In»... r .2“... .a .1 . .3.“ 2.9. m...“ L... 2% n“ . ‘ an: , a: .. . 4 my I I 41. ‘3 .tv , . 4... 1%.?» .L :4 x C; T ‘14 14;. 5-??? ,r?‘ )‘Jrfi .flcl. .J ‘ s, 2&9 3. .W5 i) .‘ru .v 13 L .W .m? ,4: . , :3 n W 61...“. f:l.fl...n : uh. . . fit 53%“ .. I. 0. 4.3.19 . x... 5.. I"... LIBRARY Michigan State University This is to certify that the dissertation entitled BIRD USE OF DIKED AND UNDIKED COASTAL WETLANDS IN MICHIGAN presented by MICHAEL JOSEPH MONFILS has been accepted towards fulfillment of the requirements for the PhD. degree in Fisheries and Wildlife W Mafor Professor’s Signature an / 2007 I Date MSU is an Affirmative Action/Equal Opportunity Employer PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5/08 K:lProj/Aoc&Pres/CIRCIDateDue.Indd BIRD USE OF DIKED AND UNDIKED COASTAL WETLANDS IN MICHIGAN By Michael Joseph Monfils A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Fisheries and Wildlife 2009 neg; dmi ABSTRACT BIRD USE OF DIKED AND UNDIKED COASTAL WETLANDS IN MICHIGAN By Michael Joseph Monfils Some Great Lakes coastal wetlands were diked to permit water level manipulations and management for waterfowl and other wetland wildlife during periods of low lake levels. Diking of coastal wetlands can alter biogeochemical cycling, flood storage, sediment movements, plant diversity, and fish and wildlife habitat. I evaluated breeding and migrant bird use during 2005-2007 at 10 diked and nine undiked sites within two coastal wetland complexes in Michigan: Saginaw Bay of Lake Huron and St. Clair River delta on Lake St. Clair. My goal was to test the hypothesis that diked coastal wetlands support greater densities and diversities of wetland birds compared to undiked sites. Breeding bird surveys consisted of 10-min point counts at random locations in emergent marsh and 30-min timed-area surveys of randomly selected areas of open water/aquatic bed wetland. Aerial surveys were done using fixed-wing aircrafi or helicopter during early fall, spring, and late summer to compare migrant waterfowl use of diked and undiked wetlands. Fall ground surveys were conducted to compare migrant waterfowl, shorebird, and waterbird use of diked and undiked sites. I measured vegetation and physical characteristics during the breeding season in emergent marsh near point count stations, and along open water-emergent marsh interfaces surveyed during migrant bird ground surveys. Vegetation and physical variable sampling revealed that diked sites were dominated by cattail, had greater water depths and percent cover of open water and floating plants, and more organic soils compared to undiked wetlands, while undiked sites had shallower water depths, greater percent cover and density of common reed and bulrush, and more inorganic soils than diked wetlands. Bird use was largely Similar between the wetland types during breeding and migration periods. Bird species richness was comparable between diked and undiked sites and similarity indices indicated high similarity in bird communities during the breeding season and early fall migration. Wood Ducks were observed in greater densities in diked and F orster’s Tern and Ring-billed Gull in undiked wetlands during breeding and migration surveys. Breeding surveys indicated that diked wetlands benefited Canada Goose, American and Least Bittems, and Common Moorhen, while Mallard, American Coot, and Herring Gull appeared more abundant at undiked sites. Although shorebird use was similar between wetland types, linear densities (birds/km edge) of Mallards and dabbling ducks were greater in undiked than diked wetlands during early fall. Water level manipulations, such as reduced water depths and periodic complete drawdowns, could increase use of diked wetlands by breeding wetland birds and migrant dabbling ducks and shorebirds. Given an uncertain future for Great Lakes coastal wetlands due to climate change and invasive species, diked wetlands may provide opportunities to maintain and improve habitat for priority wetland birds. Experimental studies are needed to identify water level management strategies that increase use by priority bird species and maximize overall wetland functioning. ACKNOWLEDGEMENTS This project was supported by the Federal Aid in Restoration Act under Pittman- Robertson project W—147-R. Partial funding for this research provided by a grant from the US. Department of Interior, Fish and Wildlife Service, via the Upper Mississippi River and Great Lakes Region Joint Venture. Additional research support provided by the Rocky Mountain Goats Foundation and Michigan State University Graduate School. Special thanks to my co-advisors, Patrick Brown and Kelly Millenbah, for their guidance, support, and encouragement. I also want to thank my other guidance committee members, Thomas Burton, Daniel Hayes, and Gregory Soulliere (US Fish and Wildlife Service), for their valuable input and advice. I greatly appreciate the substantial statistical advice provided by Daniel Hayes. Gregory Soulliere, Ernie Kafcas (Michigan Department of Natural Resources [MDNR]), and John Schafer (MDNR) initiated this research and were important cooperators during the project. Many MDNR personnel provided advice, input, equipment, and logistical support, including Barbara Avers, Donald Avers, Michael Donovan, Tim Gierman, Arnold Karr, and David Luukkenon. Donald Uzarski conducted nutrient analysis on water samples and provided advice on multivariate statistics. I appreciate the support of my colleagues at the Michigan Natural Features Inventory (MNFI) during the course of my degree. Several MNFI staff assisted with field surveys: Justin Bobick, Amy Boetcher, Kimberly Borland, Joelle Gehring, Jamie List, Rebecca Loiselle, Brandon Noel, Marie Perkins, Cole Provence, Michael Sanders, Ellen Ter Haar, and Sara Warner. I could not have accomplished this work without the encouragement of Anna and Madeline Monfils and support of my parents. iv PREFACE This dissertation is divided into four chapters. Chapter 1 provides an introduction to the issue of Great Lakes coastal wetland diking, including potential impacts to wetland functioning and benefits of wetland management, and detailed descriptions of the study areas. Chapter 1 also includes a description of the overall study design, summary of water level fluctuation and water chemistry data, and discussion of possible effects of diking on coastal wetlands. Readers interested in the primary results of my research are referred to Chapters 2 and 3, which were written as independent chapters to facilitate publication. I evaluate breeding bird use of diked and undiked coastal wetlands in Chapter 2, while in Chapter 3, I compare migrant bird use of diked and undiked wetlands. In Chapter 4, I provide a brief summary of the implications of my research with regard to the management of diked wetlands for birds, including differences in bird use of diked and undiked wetlands, management recommendations, and research needs. TABLE OF CONTENTS LIST OF TABLES ........................................................................................................... viii LIST OF FIGURES ......................................................................................................... xiv CHAPTER 1 INTRODUCTION AND STUDY AREA DESCRIPTIONS ............................................. 1 Introduction ................................................................................................................... 1 Study Design ................................................................................................................. 6 Study Area Descriptions ............................................................................................. 1 1 St. Clair Flats ........................................................................................................ l l Saginaw Bay ......................................................................................................... 13 Methods ....................................................................................................................... 17 Water Level Fluctuations ...................................................................................... 17 Water Chemistry ................................................................................................... 17 Results ......................................................................................................................... 19 Water Level Fluctuations ...................................................................................... 19 Water Chemistry ................................................................................................... 26 Discussion ................................................................................................................... 30 Literature Cited ........................................................................................................... 34 CHAPTER 2 BREEDING BIRD USE OF DIKED AND UNDIKED COASTAL WETLANDS IN MICHIGAN ................................................................................................................. 37 Introduction ................................................................................................................. 37 Methods ....................................................................................................................... 42 Point Counts .......................................................................................................... 45 Timed-area Surveys .............................................................................................. 46 Vegetation and Physical Variable Sampling ......................................................... 47 Analysis ................................................................................................................. 49 Results ......................................................................................................................... 56 Point Counts .......................................................................................................... 56 Timed-area Surveys .............................................................................................. 62 Vegetation and Physical Variable Sampling ......................................................... 69 Discussion ................................................................................................................... 74 Breeding Bird Use of Diked and Undiked Wetlands ............................................ 74 Vegetation and Physical Characteristics of Diked and Undiked Wetlands .......... 77 Management Implications ..................................................................................... 79 Research Needs ..................................................................................................... 82 Literature Cited ........................................................................................................... 85 vi CHAPTER 3 MIGRANT BIRD USE OF DIKED AND UNDIKED COASTAL WETLANDS IN MICHIGAN ................................................................................................................. 92 Introduction ................................................................................................................. 92 Methods ....................................................................................................................... 97 Aerial Waterfowl Surveys ..................................................................................... 97 Fall Migration Ground Surveys .......................................................................... 101 Vegetation and Physical Variable Sampling ....................................................... 102 Analysis ............................................................................................................... 104 Results ....................................................................................................................... 110 Aerial Waterfowl Surveys ................................................................................... 110 Fall Migration Ground Surveys .......................................................................... 110 Vegetation and Physical Variable Sampling ....................................................... 119 Discussion ................................................................................................................. 124 Bird Use of Diked and Undiked Wetlands ......................................................... 124 Vegetation and Physical Characteristics of Diked and Undiked Wetlands ........ 126 Management Implications ................................................................................... 127 Research Needs ................................................................................................... 130 Literature Cited ......................................................................................................... 133 CHAPTER 4 MANAGEMENT IMPLICATIONS ............................................................................... 139 Bird Use During Spring Migration ........................................................................... 147 Bird Use During Breeding Season ............................................................................ 148 Bird Use during Fall Migration ................................................................................. 151 Management Discussion ........................................................................................... 154 Literature Cited ......................................................................................................... 157 APPENDIX A COMMON AND SCIENTIFIC NAMES FOR BIRD SPECIES OBSERVED DURING SURVEYS ...................................................................................................... 159 APPENDIX B DATA TABLES FROM BREEDING BIRD SURVEYS AND ANALYSES .............. 163 APPENDIX C DATA TABLES FROM MIGRANT BIRD SURVEYS AND ANALYSES ................ 178 vii Table LIST OF TABLES Page Study sites surveyed and research activities conducted at St. Clair Flats and Saginaw Bay, Michigan during 2005-2007. An “X” indicates that a specific research activity was conducted at the site. Approximate areas and water management capability of Sites are listed ...................................................... 9 Means i SE by wetland type, study area, site, and period for water chemistry parameters measured during sampling conducted at St. Clair Flats and Saginaw Bay, Michigan in 2007. Data are partitioned into early (early May — mid Jul) and late (mid July — late September) periods and sample size is in parentheses ............................................................................................................ 27 Means (mg/L) t SE by wetland type, study area, and site for nitrate-N, ammonium-N, and soluble reactive phosphorous (SRP) in water samples collected at St. Clair Flats and Saginaw Bay, Michigan in late summer 2007. The number of samples for each parameter are listed in parentheses ................... 29 Least squares geometric means and lower and upper 95% confidence limits by wetland type for breeding bird densities (birds per ha) measured during point counts conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05) ........................................................................... 57 Avian species unique to diked and open wetlands and common to both types during breeding bird point counts conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007 .............................................. 58 First and second dimension coordinates for birds species/ groups included in correspondence analysis conducted using data from 605 breeding bird point counts (294 diked and 311 undiked) at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007 ................................................................ 62 Least squares geometric means and lower and upper 95% confidence limits by wetland type for area] bird densities (birds per ha open water) measured during timed-area surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05) ........................................ 63 viii Table 10 ll 12 13 14 15 Page Least squares geometric means and lower and upper 95% confidence limits by wetland type for linear bird densities (birds per km of edge) measured during timed-area surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05) ........................................ 65 Avian species unique to diked and open wetlands and common to both types during timed-area surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007 .............................................. 66 First and second dimension coordinates for birds species/ groups included in correspondence analysis conducted using data from 287 timed-area surveys (144 diked and 143 undiked) at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007 ................................................................................. 69 Least squares geometric means and standard errors for vegetation variables measured during quadrat sampling conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2006-2007. P-values for differences between wetland types provided. Bolded p-values indicate a Significant difference between wetland types (p<0.05) .......................................................... 70 Eigenvectors for first two principal components obtained through PCA of habitat data collected at 179 point count stations located at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2006-2007 .................................. 73 Approximate total area, water management capability, number of transects/ routes surveyed, and estimated area covered during migrant bird surveys at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007 ...... 99 Least squares geometric means and lower and upper 95% confidence limits by wetland type for waterfowl and waterbird densities (birds per ha wetland) measured during aerial surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05) ...................................... 111 Least squares geometric means and lower and upper 95% confidence limits by wetland type for areal bird densities (birds per ha wetland) measured during late summer/early fall ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05) ...................... 112 ix Table l6 l7 l8 19 20 21 Page Least squares geometric means and lower and upper 95% confidence limits by wetland type for linear bird densities (birds per km edge) measured during late summer/early fall ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05) ...................... 114 Avian species unique to diked and open wetlands and common to both types during late summer/early fall ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007 ............................. 115 First and second dimension coordinates for birds species/ groups included in correspondence analysis conducted using data from 45 fall migration ground surveys done along 21 routes at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007 ............................................................................... 119 Least squares means and lower and upper 95% confidence limits (CL) for vegetation and habitat variables measured during three-m2 plot sampling conducted during fall ground surveys for birds at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05) ...................................... 120 Eigenvectors for first two principal components obtained through PCA of habitat data collected during 45 fall migration ground surveys conducted along 21 routes at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007 ........................................................................................... 123 Matrix indicating potential relationships between bird species/groups and wetland conditions at St. Clair Flats and Saginaw Bay, Michigan, wetlands during breeding and migration periods. Estimated availability of wetland features at diked and undiked sites is coded as follows: no shading = absent to low; gray shading = low to medium; black shading = medium to high; and ? = uncertain status. Positive (+) signs indicate wetland features used by a bird species/group, based on this study or other research, and a “7” designates uncertainty due to limited data ............................................................................ 143 Common and scientific names of avian species observed during bird surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007. Species are listed by wetland use category .......... 160 Estimated mean densities i SE for priority and common marsh bird species observed during surveys at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan in 2005-2007 by wetland type and distance category ............ 164 Table B-3 B-S B-6 Page Estimated frequency of occurrence (number of points with species present/number of points surveyed) for priority and common marsh bird species observed during surveys at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan in 2005-2007 by wetland type and distance category ............ 166 Akaike’s Information Criterion (AIC) statistics and P-values for mixed models used to compare bird densities in diked and open wetlands during point counts conducted at St. Clair Flats and Saginaw Bay, Michigan, 2005-2007. Models that included a repeated measures component are listed by covariance structure. Bolded values indicate the most desirable model for a given variable based on AIC statistics. An asterisk “*” was placed after an AIC value if the G matrix for the given model was not positive definite. The notation “---“ indicates that the model did not converge .............. 168 Akaike’s Information Criterion (AIC) Statistics and P-values for mixed models with lower bounds set for covariance parameters with zero estimates to achieve positive definite G matrices. Models were used to compare bird densities between diked and open wetlands during point counts conducted at St. Clair Flats and Saginaw Bay, Michigan, 2005-2007. Models that included a repeated measures component are listed by covariance structure. Bolded values indicate the most desirable model for a given variable based on AIC statistics. The notation “---“ indicates that the model did not converge .............................................................................................................. 170 Mean areal densities (birds/ha), standard error, and frequency of occurrence (in parentheses) by study area and wetland type for bird species observed during breeding bird point counts conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007. Frequency of occurrence is the proportion of point counts that the species was observed ......................... 172 Mean areal densities (birds/ha), standard error, and frequency of occurrence (in parentheses) by study area and wetland type for bird species observed during timed-area surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007. Frequency of occurrence is the proportion of open water areas that the species was observed ...................... 176 xi Table C-1 C-2 C-3 Page Akaike’s Information Criterion (AIC) statistics and P-values for mixed models used to compare migrant bird areal densities (birds per ha wetland) in diked and undiked wetlands during ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, 2005-2007. Models that included a repeated measures component are listed by covariance structure. Bolded values indicate the most desirable model for a given variable based on AIC statistics. An asterisk “*” was placed after an AIC value if the G matrix for the given model was not positive definite. The notation “---“ indicates that the model did not converge ................................................................................. 179 Akaike’s Information Criterion (AIC) statistics and P-values for mixed models with lower bounds set for covariance parameters with zero estimates to achieve positive definite G matrices. Models were used to compare migrant areal bird densities (birds per ha wetland) between diked and undiked wetlands during ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, 2005-2007. Models that included a repeated measures component are listed by covariance structure. Bolded values indicate the most desirable model for a given variable based on AIC statistics. The notation “---“ indicates that the model did not converge. Data from the original model was reported if the G matrix was positive definite (denoted by the “I” symbol) .................................................................. 181 Akaike’s Information Criterion (AIC) statistics and P-values for mixed models used to compare linear densities of migrant bird (birds per km edge) in diked and undiked wetlands during ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, 2005-2007. Models that included a repeated measures component are listed by covariance structure. Bolded values indicate the most desirable model for a given variable based on AIC statistics. An asterisk “*” was placed after an AIC value if the G matrix for the given model was not positive definite. The notation “---“ indicates that the model did not converge ................................................................................. 184 Akaike’s Information Criterion (AIC) statistics and P-values for mixed models with lower bounds set for covariance parameters with zero estimates to achieve positive definite G matrices. Models were used to compare linear migrant bird densities (birds per km edge) between diked and undiked wetlands during ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, 2005-2007. Models that included a repeated measures component are listed by covariance structure. Bolded values indicate the most desirable model for a given variable based on AIC statistics. The notation “-—-” indicates that the model did not converge. Data from the original model was reported if the G matrix was positive definite (denoted by the “T” symbol) .................................................................................................... 186 xii Table C-5 06 Page Mean densities (birds/ha), standard errors, and frequencies (in parentheses) by study area, wetland type, and survey period for several waterfowl and waterbird species observed during 14 aerial surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007. Frequencies are the proportions of transects with the species present ................ 189 Mean densities (birds/ha), standard error, and frequency of occurrence (in parentheses) by study area and wetland type for bird species observed during late summer/early fall ground surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007. Frequency of occurrence is the proportion of surveys that the species was observed .............. 191 xiii Figure LIST OF FIGURES Page Locations of St. Clair F lats (Lake St. Clair) and Saginaw Bay (Lake Huron) coastal wetland study sites investigated during 2005-2007 in Michigan. Abbreviations used in text and tables are provided in parentheses ....................... 12 Water level fluctuations by week and year during late spring and summer at the East Marsh and West Marsh diked sites at St. Clair Flats, Michigan 2005- 2007. The y-axis references selected heights on staff gages, rather than true water elevation (i.e. meters above sea level) of the study sites ............................ 20 Water level fluctuations by week of year during late spring and summer at undiked Algonac and St. Clair Shores NOAA water gage locations near St. Clair Flats, Michigan 2005-2007. The y-axis represents true water elevations above sea level for the St. Clair River (Algonac gage) and Lake St. Clair (St. Clair Shores gage) ................................................................................................. 21 Long-term (1918-2007) average water levels in meters above sea level for Lake St. Clair and Lakes Michigan and Huron by month. Error bars indicate record high and low water levels for each month. Data obtained from the US Army Corps of Engineers website (www.Ire.usace.army.mil/greatlakes/ hh/greatlakeswaterlevels/historicdata/greatlakeshydrographsl) ............................ 22 Water level fluctuations by week of year during late spring and summer at diked sites on Saginaw Bay, Michigan 2005-2007: Fish Point Refuge (FPR), Nayanquing Point (NPE, NPN, and NPS), and Wigwam Bay (WBD). The y-axis references selected heights on staff gages, rather than true water elevation (i.e. meters above sea level) of the study sites ...................................... 24 Water level fluctuations by week of year during late spring and summer at open Essexville NOAA water gage on Saginaw Bay, Michigan 2005-2007. The y-axis represents true water elevation in meters above sea level for Saginaw Bay, Lake Huron .................................................................................... 26 Illustration of study design used for breeding bird point counts conducted in Great Lakes coastal wetlands in Michigan (St. Clair Flats and Saginaw Bay), 2005-2007. Independent study sites (lettered polygons) were sampled within each study area, with approximately half of the points occurring in each of two wetland types (diked — shaded, undiked — not Shaded). Points (black dots) were situated randomly within each study site, and three surveys (early, mid, and late season) were conducted at each point. ............................................ 43 xiv Figure 10 ll 12 Page Illustration of study design used for timed-area surveys for breeding birds conducted in Great Lakes coastal wetlands in Michigan (St. Clair Flats and Saginaw Bay), 2005-2007. Independent study sites (lettered polygons) were sampled within each study area, with approximately half of the open water areas occurring in each of two wetland types (diked -— shaded, undiked — not shaded). Open water areas (polygons) were randomly selected (shaded polygons) within each study site. During each of four survey periods, a new set of open water areas were randomly selected. .................................................. 44 Biplot of site and bird group coordinates for dimensions 1 and 2 from correspondence analysis conducted using point count data collected at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan, coastal wetlands, 2005- 2007. Site coordinates are coded by wetland type (“+” diked; “o” undiked). Bird group coordinates are coded with an “*” and labeled as follows: AF = aerial-foraging songbirds, BI = bitterns, CM = American Coots and Common Moorhens, GR = Pied-billed Grebes, HE = herons, N0 = non-wetland birds, RA = rails, SA = wetland-associated songbirds, SW = wetland-dependent songbirds, TG = terns and gulls, and WA = waterfowl ........................................ 61 Biplot of site and bird group coordinates for dimensions 1 and 2 from correspondence analysis conducted using timed-area survey data collected at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan, coastal wetlands, 2005-2007. Sites scores are coded by wetland type (“+” diked; “O” undiked). Bird groups are coded with an “*” and labeled as follows: BI = bitterns, CM = American Coots and Common Moorhens, DA = dabbling ducks, D1 = diving ducks, GR = Pied-billed Grebes, GS = Canada Geese and Mute Swans, HE = herons, RA = rails, SH = shorebirds, TG = terns and gulls, and WD = Wood Ducks .................................................................................................................... 68 Bi-plot of PC 1 X PC 2 from principal components analysis conducted using 14 vegetation and physical variables gathered during quadrat sampling at 179 random avian point count stations at St. Clair Flats and Saginaw Bay, Michigan, 2006-2007. Point scores are coded by wetland type (“+” diked; “o” undiked) ......................................................................................................... 72 Locations of St. Clair Flats (Lake St. Clair) and Saginaw Bay (Lake Huron) aerial waterfowl transects (black lines) surveyed during 2005-2007 in Michigan. Abbreviations used in text and tables are provided in parentheses ............................................................................................................ 98 XV Figure 13 14 15 16 17 Page Biplot of site and bird group coordinates for dimensions 1 and 2 from correspondence analysis conducted using fall migration ground survey data collected at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan, coastal wetlands, 2005-2007. Site coordinates are coded by wetland type (“+” diked; “o” undiked). Bird group coordinates are coded with an “*” and labeled as follows: B1 = bitterns, CM = American Coots and Common Moorhens, DA = dabbling ducks, D1 = diving ducks, GR = Pied-billed Grebes, GS = Canada Geese and swans, HE = herons, RA = rails, SH = shorebirds, TG = terns and gulls, and WD = Wood Ducks ................................ 118 Bi-plot of PC 1 X PC 2 from principal components analysis conducted using 13 vegetation and physical variables gathered during plot sampling during 45 fall migration bird surveys of 21 routes at St. Clair Flats and Saginaw Bay, Michigan, 2005-2007. Point scores are coded by wetland type (“+” diked; “O” undiked) ............................................................................ 122 Typical simplified profiles of diked and undiked St. Clair Flats and Saginaw Bay coastal wetlands during early spring 2005-2007. Lines indicate vegetation zones used by bird Species/ groups during spring migration ............. 140 Typical simplified profiles of diked and undiked St. Clair Flats and Saginaw Bay coastal wetlands during late spring to mid summer 2005-2007. Lines indicate vegetation zones used by bird species/ groups during the breeding season .................................................................................................................. 14] Typical simplified profiles of diked and undiked St. Clair Flats and Saginaw Bay coastal wetlands during late surmner to early fall 2005-2007. Lines indicate vegetation zones used by bird species/ groups during early fall migration ............................................................................................................. 