the nest site was the single variable model that best explained variation in nest site selection (Table 3.3). No other models had ∆AIC values ≤ 2. Additionally, I calculated scaled odds ratios for several variables contained in the top 5 models (Table 3.4). Generally, the percentage of vegetation coverage surrounding the nests positively influenced nest presence, but the confidence interval for the scaled odds ratio of this parameter included 1. Thus, the exact nature of this relationship was inconclusive. This trend existed for all covariates hypothesized to

influence nest site selection (Table 3.4). Therefore, none of the measured and evaluated covariates appeared to have a biologically meaningful effect on nest site selection.

Nest Survival

I excluded three of 45 nests from survival estimates because of presumed researcher induced abandonment. Mayfield estimates for daily nest survival (DSR) was

0.95 (SE = 0.04, n = 25) in 2011, 0.95 (SE = 0.05, n = 17) in 2012, and thus 0.95 overall

(SE = 0.03, n = 42; Table 3.2).

Of 48 candidate models investigated to explain variation in nest success, there

was one top model and two competing models with <2 ∆AIC values (Table 3.6). Area of

emergent vegetation island was the most important variable, being present in the top three

models. Given these competing models, I assessed variable importance by calculating

Akaike weights, model averaged estimates, and odds ratios for each parameter in the top

three models. I also performed a two-way ANOVA only on nests that were present on

islands (n = 28) to test if island area varied between successful and unsuccessful nests

and years. Island area was the most influential explanatory variable; it was 4.6 times

more important than the next closest parameter, distance to water (Table 3.7). Odds

ratios showed that daily nest survival was 2 times more likely with every 1 m² increase in

island area (Table 3.7). The two way ANOVA revealed that nest fate (i.e., successful or

not) was influenced by island area alone (F1, 27 = 5.28, P = 0.031), and there was neither

an effect of year (F1, 27 = 0.49, P = 0.489) nor an interaction of nest fate and year (F1, 25 =

2.19, P = 0.150). Mean area of islands with successful nests was nearly three times

greater (18.73 ± 4.48 m2 [SE, n = 11]) than unsuccessful nests (8.24 ± 1.87 m2 [n = 17]).

The confidence intervals for scaled odds-ratios of all other parameters in competitive 89

models (i.e., distance to water and salinity) included 1, so I cannot explain the possible influence of these parameters on daily nest survival (Table 3.7).

Nest Losses

I suspected that 24 nests failed because of predation based on evidence observed at the nest site. I found nest bowls that contained eggshell fragments well before expected hatch date or mammal tracks in vicinity of nests. All of the tracks observed were raccoon (Procyon lotor). I attributed one nest loss to flooding and four to undetermined causes.

Discussion

Nest Site Selection

My results suggest that suitable nest sites existed for mottled ducks in the ACE

Basin, especially on emergent islands given that 43% of island nests were successful.

Additionally, vegetation cover immediately surrounding the nest appeared to be an

important cue influencing nest site selection. This finding is consistent with patterns in

nest site use by mallard and blue-winged teal (A. discors) elsewhere in North America

(Gloutney and Clark 1997, Clark and Shutler 1999, Hoekman et al. 2002). Microhabitats

containing nests generally had greater vegetation cover than random sites, but modeling

results were inconclusive regarding the direction of effect. Regardless, vegetation

surrounding nests of mottled ducks generally was live, dense, and completely covering

nest bowls vertically and horizontally. Other factors related to nest site selection did not

seem critical such as vegetation height, distance to water, or island area. However, island

area did significantly influence nest success.

My study area was comprised predominately of wetlands, which were often bordered by small upland old field habitats or pine and mixed-forest communities.

However, the area did not contain abundant savannah or other extensive upland herbaceous vegetation characteristic of northern prairies that support most North

American breeding ducks (Gloutney and Clark 1997, Clark and Shutler 1999, Reynolds et al. 2001). Habitat features at local and regional scales of the northern prairies have been shown to influence nest survival in several avian species (Gloutney and Clark 1997,

Stephens et al. 2003, 2005). I only measured variables at local scales; hence, I cannot hypothesize if nest site selection by mottled ducks during my study area was driven more by local, landscape, or both spatial influences. Other potentially important ecological or environmental variables that may have influenced nest site choice by females may be wetland invertebrate abundance and communities, other nesting areas besides those surveyed for nests, local predator communities, competition with other locally nesting birds, or a combination of these and other factors to be evaluated by further research.

Nest Survival

Estimated nest survival (19%) of mottled ducks in my study was within the range

of results reported for other studies elsewhere in the species range (15-30%; Holbrook

1997, Walters 2000, Dugger et al. 2010, Varner et al. 2013; Table 3.8), and it also

exceeded the 15% threshold of nest success deemed crucial to maintaining midcontinent

mallard populations (Cowardin et al. 1985, Hoekman et al. 2002). Generally, nest

success of mottled ducks throughout its endemic range and in South Carolina is

comparable to that for prairie nesting ducks (10-20%; Table 3.8).

Perhaps the greatest revelation from my study was the apparent importance of islands and island area to mottled duck nest success. Other environmental characteristics, such as distance to water, nest initiation date, or year were less influential. Several models of daily nest survival were competitive, but most of them contained parameters that did not provide much support for explaining variation in nest success. Nonetheless, island size was the most important covariate influencing nest success of mottled ducks, and 64% of all mottled duck nests occurred on islands. The remainder of nests (n = 15)

were in wetlands not flooded or in vegetation along levees within wetlands or adjacent to

managed tidal impoundments. Island habitats where mottled duck nests occurred were

small mounds or hummocks of vegetation that averaged 12 m2 (range = 2.2-15.9 m2).

What remains unclear is whether there are enough of these islands to sustain or increase breeding mottled duck populations, or whether other nesting substrates (wetlands that were dry, levees) are also important to these birds and can benefit populations.

Understanding relations between density of breeding mottled ducks and available nesting habitats will be valuable information to acquire in the future in the ACE Basin.

