HABITAT VALUE OF CREATED WETL ANDS TO WATERBIRDS IN GOLF COURSE LANDSCAPES By C. LeANN WHITE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003
Copyright 2003 by C. LeAnn White
This thesis is dedicated to FloridaÂ’s waterbirds.
iv ACKNOWLEDGMENTS I am extremely grateful to my committ ee for their expertis e and support throughout my thesis work. Dr. FrederickÂ’s and Dr. PercivalÂ’s knowledge of waterbird ecology was instrumental in the development and design of my project. To my advisor, Dr. Main, I am particularly grateful for guidance in my development into an independent scientist. I thank the Wildlife Links Program (inc luding the National Fish and Wildlife Federation and the US Golf Course Associat ion) and the Bonita Bay Group for funding of this project. I am indebted to those associated with Bonita Bay Group and Watermark Communities Incorporated for the use of their facilities and equipment and for their patience while I conducted field work on thei r golf courses. I thank Maena Voigt and Kristin Miller for their assistance and patie nce with sometimes tedious field work. I thank Dr. Ken Portier for his guidance and assistance with appropriate statistical procedures. I also express my gratitude to Ginger Allen for her as sistance, guidance, and patience during all of my stints in Immokalee. Finally, I thank my family and friends for their encouragement throughout this process; and for never ceasing to ask when I w ould be finished. I am especially grateful to Matthew Reetz for constantly reminding me that I would finish; and of course, for the use of his expensive paper.
v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv ABSTRACT.....................................................................................................................vi ii CHAPTER 1 INTRODUCTION........................................................................................................1 Wetland Losses and Waterbird Population Declines....................................................1 Created Wetlands as Habitat for Waterbirds................................................................3 Project Need..................................................................................................................6 Study Objectives...........................................................................................................6 2 REVIEW OF HABITAT SELECTION BY WATERBIRDS......................................9 Overview of Waterbird Habitat Selection....................................................................9 Foraging Site Selection...............................................................................................10 Effect of Landscape Features on Foraging Site Selection..........................................10 Effect of Surface Area on Foraging Site Selection.....................................................12 Food Availability........................................................................................................13 Effects of Water Depth on Foraging Site Selection....................................................14 Effective Foraging Area: Effects of Littora l Zone Area on Foraging Site Selection.15 Effects of Water Productivity on Foraging Site Selection..........................................16 Effects of Shoreline Vegetati on on Foraging Site Selection......................................17 Effects of Surrounding Habitat on Foraging Site Selection.......................................19 Species Interactions....................................................................................................19 3 METHODS.................................................................................................................23 Study Species..............................................................................................................23 Study Area: Golf C ourse Selection.............................................................................24 Waterbird Surveys......................................................................................................25 Foraging Guild Classification.....................................................................................26 Surface Area...............................................................................................................26 Perimeter.....................................................................................................................2 7 Effective Foraging Area.............................................................................................27 Shoreline and Surrounding Vegetation.......................................................................28 Productivity Measurements........................................................................................31
vi Water Depth................................................................................................................32 Distance to Nearest Colony........................................................................................33 Analysis of Pond Similarity: Effective Foraging Area and Shoreline Vegetation.....34 Analysis of Site Preference.........................................................................................35 Interpretation of Regression Models..........................................................................37 Analysis of Golf Course Comparison.........................................................................37 4 RESULTS...................................................................................................................46 Waterbird Observations..............................................................................................46 Foraging Guilds..........................................................................................................47 Shoreline Vegetation Variables..................................................................................47 Analysis of Pond Similarity: Effective Foraging Area and Shoreline Vegetation.....47 Productivity Measurements........................................................................................48 Foraging Guild Site Selection.....................................................................................49 Probability of Observing Diving Birds................................................................52 Probability of Observing Open Water Waders....................................................53 Probability of Observing Dense Vegetation Waders...........................................54 Probability of Observing Di pping and Dabbling Foragers..................................55 Probability of Observi ng Moist-soil Foragers.....................................................56 Probability of Observing Aerial Foragers...........................................................57 Analysis of Golf Course Selection.............................................................................58 5 DISCUSSION.............................................................................................................86 Effects of Pond Size on Waterbird Site Selection......................................................87 Effects of Effective Foraging Area on Waterbird Site Selection...............................88 Effects of Vegeta tion on Waterbird Site Selection.....................................................88 Habitat Selection by Foraging Guilds: Pond Variables..............................................89 Diving Birds........................................................................................................89 Open Water Waders............................................................................................91 Dense Vegetation Waders...................................................................................93 Dipping and Dabbling Foragers..........................................................................94 Moist-soil Foragers..............................................................................................97 Aerial Foragers....................................................................................................97 Habitat Selection by Waterbirds: Comparison of Golf Courses.................................98 Comparison with Natural Areas...............................................................................100 Potential Health Risks Associated with Eutrophic Conditions.................................102 Conclusions and Suggestions for Future Research...................................................103 Management Recommendations...............................................................................105 Diving Birds......................................................................................................106 Open Water Waders..........................................................................................107 Dense Vegetation Waders.................................................................................108 Dipping and Dabbling Foragers........................................................................109 Moist-soil Foragers............................................................................................110 Aerial Foragers..................................................................................................111
vii LIST OF REFERENCES.................................................................................................112 BIOGRAPHICAL SKETCH...........................................................................................121
viii Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science HABITAT VALUE OF CREATED WETL ANDS TO WATERBIRDS IN GOLF COURSE LANDSCAPES By C. LeAnn White August 2003 Chair: Martin B. Main Major Department: Wildlife Ecology and Conservation As increasing human pressures continue to alter and degrade natural wetland habitats, alternative habitats such as creat ed wetlands may become increasingly important to wetland-dependent species. However, de spite technological a dvances in wetlands creation, this area of wildlife ecology and manage ment is still in its infancy and questions remain as to whether created wetlands are ecologically equivalent to natural areas as habitat for fish and wildlife. Therefore, al ong with preserving and restoring the function of existing wetlands it is important to dete rmine the habitat value of created wetlands such as those in urban landscapes. In Florida, wetland impoundments are ofte n created as water supply reservoirs and/or as storm-water detent ion areas to cope with pronounc ed seasonal differences in rainfall. Golf course ponds and impoundment s in Florida are often used both to store water for irrigation as well as prevent extensive flooding during the rainy season (approximately May-October). These semipe rmanent water bodies may also provide
ix habitat for waterbirds. As natural wetland habitats continue to be altered by human development of the landscape, future populat ions of waterbirds may increasingly depend on their ability to find suitable habitat in ur banized settings such as golf courses. To determine the habitat value of construc ted golf course ponds to waterbirds, I quantified the abundance and di versity of waterbirds using 183 golf course ponds in southwest Florida from January through Ap ril 2001 and 2002. I also quantified habitat and hydrological features of the ponds to de termine their influen ce on waterbird site selection. The variables quantified during this study were chosen based on existing literature of habitat selection by waterbirds, particularly t hose considered important for foraging site selection. Hydr ological variables included trop hic status (total phosphorus, total nitrogen, and chlorophyll a ) and effective foraging area available in the littoral zone of the ponds. Habitat features included shor eline and littoral zone vegetation type and cover, and adjacent land-use features (e.g., gol f course, residential housing, construction). A total of 10,474 birds representing 42 sp ecies were recorded during the 2-year period. The species were categorized into 6 foraging guilds, which were used both for analyses and management reco mmendations. In general, the results from this study indicate that golf course ponds are capable of attracting many species of waterbirds. However, analysis of site pref erence resulted in a wide range of habitat variables selected by each foraging guild. This finding coupled wi th the low densities of birds (<2 birds/ha for most species) suggests that the value of golf course ponds may be enhanced through habitat modifications designed to appeal to specific guilds.
1 CHAPTER 1 INTRODUCTION Wetland Losses and Waterbird Population Declines Throughout history, wetlands have been m odified for human development (Pineau 2000; Tiner 1984) resulting in a lo ss of more than half of th e wetlands in the coterminous United States between the 1700s and the mid-1980s (Tiner 1984). It is estimated that a net loss of more than 3.7 million hectares (9.1 million acres or 8.5% loss) occurred from the 1950s to the 1970s alone (Fraye r et al. 1983). Most of this loss can be attributed to the conversion of wetlands to agriculture (87% of recent national wetland losses, Tiner 1984), due to economic incentive to maximize earnings from all available land. In urban areas, wetlands losses are ofte n the result of developments along rivers, lakes, and coastlines that often have associated wetla nds. As urban developments continue to expand, inland wetlands are also lost to commer cial and industrial development, or filled for housing (Weller 1999). Once wetlands are drained or filled to provide additional living space for humans, the need for flood control (i.e., further draining, filling, and channelization) and competition for water s upplies is also increased (Kushlan 2000a). Although wetlands were once considered wastelands, their environmental (e.g., water quality maintenance), so cio-economic (e.g., flood contro l), and fish and wildlife values are more widely recognized today (T iner 1984). For many species of wildlife, wetlands are essential habitat for all or part of their life functions (i.e., foraging, nesting, prey production, or drinking wate r). Wetlands are particularly important for birds, as about one-third of North AmericaÂ’s avifauna uses wetlands (Kroodsma 1978). However,
2 all wetlands vary in their suitability (i.e., habitat quality) for wetland avifauna (Weller 1999) and many of the remaining wetlands have been altered functionally through human intervention (Kushlan 2000a). One of the most important alterations to wetland function is changes in hydroperiod, which influence almo st every aspect of a wetland from plant growth and germination, to prey pro duction, and bird use (Kushlan 2000a). As the result of the Swamp Lands Acts during the 1800s, Florida and many other states drained much of their wetland acreage for agriculture by c onstructing levees and drainage ditches (Tiner 1984). Further drainage took place in response to severe flooding in south Florida during the 1920s and 1940s, wh ich led to a massive flood control project conducted by the United States Army Corps of Civil Engineers. By reducing flooding, the south Florida project also accelerated fil ling of other wetlands for urban expansion of coastal cities in Dade, Broward, and Palm Beach counties (Tiner 1984). By the mid1980s, more than 46% of FloridaÂ’s wetlands were lost (Mitsch & Gosselink 1993). The function of many of the remaining wetlands in Florida has also been degraded through alterations of hydrological regimes intended to alter the flow of water or timing of inundation (Tiner 1984). Such changes to th e hydrological conditions may decrease the availability of prey particul arly for wading birds, which ha s also been suggested as a limiting factor on nesting su ccess for these species (Fre derick & Collopy 1988a). Whether through loss of area or degradation of function, alterations to wetlands have negatively affected colonial wading bird populations, which d eclined from an estimated 1.5 million birds in 1935, to about 0 .25 million in the 1980s (Tiner 1984). The magnitude of decline in wading bird populations after anthropogenic alterations of freshwater wetlands in Florid a illustrates the utility of these birds as
3 indicators of the quality and quantity of remaining wetland resources (Custer & Osborn 1977; Maxwell & Kale 1977) . Furthermore, with almost all of the Atlantic coast wetlands altered to varying degrees (Tiner 1984) , there are also cons ervation concerns for other wetland dependent species such as shorebirds (Brusa ti et al. 2001; Senner & Howe 1984) and waterfowl (Rusch et al. 1989). Therefore, conservation efforts for wetland dependent birds should be aimed at preser ving remaining natural habitat as well as determining the ability of these species to use altered wetland hab itats (Kushlan 2000a). Created Wetlands as Habitat for Waterbirds Historically, wetlands have been constructe d for a variety of pur poses such as water storage, flood control, sewage treatment, aquaculture, mos quito control, and waterfowl management (Rey et al. 1991; Roberts 1991). However, it was not until the 1980s that man-made wetlands began to be extensively used to mitigate for losses of natural wetlands (Roberts 1991). Presently, under Section 404 of the Clean Water Act, mitigation in the form of wetland creation or restoration is often required to replace wetlands degraded or destroyed by huma n activities (Brown 1991; Roberts 1991). Wetland restoration, enhancement, and crea tion are now common practices used to reverse, reduce, or compensate for wetland losses due to development or other human activities (Simenstad & Thom 1996). However, it has been argued that the practice of creating a wetland to replace a natural wetland is appropriate only if the created wetland will be equivalent to, or of higher quality than, the natural wetland (Brown 1991). The ability of a created wetland to support life functions (nes ting, foraging, etc.) and not merely reproduce the structural characteristic s of a wetland serves as a more adequate standard of the functional equivalency of a created wetland (Simenstad & Thom 1996).
4 As increasing human pressures continue to reduce the amount and alter the hydrological conditions of wetlands nationwide, alternative habitats such as created wetlands may become increasingly important to wetland-dependent sp ecies. Birds are among the top species affected by the degrad ation of wetlands and a number of studies have been conducted in created wetlands to determine their value to wetland avifauna (e.g., wading birds and shorebirds, Brei ninger & Smith 1990; Melvin & Webb 1998). Despite technological advances in wetlands creation, this area of wildlife ecology and management is still in its infancy and questi ons remain as to whether man-made wetlands are ecologically equivalent to natural areas as habitat for fish and wildlife. Furthermore, little information is available as to the wildlife value of wetlands created for mitigation or other purposes, such as recrea tional, flood control, or aesth etic components of urbanized landscapes. Waterbird populations are directly in fluenced by the amount and quality of available foraging habitat (Bancroft 1989; Kushlan 1978). Factors that influence the quality or suitability of foraging habitat va ry among species because of differences in morphological and behavioral ad aptations. These adaptations influence preferences for different physical features such as water depth and, ultimately, reduce direct competition for food resources among species (Jenni 1969; Kushlan 1978). Differences in habitat requirements make conserving foraging habitat more complex than protection of nesting habitat for some species of waterbirds (e.g., wading birds Kushlan 2000b). Thus, as natural wetlands acreage continues to decline, waterbird use of alternative foraging sites must be considered.
5 Artificial foraging habitats such as ditche s (Kushlan 2000a) and rice fields (Fasola & Ruiz 1996) have become increasingly impor tant to wetland birds and may occasionally offer more prey than nearby natural habitats . Artificial habitats may be particularly important to herons (Kushlan 2000a); which are extremely mobile, ecologically flexible, and able to take advantage of various human-a ltered or created wetland habitats that offer sufficient prey (Kushlan 2000a). In Florida, wetland impoundments are ofte n created as water supply reservoirs and/or as storm-water detent ion areas to cope with pronounc ed seasonal differences in rainfall. Water supply reservoirs function as short-term sources of water for agriculture and other industries. Detent ion areas are designed to regu late storm-water run-off by temporarily holding the water so that it can be released graduall y, thereby reducing the impact on downstream systems (Adams et al. 1985; Boman et al. 2002). Golf course ponds and impoundments in Flor ida are often used both to store water for irrigation and to control (i.e., preven t extensive) flooding during the rainy season (approximately May to October). These se mipermanent water bodies may also provide habitat for waterbirds. As natural wetland habitats continue to be altered by human development of the landscape, future populat ions of waterbirds may increasingly depend on the ability of urbanized settings such as golf courses to provide them with suitable habitat. With the highest per capita density of golf courses in the United States (866 holes/100,000 people; Ft. Myers News Press, 5/8/99), southwest Florida provides an exceptional opportunity to quantif y waterbird use and describe site selection (for foraging and other activities) of wate rbird use of created water habitats on golf courses.
6 Project Need Wetland habitats are extremely variable in the quality of habitat they provide to waterbirds and are also one of the most heavily altered types of major bird habitats Â– only grasslands have suffered a grea ter loss in area or quality in North America (Weller 1999). The declining amount of this va luable habitat is of crucial importance for conservation of waterbirds and survival of these species may ultimately reside in their ability to use altered or created wetland habitats, such as those found on golf courses. Study Objectives I focused my study on waterbird site se lection in create d urban wetlands, specifically on golf course ponds. An understand ing of waterbird habitat requirements is key to interpreting waterbird use of these cr eated wetlands. Unders tanding the process of selection and, ultimately, the use of a habitat is essential for determining which factors contribute to habitat suitability for various waterbird species. Therefore, the first objective of my study was to review existing li terature on waterbird habitat selection to determine which habitat features have been shown to influence waterbird use and should be quantified during this study. For example, prey suitabilit y, abundance, and availability were not measured directly in this study, but because these variables are crucial for site selection by waterbirds, indirect measur es (e.g., water depth, productivity, size of effective foraging area, and vegetation st ructure and density) were quantified to determine their influence on waterbird fora ging site selection of golf course ponds. The second objective of the study was to quantify waterbird use of created golf course ponds and to quantify the influence of habitat features with in golf courses on use of these ponds by different waterbird species. Habitat features chos en for the study were based on existing literature of habitat selection by waterbirds (Chapter 2) particularly
7 those considered important for foraging site selection. Foraging habitat requirements for waterbirds are often more complex and, therefor e, harder to achieve in created habitats than nesting habitat requirem ents. Furthermore, most gol f course ponds provided little nesting substrate and experience low to moderate levels of human disturbance, which can be disruptive to nesting wate rbirds (Rodgers & Smith 1995). However, I considered the importance of nesting habitat and recorded nesting activities were recorded when observed. Based on these considerations, I tested the following hypothesis: Hypothesis: Species diversity and abundance of waterbirds on golf courses is influenced by the amount and quality of fo raging habitat availabl e on the golf course. The available foraging area on golf courses is influenced by total water body surface area (of entire golf course), and individual pond surface area, perimeter, trophic status, effective foraging area, and shoreline and su rrounding vegetation stru cture and density. Predictions :If the above stated hypothe sis is true, I predict: Waterbird abundance and diversity will be directly related to combined water body surface area for a golf course. Waterbird abundance and diversity will be positively correlated with pond surface area, larger ponds will have greater bird use. Golf course ponds with larger perimeters will have greater waterbird abundance and diversity. Waterbird abundance and diversity will be gr eater in golf course ponds with higher productivity (eutrophic; hi gh levels of chlorophyll a ) than in ponds with low productivity (oligotrophic; low levels of chlorophyll a ). Wading bird and shorebird diversity and abundance will be positively correlated with size of the littoral zone . Golf course ponds with la rger littoral zones will have greater abundance of these species. Waterbird abundance and diversity will be gr eater in wetlands with intermediate vegetation structure and density.
8 Waterbird abundance and diversity will be gr eater in golf course ponds adjacent to vegetation of intermediate structure and de nsity than ponds adjacent to dense stands of natural vegetation or heavily disturbe d landscapes (i.e., housing or golf course landscapes).
9 CHAPTER 2 REVIEW OF HABITAT SELECTION BY WATERBIRDS Overview of Waterbird Habitat Selection Collectively, wetlands are among the most variable habitats in terms of their number, size, and quality. This variation resu lts in differences in their attractiveness to feeding, nesting, and roosting waterbirds, whic h may be reflected both in the number of individuals and diversity of species using them. The decisi on by a bird to use a habitat is based on innate and learned be haviors, and by trial-and-erro r testing of features and resources within a habitat. It has been suggested that birds may follo w a hierarchically structured system of sequential decisions during hab itat selection (Cody 1985; Well er 1999). At the landscape or macrohabitat scale, birds may either reside in or migrate to a geographic region with climatic regimes and overall landscape or ve getation features becau se of instinctive preferences, prior experience, tr ial-and-error testing of habitat, or a combination of these factors. Within a geographic region, bi rds may select among various landscape alternatives such as ar eas with certain types, spatial re lationships, or sizes of wetlands. At a still more refined level, birds may fo cus their search on a single wetland or wetlands complex that provides suitable habitat for an a nnual life stage such as nesting, feeding, or roosting. Lastly, birds select microhabitats such as specific foraging or nesting sites within the chosen wetland comp lex that will meet their immediate needs (Weller 1999). The relative importance of features at microha bitat (i.e., local) versus macrohabitat (i.e., landscape) scales may vary according to the proportion of suitable habitat in the
10 landscape (Donovan et al. 1997; Flather & Saue r 1996; Lichstein et al. 2002; McGarigal & McComb 1995; Schmiegelow et al. 1997). Th erefore, it is impor tant to consider multiple scales when quantifying use of wetlands habitat by waterbirds. The principal features of habitats sele cted by foraging waterbirds are often the presence of surface water and supply of suitabl e prey organisms. Heavily-used wetlands are generally larger in size, interconnected with other wetlands, ha ve fluctuating water levels, include areas of prey refugia, a nd provide suitable area s for foraging and/or nesting (Kushlan 2000a). The perceived preda tion-risks and the pres ence of other species may also either attract or deter birds from using a wetland. Foraging Site Selection Selection of foraging sites by a given speci es is influenced by a number of factors including species morphology (e .g., straight, elongated bills vs. slender, curved bills, Kushlan 1978), foraging behavior (e .g., visual vs. tactile forage rs), and prey availability (i.e., prey density and vulnerab ility to capture, Gawlik 2002). The quality and quantity of foraging habitat is extremely important a nd has been shown to influence reproductive success (Bancroft 1989; Frederick & Spal ding 1994; Powell 1983), population size (Custer & Osborn 1977; Gibbs et al. 1987) , and location of nesting wading bird colonies (Fasola & Barbieri 1978). Understanding which features of wetlands are selected by foraging wading birds and other waterbirds will aid in water re source management decisions aimed at conservation of thes e species (Gawlik 2002; Jelks 1991). Effect of Landscape Features on Foraging Site Selection Spatial and temporal fluctuations in prey availability give rise to the geographic component of habitat selecti on in waterbirds because prey must be available within certain distances from breeding colonies or roosting sites and must vary temporally in
11 ways that the birds can adjust to them (Kushlan 2000b). Wetlands are among the most variable habitats in terms of productivity because of their dynamic water regimes. Therefore, mobility is essential to waterbirds that must often track and exploit temporally and spatially changing resources in multiple wetlands or wetlands complexes to complete annual life cycle stages (Weller 1999). The spatial arrangement of available hab itat such as the distance from one food patch to another and from nesting and roos ting sites is an important component of foraging site selection fo r waterbirds. For example, Western Sandpipers ( Calidris mauri ) move an average of 2.2 km between feedi ng and roosting areas (Warnock & Takekawa 1995) and wading birds can move up to 30 km before nesting success is negatively affected (Bancroft et al. 1994; Freder ick & Collopy 1988). Distance from natural wetlands also influences both benthos (Sacco et al. 1994) and avian recruitment in created wetlands (Brusati et al. 2001). Therefor e, the location of creat ed wetlands relative to other used wetlands may, in part, de termine their ecological functionality. It is unlikely that a single wetland will provide suitable habitat for all species of waterbirds (Frederickson & Taylor 1982; We ller 1999), or for all the annual life cycle needs of a single species (i.e., nesting, migra tion, and molting or wintering habitat needs, Weller 1999). This may be especially true fo r created wetlands, which tend to be smaller in size, more uniform in habitat component s, and hydrological regimes than are natural wetlands. Artificial wetlands located within a matrix of natural or other created wetlands will increase the diversity of available habitats on a local scale and likely increase the attractiveness of created wetlands to wate rbirds (Brusati et al. 2001; Frederick & McGehee 1994).