142 xvi CHAPTER 1 INTRODUCTION AND STUDY AREA DESCRIPTIONS INTRODUCTION Great Lakes coastal wetlands provide vital breeding, migration, and wintering habitat for an array of birds. Approximately three million swans, geese, and ducks travel along migration corridors that cross the Great Lakes region (Great Lakes Basin Commission 1975, Bellrose 1980). Great Lakes coastal wetlands are also valuable stopover habitats for migrant shorebirds that breed in the boreal and arctic regions of North America (Brown et al. 2000). These wetlands are some of the region’s largest remaining emergent marshes and provide vital nesting habitat to wetland birds, including rare and declining species such as American Bittem (Botaurus lentiginosus), Least Bittem (Ixobrychus exilis), Common Moorhen (Gallinula chloropus), King Rail (Rallus elegans), Black Tern (Chlidom'as niger), and Forster’s Tern (Sternaforsteri). Prince and Flegel (1995) summarized breeding bird atlas data from Michigan and Ontario. Eighty bird species used coastal wetlands of Lake Huron as breeding habitat (Prince and F legel 1995). Dikes and water control structures have long been used by wildlife managers to enhance wetlands for wildlife (Kadlec 1962), especially breeding and migrating waterfowl. Impounded wetlands are typically managed as hemi-marshes to maximize breeding bird use (Weller and Spatcher 1965, Weller and Fredrickson 1974, Kaminski and Prince 1981a, b, Murkin et al. 1982) or shallow-water marshes dominated by moist- soil vegetation to attract migrant birds (Fredrickson and Taylor 1982). Hemi-marshes are marshes with approximately equal proportions of emergent vegetation and open water produced by natural water level fluctuations and mammal herbivory. Historically, Great Lakes coastal wetlands moved landward and lakeward with the rise and fall of the Great Lakes. Between the 19505 and 19705, many Great Lakes coastal marshes were isolated from these normal water level fluctuations through dike construction. These projects were initiated primarily to maintain elevated water depths and enhance wildlife use during periods of historic low water levels. Shoreline armoring, wetland diking and tiling to drain wetlands for agricultural use, and other land-use changes now prevent the landward movement of coastal wetlands in much of the Great Lakes during periods of high water levels (Prince et al. 1992, Gottgens et a1. 1998). The potential problems associated with isolating coastal wetlands from the Great Lakes include impaired or eliminated flood conveyance and storage, sediment control, and water quality improvement firnctions, altered nutrient flow, reduced or degraded habitat for shorebirds, rare species, fish, and invertebrates, and increased impacts from trapped carp (Cyprinus carpio) (Jude and Pappas 1992, Wilcox 1995, Wilcox and Whillans 1999). By separating coastal wetlands from the fluctuations of the Great Lakes, dike construction often stabilizes water levels. Stable water levels typically compress wetland vegetation zones and encourage dominance by shrubs and highly competitive species, such as willow (Salix spp.), alder (Alnus spp.), cattail (Typha spp.), reed canary grass (Phalaris arundinacea), and purple loosestrife (Lythrum salicaria). Irregular water levels may result in higher levels of diversity both within and among habitats (Keddy and Reznicek 1986, Wilcox 1993, Wilcox et al. 1993, Keough et al. 1999). Comparisons of plant communities in diked and undiked Great Lakes coastal wetlands have yielded varied results. Herrick and Wolf (2005) documented increased amounts of invasive Species in standing vegetation and seed banks of diked compared to undiked wetlands in Saginaw Bay, Michigan and Green Bay, Wisconsin, but noted that current conditions in undiked wetlands appear to favor an invasive haplotype of common reed (Phragmites australis). Conversely, Galloway et al. (2006) found greater species richness and percent cover of native species and lower species richness and percent cover of invasive species in diked compared to undiked coastal wetlands. Herrick et al. (2007) found more seeds from a greater number of species in the soils of diked compared to undiked wetlands and stated that diked wetlands may serve as “traps” for plant seeds, meaning seeds are held in place by dikes due to reduced water exchange with the lakes. In comparisons between vegetation in diked and undiked Lake Erie coastal wetlands during a high water year, Thiet (2002) found greater wetland plant diversity in diked wetlands compared to a nearby undiked site. An actively managed diked marsh in southwest Lake Erie maintained emergent vegetation, patchiness, and edge habitat similar to historic conditions during periods of high Great Lakes water levels, while the same measures declined in marshes connected to Lake Erie (Gottgens et a1. 1998). Research conducted by several authors on animal use of Great Lakes coastal wetlands provides insights into the possible effects of diking on animal communities. McLaughlin and Harris (1990) compared aquatic insect emergence in one diked and one undiked wetland on Green Bay, Wisconsin and recorded more insect taxa and greater total insect biomass emergence from the diked wetland. Burton et al. (2002) noted that both plant community composition and exposure to wave action were important in determining invertebrate diversity and biomass in Great Lakes marshes. Invertebrates were distributed along gradients of decreased mixing of pelagic water and increased sediment organic matter from outer to inner marsh and between littoral and adjacent inland marshes. Some invertebrates were more common on gradient ends, but most species were generalists found across all habitat types (Burton et a1. 2002). Whitt’s (1996) study of avian breeding use of Saginaw Bay coastal wetlands included study sites that were both open to and inland from Lake Huron. Although species richness was similar between coastal and inland cattail marshes, bird densities in marshes located far offshore were lower than most other sites. Whitt (1996) suggested this difference may be due to the effects of storm surges during the breeding season that can destroy nests, and stated that further study is needed to compare avian use of protected marshes with those exposed to storm surges. Galloway et al. (2006) conducted a one-year study of breeding bird use of diked and undiked Great Lakes coastal wetlands along Lakes Ontario, Erie, and St. Clair. In pooled comparisons of diked and undiked Sites, they observed greater abundance and species richness for several groups of birds in diked wetlands. Galloway et al. (2006) also noted the need for additional research to account for long-term variation in bird and vegetation communities associated with Great Lakes water level cycles and management activities. No research has been conducted in the Great Lakes region to assess the effects of coastal wetland diking on bird communities during migration periods. Ecological studies of the effects of coastal wetland isolation from natural, highly variable water level fluctuations are needed so that informed decisions can be made about the management and restoration of Great Lakes marshes. The goal of this project was to compare bird use, habitat composition and structure, and physical and chemical attributes of several diked and undiked wetlands in Michigan to gain insights into the effects of wetland diking on avian communities. I tested the hypothesis that coastal impoundments with managed water levels provide enhanced habitat for wetland birds compared to undiked wetlands. I view this research as one of many comparisons needed over the long-term to better understand how diked and undiked wetlands function during the full cycle of Great Lakes water levels. STUDY DESIGN My objectives for this project were to 1) compare indices of bird abundance and diversity between diked and undiked coastal wetlands, 2) gather information on the vegetation structure and composition of diked and undiked wetlands and investigate potential relationships with bird abundance and diversity, and 3) characterize the physical and chemical environment of diked and undiked wetlands. Indices of bird use, vegetation, and physical and chemical attributes were compared between diked and undiked wetlands to investigate the potential effects of diking on Great Lakes coastal wetlands and test the hypothesis that impounded coastal wetlands provide improved habitat for wetland birds compared to undiked wetlands. A study of invertebrate abundance and composition was undertaken by another investigator through detailed comparisons of diked and undiked wetlands at the St. Clair Flats (see Provence 2008). I focused my research in two of Michigan’s most important coastal wetland complexes, the St. Clair River delta, also known as the St. Clair Flats (SCF), and Saginaw Bay (SAG). The St. Clair Flats is a 17,500 ha wetland complex in the US. and Canada where the St. Clair River flows into Lake St. Clair. About one-third of the St. Clair Flats is diked and approximately one-third is in US territory (Bookhout et al. 1989). Lake Huron’s Saginaw Bay contains a substantial concentration of Michigan’s coastal marshes (about 2,500 ha) (Bookhout et al. 1989), which occurs as a nearly continuous strip along the perimeter of the bay (Prince et al. 1992). The St. Clair Flats and Saginaw Bay are two of the four major coastal wetland complexes identified by Krieger et al. (1992) in their call for more research in Great Lakes coastal wetlands. I selected these wetland complexes for several reasons: 1) they are two of Michigan’s largest and most intact wetland complexes, 2) rare and declining waterbird Species of management importance use these complexes for breeding, 3) their importance as migratory stop-overs for waterfowl, waterbirds, and shorebirds, and 4) the presence of both managed diked wetlands and unmanaged undiked wetlands. I classified diked wetland sites into three water level management categories: active, opportunistic, and passive. Active management occurred at sites where pump stations were used to manipulate water levels on a regular basis. Opportunistic water management took place at sites with pumps that can only function when Great Lakes water levels are above a minimum height, so water was only pumped into the diked wetlands when conditions allowed. Passive water level management occurred at sites with dikes and water control structures, but without water pumping capabilities. Water levels in these wetlands were independent of Great Lakes levels; however, pumping was not an option and water inputs came from precipitation or through control structures. 1 selected sites to ensure that diked wetlands were sampled in all three water level management categories; however, I was not able to make comparisons among the three management regimes due to the low number of sites within each category. Several bird surveys were used to produce indices of bird abundance, species richness, and diversity. Indices, rather than total population estimates, were used, because total population estimates are expected to vary based on size of the study sites and are less important than relative differences in use by Species of management importance. I assumed densities (birds/ha of wetland or birds/km of edge) and other indices would vary based on species’ food preferences and differences in vegetation and physical aspects of the study sites, so differences in bird communities between diked and undiked wetlands should have been evident in the indices if they existed. Both breeding and migrant bird surveys were conducted at 19 study sites (Table 1). I measured vegetation composition and structure at sites where bird surveys were conducted to evaluate possible relationships between habitat conditions and bird use. Provence (2008) sampled the invertebrate community at diked and Open wetlands of the St. Clair Flats. Staff gages were installed and monitored at several diked wetlands to compare their water level fluctuations with undiked wetlands, since water level changes can affect use by breeding and migrant birds. I used data from National Oceanic and Atmospheric Administration (NOAA) water level stations on Lake St. Clair/St. Clair River and Saginaw Bay to characterize water level fluctuations at the undiked Sites. Basic water chemistry parameters and nutrient levels were collected in 2007 to describe conditions at the Study sites, since other researchers observed differences in water and soil chemistry of diked compared to undiked coastal wetlands (Robb 1989, Herrick and Wolf 2005). E><>< ><><>< ><><>< ><><>< ><><>< ><><>< E><>< X x III III x x x x E><>< om: S: 0: SN wmm w; ZZZZZZ E><>< XXX XXX j><>< ><><>< m2 mwm mom Z A m XXX XXX mm 0 XXX XXX ><><>< XXX XXX ><><>< XXX XXX E><>< E><>< E><>< j><>< X XX ><>< ><><>< we mm X: nw~ mm 0 O L O m E><>< XXX XXX ><><>< ><><>< i><>< XXX XXX omm wvw Z 2 XXX XXX E><>< XXX XXX XXX XXX E><>< mom omm < < boom coca moon e586 zen boom coom moon Eofiofia 36¢. whet—am “H.832 boom ooom 33.. 33-383 Room meow meow $550 Emom mxoizm magnum $5 82 van—33 48am? EoEomeNE 83>? 53 DE» <3 BE 25 0.: man Om? mOH HAZ mmZ ZmZ mmZ ME”— «in 33: ED <23 <35 woo—=25 woo—5 Baez: v8.5 Nam Bufiwam $2.» .520 um Beam: 03 8% .«o mom—£88 aeofiowmgfi 633 use 308 omemxoamaq. :86 2: E 385:8 33 £28.“ 5332 058% m :2: 830%.: L? :< .hoomfioom wage 53:32 chum 39:me use mum—m :30 am 3 383:8 8:385 5388 p5 combine. 8% Spam A 035. 5235806 5555 .85 5on 8m .28 Zn Z 28 6>mmmamum .ouflfifiomacuo d>uo8an< .qmwfioaz 5 38-38 waist woummwmgfi moxm beam 55:03 338 90.83 83.6 ham 39:me Ea CEO am 85d mum—m .520 .um mo 23:80; A oBmE 85“: .9280 .5 £5 . 1. ... D, «a 2:9 wowmonmso . A .;. 35”: A 60b news. .5... . <5“: 6.x .mmhmm ‘ 5%. 6m czw3< 8th 5 55 .5 £5 Em“: C52 .22 .252 .528... 6959.... an. mciccgmz 2.... . . .5 £5 .- 4 . 3.5 5m .265? 25v 92885 .2 693 825 >mm Em3m5> ,. Mwfimv A0m>>v :30 mew. >mm E9593 1 111.2%... . , . 5.2.: 9mm. wooEmomsi M 9m 2m 2:: \\ . n 5&5 aw\.w_ COwC_v_Q_D 12 Dickinson Island/Fisher and Goose Bays (DIS): Dickinson Island is located northwest of Harsens Island and was dominated by emergent wetlands. Marshes were also found to the immediate west and southwest along the margins of Fisher and Goose Bays. Emergent marshes were dominated by bulrushes (Schoenoplectus acutus and S. pungens), common reed, and cattail to a lesser degree. Areas of non-persistent emergent vegetation dominated by arrowhead (Sagittaria spp.), pickerelweed (Pontederia cordata), and wild rice (Zizania spp.) were present in Mud Lake and other protected areas. Scattered water lilies, stoneworts, and aquatic macrophytes were present in aquatic bed zones of protected sites. Chara spp. typically dominated the aquatic bed vegetation. Little and Big Muscamoot Bays (LMU): Little and Big Muscamoot Bays are found west of Harsens Island between the North and South Channels of the St. Clair River. The vegetation was similar to that of the Dickinson Island area, with zones of bulrush, common reed, cattail, non-persistent emergents, and aquatic bed wetland. Saginaw Bay Fifteen wetlands were studied on Saginaw Bay, of which eight were diked and seven were undiked and open to Lake Huron water level fluctuations (Figure 1). Fish Point: I studied both diked and undiked wetlands at Fish Point State Wildlife Area, which is in Tuscola County. I conducted surveys at the east diked unit of the refuge (F PR) all three years. Cattail and aquatic bed vegetation were the dominant wetland zones, although areas of wet meadow (sedges and grasses), common reed, and scrub-shrub (Salix spp. and Cornus spp.) vegetation were also present. White and yellow water lilies, water milfoil (Myriophyllum spp.), pondweeds (Potamogeton spp.), and 13 Chara spp. dominated the aquatic bed zone. Waterfowl nesting islands were constructed and level ditching was conducted many decades ago to enhance waterfowl habitat. Small pockets of cottonwood (Populus deltoides) existed, often on old nesting islands or dredge spoils. A pump station is present at this site, although pumping only occurred in 2006 due to low Lake Huron levels. A second small diked wetland was investigated near Austin Road (FPA) during timed-area and fall ground surveys for migrant birds; point counts for breeding birds were not conducted due to its small size and limited emergent marsh. Vegetation was similar to F PR and consisted of aquatic bed wetland, cattail marsh, and wet meadow dominated by sedges, rushes, and spikerushes (Eleocharis spp.). Two areas of undiked wetland were surveyed: one east of FPR near Berger Road (F PB), and a large area of flinging coastal wetland (FPC) to the southwest of F PR and FPA (Figure 1). Both undiked wetlands were dominated by emergent marshes of common reed, cattails, and bulrushes. Small pockets of wet meadow with sedges, rushes, and spikerushes were also present. I only conducted point counts at the F PB site in 2005 and the FPC site was only used for aerial waterfowl surveys (Table 1). Nayanquing Point: Four diked wetland areas were studied at Nayanquing Point State Wildlife Area (Bay County): East Marsh (NPE), North Marsh (NPN), South Refuge Unit (NPS), and Triangle Refuge Unit (N PT). The NPN, NPS, and NPT sites have water pumps that permit pumping opportunistically when Lake Huron levels allow, while NPE is a passively managed impoundment formed inside of a natural beach ridge with a small dike and water control structure. All sites were dominated by cattail marsh and aquatic bed wetland consisting of water lilies and aquatic macrophytes. Small areas of wet meadow were present at the NPE and NPN sites. Areas of non-persistent emergents l4 dominated by pickerelweed and arrowhead were present at NPN, NPS, and NPT. Small areas of common reed and hardstem bulrush were also found at NPE. I only conducted point counts at NPE and NPN, because of the limited amount of emergent marsh at NPS and small size of NPT. I conducted timed-area and fall ground surveys at all sites except NPE, which had limited open water/aquatic bed habitat. Pinconning (PIN): I surveyed undiked coastal wetland associated with the mouth of the Pinconning River (Bay County) during aerial waterfowl surveys. This area was dominated by mixed emergent marsh stands of common reed, cattail, and bulrush. Quanicassee (QUA): This site consists of undiked wetland to the northwest of the Quanicassee River mouth and is located in the Quanicassee State Wildlife Area in Tuscola and Bay Counties. The vegetation was dominated by common reed, often found in conjunction with other emergent species, such as three-square and hardstem bulrush, rushes, and cattail. Fringing zones of bulrush and cattail occurred in deeper water. Tobico Marsh (T OB) .' Tobico Marsh is an impounded wetland located in the Bay City State Recreation Area in Bay County. Historically this was a protected coastal wetland located behind a beach ridge. A small dam and control structure was installed to regulate water levels. Tobico Marsh was dominated by cattail marsh and aquatic bed wetland, with some areas of wet meadow and shrub wetland around the perimeter. I only visited this site during aerial waterfowl surveys. Wigwam Bay: I surveyed two undiked and one diked wetland sites in the Wigwam Bay State Wildlife Area in Arenac County. The Pine River site (PIR) encompassed undiked coastal wetlands north and south of the confluence of the Pine River on Saginaw Bay in Arenac County. Dominant vegetation consisted of bulrush 15 (three-square and hard-stem), cattail, and wet meadow zones. Wet meadows were dominated by sedges, grasses, rushes, and spikerushes. A large diked wetland site (WBD) is located on the north side of Saginaw and Wigwam Bays. Pump stations are not present, but water control structures regulate inflows and outflows. Emergent vegetation primarily consisted of cattail marsh and sedge meadow, both of which often occurred as floating mats. Large areas of aquatic bed wetland were dominated by white and yellow water lilies and aquatic macrophytes (e. g. Utricularia spp. and Potamogeton spp.). Sporadic hard-stem bulrush and wild rice were also present, and forested and scrub-shrub wetland was found in the northwestern portion of the impoundment. Point counts were conducted at a second undiked wetland site (WBU) located east of WBD in 2005. This area is dominated by wet meadow vegetation with fringing zones of bulrush and cattail. Wildfowl Bay (WIL): I investigated protected undiked wetlands in the Wildfowl Bay State Wildlife Area in Huron County. These wetlands formed behind Heisterman, Maison, and Middle Grounds Islands. Several wetland vegetation zones were present, including bulrush and cattail marshes, common reed stands, wet meadows, and non- persistent emergent areas consisting of arrowhead, pickerelweed, and wild rice. l6 METHODS Water Level Fluctuations I monitored staff gages at a subset of diked coastal wetlands to characterize the fluctuation of water levels during spring and summer and to compare these fluctuations with changes observed in Great Lakes water levels. In 2005, gages were read at least once per month at the EMA, WMA, F PR, and WBD sites between early May and early September. I monitored gages at least monthly at two St. Clair Flats sites (EMA, WMA) and five Saginaw Bay sites (N PE, NPN, NPS, FPR, WBD) from early May through early September in 2006 and 2007. I used hourly NOAA water level monitoring station data to characterize fluctuations at the open wetland sites. Data from two stations, Algonac, Michigan and St. Clair Shores, Michigan, were used to represent water levels in undiked wetlands at St. Clair Flats. The Essexville, Michigan station located at the confluence of the Saginaw River and Saginaw Bay was used to evaluate fluctuations at Saginaw Bay undiked sites. I averaged water level data from NOAA stations by year and week to allow comparisons with diked sites. Water Chemistry While conducting bird surveys in 2007, I gathered data on the following water parameters: temperature, dissolved oxygen (DO), pH, turbidity, alkalinity, and nutrient levels (nitrate-N, ammonium-N, and soluble reactive phosphorus [SRP]). I only intended to use the water chemistry data to characterize the study sites. Data collection varied by time of day and season and was not intensive enough to permit statistical comparisons l7 between the wetland types. I measured water temperature and DO with a YSI 55® DO meter, pH using an Oakton pH Testr 3+®, turbidity via an Oakton® T-100 turbidity meter, and alkalinity using Hach® single parameter drop titration kits. I summarized data for these parameters by study area, wetland type, site, and time period (early [May — mid July] and late [mid July — late September] season). In late August and September, I collected water samples for nutrient analysis at four sites at St. Clair Flats (two diked and two undiked) and six sites on Saginaw Bay (three diked and three undiked). I gathered three water samples from each of three vegetation zones (common reed, cattail, and bulrush), when present, at the vegetation-open water interface using sterilized bottles or plastic bags. Water samples were immediately placed on ice in the field. I filtered samples using 0.5 micron membrane and then froze them for later analysis. Dr. Donald Uzarski (Central Michigan University) conducted analyses for nitrate-N, ammonium-N, and SRP using procedures recommended in the Standard Methods for the Examination of Water and Wastewater (American Public Health Association 1992). Quality assurance/quality control procedures followed protocols recommended by the US. Environmental Protection Agency. 18 RESULTS Water Level Fluctuations Staff gage monitoring at diked St. Clair Flats sites indicated highest water levels in spring and declining levels throughout much of the growing season (Figure 2). Levels consistently declined at EMA throughout the monitoring period, with lowest levels in August or September. At WMA, lowest water levels were in July or August, with increasing levels occurring in the late summer in response to precipitation and/or pumping to increase water levels for fall waterfowl hunting. Results from the Algonac and St. Clair Shores gaging stations were consistent with those of the SCF diked wetlands in 2005 and 2007 (Figure 3). Similar to the diked wetlands, water elevations during the monitoring period were highest in spring, declined throughout the spring and summer, and were lowest in September. The overall drop in water levels in the undiked gages in 2005 and 2007 was lower compared to the diked wetlands. Water levels observed in the undiked gages in 2006 were lower in spring, increased during the spring and early summer to a peak in late July, and then declined in the late summer. This pattern is similar to the annual cycle typically observed in Great Lakes water levels (Figure 4). Lake St. Clair water levels are usually lowest in late winter, increase during spring and early summer, peak in July, and then decrease during late summer and fall. 19 Elevation (m) ........ 2005 - I -2006 +2007 12 #..LL _. 2-- ””2222 m”-.. _-_._,-.-_-_. A 1.1 ‘i‘ M- -2 .,.__ ,-,......2---22,..-M-__...._M...- _-_.___.--__.~ g ,0 ' ------ -~.‘.;_.___t__.__ _ .5 0 9 fig“ Wit-L... ‘ ‘I. A A E ' ' "'2'. ....... ' " ""'"' 2°05 5 0'3 \ "Era-r" ~'_, “ - . -2006 0.7 N +2007 06 geisha»leis:‘9.27233.Féaiévk’e' ggassésssaéééé§9§9§sa 222 now “PP“ <<< mm West Marsh Figure 2. Water level fluctuations by week and year during late spring and summer at the East Marsh and West Marsh diked sites at St. Clair Flats, Michigan 2005-2007. The y- axis references selected heights on staff gages, rather than true water elevation (i.e. meters above sea level) of the study sites. 20 175.4 - ~—~ 175.2 175.1 175.0 . 1749 _ + 174.8 ' Lake Elevation (m) 175.1 1 175.0 174.9 174.8 174.7 Lake Elevation (m) St. Clair Shores Figure 3. Water level fluctuations by week of year during late spring and summer at undiked Algonac and St. Clair Shores NOAA water gage locations near St. Clair Flats, Michigan 2005-2007. The y-axis represents true water elevations above sea level for the St. Clair River (Algonac gage) and Lake St. Clair (St. Clair Shores gage). 2l 176.5 175.0 ' 174.5 -0. U .. _. . C. .0. - W... C] 1 73 . 5 "i ”*‘_"7‘~""‘ '7 __... F ‘ "'1'” " " " "'7' "' " ' "'-T”'"‘""‘T ' "’ " ' ' ”"' “'"‘ ‘ ’T”’”"" ""T_ “ "’ ”"5" ""““'.‘" """wj Lake Elevation (m) Mar Lake St. Clair _L . _ 1 -._ l .3 i i I . i i .l l T 3 i : E a g s . l 1 i 1 s I I l l - 1 l , 2' 3 z . . ' l i 7- l l 5 5 L i l '1 1 ‘s l l L 7 l '. i i i: l i‘ 1 . g I I l i I i a E i i : I f .i 1; ? : 2 ' . ! l ‘ i : i l : 5 I ’ l I l 1 . I . 1 l ’ s 1 i l . , ; l i I - 2 , i g 1 I l l E i . i . l a i l l l l I .- l 1 i i Lakes Michigan and Huron Figure 4. Long-term (1918-2007) average water levels in meters above sea level for Lake St. Clair and Lakes Michigan and Huron by month. Error bars indicate record high and low water levels for each month. Data obtained from the US. Army Corps of Engineers website (www.1re.usace.army.mil/greatlakes/hh/greatlakeswaterlevels/ historicdata/greatlakeshydrographs/). 22 Water level fluctuations at diked SAG sites were similar to diked wetlands at SCF. Levels were usually highest in the spring and declined throughout the monitoring period (Figure 5). Exceptions occurred at those sites with water pumping stations. At F PR in 2006, water levels increased in spring and early summer to a peak in July due to water pumping, and then decreased during late summer after pumping stopped. Water level increases at NPN and NPS in August and September were due to pumping in preparation for the fall waterfowl hunting seasons. Water elevations recorded in 2005 and 2006 at the Essexville station indicated increasing water levels in spring and early summer to peaks in July or August, and then decreasing water levels thereafter (Figure 6). In 2007, water levels were generally stable from about early May through mid July and then decreased in the late summer. Water level patterns observed at the Essexville station during the study were generally consistent with long-term averages (Figure 4). 23 Figure 5. Water level fluctuations by week of year during late spring and summer at diked sites on Saginaw Bay, Michigan 2005-2007: Fish Point Refuge (FPR), Nayanquing Point (NPE, NPN, and NPS), and Wigwam Bay (WBD). The y-axis references selected heights on staff gages, rather than true water elevation (i.e. meters above sea level) of the study sites. 24 I 2005 __----——~- .. g —2006 I *Tm’l _ . . . . . . . . 1 .. . . . I . . _ . . n _ . l ., _ . . T . _ . . . . . ., . _ .- . . _. u . w M _ . . T . . f . . . W . . . . 1 . . . .. r _. . . . _ . . . m .- .. . . n 3 _ _ . _ . . .. .- f . . . _ . l . . . .e a q a a 095755A321 1 0 O 0 0 0 O 0 0 0 :5 5.55m. 25$ :58 1.8 .533. 8-32 . 3-9:. 5.9.... 53.. «N33 23.. 9-3.. .03.. 8.55 M. 2-5... 3.3.. 933 .. 8.5.2 8.5.2 wins. 346.2 3.22 Fish Point Refuge 0. 1 "'—n-NPE .- 2006 ‘ —-B—NPE 9. 0 2007 "-o-NPN 2006 -e— NPN 8. 7. 6. 5. 4. 3. 000000 E: :o_umo>o_m 2007 . - t -NPS 2. 0 4|. O 2006 + NPS 2007 0 View - 55$ «5% .533. 5-33. 3-33. 5.33.. 5-3.. 403.. 2-3.. 2-3.. @33 _ 8-2:... . 23:2. . NYE-:- mean auras. 84.92 3.5.2 9.5.2 Ems. Nayanquing Point (East, North, and South Units) I 2005 - I - 2006 —B- 2007 . . _ . . .3 w . u w . . _ . _ ...- . M M . .. . . w . _ _ . ~ . . . . a , _ M — v.- . .n . . . . _ . m . . r . . . .I . . . . .. n u 001] ._ . . _ . I. _ _ M I . M . . I . . _ . . _ . M .H _l . . ... . m .II » M . ... T .. . . . k. m _. L . m _.o . . .3 . .r . _ — . . . .I. _ . .o.. . . . . .n . .- . m m.. m n . . .o. ,. M W W... . .l . . .... . . _ . “.- A . ._ L It. _ .. . .. ._._. . .4 m . I. . M . . 1 . V . .. . V ._ .ll . _ v _ . . . o , . . .. u . W . .. I. . Tl . .um . . _. _ .... . m . . .1. . . .. . _. . .. . . .i . ‘ ~ . .. .u. .r _ m we. . . .. .u . _. o T. . . o . . w . m . . . m .. .. u .i . . . . . u a . . w m 0 . . . . . MT _ — I . . . . , n . 1 . .7 . . . .. . u _. .. T . . _ .. . — .l» h N _. A . a 0 4321098765 I I I I I I I I C 0 1111100000 :5 5.355 25% :.&m vamm .532 5-33. 3-33. 592 5-3.- » 3m.33 5-3.. 9-3.. 22. 8.3.- 37:...- «23.. m5... mam-.2). No.5... m 33.2 3.5.). 7.6.2 Wigwam Bay 25 1764 1..--.. .---.....-. --..-----.____..-.-----. .1 _ --..-..-.--_---.. ..........------..-. A 176.3 V 176.2 1 76-1 1760 ........ 2005 175.9 — ..-._-_- “ - . -2oos 175.8 " ' -.--------- 175.7 -- _ -- -1------.._.-_----_-_-.- +2007 175.6 "””’ T 1 I r r f "1""*"-""'“T"""'7'"'"7""“7'7'. 3. .v : ,r y , g , Elevation Figure 6. Water level fluctuations by week of year during late spring and summer at open Essexville NOAA water gage on Saginaw Bay, Michigan 2005-2007. The y—axis represents true water elevation in meters above sea level for Saginaw Bay, Lake Huron. Water Chemistry Diked wetlands tended to have lower dissolved oxygen (DO) levels and pH compared to undiked wetlands, regardless of sample period and study area (Table 2). Mean water temperatures were similar between diked and undiked wetlands. Diked St. Clair Flats sites consistently had greater alkalinity compared to the undiked sites; however, alkalinity varied within and between wetland types at Saginaw Bay. Turbidity was lower in diked compared to undiked wetlands at St. Clair Flats, but varied by site at Saginaw Bay wetlands with overall means being similar. Within each of the study areas, nitrate-N levels in diked wetlands tended to be lower than undiked wetlands (Table 3). Average ammonium-N levels were slightly greater in undiked than diked sites at the St. Clair Flats, but appeared similar between the wetland types on Saginaw Bay. Mean SRP levels were low and similar among wetland types, study areas, and sites (Table 3). 26 a: 3“; A339... 3: 8.3.3.“ a: «63% a: $386 33 3 2a»: 8: 2.933 a: 232 8: Goa-.0... ram 2.2 3 o. :3 5 23mm 5 8.38s 5 moan”: 3 $3.35 33 av $3: 6 cacao: 9.: 0.93.3 a: 3333 ram 52 a: 3:”? a: Zmfl a: 2.3.: a: moan-cm a: coon-“Rs 33 as: 9.3 v. :3- cé $5 $8 8.93? 33 goat-am as 3.3%.: 33 6: :«me 88 $3M; EV nine Ev 3.53 :5 ME 68-5 - 0% 68 0.3.5 33 35 Ga 8.03% 83 23.3 68 «teams.» 83 88 SH; 88 $3 :8 8.3qu :8 333 :8 2.3%.? ram :35 53.13 a: Num- 9.: 8.393 a: 35.8. 2: 33%.” 33 Ag 0. :3 Es «£2 Ag 8.335 as; 0.92.2 A3 3.3.9; 35 a): a: 33m a: E 2 Am: NEH-”m3 a: 25.3 a: who-$3 33 as w. 33 39 32.- 69 gigs as 3...;- 80 3.3.84. ram m5 =25 .. mom as :5: as Eng- 5 2.93am GS boa-«mm as 3.3%.... 23 at v.32 a: ”flaw :5 3.3:; at 33.2 a: 8.35m ram :35 AN: 212 a: :38 a: amass a: «95%. a: fiance-V 33 :3 ”.36.... as Sfimm as 8.3.3.5 as $3.8 :9 moo-nova ram <23 9: :53 a: $5 Q: 8.93: a: 33.8 a: «2&3 33 3; 33a 3; $8. Ag 8.3m? a3 23.2 E; 333 ram SE .35 .. “am 33 has; 3E5 £532 mm CL name 553 SEE BE.- 25 2233 now-O5 @038me 98 moxm dude. $0855.89 E $ 05m 038.8 can muotom Conn—Baum 8m— : b3. REV 8m— 65 93. RE I $2 383 330 BE wagging can 8mm Room 5 newEom—z Sam Bacmwam new 33m .520 um 8 680358 mama—Sam mega 35305 $8085qu hammfiono .863 8m cotoa 98 .83 .33 .396 dab 28:25 ‘3 mm “m mgoz .N 2an 27 Sag-sac Eta :5 as 8m an _ao§§u.=3 ass 68:22 am $333553 vase-5:0".«DO amazes-5&5 13980 .5 E3055 :5 53.5 .5 5555..- 55 13.5 $5 Saw-Bug? and: 850353 can 23.5 .5 wascfihaZuEz an: 0305 5:8 .5 mesa-53352 amas- 5:2 .5 waged-@3252 3.32 “am .5 mafia-52552 swag .5 55555 35 cum-E .5 E355 mam Boga-.32 am a5 253:3: .35 38o a5 saw-$52 acacia-anma amas- §3u<23 Ema: damn/Em ”823323“ 0% €39 3% 333 :3 33 9533.33.” 33 33:3 @a $335 23 88 m3; am: 35 95: 8.38.5 Gm: 3.30.2 2m: 3.3m? ram 852: am: 933 an: 352 am: 8.30: am: 33”.: a: c 2.350 25 at «.32 am: 3c: av: 8385 3: 3.35.2 $3 3.3%.... ram 855 Bob 68 53: $8 39; G8 8.38.5 68 5.318 68 2.35.5 83 35 3.: E; 2.32.» no 930: as $38.5 ram 356 9.3 N. 33 $3 39; $6 8.3-3 $3 @335 cé 338.2 33 63 :5: as 8.3:.» an ”.325 as $.32; ram .55 as 533 as 39: as 2.325 as 93»: a as $3555 33 € 035% a: 2.3%.” a: a. 5.2 a: 338.: ram <20 a: 2.55 a: 33. a: 8.3%.” a: 235.8 a: 2.323 23 a: 3m: €6.38.» 55 N. 35.: § 8. 