On islands in both flooded and dry wetlands and in other nesting substrates, vegetative characteristics at mottled duck nest sites were comparable to those for nesting females of the species in coastal Florida (Varner et al. 2013), Texas (Stutzenbaker 1988), and Louisiana (Durham and Afton 2003). Mottled ducks in the ACE Basin typically nested in sites with dense herbaceous vegetation that provided birds’ virtually complete vertical and horizontal cover. Although species of vegetation selected by nesting females in my study differed from other breeding areas (Stutzenbaker 1988, Varner et al. 2013,

Dugger et al. 2010), dense herbaceous structure around the nest may be an important

proximate and settling cue for nesting mottled ducks (Arnold 1993, Grant and Shaffer

2012).

Island nesting by North American ducks is not abnormal and may hinge on breeding location, alternative nesting substrates, predator populations, and other factors

(Hammond and Mann 1956, Kaminski and Prince 1976, Vermeer 1970, Newton and

Campbell 1975, Duebbert et al.1983, Giroux 1981, Kaminski et al. 2013). Island nesting waterfowl often have greater nest success than birds nesting on mainlands (Newton and

Campbell 1975, Duebbert et al. 1983). Mallards, gadwall (Anas strepera), and Canada geese (Branta canadensis) that breed at northern latitudes are species more prone than others to nest densely on islands (Hammond and Mann 1956, Stieglitz and Wilson 1968,

Kaminski and Prince 1976, Giroux 1981, Vermeer 1970). Island nesting in some circumstances is likely a deliberate choice to achieve greater reproductive success than might be attained otherwise (Duebbert et al. 1983). Two studies have researched the effect of small islands similar in size to mine on nest site selection, nest success, and nest density of northern breeding ducks and geese (Kaminski and Prince 1976, Giroux 1981).

Kaminski and Prince (1976) reported Canada geese in southwestern Michigan most frequently used muskrat lodges with widths averaging 1.6 m, while Giroux (1981) reported nest success rates (57%) for several species of ducks on small islands of vegetation ranging from 0.6 ha to 19.1 ha. The islands used by mottled ducks in the ACE

Basin and muskrat lodges used by Canada geese in Michigan were similar in structure, area, and use, but were smaller on average than earthen islands described by Giroux

(1981). Islands supporting successful mottled duck nests that were in my study may have

offered nesting females with a hedge against predation and nest loss, because expansive grasslands or wet prairie are not present in the ACE Basin.

Only a few studies have investigated mottled duck nesting ecology on islands

(Stieglitz and Wilson 1968, Stutzenbaker 1988, Holbrook et al.2000, Walters et al. 2001).

In coastal Florida, 98% of local mottled ducks nested on islands and Mayfield nest success was 76% (Stieglitz and Wilson 1968). Although estimated nest survival of mottled ducks in my study was 24% overall, it was greater for island nests (30%) than non-island micro-habitats (13%). Stutzenbaker (1988) found 315 nests amid dense cordgrass near permanent ponds within marshes. Holbrook et al. (2000) found nest success (Mayfield) to range from 6-67% for >300 mottled duck nests in the Atchafalaya

River Delta, Louisiana, and Walters et al. (2001) reported that nest success (Mayfield) ranged from 21-25% for 120 nests on spoil island habitats in the Mississippi River Delta.

Similar to other dabbling ducks nesting elsewhere, island nesting mottled ducks experience greater success in the ACE Basin. All islands searched for nests existed within managed tidal impoundments with controlled water levels. Water levels varied among impoundments but averaged 20 cm ± 6.2 (SE) cm deep (range = 5–90 cm, n = 5).

Wetland managers generally maintain 5-25 cm water within impoundments throughout

the year to promote growth of emergent and submersed vegetation used by fall migrating

and wintering waterfowl (Gordon et al. 1989). Hence, islands that contained mottled

ducks were essentially surrounded by ≈20 cm water during the breeding season. I do not

know if some minimum water depth is needed to deter raccoons or other potential

predators of mottled duck nests. If mottled ducks cue in and settle to nest relative to

some combination of island size and water depths, future research should address this understanding to improve emergent wetland management for nesting mottled ducks.

Despite the apparent importance of islands to nesting mottled ducks, there is one primary caveat in comparing my results to studies by Stieglitz and Wilson (1968),

Holbrook et al. (2000), and Walters et al (2001). Islands studied by these researchers were large barrier islands that ranged from 200-25,000 m2, whereas islands in my study were small mounds or hummocks of vegetation that averaged 12 m2 (range = 2.2-15.9

m2). Thus, islands in my study area, differed from those investigated by researchers in

Florida and Louisiana (Stieglitz and Wilson 1968, Holbrook et al. 2000, and Walters et al. 2001). In my study, islands or hummocks that contained nests existed within managed tidal impoundments with controlled water levels, and they were closer to the mainland

(464 m) than those described by Stieglitz and Wilson (1968) and Holbrook et al (2001).

These among-area contrasts offer some explanation in the departures in nest success

between other studies (Stieglitz and Wilson 1968, Holbrook et al. 2000) and mine,

especially considering that vegetation characteristics at mottled duck nest sites were

similar in all three studies. Mammalian and avian predators in my study may have had

greater probability to locate and access mottled duck nests than those contained on larger

and more remote barrier islands.

Studies of survival of duck nests are equivocal in relation to vegetation

characteristics, such as vegetation height and density at nest sites (Nudds 1991, Esler and

Grand 1993, Stephens et al. 2005, Walker et al. 2008, Varner et al. 2013). My models

did not provide much support for a relationship between vegetation characteristics and

nest survival; my results are similar to patterns in two other contemporary studies of

nesting mottled ducks in Florida (Dugger et al. 2010, Varner et al. 2013). Varner et al.

(2013) found that nest age affected success of mottled duck nests, but several other habitat-related variables seemed inconsequential. In contrast, nest age had seemingly little effect on nest success in my study. Likewise, I did not find much support for an influence of vegetation characteristics between successful and unsuccessful nests upon pooling data across years.

Contrary to my a priori predictions, distance to water and wetland salinities had little influence on nesting success of mottled ducks. However, I realize that salinities are more likely to influence survival of mottled duck ducklings (Moorman et al. 1991).