12 Effect of Surface Area on Foraging Site Selection Wetland size is an important factor in determining fora ging site selection for many species of waterbirds. Larger areas often provide more fora ging habitat for an individual species of waterbirds and/or may also provide a wider variety of habitats for a greater number of species. Increased species richne ss with size of wetland ha s been reported for birds in many types of wetlands and lakes (Celada & Bogliani 1993; Gibbs et al. 1991; Hoyer & Canfield 1990; Nudds 1992). Proposed explanations for the positive speciesarea relationship include: 1) a sampling effect whereby increased area increases sample size, thereby resulting in dete ction of more species; 2) increased area may correlate with increased habitat heterogene ity (Williamson 1988); and 3) island biogeography theory, which proposes that dispersing animals have a better chance of fi nding larger habitats than small ones (MacArthur & Wilson 1967; Skagen & Knopf 1994, but see Williamson 1988). In temporally and spatially dynamic habitats such as wetlands, th e ability to exploit resources opportunistically may be benefici al to waterbirds (C olwell & Oring 1988). Skagen and Knopf (1994) reported that shor ebirds on the Great Plains were able to colonize habitat patches opport unistically as exposed and shallow wetland foraging sites became available. A strong correlation was al so found between the number of shorebirds and the area of available wet/mud shallow wate r habitats, which may be related to the birdsÂ’ ability to locate larg er patches within a wetland comp lex (Skagen & Knopf 1994). In a study of waterbird use of Florida Lakes, species richness increased with surface area of the lake, however , the relationship was thought to be due to an increased sampling area or near-shore habitat heterogene ity (Hoyer & Canfield 1990). Gibbs et al. (1991) found that piscivorous birds were par ticularly sensitive to wetland surface water
13 area and suggested that larger wetlands may pr ovide more open water for diving species. Fish abundance and diversity on small lakes (< 50 ha) may also be positively correlated with wetland area (Haines et al. 1986). Larger wetlands may not always have hi gher numbers of birds. For example, Brown and Dinsmore (1986) found that species richness was influenced by area in small wetlands (<5.5 ha) but not in larger wetla nds. Similarly, Jelks (1991) found higher densities of foraging wading birds in wetlands smaller than 5 ha and suggested that small wetlands may provide more foraging opportuniti es for these birds th an a few large ones covering the same area. The best management practice with respect to size of created wetlands may be a wetland complex, which permits more management opportunities (e.g., variety in vegetation stru cture and water fluctuations) for providing suitable habitat for multiple species of waterbirds. Food Availability The availability of food is probably the si ngle most crucial feat ure for determining waterbird habitat suitability. The availability of prey can directly influence the reproductive success of wading birds (Hafne r et al. 1986; Owen 1960; Powell 1983) and is likely the most crucial f eature of a heronÂ’s habitat (Kushlan 2000b). Availability includes prey density and vulnerability to cap ture (determined by characteristics of the prey, environment, and predator (Gawlik 2002; Kushlan 2000b). Availability of prey varies over time, changing seasonally, with wa ter depth, availability of foraging perches, and water conditions such as water temper ature, dissolved oxygen, turbidity. Prey availability also varies spatially, with daily or seasonal fluctuati ons in water conditions that increase or decrease av ailable habitat forcing waterbirds to move accordingly.
14 Effects of Water Depth on Foraging Site Selection Water depth and hydroperiod (the pattern of water depth over time, Mitsch & Gosselink 1993) influences almost every aspect of a wetland from the type and density of vegetation to density and availability of prey for foraging waterbirds. Waterbirds are in continuous search of sites w ith high availabilities of f ood (Weller 1999). Ideal water depths for foraging are based loosely on the type of prey and the morphology and foraging behaviors of the waterbird species. For example, diving birds such as Piedbilled Grebes ( Podilymbus podiceps ) and Double Crested Cormorants ( Phalacrocorax auritus ) that feed on fish or large invertebrate s often favor deeper and relatively stable water levels, however, availability of prey is more important than specific water depth so long as long as prey are accessi ble and safety from predators is adequate (Weller 1999). Other species, such as wading birds and s horebirds, have a smaller range of suitable water depths as most of these species are lim ited to depths no greater than their leg length (Baker 1979; Powell 1987). For dabbling ducks, depth at foraging sites positively corresponds with speciesÂ’ neck length (Poysa 1983) and these species often prefer depths of around 5-25 cm and shallow pools (Frederickson & Taylor 1982). The correct terminology for natural and created wetland or Â“waterÂ” areas is arguable, but generally follows accepted c onventions based on differences in size, permanence, and ratio of open water to vegetatio n. Lakes are generally used to describe permanent wetlands that possess a high percen tage of open water, little emergent vegetation, and well-defined shorelines. Th e term pond is used for smaller permanent and semipermanent water bodies with abrupt shorelines, deep basins, and open water (Weller 1999). Although water levels ma y fluctuate annually, lakes and ponds are generally more permanent bodies than ot her types of wetlands (Edelson & Collopy
15 1990). Water levels in other types of wetlands such as sw amps (i.e., forested.wetlands) or marshes (i.e., herbaceous wetlands) ofte n vary seasonally or annually. Decreasing water levels increases seedling germination a nd growth for a variety of plants and may also temporarily concentrate prey in these we tlands. However, the prey base is often rapidly depleted in shallow wetlands by la rge aggregations of birds (Gawlik 2002). Despite their permanence, much of the ope n water habitat of lakes and ponds is unavailable to wading or probing species, whic h must rely on perches (e.g., dead limbs, stumps, floating or matted vegetation) to gain access to prey in deeper water (Edelson & Collopy 1990; Smith et al. 1995). Another fo raging tactic employed in lakes and ponds is aerial foraging, which is ener getically expensive for wading birds, but may be used by some species during times of prey concentr ation at the water surface due to low water levels, dissolved oxygen, or water transpar ency (Edelson & Collopy 1990; Kushlan 1978; Meyerriecks 1962). However, the open water in lakes and ponds is quite suitable as foraging sites for other waterbirds that fo rage by diving for prey (e.g., Double-crested Cormorants [ Phalacrocorax auritus ] and Pied-billed Grebes [ Podilymbus podiceps ]) or dabble for submerged vegetation (e.g., Mottled Ducks [ Anas fulvigula ], Wood Duck [ Aix sponsa ], etc.) as long as there are sufficient f ood resources to support these birds. Open water habitats, which are often characterized by a deep central pool, may also benefit mobile aquatic prey of waterbirds such as fish, amphibians, and macroinvertebrates because they may be able to survive extr eme droughts by seeking refuge along the edges of the deep water areas (Smith 1994). Effective Foraging Area: Effe cts of Littoral Zone Area on Foraging Site Selection Waterbirds that forage primarily in shallo w water areas of dept hs not greater than their leg length, such as wading birds and s horebirds, are often confined to the littoral
16 zone (i.e., the area extending from the shor e to the maximum exte nt of nonpersistent emergent plants) within wetlands systems (Mitsch & Gosselink 1993). One method of measuring the amount of habitat available to birds that forage by wading in shallow water or probing in soft substrates is to measur e the effective foraging area or the area around the perimeter of the water body that is no greater than the deepest utilized foraging depth. A shoreline index or the ratio of shorelin e length to water area is also a useful measurement for determining the amount of marsh edge available to wading birds and shorebirds (Weller 1999, also see shoreline development index in Nilsson & Nilsson 1978). Effects of Water Productivity on Foraging Site Selection The trophic status of lakes and wetlands is generally determined by four water chemistry parameters: chlorophyll a , phosphorus, nitrogen, and water clarity. Chlorophyll a , which is the dominant green pigment that enables algae to make food during photosynthesis (Raven et al. 1992), is of ten used as an estimator of the organic base (i.e., algae) upon which aquatic bird populations depend (Hoye r & Canfield 1990; but see Hoyer & Canfield 1994; Nilsson & N ilsson 1978). More productive systems (as determined from nutrient concentrations, chlorophyll a , and/or other water chemistry parameters) may have a larger food base and may support more specialized species of wildlife, resulting in greater species richness (Hutchinson 1959; MacArthur 1970; Wright 1983). Excessive concentrations of nutrient s in hypereutrophic systems may result in algal blooms leading to reduced water transp arency and dissolved oxygen concentrations. These conditions may force fish closer to the surface to locate prey and/or obtain necessary levels of dissolved oxygen. The c oncentration of prey at the surface may in turn make them more vulnerable to fora ging waterbirds (Edelson & Collopy 1990).
17 However, oxygen deprivation from increased nutr ient loads may also result in fish kills, whereby algae begins to domina te the food web, and in turn decreases prey availability for many species of waterbirds. Prey av ailability may also be decreased in hypereutrophic systems due to increased turb idity, which may interfere with the visual foraging techniques of some waterbird species. In two studies of Florida lakes, abundance and species diversity of waterbirds were positively correlated with eutrophic c onditions (Hoyer & Canfield 1990, 1994). Hypereutrophic lakes may also have greater abundances of foraging wading birds than less productive lakes (Edelson & Collopy 1990; but see Nilsson & Nilsson 1978). However, increased nutrients in lakes and ma rshes may also be of environmental concern because of related potential health risks to foraging birds (Frederick & McGehee 1994). The nematode parasite Eustrongylides ignotus has been found only in areas of elevated nutrient inputs and/or disturbed soil condi tions (Spalding & Forrester 1993). This nematode can negatively impact the health of adult wading birds and survival of their nestlings (Spalding et al. 1993). Therefor e, while human induced eutrophication may potentially increase bird use of created wetla nds, potential health hazards, such as those posed by these parasitic nematodes, should be further studied (F rederick & McGehee 1994; Smith 1995). Effects of Shoreline Vegetation on Foraging Site Selection The type and structure of vegetation crea tes microhabitats from which species can choose for various activities (Weller 1999). Fo r permanent bodies of water such as lakes or ponds, much of the vegetation is confined to near shore areas where appropriate water depths and the greatest amount of light penetration can be found. Because the littoral zone is the primary area of foraging for many species of waterbirds, the structure, type,
18 and density of shoreline vegetation may str ongly influence the suitability of potential foraging habitats. Vegetation structure influences diversity and abundance of prey such as small forage fish and invertebrates, which require sufficient vegetative cover to hide from predators (Chick & Mclvor 1994). While prey densities may increas e with plant density (Stoner 1983), the availability of prey to foraging waterbirds may decrease. Dense vegetation may interfere with movement and foraging proficiency of waterbirds, thereby reducing the availability of pr ey. Tall and/or dense emerge nt vegetation may also reduce a birdÂ’s ability to remain vigilant agains t predators by hindering movement and visual detection abilities (Henderson 1981; Smith 1994; Surdick 1998) . Therefore, vegetation of moderate to low stature and structural diversity is believed to provide the best foraging opportunities for many species of wadi ng birds such as Great Egrets ( Ardea albus ). However, suitable habitat varies with the species of waterbird. For example, Green Herons ( Butorides virescens ) prefer more densely vege tated habitats for foraging , Belted Kingfishers ( Ceryle alcyon ) prefer sites with available perches, and Spotted Sandpipers ( Actitis macularia ) prefer bare mudflats (Ehrlich et al. 1988; Sibley 2000). Submersed plants are also an important co mponent of littoral zone vegetation. The type and structure of submerged plants infl uences prey density and distribution in the littoral zone (Chick & Mclvor 1994). Thes e plants can also form dense mats of vegetation that may inhibit diving birds but benefit other birds such as wading birds, which use them to access deeper water for foraging (Edelson & Collopy 1990; Weller 1999). Regardless of the specific type of hab itat preference, large numbers of waterbirds using permanent water bodies, such as crea ted ponds and lakes, depend on near-shore
19 areas with water depths suitable for foraging. Therefore, the stru cture and density of vegetation in this zone is an important influence on use of these water bodies by waterbirds (Hoyer & Canfield 1994). Effects of Surrounding Habitat on Foraging Site Selection The vegetation adjacent to a lake or pond is also a potentially important factor for determining foraging site selection. For ex ample, nearby trees or shrubs may provide perches for aerial foragers such as Belted Ki ngfishers, or roosting sites for other species such as Double-crested Co rmorants or Anhingas ( Anhinga anhinga; personal observation). The structure and density of surrounding vegetation may also influence a birdÂ’s ability to detect approaching predators, ther eby detracting birds from ponds surrounded by heavily forested vegetation. Lastly, surrounding vegetation may also indicate the amount of human disturbance bi rds will experience at various ponds. Ponds completely surrounded by human dominated feat ures such as residential lawns and turf grass will likely experience more disturbanc e from humans (e.g. golfers, home owners, people fishing, lawn maintenance crews). Ho wever, such open habitats may also be attractive to waterbirds since they may give excellent view of appr oaching mammalian or avian predators. Species Interactions When animals select habitats in which to live or reproduce their preferences can be affected by social as well as habitat co mponents (Muller et al. 1997). Intraand interspecific interactions may affect fora ging efficiency due to aggression, social interactions, and reduced prey availability. The same wetland or wetland complex ma y possess characteristics that are attractive to many species of waterbirds. Differences in size, behavior and morphology
20 among different species enable resources within wetlands to be partitioned among waterbirds, resulting in reduced competition for food and/or increased feeding efficiency (Jenni 1969; Weller 1999). Within a comm unity of waterbirds, behavioral and morphological adaptations result in differences in time of f eeding, type and size of prey selected, feeding behavior, or feeding site (e.g., specific water depth or substrate used), all of which contribute to resource parti tioning and reduced interspecific competition (Kushlan 1978; Weller 1999). Although morphological differe nces theoretically limit overlap in resource use among species, bird s with similar morphology often exhibit competitively induced ecological segregati on by adapting their behavior to prey availability and the competi tive situation of th e habitats they ar e using (Fasola 1986). Spacing among individuals is an important co mponent of habitat selection because, as the total number of birds increases in preferred foraging areas, individuals may be forced to feed in less preferred sites (Goss-Custard 1977). Spacing among individuals occurs during foraging in all wading birds a nd, minimally, is the length of the head and neck but may enlarge or contract depending on site-specific conditi ons (Kushlan 1978). Spacing is determined by defensibility, which is influenced by res ource dispersion, prey availability, number of competitors and habitat availability (Kushlan 1978). High densities of birds on a fora ging site may reduce feedi ng rates through aggressive encounters over food items, which detract from the amount of time otherwise spent foraging (Goss-Custard 1977; Kushlan 1978). Birds may also take part in social interactions that distract them fr om seeking prey (Goss-Custard 1976). The presence of other birds may also increa se the number of bird s using a particular habitat. When foraging birds achieve suffici ent net energy gain or frequency of capture
21 they will likely remain in the same we tland or wetland complex. The presence of competitors in a wetland may indicate the pr esence of high quality or abundant prey resources to birds choosing from available wetla nd habitats. In this manner, aggregations of birds may form, and indeed commonly occur. For example, Great Egrets, Little Blue Herons ( Egretta caerulea ), and Tricolored Herons ( Egretta tricolor ) were drawn to foraging sites by the presence of Snowy Egrets ( Egretta thula) in Panama (Caldwell1981) possibly because these species select the highest quality food patches (Gawlik 2002). Foraging success as well as protection fr om predators may also be increased through commensal (i.e., mixed species) fora ging which takes place in many species of waterbirds and shorebirds (Ehrlich et al. 1988) . Attendant species be nefit from increased strike and feeding rates (Kushlan 1978) because the commensal species increases exposure to prey through its own foraging activities (e.g., stirri ng up prey from the substrate). Commensal associations in clude the attendance of White Ibis ( Eudocimus albus ) by Little Blue Herons, and attendance of Double-crested Cormorants by White Ibis and Great and Snowy Egrets (Ehrlich et al. 1988; Kushlan 1978). Species interactions are an important component of waterbird foraging site selection. For example, the presence of cer tain species such as Snowy Egrets in a wetland may be an indication of high pr ey resources (Caldwell 1981; Gawlik 2002), which may therefore, attract other species to a site. However, with increased densities of birds there is a greater like lihood of rapid depletion of pr ey densities. Therefore, although spacing among individuals and the form ation of aggregations are an important influence on waterbird foraging site selection they are driven by the complexities of prey availability. The complexity of this rela tionship made assessment of the influence of
22 these variables outside the scope of this study, but their importance to site selection was recognized, particularly through commensal associations, which did occur during the study and are discussed in a later chapter (Chapter 5).
23 CHAPTER 3 METHODS The objective of this project was to identify habitat variables that significantly influence the use of golf course ponds by waterb irds in Southwest Florida. At the scale of the individual ponds, surface area, effective foraging area, trophic state, water depth, as well as shoreline and surrounding vegetation structure and density were quantified at each pond within individual golf courses to de termine whether these factors influenced use of individual ponds by waterbirds. At the scale of the golf courses, total water surface area of ponds was combined to determ ine its influence on waterbird use of the individual golf courses. Distance to the nearest active colony was also considered because of its potential influence on wading bird use of the golf courses, but was not analyzed for reasons discussed below. Study Species All species studied in this project were waterbirds and are defined, for the purposes of this study, as any water-dependent spec ies of bird (Weller 1999). These birds are highly mobile, and shift locally or regionally as new habita ts become available. I surveyed species from the orders: Ciconiiformes, Gruiformes, Pelecaniformes, Anseriformes, Podicipediformes, Coraciiformes, and Charadriiformes (Table 3-1). Birds from these orders represent a variety of bird sizes, morphologies, foraging techniques, and substrates used for foraging (e.g., ba re mudflat vs. open water). Because the degradation of wetland habitat ha s affected nearly all species of wetland dependent birds, it is important to consider more than one sp ecies when determining the functionality of
24 created ponds on golf courses for waterbirds. Therefore, all species of waterbirds observed in the ponds or within 5 m of pond edges were included during surveys. Study Area: Golf Course Selection A total of 12 golf courses were surveyed during this study, nine of which were owned by Bonita Bay Group and three by Wate rmark Communities Incorporated. All golf courses were located in Lee or Collier County in southwest Florida (Table 3-2). A total of 183 golf course ponds from these 12 courses were monitored during the study. Study ponds included only those water bodies that occurred primarily within golf course property boundaries. Ponds that primarily o ccurred within reside ntial communities were not investigated. Golf courses were selected to provide a diversity of study sites determined by a preliminary evaluation of habitat within and around the created ponds, age of the golf course, and geographic location within th e two counties. Geographic location is important because courses loca ted near large wading bird r oosting and nesting sites may experience increased use of ponds by waterbirds . Location of golf courses relative to one another was also an important criterion for selection to allow surveys of multiple golf courses on the same day, by the same obser ver, and within the primary waterbird foraging hours (i.e., sunrise until noon). Age of the golf course (Table 3-2) may also influence features within the course such as trophic status of the water bodies (see Productivity Measurements section) or the density of vegetation within and around the ponds, both of which influence bird use. Although relative lo cation and age were considered important for initial golf course se lection, they were not included in the final analysis. Instead, variables th at directly influence use by waterbirds, such as trophic
25 status of ponds and shoreline vegetation, were measured for each course and used in analyses. Waterbird Surveys A total of sixteen surveys were conducted at each pond (12 sunrise and 4 sunset) during January through April in 2001 and 2002 (e xcept for one course that was surveyed 15 times due to a scheduling conflict). Surv eys were conducted between sunrise and noon (0600-1200 EST) and as close to sunset as go lf course closing schedules would allow (1600-1930 EST). These periods were chosen because most active foraging by wading birds occurs near dawn and dusk (Kushlan 1978). Golf courses were grouped by location to allow more than one course to be surv eyed each day. Each group of courses was randomly surveyed throughout the survey pe riod, but the order of groups surveyed was alternated to reduce time-of-day bias. Surveys were conducted from a golf cart and birds were visually identified from a distance (to avoid disturban ce) using 8 x 42 binoculars. Ponds were surveyed sequentially in the order in which they app eared on the golf course because high humanuse precluded randomization of surveys. The or der in which ponds were surveyed varied depending upon time of day, with ponds surveyed from holes 1 to 18 if the survey started before course play (i.e., at sunrise), or from hole 18 to 1 to allow golfers maximum visibility of the surveyor if surveys were conducted after pl ay had begun. At each golf course pond, all waterbirds in the ponds or with in 5 m of the waterÂ’s edge were recorded. Habitat type (shore, water, human structures, etc.; Table 3-3) and first observed activity (Table 3-4) were also recorded for each bird. When a bird was stationary it was observed for 30 s because birds sometimes ceased thei r activities as I a pproached the ponds. Observations of more than 30 s were usually not needed for birds to return to their
26 activity nor were l onger observations feasible due to the high human-use factors associated with the courses. Birds that flew over the ponds but were not obviously foraging or did not stop at the pond were not included in the analyses. Foraging Guild Classification Waterbirds recorded during this study were categorized into fo raging guilds, which represent multiple species that use similar resources in similar ways. Foraging guilds are often defined by primary food preference, fora ging substrate, and foraging techniques and provide a useful approach for evalua ting the influence of habitat changes on community dynamics (De Graaf 1985). Foraging guilds were used to evaluate how changes in habitat variables and water characteristics associated with golf cour se ponds affected use of ponds by different groups of waterbirds. This allows devel opment of management recommendations to benefit groups of birds rath er than individual species. Foraging guilds were defined based on major foraging techniques, food types, and substrates listed for each species in Ehrlich et al. (1988) and De graaf et al. (1985) as well as personal observations of foraging birds on golf cour se ponds (Table 3-5). Surface Area The surface area for each golf cour se pond was provided by golf course superintendents or was obtained from the e ngineering design blueprints for each course. Three of the ponds also contained vegetate d islands in their center that were approximately circular in shape. The island perimeters were calculated with a measuring wheel and used as the circumference of the is land, which was then used to calculate the area of the island. The surface area of the is lands was then subtracted from the surface area of the ponds before being entered into the analysis. Surface area was also analyzed
