HABITAT SELECTION OF REINTRODUCED MIGRATORY WHOOPING CRANES (Grus americana) ON THEIR WINTERING RANGE. Lara E. A. Fondow

Transcription

1 i HABITAT SELECTION OF REINTRODUCED MIGRATORY WHOOPING CRANES (Grus americana) ON THEIR WINTERING RANGE by Lara E. A. Fondow A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science (Wildlife Ecology) at the University of Wisconsin-Madison 2012

2 ii APPROVED Stanley A. Temple Beers-Bascom Professor Emeritus in Conservation January 16, 2013

3 iii DEDICATION My incredible family deserves a vast amount of credit for the completion of this master s thesis. Without their love and support, this endeavor would not have been possible. I am especially indebted to my daughter, Clara, whose shining face was my greatest source of motivation.

4 i Table of Contents Preface. iii Project Background.....iv Abstract..viii Introduction...1 Methods Study Area....2 Monitoring and Data Collection....5 Habitat Selection vs. Availability.. 8 Behavior within Cover Types Philopatry Site Tenacity...15 Habitat Imprinting..16 Social Facilitation and Territoriality.. 17 Environmental Stochasticity..20 Innate Preferences...21 Results. 21 Habitat Selection vs. Availability...21 Behavior within Cover Types Philopatry...30 Site Tenacity 33 Habitat Imprinting...34

5 Social Facilitation and Territoriality 40 ii Environmental Stochasticity. 44 Innate Preferences. 46 Discussion.48 Acknowledgements...58 Literature Citations Figures and Tables Appendix A..108 Appendix B Appendix C

6 iii PREFACE Hovering at approximately 500 birds in the wild and captivity, the Whooping Crane (Grus americana) is one of the most severely endangered birds in North America. Once more widespread, only one wild population remains in existence. The breeding grounds of this population are located within Wood-Buffalo National Park in the Northwest Territories of Canada, and the wintering grounds are on Aransas National Wildlife Refuge and the surrounding saltmarshes of the Texas Gulf Coast. This population is referred to as the Aransas-Wood Buffalo Population (AWBP). The Whooping Crane Recovery Plan (Canadian Wildlife Service (CWS) and U.S. Fish and Wildlife Service (USFWS) 2007) identifies the reintroduction and establishment of additional, self-sustaining populations of Whooping Cranes within the species historical range but geographically separate from the AWBP as a key element in the recovery of the species. These additional populations help to ensure the survival of the species in the event of a stochastic catastrophe extirpating the remaining wild population. Continuing research to identify appropriate reintroduction sites and improve reintroductions techniques and protecting and managing the habitat of reintroduced populations were two necessary actions identified within the recovery plan (CWS and USFWS 2007). Previous attempts to reestablish populations of Whooping Cranes in portions of their historical range are well documented (Maguire 2008, USFWS 2011). This thesis focuses on the eastern migratory Whooping Crane reintroduction, an effort undertaken by a consortium of private and public entities that together form the Whooping Crane Eastern Partnership (WCEP).

7 PROJECT BACKGROUND iv Efforts to establish a migratory population of Whooping Cranes within an eastern migratory flyway began in 2000, after an unorthodox reintroduction technique had been tested on Sandhill Cranes (Grus canadensis) (Urbanek et al. 2005, Maguire 2008). Eggs produced in a handful of Whooping Crane captive breeding centers are hatched at the U.S. Geological Survey (USGS) Patuxent Wildlife Research Center in Laurel, Maryland. There, the crane colts begin training to follow ultra-light aircraft, and handlers maintain strict protocols to avoid imprinting on humans. When the colts reach approximately 40 days of age, they are transported to Necedah National Wildlife Refuge, in Central Wisconsin, where they continue their extensive training to fly behind ultra-lights and are housed in top-netted pens that encompass both shallow marshy areas and open, grassy uplands. Necedah NWR is approximately 7,725 ha of shallow marshland and managed impoundments, 695 ha of shrubscrub, 686 ha of grassland and 8,530 ha of forested uplands (Urbanek et al. 2005). Kelly Maguire (2008) described the percent composition of the habitat surrounding the training pen sites and visible to the young cranes: The viewscape from the first two pens used in 2001 encompassed an area of 23.7 ha with an open vista of wetland vegetation (16% of viewscape), fields (28% of viewscape), upland forest (7%), shrubland (2%), wetland forest (3%) and a distant backdrop of open water (44% of viewscape). The third site, added in 2002, was surrounded primarily by a viewscape of emergent wetland vegetation and wet meadow (70% of viewscape), open water (2% of viewscape), fields (2% of viewscape), upland forest (2% of viewscape) with a distant backdrop of wetland forest (24% of viewscape).