Perhaps wetland salinity is somewhat irrelevant to nest success, so long as other components (e.g., islands) are available for nesting. However, there may be bet-hedging tradeoffs in decision making by females that influence their choice of islands for nesting.

For example, females might select nest sites based on the presence of islands but also their proximity to wetlands with salinities tolerable to ducklings (Moorman et al. 1991).

Female mottled ducks select sites based on its potential for (nest) success, but tradeoffs between selecting secure nest sites and risks to future progeny (i.e., duckling death in highly saline wetlands) are not currently well understood and are an interesting topic for future study, especially as we accrue knowledge about brood ecology in this region.

Although the variable distance to water had little effect on daily survival rates of nests, mean distance from nests to water in our study was much closer (4.1 m) than for nests in other studies such as Florida (188 m, range = 147-229; Dugger et al. 2010),

Louisiana (185 m, range = 14-713 m; Durham and Afton 2003), and Texas (119 m, range

= 15-219 m; Stutzenbaker 1988). I attributed these differences to the fact that all nests in

my study were found within managed tidal impoundments or on levees immediately surrounding managed wetlands. In other studies, researchers discovered nests in uplands or rice fields (Durham and Afton 2006, Varner 2013). Although I spent time searching for nests in limited uplands adjacent to wetlands, greater attention could be given to nest searching these areas to ensure we did not overlook potentially important nesting habitats for mottled ducks. One area that was not searched for nesting females were natural tidal marsh and unmanaged tidal wetlands. These areas were not searched because mottled ducks were infrequently observed there and because these areas are completely inundated twice daily through tidal influence (Gordon et al. 1989). Nonetheless, current results seem to indicate that suitable nesting substrates exist in the ACE Basin for local mottled ducks. Knowledge of duckling and brood survival remains an important knowledge gap that will ultimately help us understand recruitment of this species.

Management Implications

My study was a first effort to understand relations between breeding mottled

ducks and their habitats in the ACE Basin. The importance of island area in explaining

variation in mottled duck nest success in managed tidal impoundments was especially

revealing. I lacked manpower to search natural tidal marshes and unmanaged tidal

wetlands, so I cannot conclude if those habitats were important to nesting mottled ducks.

However, based on my observations of breeding female mottled ducks in the ACE Basin,

managed tidal impoundments are critical habitats for mottled ducks.

Wetlands of the South Atlantic Costal Zone historically have been managed intensively for wintering waterfowl, given the importance of the region to myriad species of wetland dependent birds in the Atlantic Flyway (Gordon et al. 1989). I advocate 97

continuance of wetland management to fulfill needs of wintering waterfowl in the ACE

Basin. However, as we increasingly understand ecology of nesting and brood rearing mottled ducks, wetland managers may have increased opportunities to meet annual cycle needs of mottled ducks and other waterfowl in the ACE Basin. For example, creation of island habitats are a direct outcome of hydrological management of tidal impoundments.

When shallow water (5-25 cm) is maintained within a brackish or saline wetland for an entire year, with 2-3 day drawdowns if at all, islands of Spartina bakeri emerge, which

were suitable nesting sites for mottled ducks. However, knowledge of whether adequate

numbers of nesting substrates to support current or growing populations of mottled ducks

in the ACE Basin is lacking. Regardless, I recommend that where possible and practical

managers adopt management practices conducive to creating and maintaining island

habitats for nesting mottled ducks. Additionally, managers may desire to erect and

determine value of artificial nest tunnels to provide nest sites in managed tidal

impoundments secure from mammalian and avian predators (Chouinard et al. 2005).

Despite emerging knowledge of the importance of islands, predator communities

in ACE Basin wetlands are largely unknown. I documented loss of mottled duck nests to

predation, and an ongoing study investigating brood ecology may record even greater

predation of mottled duck nests (M. Kneece, unpublished M.S. research, Mississippi

State University). I recommend that future research examine predator communities present in and around wetlands used by nesting mottled ducks. Infrared cameras may be

used to identify potential predators of mottled duck nests. I also recommend determining

characteristics of islands used and not used by mottled ducks and those with successful

nests (e.g., Kaminski and Prince 1977).

Table 4.1 Mean (± SE; n) for parameters of mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

Year Parameter 2011 2012 Combined Nest initiation datea 117.00 ± 3.50 (25) 83.00 ± 4.53 (17) 103.00 ± 2.81 (42) Clutch size 7.60 ± 0.33 (14) 9.40 ± 0.50 (12) 8.40 ± 0.34 (26) Daily nest survival rate 0.95 ± 0.04 0.95 ± 0.05 0.95 ± 0.03 Nest success rateb 0.22 ± 0.16 0.26 ± 0.19 0.24 ± 0.13 aJulian date; 1 = 1 January bInterval success rate (55 – 156 days; 101 days)

Table 4.2 Predictor variables used to model daily survival rate of mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

Minimu Maxim Predictor Variable x a SEb m um Distance from nest to nearest water 4.1 1.9 0.2 24.6 Mean grass height at nest site (cm) 106.9 5.1 54.0 214.0 Vegetation at nest site (%) 81.9 3.1 4.0 99.0 Area of island at nest site 11.3 2.1 0.6 15.9 Spartina bakeri stem density (# 142.6 10.8 0 232.0 stems/0.25 m2) Spartina cynosuroides stem density (# 38.0 15.7 0 83.0 stems/0.25 m²) Scirpus robustis stem density 31.1 6.0 0 121.0 (#Stems/0.25 m²) a denotes mean b SE denotes standard error