27 separately to quantif y its relationship with bird abundance and species richness, which have been shown to be significantly positively correlated with the area of lakes (Hoyer & Canfield 1990, 1994). Perimeter The perimeter of each pond was determined in the field with a 0.91-m (3-ft) diameter, measuring wheel guided along the waterline of each pond. The extent of perimeter may be an important factor for bird s that feed along the shore (e.g., shorebirds: Charadriiformes) or near th e waterÂ’s edge (e.g., wading bird s: Ciconiiformes). This variable was also used to calculate the percen t vegetation coverage at each pond (see below). Effective Foraging Area The effective foraging area, as defined by th e area of the littoral or shallow water zone, was quantified to determine its influen ce on site selection of golf course ponds by waterbirds, particularly wading birds. Measurements for this habitat variable were taken along the same transects as those used for s horeline vegetation estimates (see below). The number of measurements per pond for this variable varied according to the size (perimeter) of the ponds. Effective foraging area was determined by measuring the area around each pond where the water level was 40 cm, the deepest water level in which wading birds effectively forage (Powell 1987). The distance from the waterline to a wa ter depth of 40 cm was measured with a pole calibrated in 10-cm intervals. For purposes of calculating the area of the littoral zone available as foraging habitat for wading birds, the cross section of the littoral zone was viewed as a right triangle with the base (40 cm) and the height measurements (lateral surface distance from the waterline to 40 cm depth) taken at each transect used to
28 calculate available foraging area (i.e., 1/2bh). These calculations were averaged to estimate the mean effective foraging area (m2) for each pond. This variable was used during analyses to evaluate the influence of available foraging ar ea on waterbird use of the golf course ponds. Shoreline and Surrounding Vegetation Shoreline vegetation is important for many foraging and nesting birds, as well as for birds seeking shelter and protective c over (Weller 1999). Duri ng this study, shoreline vegetation was delineated by the pondÂ’s waterl ine because the centers of the pond were generally too deep to allow growth of vege tation other than purely aquatic plants. Successive 5-m transects placed at the waterÂ’ s edge were used to quantify shoreline vegetation at 30-m intervals around the perime ter of the pond (Figure 3-1). The number of transects per pond varied w ith the size (perimet er) of the ponds. The transect length (i.e., 5 m) was chosen to allow roughly si multaneous measurement of amply sized plots (5 x 10 m) of vegetation adjacent to the ponds (see below). Percent cover of vegetation was visually estimated within 1-m2 quadrats perpendicular to th e waterline (i.e., extended 1 m outward) on both the water and terrestrial side of each transect at 1-, 3-, and 5-m intervals along the 5-m tran sect (Figure 3-2). This method was chosen because estimation of vegetation coverage was considered to be more accurate in 1-m2 quadrats. These three measurements were then averag ed to produce one final vegetation estimate for each transect. The estimations for each tr ansect were then averaged to produce an estimation of the total vegetative cover for each terrestrial and aquatic vegetation category at each pond. Six percent cover classes (Table 3-6; Daubenmire 1959) were used during visual estimates of vegetation that had been classifi ed into 7 terrestrial and 4 aquatic categories
29 (Table 3-7). Cover classes were converted to median percent cover values (Daubenmire 1959) and were used in conjunc tion with perimeter calculatio ns to determine the percent coverage of each of the terrestrial and aqua tic vegetation categories at each pond. These data were used to determine the influence of different types and de nsities of vegetation on site selection by waterbirds. To determine the influence of natural ha bitat and human-use features adjacent to the ponds (i.e., percentage of the pond surrounde d by turf, housing, or natural vegetation) on waterbird site selection, the 5-m transect used to quantify shoreline vegetation was moved 1 m away from the waterline (so as not to overlap with previous habitat measurements) and extended landward a distan ce of 10 m to form a 5x 10-m plot. Measurements of habitat features adjacent to the ponds were taken simultaneously with the shoreline vegetation measurements (i.e., at the same 30-m intervals). Six cover classes (Daubenmire 1959) were used during visu al estimation of the pe rcent cover of the 6 surrounding landscape categories, which also included non-vegeta tion categories to more fully characterize the areas adjacent to ponds (Table 3-7). Visual Estimation Technique and Observer Bias A 1-m2 open-ended quadrat calibrated in 25-cm intervals was used to train the surveyor to estimate percent vegetation covera ge. Percent coverage was then estimated visually for each transect without the aid of a quadrat. The visual estimation technique was deemed necessary primarily due to the la rge number of 5-m transects (total numbers ranged from 5-98 per pond) needed to quan tify the vegetation for each pond, as well as the large number of study ponds (183) to be surveyed during the study period. The surveyor also used a pacing off technique to pace off the 5-m transect used for vegetation
30 estimates. This technique was initially refi ned by comparing steps with actual distance measurements before using this method in the field. These techniques were selected because minimizing interference with golf play necessitated quick and efficient estimation techniques. However, when per cent vegetation coverage was questionable, the 1-m pole calibrated in 10-cm increments used for littoral zone estimations (see Effective Foraging Area) was used to make final decisions about cover classes. Measurements were also taken to determine whether observer estimations were within the range for each cover class (Table 3-6). These measurements were performed by pacing off a 5-m transect and visu ally estimating vegetation in the 1st, 3rd, and 5th meters of the transect. The vegetati on was then re-evaluated with the 1-m2 open-ended quadrat (described above) which was used to both more accurately determine the distance between and the vegetation within the 1st, 3rd, and 5th meters of the transect. A total of 15 estimates were taken for each combination of the vegetation categories (6 terrestrial and 4 aquatic) in each cover class. These estimates were then analyzed to evaluate observer accuracy. Using these methods, observer bias (percentage of vegetati on classified in incorrect coverage categories) associated with visual estimation of shoreline vegetative cover was estimated to range from 6.7 40.0% ( x = 13.4%) for aquatic and 6.7%-33.3% ( x = 13.8%) for terrestrial vegetation categories, respectively (T ables 3-8, 3-9). However, because subjective quadrats were measured by pacing off distances whereas the distance between objective quadrats was measured more accurately with an open-ended quadrat there was a lack of overlap between some subjective and objective quadrats (personal observation). These differences may have re sulted in misclassifi cation of vegetation
31 cover classes for some quadrats becaus e the two measurement techniques did not measure the same respective areas. Therefor e, many of the misclassified values were likely due to the sampling methods and not the ability of the observer to estimate vegetative coverage. Productivity Measurements Trophic status is an indicator of biological productivity of a b ody of water and the ability to support plants, fish, and wildlif e (LAKEWATCH 2001). Four water chemistry parameters were measured to determine the trophic status of the golf course ponds: total chlorophyll a , total phosphorus, total nitrogen, and wa ter clarity. Due to time and cost constraints, total chlorophyll a , total phosphorus, and total ni trogen were measured at randomly selected ponds at each golf course . Four ponds were sampled at golf courses with <20 surveyed ponds, and five ponds were sampled at courses with >20 surveyed ponds. Water samples were gathered according to protocol developed by the Lakewatch Team at the University of Florida (LAKE WATCH 2001). Samples were gathered from each pond during March and May 2001, fro zen, and sent to the LAKEWATCH laboratory at the University of Florida for analysis. Water clarity was measured at all golf course ponds during May 2001 with a 20-cm (8-inch) diameter Secchi disk. A correla tion between Secchi disk measurements and chlorophyll a measurements was established to estimate biological productivity from Secchi disk measurements alone, which provided at least one measurement of productivity in all of the golf course ponds su rveyed. When the presence of algae is the main factor that reduces wa ter clarity (as was the case fo r golf course ponds in this study), a correlation between Secchi disk and chlorophyll a measurements provides a useful method for estimating biological productivity (LAKEWATCH 2001). PROC
32 NLIN (SAS Institute 2001) was used to determine model parameters with natural logarithm transformed Secchi disk and chlorophyll a measurements. The model was then used to predict productivity levels in all other ponds from their Secchi disk measurements. Two Secchi disk measurements were also taken in each of the ponds during March and May 2001. The correlation model esta blished in 2001 from water clarity and chlorophyll a estimates was then used to determ ine changes in pond productivity between the two sampling periods. A single sampli ng period was used during the second study year (2002) because significant differences (T-test; t = -0.33; p = 0.74) did not exist between productivity measurements taken dur ing the two sampling periods (March and May) in 2001. To compare chlorophyll a levels between study y ears a paired t-test was performed with the MEANS procedure in SAS (SAS Institute 2001) on natural log transformed data. Water Depth Average water depth was calculated from three, non-randomly chosen points within the pond. In general, one measurement was taken at each end and one measurement was taken in the center of the pond from a boat. Measurements were taken in March 2002, during FloridaÂ’s dry season, when some of th e ponds may dry to depths <1 m. Although these measurements were not an indication of long-term water levels or seasonal fluctuation, they do provide a comparison of average depths between ponds during the study period. Deeper waters are not suitable fo r bird species that primarily forage in the littoral zone or along the shore but may be important for the growth of submergent waterfowl food (Brein inger & Smith 1990) as well as providing prey refugia for the production of fish.
33 Distance to Nearest Colony Because golf course ponds were primarily used as foraging sites by waterbirds, the distance to nesting and roosti ng sites may be an important influence on bird use of a particular golf course, particularly for col onial nesting wading bird s. However, wading bird colony locations often demonstrate interyear variation due to changing water levels, food supply, and other environmental parame ters (Frederick & Collopy 1988; Powell 1983; Rodgers et al. 1987). This suggests that to determine the location of a golf course site to the nearest active co lony I would have needed data on colony locations for 2001 and 2002, which were unavailable at the time of this study. Therefore, without same year data of colony locations, an analysis to dete rmine the distance from the golf courses to the nearest colony could not be performed. However, based on historic colony locations (e.g., Florida Fish and Wildlife Conserva tion Commission 1999 surveys), there was probably no dearth of wading bi rds in the southwest Florid a region. Nevertheless, the question of whether the distance to the nearest active colony influences bird use of golf courses is an important one and shoul d be considered for future studies. Although the results could not be used during this study, the methodology used to calculate the distances from each of the golf c ourses to the nearest colonies are described because of its potential usefulness to future st udies. This example used the most current colony locations, which were obtained from the Florida Fish and Wildlife CommissionÂ’s (FWC) 1999 wading bird survey in the fo rm of an ARC/VIEW (ESRI 1999b) point coverage shapefile. Only t hose colonies classified as active during recent (1990Â’s) surveys were considered for this example. Golf course GIS (Geographic Information System) location data was obtained from the County Property Appraiser offices in Lee and Collier counties, within which all golf
34 courses were located. Information from these offices is updated regularly for tax appraisal purposes and was more current th an available aerial photography (e.g., aerial photography for Lee County was taken in February and March 1998 and January 2000 for Collier County, which also would have required digitizing). Files were in the form of ARC/INFO (ESRI 1999a) workspace format or ARC/VIEW (ESRI 1999b) shapefiles as parcel polygon and line coverages current as of April 2002. All parcel data files were reprojected to Albers Equal-Area Conic proj ection to correspond with the FWC rookery coverage. The parcel data files were used to loca te each golf course and a rectangle was drawn around each course in ARC/VIEW (ESR I 1999b) that attempted to include all ponds within the course. Because the parcel data files did not include all of the golf course study ponds, location of ponds was of ten approximated based on other features such as buildings, roads, or section boundari es contained in separate shapefiles. The distance to the nearest active rookery was then determined for each golf course with the SPATIAL ANALYST extension in ARC/VI EW (ESRI 1999b). The Â“Find DistanceÂ” command was set to a cell size of 1 m. Each corner of the golf course rectangle was viewed at a 1:0 map scale and the ARC/VIEW (ESRI 1999b) identifying tool was used to calculate the distance to the n earest active rookery at an accu racy of 1 m. The distances from each of the four corners of the rectangle were averaged to determine the mean distance from the golf course to the nearest ac tive rookery. Analysis of Pond Similarity: Effective Foraging Area and Shoreline Vegetation Due to similarities inherent in pond c onstruction caused by permitting requirements (that often define the pond dimensions a nd amount of vegetation planted around the shorelines), as well as course maintenan ce policies (which often involve removal of
35 vegetation that seems unkempt or interferes w ith the ability of golfers and residents to view the ponds), there was concern about low variability of habitat variables among ponds. To determine whether the mean area of effective foraging habitat (littoral zone) varied significantly among ponds, the NPAR 1WAY procedure in SAS (2001) was used to conduct a Kruskal Wallis test. To asse ss the similarity of vegetative cover among ponds, a cluster analysis was performed with PROC CLUSTER (SAS Institute 2001). This procedure was used to group ponds that possessed similar coverage of terrestrial and aquatic vegetation (Table 3-7) for both 2001 and 2002 surveys. Analysis of Site Preference Bird abundance data was extremely skewed due to the high number of zero counts during waterbird surveys, making it difficult to successfully transform the data into an approximate normally shaped distribution. Exploratory analyses such as Principal Components Analysis also indicated that the habitat data for the golf course ponds could not easily be defined in one dimension (i.e ., they were not corr elated enough to group them into information categories). A generalized linear model with logistic regr ession was, therefore, determined to be the most appropriate analysis and was used to model the probability of a foraging guild (Table 3-5) being present at a pond given a particular set of pond characteristics (vegetation characteristics, surface area, etc.). Foraging guild presence was used as the response variable because there were genera lly not enough observations for each of the 42 species to perform separate analyses for each species. PROC GENMOD (SAS Institute 2001) was us ed to fit a generalized linear model to the data. This method was preferred over other procedures because the data represented a non-linear combination of both categorical (i.e., golf course and year) and
36 continuous (vegetative cover, surface area, etc.) predictor variab les appropriate for analysis by stepwise logis tic regression. SAS (2001) so ftware does not currently implement a stepwise selection for logist ic regressions. However, PROC GENMOD allowed for a manual selection of predictor vari ables comparable to a stepwise or forward selection of predictor variables. A logistic regression was used to model the likelihood (log odds ratio) of observing a waterbird (from a particular foraging gu ild) as a linear function of the habitat characteristics of a pond (shoreline and su rrounding vegetation, surface area, perimeter, etc.). The presence or absence of a guild member was used as the binomial (logit) response variable. The response variable was also redefined as the presence of waterbirds 5 birds from a foraging guild and was used in a second set of models to determine the influence of th e predictor variables on Â‘groupsÂ’ waterbirds. The predictor variables included: survey year, pond surf ace area, pond perimeter, effective foraging area in the littoral zone, per cent cover of aquatic and terr estrial shoreline vegetation (Table 3-7), surrounding habitat (Table 37), and pond trophic stat us (i.e., estimated amount of chlorophyll a ). Each predictor variable was run separately in PROC GENMOD with the binomially distributed foraging guild respons e variable (i.e., presence/absence). Predictor variables were manua lly selected based on the mode lÂ’s Goodness of Fit Values (i.e., small deviance values) and the predictor variablesÂ’ Type III Sums of Squares values ( p < 0.05). Combinations of variables we re then manually combined in PROC GENMOD to identify the model for each foraging guild consisting of a subset of predictor variables with the smallest deviance and p -values < 0.05.
37 Individual golf courses (n = 12; Table 3-2) were added as variables to final models for each foraging guild to determine if other pr edictor variables were still able to explain a significant ( p < 0.05) amount of the differences in bird abundance once variability due to the differences among golf courses were added to the model. The analysis was performed in this manner because the golf course variable was confounded with many of the other predictor variables. Interpretation of Regression Models Scenario tables were developed for each of the 6 foraging guild s to illustrate the ability of the logistic regression models to predict the presence of birds from each foraging guild on golf course ponds given diffe rent amounts of vegetative coverage and other significant predictor vari ables (see Results). The variab les included in these tables were selected by the logistic regression analys es as important variab les for explaining the presence of birds from each foraging gu ild on golf course ponds. The probability of observing a bird from each of the six foragi ng guilds is predicted from the intercept estimate and estimates for each of the signifi cant predictor variable s with the formula: P(observing bird)= e (Intercept Est + Var 1 Est (Var 1 Measurement) +Var 2 Est (Var 2 Measurement) + Â…+ Var n Est(Var n Measuremt)) 1 + e (Intercept Est + Var 1 Est (Var 1 Measurement) +Var 2 Est (Var 2 Measurement) + Â…+ Var n Est(Var n Measuremt)) Analysis of Golf Course Comparison To calculate a site-by-site comparison of golf courses for each foraging guild a Means Separation within a generalized lin ear model was performed with the PROC GENMOD procedure. The model was run 12 times for each of the 6 foraging guilds to allow each of the golf courses to be used as the reference site. The intercepts and p values (< 0.05) from each of the 12 models we re used to produce a Ranked Order of the
38 Log of the Odds Ratio of seeing a species fr om each of the foraging guilds on each golf course compared to the other 11 courses. The surface area of the study ponds was combin ed to determine th e total area (ha) of the water bodies on each golf course. A generalized linear m odel (PROC GLM, SAS Institute 2001) was then used to determin e whether total water surface area for each course was able to explain a significant am ount of variability in the average abundance and density of birds for each golf course. PR OC GLM was also used to conduct a lack of fit analysis on the means of pairs of abundance data and to determine the most appropriate equation for the model.
39 Table 3-1. Categories of waterbirds, with representative species, documented during surveys of golf course ponds. Waterbird category Order Examples Wading birds Ciconiiformes Herons , egrets, ibises, and storks Short-legged and other wading birds Gruiformes Coots, moorhens, and cranes Diving birds Pelecaniformes Cormorants, anhingas, and pelicans Waterfowl Anseriformes Ducks Grebes Podicipediformes Grebes Kingfishers Coraciiformes Kingfishers Shorebirds Charadriiformes Sandpipers, snipes, willets, yellowlegs, killdeer, gulls, and terns
40 Table 3-2. Property owners, county location, approximate latitude and longitude (in degrees and minutes) and date of comp leted construction for each of the 12 southwest Florida golf course s surveyed during 2001 and 2002. Golf course Property Florida county Approx. longitude Approx. latitude Year of completion Bay Island at Bonita Bay Bonita Bay Group Lee 81 49Â’ 26 22Â’ 1995 Burnt Store Marina and Country Club Watermark Communites Inc. Lee 82Â°02Â’ 26 45Â’ 1979 Copperleaf at the Brooks Bonita Bay Group Lee 81 47Â’ 26 24Â’ 2000 Creekside at Bonita Bay Club West Bonita Bay Group Lee 81 49Â’ 26 22Â’ 1991 Cypress at Bonita Bay Club East Bonita Bay Group Collier 81 39Â’ 26 17Â’ 1997 Gateway Golf and Country Club Watermark Communites Inc. Lee 81 47Â’ 26 36Â’ 1989 Marsh at Bonita Bay Club West Bonita Bay Group Lee 81 49Â’ 26 22Â’ 1985 Mediterra Bonita Bay Group Collier 81 46Â’ 26 20Â’ 2000 Sabal at Bonita Bay Club East Bonita Bay Group Collier 81 39Â’ 26 17Â’ 1998 Spring Run at the Brooks Bonita Bay Group 2001; Member owned 2002 Lee 81 48Â’ 26 24Â’ 1998 Twin Eagles Bonita Bay Group Collier 81 38Â’ 26 17Â’ 1998 Wildcat Run Watermark Communites Inc. Lee 81 45Â’ 26 26Â’ 1986
41 Table 3-3. Bird habitats categorie s used during 2001 and 2002 surveys. Code Habitat 1 Water 2 Shore (i.e., sloped area around the pond with no water, including sandbars where specified) 3 Rocks, stumps, logs 4 Vegetation 5 Man-made structures (rock ledges, wood duck boxes, staff gauges, etc.) 6 Rough/yard (manicured grass within 5 m of waterÂ’s edge) Table 3-4. Behaviors used to categorize bird activities duri ng 2001 and 2002 surveys. Code Behavior 1 Foraging or associated movement s (e.g., walking, running, wading water) 2 Moving but not obviously foraging (e .g., swimming, walking on shore, etc.) 3 Stationary or resting activiti es (e.g., standing, resting, roosting) 4 Drying wings 5 Flyover 6 Flushed from pond 7 Preening 8 Nesting activities (e.g., sitt ing on nest, feeding chicks)
42 Table 3-5. Foraging guilds used for analyses, which were determined by the speciesÂ’ size, major foraging habitat, technique, and food type. Foraging guild Species Aerial Foragers Terns, kingfishe rs, eagles, osprey, brown pelicans Dense Vegetation Waders Night and green herons, bitterns Diving Birds Grebes, cormorants, anhingas, mergansers, scaup, ruddy and rin g -necked ducks Dipping and Dabbling Foragers Mottled ducks, blue-wing teal, moorhens, coots Moist-soil Foragers Sandpipers, yellowlegs, stilts, willets, k illdeer, snipes, gulls Open Water Waders Herons, egrets, ibises, storks, cranes Figure 3-1. Example of 5-m transects used for vegetation sampling taken at 30-m intervals around the perimete r of a golf course pond. 5 mtransect 30-m interval 5-m transect Golfcoursepond
43 Figure 3-2. Example of 1-m2 vegetation sampling quadrat taken at 1st, 3rd, and 5th meters of each 5-m transect. Table 3-6. Cover classes, range, and median percent cover (Daubenmire 1959) used for estimation of percent coverage of sh oreline vegetation and areas surrounding golf course ponds. Cover class Range (% cover) Median (% cover) 1 0-5 2.5 2 5-25 15 3 25-50 37.5 4 50-75 62.5 5 75-95 85 6 95-100 97.5 Aquatic vegetation estimates Terrestrial vegetation estimates 1m 1m Meter3 Meter1 5 mtransect Shor eline Meter5
44 Table 3-7. Vegetation and landscape categories used to characterize aquatic and terrestrial vegetation along shorelines of golf course ponds and the areas immediately surrounding the ponds. Terrestrial categories Aquatic categories Surrounding landscape categories 1= Herbaceous (<1 m high) 1= Submerged/floating aquatics 1= Herbaceous (unmanicured vegetation) 2= Herbaceous (>1 m high) 2= Herbaceous (<1 m high) 2= Shrub/tree 3= Shrub 3= Herbaceous (>1 m high) 3= Housing/residential lawn 4= Tree 4= Man-made structures (e.g., walls, ledges, drains, etc.) 4= Golf course/turf grass 5= Mixed shrub/tree 5= Construction activity 6= Man-made structures (e.g., walls, ledges, drains, etc.) 6= Bulldozed/cleared area 7= Manicured grass Table 3-8. Percentage subjective meas urements that differed from objective measurements (= (1-Correct/15) Ã— 100) for Aquatic Ve getation Categories 1-4. Aquatic category Cover class 1 2 3 4 0-5% 6.7 0.0 13.3 0.0 6-25% 20.0 26.7 26.7 6.7 26-50% 20.0 40.0 26.7 0.0 51-75% 33.3 13.3 20.0 0.0 76-95% 6.7 6.7 6.7 6.7 95-100% 0.0 0.0 33.3 6.7 Mean (%) 14.5 14.5 21.1 3.4 1= Submerged and Floating Aquatics; 2 = Herbaceous Vegetation < 1 m high; 3 = Herbaceous Vegetation >1 m hi gh; 4 = Man-made Structures
45 Table 3-9. Percentage of subjective meas urements that differed from objective measurements (= (1-Correct/15) Ã— 100) for Terrestrial Vegetation Categories 1-7. Terrestrial category Cover class 1 2 2 4 5 6 7 0-5% 26.7 13.3 13.3 13.3 26.7 0.0 6.7 6-25% 20.0 26.7 26.7 13.3 20.0 13.3 26.7 26-50% 26.7 26.7 33.3 0.0 26.7 0.0 33.3 51-75% 26.7 13.3 20.0 13.3 6.7 0.0 20.0 76-95% 0.0 26.7 13.3 26.7 13.3 0.0 0.0 95-100% 0.0 0.0 0.0 0.0 6.7 0.0 0.0 Mean (%) 16.7 17.8 17.8 11.1 16.7 2.2 14.5 1 = Herbaceous Vegetation < 1 m high; 2 = Herbaceous Vegetation > 1 m high; 3 = Shrubs; 4 = Trees; 5 = Mixed Trees and Shrubs; 6 = Man-made Structures; 7 = Manicured Grass .
46 CHAPTER 4 RESULTS Waterbird Observations During JanuaryApril 2001 and 2002, I recorded a total of 4,808 and 5,666 waterbird observations (n = 10,474, x = 3.59, SD = 5.60) during surveys of 183 manmade ponds on 12 golf courses in southwest Florida. My observations included 30 species of waterbirds in 2001 and 40 species in 2002 (Table 4-1), which yielded a grand total of 42 species over both years. Diving birds were the most commonly obs erved waterbirds recorded during the 2year study period. Specifically, the two mo st abundant species were Double-crested Cormorants and Anhingas. Anhingas (in 2001) and Double-crested Co rmorants (in 2002) were also observed on more study lakes than any other species (Table 4-1). The most abundant species of wading bird during 2001 was the Great Egret and in 2002 was the Little Blue Heron. Killdeer ( Charadrius vociferous ) were the most abundant shorebird identified in 2001 and 2002 (Table 4-1). Killdeer were also observed on more lakes than any other shorebird. The most abundant sp ecies of waterfowl in 2001 was the Hooded Merganser ( Lophodytes cucllatus ), despite being observed on only one of the 12 golf courses. During 2002 the most abundant specie s of waterfowl recorded was the Mottled Duck ( Anas fulvigula ), which was observed on more study lakes than any other waterfowl species during both years.