8 A majority of the Whooping Cranes in this population have returned to Necedah v NWR and its vicinity to establish summer home ranges. The composition of the habitat within self-selected home ranges of these birds has not differed significantly from that of the pen sites and surrounding areas (Maguire 2008). In early autumn, once they are able to follow the ultra-light for significant distances, the colts and ultra-lights are lead on a pre-determined southward journey. While on migration, the colts are housed in top-netted traveling pens and provided food and fresh water. Though some of the sides are covered to hide human activities near the pen, the birds have an open view of the surrounding area through the other sides. While a few of the migration stops incorporate wetlands within the holding pen, the majority is in open upland sites (M. Nipper, Operation Migration, personal communication). In approximately two months, they reach their destination, the Chassahowitzka National Wildlife Refuge, on the Central Gulf Coast of Florida where they remain in or near an open-topped, predator-proof pen encompassing a tidal pool within typical salt-marsh plant communities. The soft release of these birds differs from the technique used for the Mississippi Sandhill Cranes and nonmigratory Florida Whooping Cranes in that these birds are never fitted with wing brails (Ellis et al. 1992, Nesbitt et al. 1997) to inhibit their movement out of the pen area. Supplemental food, fresh water, and interactions with costumed handlers, however, hold the birds in the vicinity of the pen site until they depart on their own initiative, despite the continual provision of food and water at the pen site (Urbanek et al. 2010). Dominant vegetation in the surrounding area includes extensive, dense monotypic stands of black needlerush (Juncus roemerianus) with scattered cabbage palm (Sabal palmetto) hammocks (Urbanek et al. 2010). Open water areas consist of tidal pools, bays, and creeks. In the first year of the

9 reintroduction, the soft-release pen was 0.6 ha in size, and the birds often chose to roost vi outside of the pen when not actively encouraged to stay in the pen by costumed handlers. Bobcats killed two birds during that winter before the inception of an active management protocol that included ushering birds into the pen before dark. Prior to the second year of releases, the pen was expanded to 1.6 ha to include a deeper pool and an area of salt grass (Distichlis spicata) that had been a favorite loafing area for the hatch-year 2001 cohort. An existing oyster bar along the edge of the larger pool was expanded to create a gently sloping area that would allow the birds to roost in their preferred water depth at varying tidal water levels. This oyster bar also gave the birds a smooth, solid surface on which to stand, that contrasted with the jagged oyster bars or soft, mucky soil substrates typical of Florida s coastal saltmarsh. The dense needlerush in and around the pen was also actively managed through prescribed burning and mechanical flattening to create more open habitat for Whooping Cranes (Urbanek et al. 2010). The saltmarshes of Chassahowitzka NWR and the surrounding central Gulf Coast region were chosen as the eastern migratory Whooping Crane wintering site due to both biological and management considerations. First, project planners wanted an area with characteristics that matched Aransas NWR as closely as possible. The primary winter food source for the AWBP is blue crabs, which are also abundant along Florida s central Gulf Coast. Other similarities between the saltmarshes of Chassahowitzka NWR and Aransas NWR, however, are few. The vegetative structure, stability of water levels, salinity, and surface substrate of the two sites vary widely. In contrast to the dense stands of needle rush and mucky tidal creeks and bays of Chassahowitzka NWR, the saltmarshes of Aransas are characterized by vegetated flats dominated by glasswort (Salicornia virginiana), saltwort

10 (Batis maritima), sea-oxeye daisy (Borrichia frutescens), wolfberry (Lycium carolinianum), saltgrass (Distichlis spicata), and smooth cordgrass (Spartina alterniflora), and wind tidal flats dominated by mudflat grass (Elocharis parvula), saltgrass, and cordgrasses. Interspersed among vegetated areas are bodies of open water of various sizes. (Chavez- Ramirez and Slack 1995). The shorter stature, less densely growing plants of Aransas NWR allow Whooping Cranes to forage, walk, and loaf in these areas without view or movement obstruction. The juvenile Whooping Cranes at Chassahowitzka NWR do not tend to walk through the dense needlerush unless led to do so by a costumed caretaker (Urbanek et al. 2010). Another essential characteristic for a wintering site was isolation from humans. The refuge closed a large area around the pen site to boating and crabbing activities to meet this criterion, which was somewhat difficult to achieve in the heavily populated state of Florida. A non-migratory Whooping Crane reintroduction had been on-going in Florida since 1993, and project planners wanted to assure that the new, migratory population would not interfere with the non-migratory population and vice-versa. Establishing the migratory population on the coastal saltmarsh aimed to meet this objective. Finally, the coastal location of the pen site helped to assure that the impressionable juvenile birds would not be influenced by nonmigratory Sandhill Cranes, which often exhibit tameness and inhabit semi-urban and residential areas that can be dangerous for cranes (Urbanek et al. 2010). While many older, returning birds land at the Chassahowitzka pen site on their initial independent southward migration, they have not remained there in the absence of juvenile birds and the supplemental food and fresh water, suggesting that this area does not provide vii