38.2 335 55 8525 I 9% Ge 35 a $3 35 38 8.385 Ge «.30. a 23 3.933 23 at 32. $8 2.3»: Q: non-#2 a8 3353 bam- :35 $3 «.33 $3 3:: $8 8325 3 5.35.2 3.3 3.303 23 as 3? 60 £335 60 5.33: 69 2.325 ram 55% av 33 € 2335 6 2.3st 5V 5.30% € 8.33m 23 3 232: E 2.35.» :5 235.8 :5 3. 336 ram 52 .380 .855 u 0% A35 5655 :95 5-532 :5 6°C 5&5. “as? 33.5 8:5 25 23:03 5%th 338.55 ES .85 62¢. .PEOU .N 2an 28 Table 3. Means (mg/L) :1: SE by wetland type, study area, and site for nitrate-N, ammonium-N, and soluble reactive phosphorous (SRP) in water samples collected at St. Clair Flats and Saginaw Bay, Michigan in late summer 2007. The number of samples for each parameter is listed in parentheses. Study Area, Site, and Nitrate-N Ammonium-N SRP Wetland Type (mg/L) (mg/L) (mg/L) SCF - Diked EMA 001350004 (6) 0033i0003 (9) 0OOlzt<00001 (8) WMA 0012i0003 (9) 0026:0004 (9) 0.003zt00015 (10) Overall 0012i0002 (15) 002950002 (18) 0002i00008 (18) SCF — Undiked DIS 0O99i0034 (9) 0.04Si0.007 (9) 0002100005 (9) LMU 012250036 (9) 0047i0010 (8) 0001i<00001 (9) Overall 0.11 110024 (18) 0.046:l:0006 (17) 0001500003 (18) SAG — Diked FPR 006250024 (6) 001150005 (6) 0003500004 (7) FPA 002810008 (9) 003530003 (9) 0.002:t00004 (9) NPE 0103i0092 (5) 0.041:l:0004 (5) 0.00Bd:0OOl2 (4) NPN 0021i0007 (2) 0042i0006 (2) 0002500005 (3) WBD 001810004 (9) 0.039zt0005 (8) 0.0022t0.0003 (8) Overall 0043i0.015 (31) 0.039i0.002 (30) 0002500002 (31) SAG — Undiked PIR 0.106i0.051 (6) 0038:0007 (6) 0.002:l:0.0005 (6) QUA 012750051 (6) 0043:t0007 (6) 0002500003 (5) WIL 002850006 (8) 004310004 (9) 0003i00010 (8) Overall 008110023 (20) 0042:0003 (21) 0002500004 (19) Total - Diked 0033220011 (46) 0035i0002 (48) 0002500003 (49) Total — Undiked 0095i0016 (38) 004410003 (3 8) 0.002zt00003 (37) 29 DISCUSSION Water levels of the undiked wetlands, as indicated by NOAA monitoring stations, were below long-term mean elevations for both Lake St. Clair and Saginaw Bay during all three years of the study. Below average water depths of the undiked wetlands likely influenced bird use, so comparisons with the diked wetlands must be viewed within the context of a period of low Great Lakes water levels. Investigations such as my study need to be conducted during the full range of water levels to gain a full understanding of the value of diked and undiked wetlands to birds. The water level changes I observed in Lake St. Clair during the breeding seasons of 2005 and 2007 were not consistent with the pattern observed over the long term, with lake levels declining throughout the spring and summer. This pattern is similar to that documented in the diked wetlands. Water level changes recorded in Lake St. Clair during 2006 were similar to long-term averages, with levels increasing in the spring and early summer to a peak in July and then decreasing in late summer and fall. The seasonal changes I observed in the elevation of Saginaw Bay wetlands were similar to those observed over the long-term. Isolation from the adjacent lakes altered the hydroperiod of diked wetlands when compared to undiked wetlands. Highest water levels in diked wetlands occurred in the spring, while peak water levels in undiked systems are usually in mid to late summer. Water level changes in diked wetlands were also more pronounced than the gradual changes that occurred in undiked systems. For example, even when both wetland types exhibited drawdowns, water levels tended to drop faster and at a greater depth in the diked wetlands. Conversely, water 30 levels often increased substantially (e. g., ~O.25-0.4O m) in diked wetlands over a short time (e.g., ~2-4 weeks) when pumping occurred. Both long- and short-term (i.e., seiche) Great Lakes water level fluctuations can influence biogeochemical cycling in coastal marshes (Burton 1985). Although modifications to the hydrology of the diked wetlands undoubtedly altered biogeochemical cycling, more intensive water chemistry sampling than conducted in this study is needed to understand these changes. My results indicated higher levels of nitrate-N in undiked compared to diked wetlands. Higher nitrate-N levels in undiked compared to isolated wetlands would be expected as increased DO levels and sediment exposure of undiked wetlands could increase organic matter decomposition and nitrification (Burton 1985). Runoff from agricultural lands containing excess fertilizer could have contributed to nitrate-N levels in undiked wetlands. Anaerobic conditions created by higher water levels in the diked wetlands probably lead to increased denitrification, thus reducing nitrate-N levels. Ammonium-N appeared to be slightly higher in undiked compared to diked wetlands at St. Clair Flats, but was similar between diked and undiked wetlands on Saginaw Bay. Robb (1989) found no significant difference in nitrate-N and ammonia-N levels in water of diked and undiked wetlands. I found similar levels of SRP in diked and undiked wetlands, while Robb (1989) recorded higher levels of orthophosphate in diked wetlands and higher total phosphate in undiked sites during comparisons of diked and undiked coastal wetlands on Lake Erie. Herrick and Wolf (2005) observed higher total N, available P, and available K in the soils of diked compared to undiked wetlands. My limited testing for nitrate-N and ammonium-N in water samples is not directly comparable to the study by Herrick and Wolf (2005), due to differing methods and timing 31 of sample collection. The nitrate-N levels I observed were lower than averages recorded by Uzarski et al. (2005) in cattail (Typha spp.) and aquatic bed zones across a range of Great Lakes coastal wetlands, but similar to values they observed in bulrush (Schoenoplectus spp.) marshes. My nitrate values were also lower than those of Robb (1989) in diked and undiked Lake Erie wetlands. I observed ammonium-N levels of about 0.03-0.04 mg/L, which are similar to values recorded by Robb (1989) and slightly lower than the mean observed by Uzarski et al. (2005) in aquatic bed wetlands. The SRP levels I observed were much lower than those reported by others in Great Lakes coastal wetlands (Robb 1989, Uzarski et al. 2005), which may be due to differences in sampling methodologies and timing of collections, or the small sample size used in my study. I found that the diked sites tended to be more acidic and less turbid compared to undiked wetlands, which is consistent with other studies (Robb 1989, Herrick and Wolf 2005). The mean pH readings that I observed in undiked wetlands that were similar to the undiked wetlands sampled by Uzarski et al. (2005), while the average pH values I recorded in diked wetlands tended to be lower. My mean turbidity values tended to be lower than averages reported by Uzarski et al. (2005), regardless of study area or wetland type. More study is needed to better understand the hydrology, water chemistry, and nutrient cycling of diked compared to undiked wetlands. Intensive water chemistry testing across the range of diked and undiked coastal wetlands and through a normal range of water level fluctuations is needed to learn how hydrological isolation affects wetland functioning. Wide variation in the functioning of diked wetlands is likely, due to differing hydrology of the sites. For example, diked wetlands with only passive (i.e., no 32 pumps) or Opportunistic (i.e., can only pump with higher Great Lakes levels) tend to have shallower water depths and more pronounced summer drawdowns compared to sites with active water pumping regimes. A better understanding of the biogeochemical cycling of diked wetlands may permit the development of water level management guidelines that optimize wetland functioning, while maintaining the capability to manage for wildlife. 33 LITERATURE CITED American Public Health Association. 1992. Standard methods for the evaluation of water and wastewater. Twentieth edition. American Public Health Association, Washington, DC, USA. Bellrose, F. C. 1980. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg, Pennsylvania, USA. Bookhout, T. A., K. E. Bednarik, and R. W. Kroll. 1989. The Great Lakes marshes. Pages 131-156 in L.M. Smith, R.L. Pederson, and RM. Kaminski, editors. Habitat management for migrating and wintering waterfowl in North America. Texas Tech University Press, Lubbock, USA. Brown, S., C. Hickey, and B. Harrington, editors. 2000. The US. shorebird conservation plan. Manomet Center for Conservation Sciences, Manomet, Massachusetts, USA. Burton, T. M. 1985. The effects of water level fluctuations on Great Lakes coastal marshes. Pages 3-13 in H. H. Prince and F. M. D’Itri, editors. Coastal wetlands, Proceedings of the first Great Lakes Coastal Wetlands Colloquium. Lewis Publishers, Chelsea, Michigan, USA. Burton, T. M., C. A. Stricker, and D. G. Uzarski. 2002. Effects of plant community composition and exposure to wave action on invertebrate habitat use of Lake Huron coastal wetlands. Lakes and Reservoirs: Research and Management 7:25 5- 269. Fredrickson, L. H., and T. S. Taylor. 1982. Management of seasonally flooded impoundments for wildlife. US. Fish and Wildlife Service Resource Publication 148, Washington, DC, USA. Galloway, M., L. Bouvier, S. Meyer, J. Ingram, S. Doka, G. Grabas, K. Holmes, and N. Mandrak. 2006. Evaluation of current wetland dyking effects on coastal wetlands and biota. Pages 187-229 in L. Mortsch, J. Ingram, A. Hebb, and S. Doka, editors. Great Lakes coastal wetland communities: vulnerability to climate change and response to adaptation strategies. Environment Canada and the Department of Fisheries and Oceans, Toronto, Ontario, Canada. Gottgens, J. F., B. P. Swartz, R. W. Kroll, and M. Eboch. 1998. Long-term GIS-based records of habitat changes in a Lake Erie coastal marsh. Wetlands Ecology and Management 6:5-17. 34 Great Lakes Basin Commission. 1975. Great Lakes Basin framework study. US. Department of Interior, Bureau of Sport Fisheries and Wildlife, Ann Arbor, Michigan, USA. Herrick, B. M., M. D. Morgan, and A. T. Wolf. 2007. Seed banks in diked and undiked Great Lakes coastal wetlands. American Midland Naturalist 158: 191-205. Herrick, B. M., and A. T. Wolf. '2005. Invasive plant species in diked vs. undiked Great Lakes wetlands. Journal of Great Lakes Research 31:277-287. Jude, D. J ., and J. Pappas. 1992. Fish utilization of Great Lakes coastal wetlands. Journal of Great Lakes Research 18:651-672. Kadlec, J. A. 1962. Effects of a drawdown on a waterfowl impoundment. Journal of Wildlife Management 43 2267-281 . Kaminski, R. M., and H. H. Prince. 1981a. Dabbling duck activity and foraging responses to aquatic macroinvertebrates. The Auk 98:115-126. Kaminski, R. M., and H. H. Prince. 1981b. Dabbling duck and aquatic macroinvertebrate responses to manipulated wetland habitat. Journal of Wildlife Management 45:1-15. Keddy, P. A., and A. A. Reznicek. 1986. Great Lakes vegetation dynamics: the role of fluctuating water levels and buried seeds. Journal of Great Lakes Research 12:25- 36. Keough, J. R., T. A. Thompson, G. R. Guntenspergen, and D. A. Wilcox. 1999. Hydrogeomorphic factors and ecosystem responses in coastal wetlands of the Great Lakes. Wetlands 19:821-834. Krieger, K. A., D. M. Klarer, R. T. Heath, and C. E. Herdendorf. 1992. A call for research on Great Lakes coastal wetlands. Journal of Great Lakes Research 18:525-528. McLaughlin, D. B., and H. J. Harris. 1990. Aquatic insect emergence in two Great Lakes marshes. Wetlands Ecology and Management 1:111-121. Murkin, H. R., R. M. Kaminski, and R. D. Titrnan. 1982. Responses by dabbling ducks and aquatic invertebrates to an experimentally manipulated cattail marsh. Canadian Journal of Zoology 60:2324-2332. Prince, H. H., and C. S. Flegel. 1995. Breeding avifauna of Lake Huron. Pages 247-272 in M. Munawar, T. Edsall, and J. Leach, editors. The Lake Huron ecosystem: ecology, fisheries, and management. SPB Academic Publishing, Amsterdam, The Netherlands. 35 Prince, H. H., P. I. Padding, and R. W. Knapton. 1992. Waterfowl use of the Laurentian Great Lakes. Journal of Great Lakes Research 18:673-699. Provence, C. D. 2008. Effects of diking and plant zonation on invertebrate communities of Lake St. Clair coastal marshes. Thesis, Michigan State University, East Lansing, USA. Robb, D. 1989. Diked and undiked freshwater coastal marshes of western Lake Erie. Thesis, The Ohio State University, Columbus, USA. Thiet, R. K. 2002. Diversity comparisons between diked and undiked coastal freshwater marshes on Lake Erie during a high-water year. Journal of Great Lakes Research 28:285-298. Uzarski, D. G., T. M. Burton, M. J. Cooper, J. W. Ingram, and S. T. A. Timmerrnans. 2005. Fish habitat use within and across wetland classes in coastal wetlands of the five Great Lakes: development of a fish-based index of biotic integrity. Journal of Great Lakes Research 31(Supplement 1):171-187. Weller, M. W., and L. H. Fredrickson. 1974. Avian ecology of a managed glacial marsh. Living Bird 12:269-291. Weller, M. W., and C. S. Spatcher. 1965. Role of habitat in the distribution and abundance of marsh birds. Department of Zoology and Entomology Special Report 43, Agricultural and Home Economics Experiment Station, Iowa State University, Ames, USA. Whitt, M. B. 1996. Avian breeding use of coastal wetlands on the Saginaw Bay of Lake Huron. Thesis, Michigan State University, East Lansing, USA. Wilcox, D. A. 1993. Effects of water-level regulation on wetlands of the Great Lakes. Great Lakes Wetlands 4:1-2, 1 1. Wilcox, D. A. 1995. The role of wetlands as nearshore habitat in Lake Huron. Pages 223-245 in M. Munawar, T. Edsall, and J. Leach, editors. The Lake Huron ecosystem: ecology, fisheries, and management. SPB Academic Publishing, Amsterdam, The Netherlands. Wilcox, D. A., J. A. Meeker, and J. Elias. 1993. Impacts of water-level regulation on the wetlands of the Great Lakes. Phase 2 Report to Working Committee 2, International Joint Committee Water-levels Reference Study, Ottawa, Ontario, Canada, and Washington, DC, USA. Wilcox, D. A., and T. H. Whillans. 1999. Techniques for restoration of disturbed coastal wetlands of the Great Lakes. Wetlands 19:83 5-857. 36 CHAPTER 2 BREEDING BIRD USE OF DIKED AND UNDIKED COASTAL WETLANDS IN MICHIGAN INTRODUCTION Great Lakes coastal wetlands provide vital breeding, migration, and wintering habitat for an array of birds. These wetlands are some of the region’s largest remaining emergent marshes and provide vital nesting habitat to wetland birds, including rare and declining species such as American Bittem (Botaurus lentiginosus), Least Bittem (Ixobrychus exilis), Common Moorhen (Gallinula chloropus), King Rail (Rallus elegans), Black Tern (Chlidonias niger), and F orster’s Tern (Sternaforsteri). Prince and Flegel (1995) summarized breeding bird atlas data from Michigan and Ontario. Eighty bird species used coastal wetlands of Lake Huron as breeding habitat (Prince and F legel 1995) Impoundments control structures have long been used by wildlife managers to enhance wetlands for wildlife (Kadlec 1962), especially breeding and migrating waterfowl. When the goal is to maximize breeding bird use, wildlife biologists often manage for hemi-marshes. Hemi-marshes are marshes with approximately equal proportions of emergent vegetation and open water produced by natural water level fluctuations and mammal herbivory. Several authors have found that hemi-marshes typically attract greater densities and diversities of wetland birds compared to marshes with more or less emergent vegetation (Weller and Spatcher 1965, Weller and 37 Fredrickson 1974, Kaminski and Prince 1981a, b, Murkin et al. 1982). Historically, Great Lakes coastal wetlands moved landward and lakeward with the rise and fall of the Great Lakes. Between the 19503 and 1970s, many Great Lakes coastal marshes were isolated from these normal water level fluctuations through dike construction. These projects were initiated primarily to maintain elevated water depths and enhance wildlife use during periods of historic low water levels. Shoreline armoring, wetland diking and tiling to drain wetlands for agricultural use, and other land-use changes now prevent the landward movement of coastal wetlands in much of the Great Lakes (Prince et al. 1992, Gottgens et al. 1998). The potential problems associated with isolating coastal wetlands from the Great Lakes include impaired or eliminated flood conveyance and storage, sediment control, and water quality improvement functions, altered nutrient flow, reduced or degraded habitat for shorebirds, rare species, fish, and invertebrates, and increased impacts from trapped Carp (Cyprinus carpio) (Jude and Pappas 1992, Wilcox 1995, Wilcox and Whillans 1999). By separating coastal wetlands from the fluctuations of the Great Lakes, dike construction often stabilizes water levels. Stable water levels typically compress wetland vegetation zones and encourage dominance by shrubs and highly competitive species, such as willow (Salix spp.), alder (Alnus spp.), cattail (Typha spp.), reed canary grass (Phalaris arundinacea), and purple loosestrife (Lythrum salicaria). Irregular water levels may result in higher levels of diversity both within and among habitats (Keddy and Reznicek 1986, Wilcox 1993, Wilcox et al. 1993, Keough et al. 1999). Comparisons of plant communities in diked and undiked Great Lakes coastal wetlands have yielded varied results. Herrick and Wolf (2005) documented increased 38 amounts of invasive species in standing vegetation and seed banks of diked compared to undiked wetlands in Saginaw Bay, Michigan and Green Bay, Wisconsin, but noted that current conditions in undiked wetlands appear to favor an invasive haplotype of common reed (Phragmites australis). Conversely, Galloway et al. (2006) found greater species richness and percent cover of native species and lower species richness and percent cover of invasive species in diked compared to undiked coastal wetlands. Herrick et al. (2007) found more seeds from a greater number of species in the soils of diked compared to undiked wetlands and stated that diked wetlands may serve as “traps” for plant seeds. In comparisons between vegetation in diked and undiked Lake Erie coastal wetlands during a high water year, Thiet (2002) found greater wetland plant diversity in diked wetlands compared to a nearby undiked site. An actively managed diked marsh in southwest Lake Erie maintained emergent vegetation, patchiness, and edge habitat similar to historic conditions during periods of high Great Lakes water levels, while the same measures declined in marshes connected to Lake Erie (Gottgens et al. 1998). Research conducted by several authors on animal use of Great Lakes coastal wetlands provides insights into the possible effects of diking on animal communities. McLaughlin and Harris (1990) compared aquatic insect emergence in one diked and one undiked wetland on Green Bay, and recorded more insect taxa and greater total insect biomass emergence from the diked wetland. Burton et al. (2002) noted that both plant- community composition and exposure to wave action were important determinants of invertebrate diversity and biomass in Great Lakes marshes. Invertebrates were distributed along gradients of decreased mixing of pelagic water and increased sediment organic matter from outer to inner marsh and between littoral and adjacent inland 39 marshes. Some invertebrates were more common on one end of these gradients, but most species were generalists found across all habitat types (Burton et al. 2002). Whitt’s (1996) study of avian breeding use of Saginaw Bay coastal wetlands included study sites that were both open to and inland from Lake Huron. Although species richness was similar between coastal and inland cattail marshes, bird densities in marshes located far offshore were lower than most other sites. Whitt (1996) suggested this difference may be due to the effects of storm surges during the breeding season that can destroy nests, and stated that further study is needed to compare avian use of protected marshes with those exposed to storm surges. Galloway et al. (2006) conducted a one-year study of breeding bird use of diked and undiked Great Lakes coastal wetlands along Lakes Ontario, Erie, and St. Clair. In pooled comparisons of diked and undiked sites, they observed greater abundance and species richness for several groups of birds in diked wetlands, but indicated that long-term research is needed to account for long-term variation in bird and vegetation communities associated with Great Lakes water level cycles and management activities. Ecological studies of the effects of coastal wetland isolation from natural, highly variable water level fluctuations are needed so that informed decisions can be made about the management and restoration of Great Lakes marshes. The goal of this project was to compare breeding bird use, vegetation composition and structure, and physical and chemical attributes of several diked and undiked wetlands in Michigan to gain insights into the effects of wetland diking on avian communities. I tested the hypothesis that coastal impoundments with managed water levels provide enhanced conditions for breeding wetland birds compared to undiked wetlands. This research is one of many 40 comparisons needed over the long-term to better understand how diked and undiked wetlands function during the full cycle of Great Lakes water levels. 41 METHODS To assess bird communities of diked and undiked coastal wetlands, I compared several indices of breeding bird use at sites in two study areas: St. Clair F lats, Lake St. Clair and Saginaw Bay, Lake Huron (Figure 1 - detailed descriptions found in Chapter 1). These study areas are two of Michigan’s largest remaining coastal wetland complexes (Bookhout et al. 1989, Krieger et al. 1992). Sixteen study sites were sampled during breeding bird surveys 2005-2007: four sites at St. Clair Flats (two diked, two open), and 12 sites at Saginaw Bay (six diked, five undiked). Undiked study sites at St. Clair Flats were located at two general areas: 1) Dickinson Island and nearby marshes on Fisher and Goose Bays, and 2) Little and Big Muscamoot Bays. All of the St. Clair Flats study sites were located in St. Clair Flats State Wildlife Area (SWA). I examined the following diked wetlands on Saginaw Bay: two sites at Fish Point SWA (East Refuge Unit [FPR], Austin Road [FPA]), four sites at Nayanquing Point SWA (East Marsh [NPE], North Marsh [NPN], South Refuge Unit [NPS], and Triangle Unit [NPT]), and one site at Wigwam Bay SWA (WIG). Undiked sites on Saginaw Bay were located at Wildfowl Bay SWA (WIL), Fish Point SWA near Berger Road (FPB), Quanicassee SWA west of the mouth of the Quanicassee River (QUA), Wigwam Bay SWA (PIR) north and south of the Pine River, and at Wigwam Bay SWA (WBO) in wetlands east of the diked unit. Two survey techniques were used to investigate breeding bird use of coastal wetlands: 1) point counts to assess bird use of emergent vegetation and 2) timed-area surveys to evaluate bird use of the open water/aquatic bed zone. Randomly selected points were surveyed three times at 13 sites (six diked, seven undiked) within the two 42 study areas (Figure 7). Open water areas were randomly selected for timed-area surveys during four periods from 14 sites (nine diked, five undiked) at the two study areas (Figure 8). STUDY AREA: ST. CLAIR FLATS A 8 Early t Mid Late C r D I STUDY AREA: SAGINAW BAY A‘ B' 0‘ D E r F‘ G r H I Early 1 Mid Late Figure 7. Illustration of study design used for breeding bird point counts conducted in Great Lakes coastal wetlands in Michigan (St. Clair Flats and Saginaw Bay), 2005- 2007. Independent study sites (lettered polygons) were sampled within each study area, with approximately half of the points occurring in each of two wetland types (diked — shaded, undiked — not shaded). Points (black dots) were situated randomly within each study site, and three surveys (early, mid, and late season) were conducted at each point. 43 STUDY AREA: ST. CLAIR FLATS Figure 8. Illustration of study design used for timed-area surveys for breeding birds conducted in Great Lakes coastal wetlands in Michigan (St. Clair Flats and Saginaw Bay), 2005-2007. Independent study sites (lettered polygons) were sampled within each study area, with approximately half of the open water areas occurring in each of two wetland types (diked — shaded, undiked - not shaded). Open water areas (polygons) were randomly selected (shaded polygons) within each study site. During each of four survey periods, a new set of open water areas were randomly selected. 44 Point Counts I conducted point counts in emergent marshes of impounded and undiked wetlands using methods similar to the Standardized North American Marsh Bird Monitoring Protocols (Conway 2005). Potential survey points were identified using ArcView 3.2, aerial photographs, and 200 by 200 m grids overlaying the study sites. Because Great Lakes water levels were below the long-term average every year of the study, I positioned potential survey points within 400 m of the shoreline or other open water areas. I assumed that emergent wetland located closer to open water/aquatic bed wetland was more likely to be inundated and occupied by marsh birds. Potential survey points had greater than or equal to 50% emergent vegetation within 200 m. Non- emergent cover consisted of open water/aquatic bed, scrub-shrub, or forested wetland. Potential survey points were not used if greater than 10% of the total area within 200 m of the point consisted of roads, dikes, buildings, upland, or wetland of a different type (e. g., undiked wetland in the case of diked points). Conway (2005) suggests surveying all points on a 400 by 400 m grid covering a study site; however, that was not feasible given the size and accessibility of our study areas. I surveyed randomly selected points that were at least 400 m apart and had standing water or saturated soils three times during the breeding season (early to mid May, mid May to early June, and early to late June); however, some points were only surveyed once or twice due to weather or other constraints. Each survey was separated by at least seven days. Surveys at St. Clair Flats were started approximately one week earlier than at Saginaw Bay. I counted all birds seen or heard during 10-min surveys conducted between 0.5 hour before sunrise and 10:00 AM. During the second half of the point count, I broadcasted calls of several 45 secretive marsh birds in the following order, as recommended by Conway (2005): Least Bittern (Ixobrychus exilis), Sora (Porzana carolina), Virginia Rail (Rallus limicola), King Rail (Rallus elegans), and American Bittem (Botaurus Ientiginosus). I noted each minute of the survey that a waterbird was detected. The approximate distance to each marsh bird (e. g., grebes, bitterns, rails, coots, moorhens) was estimated using ocular/aural estimation and a laser rangefinder. All other birds (e. g., songbirds, waterfowl, shorebirds, terns, gulls) were noted as being in one of five distance categories: 318 m, >18 — 50 m, >50 — 100 m, >100 — 200 m, and >200 m. Timed-area Surveys I evaluated breeding bird use of the open water/aquatic bed zone using a timed- area approach. Potential open water/aquatic bed survey areas were identified using aerial photographs and on-site visits. Surveys were conducted during four periods, late-May, mid-June, mid-July, and early-August, separated by two to three weeks. I only conducted surveys at St. Clair Flats sites in 2005, but surveyed sites at both St. Clair Flats and Saginaw Bay in 2006 and 2007. Surveys were done during all four periods at both study areas in 2006, but I only conducted surveys during the first three periods at Saginaw Bay sites in 2007. I randomly selected (with replacement) survey sites from the pool of potential sites for each round of surveys. Surveys were conducted in the morning between 0.5 hour before sunrise and four hours after sunrise. I waited 15 min after arrival before starting each survey, and surveyed each area for 30 min from a stationary boat, canoe, or vehicle. I selected survey stations that afforded the best view of the area, caused the least disturbance, and offered the most concealment. I recorded the location 46 of the survey station using GPS and estimated the size of the survey area using field maps drawn with the aid of a laser rangefinder, compass, and aerial photographs. All waterfowl, waterbirds, and shorebirds seen or heard within the survey area were counted. Birds flushed from the area upon arrival or seen only during the 15 min silent period were also counted. Flying waterbirds using the area for foraging (e. g., terns) were counted. I recorded the time of each observation and noted if I thought a bird or group of birds was observed previously during the survey, but excluded suspected repeat observations from analyses. The species, number of young, and estimated age class (according to Gollop and Marshall 1954, as cited in Bellrose 1980) were recorded for waterfowl broods. Vegetation and Physical Variable Sampling To characterize the habitat present at the study sites, I collected vegetation data at three randomly'selected 0.25 m2 quadrats surrounding point count stations. Quadrats were situated randomly between one and 18 m along three compass bearings (120°, 240°, and 360°). At each quadrat I estimated percent cover of dominant vegetation types, measured the water depth, depth of organic sediments, maximum height of standing live or dead vegetation, and visual obstruction (according to Robel et al. 1970), and counted the number of live and dead shrub and tree stems >2 m tall within 2.5 m of the quadrat center (Riffle et al. 2001). I estimated the depth of organic sediments by pushing a 1.2-m wooden stick (2-cm diameter, graduated in centimeters) to the bottom of the organic layer and measuring the depth of the sediments minus the water depth. Both percent cover and stem density was estimated for cattail (Typha spp.), bulrush (Schoenoplectus spp.), and common reed (Phragmites australis), which were the three dominant plant taxa observed. 47 I categorized the vegetation into the following structural groups: persistent deep-water emergents, persistent shallow-water emergents, non-persistent deep-water emergents, non-persistent shallow-water emergents, floating-leaved and free-floating vegetation (e. g., Nuphar spp., Lemna spp.), and submersed aquatic species (e. g., Potamogeton spp., Chara spp.). Cowardin et al. (1979) defined persistent emergent species as those that normally remain standing at least until the next growing season, such as cattail, bulrushes, and sedges (Carex spp.), and non-persistent emergents as those species that usually fall to the surface or below the water at the end of the growing season. Persistent deep-water emergents consisted of those species with rhizomes that can survive permanent or semipermanent inundation, such as cattail and bulrush. Species that usually grow in saturated soil or very shallow water, including sedges, rushes (Juncus, spp.), and grasses, were placed in the persistent shallow-water category. Although common reed can survive inundation, I considered it a persistent shallow-water emergent species because it often establishes in moist soils or shallow water, tends to occur near the wetland-upland interface, and its growth and survival is inhibited by long-term flooding with deep water (Roman et al. 1984, Tucker 1990, Marks et al. 1994). Species such as arrowhead (Sagittaria spp.), pickerelweed (Pontedaria cordata), and wild rice (Zizania spp.) were included in the non-persistent deep-water emergent category. Non-persistent shallow-water emergents consisted of species such as spikerushes (Eleocharis spp.), smartweeds (Polygonum spp.), and beggars tick (Bidens spp.). I estimated percent areal coverage for each vegetation category present within a quadrat. 