47 Foraging Guilds The most common behaviors of all bird s observed during both years were those associated with foraging and the least co mmon were those associated with nesting activities (Table 4-2). This indicated that go lf course lakes were being used primarily as foraging sites. The Diving Birds guild contai ned the highest number of bird observations during both 2001 and 2002 surveys, followed by Op en Water Waders (Table 4-3). The percentage of total birds from the latter guild remained th e same in both study years, but the total number of observati ons increased by 244 birds (1 7%) in 2002. The percentage of total observations increased significantly for both the Aerial Foragers and Dense Vegetation Waders guild (Table 4-3). The Mo ist-soil Foragers gu ild was the only guild to decrease in total number of birds and percentage of observations between study years (Table 4-3). Shoreline Vegetation Variables Cover of shoreline vegetation was <50% fo r most vegetative categories at each pond during both study years (Figures 4-1 a nd 4-2). However, percent cover of vegetation increased slightly in year 2 of the study for most of the vegetation categories, with significant increases in submerged and floating aquatic vegetation and aquatic herbaceous vegetation <1 m high during 2002 (T able 4-4). Significant increases were also observed in terrestrial shoreline vegetation for herba ceous vegetation <1 m high, shrubs, trees, man-made structures, and manicured grass (Table 4-4). Analysis of Pond Similarity: Effective Foraging Area and Shoreline Vegetation There was significant variability in the mean effective foraging area (i.e., littoral zone; m2) among ponds ( 2 = 6.71, p = 0.0096; 2001: Median = 0.33m2, Range = 0.08 Â– 0.57m2; 2002: Median = 0.30 m2, Range = 0.16 Â– 0.60 m2). There was less variability
48 among shoreline vegetation as indicated by clus ter analysis, which produced five groups (clusters) of ponds as having similar levels of aquatic and terrestria l shoreline vegetation during 2001 and 2002. In each case, cluster an alysis identified a major group of ponds that possessed similar coverage of aquatic (Table 4-5) and terre strial (Table 4-6) shoreline vegetation. Analysis of terrestrial shoreline vegetation dur ing 2002 appeared to be more strongly influenced by percent cover of manicured grass. Manicured grass was more along prevalent shorelin es during 2002 probably due to the higher water levels observed that year. The other 4 pond groups contained various le vels of shoreline vegetation and were generally dominated by e ither very high or very low amounts of a single vegetation category (Tables 4-5 and 4-6) . The large numbers of ponds in a single cluster indicate that most of the golf c ourse ponds provided similar vegetative cover (according to the vegetative measurements of this study). Furthermore, the mean percent cover of vegetation within the main cluste r was <10% for all vegetation categories. Productivity Measurements A hyperbolic relationship exists between Secchi depth (i.e., water clarity) and chlorophyll a concentrations for Florida lakes (LAKEWATCH 2001). Analysis of golf course productivity data result ed in the parameter estimates 1=2.0135 and 2= 0.4134 for the hyperbolic model: Chlorophyll a = e(( 1-log(secchi depth))/ 2) (LAKEWATCH 2001). This model was then used to predic t productivity (i .e., chlorophyll a ) levels from Secchi depth measurements in all of the golf course ponds , which revealed that golf course ponds in this study ranged from eutrophic to hypereutrophi c (Figure 4-3). Results from the t-test analysis also indica ted that there was a significant (t = 2.31, p = 0.02) decrease in
49 chlorophyll a levels from 2001 to 2002, however, all ponds remained either eutrophic or hypereutrophic. Foraging Guild Site Selection Logistic regression models for ponds with 1 waterbird identified different variables for each waterbird guild as important predictors of waterbird presence on golf course ponds in southwest Florida (Table 4-7). In every model, eith er total perimeter or surface area was identified as a positive signif icant predictor of bird presence. As the perimeter or surface area of ponds in creased, the chances of observing > 1 waterbirds also increased. Other predictor variables selected in the models varied with foraging guilds and included year, aquatic vegetation categ ories 1-3 (submerged and floating aquatics, herbaceous vegetation <1 m high, herbaceous vegetation >1 m high), terrestrial vegetation categories 1, 2, 4-7 (herbaceous vegetation <1 m high, herbaceous vegetation >1 m high, tree, trees and shrubs, man-ma de structures, manicured grass), and surrounding vegetation categories 1-6 (her baceous vegetation, trees and shrubs, residential lawns, golf course turf, ar eas under construction, areas cleared for construction). Results from the logistic regression analyses with the dependent variable Â‘waterbird presenceÂ’ redefined as the presence of 5 birds as the dependent va riable, are provided in Table 4-7. Either perimeter length or surf ace area was again chosen as a significant predictor for all foraging guilds with a positiv e effect on bird presence. As the perimeter or surface area of the ponds increased, the chances of observing a large number of waterbirds ( 5) increased. Many of the other predic tor variables selected in the model changed when presence of birds was redefined as 5 birds. No combination of variables
50 would converge in a model for the Dense Ve getation Waders guild, likely due to the small number of observations for this guild (Tab le 4-3). Variables selected in models for the other five guilds included year, aquatic vegetation categories 2, 3 and terrestrial vegetation categories 1, 2, 4, 5, 7. Scenario tables were also devel oped for each foraging guild (for both 1 bird and 5 birds models) to illustrate the ability of the regression models to predict the presence of birds from each foraging guild on gol f course ponds given different amounts of vegetation and other significan t predictor variables (Tables 4-24 to 4-35). Pond size, defined either by its perimeter or surface area, was positively correlated with bird presence for all foraging guilds. In general, as the surface area or perimeter of the pond increased, the probability of observing a bi rd from all of the foraging guilds also increased. For the purpose of making predictio ns about the influence of pond size on bird presence, pond size (i.e., surface area and peri meter) was divided into 4 size classes: small (< 0.4 ha; <500 m), medium (0.4-2 ha; 500-1000 m), large (2-4 ha; 1001-2000 m) and very large (4-10 ha; 2001-3000 m). Pond size most often fell into one of the first 3 categories; hence the median values from these classes were used in the scenario tables to make predictions. More specifically, the major ity of the ponds were classified as having a small perimeter (<500 m) or medium surface ar ea (0.4-2 ha [1-5 ac]). Therefore, to make predictions about bird presence, these cl asses were used more often in the scenario tables than the other two. All vegetation categories excep t aquatic category 4 (man-mad e structures in liitoal zone) and terrestrial category 3 (shrubs) were chosen by at least one of the models as a significant predictor for the presence of birds within each foraging guild. To determine a
51 more realistic influence of pr edictor variables, only percen t coverage classes that were most commonly observed on the majority of golf courses were used in the scenario tables. For example, due to the small amount s of vegetation present in the golf course ponds (the majority of each shoreline vegetation category remained below 50% coverage), the percent coverage categories used in the scenario tables were from the two lowest cover classes (i.e., 2.5% and 15%). Re gression model results can also be used to make predictions about the influence of ve getative cover categories higher than those observed in this study (e.g., >50% vegetative co ver), but the validity of these predictions needs to be tested with empirical data. Due to the overriding influence of the Â‘ golf courseÂ’ variable many of the other predictor variables were no longe r significant once this variable was added to the logistic regression models. This result suggests th at the golf course, as a unit, may have an overall effect on waterbird use of the golf course ponds. The differences among golf courses may be due to variables that were not easily captured by an alyses but may have included differences in human use, management practices, or location of the courses in relation to other resources for the birds that were not quantified during this study. The influence of the individual golf courses on foraging presence was analyzed separately because the primary objective of this study wa s to make comparisons of bird use among golf course ponds rather than am ong golf courses. However, the predictor va riables that continued to be significant once the Â‘golf cour seÂ’ variable was added to the models are also reported below for each of the foraging guilds.
52 Probability of Observing Diving Birds The presence of 1 Diving Birds was positively correlated with pond surface area as well as ponds with aquatic herbaceous ve getation <1 m high (A2) and trees and shrubs around the shoreline (T5; Table 4-8). Diving Bird presence was negatively correlated with ponds surrounded by herbaceous vegetation (S1; Table 4-8). Once the Â‘golf courseÂ’ variable was added to the model (Intercep t Estimate = -1.5421) only surface area ( 2 = 52.26, p < 0.0001) remained significant ( p <0.5 level). Similar to the model for > 1 Diving Bird, the probability of observing 5 birds from this guild was positively correlated with surface area and ponds whose shorelines possessed trees and shrubs (T5; Table 4-8). However, unlike the first model for this guild, groups of Diving Birds were also pos itively influenced by herbaceous vegetation <1 m high around the shoreline of the ponds (T 1; Table 4-8). When the Â‘golf courseÂ’ variable was added to this model only surface area ( 2 = 59.19, p <0.0001) and shoreline trees and shrubs ( 2 = 6.01, p = 0.01) continued to be signi ficant predictors of bird presence. Diving Birds were the most commonly observed foraging guild during this study and the probability of observing a Diving Bird was very high (0.68-1.0) even with low vegetative coverage (Table 4-9). This tabl e also illustrates that, for ponds >3 ha, increasing the vegetative coverage around the ponds to more than 2.5% coverage will have only minimal influences on Diving Bird s because the probability is already 1.0. Observations of groups of Diving Birds were somewhat less frequent but were still commonly observed in large ponds (>3 ha; Probability = 0.80; Table 4-10).
53 Probability of Observing Open Water Waders Two competing models existed for the presence of 1 bird from the Open Water Waders guild. These models had almost id entical Goodness of Fit deviances (Model 1 = 229.78, Model 2 = 229.61) and differed only in th e tradeoffs between the two predictor variables: aquatic herbaceous vegetation >1 m high (A3; Model 1) and shoreline trees (T4; Model 2; Table 4-11). These predicto r variables were most likely correlated with one another as indicated by th eir non-significant Chi-square values when both variables were included in the model and, th erefore, both models are discussed. In the first model the presence of 1Open Water Waders was positively correlated with pond surface area and with ponds surrounded by residential lawns (S3) and golf course turf grass (S4; Table 4-11). Open Wa ter Waders were also positively influenced by ponds with aquatic herbaceous vegetation >1 m high (A3; Table 4-11). Once the Â‘golf courseÂ’ variable was added to the model (Intercept Estimate = 1.0513) surface area ( 2 =19.56, p < 0.0001), surrounding golf course turf grass ( 2 =4.16, p = 0.04) and aquatic herbaceous vegetation >1 m high ( 2 = 4.60, p = 0.03) continued to be significant predictors of birds from this guild. The variables selected as significant positive predictors of the presence of Open Water Waders in Model 2 were the same as Model 1 with the exception of a positive correlation with trees around the shoreline (T4) rather than tall aquatic vegetation (A3; Table 4-11). When the Â‘golf courseÂ’ variable was added to this model (Intercept Estimate = 0.9113) only surface area ( 2 =18.36, p < 0.0001) and surrounding golf course turf grass ( 2 =2.19, p = 0.02) continued to be significant pred ictors of presence of birds from this guild.
54 The variables that influenced the presen ce of a group of Open Water Waders were similar to the models for 1 birds with positive correlations with both aquatic herbaceous vegetation >1 m high (A3) and ponds adjacent to residential lawns (S3; Table 4-11). However, unlike the models for 1 bird from this guild, the probability of observing groups of birds increased with increasing pond perimeter (rather than surface area) and percent coverage of trees and shrubs al ong the shorelines (T5; Table 4-11). The probability of observing 1 Open Water Waders with any combination of the second modelÂ’s predictor variables (Table 4-13) was greater than the probabilities predicted by the first model (T able 4-12). However, the probability of observing an Open Water Wader was high (0.50-1.0) compared to other guilds, regardless of which model was used to formulate predictions. These mode ls also illustrate th at vegetative coverage >2.5% will only slightly increase the like lihood of observing these birds in ponds >3ha. In comparison, the chances of observing a group of Open Water Waders were somewhat less and only reached >90% (Table 4-14) for the ponds with perimeters 1500 m and at least 2.5% coverage of other significant predictor variable s (S3, T5, A3). Probability of Observing Dense Vegetation Waders The presence of 1 Dense Vegetation Waders wa s positively correlated with increasing pond perimeter. These species were also positively influenced by the coverage of pond-edge trees and shrubs (T5) and the coverage of submerged and floating aquatic vegetation (A1; Tabl e 4-15). The presence of Dense Vegetation Waders was negatively correlated with ponds adjacent to ar eas cleared for construction (S6) and areas currently under construction (S5; Table 4-15). Once the Â‘golf courseÂ’ variable was added to the model (Intercept Estimate = -0.6190) only perimeter ( 2 =10.00, p = 0.002) and
55 shoreline trees and shrubs ( 2 = 4.13, p = 0.04) remained significant predictors of presence of birds from this guild. Due to the small absolute number of obs ervations of Dense Vegetation Waders during this study, the probability of observing a bird from this guild remained low with all combinations of predictor variables (Table 4-16). Additionally, the logistic regression model would not converge with any combinati on of variables when the response variable was 5 birds and, therefore, a second regression model or scenario table could not be produced for this guild. Probability of Observing Dipping and Dabbling Foragers The presence of 1 Dipping and Dabbling Foragers was positively correlated with pond perimeter (Table 4-17) and ponds whose shorelines were covered with manicured grasses (T7) and other short he rbaceous vegetation (T1). The presence of birds from this guild was negatively correlated with herb aceous vegetation surrounding the ponds (S1) and man-made structures along th e shorelines (T6; Table 4-17 ). When the Â‘golf courseÂ’ variable was added to the model, the mode l would not converge with perimeter as a variable. When perimeter was excluded, signi ficant variables that remained when the golf course variable was added (Intercept Estimate = -26.3809) included shoreline herbaceous vegetation <1 m ( 2 =3.94, p = 0.05) and shoreline manicured grass ( 2 = 7.65, p = 0.006). Similar to the model above for 1 birds, when the presence of Dipping and Dabbling Foragers was redefined to 5 birds, there was a positive correlation for bird presence with pond perimeter, shorelines covered with manicured grasses (T7), and herbaceous vegetation <1 m high (T1; Tabl e 4-17). Groups of Dipping and Dabbling
56 Foragers were also positively associated with ponds with aquatic herbaceous vegetation <1 m high (A2), trees along the shoreline (T4), and those that were surrounded by manicured lawns (S3; Table 4-17). All of the significant predictor variables other than trees represent short vegetative cover either around the pondÂ’s perimeter or in the area surrounding the ponds. The model would not converge with any combination of predictor variables once the Â‘golf course Â’ variable was added to the model. Due to the strong negative correlation with manmade structures (intercept = -1.2519), even small cover of man-made structur es resulted in a low probability for these birds to be observed on a golf course pond, regardless of the size of the pond. On ponds without this variable, the probability of observing a Dipping and Dabbling Forager was greatly increased (Probability = 0.49; Table 4-18). Although the pr obabilities listed in the scenario table for groups of Dipping and Dabbling Foragers are higher (Table 4-19) than those listed in the scenario table for 1 bird (Table 4-18), the probabilities of observing groups of these birds are somewhat infl ated in the tables by the lack of a strong negative correlation with man-made structures (T6) as observed for the 1 bird model. Probability of Observing Moist-soil Foragers The presence of 1 Moist-soil Forager increased with pond perimeter (Table 4-20, 4-21). The probability of observing birds fr om this guild decreased on ponds surrounded by shrubs and trees (S2), or t hose with shorelines covered w ith manicured grass (T7), and tall (>1 m high) vegetation (T2; Table 4-20, 4-21 ). Study year was also able to explain a significant amount of the variab ility in presence of Moistsoil Foragers on golf course ponds with fewer birds seen in year 2 (2002) of the study (Table 4-20, 4-21). When the Â‘golf courseÂ’ variable was added to this m odel (Intercept Estimate = -1.2849) three of the
57 five variables remained signi ficant predictors of bird pr esence including perimeter ( 2 = 35.44, p < 0.0001), surrounding shrubs and trees ( 2 = 7.39, p = 0.007) and year ( 2 = 15.80, p < 0.0001). When the Â‘presence of Moist-soil Fora gersÂ’ (i.e., dependent variable) was redefined as 5 birds from this guild, the size of the pond Â–defined by its surface area rather than perimeter Â– still played a positive role in determining the presence of birds from this guild (Tables 4-20, 422). Similar to the model for 1 birds, the probability of observing groups ( 5) of Moistsoil Foragers decreas ed on ponds adjacent to shrubs and trees (S2), or those with shorelines covere d with manicured grass (T7), or tall (>1 m high) vegetation (T2; Tables 4-20, 4-22). Ho wever, unlike the first model, groups of Moist-soil Foragers were also positively associated with ponds surrounded by areas cleared for construction (S6; Tables 4-20, 422). Once the Â‘golf courseÂ’ variable was added to this model (Intercept Estimate = -2.7923) only perimeter ( 2 = 26.82, p < 0.0001) continued to be a signif icant predictor of groups of birds from this guild. Probability of Observing Aerial Foragers The presence of Aerial Foragers on golf course ponds was positively correlated with the perimeter of the pond and study y ear. Aerial Foragers were negatively influenced by aquatic herbaceous vegeta tion < 1 m high (A2; Table 4-23). The probability of observing 1 Aerial Foragers was gr eatest on larger ponds (1500 m perimeter) with < 2.5% short aquatic vege tation (Table 4-24). The number of Aerial Foragers also differed significantly between ye ars, which is illustrated by the increased probabilities predicted for year 2 (Table 424). When the Â‘golf courseÂ’ variable was added to this model (Intercept Estimate = -1.1868) only perimeter ( 2 = 28.40, p
58 < 0.0001) and year ( 2 = 19.02, p < 0.0001) remained significant predictors of Aerial Forager presence. When the dependent variable in the logist ic regression model was the probability of observing 5 Aerial Foragers, the percent cover of aquatic herbaceous cover < 1 m high was no longer correlated with presence of Ae rial Foragers (Table 4-23). However, similar to the 1 bird model, presence of 5 aerial foragers was correlated with size of the ponds, although for this model surface area of the ponds was a significant positive predictor rather than perimeter (Table 423). Additionally, although the probability of observing groups ( 5) of Aerial Foragers was also greater in year 2 of the study, the probability was very low for both year 1 (<0. 0001) and year 2 (<0.05), regardless of the size of the pond (Table 4-25). Once the Â‘ golf courseÂ’ variable was added, the model would no longer converge with any comb ination of predictor variables. Analysis of Golf Course Selection A means separation within a generalized linear model was used to perform a siteby-site comparison of observing a bird from each of the foraging guilds on each of the golf courses. These models were used to produce a ranked order based on the log of the odds ratio of observing a species from each of the foraging guilds on each of the courses when compared to the 11 other courses. The odds ratios (i.e., the chances of seeing a bird vs. not seeing a bird) and the probabilities of seeing a bird from a foraging guild on each of the courses are presented in Tables 4-26 to 4-31. Total surface area of ponds on each course was found to be a significant predictor of the average abundance birds observed on each golf course. The lack of fit analysis determined that a quadratic equation (p = 0.1493) was the most appropriate model for the
59 surface area data. The generali zed linear model resulted in a significant (F = 4.76; p = 0.0197) relationship, indicating that the coll ective pond surface area may explain some of the variation in average bird abundance among golf courses (R2=0.3131). Total surface area of ponds on each course was not, however, a significant (F = 1.93; p = 0.1706) predictor of the average density of birds observed on each golf course.