11 suitable Whooping Crane habitat. Because they are evidently unable to meet their needs, they reject the saltmarsh and depart in search of more suitable areas. Beginning in the winter of a separate, inland pen was built at Southwest Florida Water Management District s Hálpata Tastanaki Preserve, located near Dunnellon in southwest Marion County to address problems created by older birds returning to the Chassahowitzka NWR pen site (Urbanek at al. 2010). An additional, supplemental release technique, Direct Autumn Release, was begun in These birds, also bred in captivity, spend the early part of their lives at an isolation rearing facility at the International Crane Foundation, in Baraboo, Wisconsin. They are then transported to Necedah NWR, where they are raised by costumed handlers and exposed to open upland and shallow marsh habitats and held in a pen at night. Wild Sandhill Cranes and Whooping Cranes frequent the site and often interact with the colts. In early autumn, the birds are released, one by one or in small groups, into wild crane flocks containing Whooping Cranes. They typically join these wild birds on the southward migration. viii ABSTRACT An effort to establish a migratory population of Whooping Cranes (Grus Americana) in the eastern United States has been ongoing since From the winter through winter , I examined the influences of philopatry (return of an animal to near its initial winter home range), site tenacity (return to a previously utilized home range), habitat imprinting (the selection of a home range with a similar habitat composition to that of the initial home range), social facilitation and territoriality (the attraction to or avoidance of conspecifics or other crane species), environmental stochasticity (the presence or occurrence

12 of unusual environmental factors leading to deviations from expected patterns of site ix selection), and innate preferences (genetically based influences) on the selection of winter home ranges by Whooping Cranes. I found evidence of a philopatric tendency in these birds based on their return rate (60.3% overall and 83.3% when stochastic environmental influences and different management regimes were excluded) to the Chassahowitzka softrelease site and surrounding saltmarshes. This philopatric tendency was significantly stronger in male birds. However, after a brief visit to the site, all returning cranes ultimately abandoned the saltmarsh, seeking more favorable habitat elsewhere. Site tenacity to previous self-selected winter home ranges was high (73%), when stochastic and social influences were excluded. The presence of other cranes (Whooping and Sandhill) within the self-selected home ranges of Whooping Cranes in this population suggested a strong social facilitation in site selection. I found little evidence of territory establishment or defense in this population of Whooping Cranes. In several cases, home range selection was strongly influenced by a social association, with birds being led by new social associates to that associate s previous self-selected home range. Stochastic environmental events and conditions influenced home range selection in a number of individually examined instances. The most prevalent example of the influence of stochastic factors is the phenomenon of short-stopping to establish a winter home range north of Florida along the migration route. I did not find evidence for home range selection being driven by innate preference alone, but found that the birds innate preferences for open landscapes with wetlands likely played a role in their selection of winter home ranges in inland Florida. I assessed the habitat preferences of the birds at various selection scales (Johnson, 1980). During the winters of and , I recorded the locations, habitat use

13 across ten cover types, social associations, and behaviors of all migratory Whooping Cranes in known locations in Florida using a semi-regular sampling plan. Habitat use was nonrandom at all selection scales for both daytime home ranges and roosting sites. At all selection scales, open pastures and freshwater marshes were preferred over other cover types, and wooded upland, urban development, and saltmarsh were the most selected-against cover types. Whooping Cranes behaviors within the different habitat types suggested that the some habitat types complement each other for various activities while they supplement each other for particular behaviors. The results of this study have important implications for the management of the reintroduced population of Whooping Cranes. In spite of efforts to encourage these birds to use saltmarsh as habitat they abandoned the coastal areas and self-selected more inland areas. Because of both short-term and long-term changes in habitat availability in Florida, it will be important to continue to monitor habitat selection by this population of birds to determine whether the patterns found in this study will persist and how changes in management practices affect future selection patterns and survival in the wild. There is already evidence that the cranes are utilizing areas further to the north than expected because they presumably provide habitat on a more consistent basis (R. Urbanek, USFWS, personal communication). x