48 Analysis Point Counts: I categorized bird species as wetland dependent, wetland associated, and non-wetland species (Crowley et al. 1996, Brown and Smith 1998). A list of bird species assigned to each category, as well as common and scientific names, is provided in Appendix A (Table A-1 ). Nomenclature follows the American Ornithologists’ Union (AOU) Check-list of North American Birds (AOU 1998) and subsequent supplements. I compared densities (birds per ha) of all birds, wetland dependent species, wetland associated species, non-wetland species, and individual species of management concern between diked and undiked wetlands using a 50-m boundary, which was the distance that appeared to be the best compromise between maximizing detection rates and minimizing the effects of decreasing density with increased distance for most species (see Appendix B, Tables B-1, B-2). However, I used a 100-m boundary when calculating Pied-billed Grebe (Podilymbus podiceps) and American Bittem densities, since density estimates and detection frequencies increased with distance. Observed density and frequency of detection estimates by distance category support my assumption that detection probabilities were similar between the two wetland types (Tables B-l, B-2). I did not conduct analyses (e. g., distance sampling) to adjust density estimates, because population estimates were not an objective of this project, low detection rates precluded such analysis for most of species of management concern, and the use of indices is appropriate in many situations (Johnson 2008). Before analysis, I log (natural) transformed all avian density variables. I used a mixed model (MIXED procedure, SAS Institute 2004) to compare avian variables between impounded and undiked coastal wetlands. Mixed models are an 49 effective means of analyzing multilevel data structures (Wagner et al. 2006). I used a mixed model that consisted of wetland type (diked and undiked), study area (St. Clair Flats and Saginaw Bay), and survey period (early, mid, and late season) as fixed effects, and year, site (e. g., Dickinson Island), and point (i.e., point count station) as random effects. A repeated measures component was used to account for multiple surveys at the same location. Using the above model, I evaluated three commonly used covariance structures: autoregressive order one (AR[1]), compound symmetric (CS), and unstructured (UN) (Littell et al. 1996, Kincaid 2005). I compared models containing the repeated measures component with a standard mixed model with no repeated measures. For each bird density variable, I selected the best-approximating model using Akaike’s Information Criterion (AIC). Of the three structures evaluated, UN covariance appeared to function best for the majority of the variables analyzed based on AIC values (Table B-3). Models containing the UN covariance structure were the best-approximating models in 17 of the 32 variables tested. I used the AR(1) structure in seven of the 32 best-approximating models, while only two of the best-approximating models included the CS structure. Stande mixed models lacking the repeated measures component appeared best of those examined for six of the variables. In all bird variable comparisons except Mallard (Anas platyrhynchos) density, model selection did not alter decisions regarding rejection of null hypotheses that bird densities were similar between diked and undiked wetlands (Table B-3). The models I evaluated produced similar least squares mean estimates for the bird density variables. Some analyses produced G matrices that were not positive definite when one of the covariance parameter estimates equaled zero. In those cases, I set the 50 lower boundary for the covariance parameters with zero estimates at a small value close to zero using the PARMS statement (SAS Institute 2004), which allowed the G matrices to be positive definite. Two values (0.0000001 and 0.00001) were used as the lower bound for covariance parameters with zero estimates in initial models. The new models achieved positive definite G matrices, but did not alter the original decisions regarding null hypotheses or parameter estimates (Table B-4). Values used for the lower bound of covariance parameter estimates (i.e., 0.0000001 or 0.00001) changed p—values and AIC estimates slightly, but not selection of the best-approximating models or decisions regarding null hypotheses. I used three similarity indices (Jaccard, Sorensen, and Morisita) to examine the level of similarity between the bird communities of diked and undiked wetlands. The Jaccard and Sorensen indices are calculated using species presence-absence data, while the Morisita index also incorporates species abundance. I calculated similarity indices between diked and undiked wetlands for all study areas and sites combined. I conducted correspondence analysis (CA) to evaluate possible relationships in breeding bird abundance of the emergent zone observed at diked and undiked study sites. Correspondence analysis is often used in ecological analyses of species data at different sampling sites (Legendre and Legendre 1998). I used the following 11 categories of bird species/groups in the CA: waterfowl, Pied-billed Grebes, bitterns, herons, rails, American Coots (F ulica americana) and Common Moorhens (Gallinula chloropus), gulls and terns, aerial-foraging songbirds (e. g., swallows), wetland-dependent songbirds (e.g., Swamp Sparrow [Melospiza georgiana]), wetland-associated songbirds (e. g., Common Yellowthroat [Geothlypis trichas]), and non-wetland birds (Mourning Dove [Zenaida 51 macroura]). Bird abundance (no. observed per point) was averaged by site and year prior to analysis. I only interpreted the first two dimensions and the solution was not rotated. T imed-area Surveys: I calculated two indices of abundance for breeding birds using open water/aquatic bed zones at diked and undiked sites: areal bird density (birds per ha) and linear bird density (birds per km). Areal bird densities were estimated by dividing the number of birds observed by the total area of open water/aquatic bed wetland surveyed at each site. Linear bird densities were calculated by dividing the total number of birds observed by the total amount of edge (interface of emergent vegetation and open water) surveyed at each area. I analyzed both density indices, because linear density may be an appropriate measure of bird abundance. Wetland birds ofien focus feeding, nesting, and rearing activities at the interface of emergent vegetation and open water, and the boundary used to delineate survey areas along open shorelines was sometimes arbitrary. I examined the relationship between the two density measures using Pearson product- moment correlation (CORR procedure, SAS Institute 2004), and used the chi-square test (FREQ procedure, SAS Institute 2004) to compare the frequencies of species with higher densities in diked and undiked wetlands between the two density calculations. The total area of wetland and length of edge surveyed at each site was estimated using ArcView 3.2 with 2005 color aerial imagery and on-site maps. I compared several density variables between diked and open coastal wetlands, including all birds, wetland- dependent birds, wetland-associated birds, and individual species of management interest. I log (natural) transformed avian density variables prior to analysis. 52 To compare avian densities between diked and undiked wetlands, I used a mixed model with wetland type (diked and undiked), study area (St. Clair Flats and Saginaw Bay), and survey period (1, 2, 3, or 4) as fixed effects, and year and site as random effects. Open water areas surveyed within a given location were considered replicates of that site. When analysis of an avian density variable resulted in a G matrix that was not positive definite, I set the lower bound for covariance parameter(s) that equaled zero using the same procedure described above for the point count data. Setting the lower bound for random variables did not alter parameter estimates. I calculated the same three similarity indices used for the point count data to compare bird species composition of diked and undiked wetlands in the open water/aquatic bed zone. When calculating the Morisita index, I used areal bird density (birds per ha surveyed) as the measure of abundance to account for differences in the size of the survey areas. I conducted correspondence analysis (CA) to evaluate potential relationships in breeding bird abundance observed in the open water zone at diked and undiked study sites. The following 11 categories of bird species/groups were used in the CA: dabbling ducks, diving ducks, geese and swans, Wood Ducks (A ix sponsa), Pied-billed Grebes, bitterns, herons, rails, American Coots and Common Moorhens, shorebirds, and gulls and terns. Areal densities were used as an index of bird abundance to account for differences in the size of survey areas. I averaged densities by site and year prior to analysis. I only interpreted the first two dimensions and the solution was not rotated. 53 Vegetation and Physical Variable Sampling: 1 compared several variables characterizing the vegetation composition and structure of diked and undiked wetlands using data gathered during quadrat sampling at point-count stations. I also compared water depth and estimated depth of organic sediments between the wetland types. Percent variables were arcsine-square root transformed and all other variables (e. g., densities, water depths) were log (natural) transformed. I conducted analyses using a mixed model with wetland type, study area, and survey period (early, mid, and late season) as fixed effects, and year and site as random effects. To evaluate the variation in vegetation structure and composition and physical variables among diked and undiked coastal wetlands, I conducted principal components analysis (PCA) on vegetation and physical variables gathered during quadrat sampling. I used SAS (PRINCOMP procedure, SAS Institute 2004) to conduct the PCA. Habitat data for the three surveys was averaged by point and year prior to analysis. The following variables were excluded from the PCA because they were highly correlated (r 2 0.70) with similar variables: visual obstruction, cattail density, bulrush density, and common reed density. I also did not include percent cover of exposed substrate, non- persistent deep-water emergents, and shrubs/trees, or shrub/tree density, due to low frequencies of occurrence (less than 10% of total quadrats). This resulted in a total of 14 vegetation and physical variables being used for the PCA. Percent variables were arcsine-square root transformed prior to analysis. Correlation coefficients were used to form the cross-products matrix and the ordination axes were not rotated. When evaluating the importance of the principal component loadings, I only considered loadings greater than 0.20 or less than -0.20, which is an approach similar to interpreting 54 correlation coefficient significance at a 0.01 alpha level and sample size between 100 and 200 (Hair et al. 1987, McGarigal et al. 2000). 55 RESULTS Point Counts Average densities of all birds, wetland-associated birds, and non-wetland birds observed during point counts were similar between diked and undiked wetlands, but mean density of wetland-dependent birds was greater in diked compared to undiked wetlands (p=0.0461, Table 4). American Bittem and Least Bittem mean densities were greater in diked than undiked wetlands (p=0.0012 and p=0.0024, respectively). Forster’s Tern (Sternaforsteri) was the only species observed in greater densities in undiked coastal wetlands (p=0.0057). Specific surveys were not conducted for nesting terns, but field observers noted when nesting colonies were seen. Forster’s Tern nests were only found in undiked wetlands dominated by bulrush at St. Clair F lats. Foraging F orster’s Tems were observed in diked wetlands, but no nesting colonies were observed. I provide densities and frequencies of occurrence for all bird species observed during point counts in Table B-5 (Appendix B). Bird species richness was similar between the two wetland types, with 57 species observed in diked wetlands and 53 species documented in undiked wetlands. Forty-four species were common to both types (Table 5). Thirteen species were unique to diked wetlands, with seven species considered wetland dependent, one wetland associated, and five non-wetland species. Nine species were unique to undiked coastal wetlands, of which five were considered wetland-dependent, one wetland-associated, and three as non-wetland species. Species unique to the two wetland types tended to be those that were only observed sporadically, use wetlands for aerial foraging, or breed in shrub, 56 Table 4. Least squares geometric means and lower and upper 95% confidence limits (CL) by wetland type for breeding bird densities (birds per ha) measured during point counts conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005- 2007. Bolded p-values indicate a significant difference between wetland types (p<0.05). Diked (n=294) Undiked (n=311) Lower Upper Lower Upper Bird Density Variable Mean CL CL Mean CL CL P-value All Birds 9.95 7.46 13.17 9.26 6.95 12.25 0.3318 Wetland-dependent Birds 8.19 6.23 10.68 7.18 5.45 9.38 0.0461 Wetland-associated Birds 1.00 0.55 1.59 1.06 0.61 1.64 0.8260 Non-wetland Birds 0.23 0.03 0.47 0.35 0.14 0.60 0.4132 Wetland-dependent Species Canada Goose 0.03 0.01 0.04 0.01 0.00 0.03 0.3904 Mute Swan 0.01 -0.01 0.02 0.01 -0.01 0.03 0.5091 Wood Duck 0.02 0.00 0.03 0.01 0.00 0.03 0.7895 Mallard 0.02 -0.02 0.06 0.05 0.01 0.09 0.1286 Pied-billed Grebe 0.02 0.00 0.04 0.02 0.00 0.05 0.7927 American Bittem 0.06 0.04 0.08 0.02 0.01 0.04 0.0012 Least Bittem 0.04 0.02 0.06 0.01 -0.01 0.02 0.0024 King Rail 0.00 -0.01 0.01 0.01 0.00 0.02 0.2402 Virginia Rail 0.20 0.14 0.26 0.15 0.09 0.21 0.2816 Sora 0.04 0.01 0.07 0.03 0.01 0.06 0.6132 Common Moorhen 0.04 0.02 0.07 0.02 0.00 0.04 0.0864 American Coot 0.09 0.02 0.16 0.10 0.03 0.17 0.8550 Black Tern 0.08 0.01 0.15 0.13 0.06 0.21 0.2105 Forster’s Tern 0.04 -0.04 0.12 0.21 0.12 0.30 0.0057 Tree Swallow 0.21 0.04 0.41 0.33 0.15 0.54 0.3515 Willow Flycatcher 0.04 0.02 0.07 0.02 0.00 0.05 0.2605 Sedge Wren 0.01 -0.07 0.09 0.05 -0.03 0.14 0.4389 Marsh Wren 1.89 1.22 2.76 1.31 0.79 1.96 0.2024 Swamp Sparrow 1.02 0.49 1.72 0.94 0.45 1.60 0.7846 Red-winged Blackbird 2.69 2.00 3.53 2.44 1.81 3.21 0.4379 Yellow-headed Blackbird 0.09 -0.05 0.25 0.00 -0.13 0.12 0.2766 Wetland-associated Species Caspian Tern 0.01 0.00 0.03 <0.01 -0.01 0.02 0.3155 Eastern Kingbird 0.02 0.00 0.05 0.01 -0.01 0.04 0.3972 Barn Swallow 0.08 0.00 0.16 0.17 0.09 0.25 0.0986 Yellow Warbler 0.18 0.07 0.31 0.13 0.03 0.25 0.5182 Common Yellowthroat 0.43 0.23 0.67 0.53 0.31 0.77 0.3868 Common Grackle 0.20 0.10 0.31 0.10 0.01 0.20 0.0612 Non-wetland Species Song Sparrow 0.10 -0.01 0.21 0.20 0.09 0.32 0.1751 57 Table 5. Avian species unique to diked and undiked wetlands and common to both types during breeding bird point counts conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007. Species Diked Common Undiked Wetland-dependent Species Canada Goose Mute Swan Wood Duck Mallard Blue-winged Teal X Redhead X Pied-billed Grebe American Bittem Least Bittem Great Blue Heron X Great Egret X Green Heron X Black-crowned Night-Heron Northern Hanier King Rail Virginia Rail Sora Common Moorhen American Coot Spotted Sandpiper X Ring-billed Gull Herring Gull Black Tern Forster's Tern Alder Flycatcher X Willow Flycatcher Tree Swallow Northern Rough-winged Swallow X Bank Swallow X Sedge Wren Marsh Wren Swamp Sparrow Red-winged Blackbird Yellow-headed Blackbird X ><><><>< ><><>< ><><><>< ><><>< ><><><>< ><><><>< 58 Table 5. Cont’d. Species Wetland-associated Species Killdeer Caspian Tern Black-billed Cuckoo Eastern Kingbird Warbling Vireo Purple Martin Cliff Swallow Barn Swallow Gray Catbird Yellow Warbler Common Yellowthroat Common Grackle Non-wetland Species Ring-necked Pheasant Rock Pigeon Mourning Dove Chimney Swift Northern Flicker Blue Jay Black-capped Chickadee American Robin European Starling Cedar Waxwing Yellow-rumped Warbler American Redstart Scarlet Tanager Song Sparrow Northern Cardinal Rose-breasted Grosbeak Indigo Bunting Brown-headed Cowbird Baltimore Oriole American Goldfinch Total Number of Species Diked 13 59 Common ><><>< >< ><><><><><><><><>< >< ><><><>< >< 44 Undiked X forest, or edge habitats. I calculated a Jaccard index value of 0.66 and Sorensen index of 0.80 between diked and undiked wetlands, indicating high similarity in species composition between the wetland types. Morisita similarity index between diked and undiked wetlands was 0.98, which indicates high similarity in species composition and abundance. The first dimension of the correspondence analysis explained 31.3% of the variation in bird abundance during point counts and the second dimension 25.6% of the variation. Correspondence analysis did not reveal distinct groupings of bird use at diked and undiked sites (Figure 9). Gulls/tems, Pied-billed Grebe, and coots/moorhens had positive dimension 1 coordinates and non-wetland birds and herons had negative values (Table 6). Bird groups that use large open water areas or wetland edges appeared to be associated with positive dimension 2 coordinates, such as gulls/tems, aerial-foraging songbirds, and non-wetland birds, while species more typical of emergent marshes, such as bitterns, herons, coots/moorhens, and rails, tended to have negative coordinates (Table 6). Diked and undiked sites at St. Clair Flats seemed to be separated along the second dimension, indicating that bittems and coats/moorhens were more abundant in diked compared to undiked wetlands, and undiked wetlands tended to have greater numbers of tems/gulls, Pied-billed Grebes, and waterfowl compared to diked sites. Diked and undiked Saginaw Bay sites were not separated along either dimension. 60 i i l 1‘ 1 0 -- TG* " NO 0'5 -.. * o >I< “ ,3. + GR g; 00 AF + E . 2+: 0 to * 0 £1 SA * N 00 WA 5 ~ +0 o + o + + g» to 9: sw a) + 0 >1: .§ RA or: + CM 0 o + +1. —o.5 + “E >k + -- * BI -10 —— __ i i i i —1.0 —0.5 0.0 0.5 1.0 Dimension 1 (31.3%) Figure 9. Biplot of site and bird group coordinates for dimensions 1 and 2 from correspondence analysis conducted using point count data collected at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan, coastal wetlands, 2005-2007. Site coordinates are coded by wetland type (“+” diked; “o” undiked). Bird group coordinates are coded with an “*” and labeled as follows: AF = aerial-foraging songbirds, B1 = bitterns, CM = American Coots and Common Moorhens, GR = Pied—billed Grebes, HE = herons, N0 = non-wetland birds, RA = rails, SA = wetland-associated songbirds, SW = wetland-dependent songbirds, TG = tems and gulls, and WA = waterfowl. 61 Table 6. First and second dimension coordinates for birds species/groups from correspondence analysis conducted on data from 605 breeding bird point counts (294 diked and 311 undiked) at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bird Species/Group Dimension 1 Dimension 2 Waterfowl (WA) 0.2682 0.1519 Pied-billed Grebe (GR) 0.8915 0.4398 Bittems (BI) 0.2813 -0.6925 Herons (HE) -0.5228 -0.4930 Rails (RA) -0.03 84 -0.2811 American Coots/Common Moorhens (CM) 0.6442 -0.2082 Gulls/T ems (GT) 0.9188 0.9214 Aerial-foraging Songbirds (AF) -0.3587 0.2352 Wetland-dependent Songbirds (SW) 0.0603 -0.l353 Wetland-associated Songbirds (SA) 02674 0.1975 Non-wetland Birds (NO) -0.7244 0.4769 Timed-area Surveys Mean areal densities of all birds, wetland-dependent birds, and wetland-associated birds were similar between diked and undiked coastal wetlands (Table 7). Average Canada Goose (Branta canadensis), Wood Duck (Aix sponsa), and Common Moorhen (Gallinula chloropus) areal densities were greater in diked compared to undiked wetlands (p<0.0001, p=0.0002, and p=0.0168, respectively). Mean areal densities of American Coot (F ulica americana), Ring-billed Gull (Larus delawarensis), Herring Gull (Larus argentatus), and Forster’s Tern were greater in undiked than diked wetlands (p=0.0378, p=0.0025, p=0.0457, and p=0.0004, respectively). Table B-6 (Appendix B) provides areal densities and frequencies of occurrence for all bird species observed during timed- area surveys by study area and wetland type. 62 Table 7. Least squares geometric means and lower and upper 95% confidence limits (CL) by wetland type for area] bird densities (birds per ha open water) measured during timed-area surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05). Diked (n=144) Undiked (n=143) Lower Upper Lower Upper Bird Density Variable Mean CL CL Mean CL CL P-value All Birds 3.17 2.27 4.32 2.16 1.43 3.11 0.1159 Wetland-dependent Birds 3.00 2.16 4.08 2.04 1.34 2.93 0.1207 Wetland-associated Birds 0.14 0.07 0.22 0.10 0.03 0.18 0.2436 Wetland Dependent Species Canada Goose 0.31 0.23 0.40 0.03 -0.03 0.10 <0.000l Mute Swan 0.15 0.03 0.29 0.09 -0.04 0.23 0.4863 Wood Duck 0.63 0.39 0.91 0.05 -0.12 0.24 0.0002 Mallard 0.20 0.03 0.49 0.63 0.28 1.07 0.0625 Great Blue Heron 0.11 0.03 0.18 0.04 -0.03 0.12 0.1210 Great Egret 0.03 0.04 0.09 0.07 0.00 0.15 0.3591 Black-er. Ni ght-Heron 0.04 -0.04 0.13 <0.01 -0.09 0.09 0.4736 Pied-billed Grebe 0.18 0.03 0.34 0.17 0.02 0.35 0.9821 Common Moorhen 0.07 0.03 0.11 0.02 -0.02 0.06 0.0168 American Coot 0.04 -0.01 0.10 0.11 0.05 0.17 0.0378 Spotted Sandpiper 0.03 0.00 0.06 <0.01 -0.03 0.04 0.1466 Greater Yellowlegs 0.01 0.00 0.02 0.01 0.00 0.02 0.9567 Lesser Yellowlegs 0.03 -0.01 0.08 0.01 -0.03 0.06 0.4527 Dunlin <0.01 -0.03 0.02 0.02 0.00 0.05 0.1581 Ring-billed Gull 0.01 -0.01 0.02 0.04 0.03 0.06 0.0025 Herring Gull <0.01 -0.02 0.03 0.04 0.01 0.07 0.0457 Black Tern 0.36 0.09 0.69 0.18 -0.08 0.50 0.3762 Forster’s Tern 0.04 -0.04 0.13 0.20 0.10 0.31 0.0004 Wetland Associated Species Killdeer 0.04 -0.02 0.10 0.01 -0.05 0.07 0.2842 Caspian Tern 0.10 0.06 0.14 0.09 0.05 0.13 0.5782 63 Average linear (birds per km of edge surveyed) and areal (birds per ha surveyed) densities for species observed during timed-area surveys were correlated (1:0.998, p<0.0001), and a chi-square test revealed no difference between linear and areal densities (p=0.3581) in the number of species with greatest densities in diked and undiked wetlands. I found greater mean linear densities of all birds combined (p=0.01 71) and wetland-dependent birds (p=0.0241) in undiked compared to diked wetlands (Table 8), while areal densities for these variables were similar between the two wetland types. Average wetland-associated bird linear densities were similar between the wetland types (p=0.4016), which is consistent with results of areal density analysis. Similar to the results of areal density analyses, Canada Goose and Wood Duck linear densities were greater in diked wetlands (p<0.0001). I observed greater mean Mallard linear densities in undiked than diked wetlands (p=0.0001). Mean linear densities of Ring-billed Gull, Herring Gull, and F orster’s Tern were greater in undiked compared to diked wetlands (p=0.0027, p=0.0278, and p=<0.0001, respectively), which is the same pattern observed in the areal density analysis. Total species richness during timed-area surveys was 32 species for both wetland types, with 25 species common to diked and undiked wetlands (Table 9). The seven species unique to diked wetlands were considered wetland-dependent. Of the seven species unique to undiked coastal wetlands, six were considered wetland-dependent and one species wetland-associated. Bird species unique to the wetland types were observed irregularly in low numbers. J accard and Sorensen similarity index values for diked and undiked wetlands were 0.64 and 0.73, respectively, which indicates high similarity in species composition between the wetland types. I calculated a Morisita index value of 64 Table 8. Least squares geometric means and lower and upper 95% confidence limits (CL) by wetland type for linear bird densities (birds per km of edge) measured during timed-area surveys conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05). Diked (n=144) Undiked (n=143) Lower Upper Lower Upper Bird Density Variable Mean CL CL Mean CL CL P-value A11 Birds 7.02 5.55 8.82 9.92 7.87 12.44 0.0171 Wetland-dependent Birds 6.65 5.27 8.33 9.26 7.36 11.58 0.0241 Wetland-associated Birds 0.29 0.17 0.43 0.37 0.24 0.52 0.4016 Wetland Dependent Species Canada Goose 0.61 0.41 0.82 0.10 -0.04 0.25 <0.000l Mute Swan 0.30 0.04 0.64 0.30 0.00 0.68 0.9722 Wood Duck 1.23 0.78 1.80 0.13 -0.12 0.45 <0.000l Mallard 0.37 0.04 0.82 2.10 1.28 3.20 0.0001 Great Blue Heron 0.23 0.08 0.39 0.15 0.00 0.31 0.3808 Great Egret 0.05 -0.10 0.23 0.28 0.07 0.53 0.1001 Black-er. Night-Heron 0.07 -0.05 0.22 <0.01 -0.14 0.15 0.4488 Pied-billed Grebe 0.44 0.06 0.95 0.55 0.13 1.15 0.6565 Common Moorhen 0.14 0.08 0.20 0.08 0.02 0.14 0.1694 American Coot 0.08 -0.04 0.22 0.23 0.08 0.39 0.1020 Spotted Sandpiper 0.03 0.01 0.06 0.01 -0.02 0.03 0.0720 Greater Yellowlegs 0.01 -0.01 0.04 0.02 0.00 0.05 0.6089 Lesser Yellowlegs 0.03 -0.01 0.07 0.02 -0.02 0.06 0.6951 Dunlin <0.01 -0.04 0.04 0.05 0.00 0.09 0.1072 Ring-billed Gull 0.02 -0.05 0.09 0.16 0.08 0.25 0.0027 Herring Gull <0.01 -0.1 1 0.11 0.19 0.05 0.35 0.0278 Black Tern 0.56 0.14 1.14 0.48 0.04 1.12 0.8322 Forster’s Tern 0.06 -0.13 0.30 0.68 0.37 1.07 <0.0001 Wetland Associated Species Killdeer 0.04 -0.02 0.10 0.02 -0.04 0.08 0.3921 Caspian Tern 0.25 0.15 0.35 0.35 0.24 0.46 0.2074 65 Table 9. Avian species unique to diked and open wetlands and common to both types during timed-area surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005-2007. Species Diked Common Undiked Wetland-dependent Species Canada Goose Mute Swan Wood Duck Mallard Blue-winged Teal Northern Shoveler Northern Pintail Green-winged Teal Canvasback Redhead Scaup (species unknown) Hooded Merganser Pied-billed Grebe Double-crested Cormorant American Bittem Least Bittem Great Blue Heron Great Egret Green Heron Black-crowned Night-Heron Virginia Rail Sora Common Moorhen American Coot Spotted Sandpiper Greater Yellowlegs Lesser Yellowlegs Least Sandpiper Dunlin Wilson’s Snipe Ring-billed Gull Herring Gull Black Tern Forster's Tern Belted Kingfisher 66 ><><><><>< >< ><><><><><>< ><><><><>< ><><><><>< ><><>< Table 9. Cont’d. Species Diked Common Undiked Wetland-associated Species Bald Eagle X Killdeer X Caspian Tern X Common Tern X Total Number of Species 7 25 7 0.62 between diked and undiked sites. Although the three indices suggested substantial similarity in the breeding bird communities in the Open water zone of diked and undiked wetlands, they were all lower compared to values observed for the emergent zone. Dimension 1 of the correspondence analysis explained 40.5% of the variation in bird densities of the open water/aquatic bed zone and the second dimension 21.2% of the variation. Correspondence analysis separated the bird species/ groups into two clusters along the first dimension, with dabbling ducks and rails on the negative end and the remaining groups clumped from approximately 0.4 to 0.8 on the positive end (Figure 10, Table 10). Shorebirds, diving ducks, and terns/ gulls had positive dimension 2 coordinates, while Wood Ducks, herons, and geese/swans had negative values (Table 10). The majority of the diked sites had positive dimension 1 and negative dimension 2 coordinates, which indicated greater densities of Wood Ducks, herons, and geese/swans compared to the other sites. Most undiked sites seemed to be associated with greater densities of terns/gulls, diving ducks, and Pied-billed Grebes (Figure 10). However, there were a small number of diked sites associated with the same bird groups. 67 I l 1 i I1 1.5 —~ SH * ¢ 1.0 -~ 0' :5 ”a O * TG g?! o g 05‘- 00 Gig N O O * CM 5 RA 2k + 0 0.0 + g >1: + Bl *+ a DA + + + _0.5 __ G8 *+ HE * ++ wo a: + —1.0-- + + i i J. i J. -1.5 —1.0 —o.5 0.0 0.5 1.0 Dimension 1 (40.5%) Figure 10. Biplot of site and bird group coordinates for dimensions 1 and 2 from correspondence analysis conducted using timed-area survey data collected at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan, coastal wetlands, 2005-2007. Sites scores are coded by wetland type (“+” diked; “o” undiked). Bird groups are coded with and labeled as follows: B1 = bitterns, CM = American Coots and Common Moorhens, DA = dabbling ducks, D1 = diving ducks, GR = Pied-billed Grebes, GS = Canada Geese and Mute Swans, HE = herons, RA = rails, SH = shorebirds, TG = terns “*9, an and gulls, and WD = Wood Ducks. 68 Table 10. First and second dimension coordinates for birds species/ groups from correspondence analysis conducted on data from 287 timed-area surveys (144 diked and 143 undiked) at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bird Species/Group Dimension 1 Dimension 2 Dabbling Ducks (DA) -1.0322 -0.0864 Diving Ducks (DI) 0.4771 0.9957 Canada Geese/Swans (GS) 0.7437 -0.4365 Wood Ducks (WD) 0.7897 -0.7924 Pied-Billed Grebes (GR) 0.3735 0.2695 Bittems (BI) 0.6412 -0.0447 Herons (HE) 0.4882 -0.7209 American Coots/Common Moorhens (CM) 0.5761 0.2881 Rails (RA) -0.9586 0.2236 Shorebirds (SH) 0.4946 1.3445 Gulls/T ems (GT) 0.4993 0.7940 Vegetation and Physical Variable Sampling Hydrological and biogeochemical changes resulting from coastal wetland diking appear to have caused differences in vegetation and physical parameters measured at diked and undiked sites (Table 11). Mean percent cover of open water/aquatic bed wetland (p=0.0003), floating vegetation (p=0.0020), persistent deep-water vegetation (p=0.0258), and cattail (Typha) (p=0.0001) was greater in diked than undiked coastal wetlands. Average percent cover of several variables was greater in undiked compared to diked sites: persistent shallow-water vegetation (p=0.0033), non-persistent shallow-water vegetation (p=0.0005), bulrush (Schoenoplectus) (p<0.0001), common reed (Phragmites) (p=0.0227), surface litter (p=0.0038), and exposed sediments (p=0.0171). Percent cover of total emergent and submersed vegetation was similar between wetland types =0.7578 and p=0.1393, respectively). Mean density of cattail stems was greater (p<0.0001) in diked wetlands, while densities of bulrush and common reed were greater 69 Table 11. Least squares geometric means and lower and upper 95% confidence limits (CL) for vegetation and physical variables measured during quadrat sampling conducted at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2006-2007. Bolded p- values indicate a significant difference between wetland types (p<0.05). Diked (n=771) Undiked (n=750) Lower Upper Lower Upper Vegetation/Physical Variable Mean CL CL Mean CL CL P-value Percent Cover Emergent Vegetation 23.9 15.2 34.0 25.7 16.3 36.4 0.7578 Open Water/Aquatic Bed 73.8 61.6 84.5 40.0 26.9 53.9 0.0003 Submersed Vegetation 1.1 0.2 2.6 0.2 0.0 1.1 0.1393 Floating Vegetation 1.9 0.7 3.7 <0.1 0.2 0.5 0.0020 Persistent Deep-water 16.9 9.9 25.3 6.3 2.1 12.7 0.0258 Persistent Shallow-water 1.0 0.0 3.8 8.0 3.4 14.3 0.0033 Non-persistent Deep-water <0. 1 <0. 1 <0.1 <0.1 <0. 1 0.1 0.0601 Non-persistent Shallow-water 0.1 <0.1 0.4 0.8 0.4 1.3 0.0005 Cattail 16.3 9.7 24.3 1.8 0.1 5.7 0.0001 Bulrush <0.1 <0.1 0.2 1.8 1.0 2.8 <0.0001 Common Reed 0.2 0.1 1.7 3.4 1.0 7.2 0.0227 Surface Litter 13.0 6.5 21.2 31.0 20.9 42.2 0.0038 Exposed Sediments <0.1 <0.1 0.1 0.3 0.1 0.6 0.0171 Stem Density Cattaill 11.78 6.76 20.06 1.58 0.52 3.38 <0.0001 Bulrush] 0.10 -019 0.49 2.88 1.80 4.37 <0.0001 Common Reed' 0.46 -014 1.48 2.80 1.16 5.67 0.0134 Trees and Shrubsz 0.24 0.08 0.42 0.04 0.10 0.20 0.0837 Vegetation Height (m) 1.55 1.22 1.92 1.44 1.11 1.82 0.6628 Visual Obstruction (m) 1.17 0.85 1.56 0.81 0.52 1.16 0.1271 Water Depth (m) 0.30 0.22 0.39 0.09 0.02 0.17 0.0002 Organic Sediment Depth (m) 0.40 0.30 0.50 0.24 0.15 0.34 0.0069 I No. stems per 0.25 m2 quadrat. 2 No. stems >2 m tall per 20 m2 (within 2.5 m radius of quadrat center). 70 in undiked wetlands (p<0.0001 and p=0.0164, respectively). Mean depths of water and organic sediment were greater in diked compared to undiked wetlands (p=0.0002 and p=0.0069, respectively). The first component from the PCA explained 37.1% of the vegetation and physical variable variation among avian point count stations, while the second component explained 21.3% of the variation. The first axis appeared to represent a gradient from deep open water/aquatic bed wetland on the negative end to dense shallow-water marsh on the positive end (Figure 11). Principal component 1 (PC 1) was negatively related to percent open water, water depth, percent submersed vegetation, and percent floating vegetation, and positively related to percent cover of litter, persistent shallow-water vegetation, total emergent vegetation, and common reed, and vegetation height (Table 12). The second axis seemed to represent a gradient from cattail marsh on the positive end to common reed marsh on the negative end. The second principal component (PC 2) was positively related to percent cover of cattail, percent cover persistent deep-water emergents, and organic sediment depth, and negatively related to percent cover of persistent shallow-water emergents and common reed (Table 12). Although there was substantial overlap between diked and undiked point count stations in PC scores, undiked wetlands tended to have higher PC 1 scores and lower PC 2 scores compared to diked wetlands (Figure 11). The PCA indicates a tendency for undiked sites to have shallower water, denser vegetation, more common reed, and taller vegetation compared to diked wetlands, while diked sites typically had greater water and organic sediment depths, greater percent cover of open water, submersed vegetation, and floating plants, and more cattail compared to undiked wetlands. 71 h __ % Cattail ' ' ' ' ' ' % Persistent Deep 50 + -2 Organic Depth ++ o 0 ”44+ 2.5 ‘i— + 4%; ++° o .0. $3 +51% "'t + + o or; + + £1445 + o "" + N __ + + 0 $ + O .._ 0 + + £8" fi‘iufiao 0. + O O -2 5 -— + 0 ‘8 -_ % Persistent Shallow -5.0 i- __ % Common Reed 1 1 I l 1 1 -7.5 -5.0 -2.5 0.0 2.5 5.0 7.5 PC 1 (37.1%) % Open Water % Litter Water Depth ‘n—p % Persistent Shallow % Submersed % Emergents % Floating % Common Reed Height Figure 11. Bi-plot of PC 1 X PC 2 from principal components analysis conducted using 14 vegetation and physical variables gathered during quadrat sampling at 179 random avian point count stations at St. Clair Flats and Saginaw Bay, Michigan, 2006-2007. Point scores are coded by wetland type (“+” diked; “O” undiked). 72 Table 12. Eigenvectors for first two principal components obtained through PCA of habitat data collected at 179 point count stations located at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2006-2007. Habitat Variable Principal Component 1 Principal Component 2 Percent Cover Emergent Vegetation 0.3300 0.2987 Open Water/Aquatic Bed -0.3871 -0.0511 Submersed Vegetation -0.3056 -0.1844 Floating Vegetation -0.2196 0.0294 Persistent Deep-water -0.0208 0.5271 Persistent Shallow-water 0.3446 -0.2264 Non-persistent Shallow-water 0.1968 -0.0060 Cattail -0.0354 0.5477 Bulrush 0.0424 -0. 