Table 4.3 Nest site selection candidate models for mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

a b c d e Model k AICC ∆AICC wi Percent vegetation 2 50.02 0.00 0.63 Percent bare ground 2 52.17 2.15 0.21 Percent vegetation + vegetation height 3 53.01 2.99 0.14 Island area + percent vegetation 3 53.18 3.16 0.13 Distance to water + percent bare ground 3 53.48 3.46 0.11 percent bare ground + island area 3 54.13 4.11 0.08 Island area + vegetation height + percent vegetation 4 54.96 4.94 0.05 Number of stems of spartina bakeri 2 55.16 5.14 0.04 percent vegetation + percent bare ground + island area 4 55.15 5.13 0.04 Distance to water × island area 3 55.97 5.95 0.03 Vegetation height 2 56.41 6.39 0.02 Presence of nest on island 2 56.91 6.89 0.01 Number of stems of saltmarsh bulrush 2 56.98 6.96 0.01 Distance to water 2 57.07 7.05 0.01 Island area 2 57.15 7.13 0.01 Number of stems of cord grass 2 57.35 7.33 0.01 Island area + number of stems of spartina bakeri 3 57.63 7.61 0.01 Vegetation height + island area 3 57.95 7.93 0.01 Null 1 58.23 8.21 0.01 Distance to water + island area 3 58.75 8.73 0.01 Vegetation height + island area + distance to water 4 59.60 9.58 0.01 a +denotes additive effect. b Number of estimated parameters. c Akaike’s Information Criterion adjusted for small sample sizes. d Difference between current model and the model with lowest AICc. e Relative likelihood of the current model (i) based on AICc value. 100

Table 4.4 Model-averaged parameter estimates, odds ratios, and importance weights of parameters used to describe local scale nest site selection of mottled ducks in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

Odds-ratio unit Estimated Odds-ratio 95% Parameter Estimate Odds-ratio 95% UCL change Odds-ratio LCL Importance Weight Intercept 3.01 Percent vegetation 1.01 5 0.99 0.91 1.09 0.22 Percent bare ground -0.28 1 0.97 0.94 1.02 0.20

101 Vegetation height 1.00 10 1.00 0.99 1.01 0.14

Distance to water 0.54 5 1.01 0.96 1.08 0.10 Island area 0.47 1 1.03 0.94 1.12 0.07

Table 4.5 Descriptive statistics (± SE) used to describe mottled duck nests and matched local pair sites in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

Parameter Nest Sites Local Sites

Distance to water 5.85 ± 1.98 3.01 ± 0.94 (m) Vegetation height 407.15 ± 22.72 379.15 ± 26.31 (cm) Percent vegetation 78 ± 3.32 72 ± 3.65 (%) Island area (m2) 2.98 ± 0.75 2.38 ± 0.90

102

Table 4.6 Candidate models used to explain daily survival rates of mottled duck nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

a b c d e Model k AICC ∆AICC wi Island area 2 151.22 0.00 0.22 Distance to water + island area 3 152.64 1.42 0.11 Wetland salinity + island area 3 153.05 1.84 0.09 Initiation date 2 153.26 2.05 0.08 Island area + percent Spartina bakeri 3 153.34 2.13 0.07 Percent vegetation + percent bare ground 4 153.50 2.29 0.07 NULL 1 154.04 2.83 0.05 Julian date + age + island area 4 154.10 2.89 0.05 Initiation date2 2 154.10 2.89 0.05 Age + initiation date 3 154.99 3.78 0.03 Age 2 155.24 4.03 0.03 Age2 2 155.42 4.21 0.03 Age + initiation date2 3 155.56 4.34 0.02 Percent vegetation 2 155.96 4.74 0.02 Vegetation height 2 156.13 4.92 0.02 Age + percent vegetation 3 156.90 5.69 0.01 Age + year 3 157.23 6.02 0.01 a +denotes additive effect. b Number of estimated parameters. c Akaike’s Information Criterion adjusted for small sample sizes. d Difference between current model and the model with lowest AICc. e Relative likelihood of the current model (i) based on AICc value.

103

Table 4.7 Model averaged estimates and odds ratio estimates of parameters used to describe daily nest survival of mottled ducks nests in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

Odds-ratio Odds- Odds- Importa Estimated Parameter Estimate unit ratio 95% ratio 95% nce Odds-ratioc change LCL UCL Weight (Intercept 2.83 ) Island 0.56 1 1.94* 1.09 4.05 0.69 area Distance 0.17 10 1.30 0.91 2.26 0.15 to water Julian -0.30 1 0.61 0.37 1.01 0.13 date Salinitya -0.12 1 1.24 0.75 2.03 0.12 Ageb 0.13 1 1.12 0.75 1.69 0.13 Year 0.08 1 1.16 0.07 2.12 0.03 a Salinity of nearest water to nest b Age of nest cAsterisk denotes a confidence interval not including 1

104

Table 4.8 Comparisons of estimated nest success for mottled ducks across the birds’ geographic range (adapted from Durham and Afton 2003).

Nest Success (%) Apparen Mayfiel Location n Reference t d Stieglitz and Wilson Merritt Island, FL 90 76.7 57.0 (1968) 26 Atchafalaya River delta, LA 47.5 30.6 Holbrook (2000) 5 27 Mississippi River delta, LA 40.0 20.0 Walters (2000) 9 South and east-central FL 77 44.2 58.3 Varner et al. (2013) Coastal South Carolina 42 33 23.9 This Study TX 51 27.5 11.0 Engeling (1950) Interior FL 25 16.0 9.5 Dugger et al. (2010) 14 TX and LA 24.7 9.0 Stutzenbaker (1988) 6 Cameron and Calcasieu Durham and Afton 66 21.0 6.0 Parishes, LA (2003) Cameron Parish, LA 30 16.6 5.0 Baker (1983)

105

Figure 4.1 Study area (Ashepoo, Combahee, and Edisto River Basin) within South Carolina, where mottled ducks were captured, radio-marked, and tracked 2010-2012.

106

Figure 4.2 Private, state, and federal lands where female mottled ducks were captured, radio-marked, and tracked in the Ashepoo, Combahee, and Edisto River Basin, South Carolina, 2010-2012.

107

108

Figure 4.3 Influence of island area (m2; ± SE) on successful and failed mottled duck nests and a matched random site (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

109

Figure 4.4 Influence of distance to water (m; ± SE) on successful and failed mottled duck nests and matched random site measurement (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

110

Figure 4.5 Influence of vegetation height (cm; ± SE) on successful and failed mottled duck nests and a matched random site measurement (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

111

Figure 4.6 Influence of percent bare ground (± SE) on successful and failed mottled duck nests and a matched random site measurement (local pair site) in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

Figure 4.7 Seasonal trend in clutch size for nesting mottled ducks in the Ashepoo, Combahee, and Edisto River Basin, South Carolina during spring 2011 and 2012.