60Table 4-1. Waterbirds recorded during surveys of 183 golf course ponds in southwest Florida during 2001 and 2002. Total abundance, density (birds/total ha for all golf course ponds), and number of lakes where species were observed are listed for each year. 2001 2002 Species (scientific name) Observed Density Number Observed Density Number Diving Birds Anhinga ( Anhinga anhinga ) 462 1.971 111 481 2.052 107 Double-crested Cormorant ( Phalacrocorax auritus ) 1262 5.383 105 1816 7.746 119 Hooded Merganser ( Lophodytes cucllatus ) 220 0.938 9 20 0.085 7 Lesser Scaup ( Aythya affinis ) 0 78 0.333 3 Pied-billed Grebe ( Podilymbus podiceps ) 136 0.580 38 111 0.473 25 Ring-necked Duck ( Aythya collaris ) 0 1 0.004 1 Ruddy Duck ( Oxyura jamaicensis ) 0 1 0.004 1 GUILD SUMMARY 2080 8.872 152 2508 10.697 150 Open Water Waders Glossy Ibis ( Plegadis falcinellus ) 84 0.358 24 165 0.704 21 Great Egret ( Ardea albus ) 312 1.331 107 221 0.943 79 Great Blue Heron ( Ardea herodias ) 167 0.712 85 173 0.738 79 Little Blue Heron ( Egretta caerulea ) 280 1.194 100 397 1.693 108 Sandhill Crane ( Grus canadensis ) 3 0.013 2 4 0.017 2 Snowy Egret ( Egretta thula ) 241 1.028 74 289 1.233 68 Tricolored Heron ( Egretta tricolor ) 176 0.751 73 244 1.041 78
61Table 4-1. Continued 2001 2002 Species (scientific name) Observed Density Number Observed Density Number Open Water Waders (continued) White Ibis ( Eudocimus albus ) 94 0.401 31 114 0.486 29 Open Water Waders (continued) Wood Stork ( Mycteria americana ) 41 0.175 18 35 0.149 14 GUILD SUMMARY 1398 5.963 162 1642 7.004 155 Dense Vegetation Waders American Bittern ( Botaurus lentiginosus ) 0 1 0.004 1 Green Heron ( Butorides virescens ) 27 0.115 21 69 0.294 35 Black-crowned Night-Heron ( Nycticorax nycticorax )8 0.034 4 14 0.060 4 GUILD SUMMARY 35 0.149 24 84 0.358 38 Dipping and Dabbling Foragers American Coot ( Fulica americana) 24 0.102 2 24 0.102 2 Blue-winged Teal ( Anas discors ) 65 0.277 16 65 0.277 8 Common Moorhen ( Gallinula chloropus ) 251 1.071 17 260 1.109 28 Hybrid (Mottled Duck and Mallard) 0 1 0.004 1 Mottled Duck ( Anas fulvigula ) 189 0.806 58 286 1.220 70 Wood Duck ( Aix sponsa ) 0 2 0.009 1 GUILD SUMMARY 529 2.256 68 638 2.721 80
62Table 4-1. Continued 2001 2002 Species (scientific name) Observed Density Number Observed Density Number Moist-soil Foragers Black-bellied Plover ( Pluvialis squatarola ) 0 3 0.013 2 Black-necked Stilt ( Himantopus mexicanus ) 0 4 0.017 2 Bonaparteâ€™s Gull ( Larus philadelphia ) 1 0.004 1 0 0 Common Snipe ( Gallinago gallinago ) 16 0.068 12 19 0.081 10 Killdeer ( Charadrius vociferous ) 297 1.267 99 200 0.853 60 Laughing Gull ( Larus atricilla ) 0 8 0.034 3 Ring-billed Gull ( Larus delawarensis ) 130 0.554 19 32 0.136 9 Unidentified Shorebird 118 0.503 22 244 1.041 36 Willet ( Catoptrophorus semipalmatus ) 4 0.017 4 3 0.013 1 Greater or Lesser Yellowlegs ( Tringa melanoleuca/flavipes ) 153 0.653 58 135 0.576 45 GUILD SUMMARY 719 3.067 119 648 2.764 91 Aerial Foragers Bald Eagle ( Haliaeetus leucocephalus ) 0 4 0.017 4 Belted Kingfisher ( Ceryle alcyon ) 38 0.162 33 119 0.508 67 Brown Pelican ( Pelecanus occidentalis ) 2 0.009 2 0 0 Forsterâ€™s Tern ( Sterna forsteri ) 4 0.017 2 3 0.013 2
63Table 4-1. Continued 2001 2002 Species (scientific name) Observed Density Number Observed Density Number Aerial Foragers (continued) Least Tern ( Sterna antillarum ) 0 2 0.009 1 Osprey ( Pandion haliaetus ) 0 16 0.068 10 Royal Tern ( Sterna maxima ) 3 0.013 2 2 0.009 1 GUILD SUMMARY 47 0.200 38 146 0.623 72 ANNUAL SUMMARY 4808 20.508 5666 24.167
64 Table 4-2. Percentage of birds engaged in each recorded behavior for 2001 and 2002 surveys. Behavior 2001 2002 Foraging or associated movements 46.2 45.7 Moving but not obviously foraging 9.6 5.6 Stationary/Resting activities 33.4 36.2 Drying Wings 4.5 3.3 Flushed 2.1 2.0 Preening 4.1 6.8 Nesting Activities 0.2 0.4 Table 4-3. Percentage of total bird observat ions for each foraging guild recorded during 2001 and 2002 surveys. P -values were calculated w ith Chi-Square tests used to determine whether changes in th e proportion of total birds from each foraging guild were sign ificant between years. 2001 2002 Foraging Guild Total % Total % 2 p Diving Birds 2,080 43.3 2,508 44.2 1.019 0.313 Open Water Waders 1,398 29.1 1,642 29.1 0.007 0.932 Dense Vegetation Waders 35 0.73 84 1.5 13.184 0.0003 Dipping/Dabbling Foragers 529 11.0 638 11.3 0.174 0.676 Moist-soil Foragers 719 15.0 648 11.4 25.360 <0.0001 Aerial Foragers 47 0.94 146 2.6 36.778 <0.0001
65Figure 4-1. Number of ponds in 2001 and 2002 in each cover class for aqua tic cover categories 1 through 4. 0 20 40 60 80 100 120 140 160 180 200 0-56-2526-5051-7576-9596-100Aquatic 4 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 160 180 200 0-56-2526-5051-7576-9596-100Aquatic 3 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 160 0-56-2526-5051-7576-9596-100Aquatic 1 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 1600-56-2526-5051-7576-9596-100Aquatic 2 Cover (%)Number of Ponds 2001 2002
66Figure 4-2. Number of ponds in 2001 and 2002 in each cover class for terres trial cover categories 1 through 7. . 0 20 40 60 80 100 120 140 0-56-2526-5051-7576-9596-100Terrestrial 1 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 160 180 0-56-2526-5051-7576-9596-100Terrestrial 2 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 160 180 200 0-56-2526-5051-7576-9596-100Terretrial 3 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 160 180 200 0-56-2526-5051-7576-9596-100Terrestrial 4 Cover (%)Number of Ponds 2001 2002
67Figure 4-2. Continued 0 20 40 60 80 100 120 140 160 180 200 0-56-2526-5051-7576-9596-100Terrestrial 5 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 160 180 200 0-56-2526-5051-7576-9596-100Terrestrial 6 Cover (%)Number of Ponds 2001 2002 0 20 40 60 80 100 120 140 0-56-2526-5051-7576-9596-100Terrestrial 7 Cover (%)Number of Ponds 2001 2002
68 Table 4-4. Comparison of percent cover (median, mean, and range) for shoreline vegetation measured on golf course ponds in southwest Florida during 2001 and 2002. Chi-Square statistic and P -values were calculated for comparisons of median % cover between 2001 and 2002 with the Kruskal-Wallis test. Vegetation category Median (%) Mean (%) Range (%) Aquatic 2001 20022001 2002 2001 2002 2 p Submerged/ floating aquatic 0.8 12.3 7.9 20.0 0 82.40 77.3 74.38 <0.0001 Herbaceous (<1 m high) 0 1.1 2.4 3.8 0 32.90 53.8 43.21 <0.0001 Herbaceous (>1 m high) 0 0 1.4 1.0 0 59.20 48.0 1.29 0.26 Man-made structures 0 0 0.3 0.1 0 9.7 0 3.0 0.42 0.52 Terrestrial Herbaceous (<1 m high) 0 43.4 5.3 8.9 0 77.00 66.6 36.76 <0.0001 Herbaceous (>1 m high) 0 0 2.2 1.6 0 32.50 47.7 0.50 0.50 Shrub 0 0 0.3 1.1 0 21.70 28.5 14.21 0.0002 Tree 0 0 0.2 0.5 0 7.5 0 29.9 4.35 0.04 Mixed shrub/ tree 0 0 2.0 0.5 0 56.90 43.3 1.04 0.31 Man-made structures 0 0 0.1 0.3 0 9.8 0 6.7 26.63 <0.0001 Manicured grass 0 16.4 16.4 31.3 0 97.50 97.5 48.25 <0.0001
69 Table 4-5. Results of cluster analysis of mean percent cover of aquatic shoreline vegetation measured for golf course ponds in southwest Florida during 2001 and 2002. Cluster 1 2001 (n = 117) 2002 (n = 103) Aquatic vegetation category x SD x SD Submerged/floating aquatic 2.20 3.85 5.87 4.95 Herbaceous (<1 m high) 0.13 0.47 2.14 3.48 Herbaceous (>1 m high) 0.04 0.38 0.24 0.93 Man-made structures 0.29 0.98 0.17 0.47 Cluster 2 2001(n = 36) 2002 (n = 52) Aquatic vegetation category x SD x SD Submerged/floating aquatic 3.86 5.10 32.32 8.28 Herbaceous (<1 m high) 10.43 6.92 4.29 6.53 Herbaceous (>1 m high) 4.47 5.85 0.59 1.60 Man-made structures 0.10 0.32 0.11 0.33 Cluster 3 2001 (n = 24) 2002 (n = 22) Aquatic vegetation category x SD x SD Submerged/floating aquatic 28.15 8.57 58.83 8.41 Herbaceous (<1 m high) 1.92 6.18 4.02 5.36 Herbaceous (>1 m high) 1.35 3.31 0.85 2.77 Man-made structures 0.57 2.00 0.01 0.01 Cluster 4 2001 (n = 5) 2002 (n = 3) Aquatic vegetation category x SD x SD Submerged/floating aquatic 73.32 5.62 15.52 13.05 Herbaceous (<1 m high) 0.05 0.12 46.17 6.69 Herbaceous (>1 m high) 0.83 1.86 1.96 0.80 Man-made structures 0.05 0.10 0.03 0.05 Cluster 5 2001 (n = 1) 2002 (n = 3) Aquatic vegetation category x SD x SD Submerged/floating aquatic 0.32 . 15.76 17.62 Herbaceous (<1 m high) 7.32 . 12.80 18.56 Herbaceous (>1 m high) 59.23 . 33.22 12.88 Man-made structures 0 . 0 0
70 Table 4-6. Results of cluster analysis of mean percent cover of terrestrial shoreline vegetation measured on golf course ponds in southwest Florida during 2001 and 2002. Cluster 1 2001 (n = 137) 2002 (n = 87) Terrestrial vegetation category x SD x SD Herbaceous (<1 m high) 3.64 6.27 6.98 7.53 Herbaceous (>1 m high) 1.92 5.20 1.27 6.13 Shrub 0.24 1.92 0.44 2.23 Tree 0.04 0.26 0.42 3.23 Mixed shrub/tree 0.26 3.01 0.09 0.84 Man-made structures 0.02 0.11 0.29 0.92 Manicured grass 1.37 5.13 4.17 5.86 Cluster 2 2001 (n = 21) 2002 (n = 31) Terrestrial vegetation category x SD x SD Herbaceous (<1 m high) 5.64 7.58 5.37 5.22 Herbaceous (>1 m high) 4.90 9.90 3.19 6.33 Shrub 1.01 2.25 1.72 .356 Tree 0.45 1.10 0.77 4.01 Mixed shrub/tree 0.97 3.68 2.58 8.84 Man-made structures 0.96 2.33 0.32 1.22 Manicured grass 58.17 12.43 64.09 10.41 Cluster 3 2001 (n = 13) 2002 (n = 30) Terrestrial vegetation category x SD x SD Herbaceous (<1 m high) 2.32 4.67 11.16 9.50 Herbaceous (>1 m high) 0.09 0.32 1.52 2.80 Shrub 0.17 0.62 3.16 7.37 Tree 0.01 0.02 0.66 2.36 Mixed shrub/tree 0 0 0.26 1.13 Man-made structures 0 0 0.15 0.47 Manicured grass 92.70 5.39 36.09 10.17 Cluster 4 2001 (n = 7) 2002 (n = 24) Terrestrial vegetation category x SD x SD Herbaceous (<1 m high) 4.58 8.47 0.70 1.75 Herbaceous (>1 m high) 1.56 2.68 0.34 1.23 Shrub 0 0 0.07 0.23 Tree 1.07 2.83 0.01 0.04 Mixed shrub/tree 41.81 10.16 0 0 Man-made structures 0 0 0.33 0.83 Manicured grass 45.10 10.17 93.34 4.78
71 Table 4-6. Continued Cluster 5 2001 (n = 5) 2002 (n = 11) Terrestrial vegetation category x SD x SD Herbaceous (<1 m high) 57.37 16.33 45.25 11.88 Herbaceous (>1 m high) 6.27 8.17 2.62 5.74 Shrub 0.76 1.71 1.34 3.56 Tree 0.76 1.71 0.32 0.84 Mixed shrub/tree 1.53 3.42 0 0 Man-made structures 0.15 0.34 0.11 0.31 Manicured grass 0 0 2.91 5.51 Figure 4-3. Estimated chlorophyll a levels ( g/L) concentrations for each golf course pond grouped by trophic state. Trophic le vel descriptions (oligotrophichypereutrophic) follow those of Forsberg and Ryding (1980). 0 20 40 60 80 100 120 140Oligotrophic < 3Mesotrophic 3-7Eutrophic 7-40Hypereutrophic > 40Number of Ponds 2001 2002
72Table 4-7. Summary of the Logistic Regressi on models predicting presence of waterbird fo raging guilds at surveyed golf course p onds in Southwest Florida in 2001 and 2002. Diving Birds Open Water Waders Dense Vegetation Waders Dipping and Dabbling Foragers Moist-Soil Foragers Aerial Foragers Predictor Variable 1A 5B 1A 5B 1A 5D 1A 5B 1A 5B 1A 5B Perimeter + + + + + + Surface area + + + + + Year + + Submerged and floating aquatics (A1) + Aquatic herbaceous vegetation < 1 m high (A2) + + Aquatic herbaceous vegeta tion > 1 m high (A3) +C + Shoreline herbaceous vegetation <1 m high (T1) + + + Shoreline herbaceous vegetation > 1 m high (T2) Shoreline trees (T4) +C + Shoreline trees and shrubs (T5) + + + + Shoreline man-made structures (T6) Shoreline manicured grass (T7) + + Surrounding herbaceous vegetation (S1) Surrounding trees and shrubs (S2) Surrounding residential lawns (S3) + + + Surrounding golf course turf (S4) + Surrounding areas under construction (S5) Surrounding areas cleared for construction (S6) + Predictor variables included perimeter, surface area, year, aquatic vegetation cate gories 1-4, terrestrial vegetation categorie s 1-4, surrounding categories 1-6, average effective foraging area, and estimated chlorophyll a . A Â‘+Â’ sign indicates a positive correlation and the Â‘-Â‘ indicates a negative relationship. A = Model 1 results were correlated with the presence of 1 bird from specified foraging guild. B = Model 5 results were correlated with the presence of 5 birds from specified foraging guild. C = Variables selected by 2 comp eting models. D = Model would not converge with any combination of variables.
73 Table 4-8. Results of logistic regression mode ls identifying significant predictor variables correlated with the presence of Divi ng Birds on golf course ponds in Southwest Florida surveyed during 2001 and 2002. Variable Model* df Log odds 2 p 1 1 1.9669 46.52 <0.0001 Surface area 5 1 0.8446 55.84 <0.0001 1 1 0.0640 5.15 0.0232 Aquatic herbaceous vegetation <1 m high (A2) -----1 1 0.0981 4.69 0.0303 Shoreline shrubs and trees (T5) 5 1 0.0529 9.23 0.0024 -----Shoreline herbaceous vegetation <1 m high (T1) 5 1 0.0205 4.04 0.0444 1 1 0.0148 3.96 0.0466 Surrounding herbaceous vegetation (S1) -----The limit for selection in the model was set at p = 0.05. Chi-Square and P -values were taken from LR Statistics for Type III Analysis. *Model 1 results predic t the presence of 1 bird (Intercept Estimate = -0.0234), Model 5 results predict the presence of 5 birds (Intercept Estimate = -1.4164). Table 4-9. Probability of observing a Diving Bird at a golf course under different scenarios using significant predictor variables that were previously identified through logistic regression. P(observing Diving bird)= e (Â–0.0234 + 1.9669 (Surface Area) + 0. 0640(A2) + 0.0981(T5) 0.0148 (S1)) 1 + e (Â–0.0234 + 1.9669 (Surface Area) + 0. 0640(A2) + 0.0981(T5) 0.0148 (S1)) Surface area (ha) Aquatic herbaceous vegetation <1 m high (A2) (% cover) Shoreline trees and shrubs (T5) (% cover) Surrounding herbaceous vegetation (S1) (% cover) Probability .20 (~1ac) 2.5 2.5 2.5 0.68 1.01(~2.5ac) 2.5 2.5 2.5 0.91 1.01 15 2.5 2.5 0.96 1.01 2.5 15 2.5 0.97 1.01 2.5 2.5 15 0.90 3.04 (~7.5ac) 2.5 2.5 2.5 1.0
74 Table 4-10. Probability of observing 5 Diving Birds at a golf course under different scenarios using significant predictor variables that were previously identified through logistic regression. P(observing 5 Diving birds) = e (Â–1.4164 + 0.8446(Surface Area) + 0.0529(T5) + 0.0205(T1)) 1 + e (Â–1.4164 + 0.8446(Surface Area ) + 0.0529(T5) + 0.0205(T1)) Surface area (ha) Shoreline herbaceous vegetation <1 m high (T1) (% cover) Shoreline trees and shrubs (T5) (% cover) Probability .20 (~1ac) 2.5 2.5 0.26 1.01(~2.5ac) 2.5 2.5 0.41 1.01 2.5 15 0.57 1.01 15 2.5 0.47 3.04 (~7.5ac) 2.5 2.5 0.79 Table 4-11. Results of logist ic regression models identifying significant predictor variables correlated with the presence of Open Water Waders on golf course ponds in Southwest Florida su rveyed during 2001 and 2002. Variable Model* df Log odds 2 p 1A 1 1.2999 20.79 <0.0001 1B 1 1.2352 18.39 <0.0001 Surface area ---------------Perimeter 5 1 1A 1 0.2705 5.31 0.0212 -----Aquatic herbaceous vegetation >1 m high (A3) 5 1 0.0745 6.32 0.0119 ----------Shoreline trees (T4) 5 1 0.0479 7.53 0.0061 1A 1 0.0445 19.51 <0.0001 1B 1 0.0460 21.19 <0.0001 Surrounding residential lawns (S3) 5 1 0.0181 8.40 0.0037 1A 1 0.0250 14.50 0.0001 1B 1 0.0271 17.15 <0.0001 Surrounding golf course turf (S4) -----The limit for selection in the model was set at p = 0.05. Chi-Square and P -values were taken from LR Statistics for Type III Analysis. *Model 1A results predict the presence of 1 bird (Intercept Estimate = -1.1074) for the first of 2 competing models for this guild; Model 1B results pred ict the presence of 1 bird (Intercept Estimate = -1.1446) for the seco nd of 2 competing models for this guild; Model 5 results pred ict the presence of 5 birds (Intercept Estimate = -2.4484).
75 Table 4-12. Model 1: Probability of observing an Open Water Wader at a golf course under different scenarios using signifi cant predictor variables that were previously identified thr ough logistic regression. P(observing Open Water Wader) = e(Â–1.1074 + 1.2999(Surface Area) + 0. 0445(S3) + 0.0250(S4) + 0.2705 (A3)) 1 + e (Â–1.1074 + 1.2999(Surface Area) + 0. 0445(S3) + 0.0250(S4) + 0.2705 (A3)) Surface area (ha) Aquatic herbaceous vegetation >1 m high (A3) (% cover) Surrounding residential lawns (S3) (% cover) Surrounding golf course turf (S4) (% cover) Probability .20 (~1ac) 2.5 2.5 2.5 0.50 1.01(~2.5ac) 2.5 2.5 2.5 0.74 1.01 2.5 15 2.5 0.83 1.01 2.5 2.5 15 0.80 1.01 15 2.5 2.5 0.99 3.04 (~7.5ac) 2.5 2.5 2.5 0.98 Table 4-13. Model 2: Probability of observing an Open Water Wader at a golf course under different scenarios using signifi cant predictor variables that were previously identified thr ough logistic regression. P(observing Open bird) = e (Â–1.1446 + 1.2352(Surface Area) + 0. 0460(S3) + 0.0271(S4) + 5.0516(T4)) 1 + e (Â–1.1446 + 1.2352(Surface Area) + 0. 0460(S3) + 0.0271(S4) + 5.0516(T4)) Surface area (ha) Shoreline trees (T4) (% cover) Surrounding residential lawns (S3) (% cover) Surrounding golf course turf (S4) (% cover) Probability .20 (~1ac) 2.5 2.5 2.5 1.0 1.01(~2.5ac) 2.5 2.5 2.5 1.0 1.01 2.5 15 2.5 1.0 1.01 2.5 2.5 15 1.0 1.01 15 2.5 2.5 1.0 3.04 (~7.5ac) 2.5 2.5 2.5 1.0
76 Table 4-14: Probabil ity of observing 5 Open Vegetation Waders presence in different scenarios using significant predictor variables that were previously identified through logistic regression. P(observing 5 Open birds) = e (-2.4484 + 0.0032(Perimeter) + 0.0181(S3) + 0.0479(T5) + 0.0745(A3)) 1 + e (-2.4484 + 0.0032(Perimeter) + 0.0181(S3) + 0.0479(T5) + 0.0745(A3)) Perimeter (m) Aquatic herbaceous vegetation >1 m high (A3) (% cover) Shoreline trees and shrubs (T5) (% cover) Surrounding residential lawns (S3) (% cover) Probability 300 2.5 2.5 2.5 0.24 300 2.5 2.5 15 0.29 300 2.5 15 2.5 0.37 300 15 2.5 2.5 0.45 700 2.5 2.5 2.5 0.54 1500 2.5 2.5 2.5 0.94 Table 4-15. Results of logistic regres sion model (Intercept Estimate = -2.5421) identifying significant predictor variable s correlated with the presence of Dense Vegetation Waders on golf course ponds in Southwest Florida surveyed during 2001 and 2002. Variable df Log odds 2 p Perimeter 1 0.0012 22.54 <0.0001 Submerged/floating aquatics (A1) 1 0.0152 4.27 0.0387 Shoreline trees and shrubs (T5) 1 0.0360 4.27 0.0388 Surrounding area under construction (S5) 1 -0.1049 4.31 0.0379 Surrounding area cleared for construction (S6) 1 -0.1051 7.42 0.0064 The limit for selection in the model was set at p = 0.05. Chi-Square and P -values were taken from LR Statistics for Type III Analysis.
77 Table 4-16. Probability of obs erving a Dense Vegetation Wa der at a golf course under different scenarios using si gnificant predictor variables that were previously identified through logistic regression. P(observing Dense bird)= e ( Â–2.5421+ 0.0012 (Perimeter) 0.1051 (S6) 0.1049 (S5) + 0.0360 (T5) + 0.0152 (A1)) 1 + e (Â–2.5421+ 0.0012 (Perimeter) 0.1051 (S6) 0.1049 (S5) + 0.0360 (T5) + 0.0152 (A1)) Perimeter (m) Submerged and floating aquatics (A1) (% cover) Shoreline trees and shrubs (T5) (% cover) Surrounding areas under construction (S5) (% cover) Surrounding areas cleared for construction (S6) (% cover) Probability 300 2.5 2.5 2.5 2.5 0.07 300 2.5 2.5 2.5 15 0.02 300 2.5 2.5 15 2.5 0.02 300 2.5 15 2.5 2.5 0.11 300 15 2.5 2.5 2.5 0.08 700 2.5 2.5 2.5 2.5 0.11 1500 2.5 2.5 2.5 2.5 0.24
78 Table 4-17. Results of logistic regressi on model identifying significant predictor variables correlated with the presence of Dipping and Dabbling Foragers on golf course ponds in Southwest Flor ida surveyed during 2001 and 2002. Variable Model*df Log Odds 2 p 1 1 0.0008 11.00 0.0009 Perimeter 5 1 0.0012 21.10 <0.0001 -----Aquatic herbaceous vegetation <1 m high (A2) 5 1 0.0639 10.07 0.0015 1 1 0.0445 19.24 <0.0001 Shoreline herbaceous vegetation <1 m high (T1) 5 1 0.0411 10.26 0.0014 -----Shoreline trees (T4) 5 1 0.1743 8.96 0.0028 1 1 -1.2519 19.33 <0.0001 Shoreline man-made structures (T6) -----1 1 0.0209 29.35 <0.0001 Shoreline manicured grass (T7) 5 1 0.0241 25.03 <0.0001 1 1 -0.0185 5.72 0.0167 Surrounding herbaceous vegetation (S1) ----------Surrounding residential lawns (S3) 5 1 0.0177 4.93 0.0265 The limit for selection in the model was set at p = 0.05. Chi-Square and P -values were taken from LR Statistics for Type III Analysis. *Model 1 results predic t the presence of 1 bird (Intercept Estimate = -1.3645), Model 5 results predict the presence of 5 birds (Intercept Estimate = -4.1977).