1466 Common Reed 0.3017 -0.2179 Surface Litter 0.3562 0.0121 Vegetation Height (m) 0.2814 0.1853 Water Depth (m) -0.3546 -0.0112 Organic Sediment Depth (m) -0.1260 0.3750 73 DISCUSSION Breeding Bird Use of Diked and Undiked Wetlands Wildlife managers in the Great Lakes region built dikes around wetlands to provide the capability to manage water levels and enhance conditions for wetland birds. This study evaluated bird use of diked wetlands through comparisons with wetlands open to Great Lakes water level fluctuations, and examined bird use in the context of habitat conditions. I found greater densities of some wetland-dependent breeding bird species in diked coastal wetlands, while others were observed in lower densities compared to undiked sites. Most of the breeding bird density variables were not different between wetland types. Total wetland-dependent bird densities were greater in diked wetlands during point counts, but areal densities observed during timed-area surveys were similar and linear densities were greater in undiked than diked wetlands. Galloway et al. (2006) observed greater abundance of several groups of birds in diked wetlands, including marsh-nesting obligates, marsh—nesting generalists, and area-sensitive marsh-nesting obligates, in pooled comparisons of diked and undiked coastal wetlands of the southern Great Lakes. I found comparable species richness in diked and undiked wetlands during both point counts and timed-area surveys. Similarity index results suggested breeding bird species composition and abundance was similar between wetland types. Approximately two-thirds of the species documented during breeding surveys were common to both wetland types, and unique species primarily consisted of species observed in low numbers, such as nonbreeding species or late migrants, or those that use adjacent habitats, such as forests, shrub lands, or grasslands. Galloway et al. (2006) 74 found greater cumulative species richness in diked compared to undiked wetlands for several of the marsh bird groups they compared, and only aerial forager species richness was greater in undiked wetlands. Although Galloway et al. (2006) observed greater bird abundance and species richness for several bird groups in overall comparisons of diked and undiked sites, they found few differences in wetland bird use in paired comparisons of nearby diked and undiked wetlands. Differences in the results of my study and Galloway et al. (2006) could be due to variation in management and hydrologic regimes, human disturbance levels, invasive species impacts, and surrounding landscape. Galloway et a1. (2006) also only sampled during one field season, which may not have accounted for long-term or annual variation in bird use and wetland conditions. Most of the breeding species observed in greater densities in diked than undiked coastal wetlands use deep-water marshes for some part of their life cycle. Canada Geese and Wood Ducks were observed in greater densities in diked wetlands during timed-area surveys. Higher water levels in the diked wetlands likely provided attractive brood rearing habitat for both species proximal to nesting sites. Canada Geese regularly nest on dikes and were observed feeding on dikes and in nearby row-crop fields. Most of the diked wetlands had Wood Duck nest boxes, while the undiked wetlands did not. Wood Ducks may have been attracted to dense cover provided by emergent and floating-leaved plants of the diked wetlands, and the greater abundance of aquatic invertebrates (Provence 2008), which are an important food source for nesting females and broods (Drobney and Fredrickson 1979). Densities of American and Least Bittems were greater in diked coastal wetlands. Although he only surveyed diked wetlands, Yocum (2007) found Least Bittems to be abundant at some sites. Least Bittems tend to use deeper water 75 marshes when compared to American Bittem (Weller 1961, Weller and Spatcher 1965), and Bogner and Baldassarre (2002) suggested vegetation type and cover ratios (emergentzopen water) may be more important factors to Least Bittems populations than marsh size. Weller (1961) found Least Bittem nests primarily in cattail and bulrush marshes, usually near open water patches, and only occasionally in common reed. During bird surveys in cattail, marsh meadow, and common reed wetlands, Meyer (2003) only observed Least Bittems in common reed stands. American Bittems in Maine seemed to prefer impounded and beaver-created wetlands over wetlands of glacial origin (Gibbs et al. 1992). Higher water levels and greater percent open water in the diked wetlands may have increased interspersion of emergent vegetation and open water, which would be attractive to American and Least Bittems. Dikes surrounding the isolated coastal wetlands may have provided nesting bittems protection from wave action and seiches. Higher water levels in the diked wetlands may have created a more stable environment for invertebrates, amphibians, and small fish that bittems use for food. Although densities of Common Moorhen were similar between diked and undiked wetlands during point counts, areal densities were higher in diked wetlands during timed- area surveys. Common Moorhens typically breed in permanently flooded deep-water marshes consisting of tall emergent vegetation interspersed with open areas containing floating-leaved and submersed vegetation or mudflats (Bannor and Kiviat 2002). Mallard, American Coot, Ring-billed Gull, Herring Gull, and Forster’s Tern were the only breeding species observed in greater densities in undiked compared to diked wetlands. Mean linear density of Mallard was greater in undiked sites. Mallards prefer to forage in shallow water (Fredrickson and Taylor 1982), and undiked wetlands had 76 lower water depths than diked sites, which could account for differences in Mallard abundance. American Coot areal densities in diked and undiked emergent marsh were similar during point counts; however, densities recorded in the open water/aquatic bed zone during timed-area surveys were greater in undiked coastal wetlands. Weller and Fredrickson (1974) suggested that American Coots pioneer new habitats quickly, while Common Moorhens tend to move into sites several years after reflooding. Fish are an important component of the diets of Ring-billed Gull, Herring Gull, and F orster’s Tern (see Ryder 1993, Pierotti and Good 1994, McNicholl et al. 2001). Studies conducted in Lake Erie coastal wetlands indicated differences between diked and undiked wetlands in total fish species richness and abundance, age class frequencies, lengths, and body condition indices for some species (Johnson et al. 1997, Markham et a1. 1997). Fish abundance and composition were not measured in my study, but it would be useful to know the relative abundance of forage fish in diked and undiked wetlands to understand the effects of coastal wetland diking on these bird species. Foraging in diked wetlands may have been more difficult for gulls and Forster’s Tems due to greater coverage of floating vegetation. F orster’s Tems were only observed nesting in undiked wetlands where dead bulrush stems from the previous growing season collected, which provided a substrate for their floating nests. Bulrush percent coverage and stem density were lower in diked than in undiked wetlands. Vegetation and Physical Characteristics of Diked and Undiked Wetlands I observed greater percent cover of open water, floating vegetation, persistent deep-water emergents, and cattails, and greater mean cattail density in diked compared to 77 undiked wetlands, and these differences were likely due to higher, more stable water levels. Although water levels of diked wetlands often dropped dramatically during the summer, the majority of the wetlands remained inundated throughout the season. Percent cover and density of bulrush and common reed were greater in undiked than diked wetlands. Albert and Brown (2008) observed similar results when comparing the vegetation at several of the same diked and undiked study sites. Most of variables in my study that differed between diked and undiked sites also tended to have high loadings in PC 1 and PC 2 of the PCA. My PCA indicated some separation of diked and undiked point count stations and generally supported the results of parametric comparisons. In vegetation comparisons between diked and undiked wetlands, Herrick and Wolf (2005) similarly found greater cattail cover in diked wetlands and greater common reed cover in undiked wetlands. Lower mean percent cover and stem density of common reed in diked than undiked wetlands may be due to higher water levels and activities (e. g., herbicide application, burning) used to control common reed in some diked areas. Several studies have suggested that wetland plant species are distributed along gradients of disturbance, fertility, and organic matter content based on competitive abilities (e.g., Wilson and Keddy 1986, Gaudet and Keddy 1988, 1995, Day et al. 1988, Moore et al. 1989), with species such as cattails outcompeting other species in areas with high fertility and low disturbance (Wisheu and Keddy 1992). Diked wetlands likely experience less disturbance than undiked sites due to higher water levels and infrequent complete drawdowns, and greater fertility due to high organic content of soils and trapped nutrients, which could lead to dominance by cattail. Herrick et al. (2007) suggested that diked coastal wetlands serve as traps for organic matter and nutrients. 78 I found no difference in percent cover of submersed plants between the wetland types; however, sampling was focused in emergent marsh where point counts were conducted. Sampling of submersed vegetation within the open water/aquatic bed zone may have produced different results. Aquatic bed zones of the diked wetlands, including excavated channels, typically had dense submersed vegetation. When I conducted PCA of habitat data gathered at point count stations, percent cover of submersed vegetation was an important variable in PC 1. There was some separation of diked and undiked wetlands along the first axis, which indicated that diked wetlands were associated with greater percent cover of submersed vegetation compared to undiked sites. Prince (1985) observed that bird species richness and nesting density were negatively related to percent open water during surveys of diked and undiked wetlands, and that the lack of submersed vegetation limited breeding bird use in some wetlands. Management Implications Breeding bird use of diked and undiked coastal wetlands in Michigan was largely similar, despite clear differences in vegetation and physical variables. American Bittem, Least Bittem, and Common Moorhen, all rare species known to use deep-water marshes, appeared to benefit from diked wetland management. Several years of low Great Lakes water levels have limited the availability of deep-water cattail marshes in undiked wetlands of both study areas, which may explain greater densities of the above species in diked wetlands. Although deep-water bulrush marshes were common in undiked wetlands, they may have been of lower value to nesting bittems and Common Moorhens than diked cattail marshes. Standing dead bulrushes from the previous season that could 79 be used for cover and nest building are usually removed by ice scour, and new bulrush growth occurs later in the season than cattail. A common criticism of diked coastal wetlands is that their management focuses on waterfowl or game species, potentially at the detriment of rare and/or non-garne bird species. The results of my study do not support this criticism. Least Bittem is a State- threatened species, American Bittem and Common Moorhen are State special concern species, and Common Moorhen was recommended for listing as threatened in Michigan (Brewer et al. 2005). I also found no difference between diked and undiked wetlands in the densities of Black-crowned Night-Heron (State special concern), King Rail (State endangered), Marsh Wren (State special concern), and Yellow-headed Blackbird (State special concern). F orster’s Tern (State special concern species) was the only rare species observed in greater densities in undiked wetlands. Albert and Brown (2008) reviewed aerial photographs taken prior to dike construction at three of the sites used in my study, and they found that much of these areas appeared to be wet meadows mixed with densely vegetated emergent marsh. At least some of the diked wetlands may not have been used extensively by breeding F orster’s Tems before diking, given a predominance of wet meadow vegetation. Along with the rare species described above, I observed eight other species considered species of greatest conservation (SGCN) need in Michigan’s Wildlife Action Plan (Eagle et al. 2005). I found no difference between diked and undiked wetlands in the densities of seven of the eight SGCN. I observed greater area] densities of American Coot in undiked wetlands during timed-area surveys, but linear densities from timed-area surveys and areal densities from point counts were similar between wetland types. The diking and management of coastal wetlands did not seem to cause 80 substantial negative impacts to rare or nongame breeding bird species in the wetlands I investigated. Invasive populations of common reed have substantially expanded in Great Lakes coastal wetlands during the recent period of low water levels (Tulbure et al. 2007, E. Kafcas, Michigan Department of Natural Resources, person. commun.). Most climate change models predict decreasing Great Lakes water levels in the future (Mortsh et al. 2000, 2006, Lofgren et al. 2002, Croley 2003), which could further increase common reed expansion in undiked wetlands and potentially reduce the value of these areas for birds species of management concern. Although the construction of dikes may have provided avenues for the expansion of invasive species (e. g., common reed) in coastal wetlands, diked wetlands now provide the opportunity to manage against invasive plant species like common reed. Given that the future status of coastal wetlands is uncertain due to the effects of climate change and invasive species, diked wetlands may provide important management opportunities to maximize use by wetland birds. Greater linear density of Mallards in undiked than diked wetlands was not predicted, because they are a focal species in diked wetland management. My results are also surprising given that invertebrate abundance was greater in diked compared to undiked sites at St. Clair Flats (Provence 2008), which included several taxa known to be important food items for Mallards during the breeding season. F redrickson and Taylor (1982) noted that the preferred foraging depth for Mallards is approximately 10-15 cm in seasonally flooded impoundments. Although invertebrates seemed abundant in diked wetlands during this study, Mallards may have had better access to food in undiked wetlands due to shallower water depths. Managing the diked wetlands for shallower 81 water depths could enhance use by Mallards and many other wetland bird species by improving access to abundant invertebrate foods. Periodic complete drawdowns of the diked wetlands could potentially improve habitats for breeding birds. Kadlec and Smith (1992) noted three potential benefits of drawdowns as nutrient release due to the decomposition of organic sediments, consolidation of loose sediments due to drying, and germination and establishment of emergent vegetation, including annual species. Drawdowns could reduce the buildup of organic matter, release nutrients and stimulate plant growth, and improve vegetation and structural diversity of the diked marshes. Recommended frequencies for drawdowns have ranged from 5 to 7 years (Harris and Marshall 1963, Whitman 1976). Areas with multiple impoundments should not be drawn down in the same season, since dewatering could cause short-term impacts to invertebrates (Kadlec 1962) and breeding bird use. Since drawdowns can encourage growth of invasive plant species (Fredrickson and Taylor 1982), I suggest close monitoring of the vegetation response if drawdowns are conducted. Research Needs My study occurred during a period of low Great Lakes water levels, and water level fluctuations and depths are known to affect bird use of wetlands. Timmerrnans et al. (2008) found annual abundances of several wetland bird species were positively correlated with annual water level changes in Lakes Michigan, Huron, and Erie. Steen et al. (2006) felt that the stabilization of water levels was an important factor contributing to the decline of some bird species using Lake Ontario coastal wetlands. Bird use of diked 82 and undiked wetlands during normal to high water levels could differ from the results of my study, and more research is needed during other parts of the Great Lakes water level cycle to investigate if patterns of bird use change under different hydrological conditions. Long-term studies would be beneficial to understand changes in Great Lakes coastal wetlands that occur over 5-20 years. Research is needed to understand the effects of differences in wetland conditions (e. g., water depths, floating vegetation mats, interspersion) between diked and undiked wetlands on breeding bird use. More study is required to determine if the pattern of higher invertebrate abundance in diked compared to undiked wetlands that Provence (2008) observed at St. Clair Flats applies to wetlands in other parts of the Great Lakes, and to examine if wetland bird density and diversity is linked to food abundance and availability. Fish and amphibian populations are also likely affected by the diking of coastal wetlands, and the effects on their populations and the secondary effects on bird populations are not understood. Management guidelines need to be developed to maximize wildlife benefits in diked wetlands in the context of changing coastal wetland conditions associated with climate change and invasive species expansion, and for specific species of concern (e. g., game, threatened, endangered, SGCN). Diked wetlands provide opportunities to conduct experimental studies that test the success of water level management regimes (e. g., lower water levels, periodic drawdowns) for selected management goals (e. g., breeding use by focal species, diverse vegetation). For example, Mallards are often a focal species for management and invertebrate abundance was greater in some diked wetlands during this study (Provence 2008), but Mallard densities tended to be greater in undiked than diked wetlands. Water levels could be experimentally lowered in the diked wetlands to 83 evaluate if Mallard densities increase when preferred water depths for foraging are provided. 84 LITERATURE CITED Albert, D. A., and P. W. Brown. 2008. Analysis of vegetation in adjacent diked-undiked coastal wetlands. Michigan Natural Features Inventory Report 2008-14, Lansing, USA. American Ornithologists' Union. 1998. Check-list of North American Birds. Seventh edition. American Ornithologists' Union, Washington, DC, USA. Bannor, B. K., and E. Kiviat. 2002. Common Moorhen (Gallinula chloropus). Account 685 in A. Poole and F. Gill, editors. The birds of North America. The Birds of North America, Philadelphia, Pennsylvania, USA. Bellrose, F. C. 1980. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg, Pennsylvania, USA. Bogner, H. E., and G. A. Baldassarre. 2002. Home range, movement, and nesting of least bittems in western New York. Wilson Bulletin 114:297-308. Bookhout, T. A., K. E. Bednarik, and R. W. Kroll. 1989. The Great Lakes marshes. Pages 131-156 in L.M. Smith, R.L. Pederson, and RM. Kaminski, editors. Habitat management for migrating and wintering waterfowl in North America. Texas Tech University Press, Lubbock, USA. Brewer, R., R. J. Adams, Jr., J. A. Craves, D. N. Ewert, D. J. Flaspohler, M. J. Hamas, and M. J. Monfils. 2005. Recommendations of the technical advisory committee (birds) for endangered and threatened species in Michigan. Michigan Department of Natural Resources, unpublished report, Lansing, USA. Brown, S. C., and C. R. Smith. 1998. Breeding season bird use of recently restored versus natural wetlands in New York. Journal of Wildlife Management 62: 1480- 1491. Burton, T. M., C. A. Stricker, and D. G. Uzarski. 2002. Effects of plant community composition and exposure to wave action on invertebrate habitat use of Lake Huron coastal wetlands. Lakes and Reservoirs: Research and Management 7:255- 269. Conway, C. J. 2005. Standardized North American marsh bird monitoring protocols. US. Geological Survey Wildlife Research Report 2005-04, Arizona Cooperative Fish and Wildlife Research Unit, Tucson, Arizona, USA. 85 Cowardin, L. M., V. Carter, F. C. Golet, E. T. LaRoe. 1979. Classification of wetlands and deepwater habitats of the United States. US. Fish and Wildlife Service, Washington, DC, USA. Croley, T. E., II. 2003. Great Lakes climate change hydrologic impact assessment: International Joint Commission Lake Ontario-St. Lawrence River regulation study. National Oceanic and Atmospheric Administration Technical Memorandum GLERL-126, Ann Arbor, Michigan, USA. Crowley, S., C. Welsh, P. Cavanaugh, and C. Griffin. 1996. Weighing — birds: habitat assessment procedures for wetland-dependent birds in New England. University of Massachusetts, Department of Forestry and Wildlife Management, Amherst, USA. Day, R. T, P. A. Keddy, J. McNeill, and T. Carleton. 1988. Fertility and disturbance gradients: a summary model for riverine marsh vegetation. Ecology 69: 1044- 1054. Drobney, R. D., and L. H. Fredrickson. 1979. Food selection by wood ducks in relation to breeding status. Journal of Wildlife Management 43:109-120. Eagle, A. C., E. M. Hay-Chmielewski, K. T. Cleveland, A. L. Derosier, M. E. Herbert, and R. A. Rustem, editors. 2005. Michigan's wildlife action plan. Michigan Department of Natural Resources. Lansing, USA. . Accessed 13 February 2009. Fredrickson, L. H., and T. S. Taylor. 1982. Management of seasonally flooded impoundments for wildlife. US. Fish and Wildlife Service Resource Publication 148, Washington, DC, USA. Galloway, M., L. Bouvier, S. Meyer, J. Ingram, S. Doka, G. Grabas, K. Holmes, and N. Mandrak. 2006. Evaluation of current wetland dyking effects on coastal wetlands and biota. Pages 187-229 in L. Mortsch, J. Ingram, A. Hebb, and S. Doka, editors. Great Lakes coastal wetland communities: vulnerability to climate change and response to adaptation strategies. Environment Canada and the Department of Fisheries and Oceans, Toronto, Ontario, Canada. Gaudet, C. L., and P. A. Keddy. 1988. A comparative approach to predicting competitive ability from plant traits. Nature 334:242-243. Gaudet, C. L., and P. A. Keddy. 1995. Competitive performance and species distribution in shoreline plant communities: a comparative approach. Ecology 76:280-291. 86 Gibbs, J. P., S. Melvin, and F. A. Reid. 1992. American Bittem (Botaurus lentiginosus). Account 18 in A. Poole, P. Stettenheim, and F. Gill, editors. The birds of North America, The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Ornithologists’ Union, Washington, DC, USA. Gollop, J. B., and W. H. Marshall. 1954. A guide for aging duck broods in the field. Mississippi F lyway Council Technical Section. Gottgens, J. F., B. P. Swartz, R. W. Kroll, and M. Eboch. 1998. Long-term GIS-based records of habitat changes in a Lake Erie coastal marsh. Wetlands Ecology and Management 6:5-17. Hair, J. F., Jr., R. E. Anderson, and R. L. Tatham. 1987. Multivariate data analysis. Second edition. MacMillan Publishing, New York, New York, USA. Harris, S. W., and W. H. Marshall. 1963. Ecology of water-level manipulations on a northern marsh. Ecology 44:331-343. Herrick, B. M., M. D. Morgan, and A. T. Wolf. 2007. Seed banks in diked and undiked Great Lakes coastal wetlands. American Midland Naturalist 158:191-205. Herrick, B. M., and A. T. Wolf. 2005. Invasive plant species in diked vs. undiked Great Lakes wetlands. Journal of Great Lakes Research 31 :277-287. Johnson, D. H. 2008. In defense of indices: the case of bird surveys. Journal of Wildlife Management 72:857-868. Johnson, D. L., W. E. Lynch, and T. W. Morrison. 1997. Fish communities in a diked Lake Erie wetland and an adjacent undiked area. Wetlands 17:43-54. Jude, D. J ., and J. Pappas. 1992. Fish utilization of Great Lakes coastal wetlands. Journal of Great Lakes Research 18:651-672. Kadlec, J. A. 1962. Effects of a drawdown on a waterfowl impoundment. Journal of Wildlife Management 43 2267-28 1 . Kadlec, J. A., and L. M. Smith. 1992. Habitat management for breeding areas. Pages 590-610 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. 1981a. Dabbling duck activity and foraging responses to aquatic macroinvertebrates. The Auk 98:115-126. 87 Kaminski, R. M., and H. H. Prince. 1981b. Dabbling duck and aquatic macroinvertebrate responses to manipulated wetland habitat. Journal of Wildlife Management 45:1-15. Keddy, P. A., and A. A. Reznicek. 1986. Great Lakes vegetation dynamics: the role of fluctuating water levels and buried seeds. Journal of Great Lakes Research 12:25- 36. Keough, J. R., T. A. Thompson, G. R. Guntenspergen, and D. A. Wilcox. 1999. Hydrogeomorphic factors and ecosystem responses in coastal wetlands of the Great Lakes. Wetlands 19:821-834. Kincaid, C. 2005. Guidelines for selecting the covariance structure in mixed model analysis. Paper 198-30 in Proceedings of the Thirtieth Annual SAS® Users Group International Conference. SAS Institute, 10-13 April 2005, Philadelphia, Pennsylvania, USA. Krieger, K. A., D. M. Klarer, R. T. Heath, and C. E. Herdendorf. 1992. A call for research on Great Lakes coastal wetlands. Journal of Great Lakes Research 18:525-528. Legendre, P., and L. Legendre. 1998. Numerical ecology. Second English edition. Elsevier, Amsterdam, The Netherlands. Littell, R. C., G. A. Milliken, W. W. Stroup, and R. D. Wolfinger. 1996. SAS® system for mixed models. SAS Institute, Cary, North Carolina, USA. Lofgren, B. M., F. H. Quinn, A. H. Clites, R. A. Assel, A. J. Eberhardt, and C. L. Luukkonen. 2002. Evaluation of potential impacts on Great Lakes water resources based on climate scenarios of two GCMs. Journal of Great Lakes Research 28: 537-554. Markham, C. A., W. E. Lynch, Jr., D. L. Johnson, and R. W. Petering. 1997. Comparison of white crappie populations in diked and undiked Lake Erie wetlands. Ohio Journal of Science 97:72-77. Marks, M., B. Lapin, and J. Randall. 1994. Element stewardship abstract for Phragmites australis: threats, management, and monitoring. Natural Areas Journal 14:285- 294. McGarigal, K., S. Cushman, and S. Stafford. 2000. Multivariate statistics for wildlife and ecology research. Springer, New York, New York, USA. McLaughlin, D. B., and H. J. Harris. 1990. Aquatic insect emergence in two Great Lakes marshes. Wetlands Ecology and Management 1:111-121. 88 McNicholl, M. K., P. E. Lowther, and J. A. Hall. 2001. Forster’s Tern (Sternaforsteri). Account 595 in A. Poole and F. Gill, editors. The birds of North America. The Academy of Natural Sciences, Philadelphia, Pennsylvania, USA. Meyer, S. W. 2003. Comparative use of Phragmites australis and other habitats by birds, amphibians, and small mammals at Long Point, Ontario. Thesis, University of Western Ontario, London, Canada. Moore, D. R. J ., P. A. Keddy, C. L. Gaudet, I. C. Wisheu. 1989. Conservation of wetlands: do infertile wetlands deserve a higher priority? Biological Conservation 47:203-217. Mortsch, L., H. Hengeveld, M. Lister, B. Lofgren, F. Quinn, M. Slivitzky, and L. Wenger. 2000. Climate change impacts on the hydrology of the Great Lakes-St. Lawrence system. Canadian Water Resources Journal 25:153-179. Mortsch, L., J. Ingram, A. Hebb, and S. Doka, editors. 2006. Great Lakes coastal wetland communities: vulnerability to climate change and response to adaptation strategies. Environment Canada and the Department of Fisheries and Oceans, Toronto, Ontario, Canada. Murkin, H. R., R. M. Kaminski, and R. D. Titrnan. 1982. Responses by dabbling ducks and aquatic invertebrates to an experimentally manipulated cattail marsh. Canadian Journal of Zoology 60:2324-2332. Pierotti, R. J ., and T. P. Good. 1994. Herring Gull (Larus argentatus). Account 124 in A. Poole and F. Gill, editors. The birds of North America. The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Omithologists’ Union, Washington, DC, USA. Prince, H. H. 1985. Avian communities in controlled and uncontrolled Great Lakes wetlands. Pages 99-119 in H. H. Prince and F. M. D’Itri, editors. Coastal wetlands, Proceedings of the first Great Lakes Coastal Wetlands Colloquium. Lewis Publishers, Chelsea, Michigan, USA. Prince, H. H., and C. S. Flegel. 1995. Breeding avifauna of Lake Huron. Pages 247-272 in M. Munawar, T. Edsall, and J. Leach, editors. The Lake Huron ecosystem: ecology, fisheries, and management. SPB Academic Publishing, Amsterdam, The Netherlands. Prince, H. H., P. I. Padding, and R. W. Knapton. 1992. Waterfowl use of the Laurentian Great Lakes. Journal of Great Lakes Research 18:673-699. Provence, C. D. 2008. Effects of diking and plant zonation on invertebrate communities of Lake St. Clair coastal marshes. Thesis, Michigan State University, East Lansing, USA. 89 Riffle, S. K., B. E. Keas, and T. M. Burton. 2001. Area and habitat relationships of birds in Great Lakes coastal wet meadows. Wetlands 21:492-507. Robel, R. J ., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. Journal of Range Management 23:295-297. Roman, C. T., W. A. Niering, and R. S. Warren. 1984. Salt marsh vegetation change in response to tidal restriction. Environmental Management 8: 141-150. Ryder, J. P. 1993. Ring-billed Gull (Larus delawarensis). Account 33 in A. Poole, P. Stettenheim, and F. Gill, editors. The birds of North America. The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Omithologists’ Union, Washington, DC, USA. SAS Institute. 2004. SAS OnlineDoc® 9.1.3. SAS Institute, Cary, North Carolina, USA. Steen, D. A., J. P. Gibbs, and S. T. A. Timmermans. 2006. Assessing the sensitivity of wetland bird communities to hydrologic change in the eastern Great Lakes region. Wetlands 26:605-611. Thiet, R. K. 2002. Diversity comparisons between diked and undiked coastal freshwater marshes on Lake Erie during a high-water year. Journal of Great Lakes Research 28:285-298. Timmermans, S. T. A., S. S. Badzinski, and J. W. Ingram. 2008. Associations between breeding marsh bird abundances and Great Lakes hydrology. Journal of Great Lakes Research 34:351-364. Tucker, G. C. 1990. The genera of Arundinoideae (Gramineae) in the southeastern United States. Journal of the Arnold Arboretum 71:145-177. Tulbure, M. G., C. A. Johnston, and D. L. Auger. 2007. Rapid invasion of a Great Lakes coastal wetland by non-native Phragmites australis and T ypha. Journal of Great Lakes Research 33 (Special Issue 3):269—279. Wagner, T., D. B. Hayes, and M. T. Bremigan. 2006. Accounting for multilevel data structures in fisheries data using mixed model. Fisheries 31 :180-187. Weller, M. W. 1961. Breeding biology of the Least Bittem. Wilson Bulletin 73:11-35. Weller, M. W., and L. H. Fredrickson. 1974. Avian ecology of a managed glacial marsh. Living Bird 12:269-291. 90 Weller, M. W., and C. S. Spatcher. 1965. Role of habitat in the distribution and abundance of marsh birds. Department of Zoology and Entomology Special Report 43, Agricultural and Home Economics Experiment Station, Iowa State University, Ames, USA. Whitman, W. R. 1976. Impoundments for waterfowl. Canadian Wildlife Service Occasional Paper 22, Ottawa, Ontario, Canada. Whitt, M. B. 1996. Avian breeding use of coastal wetlands on the Saginaw Bay of Lake Huron. Thesis, Michigan State University, East Lansing, USA. Wilcox, D. A. 1993. Effects of water-level regulation on wetlands of the Great Lakes. Great Lakes Wetlands 421-2, 11. Wilcox, D. A. 1995. The role of wetlands as nearshore habitat in Lake Huron. Pages 223-245 in M. Munawar, T. Edsall, and J. Leach, editors. The Lake Huron ecosystem: ecology, fisheries, and management. SPB Academic Publishing, Amsterdam, The Netherlands. Wilcox, D. A., J. A. Meeker, and J. Elias. 1993. Impacts of water-level regulation on the wetlands of the Great Lakes. Phase 2 Report to Working Committee 2, International Joint Committee Water-levels Reference Study, Ottawa, Ontario, Canada, and Washington, DC, USA. Wilcox, D. A., and T. H. Whillans. 1999. Techniques for restoration of disturbed coastal wetlands of the Great Lakes. Wetlands 19:835-857. Wilson, S. D., and P. A. Keddy. 1986. Species competitive ability and position along a natural stress/disturbance gradient. Ecology 67:1236-1242. Wisheu, I. C., and P. A. Keddy. 1992. Competition and centrifugal organization of plant communities: theory and tests. Journal of Vegetation Science 3: 147-156. Yocum, B. J. 2007. Breeding biology of and nest site selection by Least Bittems (Ixobrychus exilis) near Saginaw Bay, Michigan. Thesis, Central Michigan University, Mount Pleasant, USA. 91 CHAPTER 3 MIGRANT BIRD USE OF DIKED AND UNDIKED COASTAL WETLANDS IN MICHIGAN INTRODUCTION Great Lakes coastal wetlands provide vital breeding, migration, and wintering habitat for an array of birds. Approximately three million swans, geese, and ducks travel along migration corridors that cross the Great Lakes region (Great Lakes Basin Commission 1975, Bellrose 1980). Great Lakes coastal wetlands are also valuable stopover habitats for migrant shorebirds that breed in the boreal and arctic regions of North America (Brown et al. 2000). These wetlands are some of the region’s largest remaining emergent marshes and provide vital nesting habitat to wetland birds, including rare and declining species such as American Bittern (Botaurus lentiginosus), Least Bittem(1xobrychus exilis), Common Moorhen (Gallinula chloropus), King Rail (Rallus elegans), Black Tern (Chlidonias niger), and F orster’s Tern (Sternaforsteri). Prince and Flegel (1995) summarized breeding bird atlas data from Michigan and Ontario. Eighty bird species used coastal wetlands of Lake Huron as breeding habitat (Prince and Flegel 1995) Impoundments control structures have long been used by wildlife managers to enhance wetlands for wildlife (Kadlec 1962), especially breeding and migrating waterfowl. Impounded wetlands are typically managed as hemi-marshes to maximize breeding bird use or shallow-water marshes dominated by moist-soil vegetation to attract 92 migrant birds (Weller and Spatcher 1965, Fredrickson and Taylor 1982, Murkin et al. 1997). Hemi-marshes are marshes with approximately equal proportions of emergent vegetation and open water produced by natural water level fluctuations and mammal herbivory. Historically, Great Lakes coastal wetlands moved landward and lakeward with the rise and fall of the Great Lakes. Between the 19503 and 19705, many Great Lakes coastal marshes were isolated from these normal water level fluctuations through dike construction. These projects were initiated primarily to maintain elevated water depths and enhance wildlife use during periods of historic low water levels. Shoreline armoring, wetland diking and tiling to drain wetlands for agricultural use, and other land- use changes now prevent the landward movement of coastal wetlands in much of the Great Lakes (Prince et al. 1992, Gottgens et al. 1998). The potential problems associated with isolating coastal wetlands from the Great Lakes include impaired or eliminated flood conveyance and storage, sediment control, and water quality improvement functions, altered nutrient flow, reduced or degraded habitat for shorebirds, rare species, fish, and invertebrates, and increased impacts from trapped Carp (Cyprinus carpio) (Jude and Pappas 1992, Wilcox 1995, Wilcox and Whillans 1999). By separating coastal wetlands from the fluctuations of the Great Lakes, dike construction often stabilizes water levels. Stable water levels typically compress wetland vegetation zones and encourage dominance by shrubs and highly competitive species, such as willow (Salix spp.), alder (Alnus spp.), cattail (Typha spp.), reed canary grass (Phalaris arundinacea), and purple loosestrife (Lythrum salicaria). Irregular water levels may result in higher levels of diversity both within and among habitats (Keddy and Reznicek 1986, Wilcox 1993, Wilcox et al. 1993, Keough et al. 1999). 