(n = 14, 2011, R2 = 0.37; n = 12, 2012, R2 = 0.02)

112

References

Ackerman, J. T., A. L. Blackmer., J. M. Eadie. 2004. Is predation on waterfowl nests density dependent? Tests at three spatial scales. Oikos 107: 128–140.

Aebischer, N. J., P. A. Robertson, and R. E. Kenward. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology 74: 1313-1325.

Akaike, H. 1985. Predictions and entropy. Pages 1-24 in A. C. Atkinson and S. E. Fienberg, editors. A celebration of statistics: the ISI centenary volume. Springer- Verlag, New York, USA.

Allen, J. A. 1981. Nesting and productivity of Mottled Ducks in the marshlands of southwestern Louisiana. M.Sc. Thesis. Louisiana State University. Baton Rouge, LA. 59 pp.

Amundson, C. L., and T. W. Arnold. 2011. The role of predator removal, density- dependence, and environmental factors on mallard duckling survival in North Dakota. Journal of Wildlife Management 75: 1330-1339.

Anderson, D. R., and K. P. Burnham. 2002. Avoiding pitfalls when using information- theoretic methods. Journal of Wildlife Management 66: 912-918.

Arnold, T. W. 2010. Uninformative parameters and model selection using Akaike’s Information Criteria. Journal of Wildlife Management 74: 1175-1178.

Baker, O. E. 1983. Nesting and brood-rearing habits of the Mottled Duck in the coastal marsh of Cameron Parish, Louisiana. Thesis, Louisiana State University, Baton Rouge, USA.

Baker, O. E., P. McKenzie, P. J. Zwank. 1984. Chase flights as an index to nest density in mottled ducks. Proceedings of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 38: 278-284.

Ballard, B. M., M. T. Merendino, R. H. Terry, T. C. Tacha. 2001. Estimating abundance of breeding mottled ducks in Texas. Wildlife Society Bulletin 29: 1186-1192.

Barras, S. C. 1999. Evaluation of sampling methods, quantitative indices, and predictive models for estimating mallard breeding populations. Dissertation, Utah State University, Logan, USA.

Bellrose, F. C. 1980. Ducks, geese, and swans of North America. Third Edition. Stackpole Books, Harrisburg, Pennsylvania, USA. 113

Bielefeld, R. R., M. G. Brasher, T. E. Moorman, and P. N. Gray. 2010. Mottled Duck (Anas fulvigula), The Birds of North America (A. Poole, Ed.). Ithaca: Cornell Lab of Ornithology.

Bielefeld, R. R. & Cox, R. R. Jr. 2006. Survival and cause-specific mortality of adult female mottled ducks in east-central Florida. Wildlife Society Bulletin 34: 388– 394.

Bowman, T. D. 1987. Ecology of male black ducks molting in Labrador. Thesis, University of Maine, Orono, USA.

Brasher, M. G., R. M. Kaminski, and L. W. Burger. 2002. Evaluation of indicated breeding pair criteria to estimate mallard breeding populations. Journal of Wildlife Management 66: 985-992.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and inference: an information-theoretic approach. Springer-Verlag, New York.

Carney, S. M. 1992. Species, age, and sex identification of ducks using wing plumage. U.S. Fish and Wildlife Service, Washington, D.C., USA

Chouinard, M. P., Jr. and T. W. Arnold. 2007. Survival and habitat use of mallard (Anas platyrhynchos) broods in San Joaquin Valley, California. Auk 124: 1305-1316.

Clark, R.G., and Shutler, D. 1999. Avian habitat selection: adaptive nest-site choice by ducks. Ecology, 80.

Coluccy, J. M., T. Yerkes, R. Simpson, J. W. Simpson, L. Armstrong, and J. Davis. 2008. Population dynamics of breeding mallards in the great lakes states. Journal of Wildlife Management 72:1181-1187. Cowardin, L. M., D. S. Gilmer, and C. W. Shaiffer. 1985.

Cowardin, L. M., D. S. Gilmer, and C. W. Shaiffer. 1985. Mallard recruitment in the agricultural environment of North Dakota. Wildlife Monographs, no. 92.

Cox, R. R., Jr., and A. D. Afton. 1997. Use of habitats by female northern pintails wintering in southwestern Louisiana. Journal of Wildlife Management 61: 435- 443.

Cox, R. R., Jr., and A. D. Afton. 1998b. Effects of capture and handling on survival of female northern pintails. Journal of Field Ornithology 69: 276-287.

114

Cuthbert, Robert B. and Stephen G. Hoffius, Eds. 2009. Northern Money, Southern Land: The Lowcountry Plantation Sketches of Chlotilde R. Martin. Columbia: University of South Carolina Press.

Dahl, T. 1999. South Carolina’s wetlands: status and trends, 1982-1989. Accessed 13 January, 2014.

Dahl, T. E. 2006. Status and trends of wetlands in the conterminous United States 1998 and 2004. U.S. Department of the Interior. U.S. Fish and Wildlife Service, Washington, D.C.

Daubenmire, R. 1959. A canopy-coverage method of vegetational analysis. Northwest Science 33: 43-64.

Davis, B. E. 2007. Habitat use, movements, and survival of radiomarked female mallards in the lower Mississippi Alluvial Valley. Thesis, Louisiana State University, Baton Rouge, USA.

Davis, B. E. 2012. Habitat use, movements, and ecology of female mottled ducks on the gulf coast of Louisiana and Texas. Dissertation, Louisiana State University, Baton Rouge, USA.

Davis, B. E., A. D. Afton, and R. R. Jr. Cox. 2009. Habitat use by female mallards in the lower Mississippi Alluvial Valley. Journal of Wildlife Management 73: 701-709.

Davis, B. E., and A. D. Afton. 2010. Movement distances and habitat switching by female mallards wintering in the lower Mississippi alluvial valley. Waterbirds 33: 349-356.

Davis, J. B., R. M. Kaminski, and S. E. Stephens. 1998. Wood duck egg shell membranes predict duckling numbers. Wildlife Society Bulletin 26: 299-301.