79 Table 4-18. Probability of obs erving a Dipping or Dabbling Forager at a golf course under different scenarios using signifi cant predictor variables that were previously identified thr ough logistic regression. P(observing Aquatic Vegetation Forager)= e (-1.3645 +.0008 (Perimeter) + .0209(T7) + .0445(T1) Â– 1.2519 (T6) -0.0185 (S1)) 1 + e (-1.3645 + .0209(T7) + .0445(T1) Â– 1.2519 (T6) + .0008 (Perimeter) -0.0185 (S1)) Perimeter (m) Shoreline manicured grass (T7) (% cover) Shoreline herbaceous vegetation <1 m high (T1) (% cover) Man-made structures on shoreline (T6) (% cover) Surrounding herbaceous vegetation (S1) (% cover) Probability 300 2.5 2.5 2.5 2.5 0.02 300 15 2.5 2.5 2.5 0.02 300 2.5 15 2.5 2.5 0.03 300 2.5 2.5 15 2.5 <0.0001 300 2.5 2.5 2.5 15 0.01 300 2.5 2.5 0 2.5 0.27 700 2.5 2.5 2.5 2.5 0.02 700 2.5 2.5 0 2.5 0.33 1500 2.5 2.5 2.5 2.5 0.04 1500 2.5 2.5 0 2.5 0.49 Table 4-19: Probabil ity of observing 5 Dipping and Dabbling Foragers at a golf course under different scenarios using signifi cant predictor variables that were previously identified thr ough logistic regression. P(observing 5 Dipping and Dabbling Foragers) = e (-4.1977 + 0.0012(Perimeter) + 0.0241(T7) + 0.0411(T1) + 0.0639(A2) + 0.1743(T4) + 0.0177(S3)) 1 + e (-4.1977 + 0.0012(Perimeter) + 0.0241(T7) + 0.0411(T1) + 0.0639(A2) + 0.1743(T4) + 0.0177(S3)) Perimeter (m) Shoreline manicured grass (T7) (% cover) Shoreline herbaceous vegetation <1 m high (T1) (% cover) Aquatic herbaceous vegetation <1 m high (A2) (% cover) Shoreline trees (T4) (% cover) Surrounding manicured lawns (S3) (% cover) Probability 300 2.5 2.5 2.5 2.5 2.5 0.05 300 15 2.5 2.5 2.5 2.5 0.06 300 2.5 15 2.5 2.5 2.5 0.07 300 2.5 2.5 15 2.5 2.5 0.10 300 2.5 2.5 2.5 15 2.5 0.30 300 2.5 2.5 2.5 2.5 15 0.06 700 2.5 2.5 2.5 2.5 2.5 0.07 1500 2.5 2.5 2.5 2.5 2.5 0.17
80 Table 4-20. Results of logistic regressi on model identifying significant predictor variables correlated with the presence of Moist-soil Foragers on golf course ponds in Southwest Florida surveyed during 2001 and 2002. Variable Model* df Log Odds 2 p 1 1 0.0017 34.11 <0.0001 Perimeter ----------Surface area 5 1 0.4749 28.99 <0.0001 1 -0.7823 9.86 0.0017 Year -----1 1 -0.0564 6.37 0.0116 Shoreline herbaceous vegetation >1 m high (T2) 5 1 -0.0801 3.88 0.0488 1 1 -0.0112 9.57 0.0020 Shoreline manicured grass (T7) 5 1 -0.0128 6.94 0.0085 1 1 -0.0420 16.24 <0.0001 Surrounding trees and shrubs (S2) 5 1 -0.0353 6.03 0.0140 -----Surrounding areas cleared for construction (S6) 5 1 0.0283 5.28 0.0215 The limit for selection in the model was set at p = 0.05. Chi-Square and P -values were taken from LR Statistics for Type III Analysis. * Model 1 results predict the presence of 1 bird (Intercept Estim ate = 0.3834), Model 5 results predict the presence of 5 birds (Intercept Estimate = -1.5027). Table 4-21. Probability of obs erving a Moist-soil Forager at a golf course under different scenarios using significant predictor variables that were previously identified through logistic regression. P(observing Moist-soil bird) = e (0.3834 + 0.0017(Perimeter) Â– 0.0420(S2) Â– 0.0112(T7) Â– 0.0564(T2) Â–0.7823(Year Indicator)) 1 + e (0.3834 + 0.0017(Perimeter) Â– 0.0420(S2) Â– 0.0112(T7) Â– 0.0564(T2) Â–0.7823(Year Indicator)) Perimeter (m) Shoreline herbaceous vegetation >1 m high (T2) (% cover) Shoreline manicured grass (T7) (% cover) Surrounding shrubs and trees (S2) (% cover) Probability Year 1 (2001) Probability Year 2 (2002) 300 2.5 2.5 2.5 0.65 0.46 300 2.5 2.5 15 0.52 0.33 300 2.5 15 2.5 0.43 0.26 300 15 2.5 2.5 0.48 0.30 700 2.5 2.5 2.5 0.79 0.63 1500 2.5 2.5 2.5 0.93 0.87
81 Table 4-22. Probability of observing 5 Moist-soil Foragers at a golf course under different scenarios using si gnificant predictor variables that were previously identified through logistic regression. P(observing 5 Moist-soil birds) = e (-1.5027 + 0.4749(Surface Area) Â– 0.0128(T7) Â– 0.0353(S2) + 0.0283(S6) Â– 0.0801(T2)) 1 + e (-1.5027 + 0.4749(Surface Area) Â– 0.0128(T7) Â– 0.0353(S2) + 0.0283(S6) Â– 0.0801(T2)) Surface area (ha) Shoreline manicured grasses (T7) (% cover) Surrounding herbaceous vegetation (S2) (% cover) Surrounding areas cleared for construction (S6) (% cover) Shoreline herbaceous vegetation >1 m high (T2) (% cover) Probability .20 (~1ac) 2.5 2.5 2.5 2.5 0.16 1.01 (~2.5ac) 2.5 2.5 2.5 2.5 0.22 1.01 15 2.5 2.5 2.5 0.20 1.01 2.5 15 2.5 2.5 0.15 1.01 2.5 2.5 15 2.5 0.29 1.01 2.5 2.5 2.5 15 0.09 3.04 (~7.5ac) 2.5 2.5 2.5 2.5 0.42 Table 4-23. Results of logistic regressi on model identifying significant predictor variables correlated with th e presence of Aerial Foragers on golf course ponds in Southwest Florida surveyed during 2001 and 2002. Variable Model* df Log Odds 2 p 1 1 0.0014 34.24 <0.0001 Perimeter ----------Surface area 5 1 0.6078 13.01 0.0003 1 1 1.0676 18.36 <0.0001 Year 5 1 26.9727 7.82 0.0052 1 1 -0.0651 7.55 0.0060 Aquatic herbaceous vegetation <1 m high (A2) -----The limit for selection in the model was set at p = 0.05. Chi-Square and P -values were taken from LR Statistics for Type III Analysis. * Model 1 results predict the presence of 1 bird (Intercept Estim ate = -2.1827), Model 5 results predict the presence of 5 birds (Intercept Estimate = -32.0366).
82 Table 4-24. Probability of observi ng an Aerial Forager at a golf course under different scenarios using significant predictor variables that were previously identified through logistic regression. P(observing Aerial bird) = e (-2.1827 + 0.0014(Perimeter) Â– 0.0651(A2) + 1.0676(Year Indicator)) 1+ e (-2.1827 + 0.0014(Perimeter) Â– 0.01651(A2) + 1.0676(Year Indicator)) Perimeter (m) Aquatic herbaceous vegetation <1 m high (A2) (% cover) Probability Year 1 (2001) Probability Year 2 (2002) 300 2.5 0.13 0.30 300 15 0.06 0.16 750 2.5 0.21 0.44 750 15 0.11 0.26 1500 2.5 0.43 0.69 1500 15 0.26 0.50 Table 4-25. Probabil ity of observing 5 Aerial Foragers at a golf course under different scenarios using significa nt predictor variables that were previously identified through logistic regression. P(observing 5 Aerial birds) = e (Â–32.0366 + 0.6078(Surface Area) + 26.9727(Year Indicator)) 1 + e (Â–32.0366 + 0.6078(Surface Area ) + 26.9727(Year Indicator)) Surface area (ha) Probability Year 1 (2001) Probability Year 2 (2002) .20 (~1ac) <0.0001 0.01 1.01(~2.5ac) <0.0001 0.01 3.04 (~7.5ac) <0.0001 0.04 Table 4-26. Ranked order of the odds ratio a nd probability of obser ving a Diving Bird on each golf course compared to other courses. . Course Odds ratio Probability Rank Bay Island 1.0377E 11 1.0 A Gateway 18.9993 0.95 B Spring Run 10.0001 0.9091 B Sabal 6.9999 0.875 B Copper Leaf 6.6665 0.8696 B Mediterra 6.3331 0.8636 BC Creekside 5.5146 0.8465 BC Marsh 4.9998 0.8333 BCD Twin Eagles 3.8000 0.7917 BCDE Burnt Store 2.1111 0.6786 CDE Wildcat Run 2.0769 0.675 DE Cypress 1.2223 0.55 E Sites with the same letter are not signif icantly different from one another at p = 0.05
83 Table 4-27. Ranked order of the odds ratio a nd probability of observing an Open Water Wader on each golf course compared to other courses. Course Odds ratio Probability Rank Bay Island 2.8204E 11 1.0 A Gateway 2.8204 E 11 1.0 A Wildcat Run 39.0015 0.975 B Burnt Store 26.9990 0.9643 B Creekside 25.0006 0.9615 B Marsh 11.0001 0.9167 B Copper Leaf 10.5003 0.913 B Spring Run 7.7998 0.8864 BC Sabal 4.9998 0.8333 BCD Cypress 2.3333 0.7 CD Mediterra 1.7500 0.6364 D Twin Eagles 1.6666 0.625 D Sites with the same letter are not signif icantly different from one another at p = 0.05. Table 4-28. Ranked order of the odds rati o and probability of observing a Dense Vegetation Wader on each golf course compared to other courses. Course Odds ratio Probability Rank Gateway 0.8182 0.45 A Twin Eagles 0.6000 0.375 A Wildcat Run 0.5385 0.35 AB Burnt Store 0.4000 0.2857 ABC Bay Island 0.3684 0.2692 ABC Marsh 0.2000 0.1667 BCD Creekside 0.1304 0.1154 CD Cypress 0.1111 0.1 CD Mediterra 0.0732 0.0682 D Spring Run 0.0476 0.0454 D Sabal 0.0435 0.0418 D Copper Leaf 3.5432 E -12 0.0 E Sites with the same letter are not signifi cantly different from one another at the p = 0.05.
84 Table 4-29. Ranked order of the odds ratio and probability of observing a Dipping and Dabbling Forager on each golf course comp ared to other courses. Sites with the same letter are not significantl y different from one another at p = 0.05. Course Odds ratio Probability Rank Marsh 4.9998 0.8333 A Gateway 4.0004 0.8 A Creekside 2.7142 0.7308 A Spring Run 2.3845 0.7045 A Bay Island 2.2499 0.6923 AB Copper Leaf 0.8400 0.4565 BC Mediterra 0.7600 0.4318 C Burnt Store 0.1200 0.1071 D Cypress 0.0526 0.05 D Wildcat Run 3.5428E -12 0.0 E Sabal 3.5361 E -12 0.0 E Twin Eagles 3.5379 E -12 0.0 E Table 4-30. Ranked order of odds ratio and pr obability of observing a Moist-soil Forager on each course compared to other courses. Course Odds ratio Probability Rank Copper Leaf 14.3335 0.9348 A Gateway 8.9998 0.9 AB Cypress 5.6667 0.85 AB Sabal 3.0000 0.75 BC Mediterra 2.1428 0.6818 BC Twin Eagles 1.1819 0.5417 CD Spring Run 1.0952 0.5227 CD Burnt Store 0.7500 0.4286 DE Wildcat Run 0.7391 0.425 DE Bay Island 0.5294 0.3461 DE Creekside 0.3684 0.2692 EF Marsh 0.0909 0.0833 F Sites with the same letter are not signif icantly different from one another at p = 0.05.
85 Table 4-31. Ranked order of the odds ratio and Probability of seeing an Aerial Forager on each course compared to other courses. Course Odds ratio Probability Rank Wildcat Run 1.6666 0.625 A Sabal 1.1819 0.5417 AB Twin Eagles 0.7143 0.4167 ABC Gateway 0.6666 0.4 ABC Cypress 0.6666 0.4 BC Burnt Store 0.5555 0.3571 BC Mediterra 0.3750 0.2727 C Marsh 0.3333 0.25 CD Creekside 0.2381 0.1923 CDE Bay Island 0.2381 0.1923 CDE Copper Leaf 0.0952 0.0869 DE Spring Run 0.0732 0.0682 E Sites with the same letter are not signif icantly different from one another at p = 0.05.
86 CHAPTER 5 DISCUSSION In this study, I evaluated the effects of numerous habitat variables on waterbird use of constructed golf course ponds in southwest Florida. Forty-two recorded species of waterbirds used the golf course ponds from January through April 2001 and 2002. My analysis of recorded bird activities indicated that 46% of all waterbirds observed used golf course ponds as foraging habitat. I regard this as a conservative estimate of foraging use because the remaining 54% may have used the golf courses as foraging habitat as well, but were engaged in other activities dur ing surveys. These, results, therefore, indicated, that man-made ponds on golf cour ses provide food resources capable of attracting many species of waterbirds. The low densities of birds, however, (<2 birds/ha for most species; Table 4-1), suggests the valu e of golf course ponds to waterbirds can be enhanced through habitat modifications de signed to appeal to specific guilds. There was year-to-year vari ation in both the number a nd species of waterbirds observed in golf course ponds, with an incr ease in both abundance and species richness during 2002 (Table 4-1). Waterb irds are highly mobile speci es capable of moving large distances to track suitable re sources (Weller 1999). Therefor e, this increase may have been due to greater amounts of rainfall in the region before the second study season, perhaps because increased rain prolongs suitable foraging conditions in wetlands throughout the region (see Pineau 2000; Welle r 1999). Increased ra infall also raised water levels in the ponds, which may have im proved the suitability of golf course ponds as waterbird foraging habitat. For example, water levels fluctuate annually and often
87 substantially in smaller ponds, which may b ecome nearly dry during drought years, but may form shallow pools during wet years (perso nal observation). Increased rainfall may also have enhanced the attr activeness of golf courses by flooding other low-lying areas on the courses. Many of these areas were c overed with natural vegetation and may have created ideal conditions for concentration of prey as water receded. Wading birds, in particular, seemed to be attracted to these flooded areas (per sonal observation). Alternatively, the large annual di fferences for some species may also have been an effect of forces operating at regi onal or continental scales. Effects of Pond Size on Waterbird Site Selection Pond size, defined either by its perimeter or surface area, was positively correlated with bird presence for all foraging guilds (Table 4-17). In general, as the surface area or perimeter of the pond increased, the probability of observing a bird from all of the foraging guilds also increased. Larger ponds may provide more foraging habitat for individual species of waterbirds , as well as provide a wider va riety of habitats to support a diversity of species. Similar relationships have been report ed for birds in many types of freshwater habitats (Celada & Bogliani 1993; Gibbs et al. 1991; Hoyer & Canfield 1990; Nudds 1992). Total combined surface area for all wate r bodies on each golf course also had a positive correlation with average bird abunda nce on individual golf courses. Golf courses with more water had more birds on av erage. This suggests that waterbirds may key in on landscape-level influences and choos e courses with more water because of the increased likelihood of finding suitable si tes. However, although pond size may influence bird use of golf courses and golf course ponds, this variable may not be as important as other predictor variables such as the amount of shallow water (water depth
88 <40 cm) and suitability of vegetation in a nd around the ponds. Without other suitable habitat features, such as shallow water ar eas, many species (e.g., wading birds) may be limited in their ability to use ponds for foraging, regardless of pond size. Effects of Effective Foraging Area on Waterbird Site Selection The availability of food is the most crucial feature for determining habitat suitability for waterbirds. Availability include s not only the density of suitable prey (i.e., species, size, and capture susceptibility of th e prey) but also the accessibility of the prey items to the birds (Gawlik 2002; Kushlan 2000b) . This is particularly important for wading birds and shorebirds, which are often co nfined to water depths no greater than their leg length during foraging, indicating that the amount of shallow water in the ponds may be important to waterbird site selection. However, the size of the effective foraging area in golf course ponds was not selected as a significant predictor for any of the foraging guilds even though there was signifi cant variability of this variable among ponds. A possible explanation for this findi ng is that while there was variation among ponds in the size of the effective foraging areas, an unknown critical size threshold, which influences wading bird use of the ponds, was not reached. Effects of Vegetation on Waterbird Site Selection All vegetation categories quantified duri ng this study, except shrubs along the shoreline and man-made structures in the litto ral zone, were chosen by at least one of the logistic regression models as a significant pr edictor of the presence of birds from each foraging guild (Table 4-7). The predictor va riables that were selected by the logistic regression models influenced waterbird site selection even within the small range of vegetative cover (<50% coverage at mo st ponds) among ponds. Therefore, these variables should be considered important determinants of waterbird site selection of golf
89 course ponds and may be used to develop management recommendations to enhance the value of golf course ponds to waterbirds. I de fined Â‘presence of bird sÂ’ is defined as one or more birds and Â‘groups of birdsÂ’ is define d as five or more birds from a foraging guild. Habitat Selection by Foragi ng Guilds: Pond Variables Diving Birds Golf course ponds seemed particularly attractive as foragi ng habitat for Diving Birds. Diving Birds were the most abundant species recorded duri ng both years of the study, making up 42% of total waterbird obser vations and occurring on 168 (91%) golf course ponds. The highly productive nature of these ponds may yield increased prey densities, which may favor diving birds becaus e they are able to use a large range of water depths for foraging (ponds ranged in average depth from 1 to 15 m). Presence of Diving Birds was positively correlated with pond surface area, perhaps due to an increase in available foraging habita t and prey populations in larger ponds. For example, Double-crested Cormorants (the most commonly observed Diving Bird) are piscivorous and have been found to be partic ularly sensitive to wetland surface water area (Gibbs et al. 1991), perhaps because fish a bundance and diversity on small lakes (<50 ha) is positively correlated with wetl and area (Haines et al. 1986). Diving Bird presence was also positively correlated with short (<1 m high) aquatic herbaceous vegetation. This type of short, emergent vegetation may provide cover for small forage fish and invertebrates (Chick & Mclvor 1994) that constitute the principle diets of these piscivorous bird s. Although these prey specie s are attracted to the cover from this vegetation (i.e., aquatic vegeta tion <1 m high) it may be less effective protection against Diving Birds such as A nhingas and Pied-billed Grebes, which appear
90 able to move easily through and capture pr ey within dense vegetation (Esler 1992; Frederick & Siegel-Causey 2000). Trees and shrubs also positively affected bird abundance in this guild. Large numbers of Double-crested Cormorants and Anhingas were often observed resting in trees and shrubs when they were not fo raging. More specifically, Double-crested Cormorants were commonly observed resting in large numbers (>10) in the trees along the shore and on vegetated islands located in the center of the ponds (Chapter 3). Anhingas were generally observed resting alone or in small groups of 2-5 birds in trees or shrubs around the pondÂ’s shoreline (personal observation). It is worth noting that Anhingas require snags, rocks or trees growing in or near the edge of the water as perches while drying their wings and plumage. Since Anhingas have completely wettable plumage, drying is a critical for thermoregulatio n for this species, th erefore, perches near or over the water are said to be a critical component of habitat su itability (Frederick & Siegel-Causey 2000). Groups of Diving Birds were often observed on ponds with short herbaceous vegetation around the pondsÂ’ shoreline (within 1 m of the waterÂ’s edge). Large groups of birds may prefer shorelines covered with na tural vegetation so long as the vegetation is not excessively tall (>1 m high), which may im pede movement to and from the water or conceal mammalian predators (Gosser et al . 1997; New York State Department of Environmental Conservation 1999; Williams -Whitmer et al. 1996). Ponds whose shorelines are vegetated with natural herbaceous vegetation may also be able to buffer birds foraging near the shore or resti ng on the shore from human disturbances.