93 Comparisons of plant communities in diked and undiked Great Lakes coastal wetlands have yielded varied results. Herrick and Wolf (2005) documented increased amounts of invasive species in standing vegetation and seed banks of diked compared to undiked wetlands in Saginaw Bay, Michigan and Green Bay, Wisconsin, but noted that current conditions in undiked wetlands appear to favor an invasive haplotype of common reed (Phragmites australis). Conversely, Galloway et al. (2006) found greater species richness and percent cover of native species and lower species richness and percent cover of invasive species in diked compared to undiked coastal wetlands. Herrick et al. (2007) found more seeds from a greater number of species in the soils of diked compared to undiked wetlands and stated that diked wetlands may serve as “traps” for plant seeds. In comparisons between vegetation in diked and undiked Lake Erie coastal wetlands during a high water year, Thiet (2002) found greater wetland plant diversity in diked wetlands compared to a nearby undiked site. An actively managed diked marsh in southwest Lake Erie maintained emergent vegetation, patchiness, and edge habitat similar to historic conditions during periods of high Great Lakes water levels, while the same measures declined in marshes connected to Lake Erie (Gottgens et al. 1998). Research conducted by several authors on animal use of Great Lakes coastal wetlands provides insights into the possible effects of diking on animal communities. McLaughlin and Harris (1990) compared aquatic insect emergence in one diked and one undiked wetland on Green Bay, and recorded more insect taxa and greater total insect biomass emergence from the diked wetland. Burton et al. (2002) noted that both plant- community composition and exposure to wave action were important determinants of ' invertebrate diversity and biomass in Great Lakes marshes. Invertebrates were 94 distributed along gradients of decreased mixing of pelagic water and increased sediment organic matter from outer to inner marsh and between littoral and adjacent inland marshes. Some invertebrates were more common on one end of these gradients, but most species were generalists found across all habitat types (Burton et al. 2002). Whitt’s (1996) study of avian breeding use of Saginaw Bay coastal wetlands included study sites that were both open to and inland from Lake Huron. Although species richness was similar between coastal and inland cattail marshes, bird densities in marshes located far offshore were lower than most other sites. Galloway et al. (2006) conducted a one-year study of breeding bird use of diked and undiked Great Lakes coastal wetlands along Lakes Ontario, Erie, and St. Clair. In pooled comparisons of diked and undiked sites, they observed greater abundance and species richness for several groups of birds in diked wetlands, but indicated that long-term research is needed to account for long-term variation in bird and vegetation communities associated with Great Lakes water level cycles and management activities. No research has been conducted in the Great Lakes region to assess the effects of coastal wetland diking on bird communities during migration periods. Ecological studies of the effects of coastal wetland isolation from natural, highly variable water level fluctuations are needed so that informed decisions can be made about the management and restoration of Great Lakes wetlands. The goal of this project was to evaluate the effects of coastal wetland diking on migrant birds by comparing bird use and vegetation and physical conditions of several diked and undiked wetlands in Michigan. I tested the hypothesis that coastal impoundments with managed water levels support greater densities and more species of migrant wetland birds compared to undiked 95 wetlands. This research is one of many comparisons needed over the long-term to better understand how diked and undiked wetlands function during the full cycle of Great Lakes water levels. 96 METHODS Aerial Waterfowl Stu'veys Fourteen aerial waterfowl surveys were conducted in spring (n=5), late summer (n=5), and early fall (n=4) during 2005-2007 to evaluate staging and migrant waterfowl use of three St. Clair Flats and 12 Saginaw Bay study sites (Figure 12). Fall surveys were not attempted after duck hunting seasons began in early- to mid- October due to changes in waterfowl behavior and habitat use. The first survey conducted in fall 2005 was done using a MD-500 helicopter and traversed 22 transects (12 diked, 10 undiked) totaling approximately 76 km (21 km diked, 55 km undiked) in length (Table 13). Beginning in spring 2006, aerial surveys were done using a Cessna 172N fixed-wing aircrafi, which was more cost efficient and had a faster flight speed better suited to surveying large flocks of waterfowl that often flushed ahead of the aircraft. Sixteen transects (8 diked, 8 undiked) totaling 66.6 km (18.7 km diked, 47.9 km undiked) in length were surveyed during subsequent surveys with fixed-wing aircraft (Table 13). Methods used were similar to the standard operating procedures used for breeding surveys (U .8. Fish and Wildlife Service/Canadian Wildlife Service 1987). Transects were flown at slow speeds of about 130 — 200 km/h (approximately 80 — 125 mph) at an altitude of approximately 30 — 45 in (about 100 — 150 ft). One observer sat on each side of the aircraft and counted all waterfowl within 200 m for a total transect width of 400 m. Other waterbirds that could be identified from the air (e. g., Great Blue Heron [Ardea herodias], Great Egret [Ardea alba], American Coot [F ulica americana]) were also recorded. Transects crossing impounded wetlands were situated along the longest axis and approximately 97 .momofiaousa E 3235 0.8 838 can ~on 5 com: massage? Swan: 5 88-88 mesa Beta 96:: 0328 38mg gorge? Eton. 393: 83.: mam Bufiwam use 320 um 835 32m :20 am we £83304 .N_ gamma an“: .288 .5 an“. 239 6332530 1. A 2%: act gems. _. .3”. 5.6.3. 838 .E :6: Em“: fiaz .maz .zaz .mazy... Amoaom .E 9.50:ng . . . .E fir. \ \fifidx . . 3.5 am _ I . ’ _2’OhU__>> Az_n_v OC_CCCOC_Q fl. . Emzc 6.8.5 . EEC 23 89535 Loam _. mEn. - 1."... 22$ \ smas— 1 comm ‘ . w . e. H . 2.25 2.0.5.2 “mm fie .. >mm 55mm. 4. E 82265 ”it _, 98 Q: «.2 wfim 5.: 5.: 5.: v.0 v.0 Wm as as wH vI-lv—Iv—fl MM") mdmm adcv oéNZ Ndmm md—m mdmm @694 cans mémm We: ono— @630 oéN— 9mm— md—m WNN q—rv—A I—tv—t v—Iv—r w-nv—n omS 03. SN wmm w; Z Z Z Z Z Wm Wm vé Z. I. Z. S m.mm mag 1. 03 Mag .1 #62 mag 5R mom mom mm m m 0 o.— o._ o; NMN an m 2 5 000000 0 O O v—(I—v— F-lQ—Iv—i I O O v-dr-IF-fl 0:0 No.2: Know 99» 0:4 No.2: 52. ofiv m. :0 wdo wdo 5.2. o.mv xv mm X: 2: mm 0 O A O m omm Z fim mm Nd md m.N md ARON fimom mdm— v-‘NN I—I wvw Z < mam Wm m.N 05 ad»: «:3 ad? GS 98 mom omm < meow meow meow Eco ems”: 8355. Boom ooom moom acacia—m A65 .82 noon coca moon 308ch Mo 8&8: Z magnum garage? 3.51% «0.2 23:03 onmeomex .EofiowmcmE c335 dB <20 ma 2E PE Gm}; mOH HmZ mmZ ZmZ mEZ MEL «in 33: En— g? gm 68:2: woo—5 68:65 6825 mam Bacmmmm mam—m “EU am .hoommoom mama—v 3:283 388 gwfiog chem Bmcmwmm can age—m .520 am “a 9333 EB “gamma wagw @838 no.8 33:53 use 60325 mousotfloomnmb mo 838:: £352qu Eofiowgmfi 883 £98 :38 Begxoafiz .2 033. 99 £80ch wgéoxm .Zo bZZZn—Eogoaafi Z3268 Ea .bZEZxoa £032 95% 3 26 ZnZZ use mEZ 8m 3:388 $0892? .aoZZnZZcomch .8.Z :oZZoom mace roam com .ocoZuZ Z38 633$an .oZZmEBcoanZOuO .o>ZZo<1 ><><><><><><>< ><>< ><><>< ><><><><><><><><><><><><><><><><>< 115 Table 17. Cont’d. Species Diked Common Open Wetland-dependent Species Greater Yellowlegs Lesser Yellowlegs Semipalmated Sandpiper Least Sandpiper Baird’s Sandpiper X Pectoral Sandpiper Dunlin X Stilt Sandpiper Short-billed Dowitcher Wilson’s Snipe American Woodcock Red-necked Phalarope Bonaparte’s Gull X Ring-billed Gull Herring Gull Black Tern Forster's Tern Belted Kingfisher XX XXX >< ><><><>< ><><><><>< Wetland-associated Species Bald Eagle Merlin X Black-bellied Plover X Killdeer X Caspian Tern X Common Tern X X Total Number of Species 7 46 7 116 The first dimension of the correspondence analysis explained 45.0% of the variation and the second dimension 24.0% of the variation in bird densities among the sites during early fall migration ground surveys. Correspondence analysis separated the bird species/ groups into two groups along the first dimension, with coots/moorhens alone on the negative end and the remaining groups clumped from approximately zero to about 0.55 on the positive end (Figure 13). Dabbling ducks, coots/moorhens, and shorebirds had negative coordinates in the second dimension, while all other bird species/groups had positive coordinate values, with Wood Duck and bittems having the greatest values (Table 18). Diked and undiked sites were largely separated along the second dimension. Undiked wetlands formed two groups on the negative end of dimension two: the first consisting of undiked St. Clair Flats sites associated with greater American Coot and Common Moorhen densities, and the second made up of undiked Saginaw Bay sites that appeared related to greater dabbling duck and shorebird abundance compared to other sites (Figure 13). A small group of diked Saginaw Bay sites also had negative dimension two values and were clumped with the undiked Saginaw Bay wetlands, indicating greater dabbling duck and shorebird densities compared to other diked sites. Many of the diked wetlands appeared to be associated with greater Wood Duck and bittem densities than other sites (Figure 13). 117 db— —1— - Bl + GR + * RA Dimension 2 (24.0%) G) (n +*£ rn 0*0 IL —2 -1 O Dimension 1 (45.0%) Figure 13. Biplot of site and bird group coordinates for dimensions 1 and 2 from correspondence analysis conducted using fall migration ground survey data collected at St. Clair Flats (SCF) and Saginaw Bay (SAG), Michigan, coastal wetlands, 2005- 2007. Site coordinates are coded by wetland type (“+” diked; “O” undiked). Bird group coordinates are coded with an “"'” and labeled as follows: BI = bitterns, CM = American Coots and Common Moorhens, DA = dabbling ducks, D1 = diving ducks, GR = Pied-billed Grebes, GS = Canada Geese and swans, HE = herons, RA = rails, SH = shorebirds, T0 = terns and gulls, and WD = Wood Ducks. 118 Table 18. First and second dimension coordinates for birds species/ groups included in correspondence analysis conducted using data from 45 fall migration ground surveys done along 21 routes at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bird Species/Group Dimension 1 Dimension 2 Dabbling Ducks (DA) 0.4535 -0.5188 Diving Ducks (DI) 0.4854 0.3201 Canada Geese/Swans (GS) 0.2166 0.6395 Wood Ducks (WD) 0.1549 1.2528 Pied-Billed Grebes (GR) -0.0921 0.8069 Bittems (BI) 0.1721 1.1045 Herons (HE) 0.3419 0.5018 American Coots/Common Moorhens (CM) -1.6618 -0.l646 Rails (RA) 0.2807 0.6834 Shorebirds (SH) 0.5486 -0.0087 Tems/Gulls (TG) 0.1874 0.2280 Vegetation and Physical Variable Sampling Most vegetation and physical variables measured along fall ground survey routes were similar between diked and undiked wetlands (Table 19). I found similar mean percent cover estimates between diked and undiked wetlands for emergent, open water/aquatic bed, and submersed vegetation. I observed greater mean percent cover of floating vegetation (p<0.0001) in diked compared to undiked wetlands. Mean percent cover of cattail was greater in diked wetlands (p=0.0015), while average percent cover of bulrush was greater in undiked wetlands (p<0.0001). The number of plots with organic and inorganic soils differed between diked and undiked sites (p<0.0001). Most plots within diked wetlands were dominated by organic soils, while undiked wetlands that largely consisted of inorganic soils, such as sand or silt (Table 19). 119 Table 19. Least squares means and lower and upper 95% confidence limits (CL) for vegetation and habitat variables measured during three-m2 plot sampling conducted during fall ground surveys for birds at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Bolded p-values indicate a significant difference between wetland types (p<0.05). Diked Undiked Lower Upper Lower Upper Bird Density Variable Mean CL CL Mean CL CL P-value Percent Cover No. Samples 571 497 «- Emergent Vegetation 51.0 40.5 61.5 48.6 38.1 59.3 0.2646 Open Water/Aquatic Bed 44.1 31.8 56.8 49.2 36.1 62.4 0.3645 Submersed Vegetation 17.4 7.3 30.6 6.8 0.9 17.6 0.0907 Floating Vegetation 14.0 9.3 19.4 0.1 -0.3 1.2 <0.0001 Persistent Deep-water 22.3 12.0 34.6 22.2 10.9 36.0 0.9907 Persistent Shallow-water 5.7 1.9 11.4 3.6 0.6 9.0 0.4790 Non-persist. Deep—water 0.4 <0. 1 2.4 1 .2 <0.1 4.1 0.5264 Non-persist. Shallow-water 1.2 0.2 3.0 0.8 <0.1 2.6 0.6823 Cattail 18.1 9.0 29.5 1.5 0.1 7.3 0.0015 Bulrush 0.2 0.1 1.2 14.9 10.1 20.5 <0.0001 Common Reed 1.6 0.4 3.6 2.7 0.8 5.5 0.4144 Surface Litter <0.1 -0.3 0.2 0.1 <0.1 0.8 0.1699 Exposed Sediments 0.8 <0.] 3.3 0.2 0.4 1.9 0.2703 Sediment Type (Proportion of Total Samples) No. Samples 570 497 NA Organic 94.2 --- --- 5.8 --- --- 12 Test Inorganic 9.0 --- --- 91.0 --- --- <0.0001 Water Depth No. Samples 1666 1527 NA Depth (m) 0.27 0.15 0.41 0.20 0.07 0.35 0.4440 120 The first component from the PCA explained 27.6% of the variation in vegetation and physical variables among the fall migration ground survey routes, while the second component explained 21.3% of the variation. Principal component 1 (PC 1) primarily represented a gradient from bulrush and common reed marsh with low percent cover of floating vegetation to cattail marsh with high levels of floating vegetation (Figure 14). The first principal component was positively related to percent cover of cattail, floating vegetation, and persistent deep-water emergents, and negatively related to percent cover of bulrush, common reed, and persistent shallow-water emergents (Table 20). Principal component 2 (PC 2) represented a gradient from deep-water open wetlands dominated by submersed and non-persistent deep-water vegetation to shallower marshes dominated by deep-water persistent emergents (Figure 14). The second PC was positively related to percent cover of submersed vegetation, non-persistent deep-water emergents, and open water/aquatic bed, and water depth, and negatively related to persistent deep-water emergents (Table 20). Diked and undiked routes were largely separated along the first axis, with diked wetlands tending to have higher PC 1 scores compared to undiked wetlands (Figure 14). Diked wetland survey routes tended to be dominated by cattail with greater percent cover of floating vegetation compared to undiked sites, while undiked routes usually were dominated by bulrush and common reed. Diked and undiked survey routes had similar scores for PC 2, indicating similar variation in percent cover of submersed vegetation, non-persistent and persistent deep-water emergents, and open water/aquatic bed wetland, and water depths along the open water-emergent marsh interface (Table 20). 121 ‘- -h - % Submersed I I I I I % Non-persist. Deep 5 Z -1. % Open Water Water Depth 4 __ 0 + __ 3 J- + + 2 -*- + $3 00 o + + + + :3 1 -- O o -- Q o 0 + ++ + + N _,_ O O to + -- O 0 O + + D. + _1 __ 0 i + __ o _ i — 2 - o + -- 0 db _ 3 .1— O O .1.— — 4 W- .1- % Persistent Deep . 1 L 1 1 L . L I I I I I I I I -3 — 2 — 1 0 1 2 3 4 PC 1 (27.6%) % Bulrush % Cattail % Common Reed ‘—'—> % Floating % Persistent Shallow % Persistent Deep Figure 14. Bi-plot of PC 1 X PC 2 from principal components analysis conducted using 13 vegetation and physical variables gathered during plot sampling during 45 fall migration bird surveys of 21 routes at St. Clair Flats and Saginaw Bay, Michigan, 2005- 2007. Point scores are coded by wetland type (“+” diked; “O” undiked). 122 Table 20. Eigenvectors for first two principal components obtained through PCA of habitat data collected during 45 fall migration ground surveys conducted along 21 routes at St. Clair Flats and Saginaw Bay, Michigan, coastal wetlands, 2005-2007. Habitat Variable Percent Cover Emergent Vegetation Open Water/Aquatic Bed Submersed Vegetation Floating Vegetation Persistent Deep-water Persistent Shallow-water Non-persistent Deep-water Non-persistent Shallow-water T ypha Schoenoplectus Phragmites australis Exposed Sediments Water Depth (m) Principal Component 1 0.0319 0.0598 0.1463 0.4042 0.3402 -0.3096 -0.0992 -0.l689 0.4725 -0.3692 -0.3542 -0.0483 0.2707 123 Principal Component 2 0.0536 0.3443 0.5037 0.1539 -0.3426 0.2207 0.3947 0.0982 -0.0868 -0.2292 0.1781 -0.2636 0.3293 DISCUSSION Bird Use of Diked and Undiked Wetlands I observed few differences in migrant waterfowl densities between diked and undiked coastal wetlands during aerial surveys. Most estimated mean areal densities for waterfowl species from early fall ground surveys were also similar between wetland types, but mean linear densities of dabbling ducks and Mallards were greater in undiked than diked wetlands. No studies evaluating migrant bird use of diked and undiked coastal wetlands were found for the Great Lakes region. Brasher et a1. (2007) found that duck foraging resources were abundant during fall in both actively (i.e., with water-level control) and passively (i.e., no water-level control) managed wetlands in Ohio. I did not measure food resources in this study, but greater linear densities of dabbling ducks and Mallards in undiked wetlands could be due to more abundant foods and/or shallower water depths that provided better access to foods. Water depths along the emergent-open water interface were comparable between diked and undiked wetlands, but depths tended to be shallower at undiked sites. Water depths measured at randomly selected quadrats during the breeding season were greater at diked compared to undiked sites (see Chapter 2). Fredrickson and Taylor (1982) indicated that preferred foraging depths for Mallards in seasonally flooded impoundments was approximately 10—1 5 cm, so the shallow water depths observed in the undiked wetlands could provide better foraging habitat than diked wetlands. Canada Goose densities were greater in undiked wetlands during aerial surveys, while fall ground surveys revealed similar densities in diked and undiked wetlands. This discrepancy may be due to seasonal changes in Canada Goose densities and habitat use, high variation of densities during migration, and the low number of aerial 124 surveys conducted. Large flocks (i.e., >100 individuals) of Canada Geese were observed on undiked Saginaw Bay wetlands during spring aerial surveys. Canada Geese probably used these wetlands as roosting sites and flew to other locations (e. g., agricultural lands) to forage, so the primary determinant of habitat selection may have been secure roosting areas. Wood Duck densities were greater in diked than undiked wetlands during both aerial and early fall ground surveys. I observed the same pattern of Wood Duck densities during the breeding season, and greater use of diked sites could be related to cover provided by dense floating-leaved vegetation. My surveys indicated fall migration shorebird use of diked and undiked coastal wetlands was similar. Wilson’s Snipe was the only shorebird species observed in greater densities in diked than undiked wetlands; greater densities of Wilson’s Snipe in diked wetlands may have been related to the prevalence of organic soils and high invertebrate abundance (Provence 2008). Linear densities of Greater Yellowlegs were greater in undiked compared to diked wetlands, which was the only shorebird density variable observed in greater abundance in undiked wetlands. Ring-billed Gull and Forster’s Tern densities during fall migration ground surveys were greater in undiked than diked wetlands, which is the same pattern observed during breeding surveys of the same sites (see Chapter 2). Fish are an important component of the diets of both species (see Ryder 1993, McNicholl et al. 2001), and studies conducted in Lake Erie coastal wetlands indicated differences in total fish species richness and abundance, age class frequencies, lengths, and body condition indices for some species between diked and undiked wetlands (Johnson et al. 1997, Markham et al. 1997). I did not measure fish abundance and composition in my study, but it would be useful to know 125 the relative abundance of forage fish to understand the effects coastal wetland diking on these bird species. Foraging in diked wetlands may have been more difficult for these species due to greater coverage of floating-leaved vegetation compared to undiked wetlands. Vegetation and Physical Characteristics of Diked and Undiked Wetlands Plot sampling along ground survey routes indicated some differences in vegetation and physical variables at the open water-emergent interfaces of diked and undiked wetlands. I observed greater percent cover of floating vegetation and cattails in diked compared to undiked wetlands, and these differences were likely due to higher, more stable water levels in diked sites. Although water levels of diked wetlands ofien dropped dramatically during the summer, the majority of the wetlands remained inundated throughout the season. Percent cover of bulrush was greater in undiked than diked wetlands. Intensive quadrat sampling during the breeding season revealed similar differences in vegetation and physical characteristics of diked and undiked wetlands (see Chapter 2). Albert and Brown (2008) observed similar results when comparing the vegetation at several of the same diked and undiked study sites. Principal components analysis of the vegetation and physical data provided analogous results to parametric comparisons. Diked and undiked survey routes were primarily separated along the first axis, which indicated that diked routes were usually dominated by cattail marsh and had greater percent cover of floating vegetation than undiked sites, while undiked routes were dominated by bulrush and common reed. In vegetation comparisons between diked and undiked wetlands, Herrick and Wolf (2005) similarly found greater cattail cover in diked 126 wetlands and greater common reed cover in undiked wetlands. Lower mean percent cover of common reed in diked compared to undiked wetlands may be due to higher water levels and common reed management (e.g., herbicide application, burning) that occurred in some diked areas. In North America, cattails outcompete other plant species in wetlands with high fertility and low disturbance (Moore et al. 1989, Wisheu and Keddy 1992). Diked wetlands likely experience less disturbance than undiked sites due to higher water levels and infrequent complete drawdowns, and greater fertility due to high organic content of soils and trapped nutrients. Herrick et al. (2007) stated that diked coastal wetlands appeared to serve as traps for organic matter and nutrients. Management Implications Despite some differences in habitat, migrant bird use of diked and undiked wetlands was largely similar. A common criticism of diked coastal wetlands is that their management tends to focus on waterfowl or game species, potentially at the detriment of rare and/or non-game bird species. The results of my study do not support this criticism. Forster’s Tern (State special concern) was the only rare species observed in greater densities in undiked than diked wetlands. In addition to Forster’s Tern, I observed 13 bird species of greatest conservation need (SGCN, Eagle et al. 2005) often enough to permit statistical comparisons between diked and undiked wetlands. Mean American Black Duck density was greater in undiked wetlands, while Great Blue Heron and Wilson’s Snipe densities were greater in diked sites; the remaining 10 SGCN were similar between diked and undiked wetlands. Species richness was also similar between 127 the two wetland types, and similarity indices indicated that the bird communities of diked and undiked wetlands were comparable. Wilcox (1995) suggested that shorebird habitat provided by continually changing Great Lakes water levels may be lost when coastal wetlands are isolated through diking. My observations during fall migration revealed that shorebird use was similar between diked and undiked wetlands. Although water depths tended to be higher in diked than undiked wetlands, water levels were usually lowest in late summer, which provided pockets of mudflats and shallow water at a time when fall shorebird migration typically peaks. Conversely, Lake St. Clair and Huron water levels are usually highest in late summer. I also observed that low water conditions in diked wetlands sometimes created mats of organic matter and submersed vegetation that shorebirds used for foraging. Given recent comparisons of invertebrate abundance and composition in diked and undiked wetlands of St. Clair Flats (see Provence 2008), these diked habitats likely had high abundances of invertebrate foods for shorebirds. In some cases, pumping to increase water levels in diked wetlands in preparation for fall waterfowl hunting reduced available habitat for migrant shorebirds. Minor alterations to water management schedules could enhance habitat for migrant shorebirds at a time when available shorebird habitat in coastal wetlands could be limited. Differences in migrant shorebird use of diked and undiked wetlands could be more prominent in spring. The high spring water levels of diked wetlands probably limit use by migrant shorebirds compared to undiked wetlands, which tend to have lower water levels in spring than summer. In a study of impoundments in Delaware Bay, Parsons (2002) observed greatest migrant shorebird abundance in impoundments with low spring water levels. Potter et al. (2007) 128 noted that migrant shorebird habitat may be more limited during the fall migration compared to spring, due to vegetation coverage. However, they assumed that spring was the habitat-limited season because the migration period is short and precedes the breeding season, therefore the timing of resource availability is most critical. Invasive populations of common reed have substantially expanded in Great Lakes coastal wetlands during the recent period of low water levels (Tulbure et al. 2007, E. Kafcas, Michigan Department of Natural Resources, person. commun.). Most climate change models predict decreasing Great Lakes water levels in the future (Mortsh et al. 2000, 2006, Lofgren et al. 2002, Croley 2003), which could firrther increase common reed expansion in undiked wetlands and potentially reduce the value of these habitats for birds species of management concern. Given that the future status of coastal wetlands is uncertain, diked wetlands may provide important management opportunities to maintain and improve habitats for wetland birds and reduce impacts from invasive plant species like common reed. Greater mean linear densities of total dabbling ducks and Mallards in undiked compared to diked wetlands was not predicted because they are focal management species. I did not evaluate food resources during the early fall migration season in this study, so it is unknown whether these differences are related to food availability, access, or other factors. F redrickson and Taylor (1982) indicated that the preferred foraging depths for Mallards, Blue-winged Teal, and Green-winged Teal ranged from about 10-20 cm in seasonally flooded impoundments. Managing the diked wetlands for shallower water depths could potentially increase dabbling duck use by providing preferred foraging conditions. According to correspondence analysis, two diked wetlands at Fish 129 Point State Wildlife Area appeared to be associated with greater dabbling duck abundance than other sites. One of these sites is only passively managed and does not have a water pump, while the second site has a pump that was rarely used in recent years due to low Lake Huron water levels. Both sites typically had lower water depths compared to other diked sites, especially in late summer when exposed mudflats were often present. Periodic complete drawdowns of the diked wetlands could potentially improve habitat conditions for migrant birds. Kadlec and Smith (1992) noted three potential benefits of drawdowns: nutrient release due to the decomposition of organic sediments, consolidation of loose sediments due to drying, and the germination and establishment of emergent vegetation, including annual species. Drawdowns could reduce the buildup of organic matter, release nutrients and stimulate plant growth, and improve vegetation and structural diversity of the diked marshes. Recommended frequencies for drawdowns have ranged from 5 to 7 years (Harris and Marshall 1963, Whitman 1976). Moist-soil management requires more frequent drawdowns (Fredrickson and Taylor 1982), but could be an effective means of producing plant foods attractive to dabbling ducks during fall and spring migration. Since drawdowns can encourage growth of invasive plant species (Fredrickson and Taylor 1982), I suggest close monitoring of the vegetation response if drawdowns are conducted. Research Needs This study occurred during a period of low Great Lakes water levels, and water level fluctuations and depths are known to influence bird use of wetlands (e.g., Weller 130 and Spatcher 1965, Steen et a1. 2006, Timmermans et al. 2008). Migrant bird use of diked and undiked coastal wetlands could be different from the results of this study during normal to high water levels. More study is needed during other parts of the Great Lakes water level cycle to investigate if patterns of bird use change under different hydrological conditions. Long-term studies are also needed to understand changes in Great Lakes coastal wetlands that occur over 5-20 years. Research is needed to understand the effects of structural differences (e. g., water depths, floating vegetation mats, interspersion) in the habitats of diked and undiked wetlands on migrant bird use. Since shorebird habitat is thought to be limiting during spring migration in the Great Lakes region (Potter et al. 2007), comparisons of shorebird use between diked and undiked coastal wetlands during spring are needed to assess management actions. Investigations are needed to determine the availability of plant and animal foods for migrant waterfowl, waterbirds, and shorebirds in diked and undiked coastal wetlands, and to examine if wetland bird density and diversity is linked to those food resources. Management guidelines need to be developed to maximize wildlife benefits in diked wetlands in the context of changing coastal wetland conditions associated with climate change and invasive species expansion, and for specific species of concern (e. g., game, threatened, endangered, SGCN). Diked wetlands provide opportunities to conduct experimental studies that test the success of water level management regimes (e. g., lower water levels, periodic drawdowns) for selected management goals (e. g., increased use by important migrant bird groups). For example, even though migrant dabbling ducks are often a focus of diked wetland management, linear densities tended to be greater in undiked than diked wetlands. Water levels could be experimentally lowered in diked 131 wetlands to evaluate if dabbling duck densities increase when preferred water depths for foraging are provided. 132 LITERATURE CITED Albert, D. A., and P. W. Brown. 2008. Analysis of vegetation in adjacent diked-undiked coastal wetlands. Michigan Natural Features Inventory Report 2008-14, Lansing, USA. American Omithologists' Union. 1998. Check-list of North American Birds. Seventh edition. American Omithologists' Union, Washington, DC, USA. Bellrose, F. C. 1980. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg, Pennsylvania, USA. Brasher, M. G., J. D. Steckel, and R. J. Gates. 2007. Energetic carrying capacity of actively and passively managed wetlands for migrating ducks in Ohio. Journal of Wildlife Management 71 :2532—2541. Brown, S. C., and C. R. Smith. 1998. Breeding season bird use of recently restored versus natural wetlands in New York. Journal of Wildlife Management 62:1480- 1491. Brown, S., C. Hickey, and B. Harrington, editors. 2000. The US. shorebird conservation plan. Manomet Center for Conservation Sciences, Manomet, Massachusetts, USA. Burton, T. M., C. A. Stricker, and D. G. Uzarski. 2002. Effects of plant community composition and exposure to wave action on invertebrate habitat use of Lake Huron coastal wetlands. Lakes and Reservoirs: Research and Management 7:255- 269. Cowardin, L. M., V. Carter, F. C. Golet, E. T. LaRoe. 1979. Classification of wetlands and deepwater habitats of the United States. US. Fish and Wildlife Service, Washington, DC, USA. Croley, T. E., II. 2003. Great Lakes climate change hydrologic impact assessment: International Joint Commission Lake Ontario-St. Lawrence River regulation study. National Oceanic and Atmospheric Administration Technical Memorandum GLERL-l26, Ann Arbor, Michigan, USA. Crowley, S., C. Welsh, P. Cavanaugh, and C. Griffin. 1996. Weighing — birds: habitat assessment procedures for wetland-dependent birds in New England. University of Massachusetts, Department of Forestry and Wildlife Management, Amherst, USA. 133 Eagle, A. C., E. M. Hay-Chmielewski, K. T. Cleveland, A. L. Derosier, M. E. Herbert, and R. A. Rustem, editors. 2005. Michigan's wildlife action plan. Michigan Department of Natural Resources. Lansing, USA. . Accessed 13 February 2009. F redrickson, L. H., and T. S. Taylor. 