Devries J. H., R. W. Brook, D. W. Howerter, & M. G. Anderson. 2008. Effects of spring body condition and age on reproduction in Mallards (Anas platyrhynchos). Auk 125: 618–628.

Devries, J. H., L. M. Armstrong, R. J. MacFarlane, L. Moats, and P. T. Thoroughgood. 2008. Waterfowl nesting in fall seeded and spring seeded cropland in Saskatchewan. Journal of Wildlife Management 72: 1790-1797.

Dinsmore, S. J., G. C. White, and F. L. Knopf. 2002. Advanced techniques for modeling avian nest survival. Ecology 83: 3476-3488.

Duebbert, H. F., J. T. Lokemoen, and D. E. Sharp. 1983. Concentrated nesting of mallards and gadwalls on Miller Lake Island, North Dakota. Journal of Wildlife Management 47: 729-740. 115

Dugger, B. D., K. J. Reinecke, and L. H. Fredrickson. 1994. Late winter survival of female mallards in Arkansas. Journal of Wildlife Management 58: 94-99.

Dugger, B. D., R. Finger, and S. L. Melvin. 2010. Nesting ecology of mottled ducks Anas fulvigula in interior Florida, USA. Wildfowl 60: 95-105.

Durham, R. S. 2001. Nesting ecology of mottled ducks on agricultural lands in southwest Louisiana. Thesis, Louisiana State University, Baton Rouge, USA.

Durham, R. S. and A. D. Afton. 2003. Nest site selection and success of Mottled Ducks on agricultural lands in southwest Louisiana. Wildlife Society Bulletin 31: 433– 442.

Durham, R. S. and A. D. Afton. 2006. Breeding biology of mottled ducks on agricultural lands in southwestern Louisiana. Southeastern Naturalist 5: 311-316.

Dwernychuk, L. W., and D. A. Boag. 1972. How vegetative cover protects duck nests from egg-eating birds. Journal of Wildlife Management, 36: 955-958.

Dwyer, T. J. 1972. An adjustable radio-package for ducks. Bird Banding 43: 282-284

Dzubin, A. 1969. Assessing breeding populations of ducks by ground counts. Pages 178- 230 in Saskatoon wetlands seminar. Canadian Wildlife Service Report Service Number 6. Ottawa, Canada.

Elsey, R. M., P. L. Trosclair III, and J. T. Linscombe. 2004. The American alligator as a predator of mottled ducks. Crocodile Specialist Group Newsletter 23: 21.

Emery, R. B., D. W. Howerter, L. M. Armstrong, M. G. Anderson, J. H. Devries, and B. L. Joynt. 2005. Seasonal variation in waterfowl nesting success and its relation to cover management in the Canadian prairies. Journal of Wildlife Management 69:1181-1193.

Esler, D., and J. B. Grand. 1993. Factors influencing depredation of artificial duck nests. Journal of Wildlife Management 57: 244-248.

Gloutney, G. L., and R. G. Clark. 1991. Nest site selection by mallards and blue-winged teal in relation to microclimate. Auk 114: 381-395.

116

Gordon, D. H., B. T. Gray, R. D. Perry, M. B. Prevost, T. H. Strange, and R. K. Williams. 1989. South Atlantic Coastal Wetlands. Pages 57–92 in L. M. Smith, R. L. Pederson, and R. M. Kaminski, editors. Habitat management for migrating and wintering waterfowl in North America. Texas Tech University Press, Lubbock, Texas.

Gordon, D. H., B. T. Gray, and R. M. Kaminski. 1998. Dabbling duck habitat associations during winter in coastal South Carolina. Journal of Wildlife Management 62: 569-580.

Grand, J. B. 1992. Breeding chronology of Mottled Ducks in a Texas coastal marsh. Journal of Field Ornithology 63: 195–202.

Green, R. E. 1989. Transformation of crude proportions of nests that are successful for comparison with Mayfield estimates of nest success. Ibis 131: 306

Greenwood, R. J., A. B. Sargeant, D. H. Johnson, L. M. Cowardin, and T. L. Shaffer. 1995. Factors associated with duck nest success in the Prairie Pothole Region of Canada. Wildlife Monographs 128.

Hammond, M. C., and G. E. Mann. 1956. Waterfowl nesting islands. Journal of Wildlife Management. 20: 345-352.

Hammond, M. C. 1969. Notes on conducting waterfowl breeding population surveys. Pages 238-258 in Saskatoon Wetlands Seminar. Canadian Wildlife Service Report Series No. 6, Ottawa, Ontario, Canada.

Haukos, D., S. Martinez, and J. Heltzel. 2010. Characteristics of ponds used by breeding mottled ducks on the chenier plain of the Texas gulf coast. Journal of Fish and Wildlife Management 1: 93-101.

Heusmann, H. W., and J. R. Sauer. 1997. A survey for mallard pairs in the Atlantic flyway. Journal of Wildlife Management 61: 1191-1198.

Heyward, D. C. 1937. Seed from Madagascar. University of North Carolina Press, Chapel Hill, USA.

Hoekman, S. T., I. J. Ball, T. F. Fondell. 2002. Grassland birds orient nests relative to nearby vegetation. Wilson Bulletin 114: 450-456.

Hoekman, S. T., L. S. Mills, D. W. Howerter, J. H. Devries, and I. J. Ball. 2002. Sensitivity analysis of the lifecycle of midcontinent mallards. Journal of Wildlife Management 66: 883-900.

Holbrook, R. S. 1997. Ecology of nesting mottled ducks at the Atchafalaya River delta, Louisiana. Thesis, Louisiana State University, Baton Rouge, USA.

Holbrook, R. S., F. C. Rohwer, and W. P. Johnson. 2000. Habitat use and productivity of mottled ducks on the Atchafalaya River Delta, Louisiana. Proceedings of the Annual Conference of Southeastern Association of Fish and Wildlife Agencies 54: 292-303.

Hosmer, D. W. and S. Lemeshow. 2000. Applied logistic regression, 2nd edition. John Wiley & Sons, Inc., New York, USA.