91 Diving Birds were negatively influenced by herbaceous ve getation in areas adjacent to the ponds perhaps because much of the herb aceous vegetation in these areas (e.g., cord grasses [ Spartina spp.] and bulrushes [ Scirpus spp.]) was planted by course groundskeepers and tended to be taller than manicured areas. Th erefore, Diving Birds may have responded negatively to herbaceous vegetation in areas adjacent to the ponds because of reduced detection of pot ential ground predators (e.g., raccoons [ Procyon lotor ] and house/feral cats [ Felis silvestris catus ]) while engaged in activities such as resting, preening, or wing-drying on the shore. Indeed, a reduced ability to detect predators may affect avoidance of tall or dense vegetation in a number of waterbird species (Henderson 1981; Smith 1994; Surdick 1998). The predictor variables discussed above prim arily concern those factors that likely influence site selection by foraging or res ting Diving Birds. The Anhinga was the only waterbird species observed nesting on a ny golf course pond. The most conspicuous difference of the lone pond with nesti ng Anhingas was the presence of an island composed of willow trees ( Salix sp.), which were partially submerged in the water. Many species of waterbirds, including species such as colonial nesting wading birds, require vegetation over or surrounded by wate r to protect their nests from predators (Hafner 2000). Therefore, increasing the numbe r of vegetated islands within golf course ponds, particularly those with partially or occasionally submerged vegetation, may increase the number of birds that ar e able to find suitable nesting habitat. Open Water Waders Wading birds are attracted to food resour ces similar to those preferred by Diving Birds, but occurred at lower dens ities in this study (Table 4-1) . This may be an artifact of their different foraging requirements. Open Water Waders require suitable shallow water
92 areas that allow access to prey (Baker 1979; Powell 1987). The effective foraging area for wading birds was typically less than 2 m aw ay from the shoreline (average area <0.4 m2) due to the 4:1 slope ratio (horizontal to vertical) used to cons truct golf course ponds (ratio obtained from golf course blueprints). Nevertheless, Open Water Waders were the second most commonly observed guild of bird s on golf course ponds. Birds from this guild constituted 29% of total waterbird obs ervations and were observed on more ponds than any other guild (174 ponds or 95% of all ponds). Pond surface area and perimeter was positiv ely correlated with presence of Open Water Waders. This may be because larger ponds provided more total littoral zone and greater amounts of foraging habitat for thes e species. There was also a positive relationship between presence of Open Wa ter Waders and two surrounding vegetation categories, both of which represented types of short manicured grasses. Ponds that are adjacent to areas with short vegetation may allow increased predator detection compared to other types of surrounding ve getation (e.g., wooded areas). Open water waders were attracted to ponds with tall (> 1 m high) herbaceous vegetation in the littoral zone, perhaps because this vegetation provides a screen for the birds from human disturbance. However, increased vegetation de nsity and height may also reduce a birdÂ’s ability to remain vigilant against predators by limiting mobility and visual detection abilities (Henderson 1981; Smith 1994; Surdick 1998). As a result, wading birds may prefer tall (>1 m high) vege tation so long as it is of moderate density (Gibbs et al. 1991; Smith 1994; Surdick 1998). The positive correlation with ponds that po ssessed a thin row of trees around the shoreline perhaps reflects the need for resting and roosting sites. Trees may also provide
93 escape from terrestrial predators or human dist urbances as many birds flushed to the trees when disturbed (personal observation). Stands of trees and shrubs in areas adjacent to the ponds were not positively correlated with open water waders, suggesting that these birds may be attracted to ponds with trees around th e waterÂ’s edge, but not necessarily ponds that are adjacent to larg e areas of woody vegetation. Open Water Wader site selection may also have been influenced by the presence of other species in the golf course ponds. Fo r example, although Diving Birds were often seen foraging in the deeper, central areas of the ponds, they also were observed moving from the center of the ponds toward the shor e, presumably forcing prey to shallower areas to make them more accessible. This type of foraging activity by groups of diving birds often attracted large gr oups of foraging wading birds such as Great Egrets, Snowy Egrets, and Great Blue Herons along the shoreline (personal observation). Commensal associations have been documented for a number of species and, presumably, the foraging activity of diving birds increases pr ey exposure to wading birds (Ehrlich et al. 1988; Kushlan 1978). In this study, wading bird s appeared to attend cormorants, which may have concentrated or Â‘herdedÂ’ prey thr ough group foraging activities. However, data was not collected during this study to dete rmine foraging rates of wading birds in the presence of other species and alone, therefor e, the influence of cormorants on foraging success rates could not be analyzed. Dense Vegetation Waders Dense vegetation waders were the leas t commonly observed foraging guild during both years of surveys, making up only 1% of th e total bird observations and occurring on only 26% of golf course ponds. The presence of dense vegetation waders was positively correlated with the presence of trees and shrubs around the sh oreline of the ponds (Table
94 4-15), which provide protective cover as well as resting and/or roosting sites for these species (Kushlan 2000a; Sibley 2000). Thes e habitat components , however, were not common as indicated by the mean percent cove rage for this type of vegetation, which was less than 3% for all ponds surveyed (Table 4-4). It is also worth noting that this guild is also the most cryptic of all guilds recorded and, therefore, also the least likely to be recorded if present. Dense Vegetation Waders were also positiv ely correlated with the presence of submerged and floating aquatic vegetation (Tab le 4-15). Algae and submerged plants in the golf course ponds often formed rings around shoreline (Zinn & Redles 2002), possibly because near-shore areas were shal lower and provided greater light penetration for plant growth (Weller 1999). Aquatic ve getation may attract Dense Vegetation (and other shallow water waders) by facilitating the co ncentration of prey ite ms near the shore, because small forage fish often use these areas as feeding grounds and refuge from predators (Chick & Mclvor 1994). The presence of Dense Vegetation Wade rs was negatively correlated with ponds adjacent to areas cleared for construction and areas currently under construction (Table 415). Therefore, the small numbers of observa tions of birds from this guild relative to other guilds may be due, in part, to the small amount of dense vegetation around the ponds (>170 ponds had 5% coverage or less; Figu re 4-9) as well as construction activities associated with residential deve lopment around some of the ponds. Dipping and Dabbling Foragers Dipping and Dabbling Foragers made up 11% of total waterbird observations and were observed on 94 (51%) of the golf cour se ponds. Dipping and Dabbling Foragers most often forage by surface dipping and dabbling in shallow water. Therefore, birds
95 from this guild may not have been as prevalent as other guilds because of the depth of ponds and lack of shallow littoral zone habitat that eliminated much of the guild-specific foraging habitat. Although the presence of Dipping and Dabbli ng Foragers was positively correlated with pond perimeter, birds from this foragi ng guild were less likely than most other guilds to be observed on golf course ponds, regardless of th e size of the pond, when manmade structures were present along the shoreline (Table 4-18). Conversely, the probability of observing these birds increas ed significantly for ponds without these structures (Table 4-18). Many of the man-ma de structures observed along the shoreline of the ponds were the visible, upper portions of drains and culverts used to connect the flow of water among ponds, which often form ed walls and ledges at the waterÂ’s edge, particularly when water levels were low. Other structures recorded during habitat mapping included wood and cement walls that we re often planted with flowers and other vegetation for aesthetic purposes, but were functionally used to increase the elevation of the course near the ponds. The ledges formed by these structures were usually 0.5 m above the water and may have formed physical barriers that restrict ed the movements of Dipping and Dabbling Foragers between th e water and shore (Gosser et al. 1997; Williams-Whitmer et al. 1996). The presence of Dipping and Dabbling Foragers was positively correlated with ponds whose shorelines were covered with ve ry short vegetation (T able 4-17). Mottled Ducks and Blue-winged Teal ( Anas discors ) often rested on the shores of the ponds to perform daily activities such as preening. Short vegetation along the shore of the ponds may allow increased detection of potentia l predators, such as alligators ( Alligator
96 mississippiensis ) at least by comparison with tall (>1 m high) herbaceous along shorelines. The positive influence of residential lawn s adjacent to golf course ponds on groups of Dipping and Dabbling Foragers may also be re lated to the height of this vegetation. These areas were usually relatively free from predators and offered more open areas for vigilance compared to ponds surrounded by wooded areas or tall herbaceous vegetation (which negatively influenced presence of these birds). Fu rthermore, ponds surrounded at least partially by residential areas may experi ence less human disturbances (because these areas do not usually experience daily mainte nance) than when completely surrounded by the golf courses. The finding that groups of Dipping and Dabbling Foragers often used ponds with short littoral zone vegeta tion and trees along the shorel ine may be a result of the protective cover (at least from human disturbances) that these areas provide. Therefore, even a thin row of trees along the shoreline or herbaceous vegetation in the littoral zone of the ponds may provide a buffer from human disturbances such as maintenance crews and golfers. Short, robust emergent herbaceous vegetation (e.g., spikerush [ Eleocharis spp.]) in littoral zone may also provide perche s for these birds, allowing them to stay near the water and remain inconspicuous when not foraging. This was particularly common for Common Moorhens ( Gallinula chloropus ), which were often seen standing on short or fallen vegetation (e.g., bulrush or fire flag [ Thalia geniculata ]) in the water. Some of the herbaceous vegetation, such as bulrush seed s, may also have been eaten by species in this guild, which often feed on the seed s of such plants (Bent 1926; Greij 1994).
97 Moist-soil Foragers Moist-soil Foragers made up a small per centage (13%) of the total waterbirds observations but were observed on 74% ( 135 ponds) of the golf course ponds. I hypothesize that the positive in fluence of pond size on Moist-soil Forager presence was due to the increased foraging habitat provided by ponds with la rger perimeters (Gibbs et al. 1991). Although analyses determined surf ace area to be a better predictor for groups of Moist-soil Foragers than pond perimeter, ponds with larger surf ace areas generally had larger perimeters and thereby increased fora ging habitat as well. There was a negative correlation of this guildsÂ’ presence with ponds with shorelines covered with manicured grass or tall herbaceous vegetation. Shoreb irds feed in muddy areas and other soft sediments and manicured turf grasses or tall vegetation probably interfered with their ability to forage. Moist-soil Foragers avoided ponds adjacent to areas with trees and shrubs, possibly because these areas conceal potential predators (e.g., ha wks and falcons) of these typically small shorebirds. The positive correl ation of groups of Mois t-soil Foragers with ponds adjacent to areas cleared for construc tion also indicates that these birds are sensitive to areas with heavy vege tative cover adjacent to the ponds. Aerial Foragers Aerial Foragers made up only 2% of the to tal bird observations, but were observed on almost half (48%) of the ponds surveyed. Birds from this guild were most often observed on large ponds, which was probably an artifact of large ponds being more likely to possess a suitable perching site. Solitary aerial foragers were negatively correlated with short (<1 m high) aquatic herbaceous vegetation (Table 4-23). However, the absence of a negative correlation between so litary aerial foragers and tall aquatic
98 vegetation or the presence of groups (> 5) of aerial foragers, suggests the negative correlation with short aquatic vegetation may have been an artifact of small sample size. Habitat Selection by Waterbirds: Comparison of Golf Courses The Â‘golf course variableÂ’ was a signifi cant predictor variable for all 6 foraging guilds. Additionally, when this variable was added to the logistic regression models, many of the other predictor va riables were no longer signif icant predictors of bird presence (Tables 4-11 to 4-16). This is an interesting finding because the overall variability within the predictor variable categories was low for all golf courses (Figures 4-1 to 4-10). Furthermore, the cluster anal ysis suggested that the majority of the ponds could be placed in the same group based on ve getative cover (Tables 4-5 to 4-8). This implies that each of the golf courses provided similar habitats. However, one important difference among golf courses was the total co mbined surface area of the water bodies. This variable was found to have a signifi cant relationship with average bird abundance and was able to explain some of the va riation in average bi rd abundance among golf courses. However, total combined surface area was unable to explain differences in average bird density among golf courses. Therefore, increases surface area leads to increased abundance but not higher densities. Distance from one habitat to another also influences environmental heterogeneity, and spatially closer habitats are generally more similar to each ot her than distant ones (see Williamson 1988). Therefore, artificial we tlands within the same golf course were probably more similar to one another than to those located in other golf courses. This may result in a Â‘golf course effectÂ’ whereby the golf course as a whole (i.e., spatial location, golf play, management practices, et c.) provides different or unique habitat characteristics for waterbirds. Furthermore, the probability of observing birds from the 6
99 foraging guilds was significantly different among golf courses (Tables 4-26 to 4-31), indicating that there were di fferences in waterbird hab itat selection among the golf courses. One potential explanation for the differences in bird presence among golf courses is that the variables measured during this study we re not discriminatory at the spatial scale at which they were measured (i.e., individual pond scale). The Â‘golf courseÂ’ variable was a significant predictor of bird presence for a ll foraging guilds implying that this variable served as a gross classification of habitat types present at ea ch golf course. It has been suggested that birds follow a hierarchically structured sy stem during habitat selection (Cody 1985; Weller 1999). Therefore, the bi rds observed on golf courses may either reside in or migrate to Sout hwest Florida for a variety of reasons, including the regionÂ’s landscape or vegetation features, instinctive pr eferences, and prior experience. Once the birds reach Southwest Florida they may then select suitable hab itat sites at the golf course scale or larger. It seems likely that the high mobility of waterbirds allows them to select an area with many potential habi tat sites such as a golf course with many ponds. The birds may then select from available ponds based on more specific habitat characteristics such as pond size or vegetation structure and/or density. Birds may also select golf courses based on other landscape-scale features such as the total combined surface area of water bodies. However, ot her landscape scale features should be explored in future studies to improve understanding of bird use of golf courses. For example, distance from each of the cour ses to FloridaÂ’s coast and other natural foraging sites may have influenced use of golf course ponds by Moist-soil Foraging species. Because shorebirds of ten leave preferred habitats such as beaches and mudflats
100 during high or low tides (Bur ger et al. 1997; Erwin et al . 1994), surrounding habitats, particularly those located in close proximity to natura l foraging habitats, may be important alternative habitats to species from the this guild (Brusati et al. 2001). Therefore, the influence of distance from other foraging sites on waterbird use of golf courses should be further investigated. Comparison with Natural Areas A common question asked of st udies in urban landscapes is how the densities of wildlife documented in urbanize d or man-made habitats co mpares with that found in natural areas. This comparison is difficu lt and perhaps unfair for both the urban landscape and the natural one for many reasons , which can be grouped into two primary categories, spatial and tempor al. Spatially, the comparison is complex because of the difficulty in determining from which natura l setting bird density should be compared. For example, if one were to compare bird densities from golf c ourse ponds with those from a large natural setting such as the Florid a Everglades, the densities may be higher or lower simply due to the manner in which comparisons are made. Wildlife is not normally evenly distributed in nature, and the vastne ss of the Everglades would provide very low overall densities that might poor ly reflect the importance of th is ecosystem. Similarly, if waterbird densities in the Everglades were measured only in areas of concentrated feeding activity or at roos ting sites, the comparisons would provide equally poor comparisons with overall dens ities of birds using golf c ourse ponds. Another example may be a comparison between golf course ponds and a lake, such as FloridaÂ’s Lake Okeechobee. As with the Everglades, the large size of Lake Okeechobee makes direct comparisons of little practical use. This is because the ratio of deep water to shallow littoral zone habitat on Lake Okeechobee far ex ceeds that of the much smaller lakes on
101 golf courses. What this particular compar ison does point out, however, is that the extensive shoreline provided by multiple golf course ponds provides excellent opportunities to provide habitat for multiple species of waterbirds. Temporally, the comparison with natural sy stems is also complicated because of the difficultly in determining the time of year or season from which to make a comparison with other studies. Since movement of waterbirds is often related to rainfall patterns (Pineau 2000; Weller 1999), comparisons should be made during years with similar rainfall levels and at similar times during the year or across a range of years, particularly in regions that experience distinct dry and wet seasons. Many naturally fluctuating wetlands may provide suitable habita t for a large number of species but for a shorter period of time than the more perman ent golf course ponds. As the water levels recede, the prey in shallow, natural wetland s may be depleted because there is reduced prey refugia compared to permanent wate r bodies (Edelson & Co llopy 1990), such as golf course ponds. Many natural areas often ex perience a boom and bust type of bird use as waterbirds deplete the prey resources and are forced to use other habitats until these sites are re-flooded and the prey base has recovered (Hafner & Britton 1983; Kushlan 1978, 2000a). These temporal fluctuations furt her increase the difficulty of comparisons with natural wetlands. Another temporal con cern is that comparisons would have to be made during similar times of the year based on migratory patterns in each region. Many waterbird species are migratory, which may infl ate or decrease the number of birds in a region during different times of the year. Th e tendency for waterbirds to search for suitable foraging areas in response to changing water levels also begs the question as to
102 whether golf course ponds, with their semi permanent wetlands status, may provide important foraging locations in times of severe drought. A more appropriate comparison of the wa terbird densities ob served during this study would be with densities observed on ot her golf course ponds. Temporal concerns other than regional migration may be less of a problem when comparing golf courses because they are more permanent than natura l wetlands and less likely to be depleted of all prey resources. These comparisons may yi eld information about regional differences in waterbird diversity and abundance on gol f course ponds. Likewise, comparisons between densities of waterbirds on the same courses over time would provide opportunity to monitor the success of habitat modificat ions designed to increase use of ponds by waterbirds. Potential Health Risks Associat ed with Eutrophic Conditions I was unable to investigate the potential hea lth risks to birds foraging in golf course ponds during this study. However, one concern is that eutrophic conditions, which were characteristic of golf course ponds, have b een associated with parasites such as Eustrongylides ignotus (Spalding & Forrester 1993). This nematode parasite can have major impacts on the health of nestling and a dult birds (Spalding et al. 1993). Eutrophic conditions may also attract elev ated numbers of birds due to their increased biological productivity. By increasing the number of fora ging birds in a small area the transmission rates of diseases such as avian chlolera and botulism ma y also be magnified through increased contact among birds (Enright 1971; Samuel et al. 1999). The finding that golf course ponds attract waterbirds emphasizes the need for research to evaluate the potential health risk factors associ ated with these sites.
103 Conclusions and Suggestions for Future Research Waterbird habitat selection of golf course s probably takes place at many spatial scales beginning with the selection of a geogr aphic region in which to search for suitable habitats (Cody 1985; Weller 1999). It is likely that birds then search for areas that provide multiple foraging opportunities such as a wetland complex or a golf course with large water body surface area. Other factors su ch as the distance from other foraging habitats (e.g., such as the coast for shorebirds ) or from roosting and nesting sites may also influence waterbird use of a particular golf course. At a more refined spatial scale, waterb irds may then select which pond from the golf course/wetland complex will meet th eir immediate needs (e.g., foraging, resting, nesting). The principle features of golf cour se ponds selected by waterbirds during this study were size (i.e., surface area or perimeter) and cover of shoreline and surrounding vegetation. Wetland size is an important factor for fora ging site selection for many species of waterbirds. Larger areas often provide more foraging habitat for a single species of waterbirds and/or may also provide a wider variety of habitats for a greater number of species. On the other hand, sma ller wetlands have the pot ential to be managed individually for suitable habitat features not found in larger wetlands, such as particular type or density of vegetation. The type and structure of vegetation on gol f course ponds is also important because it creates microhabitats that can be chosen for various activities (Weller 1999). For permanent bodies of water such as golf course ponds, much of the vegetation is confined to near shore areas that are also the primar y foraging area for many species of waterbirds (e.g., waders). Not surprisingly, different spec ies of waterbirds have different habitat needs when using golf course ponds. Some species habitat needs (e.g., Diving Birds,
104 Open Water Waders and Dense Vegetation Waders) overlap, such as the need for roosting and resting sites that provide protection from pred ators and human disturbances, but which are located at a minimal distance from foraging sites. Others species (e.g., Moist-soil Foragers) seem to prefer ponds with minimal amounts of vegetation both in and around the ponds. The vegetation surrounding golf course ponds is also important because nearby trees or shrubs may provide r oosting or resting sites (e.g., for groups of Diving Birds and Open Water Waders). The structure and density of surrounding vegetation may also influence a birdÂ’s ability to detect approaching predators, thereby deterring birds from using ponds surrounded by forested areas (e .g., Moist-soil Foragers). However, waterbirds preference for vegetation in area s adjacent to the ponds is somewhat more complicated due to the different preference s for each of the 6 foraging guilds. One commonality seems to be the preference for ve ry short herbaceous vegetation. However, the negative relationship with these natural areas should be treated very cautiously, because they do not indicate natural vegetation surrounding golf course ponds discourages use of the ponds by all species of waterbirds. What they do imply is that experimental evidence is needed to determ ine their relationship with other foraging guilds for which no relationship could be es tablished. In any case, managing ponds or sections of ponds to provide a diversity of habitats will benefit the most species while also providing greater manageme nt options for golf courses. One of the weaknesses of this analysis was that there was relatively little variability in measured variables on golf course ponds. Fo r this reason, the effects of some variables may have been masked or diminished by th e relatively monomorphic habitat studied, yet
105 would be of importance in areas with a broa der range of habitat t ypes and densities. Although the probability of observing a bird fr om each guild varied for each golf course, the probability of observing a diving bird a nd an open water wader was >50% on all of the courses surveyed. These results illustrate that golf course ponds are widely used and provide potentially important habitat in increasingly urbani zed areas. The large number of waterbird species (42 sp ecies total) also indicate s that golf course ponds are biologically productive and have the potential to provide habitat for a diversity of waterbirds. The extent to wh ich waterbirds used golf course ponds is primarily related to the availability of suitable habitat features influencing security and foraging success. The low densities of birds, <2 birds/ha for most species (Table 4-1), suggest there is excellent opportunity to increase the value of golf c ourse ponds to waterbirds through relatively simple habitat modifications. As a corollary, one area for future st udy includes experimentally testing the management recommendations from this study (i.e., increasing the littoral zone and creating vegetated islands for nesting; see belo w) to better understa nd their influence on waterbird site selection on golf courses. However, it is impor tant to note that this study quantified variables known to influence waterb ird habitat selection in natural wetlands. Therefore, habitat variables other than thos e measured during this study and/or measured at other scales (e.g., regional scale) shoul d be also included in future studies. Management Recommendations The range of predictor variables select ed for each foraging guild suggests that diversity in design and management of gol f course ponds can provide a local wetland complex that provides suitable habitat for multiple species of waterbirds. Providing a diversity of habitats (includi ng size, water depth, and vege tative cover), whether among
106 individual ponds or within the same pond, would provide th e greatest benefits to the largest number of species. Management opti ons designed to improve waterbird habitat may in some instances provide financial savings for golf courses through reduction of costs associated with turf maintenance in lo w-lying wet areas and other problem areas. Recommendations for improving foraging and re sting/roosting habitat on golf course ponds in south Florida are provided for each of the six waterbird guilds evaluated. Diving Birds Diving Birds were the most commonly obs erved waterbirds, indicating that golf course ponds possess many of the habitat featur es attractive to these species. Diving birds primarily forage on fish. Conseque ntly, stocking newly constructed ponds with fish, or ensuring connections with water bod ies that support fish, will provide suitable forage resources for these species. Roosting and resting sites that provide prot ection from predators and are located at a minimal distance from foraging sites are also essential waterbird habitat (Weller 1999). Planting trees and shrubs along the shoreline of golf course ponds could be nefit roosting and resting Diving Birds by allowing them to be off of the ground and away from predators and human disturbances. Planting shorelines with native, short herbaceous vegetation may benefit resting Diving Birds by reducing the amount of human disturbances (through routine maintenance) compared to shorelines with manicured vegetation. Removing tall vegetation from ar eas adjacent to the ponds may also benefit resting Diving Birds by increasing their abil ity to detect approaching predators. A final consideration involves increasing th e amount of nesting habitat available on golf courses. Most waterbirds require ne sting sites that afford protection against predators (including humans), which is gene rally provided by water surrounding the nest
107 site (Hafner 2000). Although A nhingas were the only species observed nesting during this study, the construction of vegetated islands with partially or occasionally submerged vegetation may benefit many species of nesting waterbirds on golf courses. Nest trees should be tolerant of flooding, but may includ e a variety of species such as cypress ( Taxodium spp.), red maple ( Acer rubrum ), live oak ( Quercus virginiana ), and southern willow ( Salix spp). Many nesting waterbird species are sensit ive to human disturbances (Rodgers & Smith 1995; Weller 1999), therefore, to reduce the perceive d threat from humans, nesting islands should probably be cons tructed in larger ponds so that the distance from the shore is maximized. It is also important to cons ider the adverse effect s of attracting nesting waterbirds such as the noise associated with large numbers of breed ing birds. The large amounts of guano associated with dense aggr egations of waterbirds may also cause problems such as foul odors destruction of ne st trees through increa sed acidification of the soil. Open Water Waders The greatest limitation to Open Water Wa ders was available foraging habitat, specifically suitable shallow water areas, which allow access to (Baker 1979; Powell 1987) and concentrate prey. Based on blueprin ts of the golf courses surveyed, the pond sides were constructed with 4: 1 slope (horizontal to vertical distance). In order to increase the amount of shallow-littoral zone surrounding the golf course ponds, the overall ratio could be increased (i.e., increased horizontal relati ve to vertical distance). This would likely increase the area suitabl e for foraging by making prey more accessible to birds such as Open Water Waders that re quire shallow (e.g., <40 cm) water depths for foraging. The effective foraging area for thes e birds could also be increased through the
108 construction of berms to create shelved or st air-stepped areas of di fferent elevation. The general premise of this type of structure is, that as water levels recede in the ponds, water and prey items become trapped in these shallow areas. Increased shallow water zones around the pondÂ’s perimeter may also help eliminate easy escape opportunities for fish. However, it may also be beneficial to place such modifications on portions of the shoreline not heavily used by golfers (e.g., those near residential areas or natural areas) to redu ce the amount of human disturbance on these birds. Increasing the littoral zone in maintenance problem areas (e.g., sections that are often too wet to mow) may also provide financ ial savings to the golf course because they would no longer have to rou tinely maintain these areas. Planting trees and shrubs al ong the shoreline of the ponds could benefit resting and roosting sites Open Water Waders by providi ng them escape cover from predators or human disturbances. Planting short vegeta tion in areas adjacent to ponds managed for Open Water Waders could also benefit resti ng as well as foraging birds through increased predator detection. Dense Vegetation Waders Dense Vegetation Waders were the least commonly observed foraging guild on golf courses. A primary limiting factor for these species may have been the amount of dense vegetation present in the ponds, which could be increased by increasing the amount of trees and shrubs on the shoreline of ponds or sections of ponds. Dense Vegetation Waders are sensitive to human disturbance and were negatively impacted by construction activities and areas ad jacent to the ponds that were de void of vegetation. A suggestion for these birds may be the construction of isla nds with dense vegetati on in the center of the ponds to reduce the impact of human activities adjacent to the ponds.