1982. Management of seasonally flooded impoundments for wildlife. US. Fish and Wildlife Service Resource Publication 148, Washington, DC, USA. Galloway, M., L. Bouvier, S. Meyer, J. Ingram, S. Doka, G. Grabas, K. Holmes, and N. Mandrak. 2006. Evaluation of current wetland dyking effects on coastal wetlands and biota. Pages 187-229 in L. Mortsch, J. Ingram, A. Hebb, and S. Doka, editors. Great Lakes coastal wetland communities: vulnerability to climate change and response to adaptation strategies. Environment Canada and the Department of Fisheries and Oceans, Toronto, Ontario, Canada. Gottgens, J. F., B. P. Swartz, R. W. Kroll, and M. Eboch. 1998. Long-term GIS-based records of habitat changes in a Lake Erie coastal marsh. Wetlands Ecology and Management 6:5-17. Great Lakes Basin Commission. 1975. Great Lakes Basin framework study. US. Department of Interior, Bureau of Sport Fisheries and Wildlife, Ann Arbor, Michigan, USA. Hair, J. F., Jr., R. B. Anderson, and R. L. Tatham. 1987. Multivariate data analysis. Second edition. MacMillan Publishing, New York, New York, USA. Harris, S. W., and W. H. Marshall. 1963. Ecology of water-level manipulations on a northern marsh. Ecology 44:331-343. Herrick, B. M., M. D. Morgan, and A. T. Wolf. 2007. Seed banks in diked and undiked Great Lakes coastal wetlands. American Midland Naturalist 158:191-205. Herrick, B. M., and A. T. Wolf. 2005. Invasive plant species in diked vs. undiked Great Lakes wetlands. Journal of Great Lakes Research 31 :277-287. Johnson, D. L., W. E. Lynch, and T. W. Morrison. 1997. Fish communities in a diked Lake Erie wetland and an adjacent undiked area. Wetlands 17:43-54. Jude, D. J ., and J. Pappas. 1992. Fish utilization of Great Lakes coastal wetlands. Journal of Great Lakes Research 18:651-672. Kadlec, .I. A. 1962. Effects of a drawdown on a waterfowl impoundment. Journal of Wildlife Management 43 2267-281 . 134 Kadlec, J. A., and L. M. Smith. 1992. Habitat management for breeding areas. Pages 590-610 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. Keddy, P. A., and A. A. Reznicek. 1986. Great Lakes vegetation dynamics: the role of fluctuating water levels and buried seeds. Journal of Great Lakes Research 12:25- 36. Keough, J. R., T. A. Thompson, G. R. Guntenspergen, and D. A. Wilcox. 1999. Hydrogeomorphic factors and ecosystem responses in coastal wetlands of the Great Lakes. Wetlands 19:821-834. Kincaid, C. 2005. Guidelines for selecting the covariance structure in mixed model analysis. Paper 198-30 in Proceedings of the Thirtieth Annual SAS® Users Group International Conference. SAS Institute, 10-13 April 2005, Philadelphia, Pennsylvania, USA. Legendre, P., and L. Legendre. 1998. Numerical ecology. Second English edition. Elsevier, Amsterdam, The Netherlands. Littell, R. C., G. A. Milliken, W. W. Stroup, and R. D. Wolfinger. 1996. SAS® system for mixed models. SAS Institute, Cary, North Carolina, USA. Lofgren, B. M., F. H. Quinn, A. H. Clites, R. A. Assel, A. J. Eberhardt, and C. L. Luukkonen. 2002. Evaluation of potential impacts on Great Lakes water resources based on climate scenarios of two GCMs. Journal of Great Lakes Research 28: 537-554. Markham, C. A., W. E. Lynch, Jr., D. L. Johnson, and R. W. Petering. 1997. Comparison of white crappie populations in diked and undiked Lake Erie wetlands. Ohio Journal of Science 97:72-77. Marks, M., B. Lapin, and J. Randall. 1994. Element stewardship abstract for Phragmites australis: threats, management, and monitoring. Natural Areas Journal 14:285- 294. McGarigal, K., S. Cushman, and S. Stafford. 2000. Multivariate statistics for wildlife and ecology research. Springer, New York, New York, USA. McLaughlin, D. B., and H. J. Harris. 1990. Aquatic insect emergence in two Great Lakes marshes. Wetlands Ecology and Management 1:111-121. McNicholl, M. K., P. E. Lowther, and J. A. Hall. 2001. F orster’s Tern (Sternaforsteri). Account 595 in A. Poole and F. Gill, editors. The birds of North America. The Academy of Natural Sciences, Philadelphia, Pennsylvania, USA. 135 Moore, D. R. J ., P. A. Keddy, C. L. Gaudet, I. C. Wisheu. 1989. Conservation of wetlands: do infertile wetlands deserve a higher priority? Biological Conservation 47:203-217. Mortsch, L., H. Hengeveld, M. Lister, B. Lofgren, F. Quinn, M. Slivitzky, and L. Wenger. 2000. Climate change impacts on the hydrology of the Great Lakes-St. Lawrence system. Canadian Water Resources Journal 25: 1 53-179. Mortsch, L., J. Ingram, A. Hebb, and S. Doka, editors. 2006. Great Lakes coastal wetland communities: vulnerability to climate change and response to adaptation strategies. Environment Canada and the Department of Fisheries and Oceans, Toronto, Ontario, Canada. 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. Neuman, K. K., L. A. Henkel, and G. W. Page. 2008. Shorebird use of sandy beaches in central California. Waterbirds 31:115-121. Parsons, K. C. 2002. Integrated management of waterbird habitats at impounded wetlands in Delaware Bay, U.S.A. Waterbirds 25(Special Publication 2):25-41. Potter, B. A., R. J. Gates, G. J. Soulliere, R. P. Russell, D. A. Granfors, and D. N. Ewert. 2007. Upper Mississippi River and Great Lakes Region Joint Venture shorebird habitat conservation strategy. U. S. Fish and Wildlife Service, Fort Snelling, Minnesota, USA. Prince, H. H., P. I. Padding, and R. W. Knapton. 1992. Waterfowl use of the Laurentian Great Lakes. Journal of Great Lakes Research 18:673-699. Prince, H. H., and C. S. Flegel. 1995. Breeding avifauna of Lake Huron. Pages 247-272 in M. Munawar, T. Edsall, and J. Leach, editors. The Lake Huron ecosystem: ecology, fisheries, and management. SPB Academic Publishing, Amsterdam, The Netherlands. Provence, C. D. 2008. Effects of diking and plant zonation on invertebrate communities of Lake St. Clair coastal marshes. Thesis, Michigan State University, East Lansing, USA. Roman, C. T., W. A. Niering, and R. S. Warren. 1984. Salt marsh vegetation change in response to tidal restriction. Environmental Management 8: 141-150. Ryder, J. P. 1993. Ring-billed Gull (Larus delawarensis). Account 33 in A. Poole, P. Stettenheim, and F. Gill, editors. The birds of North America. The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Omithologists’ Union, Washington, DC, USA. 136 SAS Institute. 2004. SAS OnlineDoc® 9.1.3. SAS Institute, Cary, North Carolina, USA. Steen, D. A., J. P. Gibbs, and S. T. A. Timmermans. 2006. Assessing the sensitivity of wetland bird communities to hydrologic change in the eastern Great Lakes region. Wetlands 26:605-611. Thiet, R. K. 2002. Diversity comparisons between diked and undiked coastal freshwater marshes on Lake Erie during a high-water year. Journal of Great Lakes Research 28:285-298. Timmermans, S. T. A., S. S. Badzinski, and J. W. Ingram. 2008. Associations between breeding marsh bird abundances and Great Lakes hydrology. Journal of Great Lakes Research 34:351—364. Tucker, G. C. 1990. The genera of Arundinoideae (Gramineae) in the southeastern United States. Journal of the Arnold Arboretum 71:145-177. Tulbure, M. G., C. A. Johnston, and D. L. Auger. 2007. Rapid invasion of a Great Lakes coastal wetland by non-native Phragmites australis and Typha. Journal of Great Lakes Research 33 (Special Issue 3):269—279. US. Fish and Wildlife Service and Canadian Wildlife Service. 1987. Standard operating procedures for aerial waterfowl breeding ground population and habitat surveys. Unpublished manual as revised, US. Fish and Wildlife Service, Laurel, Maryland, USA. Wagner, T., D. B. Hayes, and M. T. Bremigan. 2006. Accounting for multilevel data structures in fisheries data using mixed model. Fisheries 31 :180-187. Weller, M. W., and C. S. Spatcher. 1965. Role of habitat in the distribution and abundance of marsh birds. Department of Zoology and Entomology Special Report 43, Agricultural and Home Economics Experiment Station, Iowa State University, Ames, USA. Whitman, W. R. 1976. Impoundments for waterfowl. Canadian Wildlife Service Occasional Paper 22, Ottawa, Ontario, Canada. Whitt, M. B. 1996. Avian breeding use of coastal wetlands on the Saginaw Bay of Lake Huron. Thesis, Michigan State University, East Lansing, USA. Wilcox, D. A. 1993. Effects of water-level regulation on wetlands of the Great Lakes. Great Lakes Wetlands 421-2, 11. 137 Wilcox, D. A. 1995. The role of wetlands as nearshore habitat in Lake Huron. Pages 223-245 in M. Munawar, T. Edsall, and J. Leach, editors. The Lake Huron ecosystem: ecology, fisheries, and management. SPB Academic Publishing, Amsterdam, The Netherlands. Wilcox, D. A., J. A. Meeker, and J. Elias. 1993. Impacts of water-level regulation on the wetlands of the Great Lakes. Phase 2 Report to Working Committee 2, International Joint Committee Water-levels Reference Study, Ottawa, Ontario, Canada, and Washington, DC, USA. Wilcox, D. A., and T. H. Whillans. 1999. Techniques for restoration of disturbed coastal wetlands of the Great Lakes. Wetlands 19:835-857. Wisheu, I. C., and P. A. Keddy. 1992. Competition and centrifugal organization of plant communities: theory and tests. Journal of Vegetation Science 3:147-156. 138 CHAPTER 4 MANAGEMENT IMPLICATIONS Management actions that isolate wetlands from their natural ecological forces have become controversial. The construction of dikes in Great Lakes coastal wetlands separates the hydrology of diked wetlands from the adjacent lakes, which ultimately leads to changes in biogeochemical cycles, water levels and fluctuations, vegetation, and food resources for birds (Provence 2008). Between the 19505 and 19703, several coastal wetlands were isolated fiom normal water level fluctuations through dike construction at important migrant waterfowl stop-over areas (see Bellrose 1980, Bookhout et al. 1989). These projects were initiated primarily to maintain elevated water depths and enhance wildlife use during periods of historic low Great Lakes water levels. The vegetation and physical characteristics of diked and undiked wetlands were different during avian breeding and migration periods (Figures 15-17). Diked coastal wetlands were generally dominated by cattail and had higher water depths and more organic sediment, open water, and floating vegetation compared to undiked wetlands (Table 21). Conversely, undiked wetlands had shallower water depths, inorganic soils, and more common reed, bulrush, and litter than diked wetlands. The timing and intensity of water level fluctuations can influence bird use, and diked wetlands had fluctuations similar to those of inland wetlands, with water levels highest in early spring, declining during the breeding season, and lowest in late summer. Undiked wetlands exhibited similar water level fluctuations to diked wetlands in 2005 and 2007, with water levels 139 .aoZZSwZE macaw waist masonEoZooan BB 3 wow: 8:3 538%? 8865 8:5 .hoONfiooN macaw >13 warn—Z6 menace? Z938 mam Bmfiwmm v5 82m “EU .5 68:ch Zia Z6856 .Zo 8589 35:35 Roar? .mZ 2:me .5 8:325 952 5&8: .8 U028 as: 88.5% 3. £83 L LL: 38%: 33 a}. .a 32-323.» LEE, .23 :82 £222 880 .56 N83 L25 e83 583 685 as: ”am: Eve .5 LBS 9&3 wanes as. L58 2: Box 8888 Lea 3535 ”$5, Lam mag ”Sam 88 .5. Loos? .5 83m ”EL/mm any abuse ”ME: 585 :2. ”as? $5.3 zofifimog oe Em mZOZHmamm< mmZUmZan QMZmZ L ...L. ......L...__ L:.L.LL-...-....-....----!I.-- LLL_.L LL Z .1 LLLFLIL Z LLL LL LLL_ LL LL. L LLL LLL LLIILLLLL II it 858% mentozooboz... 065225 3325 mommafia wean .3.-2.22::— an... «11....n.1LLm_.n.l..|.:lflerd-lulndwrituwbtr,5§ ...-u... u...Iu.fififiE-.:nn.tw ...-....MMmN..mr.. .3.. 8:9 LLLLL __;L. . .LL . L L _ L L L l meotozooaoz... Eomezaoéoofi ll 325 8:38 vaoza wEZQ i i 140 dammom @5505 05 magi 358336on ES 3 com: mesa cos—30w?» Baumcfi 8:3 SOON-mos .8883 BE 8 watam 3m— mfihfi €5.83 3300 5mm Bufiwmm Ea 32m ~30 .Lm @855: 98 wow—Lu mo 3an comm—9:6 18%» H .2 05an— 5 ELLE-BE £02 5&8: .8 ”028 was vow-saga ...L. fieam L 2: 3882 5% EL. .5 Laos-323.» L-me.» L3 552 £334 380 .58 L005 L25 LL83 HBELL EouLmLao-L “Em-L LEE .a “8.3 new: mass-L L? L58 = Bum 8888 L5 «LchLLLLLSLS Lam LLLULLLEUL Loco.e<602< LB omcom ”EL/mm an; Mafia ”when Sam .a< “as? flat. zofi OH Em mzofisL/mmmmLL. mmLommm Bum L LLLL L LLL.-- LL L LLLLLL LLLL LLL._LLL LLLL LLLLL LLLLLLLLLL LLLLL 1 LL LLL LL 5558 005203 5:329 LozouLou2< 1 am: II II M ELL/L3 mama ESE-HE 0&5 . e. .52. 1L: - .LL. L---: :----LL..........L.....L_.‘L.L-..---LLL.-.L..L--:: .L LLL LLL _ , g _ I machoEDQOB kl] Ill BELLE; >28 mkgqmgmm-L limomLLLoonouzLL. oo mam mag ”Leic— uooU .E< ”003?. 55 owwom ”Mk/mm Ech- mLBmhom ”mu-On— EuEm .E< L532 mZOLHmBmm< mmLUmEm OMS LL . L. LLLLLLLLEL- ,..L....,_...LL_..L_.....LL LLLL . . . LL......L.L.L....L..__._.LLL.._. LLL L L.LLLL LLL LL LLLLLL. L LLL LLL LLLL LLLL 83:9 w:_5nm0\m:o$E\mEo-w \OEOU\OUE .Sac 3:8: 8 26 bnmmtuofi. moumnwmmow :9. a 93 £0888 550 Ho $53 £8 :o 3me “asp—@862? EB a 3 wow: wok—58w 28:03 38%.: mama At 328m £53m Sateen: n m 98 EwE 8 8238 n mica“? x83 H562: 9 32 H mfivfim haw ”>22 8 :5QO n mafia”? o: ”via—Lem mm 368 mm 8% 33:6: 98 cue—6 a 3:58.“ @2233 .Lo bmmnm=§m 68955”; .mvotom :23“me Ea wfiuoofi mats“. muss—Ho? .aawfiog Sam Banmmaw can 32m HELD Am E 305650 65:03 98 3383862; EB 5953 maquoflmLE Luca—Son wnunomcfi x552 4N Dish 143 Table 21. Cont’d. uoumafia mopeaw Forster’s Tern Tern Least Bittem Herons mopeaw Ducks American Bittem Table 21. Cont’d. 145 Table 21. Cont’d. Food uopmafia A «51:2;qudi qsu awxqauawl (um oz<) daea De th (um 0Z>) mucus Microhabitat [gos pasodxa new pasmuans ma qsmlna mopeaw 19M Baum] :1 1191191) Supeou Ve etation Zone 5991 l/sqnxqs daaq waisgsmd-uo N MOIIWS 1091SISJQd-UON pasx-nuqns 31111901 :1 mopeaw 19M qsmlng llmmC) sauwfiqu Fall Migration Black Tern F orster’s Tern Caspian Tern 146 declining during summer months, although the decrease in water levels at diked wetlands was more pronounced compared to undiked sites. In 2006, water levels in undiked wetlands peaked during late summer, which was consistent with long-term averages on the Great Lakes. Bird Use During Spring Migration Due to time and resource constraints, my surveys for birds during spring migration were limited to aerial waterfowl surveys. Only 14 aerial waterfowl surveys were conducted, which did not permit comparisons between diked and undiked wetlands by time period (i.e., spring, late summer, early fall). However, some predictions can be made regarding bird use of diked and undiked wetlands based on observed vegetation, water depths, and water level fluctuations (Figure 15). Canada Goose densities were greater in undiked compared to diked wetlands during migration, and large flocks (i.e., >100 individuals) of Canada Geese were observed on undiked Saginaw Bay wetlands during spring aerial surveys. Canada Geese probably used these wetlands as roosting sites and flew to other locations (e. g., agricultural lands) to forage, so the primary determinant of habitat selection may have been secure roosting areas. Wood Ducks seemed to positively respond to the deep, aquatic bed zones of diked wetlands, and were observed there in greater densities than on undiked sites regardless of season. Dabbling and diving duck species used both wetland types in spring, with diving ducks focused in deeper water depths and dabbling ducks in shallower areas (Figure 15). Gadwall density was greater in diked compared to undiked sites, and Gadwalls may have been attracted to submersed aquatic plants in diked wetlands, which are an important food source for this 147 species (Table 21, Bellrose 1980). Based on my personal observations and the high spring water depths of diked wetlands, shorebird use was likely lower in diked compared to undiked sites (Figure 15). Parsons (2002) observed greatest migrant shorebird abundance in Delaware Bay impoundments with low spring water levels. Potter et al. (2007) assumed that the timing of resource availability for shorebirds is most critical during spring migration in the Great Lakes and Upper Mississippi River region, because the migration period is short and precedes the breeding season. Depending on the spring migration management goals, changes to water level manipulations may be required in diked wetlands to maximize use by some bird groups. For example, increasing densities of most shorebird and dabbling duck species in diked wetlands would require lower spring water levels. More study is needed to better understand bird use of diked and undiked coastal wetlands during spring migration, as well as to determine management guidelines that increase use by focal bird groups. Bird Use During Breeding Season Although there were clear differences in the vegetation and physical conditions of diked and undiked coastal wetlands, breeding bird use was largely similar (Figure 16). Bird species richness was similar between diked and undiked wetlands and similarity indices suggested the breeding bird communities were comparable. This similarity may reflect the ability of most wetland bird species to adapt to dynamic wetland conditions. Wetland birds often use complexes of wetlands to meet life requisites (see Dzubin 1969, Brown and Dinsmore 1986, Weller 1999), and all of the diked wetlands were within large complexes of undiked wetlands. Diking also did not appear to negatively impact use by 148 non-game bird species of conservation concern (e. g., listed or species of greatest conservation need [SGCN, Eagle et al. 2005]). While overall bird use between the two wetland types was similar, more study is needed to compare reproductive success (e. g., nest densities, nest success) of priority species in diked and undiked wetlands. Several species seemed to respond to differences in vegetation, hydrology, and/or food resources between the wetland types. Management of diked wetlands appeared to benefit several species that use deep-water marshes for breeding. I observed greater densities of Canada Goose and Wood Duck in diked compared to undiked wetlands during the breeding season. Higher water levels in diked wetlands likely provided attractive brood rearing habitat with abundant food for both species near nesting sites. American Bittem, Least Bittem, and Common Moorhen were observed in greater densities in diked than undiked wetlands. Higher water levels and greater percent open water in the diked wetlands may have increased interspersion of emergent vegetation and open water, which could have provided attractive breeding habitat for these species (Table 21). Deeper water levels in diked compared to undiked wetlands may have created a stable environment for the invertebrates, amphibians, small fish, and submersed vegetation used by these species for food. Some bird species appeared more abundant in undiked compared to diked wetlands. Mallard linear density was greater in undiked sites during timed-area surveys. Mallards prefer to forage in shallow water (Fredrickson and Taylor 1982), so shallower water depths at undiked wetlands could account for greater Mallard abundance. Although invertebrates were more abundant at diked than undiked sites at St. Clair Flats (Provence 2008), deep water may have limited access to invertebrate and plant foods 149 used by Mallards and other wetland birds. I observed greater densities of Ring-billed Gull, Herring Gull, and Forster’s Tern in undiked than diked wetlands, which could be related to differences in forage fish abundance between the wetland types, conditions limiting the ability of these species to forage in diked wetlands (e. g., floating and/or submersed plants), or proximity to nesting sites. F orster’s Tems were only observed nesting in undiked wetlands where dead bulrush stems from the previous growing season collected, which provided a substrate for their floating nests (Figure 16). Percent cover and density of bulrush were lower in diked compared to undiked wetlands (Table 21). If maximizing use of diked wetlands by breeding birds is a goal, water level manipulations could potentially increase use by some species. Managing diked wetlands for shallower water depths could enhance use by Mallards and many other wetland bird species by improving access to abundant invertebrate and plant foods. Periodic (e.g., every 5-7 yrs) complete drawdowns of diked wetlands could improve habitats for breeding birds by reducing the buildup of organic matter, releasing nutrients and stimulating plant growth, and improving vegetation and structural diversity of the diked wetlands. Given that common reed was more prevalent in undiked than diked wetlands, it was surprising that more differences in breeding bird use were not observed between the wetland types. Due to low lake levels, I selected point count locations within 400 m of open water, which may have reduced potential effects to bird use caused by common reed in undiked wetlands. Meyer (2003) investigated wildlife use of common reed and native vegetation in one coastal wetland complex, but more research is needed to understand the effects of common reed on breeding birds. Since recent common reed expansions have 150 occurred with low Great Lakes water levels, studies need to evaluate bird use of common reed under a variety of water depths. Some diked wetlands contained pockets of common reed, which provide opportunities to study bird use of common reed under flooded conditions. Bird Use During Fall Migration Vegetation and substrates along open water-emergent marsh interfaces surveyed during ground surveys differed between diked and undiked wetlands, which was consistent with quadrat sampling done during the breeding season (Table 21). Cattail, floating vegetation, and organic soils were more common in diked wetlands, while bulrush and inorganic soils were observed more often in undiked sites (Figure 17). Despite these differences, migrant bird use of diked and undiked wetlands was similar. Most bird density variables and total species richness were similar between the wetland types, and similarity indices also implied similar bird communities. Some authors (e. g., Wilcox 1995) have suggested that diked wetlands may negatively impact shorebirds and rare species. I found that most densities of rare species (i.e., threatened, endangered, special concern), SGCN, and shorebirds were similar between diked and undiked wetlands during fall migration. Migrant wetland birds are known to use a variety of wetland types within larger complexes to meet life needs (e.g., foraging, resting, escape cover). All the diked wetlands studied were located within large undiked wetland complexes, so the similarity of migrant bird use during late summer/early fall suggests that birds were using larger wetland complexes, rather than individual wetlands. More 151 study is needed to understand why bird species were using the diked and undiked wetlands during migration. A few species appeared to respond to differences in the vegetation and physical conditions of diked and undiked wetlands. Canada Goose densities were greater in undiked wetlands during aerial surveys, while fall ground surveys revealed similar densities in diked and undiked wetlands. This discrepancy may be due to seasonal changes in Canada Goose densities and habitat use, such as high numbers in undiked Saginaw Bay wetlands during spring migration, high variation of densities during migration, and the low number of aerial surveys conducted. Wood Duck densities were greater in diked compared to undiked wetlands during both aerial and early fall ground surveys, which is consistent with breeding season surveys. Linear densities (birds/km of edge surveyed) of dabbling ducks and Mallards were greater in undiked compared to diked wetlands. Wilson’s Snipe was the only shorebird species observed in greater densities in diked wetlands and may have been attracted to organic soils and high invertebrate abundance. Greater Yellowlegs were observed in greater linear densities in undiked wetlands, but was the only shorebird species more abundant in undiked wetlands. Similar to breeding season surveys, Ring-billed Gull and Forster’s Tern densities were greater in undiked than diked wetlands during early fall surveys. More study is needed to determine if fish abundance and composition or wetland microhabitats influenced use by Ring-billed Gull and Forster’s Tern. Greater abundance of dabbling ducks and Mallards in undiked wetlands could be related to food abundance or foraging habitat. These species seemed to be attracted to sites with shallow water depths (Figure 17), which may have offered preferred foraging 152 conditions (F redrickson and Taylor 1982). Other studies (e. g., behavioral, radio- tracking) are needed to understand why (e. g., feeding, resting) waterfowl were using diked and undiked wetlands. Assuming that maximizing use by dabbling ducks is a management goal, managing diked wetlands for shallow water levels (e. g., 10-20 cm) could increase dabbling duck densities. Periodic drawdowns could also promote growth of annual moist-soil plant species that provide valuable food for dabbling ducks and other bird species. However, close monitoring would be required, since drawdowns can also encourage invasive species, such as common reed. My observations during fall migration revealed that shorebird use was similar between diked and undiked wetlands. Although water depths tended to be higher in diked compared to undiked wetlands, water levels in diked wetlands were usually lowest in late summer, which provided pockets of mudflats and shallow water at a time when fall shorebird migration typically peaks (Figure 17). Low water conditions in diked wetlands sometimes created mats of organic matter and submersed vegetation that shorebirds used for foraging. Dabbling ducks, such as Green-winged Teal, Blue-winged Teal, and Mallard, were observed using the same habitats (Figure 17). In some cases, pumping water into impoundments in preparation for fall waterfowl hunting flooded these shallow-water habitats and reduced use by shorebirds and dabbling ducks. Maintaining shallow water depths in impoundments further into early fall (e.g., mid to late September) could sustain habitat for shorebirds for the duration of most species’ fall migration period. 153 Management Discussion The primary objective of my study was to compare bird use and habitats of several diked and undiked Great Lakes coastal wetlands, rather than answer the larger question as to whether the practice of coastal wetland diking should be continued at current levels, possibly expanded, or ended. While my work provides insight into one component of this larger question, additional studies, such as wetland function comparisons, other fish and wildlife investigations, cost-benefit analyses, and value assessments, are also required. Comparisons of bird use during normal to high Great Lakes water levels are also needed. This study was also not designed to evaluate specific management practices used in diked wetlands; however, some general recommendations can be made based on my findings and likely management goals. Biologists can manage water levels in diked coastal wetlands to achieve a variety of goals, such as improved conditions for wildlife (e.g., game, endangered, or threatened, species), vegetation diversity, recreation, and invasive species eradication. Management recommendations will vary depending on stated goals, and some goals may not be compatible at a given site. For example, managing to provide foraging habitat for spring migrant shorebirds will likely impact breeding habitat for some bird species (e.g., bitterns, Common Moorhen). In general, the differences in bird use and abundance between diked and undiked wetlands, though minor, seemed to be related to differences in water depths and vegetation resulting from coastal wetland diking. There appears to be potential for managers to increase use of diked wetlands by some bird species by providing shallow water depths and conducting periodic drawdowns. Occasional drawdowns of diked wetlands could improve habitats for breeding birds by improving 154 vegetation and structural diversity, and shallow water depths increase access to invertebrate and plant foods by many wetland bird species. Densities of Mallards, other dabbling ducks, and shorebirds in diked wetlands would likely equal or surpass those of undiked wetlands if managers provide shallow water in diked wetlands during peak migration periods. Drawdowns could also be used to increase the production of annual plant seeds as a food source for migrating dabbling ducks. Where complexes of diked wetlands exist, water levels could be manipulated to provide a variety of wetland conditions (e. g., deep marsh, shallow marsh, mudflat) among the impoundments that address multiple management goals (e. g., multiple bird groups, both breeding and migration periods). Water levels could be changed within each diked wetland periodically on a rotational basis to mimic natural water level fluctuations, while maintaining a diversity of wetland types for birds. Drawdowns must be conducted with caution and constantly monitored to avoid the spread of invasive plant species. Late summer drawdowns are recommended to reduce the expansion of common reed (Avers et a1. 2007). Management guidelines need to be developed for focal species (e.g., game, threatened and endangered, SGCN) in the context of changing coastal wetland conditions, such as climate change and invasive species expansion. Most climate change models predict decreasing Great Lakes water levels in the future (Mortsh et al. 2000, 2006, Lofgren et al. 2002, Croley 2003). Long-term low water levels could increase common reed expansion in undiked wetlands and reduce their value to bird species of management concern. Diked wetlands may provide management opportunities to maximize use by wetland birds and reduce impacts from invasive plant species like 155 common reed. Experimental studies could be implemented in diked wetlands to test the success of new or modified water level management regimes for selected management goals (e. g., use by focal bird species, diverse vegetation). 156 LITERATURE CITED Avers, B., R. Fahlsing, E. Kafcas, J. Schafer, T. Collin, L. Esman, E. Finnell, A. Lounds, R. Terry, J. Hazelman, J. Hudgins, K. Getsinger, and D. Scheun. 2007. A guide to the control and management of invasive Phragmites. Michigan Department of Natural Resources, Lansing, Michigan Department of Environmental Quality, Lansing, Ducks Unlimited, Ann Arbor, Michigan, US. Fish and Wildlife Service, East Lansing, Michigan, US. Army Corps of Engineers, Vicksburg, Mississippi, Michigan Department of Transportation, Lansing, and Aquatic Ecosystem Restoration Foundation, Marietta, Georgia, USA. Accessed 25 March 2009. Bellrose, PC. 1980. Ducks, geese, and swans of North America. Stackpole Books, Harrisburg, PA. Bookhout, T. A., K. E. Bednarik, and R. W. Kroll. 1989. The Great Lakes marshes. Pages 131-156 in L.M. Smith, R.L. Pederson, and RM. Kaminski, editors. Habitat management for migrating and wintering waterfowl in North America. Texas Tech University Press, Lubbock, USA. Brown, M., and J. J. Dinsmore. 1986. Implications of marsh size and isolation for marsh bird management. Journal of Wildlife Management 50:392-397. Croley, T. E., II. 2003. Great Lakes climate change hydrologic impact assessment: International Joint Commission Lake Ontario-St. Lawrence River regulation study. National Oceanic and Atmospheric Administration Technical Memorandum GLERL-126, Ann Arbor, Michigan, USA. Dzubin, A. 1969. Comments on carrying capacity of small ponds for ducks and possible effects of density on Mallard production. Pages 138-160 in Saskatoon wetlands seminar. Canadian Wildlife Service, Report Series 6, Saskatoon, Saskatchewan, Canada. Eagle, A.C., E.M. Hay-Chmielewski, K.T. Cleveland, A.L. Derosier, M.E. Herbert, and RA. Rustem, (eds). 2005. Michigan's Wildlife Action Plan. Michigan Department of Natural Resources. Lansing, MI. http://www.michigan.gov/ dnrwildlifeactionplan. Date accessed: February 13, 2009. Fredrickson, L.H. and TS. Taylor. 1982. Management of seasonally flooded impoundments for wildlife. US. Fish and Wildlife Service Resource Publication 148. 157 Lofgren, B. M., F. H. Quinn, A. H. Clites, R. A. Assel, A. J. Eberhardt, and C. L. Luukkonen. 2002. Evaluation of potential impacts on Great Lakes water resources based on climate scenarios of two GCMs. Journal of Great Lakes Research 28: 537-554. Meyer, S. W. 2003. Comparative use of Phragmites australis and other habitats by birds, amphibians, and small mammals at Long Point, Ontario. Thesis, University of Western Ontario, London, Canada. Mortsch, L., H. Hengeveld, M. Lister, B. Lofgren, F. Quinn, M. Slivitzky, and L. Wenger. 2000. Climate change impacts on the hydrology of the Great Lakes-St. Lawrence system. Canadian Water Resources Journal 25: 153-179. Mortsch, L., J. Ingram, A. Hebb, and S. Doka, editors. 2006. Great Lakes coastal wetland communities: vulnerability to climate change and response to adaptation strategies. Environment Canada and the Department of Fisheries and Oceans, Toronto, Ontario, Canada. Parsons, KC. 2002. Integrated management of waterbird habitats at impounded wetlands in Delaware Bay, U.S.A. Waterbirds 25(Special Publication 2):25-41. Potter, B.A., R.J. Gates, G.J. Soulliere, R.P. Russell, D.A. Granfors, and D.N. Ewert. 2007. Upper Mississippi River and Great Lakes Region Joint Venture Shorebird Habitat Conservation Strategy. U. S. Fish and Wildlife Service, Fort Snelling, MN. Provence, C .D. 2008. Effects of diking and plant zonation on invertebrate communities of Lake St. Clair coastal marshes. MS. Thesis, Michigan State University, East Lansing, MI. Weller, M. W., and L. H. Fredrickson. 1974. Avian ecology of a managed glacial marsh. Living Bird 12:269-291. Weller, M. W. 1999. Wetland birds: habitat resources and conservation implications. Cambridge University Press, Cambridge, United Kingdom. Wilcox, D. A. 1995. The role of wetlands as nearshore habitat in Lake Huron. Pages 223-245 in M. Munawar, T. Edsall, and J. Leach, editors. The Lake Huron ecosystem: ecology, fisheries, and management. SPB Academic Publishing, Amsterdam, The Netherlands. 158 APPENDIX A Common and scientific names for bird species observed during surveys. 