Huseby, J. T., W. D. Svedarsky, and R. D. Crawford. 2001. Dabbling duck production from a wild rice farming landscape in northwest Minnesota. Wildlife Society Bulletin 29: 838-845.

Jehle, G., A. A. Y. Adams, J. A. Savidge, and S. K. Skagen. 2004. Nest survival estimation: a review of alternatives to the Mayfield estimator. The Condor 106: 472-484.

Johnson, D. H. 1979. Estimating nest success: the Mayfield method and an alternative. Auk 96: 651-661.

Johnson, D. H. 1980. The comparison of usage and availability measurements for evaluating resource preference. Ecology 61: 65–71.

Johnson, D. H., J. D. Nichols, and M. D. Schwartz. 1992. Population dynamics of breeding waterfowl. Pages 446-485 in Ecology and management of breeding waterfowl, B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu, editors. University of Minnesota Press, Minneapolis, USA.

Johnson, D. H., and J. W. Grier. 1988. Determinants of breeding distributions of ducks. Wildlife Monographs 100: 3–37.

Johnson, D. H., G. L. Krapu, K. J. Reinecke, and D. G. Jorde. 1985. An evaluation of condition indices for birds. Journal of Wildlife Management 49: 569-575.

Johnson, F. A., F. Montalbano III, and T. C. Hines. 1984. Population dynamics and status of the mottled duck in Florida. Journal of Wildlife Management 48: 1137–1143.

118

Johnson, F. A., F. Montalbano III, J. D. Truitt, and D. R. Eggeman. 1991. Distribution, abundance, and habitat use by mottled ducks in Florida. Journal of Wildlife Management 55: 476–482.

Johnson, T. W. 1974. Study of mottled duck broods in Merritt Island National Wildlife Refuge. Wilson Bulletin 86: 68.

Johnson, W. P., F. C. Rohwer, and M. Carloss. 1996. Evidence of nest parasitism in mottled ducks. Wilson Bulletin 108: 187-189.

Joyner, D. E. 1976. Effects of interspecific nest parasitism by Redheads and Ruddy Ducks. Journal of Wildlife Management 40: 33-38.

Kaminski, M. R., G. A. Baldassarre, J. B. Davis, E. R. Wengert, R. M. Kaminski. 2013. Mallard survival and nesting ecology in the lower great lakes region, New York. Wildlife Society Bulletin 37: 778-786.

Klett, A. T., H. F. Duebbert, C. A. Faanes, and K. F. Higgins. 1986. Techniques for studying nest success of ducks in upland habitats in the prairie pothole region. United States Fish and Wildlife Service, Resource Publication 158, Washington, DC, USA.

Klett, Albert T., Terry L. Shaffer, and Douglas H. Johnson. 1988. Duck nest success in the prairie pothole region. Journal of Wildlife Management 52(3): 431-440.

Korschgen, C. E., K. P. Kenow, A. Gendron-Fitzpatrick, W. L. Green, and F. J. Dein. 1996. Implanting intra-abdominal radiotransmitters with external whip antennas in ducks. Journal of Wildlife Management 60: 132–137.

Kovacik, C. F. 1979. South Carolina rice coast landscape changes. Proceedings Tall Timbers Ecology Management Conference 16: 47-65.

Krapu, G. L. 1981. The role of nutrient reserves in mallard reproduction. Auk 98: 29-38.

Kirsch, L. M. 1969. Waterfowl production in relation to grazing. Journal of Wildlife. Management 33(4): 821-828.

LaHart, D. E., and G. W. Cornwell. 1970. Habitat preference and survival of Florida duck broods. 24th Annual Conference of the Southeastern Association of Game and Fish Commissioners 117-120.

Lariviere, S., F. Messier. 1998. Effect of density and nearest neighbors on simulated waterfowl nests: can predators recognize high-density nesting patches? Oikos 83:12-20.

119

Lyon, B. E., Eadie J. M. 2008. Conspecific brood parasitism in birds: a life-history perspective. Annual Review of Ecology, Evolution, and Systematics 39: 343–363.

Mayfield, H. F. 1961. Nesting success calculated from exposure. Wilson Bulletin 73: 255-261.

Mayfield, H. F. 1975. Suggestions for calculating nest success. Wilson Bulletin 87: 456- 466.

McLandress, M. R., G. S. Yarris, A. E. H. Perkins, D. P. Connelly, and D. G. Raveling. 1996. Nesting Biology of Mallards in California. Journal of Wildlife Management 60: 94-107.

McPherson, R. J., T. W. Arnold, L. M. Armstrong, C. J. Schwarz. 2003. Estimating the nest success rate and the number of nests initiated by radio-marked mallards. Journal of Wildlife Management 67: 843-851.

McRoberts, J. T., N. T. Quintana, W. A. Smith, W. B. Ballard, F. P. Kehoe, T. G. Dilworth. 2012. Greater Scaup, Aythya marila, nest site characteristics on Grassy Island, New Brunswick. The Canadian Field Naturalist 126: No. 1.

Merendino M. T., Lobpries D. S., Neaville J. E., Ortego J. D., Johnson W. P. 2005. Regional differences and long-term trends in lead exposure in mottled ducks. Wildlife Society Bulletin 33: 1002–1008.

Moorman, T. E. and P. N. Gray. 1994. Mottled Duck. in Poole, A. and F. Gill, editors. The Birds of North America, Academy of Natural Science. Philadelphia, USA.

120

Nareff, G. E. 2009. Ecological value and bird use of managed impoundments and tidal marshes of coastal South Carolina. Thesis, University of Georgia, Athens, USA.

Newton, I., and C. R. G. Campbell. 1975. Breeding of ducks at Loch Leven, Kinross. Wildfowl 26:83- 102.

Oldenburger, S. L. 2008. Breeding ecology of mallards in the Central Valley of California. Thesis. University of California Davis, Davis, USA.

Olson, G. H., F. J. Dein, M. G. Haramis & D. G. Jorde. 1992. Implanted radio transmitters in wintering Canvasbacks. Journal of Wildlife Management 56: 325– 328.