109 Due to concerns about excess primary pr oductivity caused by leaching and run-off from pesticides, herbicides, and fertilizers a pplied to the turf gra ss, chemicals are often applied to the ponds on a regular basis to cont rol algae and other aquatic plants (Zinn & Redles 2002). However, moderate amounts of algae and submerged aquatic plants may attract Dense Vegetation by facilitating the c oncentration of prey items near the shore, because small forage fish often use the litto ral zone as feeding grounds and refuge from predators (Chick & Mclvor 1994). Therefore, reduction in the quantit y and/or extent of chemical treatments in ponds managed to provide habitat for Dense Vegetation Waders may prevent beneficial algae and submerged aquatic vegetation growing in the littoral zone from being harmed. Another possible so lution to reduce nutrien t concentrations in the ponds without the use of chemicals may be to grow large amounts of macrophytic vegetation to absorb the nutrients and then re move it regularly. This type of treatment may be beneficial to several of the guilds if were done in a staggered manner across all ponds in the golf course. For example, when the macrophytes were initially removed, the conditions may favor Open Water Waders and when the macrophytes were at peak densities climax, the conditions would pr obably favor Dense Ve getation Foragers. Dipping and Dabbling Foragers Dipping and Dabbling Foragers most ofte n forage by surface dipping and dabbling in shallow water for the foliage and tubers of submerged plant beds (Weller 1999). Therefore, foraging birds could benefit from increased submerged, aquatic vegetation in the littoral zone. Foraging Dipping and Dabblin g Foragers may also benefit from short, emergent vegetation in the littoral zone because the seeds produced from plants such as bulrushes are often eaten by species from this guild (Bent 1926; Greij 1994).
110 Minimizing the number of man-made stru ctures, such as cement ledges and drainage structures, around the shoreline of golf course ponds could benefit Dipping and Dabbling Foragers because these structures impede movement to and from the shore (Williams-Whitmer et al. 1996). Keeping vege tation short along the shorelines and in areas adjacent to the ponds may also benef it resting Dipping and Dabbling Foragers by increasing visibility of nearby predators. Pl anting short, robust herbaceous vegetation (e.g., spikerushes) in the pondsÂ’ littoral zo nes may provide perches for Dipping and Dabbling Foragers (e.g., American Coots [ Fulica Americana ]), allowing these birds to stay near the water and remain inconspicuous when not foraging. Planting a thin row of trees and shrubs along the shoreline of the ponds may also benefit Dipping and Dabbling Foragers by reducing human disturbances (e .g., maintenance and golf play) to resting birds. However, the density of trees near the shoreline should be considered carefully because dense cover of this type of vegeta tion may reduce attractiveness for some species by decreasing predator visibility or obstru cting flight paths (Gosser et al. 1997). Moist-soil Foragers Species from the Moist-soil Foragers guild forage along the shoreline of the golf course ponds by either ground gleaning or probi ng for food beneath the surface of the substrate either in or near shallow water (fro m Ehrlich et al. 1988). Therefore, one of the most important management suggestions for Moist-soil Foragers is the management of ponds or sections of ponds with minimal shorelin e vegetation. To increase use of certain golf course ponds or sections of ponds by foraging Moist-soil Foragers, herbaceous vegetation could be planted farther away (>1 m) from the waterÂ’s edge to maximize the available moist-soil foraging habitat for these species. Foraging and resting Moist-soil
111 Foragers may also benefit from reduced vegetation in areas adjacent to the ponds because these species most often used ponds that were adjacent to areas devoid of vegetation. Aerial Foragers Aerial Foragers were most often obser ved using a Â‘high divingÂ’ technique to forage, whereby the bird drops fr om a height into the water to catch prey (from Ehrlich et al. 1988). When not actively foraging, these bird s usually require some type of perch to consume prey or rest. Aerial Foragers we re observed using both man-made structures (e.g., out-of-bounds markers) and natural ve getation (e.g., shrubs and trees) along the shorelines as perches. One suggestion for th ese birds would be to increase the number of perches along the shoreline of ponds managed for these species. These birds also seem to prefer larger ponds, indicating that perches (either construc ted or natural) may benefit more Aerial Foragers when placed around ponds w ith either larger perimeters or surface area.
112 LIST OF REFERENCES Adams, L. W., L. E. Dove, and T. M. Fr anklin. 1985. Use of urban stormwater control impoundments by wetland birds. Wilson Bulletin 97:120-122. Baker, M. C. 1979. Morphological correlates of habitat selection in a community of shorebirds (Charadriifo rmes). Oikos 33:121-126. Bancroft, G. T. 1989. Status and conserva tion of wading birds in the Everglades. American Birds 43:1258-1265. Bancroft, G. T., A. M. Strong, R. J. Sa wicki, W. Hoffman, and S. D. Jewell. 1994. Relationships among wading bird fora ging patterns, colony locations, and hydrology in the Everglades. In S. M. Davis, and J. C. Ogden (Eds.) Everglades: the ecosystem and its restoration: 615-657. St. Lucie Press, Delray Beach, Florida. Bent, A. C. 1926. Life histories of North Am erican marsh birds. U.S. National Museum Bulletin 135. Boman, B., C. Wilson, M. Jennings, and S. S kula. 2002. Detention/retention for citrus stormwater management. Florida Cooperati ve Extension Service, Institute of Food and Agricultural Science. Univer sity of Florida, Gainesville. Breininger, D. R., and R. B. Smith. 1990. Wa terbird use of coastal impoundments and management implications in eastcentral Florida. Wetlands 10:223-241. Brown, M., and J. J. Dinsmore. 1986. Implica tions of marsh size and isolation for marsh bird management. Journal of Wildlife Management 50:392-397. Brown, M. T. 1991. Evaluation of created we tlands through comparisons with natural wetlands. U.S. Environmental Protection Agency, Environmental Research Laboratory. Corvallis, Oregon. Brusati, E. D., P. J. DuBowy, and T. E. Lacher, Jr. 2001. Comparing ecological functions of natural and created wetla nds for shorebirds in Texas. Waterbirds 24:371-380. Burger, J., L. Niles, and K. E. Clark. 1997. Importance of beach, mudflat, and marsh habitats to migrant shorebirds on De laware Bay. Biological Conservation 79:283292.
113 Caldwell, G. S. 1981. Attraction to tropical mixed-species heron flocks: proximate mechanism and consequences. Behavior al Ecology and Sociobiology 8:99-103. Celada, C., and G. Bogliani. 1993. Breeding bi rd communities in fragmented wetlands. Bolettino Zoologia 60:73-80. Chick, J. H., and C. C. Mclvor. 1994. Fi sh distribution among beds of different macrophyte types: viewing a littoral zone la ndscape. Canadian Journal of Fish and Aquatic Sciences 51:2873-2882. Cody, M. L. 1985. Habitat selection in birds. Academic Press, Orlando, Florida. Colwell, M. A., and L. W. Oring. 1988. Habita t use by breeding and migrating shorebirds in southcentral Saskatchew an. Wilson Bulletin 100:554-566. Custer, T. W., and R. G. Osborn. 1977. Wadi ng birds as biological indicators: 1975 colony survey. Department of the Interior , Fish and Wildlife Service. Washington D.C. Fish and Wildlife Special Report 206. Daubenmire, R. F. 1959. A canopy-coverage me thod of vegetational analysis. Northwest Scientist 33:43-64. De Graaf, R. M., Tilghman, N.G., and S. T. Anderson. 1985. Foraging guilds of North American birds. Ecological Management 9:493-536. Donovan, T. M., P. W. Jones, E. M. Annand, and F. R. Thompson. 1997. Variation in local-scale edge effects: mechanis ms and landscape context. Ecology 78:20642075. Edelson, N. A., and M. W. Collopy. 1990. Fo raging ecology of wading birds using an altered landscape in central Florida. Fl orida Institute of Phosphate Research. Bartow, Florida. Ehrlich, P. R., D. S. Dobkin, and D. Wheye 1988. The Birder's Handbook. Simon and Schuster Inc., New York. Enright, C. A. 1971. A review of resear ch on type C botulism among waterbirds. Colorado Cooperative Wildlife Research Unit. Colorado State University. Fort Collins. Erwin, R. M., J. S. Hatfield, M. A. Howe , and S. S. Klugman. 1994. Waterbird use of saltmarsh ponds created for open marsh water management. Journal of Wildlife Management 58:516-524. Esler, D. 1992. Habitat use by piscivorous birds on a power plant cooling reservoir. Journal of Field Ornithology 63:241-392.
114 ESRI. 1999a. ARC/INFO. Environmental Syst ems Research Institute, Redlands, California. ESRI. 1999b. ARC/VIEW. Environmental Syst ems Research Institute, Redlands, California. Fasola, M. 1986. Resource use of foraging he rons on agricultural and nonagricultural habitats in Italy. Colonial Waterbirds 9:139-148. Fasola, M., and F. Barbieri. 1978. Factors a ffecting the distribution of heronries in northern Italy. Ibis 120:537-540. Fasola, M., and X. Ruiz. 1996. The value of rice fields as substitutes for natural wetlands for waterbirds in the Mediterranean region. Colonial Waterbirds (Special Publication) 19:108-114. Flather, C. H., and J. R. Sauer. 1996. Usi ng landscape ecology to test hypotheses about large-scale abundance patterns in migratory birds. Ecology 77:28-35. Forsberg, C., and S. O. Ryding. 1980. Eutr ophication parameters and trophic state indexes in 30 Swedish waste-receiving lakes. Archiv fur Hydrobiologie 89:189207. Frayer, W. E., T. J. Monahan, D. C. Bowde n, and F. A. Graybill. 1983. Status and trends of wetlands and deepwater habitats in the conterminous United States, 1950s to 1970s. Department of Forest and Wood Sc iences. Colorado State University, Ft. Collins. Frederick, P. C., and M. W. Collopy. 1988. Reproductive ecology of wading birds in relation to water conditions in the Flor ida Everglades. Florida Cooperative Fish and Wildlife Research Unit, School for Fo restry and Research and Conservation. University of Florida, Technical Report. 30. Frederick, P. C., and S. M. McGehee. 1994. Wading birds use of waterwater treatment wetlands in central Florida, USA. Colonial Waterbirds 17:50-59. Frederick, P. C., and D. Siegel-Causey. 2000. Anhinga ( Anhinga anhinga ) In The Birds of North America, No. 522. (A. Poole, a nd F. Gill, Eds) The Birds of North America, Inc., Philadelphia. Frederick, P. C., and M. G. Spalding. 1994. Factors affecting reproductive success of wading birds (Ciconiiformes) in the Everglades In M. Davis, and J. C. Ogden (Eds). The Everglades: The Ecosystem and its restoration: 50-59. St. Lucie Press, West Palm Beach, Florida.
115 Frederickson, L. H., and T. S. Tayl or. 1982. Management of seasonally flooded impoundments for wildlife. United States De partment of the Interior, Fish and Wildlife Service. Washington, D.C. Gawlik, D. E. 2002. The effects of prey ava ilability on the numeri cal response of wading birds. Ecological Monographs 73:329-346. Gibbs, J. P., J. R. Longcore, D. G. McAu ley, and J. K. Ringelman. 1991. Use of wetland habitats by selected nongame water bird s in Maine. United States Fish and Wildlife Service. Washington, D.C. Gibbs, J. P., S. Woodward, M. L. Hunter, and A. E. Hutchinson. 1987. Determinants of great blue heron col ony distribution in coastal Maine. Auk 104:38-47. Goss-Custard, J. D. 1976. Variation in the dispersion of redshank, Tringa totanus , on their winter feeding grounds. Ibis 118:257-263. Goss-Custard, J. D. 1977. The Ecology of the Wash: III. Density-rel ated behaviour and the possible effects of a loss of feedi ng grounds on wading birds (Charardrii). Journal of Applied Ecology 14:721-739. Gosser, A. L., M. R. Conover, and T. A. Messmer. 1997. Managing problems caused by urban Canada geese. Berryman Institut e Publication 13. Utah State University, Logan. Greij, E. D. 1994. Common Moorhen. Pages 145-157 In T. C. Tacha, and C. E. Braun, (Eds.) Migratory Shore and Upland Ga me and Bird Management in North America: 145-157. International Associat ion of Fish and Wildlife Agencies. Lawrence, Kansas. Hafner, H. 2000. Heron Ne st Site Conservation. In J. A. Kushlan and. H. Hafner (Eds.) Heron Conservation: 201-217. Academic Press, San Diego, California. Hafner, H., and R. Britton. 1983. Changes of foraging sites by nes ting Little Egrets Egretta garzetta L . in relation to food supply. Co lonial Waterbirds 6:24-30. Hafner, H., P. J. Dugan, and V. Boy. 1986. Us e of artificial and natural wetlands as feeding sites by little egrets ( Egretta garzetta L.) in the Camargue southern France. Colonial Waterbirds 9:149-154. Haines, T. A., S. J. Pauwels, and C. H. Jagoe. 1986. Predicting and evaluating the effects of acidic precipitation on water chemistr y and endemic fish populations in the northeastern United States. United States Fish and Wildlife Service. Biological Report 80.40.23.
116 Henderson, E. G. 1981. Behavioral ecology of th e searching behavior of the white ibis ( Eudocimus albus ).M.S. thesis, University of South Carolina, Columbia. Hoyer, M. V. and D. E. Canfield, Jr. 1990. Limnological factors influencing bird abundance and species richness on Flor ida lakes. Lake and Reservoir Management 6:133-141. Hoyer, M. V., and D. E. Canfield, Jr . 1994. Bird abundance and species richness on Florida lakes: influence of trophic status, lake morphology, and aquatic macrophytes. Hydrobiologia 297:107-119. Hutchinson, G. E. 1959. Homage to Santa Rosa lia, or why are there so many kinds of animals. American Naturalist 93:137-145. Jelks, H. L. 1991. The distribution of foragi ng wading birds in relation to the physical and biological characteristics of freshwater wetlands in southwest Florida. M.S. thesis, University of Florida, Gainesville. Jenni, D. A. 1969. A study of the ecology of fo ur species of herons during the breeding season at Lake Alice, Alachua County, Florida. Ecological Monographs 39:245270. Kroodsma, D. E. 1978. Habitat valu es for nongame wetland birds. In P. E. Greeson, J. R. Clark, and J. E. Clark (Eds.) Wetland f unctions and values: the state of our understanding: 320-329. American Wate r Resources Association Technical Publication TPS79-2. Kushlan, J. A. 1978. Feeding ecology of wading birds. In A. Sprunt, IV, J. C. Ogden, and S. Winckler (Eds.) Wading Birds: 249-297. National Audubon Society. Research Report No. 7. New York, New York. Kushlan, J. A. 2000a. Heron feeding habitat conservation. In J. A. Kushlan, and H. Hafner (Eds.) Heron Conservation: 219-235. Academic Press, San Diego, California. Kushlan, J. A. 2000b. Research and info rmation needs for heron conservation. In J. A. Kushlan, and H. Hafner (Eds.) Heron Conservation: 331-342. Academic Press, San Diego, California. LAKEWATCH. 2001. Florida LAKEWATCH/ Pr oject COAST Annual Data Summaries for the year 2000. Department of Fisherie s and Aquatic Sciences, Institute of Food and Agricultural Sciences Library. Un iversity of Florida, Gainesville. Lichstein, J. W., T. R. Simons, and K. E. Franzreb. 2002. Landscape effects on breeding songbird abundance in managed forest s. Ecological Applications 12:836-857.
117 MacArthur, R. 1970. Species packing and co mpetitive equilibrium for many species. Theoretical Population Biology 1:1-11. MacArthur, R. H., and E. O. Wilson 1967. Th e Theory of Island Biogeography. Princeton University Press, Princeton, New Jersey. Maxwell, G. R., and H. W. Kale. 1977. Breedin g biology of five species of herons in coastal Florida. Auk 94:689-700. McGarigal, K., and W. C. McComb. 1995. Relationships be tween landscape structure and breeding birds in th e Oregon coast range. Ecological Monographs 65:235260. Melvin, S. L., and J. W. Webb, Jr. 1998. Differe nces in the avian communities of natural and created spartina alterniflora salt marshes. Wetlands 18:59-69. Meyerriecks, A. J. 1962. Diversity typifi es heron feeding. Na tural History 71:48-59. Mitsch, W. J., and J. G. Gosselink 1993. Wetlands. John Wiley and Sons, Inc., New York. Muller, K. L., J. A. Stamps, V. V. Kris hnan, and N. H. Willits. 1997. The effects of conspecific attraction and ha bitat quality on habitat selection in territorial birds ( Troglodytes Aedon ). American Naturalist 150:650-661. New York State Department of Environmen tal Conservation. 1999. When geese become a problem. Division of Fish, Wildlife, a nd Marine Resources and United States Department of Agriculture Animal a nd Plant Health Inspection Service. Castleton, New York. Nilsson, S. G., and I. N. Nilsson. 1978. Breed ing bird community densities and species richness in lake s. Oikos 31:214-221. Nudds, T. D. 1992. Patterns in breeding waterfowl communities. In B. D. J. Batt, A. D. Afton, M. G. Anderson, C. D. Ankney, D. H. Johnson, J. A. Kadlec, and G. L. Krapu (Eds.). Ecology and Manageme nt of Breeding Waterfowl: 540-567. University of Minnesota Pre ss, Minneapolis, Minnesota. Owen, D. F. 1960. The nesting success of the heron Ardea cinera in relation to the availability of food. Proceedings of the Zoological Society of London 133:597617. Pineau, O. 2000. Conservation of wi ntering and migratory habitats. In J. A. Kushlan, and H. Hafner (Eds.). Heron Conservati on: 237-250. Academic Press, San Diego, California.
118 Powell, G. V. N. 1983. Food availability and reproduction by Great White Herons, Ardea herodias : a food addition study. Colo nial Waterbirds 6:139-147. Powell, G. V. N. 1987. Habitat use by wading bird s in a subtropical estuary: Implications of hydrography. Auk 104:740-749. Poysa, J. 1983. Resource utilization patte rn and guild structure in a waterfowl community. Oikos 10:295-307. Raven, P. H., R. F. Evert, and S. E. Eic hhorn 1992. Biology of Plants. Worth Publishers, Inc., New York. Rey, J. R., T. Kain, and R. Stahl. 1991. We tland impoundment of eas t-central Florida. Florida Scientist 54:33-40. Roberts, T. H. 1991. Habitat value of manmade coastal marshed in Florida. United States Army Waterways Experiment Station. Vicksburg, Mississippi. Rodgers, J. A., Jr., and H. T. Smith. 1995. Setback distances to protect nesting bird colonies from human disturbanc e. Conservation Biology 9:89-99. Rodgers, J. A., Jr., A. S. Wenner, and S. T. Schwikert. 1987. Population dynamics of Wood Storks in North and Central Florid a, USA. Colonial Waterbirds 10:151156. Rusch, D. H., C. D. Ankney, J. R. Boyd, J. R. Longcore, F. Montalbano, III, J. K. Ringelman, and V. D. Stotts. 1989. P opulation ecology and harvest of the American black duck: a revi ew. Wilson Bulletin 17:379-406. Sacco, J. N., E. D. Seneca, and T. R. Wentworth. 1994. Infaunal community development of artificially established salt marshes in North Carolina. Estuaries 17:489-500. Samuel, M. D., J. Y. Takekawa, G. Samelius, and D. R. Goldberg. 1999. Avian cholera mortality in lesser snow geese nestin g on Banks Island, Northwest Territories. Wildlife Society Bulletin 27:780-787. SAS Institute. 2001. SAS. Statistical Analysis System Institute, Cary, North Carolina. Schmiegelow, F. K. A., C. S. Machtans, and S. J. Hannon. 1997. Are boreal birds resilient to forest fragmentation? An e xperimental study of short-term community responses. Ecology 78:1914-1932. Senner, S. E., and M. A. Howe. 1984. Conservation of nearctic shorebirds. In J. Burger, and B. Olla (Eds.). Shorebirds: br eeding behavior and populations: 379-421. Plenum Press, New York.
119 Sibley, D. A. 2000. The Sibley Guide to Birds. Alfred A. Knopf, Inc., New York. Simenstad, C. A., and R. M. Thom. 1996. Func tional equivalency trajectories of the restored Gog-Le-Hi-Te estuarine we tland. Ecological Applications 6:38-56. Skagen, S. K., and F. L. Knopf. 1994. Migra ting shorebirds and habitat dynamics at a prairie wetland complex. Wilson Bulletin 106:91-105. Smith, J. P. 1994. The reproductive and foraging ecology of wading birds (Ciconiiformes) at Lake Okeechobee, Flor ida. Ph.D. dissertat ion. University of Florida, Gainesville. Smith, J. P. 1995. Foraging flights and habitat use of nesting wading birds (Ciconiiformes) at Lake Okeechobee, Fl orida. Colonial Wa terbirds 18:139-158. Smith, J. P., J. R. Richardson, and M. W. Collopy. 1995. Foraging habitat selection among wading birds (Ciconiiformes) at Lake Okeechobee, Florida, in relation to hydrology and vegetative cover. Archives of Hydrobiology (Special Issues Advanced Limnology) 45:247-285. Spalding, M. G., G. T. Bancroft, and D. J. Forrester. 1993. The epizootiology of Eustrongylidosis in wading birds (Ciconiif ormes). Journal of Wildlife Diseases 29:237-249. Spalding, M. G., and D. J. Forrester. 1993. Pathogenesis of Eustrongylides ignotus (Nematoda: Dioctophymatoidea) in Cicon iiforms. Journal of Wildlife Diseases 29:250-260. Stoner, A. W. 1983. Distribution of fishes in seagrass meadows: role of macrophyte biomass and species compositi on. Fishery Bulletin 81:837-846. Surdick, J. A., Jr. 1998. Biotic and abiotic i ndicators of foraging site selection and foraging success of four ci coniiform species in the fr eshwater Everglades of Florida. M.S. thesis. University of Florida, Gainesville. Tiner, R. M., Jr. 1984. Wetlands of the United States: Current status and recent trends. U.S. Department of the Interior Fish and Wildlife Service.Washington, D.C. Warnock, S. E., and J. Y. Takekawa. 1995. Habita t preferences of wintering shorebirds in a temporally changing environment: wester n sandpipers in the San Francisco Bay estuary. Auk 112:920-930. Weller, M. W. 1999. Wetland birds: Habitat resources and conser vation implications. University Press, Cambridge.
120 Williamson, M. 1988. Relationship of species number to area, distance, and other variables. In A. A. Myers, and P. S. Giller (Eds.) Analytical Biogeography: 91115. Chapman and Hall, New York. Williams-Whitmer, L. M., M. C. Brittingham, and M. J. Casalena. 1996. Geese, ducks, and swans. Wildlife Damage Control 6. College of Agricultural Sciences Cooperative Extension, Penn State University, College Station. Wright, D. H. 1983. Species-energy theory: an extension of species-area theory. Oikos 41:496-506. Zinn, S., and R. Redles. 2002. Keeping lakes and wetlands healthy: Aquagenix succeeds with service. Turf South Magazine: Jan. 2002
121 BIOGRAPHICAL SKETCH C. LeAnn White was born in 1977 in Williamsburg VA. She attended the College of William and Mary from 1995 to 1998 earni ng a bachelorÂ’s degree in biology. A lifelong curiosity about nature and the desire to answ er questions pertaining to it encouraged her to pursue a research-oriented degree. She entered the graduate program at the University of Florida in the Fall of 1999 and earned a Master of Science degree from the Department of Wildlife Ec ology and Conservation in August of 2003.