159 Table A-1. Common and scientific names of avian species observed during bird surveys conducted at St. Clair Flats and Saginaw Bay, Michigan coastal wetlands during 2005- 2007. Species are listed by wetland use category. Species Scientific Name Wetland-dependent Species Canada Goose Mute Swan Trumpeter Swan Wood Duck Gadwall American Wigeon American Black Duck Mallard Blue-winged Teal Northern Shoveler Northern Pintail Green-winged Teal Canvasback Redhead Ring-necked Duck Scaup (species unknown) Bufflehead Hooded Merganser Ruddy Duck Pied-billed Grebe Double-crested Cormorant American Bittem Least Bittem Great Blue Heron Great Egret Green Heron Black-crowned Night-Heron Northern Harrier Branta canadensis Cygnus olor Cygnus buccinator Aix sponsa Anas strepera Anas americana Anas rubripes Anas platyrhynchos Anas discors Anas clypeata Anas acuta Anas crecca A ythya valisineria A ythya americana A ythya collaris A ythya spp. Bucephala albeola Lophodytes cucullatus Oxyurajamaicensis Podilymbus podiceps Phalacrocorax auritus Botaurus lentiginosus Ixobrychus exilis Ardea herodias Ardea alba Butorides virescens Nycticorax nycticorax Circus cyaneus King Rail Rallus elegans Virginia Rail Rallus limicola Sora Porzana carolina Common Moorhen Gallinula chloropus American Coot F ulica americana Sandhill Crane Grus canadensis Semipalmated Plover Charadrius semipalmatus Spotted Sandpiper Actitis macularius 160 Table A-1. Cont’d. Species Wetland-dependent Species, Cont’d Solitary Sandpiper Greater Yellowlegs Lesser Yellowlegs Semipalmated Sandpiper Least Sandpiper Baird’s Sandpiper Pectoral Sandpiper Dunlin Stilt Sandpiper Short-billed Dowitcher Wilson’s Snipe American Woodcock Red-necked Phalarope Bonaparte’s Gull Ring—billed Gull Herring Gull Black Tern Forster's Tern Belted Kingfisher Alder Flycatcher Willow Flycatcher Tree Swallow Northern Rough-winged Swallow Bank Swallow Sedge Wren Marsh Wren Swamp Sparrow Red-winged Blackbird Yellow-headed Blackbird Wetland-associated Species Bald Eagle Merlin Black-bellied Plover Killdeer Caspian Tern Common Tern Black-billed Cuckoo Eastern Kingbird 161 Scientific Name Tringa solitaria T ringa melanoleuca T ringa flavipes Calidris pusilla Calidris minutilla Calidris bairdii Calidris melanotos Calidris alpina C alidris himantopus Limnodromus griseus Gall inago delicata Scolopax minor Phalaropus lobatus C hroicocephalus philadelphia Larus delawarensis Larus argentatus C hlidonias niger Sternaforsteri Megaceryle alcyon Empidonax alnorum Empidonax traillii T achycineta bicolor Stelgidoptetyx serripennis Riparia riparia C istothorus platensis Cistothorus palustris Melospiza georgiana Agelaius phoeniceus Xanthocephalus xanthocephalus Haliaeetus leucocephalus Falco columbarius Pluvialis squatarola Charadrius vociferus Hydroprogne caspia Sterna hirundo C occyzus erythropthalmus Tyrannus lyrannus Table A-1. Cont’d. Species Wetland-associated Species, Cont’d Warbling Vireo Purple Martin Cliff Swallow Barn Swallow Gray Catbird Yellow Warbler Common Yellowthroat Common Grackle Nonwetland Species Ring-necked Pheasant Rock Pigeon Mourning Dove Chimney Swift Northern Flicker Blue Jay Black-capped Chickadee American Robin European Starling Cedar Waxwing Yellow-rumped Warbler American Redstart Scarlet Tanager Song Sparrow Northern Cardinal Rose-breasted Grosbeak Indigo Bunting Brown-headed Cowbird Baltimore Oriole American Goldfinch 162 Scientific Name Vireo gilvus Progne subis Petrochelidon pyrrhonota Hirundo rustica Dumetella carolinensis Dendroica petechia Geothlypis trichas Quiscalus quiscula Phasianus colchicus Columba livia Zenaida macroura C haetura pelagica Colaptes auratus C yanocitta cristata Poecile atricapillus T urdus migratorius Sturnus vulgaris Bombycilla cedrorum Dendroica coronata Setophaga ruticilla Piranga olivacea Melospiza melodia Cardinalis cardinalis Pheucticus ludovicianus Passerina cyanea Molothrus ater Icterus galbula Carduelis tristis APPENDIX B Data tables from breeding bird surveys and analyses. 163 Lodvfimod modnnde hodflmod cmdfimoL Lodvnumod modfloLd hodfiofio hmdflth C LodVfimod 8.36 L .o wodnuwhd Nvdfimmm LodvnuLod No.91 L.o oodnnhvd Lumdfiowd ..LUm :oEEoU Lodvnumcd Nodfiood oodumowd vmdflvfio LodvaLod Nod”: L.o hodfiwvd mmdfimwd 0i 02 0295* 02 0EES> 5680a 00m 22022 8052 532% 822 85258: $8 005858 SE3 088800 0330838080. 20002 $80002 0080000m 030800 80: E0 3008 05 ~08 83308 ..---.. 80080: 05. 008-30 03:80 “0: 33 6008 003.0. 05 80,0 x508 O 05 t 03g UH< :0 8000 00008 003 is. 0880800 :< 000.5006 02 80 00mg 030.§> :0>_w m 00.8 _0008 038600 808 0.0 08208 00205 000—0m 0830080 00808950 3 008: 08 80800800 885008 00800008 0 000208 005 20002 .hddN-mddm 803022 Sam Bafiwam 0:0 305 .830 Am 8 0080:0800 08:00 8800 w880 0.08000? 8000 000 00000 8 32800 005 0.80800 9 00m: £0008 0088 800 m0:_0>-m 080 800080. AUZV 800800.80 80008.80: 0.008? .m-m 030,—. 168 82.0 3: 3:... 3: 32.0 m. :0 :03 33. 32.0% meow 0000085 000003002 ddmdd ndow Nu0dd Ndw0 -- -- mwmdd 0000 _ 030000 008800 005d mama 033d wd0w m Smd 0.03 00mm... mfimw 0008030=0> 008800 amamd 0.003 «wad .0.va -- -- 300d ddmm 005.003 320% d0ddd 3.43% 003d 0.0% 008d .0.me 58d $.me 30:03m 005 mummd Ndwn- -- -- 0mdvd 3.80- Nhamd add»- 000w8v~ 800000 0vad ...wémoT mm—md 00002- mwmvd 16de 0 30d .56de 80H 008000 00603 00008800000003 Emmd 0.0.005- -- -- -- -- 005d ..wfivw- 0000—005 000000-30=0> mvwvd vad— fined vdmd— mwmvd 0.02: Nmmvd 0mg: 000x005 00805-030 000:. 0.03 :dwd Mada: wowud :30 050d 030 30825 0802i mudnd N00: 03 _ d QQZ mmdmd ~.w0_ _ vmdmd _.m0_ _ 00.5? 0802 393 0.22. $0.... .8002- :23 000% Nwwmd ...ddmm- -- -- md0Nd ..vdmm- mbwmd ... _ dmm- 00030obm 32:3 m—mmd 00$ 323 M03 mommd 02:. NNOmd 0.30 BBRBm 0000. 000003 80000000000003 020.70 02 020>i 0?. 020.50 0?‘ 020070 03“ 0308; .3800 00m 2082 802 0550 200 0282520 $00 0508000 0 003 00000800 030000900800 20002 00.80002 00000030 0.800 .m-m 030,—- 169 mowmd m. 5m o _ mmd mdom 23d odom vwmmd Ndom _ooood ~nmwmd v. 5m oommd mdom Sand odom mwwmd Nmom _ooooood =mm 359$ mmmmd v.32- -- -- hmmmd www.2- movud ado:- _ooood Somd ”mom T -- -- ommmd mdwm T Novnd Noe:- Soooood mam wag omood m.m 3. good «doo- ~mood b; $- nmood o.m S- _oooo.o Snood m . m G- vmood «.63- _moo.o h; S- smood o. m S- _oooooo.o EoEm fine-H woood odmoT flood @33- ooood mdmoT noood vdvoT Soood woood fiomoT flood owed? ooood mdmoT noood vdvoT _ooooood E035 505:3 owowd odva -- -- -- -- onohd WENT Soood mSwd mdva -- -- -- -- baa-d mix-«- _oooooo.o 380 825-35 Smod wNmT o—m—d p.37 mmmod mi:- oomod N43- Soood mmmod onT com—d v.3:- ovmod mil- oomod mi:- _oooooo.o 2232 2 Ed wag- ooohd 063- ~Vmood fimnw- mvood fimnw- _oooo.o moood o.§w- moobd e63- vmood fimhw- ovood fimmw- Soooood x25 woo? mmfd m._ooT :de N.Nbo_- mvofio odooT 32d odoo- _oooo.o amid hgooT Somd 9.3.3- 32d odooT m Kod odoo- _oooooo.o 53m 832 omomd EEm- -- -- ....- -- mnovd o. _ mm- ~ooood voomd o.—mm- :- -- -- -- muovd o. _ mm- _oooooo.o 380 «9280 $6on “cooaomou-ccm—EB 039d 02 mag-m 02 0:13.; 02 egg-m 0?. BBoESam oogtgoo mo 2252 A23 8502?: $8 25am 52$ @5853“: B 2me> £28m Em @822 @8285 252580 o>mmm2w28=< mEooE 85802 88.26% 630280 “on Ed 388 05 $5 832?: ..--J. c238: 2E. .moumuflm U_< no woman Boat? :QZw Q So .358 03863 62: 23 8365 mos—Q, vow—om 38033 85938 3 37.: 08 328955 3.5988 @833“ a cows—2: a5 2033 .noom-moom dawEBE Sam Bufiwmm Ea mam—m “mm—U am 8 383:8 $580 Smog waist moan—83 some 98 Bid 5223 333% BE oBQEoo 2 wow: 803 $302 .8338 O 325% 9560a 3028 8 83:53 PEN 5:5 €808qu oocutgoo 8m 8m $283 532 £32 2358 8me So Big-m can moumwfim 67b :25th cognac?“ 98:52 .v-m 033. 170 £33 :37 ~22 352- $23 awe- m 3.0 32:- 582 33$ 387 $3... «.32. $23 22:- m 36 32:. 8883 Efi 5&me momoomm 633003998303 32 23- 85... 3%- Booed 38¢ 33. 82... 33- 8883 28.85 323-323 23¢ 28.. £3... Sad- 88% 393 38.- $9... $3.- 5883 :23 omBm 83¢ a. _ mm- 2.3... 3%. 52 3%. 889° .3de 3mm- 83... cam. and $2. 588$ 238% 32:3 32.... 3:. «~85 $3. $23 $3 883 $8... 33. ”Sod $3 $85 $3 8883 Efi Mafia 33¢ 22. 8%... 38- ~33 2%.. £86 23. 583 83¢ 32. $8... 5.3. ammo N. :2. 28.0 $3. 5883 5&8: 8888 88¢ 32. ..- ... £35 33. ~23. 23.- 583 N83 22. ... 23¢ 33. "as... 3%. 58°85 eom £252 23 85258: 3% ofioeim 95$ $2835 85:?8 .8 U332 fiance—dam UgaEOU o>mmmouwouou3< figgnuogod kn osmtdxr hummGoQ vim 2282 8:582 Banned .Paoo «é 29¢ 171 «88 838.8 :88 8.38.8 :88 838.8 88 8.38.8 E8 838.0 28.8 5.388 :88 838.8 68.8 8.38 3 88 8.38.8 388 838.8 88 832.8 5.8 8.388 :88 8.38.8 388 838.8 58 8.38.8 88 838.8 :88 8.38.8 2:8 8.32.8 :88 8.38.8 A88 8388 828 8.338 :88 5.38.8 Sammfiamm 0030mm Boo 505:2 5:802 5:880 0.5m fix .2585 fig 89 BEN: E058 Z 5003-0an Z 0085808005 5003 :0000 :88 8.38.8 0200 :85 :88 8.38.8 :80: 020 sec :88 838.8 688 8.38.8 88 838.8 .520 083 $88 5.38.8 A8 .8 838.8 8.8 538.8 888 8328 82:0 55:2 38.8 538.8 A88 8.38.8 688 838.8 320 8:56»: :88 5.38.8 228m :88 8.38.8 :3 8053-020 :58 8.388 :88 838.8 88 83:8 88 8.38.8 2232 :88 838.8 :88 838.8 88 838.8 :88 5.38.8 0.25 895 8.8 838.8 28.8 8.388 53m 232 :88 835.8 88 838.8 88 838.8 88 838.8 880 8930 3605mm 802033050303 :25 «000—5 5&0 ~00me 0080mm mam 305me 30E :20 gm 8002030 003 0060mm 05 005 $550 BBQ 80 nomtaoa 05 mm 005.5800 80 5:03.005 .mooméoom mcgv 028303 .8008 5&232 55 305mg can 305 :20 am 8 60003980 3550 Ewan EB $50005 mama—u «003030 860% BB 8m 093 98:03 28 0000 Exam an $00,055.88 :0 00503000 .«o 80:250.: 28 .0800 30288 63358 0.02300 100.8 .802 .m-m 030,—. 172 9:8 838.8 :88 8.38.8 88 838.8 :98 8.388 :88 838.8 :88 838.8 88 838.8 628 8.382 5.8 2.38; A88 83;; 8.8 838.8 :88 :388 8.8 838.8 88.8 8.388 28.8 8.388 28.8 8.388 5&0 988 832.8 58 8.38.8 :88 838.8 38.8 838.8 28.8 8.388 88 838.8 88.8 838.8 8.8 2303 E8 833; 60.8 n _ 3” _ .m 88 838.8 :88 838.8 38 83:8 88 838.8 :88 538.8 8.8 8.338 :88 838.8 :88 838.8 .35 38 838.8 88 8.38.8 :88 838.8 :88 538.8 :88 8.38.8 28.8 8.383 5.8 8.383 808 R883 :88 8.38.8 A88 8.33.8 8&8 2.388 :88 5.38.8 5&0 8.8 332.8 28.8 8.388 98.8 838.8 88.8 2.388 3.8 8.3:; 3.8 8.388 :88 838.8 88 838.8 :88 8.38.8 :88 838.8 88 832.0 98.8 832.8 00x5 xmm Bafiwmm 30E :05 um 30:03m 50m 30:03m tau 5%: 2&5 02$ 05333 EEmEM EBmmm 08—030 3:50.005 50,—. 58ng 0002—9 m0m00mm 003680065033 930—005 00000:-30:0> 38.8.0 883,808 Bow—mam 953m :23 £082 :83 0wc0m Bo=§>m xqmm Bozmam c0wE3éw=om E0582 BERBW 00C: Enofimobm 32:3 Snowmobm 00E< E05 0.00809; 50,—. 0.02m :5 mafia =5 8:583 0060mm fibcou .m-m 050,—. 173 dc8 838.8 :88 838.8 88 838.8 88 838.8 :88 838.8 88 8.38.8 8888 838.8 888 838.8 888 838.8 :88 838.8 88 838.8 :88 838.8 88 838.8 :88 838.8 :88 838.8 :88 838.8 :88 838.8 :88 838.8 :88 838.8 88 838.8 A 8.8 838.8 Atdv modficNd :88 838.8 :88 838.8 :88 838.8 888 838.8 :88 838.8 :88 838.8 :88 8.38.8 :88 838.8 :88 838.8 :88 838.8 :88 838.8 wcsgm owmvfi x3880 voammofibgm 85980 2.65.52 38.8mm waom Swacfl. 623m 533m Swerve/w $383 BQEE¢$O=o> @8383 830 mam—5m fioqoam :Eom 53805 823.020 wommmoioflm .3.. 02m Exam—m €05.52 8% 8:889 9:5 wago—z coowE xoom “Gamma—E coxoo:-wfim 86on 233352 :88 838.8 E8 3388 A28 8382 88 838.8 0285 8568 3.8 838.8 888 838.8 38 :38; 38 838.8 §E§2§ 8888 9:8 8383 A88 838.8 88 832.8 888 838.8 5333 323 888 838.8 888 838.8 285 8.6 83on wogmmoOmmméSwgB :25 Ban 580 835 86on 8m 3288 £8 :20 .8 .380 .38 25 174 $88 83_ S 88 838 S :88 838.8 888 838.8 888 838.8 :25 89.5 888 838.8 88 8.38.8 :88 838.8 :25 8x5 mam Bmcmwmm £8 :20 .8 gamma—00 585:2 2080 oHoEuEm 25300 wovmonfiaem momoomm .PEOU .m-m 2an 175 .88 838.0 5.8 830.... ....8 838.0 .88 8.3383 8.8 83. .... .88 8.3383 8.8 8,338.3 808 w... .380 8.8 838... 8.8 838.8 .88 83...... 88 838... .88 838... .28 830.... .88 830.8 .88 838.8 .88 838.8 8.8 83...... ....8 83...... .88 8.3383 .88 8338.3 .88 838... .88 83?. .28 8.3.0.. .88 832.8 808 0.3%... .88 838.8 .88 838... 2 ..8 838.8 .88 838.8 88 8338.3 .88 838... 8.8 838.8 88 8338.3 .88 838... c ..8 838.8 .28 838.8 .88 838.8 88 8338.3 .88 8.30.... .88 838.8 .88 8.38... 8.8 8.38.8 88 838.8 .88 8338.3 E8 838... .88 0.38.. 888 83. ..o .88 8.3%... 8.3.03.2 500308-805 8.0... 000.0 .0.wm. 000.0 0803 02m. 000.0 0.055 800.. E055 0008084... 8080.80 50.00.00.580 050.0 5023-505 .0mc0m.0.>. 50500... €305.00 00.00%. 0:00m 502.53. 0.0030380 .00 H 50w...3-..00.0 508$ E0582 .0.0>0..m E0582 .00 H 50w:.3-0=.m 5.0..03. 0.0.5 5003 ..03m 083. 00000 050.00 8.00.5 80500005500503 . . 3.85.0 8m"... .08.... 838 5.0 .88 .80.... 8.8.0 30m 30:.w0m 0.0... ..0.0 ..m 508030 003 00.00% 05 .05 000.0 .803 0000 .0 005.000... 05 m. 02.0.8000 .8 30:03.00... 8.03-33 w......5 0500503 .08000 00350:). $0.. 30:.w0m 5..0 0.0.”. ..0.0 ..m .0 500005000 00.080 00.0508... w....=5 50.3.0050 00.00% 5.... 8.. 09¢ 500.003 5..0 00.0 350.0 3 800055.00 .... 02.0.8000 .8 30:05.00... 5..0 ..0...0 5.05008 a€505.50 005.0005 .00.0 0003. 8-0 0.00... 176 .88 ....ovflo? E0... 88:80 ..m8 8.3:... .88 3.300... .....8 8.38... ....8 ....ovflodv Eve 5.0900 .88 80%;? .88 .0080... .825. .38 8.3.0.? .88 8.82.0... 200m. 200 8.00.5 50008800500503 .88 ....ovfiodv .88 5.3.0.0 5.00050. 02.00 .28 8.3.... .38 ....ovflo... .38 83$... .88 3.8a... Eve £008... ....8 8.32... .38 00.38.... .008 0.08.0... .88 2.830.. E“; V.20.... ...8 8.3%... .28 5.38... .88 ....ovflodv =5 050%. .008 3.3. .... .88 5.3.0... .88 5.38... .88 8.90.0.9 88 02.5.0.5. .58 3.3.0... 2.8m €8.05 .88 8.38... .88 2.3.0.... 3.5... .88 8.82.0... 2.0.0.00 .30.. .88 8.008... .88 0.00m... mmo.ao..0> .38.. .88 3.38... .88 8.38... 0.232.; .388 .88 8.3.0... ...08 8.908.? .88 2.3. .... .00....05m .000on .88 8.33... .88 8.33... .28 0.8.2... .88 3.38... .80 53.38... .88 8.38... .0 . .8 88.8... .m . .8 8.38... .m . .8 8.3:... 8882 88800 .38 3.03.0... .88 8.38... 05m .88 88.8... .88 5.908.? :3. 0.50.3 .. T8 5..0 Gmua 08.8 .88 5..0 .00“... .03.... 8.8% x0m B0.._w0m $0.... :05 ..m 3.50 0-0 2...; 177 APPENDIX C Data tables from migrant bird surveys and analyses. 178 came ‘28. ... ... Sam... 3%. ERA. Eden- :8»: 820 some *3: m8; *3: :38 .32 93° .32 saw “85 58.? 56:- 88... $.22- 586v 3E- 88.? ...SE- 8:5 ”am 320 3w 5. $98 3:... ..3: ~82. .. _ .wmw 893 .33 an; ngaoeo 82d “.32 2S... *3: 3m; .32 82¢ ...52 :3 833-85 EM; .833 an”... is: £36 .33 22d *wfim figs: Em... 3S7 .... .... mane $.22- ”23 5.3:- =§80 88.9 23 88.? 3.9. Soodv nae Soodv ”.3 x25 885 83c .38- ES... ES”- :25 LEN- $63 $.62- 5% as: 235 32. NR; ...m.mS- EMS 3.3. 286 LE- 3.80 $88 86on “covcoqouéqmzok/ $25 a3- ... -.. as"... «:6. 8:8 imam- mezeonm afisu 83o Ema ~33. 3:5 8:5 *3: 5‘3 *2: meseoam Bob 82; as: 083 *32 $35 *3: $3... 353 mcfisaa 33 5 62 32 Sam... #2». game 32 2&2 Exam 9.25 “€338 Efi 83° *3: man—... ......Sm «ammo *8: $53 3.2: 30333 was came 3%. as»... fizz- $86 .3? M: :3 ...wdm- WEE 320836333 923 3.2: 3:... :22 £25 :38 885 *0sz 2:5 Ecuaaméefifia Emma 1.2% 3:... ...“.MR ”8; $8». 92.3 33% v.25 =< 83¢ 02 eggd 02 2%; u? can; 02 2%?» £83 Em 2082 352 $35 23 85835:: $8 ageim 573 0332333 9:59:00 2352 85802 woumomom .owHoZSQ Ho: Ev .038 of 35 moumomcfi 3:-.. .8938 Bil gamut 9563 you 33 3on 53m 06 8.“ 592: O 05 a 025» 02 5 Sam Bow—m 83 is. x388 q< .moumufim U~< no 393 932.5, 53w a 8m 6on 03863 “88 05 8365 mos—g wow—om 982:3 macaw—goo 3 3%: 98 88388 8.532: 3338 a cows—05 85 2252 .mooméoom .cmeomE «mam Bacmmmm ES 32m “EU am 8 @8838 $255 95on waist $.83?» “.8525 was wow—6 E 35:95 a: 5m £53 323% :38 EB 58me 989:8 9 wow: £088 3me 8m mos—STA Ea moflmufim 63¢ combing cognac?” @0532 .TO 058. 179 |1| # ..x 893 1.3. «$3. .33- 29: *3: m- S33 3%:- Eo-r aammo _ a; ...ofi- SN; «.3;- Evod 3.?- Noas $.3- 302:2 83on 386836333 :53. ENE- 88? ...-.2? good ...v. 1%. 88.0 ...m. i. EB- {33$ 885 ...c. am. 2m"... 5.3m- 83o Ewa- Saod $.87 Ewe £85 £86 *2: 7 £8... 1:2- :85 ... 2: _- $86 .8: _- =5 Ema-mam $85 5.8- 38... :23- Swod 1.8- 38.0 1.. a- 33m $825? 286 $.87 BE... .3:- omfid ad _- Sm _ .o ...32- Haaaam $3 $3 ...v. G- Ea... ...-:3- 886 ES- Emwd ... _ £- mw23o=o> 5&3 £35 .3...- 5.2... :52- 803 $8- 386 3.8- $232; 5380 :25 5.9a- :2... .32- as; .5. SN- 23d 13%. Haauswm 559m $8.0 ...Zwm- 83... 34mm- 236 .2%. 3.86 ...w. am- 355m 38% 5:3 vam- 33... 3.8..- 83o max. 823 9mm».- 85 $053 8: .o 3%- 33... ...memm- 3.26 3.2-- R: .o ~3- Efiooz 8888 $85 3.; _- ES. ...m.:~- so; :3:- 886 La:- Bec 8:585 335 8: _- :3... «:3- 336 ...26 7 E: 3.87 neon-9&2 gig-E 86on “cassava-@533 33¢ 02 8_§--_ 02 can; 02 gas-m 0?. 03%; bacon Em 2232 352 32% 28 35252: 68 oEoeim 553 358985.... 2:59:00 £0602 3:582 boamomom .Paoo .70 22¢ 180 aohmo wdT mambo m.mo- mhwmo vam- mfioo mg;- _oooo.o N Koo wd T «Rho m.mo- whwwo 0.0-N- SSo ms 7 _oooooo.o 00000 000000 0060mm 00000000000000? wehmo ~400- -- -- 0035 fixe- movmo imm- ~oooo.o ochmo N60- ..- I- 336 ~00- movmo 00m- Mooooooo 00050000m 0.20.:0D coamo 0.2: N306 mfim— omCo oé: nmvmo mSC _oooo.o ooomo 0. m3 «305 mfifl oak. _ .o o6: hmvmo 002 _oooooo.o 00.000025 30,—. omeoo odfl go 0 .o o.mm_ mvmoo 0.02 mmmoo m.om~ _oooo.o oaooo odm _ 003 .o 3% _ NVmoo 0.0m _ mmmoo mom— _oooooo.o 000500003 .0000- Como 2.0mm Sumo 0..on omwmo $.de ovmmo o.mmm _oooo.o Como 2.0mm ammo €on omwmo #mwmm ovmmo o.mmm Soooooo 00—009 wfiEn0Q 0300- ammmo mg: m 326 o.~om wmmmo m0 _ m 853 0.8m _oooo.o ammo 0: m mmmuo oéom wmmmo W07.” mowvo 0.3m Soooooo 36.000003 H0009 mcamo odm- 02nd m.vo—- Nwmmo man- 2 3o wdm- ~oooo.o Kmmo mdm- eaomo méoT Nvmmo m.mm- w“ _m.o mam- _oooooo.o 000m 08060000000003 0m_m.o 0.8m omzo wag wwwfio wdom woomo 0.5mm _oooo.o cmfimo 0.8m om— 1o fin: oww _ .o w.mom woomo 0.0mm _oooooo.o 000m “00000000000003 Emmo Tom 0936 mama wmmgo “Ya-om 0m_m.o m.ww~ _oooo.o v _ mmo :Vom 090—5 Yawn wma _ .o w.mom 0m _ mo wam ~oooooo.o 000m =< 0205-0 00% 030>d 02 023-0 02 0205-0 02 030000» @0000 00m 2252 950 0232520 000 0:08:00 0:03 0002 0000000m 00009000 03000090002 0.0002 0050002 00000Q0m .2090? at. 05 3 0000003 000000 030000 003 0000.0 0 05 .0 0000000 003 $008 BEES 05 0000.0 0009 .0w00>000 000 20 _00000 000 005 00000000_ ..---.. 0000000 0&- .00000000 02 00 00000 030000, 002m 0 00.0 .0008 0300000 0008 05 0000m00m 0020> 000—0m 00000.00 00000050 3 0000: 000 «00009000 00000008 00000000 0 0002000 005 $0002 Soon-moon “00903—2 cam B00mw0m 000 0003 005 am 00 000000000 03200 00.00% w0€00 0000003 0000000 000 00000 0002500 8000-03 00 .000 00.03 003000 005 00000 008me 00000000 00 0000 0003 £0002 000.508 0 0000-30 030000 00/0200 00 000000000 0000 £05 00000000000 000000060 00.0 000 000000 0030— .005 $00000 0058 00.“ 00205-0 000 00000000 61$ 00000000 000000005 0.000002 .N-U 030B 181 5 50 00.03- «02.0 0.30- mmmwo 000mm- omowo 00.0»? 0oooo.o 550 0v.mwm- Soho o.mov- mmmwo 3.me omowo 00.90- Soooooo 0000 00000E< 00: .o 00.30a- Nv0m.o N.Nmm- 0mm _ .o 0.03”- hNE .o 0N. mvm- 0oooo.o w03 .o Rovm- mn0m.o 0000- 0mm 0 .o 0.00%- R: .o 0063- _oooooo.o 0000002 00000000 000o.o o. E T m: _.o v.0 :- So 0 .o o.0h T 000o.o 000. T _oooo.o 000o.o o._w T 0026 0.0:- 520 o.0:- 000o.o 0.0:- _oooooo.o 00000 00—00-0000 ~.Vm0o.o w. _ m T 0mmv.o 0.00m- vvvoo m.w0 T 380 0.00 T _oooo.o vm0o.o w. 0 m T 3006 0.00m- vvvoo m.w0T mhhoo 0.m0T 0oooooo.o 0000:0032 00-0005 N00m.o @000 002.6 0.0: 30mo 0.02 30900 002 0oooo.o _00m.o 0.0m _ m0ob.o m0: nw0mo Q02 mumvo N00 0 _oooooo.o 000mm 00000 _ooo.ov o.m:- 0ooo.o «Sm—- Hoooov 0.3-T _ooo.ov N.00T 0oooo.o 0ooo.ov 900T 0ooo.o NRmT 0ooo.ov 000. T 0ooo.ov N0:- 0oooooo.o 00003 005 00000 cm _0.o fiomm N036 0.3.— Nmowo ..wmm o00w.o m0 0 N _oooo.o 00 ~00 fiomm 00:6 QM: mmowd ZEN o00w.o 0.00m 0oooooo.o .000. 00w00s-0000O mom 0o 0.mm 0 mo~0.o 0.00 m0m0o 0.0m _ 330 0.. 0m 0 _oooo.o mom 0o 0N2 o—u0.o You m0~_.o Wmm 0 mN0_.o 0.000 0oooooo.o 0000- 00305-005 0920 @va ummuo 0.05 0030 0.0mm 00000 ”mum ~oooo.o 00 m 0 .o @va ommno 0.03 02; .o 0.0%” 000 0 .o 0.00-N 0oooooo.o 000002 0:00 m.~wT -- -- mm Go 0.owT 0030 o. 5T _oooo.o 0:00 0&3- -- -l mm 0 mo 0.ow T w0 0 00 o. S T 0oooooo.o =0300O _ooo.ov 0300 ~ooo.ov 0.6m 0ooo.ov 0m .00 _ooo.ov RSO _oooo.o _ooo.ov 00.00 —ooo.ov 0&0. _ooo.ov 0m .00 _ooo.ov 3.00 _oooooo.o 00000— 0003 ammo m0 _ N- 0w~m.o 0.00N- ammo 0 .0 0 m- 000N.o mémm- 0oooo.o ammo WEN- vwamé 0.00N- ammo 0.00m- 0w0~.o mémm- ~oooooo.o 003m 0002 0060mm 00000000000003 020070 D?- 00_0>-m 00¢ 02000-0 030 020070 00% 2000000» 000000— 000m 2252 0200 8038020 0000 2005800 000$ 00002 0000000m 000000000 03000000000000 20002 00000002 00000030 0.0000 .N-U 030-0- 182 0023 3:.- .Ez... 30m- 203 2:- Sowo «m:- 883 33.0 3 _ m- «$0... 3%. 203 n: m- $020 0% _ m- 5883 88- 5030 360mm 00000800000003 £85 2%. 88... 0:0- 380 v. 0%. Sood m. 0%. 883 58... 0.0-m- 38.? 9.20. 333 v. 0%. $8.0 m. 0%. 5883 an; €220 3.86 28. mam... 0.8m- 880 $2- 885 0.9:- .885 885 0. am. can"... 0.2m- 0830 302- 80:. 0.8 0- 5883 Ea- 0020 $86 2: _- 30.... ...—£- Eood 2: 0- $85 02 _- 883 $85 2: 0- £3... 32. :85 2: 0- $86 0: _- 8883 :90 Ban-0&0 $85 c.8- 223 0.20- 880 03. wwwod «.8. 883 $de c.3- 005... 08". Swod Za- mwwod v. a- 88°85 200m $825, some $2- 3%... 3:- 525 mm:- Swg 32- 883 m 83 32- ER... 00:- £26 0.2 0- $2.0 $2- 8883 3023 033 $3 v. :- mna... 03- 885 c.3- Emwd 0.30- 883 NS; 0. :- 05? 0&- 885 0%- E03 1&- Soooood $2320.» 8&3 00-0... :3. 22... 0.20- 032 30- £005 08- 883 306 $3. 50:. 2.2- 0003 30- £80 000- 58830 02320.» 5320 :26 Sam- 38... 0.0:.- omwS 003- 5:5 08m- 0885 32.9 c.00- 05; ER- 033 0:0- 50¢ 00%. 8883 30023 Ezom $05 20m- $3... 0.3-m. 22:. 25m. $86 0. SN. 8256 303 3:0. 083 ER- 285 2%. $8.00 3%.. 8883 snacfim 80am mom-comm EDUGQQoU-fiflwfi—DB 02950 02 039-0 02 2.030 02 can; 02 05%; 0050 20 2252 950 85058: $00 050.5% 020$ 000:2 000008m 000000000 0>mmm00w00000< £0002 0005-00—2 00000030 0,0000 .N-U 0300- 183 184 I i .x Nmomd ...oda- as -- ocmmd ...v. 87 emmmé 2.3? :80: 50.5 3.de 36mm 336 ...Ndmm 3.36 ...w. Sum 3.3.: «adv-m “Ewm 28.5 23.: 2.8 ovnmd ...—do 336 ...néo ocomd ...Nde 883 02m 380 vvomd Lumwm vmmmé ..mévm Bad ...—.mwm vw _ md ...vwnm Rn:- Emacs-520 memod Emmm «was... «Nfieu mmmod 01mm bwmod ...mdmm 30,—- Egg-03m w Sod .5. wow 33.: «mdam -- -- ”mood ...mdov BEBE mmeeé «awn -- -- N306 *odm flood ...hém £330 Soodv .583 33.: ”62 Soodv ...Nde Soodv ...NMm— xosm coo? Newmd Was 35.: «md— Sofie ...NSN. owned Ede 53m 822 anomd $.th nae-mm... «WEN cmwmd ...wfimm «memd 3.0mm 3000 «9380 $6on «newcomoc-wcmzokz nomad ..de 33.: «mi: mmvwd $63 $36 1:: mEEouoam ”EBB-u amend *mdnm «Ev... Luann mnmod R63. 2R6 .Ldg mEEEonm 130,—- oavmd ...N. am onwfio ...néom mummd *hdom scum... 5.wa mEEBBB =88- Ebod c. 3% @356 @va good 09% 33.: .3.va mxosm mam—snag =88- wmmod Ecov mmmud «v.3»- o _ mod 33v vowed 1.3m Eofifia? =88- mhmod odmm «~35 ......ch wmmod ...mdg wowed 363 2:5 356836.530? gmod 36m $3.: 3.3m camod Edam :36 *m.mmm mwhm ”newconov-vcwzoB mmmod 36m 33.: 3.5:.” $86 33m 836 EN? mvhm =< egg-m 02 0:13-; OZ 023d 02 023$ 0?. 03mtw> 5653 BE 2082 352 355m 23 85358: $8 ofiogm 92$ 033323? 2:59:00 2252 3:5on wofiomom .owbéoo 8: Eu BEE 05 35 886me ..---.. cargo: 2E- 9?wa 2563 “oz 33 ESE :ozw 05 Ba £me 0 05 .t 33> 02 5 Sta coon—q v.95 is. xmtfimm a< .85:me 02 no 3me East? swim a no.“ Evofi oBfifiov H88 05 83:2: 839» cow—om .9335 8.8550 3 new: 03 Encomfioo $532: 38qu a 39:05 35 £352 Seem-meow .Swwfloaz Sam Bmcmmmm 98 33”— :30 am a 383:8 @833 23on wgc €3.63 wow—6:: was. 98% E Aowvo Ex 5Q £58 25 #:3me mo 823% Sec: 2388 8 wow: £358 3me Sn moflg-m EB mozmufim 67¢ H8550 song-:85 fleas? .m-U £an I [II \ 2:3 .2: 2.2... 33%.. New; Em we; .o *3- Eve Ego-mu 2 8o 2 a $2.... 3.3 3.26 .32 ammo $.32 502.9 $6on cowmmoommm-vcwcoa 38¢ .88- 88... .3.:- 886 .30- 886 3%- use $292 OR; .6: 23... :3:- ommg *5: a; 33 ea- x85 38¢ 38 82:. ...-2:: 886 .802 886 3&2 =5 8:3-w5m :25 3% ES... :3 £26 *9: mg; ...m: aim peas? 33¢ ...m. :2 £8... 3:2 $de .55 83¢ a. E saaofim an»: 9:2 Emma emu-... 3.ch 8de 368 £52 Boom $232; 5&3 $85 5.8. was... 3.93 EVE *2: 8mg *3: $232“; 5320 mg; .53. 3.5.... 3.3;- 5:5 L. 6. Man; 3.8- .aacam 550m 326 ...SN- .2%... 3;- $33 3.8- 52d .88.. Haagm 8:on ammo 3a. 3.2 .33- ~82 v.2- 253 SM. 88 5055 33¢ .82. 8%... 3.27 252 ...mé- $on gm. 5&8: 8888 ES :32 2.9... as E; 38 $de 83 320 8:5-qu Sm _ .o ...de 8:... 3.3- 38¢ 1%- w 83 *5 852-232 sigma 86on “coccoaov-vcwuog 02?-.. 02 83¢ 02 029$ 02 can; 02 osafi> bacon Em 2282 8me usufim 23 8:85:22 Gov abuse-Am 233 0332383 952580 $3022 $5802 wowaomom 63:00 .m-0 2an 185 I — _ K. onomd osmm hmmmd QEN Ewmd wsmm .3on vomm _ooood anomd osmm bmmmd We: omwmd wfimm wovmd vdmm Soooood 880 3980 momooam “coccomovésw—BB mowod 0on wand md: mmvwd mew: woowd o. K_ Soood mowod mdfi crowd md: mmvwd még woowd o. E“ Soooood moufieonm E339 owond Wmnm owned fiuvm Numod adv». Std fidvm _oooo.o owond Wm? wooed hmvm mumod mdvm Std "din flooooood mEEBonm €on oovmd N. 8N and o .o Edam mummd ndom ooNNd ”5N oooood oovmd m. Sm amid ndom mummd ndom oound «Son Soooood mandates? 30% good E43 omvod 3.3% good .3.va 33d «.mmv _ooood good ho. 3v omvod 3.3% Nwood +02% 38d flame _oooooo.o 33:9 wEEan 30,—. wmmod Fdov mmmmd Yvon o_mo.o $.on vmvod fiddm _oooo.o wmmod Kdov mmmnd 35m ofimod $.on wovod fioom Soooood gotofia 30H mnmod Edam ouood odoN ammod Nda womod oéom Soood mnmod hodmm wuood odom ammod md—m womod oéom Soooood mohm 3368365303 mwmod :dom nobod “den oomod fioom Evod mdmm oooood NVmod Zdom Nobod flown oomod Eoom Svod m.mmm Soooood mgm oceanoQQVvSw—BB ammod hodom vomod vfivm womod Fodom movod odmm ~ooood mmmod hodom vomod v.53” womod Fodom movod odmm Soooood 2:5 =< 0.5—go 02 2:?d 02 2:97m 02 “BETA OE 950m 05mtw> 3659 BE .2252 23 8585:: $8 050.5% 25$ 533 @822 oaugm 9:59:00 omeEwBonx 2252 8:582 3303M dong? at. 05 3 38:03 38on “Eamon 83 x59: 0 2.: .2 outage“ mm? 3.38 353.8 05 Soho 8mm .owhogoo 8: Ed 688 on“ 35 moomomofi 2:-.. 8:80: 2: .8593 02 co woman 03.2.5, 55% a 8m E58 Banged ~88 05 38mg: 829 ooEom .8335 028550 3 com: 08 “cocoafioo 85308 gauge a 3333 65 £382 Hoomdoom .SwwEomE Sam Baamwmm 28 32m :30 .Hm 5 6826200 @823 958m muggy moan—$3 voice: 98 dead congen— Aowoo Ex “on £58 323% BB “:8me 80:: 089:8 3 wow: 0.53 2252 .mooEmE 0 command 0358 9628 3 83:58 80w :53 £325.83 023938 So How $559 $32 53> £088 3me So mofigi Ea mozmsfim 62o :otoEU Gowns—SE 98:82 6-0 2%; 186 o _ mmd h _ .wm- nmomd Yon- moomd timm- mwomd Z .om- ~ooood oENd tom- oNom. o Yon- moomd Zion- mwomd de- oooooood 800 S855 moomd 1.6m- vaod o.m~ T omomd mow. moomd Rom- ~ooood moomd Rom- ooNod 32—- owomd mow. ooomd Rom- _oooooo.o 5:802 :oEEoU oonmd mdo_ vomvd ofio hm _ vd o.oo ovwmd NR .ooood oRmd wdo_ vomvd o.~o mm :d o.oo oVwmd NHo oooooood mop—O 3:560?— _om_d mdm woe—d moo- mmood 1m- w—ood w; _oooo.o Sm _ .o wdm doe—d moo- mmood v6- m _o_.o M: _oooooo.o :oSIéEZ Evian—m mmomd odd- --- --- oommd v.57 ommmd fimo? _oooo.o mmomd o.oo- --.. --- oond v.57 ommmd —.No—- _oooooo.o :80: 580 vamd odmm va _ .o Ndmm mnvmd w. Sum uvoud o.mmm flooood cmvmd odmm cm: .o Ndmm mnvnd x. 3m :oNd o.mmm _oooooo.o Sam .on0 39nd fimo ovnmd fiwo owmmd fivo ooomd mdo oooood m—vud fimo ovumd fimo owmmd néo ooomd N.mo _ooooood 882 0:5 320 «vomd Emmm mound mdvm nond fimwm ¢w~md «.mnm _oooo.o gomd fimwm wound mdvm poemd fimwm Sfimd comm ~ooooood Rob oowEBéoBO womod RNmm Noood NfioN mmmod hm. _ mm mm mod mdmm _oooo.o womod Rdmm "wood NHoN mmmod ..m. _ mm hwmod mdmm Soooood Hob oowcosbfim m _oo.o o. mow voood mdam --- --- wmood odov _oooo.o w _oo.o Qmov voood mdom --- --- wmood odov _oooooo.o ohm—RE vmood mom --- ..-- mmood o.om whood Eon _oooo.o mmood mom --- -..- mmood o.om whood mom _oooooo.o .3530 Newmd hmdh oohvd md— ooomd NR. osmod ndo oooood vamd omdn oohvd md— ooomd NR oomod Woo _oooooo.o 53m 832 $625 “coocaooécwooa osfi>i DZ o2m>d 0?. 023i U~< magi 02 2355 5650 BE 2252 23 85032: God 0.525% 32$ ooxaz oSoQSm canon—Sou o>mmm2meo§< $082 85302 332mg .o.EoU 410 Bank 187 mo _ md o. _ oom—d «in- No _ .o oN wmid v.5- _ooood vo _ Nd o._ Em—d aim- mvo _ .o o.m wow _ .o v.5- Soooood Eco. Saammo 3 Ed ,3.— _N ovovd o.m~— ommvd v62 cmmmd mi: Soood o2 3d 22 R omovd owe ommvd v.32 memd oi: Soooood Boo—=2 momooom ooommoOmmméqm—Hog woood 5.mo- noood WET ooood oéo- ooood oéo- Soood moood 5.mo- Noood m.—o—- ooood oéo- ooood oéo- Soooood Eco. mzofiom omm5d 5.: ovmvd o.o5—- omm5d 5.2 $m5d 5.2 _oooo.o ovm5d 5._ 2 wound od57 omm5d 5.2 $m5d 5.2 Soooood Eco. xoflm ovood ..Ndom moood o.—o~ moood mda moood v.32 Soood ovood ..Ndom moood o4: moood mde moood Two" Soooood =20 “go—gong 582d odo «wood 3% 32d o. 25 m8 _ .o m. _5 ~ooood 532d odo a5ood 3; 32d o._5 Qua—d 025 Soooood 032m whom—c5 mmmod m. :1 Bond Yum womwd 5.52 oomod o. 2 2 Soood mmmod m. 3; mmomd ado womwd 5.52 oomod o. _N_ Soooood Hoofiocmm 584 mm Ed odmm owned odon oommd odmm wwomd Noon _oooo.o mm 5d odmm omavd odom oommd odmm wwomd Noom Soooood mwo_>>o=o> .533 momod o.m2 m5~od m.vo— 595d W552 momod o.m5_ ~ooood momod o.~o_ m5~od m.vo_ 525d m.55_ momod o.m5_ Hooooood mw2>>o=o> .6880 m5w _.o 5. _o.. mm5vd 34;- 2mm _ .o 5. 2o- wm2d moo- _oooo.o m52d 52o- mm5vd Q37 32d 52o- wm2d ado- ~ooooood 5935mm Ezom museum Housemooéamoog magi 02 magi 02 @2952 02 263$ 0?. oEuE> 5650 ohm 20%: 23 83252: $8 oEoEEm 32$ 352 Emocfim oasoofioo 0382382340.. £352 8.5302 Banned .380 #0 ”BS 188 58.8 8.3385 28.8 8.3385 8.8 8,338.3 25.8 8.325 28.8 8.385 88 8.38.2 85.8 8,338.3 5.8 :385 85.8 8338.3 8.8 8.3383 8.8 8.385 85.: 838.2 85. o 2.385 85.: 2.382 28.8 8.385 28.8 8.385 28.8 8.385 2.8 8.385 8.8 8.385 8.8 8.3385 85. o 8.353 28.8 8.385 828 8.385 28.8 8.385 2.8 8.385 8.8 2.385 5.8 8.388 35.8 8.385 85. o 8.385 28.8 8.385 82.8 8588.3 85.8 8338.3 28.8 8.385 92.8 8338.3 8.8 8338.3 85.: 8.385 85.: 8.385 858 8.325 85.: 8.325 82.8 8.3383 85.: 8.385 85. 8 22.385 85.: 8.385 85.: 8.385 8.8 8.385 808 8.385 858 8.3385 82.8 8.385 85. o 8.385 22.8 8.385 88 8588.3 858 8.385 85. o 8.385 28.8 8.385 88.8 8.385 85. o 8.385 8.8 8.385 owns :8 $ch 88.56 Anus 88m ole =8 QUE 88:8m Qua 88m :18 :8 Amnsv 88:5m fins 828m 88 :8 Amuav 88:5m ans 8:8 ole =8 Amncv 8:255 fins 882m :18 :8 GUS 888m fins 888 owns :8 GUS 885.6 Anna 8:8 85 88% 8.0.80 «880 283m 88an war/5 ”325 8:85 895883 88.5 238883 88.5 85.: 83:5 8.8 8.325 85.8 8.385 88.8 8.385 28.8 8.385 88.8 8.325 3.8 8.385 85.8 8.3385 85.: 8.38.2 2858 R385 85.: :385 28.8 8.385 85. o 2.385 85. o 8588 85. o 8.385 88.8 8.385 85.: 8.385 85.: 8.385 . 85. o 8.35: 85.8 8.385 8825 83.5 88:: 835 8m BuEwam 82m 285 am 88on 58.88 880% 05 58> 8828.: .20 822888 05 08 888:5on .5oom-moom wet—.6 288—83 8588 S8232 ism Bufiwam o8 82m 88—0 5m 8 88388 8823 E88 3 matso 252,830 880% 23883 28 30.2883 888m 8.2 3:3 88m 28 .25 28.83 .88 595m 3 5885888 .5 momocosooa o8 .288 2582588 Amémohov 838.825 582 .m-U 038.5 189 3.8 8.385 8.8 8,338.3 85. 8 8.385 25.8 8.3385 88.8 8.3385 8.8 8.3383 28.8 8.325 858 8.3_ _5 8.8 8.385 85. 8 8.385 85.: 2.385 85.: :385 28.8 8.385 8.8 8.3383 5.8 8.385 8.8 8.385 28.8 8.385 5.8 8.385 8.8 8.385 28.8 8.385 808 8.385 23.8 8.3_ _5 5.8 8.382 8.8 8.325 85. 8 8.385 21.8 8.3385 85.8 8538.3 22.8 8.385 888 8.385 22.8 85vfl55v 2 _.8 85vi55v 8.8 8.385 85.8 8.3385 88 8.385 2 _.8 8.3383 85.: :385 85.: 8.385 85. 8 8.385 82.8 8338.3 82.8 8.3383 88.8 8.385 8.8 8.3385 88.8 8.385 5.8 8.3385 88.8 8.385 A28 85vfi55v 85.8 8.385 85.8 8.385 8.8 8.385 88.8 8.385 85.: 8.325 8.8 8.3383 88.8 8.385 9N8 =8 Amncv 85:5m fins 8:8 :15 :8 Amncv 8.55% 88m 880 GUS wctmm 508E 82m 820 2H8 =8 3H3 8:55m 88 8:8 ans :8 GUS 8.555 88 8:8 8.1.8 :8 QUE 855=m ”88% 5382 Qua 8:8 225 88 .5. 88.8 8.3385 28.8 2.385 :18 :8 88 8338.3 88 8.3383 ans 3888 85.8 8.3385 E8 8385 808 8.3385 fins .8ch 5883 .e< 28.8 8.385 :15 :8 8.8 8.3383 88 3558 85.8 8,338.3 28.8 8.385 c _.8 8.3383 82.8 85v385v fins 88m 8383 588:: 85a 8582 885 8288 flm Bacmwmm m8?— ._8_U 5m .PEoU 6-0 038,—. 190 555 5.355 5555 5.3555 555 5.385 5555 5.385 820 355-85 555v 55vfi55v 0.8: 55:: 2555 5.3355 G: 55 5.355 “8505: 588: @555 55vfi55v @555 55v£55v 53:05:: 5555 5,335.3 $555 55vfi55v A5555: 862E 5585 a _ 55 5.3355 v.8: 58.85-55: A555 55vi55v €555 55vfl55v 5853: A555 55vfi55v 0.853230 5555 N535: $55 5.355 5:55 5385 5555 3.3555 35 8553-585 5555 5535.3 @555 5535.3 :95: E252 5555 5385 A855 55va55v $555 55vfi55v 5385 82582 6555 55.325 5555 3.355 5555 5.355 5555 5.355 58: 8553-02: 55. c 5.355 355 55.355 55.: 2.355 5555 2.355 5332 5555 5.3355 5555 5.355 A855 55vfi55v 5555 55vi55v :8: V.22: .5055. 5555 5.355 5555 5.385 3555 55vi55v 585:3 555:2. @555 55vfl55v 5:55 5355 3353 5555 8.3555 @555 5.355.: 5555 5.355 55.: 3.33: :8: 853 €555 55va55v 53m .2585 5555 5,335.3 5555 5.325 555 5355 9:55 5.355 53m 2:2 95.55 5 _ .35 _ 5 5555 5385 555 5.3355 5 _ 55 5.355 380 8:30 . 860mm 50303065303 mmus 8525 535 35: d 35 355:: 535 585: museum xmm BuEwmm 90$ :20 am 602030 003 505.0% 05 85 £033 90 53:38: 05 mm 00:05.00: .«o >0:0:c0:m .noom-moom wES—u 0:50:03 3980 gwfiozz .xmm Kmfiwmm “5:: 305 :20 5m :0 60:26:00 9033 350% zfl 33050583 0:: wES: "002030 860% BB :8 093 “0:030? :5 :03 3.55 3 $80558: :6 00:05:08 mo 50:03:08.: :5 “5:0 30:55 Anni—0:8 335:0: :002 6-0 030,—. 191 E8 8338.3 8.8 86vfl66v 8&8 8386 E8 8.3386 E8 8386 8&8 86“: _6 E8 8,338.3 E8 8.3386 8.8 86vi66v 8.8 8.3383 86.8 8338.3 86.8 86vfi66v 86.8 86vfl66v 8.8 8.3386 8.8 8.3383 8.8 8338.3 8&8 3.386 86.8 8.386 8.8 8338.3 2&8 8.3386 8&8 8.386 8.8 8,338.3 8.8 8.3386 8&8 8.326 E8 8386 8&8 8.386 8&8 8386 8.8 8.386 3.8 8.36; A88 8386 8.8 8386 E8 8386 8&8 8.386 86.8 8.3386 8.8 8.3386 8.8 8.3386 8&8 8.326 8&8 8.386 8&8 8.386 8&8 8386 88 8.386 8.8 8.386 8&8 8.386 86.8 8.3383 8&8 8.3386 8&8 8.3386 8.8 8.3383 8.8 86vfl66v 28.8 8.386 8&8 8.3386 8.8 86vfi66v 8&8 86v386v 86.8 8338.3 8&8 8.3386 8.8 8.3386 8.8 8338.3 8&8 8.3386 86.8 8338.3 2 E8 86v386v 2&8 8.386 2 8 8.386 8.8 8.386 8.8 8.386 8&8 8.386 5.8 8.386 E8 8.386 8.8 8338.3 E8 8.3386 $8 8.386 E8 8.386 c _.8 8.3386 8.8 8.3386 8&8 8.386 E8 8.3386 3.8 8.386 86.: 8.386 8&8 8.386 8&8 8.3386 86.8 8.3386 2:55 bafiugm .9303 Homavcam thmm Samarium 834 .SQEEBm onmEEEEom mwo_Bo__o> Momma: $2323.» .8820 36888 5:8 Samarium 338m BEE BEE—«£80m 0:80 =Evcwm Sou 525:3 5:802 8:580 Sow .5 «88$ hog: F5582 552-2% _ Z cog/80-865 .883 580 Houwm 880 .85: 0:5 820 5885 834 5035 $0593 8:88:00 3820-0350Q 86on Eoncomocécmzoa 23$ 38?: 838 883 GT8 8825 8E8 685 Nam Bacmmmm 36: £6 .8 380mm 6.58 6-0 268 192 86.8 8,338.3 8.8 8.386 :88 83:6 868 8.3383 E8 8.3383 8&8 8338.3 88.8 8.386 8.8 8.3383 86.8 8.386 8&8 8.386 86.8 8638.3 58 8.386 86.8 8,338.3 86.8 86v386v 88.8 8.386 8&8 83:6 868 8338.3 86.8 8338.3 818 8.386 8.8 8338.3 8&8 8.386 86.8 8338.3 8&8 8.386 86.8 8.3386 8.8 8338.3 63.8 8.386 86.8 8.386 86.8 8.386 68.8 8.3383 28.8 8.3386 86.8 8,338.3 86.8 8338.3 8.8 8.3386 :&8 8.386 2 _.8 8338.3 :68 8.386 86.8 8338.3 88.8 8.386 5.8 8.386 86.8 86V386v 86.8 8.3383 22.8 8.386 86.8 8338.3 Awmdv modfimfio Eng. :oEEoU EDP 58580 “82:18 858 3:28-885 8:02 2me 2mm momooaw 386836533 Honmmwcg cog—om Eek {880m 88 8.85 86 8E»: :5 885-88 :50 Moran—«gm oqoaasm coxoocéom 808.6003 50tu§< gnaw FEES, H8828 888-268 5885.8 :8 $6on Housemovécmzoa 838 88:25 838 88.5 8.18 88:8: 8.8 885 mam 39:me 33m :20 6m momuomm .PEoU 6-0 2an 193 NIVERSITY LIB MITIGWIHsllTlTHfMflH"I HIHII II” |\ ”W 3 12 9 3 0 3062 9087