Oring, L. W. and R. D. Sayler. 1992. The mating systems of waterfowl. Pages 190- 213 in Ecology and Management of Breeding Waterfowl, B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec and G. L. Krapu, editors. University of Minnesota Press, Minneapolis, Minnesota.

Pagano, A. M., and T. W. Arnold. 2009. Detection probabilities for ground-based breeding waterfowl surveys. Journal of Wildlife Management 73:392-398.

Pieron M. R., and F. C. Rowher. 2010. Effects of large-scale predator reduction on nest success of upland nesting ducks. Journal of Wildlife Management 74: 124-132.

Rearden, J. D. 1951. Identification of waterfowl nest predators. Journal of Wildlife Management 15: 386-39.

Reinecke, J. K., C. W. Shaiffer, and D. Delnicki. 1987. Winter survival of female mallards in the Lower Mississippi Valley. Transactions of the North American Wildlife and Natural Resources Conference 52:258–263.

Reynolds, R. E., T. L. Shaffer, R. W. Renner, W. E. Newton, and B. D. J. Batt. 2001. Impact of the Conservation Reserve Program on duck recruitment in the U.S. Prairie Pothole Region. Journal of Wildlife Management 65: 765–780.

Ringelman, K. M., J. M. Eadie, J. T. Ackerman. 2011. Density-dependent nest predation in waterfowl: the relative importance of nest density versus nest dispersion. Oecologia 169: 695-702.

121

Ronka, M., L. Saari, M. Hario, J. Hanninen, and E. Lehikoinen. 2011. Breeding success and breeding population trends of waterfowl: implications for monitoring. Wildlife Biology 17:225-239.

Rohwer, F. C. & S. Freeman. 1989. The distribution of conspecific nest parasitism in birds. Canadian Journal of Zoology, 67 239-253.

Rowland, Lawrence S. 1987. “Alone on the River:” The Rise and Fall of the Savannah River Rice Plantations of St. Peter’s Parish, The South Carolina Historical Magazine 88: 121-150.

Sargeant, A. B., and D. G. Raveling. 1992. Mortality during the breeding season. Pages 396-422 in B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. F. Kadlec, and G. L. Krapu, editors. Ecology and management of breeding waterfowl University of Minnesota Press, Minneapolis, USA.

South Carolina. South Carolina Historical Magazine 88(3):121-150.

Sauder, D. W., R. L. Linder, R. B. Dahlgren, and W. L. Tucker. 1971. An evaluation of the roadside technique for censusing breeding waterfowl. Journal of Wildlife Management 35:538-543.

Semel B., and P. W. Sherman. 2001. Intraspecific parasitism and nest-site competition in wood ducks. Animal Behaviour 61, 787–803.

Sorenson, M. D. 1991. The functional significance of parasitic egg laying and typical nesting in redhead ducks: an analysis of individual behavior. Animal Behavior 42:771-796.

South Carolina Department of Natural Resources, South Carolina Code of Laws. 2013. Marine Resources Act, Section 50-5-80.

Stephens, S. E., J. J. Rotella, M. S. Lindberg, M. L. Taper, and J. K. Ringelman. 2005. Duck nest survival in the Missouri Coteau of North Dakota: landscape effects at multiple spatial scales. Ecological Applications 15:2137-2149.

Stieglitz, W. O. and C. T. Wilson. 1968. Breeding biology of the Florida Duck. Journal of Wildlife Management 32:921–934.

122

Strange, T. H., D. H. Gordon, and J. A. Timmerman, Jr. 1989. Atlantic Flyway perspectives and expectations. Transactions of the North American Wildlife and Natural Resources Conference 54: 496-511.

Stutzenbaker, C. D. 1988. The Mottled Duck: Its Life History, Ecology, and Management. Texas Parks and Wildlife Department. Austin, TX. 209 pp.

South Carolina Department of Parks, Recreation, and Tourism (SCDPRT). 2014. Tourism regions (map).

U.S. Bureau of the Census. 2008. Projections of the Total Population of States: 1995 to 2025.

U. S. Department of Agriculture [USDA]. 2010. National Agricultural Statistics Service [NASS]. Quick Stats. < http://quickstats.nass.usda.gov/>. Accessed 25 July 2010.

Vermmer, K. 1970. Some aspects of the nesting of ducks on islands in Lake Newell, Alberta. Journal of Wildlife Management 34:126-129.

Verner, D. M. 2013. Habitat use, survival, and reproductive success of female Florida mottled ducks (Anas fulvigula fulvigula) using the Everglades Agricultural Area and urban habitats of south-eastern Florida, Auburn. Thesis, Auburn University, Auburn, USA.

Weller, M. W. 1959. Parasitic egg laying in the redhead (Aythya americana) and other North American Anatidae. Ecological Monographs 29: 333-365.

Weng, G. J. 2006. Ecology and population genetics of mottled ducks within the south atlantic coastal zone. Dissertation. University of Georgia, Athens, GA. 100.

Weller, M. W. 1956. A simple field candler for waterfowl eggs. Journal of Wildlife Management 20:111–113.

White, D. H., and D. James. 1978. Differential use of fresh water environments by wintering waterfowl of coastal Texas. Wilson Bulletin 90:99-11

White, G. C., and K. P. Burnham. 1999. Program MARK-survival estimation from populations of marked animals. Bird Study 46:120-138.

White, G. C., and R. A. Garrott. 1990. Analysis of wildlife tracking data. Academic Press, San Diego, USA.

Williams C. L., Fedynich A. M., Pence D. B., Rhodes O. E. Jr. 2005. Evaluation of allozyme and microsatellite variation in Texas and Florida Mottled Ducks. Condor 107: 155–161.

123

Wilson, B., P. D. Levesque. 1996. Egg carrying by a mottled duck. The Southwest Naturalist 41: 444-445.

Wolfe, D. F. 1996. Evidence of nest parasitism in Mottled Ducks. Wilson Bulletin 108: 187-189.

Yarrow, G. 2009. Wetland Management and Protection. Clemson Cooperative Extension, Extension Forestry and Natural Resources.

Zwank, P. J., P. M. McKenzie, and E. B. Moser. 1989. Mottled duck habitat use and density indices in agricultural lands. Journal of Wildlife Management 53:110-114.

124