14 1 INTRODUCTION Knowing how reintroduced migratory Whooping Cranes select wintering home ranges and habitat is a key to understanding how to manage a self-sustaining population. This study had four primary objectives. First was to determine the winter habitat preferences of birds in this population at three spatial scales. Next, I wished to determine how home range components were being utilized. Specifically, I wished to determine whether utilized habitats were complementary or supplementary for various behaviors, as suggested by Dunning et al. (1992). Complementation is said to occur when different critical resources are found in patches of different cover types. For example, animals may use one cover type for foraging and a second for roosting or resting behaviors. Supplementation is said to occur when all cover types provide similar but insufficient resources, forcing individuals to use many cover types to obtain sufficient resources (e.g., Graham 2001). Third, I wished to examine the relative influences of underlying factors that often affect habitat selection in wild animal populations. These factors included philopatry, site tenacity, habitat imprinting, social facilitation, environmental stochasticity, and innate or genetically determined preferences. Philopatry is defined as the return to the general vicinity of the natal or initial home range (Cody 1985). In this study, the initial winter home range on Florida s central Gulf Coast was not selected by the birds but was imposed on them by managers. Site tenacity refers to the return to a previously occupied home range (Hildén 1965). In this study, I examined site tenacity in terms of a tendency to return to a previously self-selected winter home range, not the one imposed by managers. Habitat selection in birds is often influenced by other birds in the population (Cody 1985). I wished to examine whether the choices made

15 by young birds in this population were influenced by older social associates and by the 2 distribution of Whooping Cranes and Sandhill Cranes on the Florida landscape. Finally, I wished to determine how the findings of this study may help to shape management decisions. METHODS Study Area The main 33,911-km 2 study area (Figure 1) encompassed all of the known wintering sites of eastern migratory Whooping Cranes in Florida (in the counties of Citrus, Hillsborough, Marion, Alachua, Lafayette, Highlands, Madison, Polk, Pasco, Osceola, Volusia, Taylor, Lake, Sumter, Hernando and Levy). Additional information was gathered on four birds that wintered in South Carolina in ; however, this information was not included in the initial data analysis due to an inability to follow the same sampling protocol or to gather complete data on these birds because of their geographic isolation from the remainder of the population. A simpler analysis was conducted for these birds using the limited available data. Figure 1 shows all known Florida wintering areas and the minimum convex polygon surrounding the wintering locations of birds soft released at Chassahowitzka NWR (outlying points represent locations of a direct autumn release bird). In Florida, where the detailed habitat selection analysis took place, Whooping Crane home ranges were located within the following five major land resource areas (MLRAs), as identified in Agriculture Handbook 296 (Natural Resources Conservation Service (NRCS) 2006): south central Florida ridge (73.8% of home ranges), southern Florida flatwoods (10.7%), southern coastal plain (9.5%), eastern Gulf Coast flatwoods (3.6%) and north central Florida ridge (2.4%). This classification system uses a combination of soil types,

16 climate, natural vegetation, physiography, and land uses to divide land into appropriate 3 categories in order to help guide decisions regarding natural resources and agriculture on the landscape. Figure 2 shows the overlap of known Whooping Crane home ranges through 2007 and the major land resource areas of FL. Because the majority of birds wintered within the south central Florida Ridge, a detailed description of the composition of this MLRA is included below. For descriptions of the additional MLRAs, see Appendix A. South Central Florida Ridge.--This MLRA supported the greatest percentage (73.8%) of Whooping Crane home ranges through the course of our study. The area, located entirely within Florida, is characterized by level to gently rolling topography, with an irregular land surface due to the many sinkholes that dot the area. The area supports many lakes but has few perennial streams. The Withlacoochee River flows through this MLRA. Geologically, the area is a young marine plain with Tertiary-age rocks below, including shale, mudstone, dolomite and limestone beds. Phosphate is mined from the limestone beds in the central part of the MLRA. Eastern migratory Whooping Cranes have temporarily utilized reclaimed mine areas but never established long-term home ranges in these areas. The area averages 46 to 56 inches of precipitation annually, with about 60 percent occurring from June through September in the form of moderate-intensity tropical storms. However, the second year of our study marked the beginning of a long-term drought that significantly impacted rivers and wetlands across much of Florida (Florida Department of Environmental Protection (DEP) 2008). In fact, due to some unique characteristics of Florida s water cycle, it is prone to periods of extreme drought or heavy rainfall, as shown in Figure 3, based on Florida DEP (2008).

17 Ultisols and Entisols are the dominant soil orders in this MLRA, and soils are 4 typically very deep, excessively drained to somewhat poorly drained, and loamy or sandy. Sand hill vegetation is supported in this area, with turkey oak, bluejack oak and longleaf pine making up the major species. Running oak, gopher apple, and bluestems and panic grasses are typical understory species (NRCS 2006). Livestock and citrus production are important to this area, with 23% of the area in private grasslands and 7% in private cropland. Conservation practices on private rangeland include brush management, pest management, prescribed burning, and prescribed grazing (NRCS 2006). These practices likely increase the areas attractiveness and safety for cranes by reducing predators as well as predator cover and improving grassland productivity. The uneven land surface of Florida s central ridge creates a diversity of marshes inhabiting different types of depressions, including former lake basins, shallow peat-filled valleys between extant lakes, and depressions landward of swamps that ring some lakes. The existence of many of these marshes arises from the shifting balance between the compaction of surface sediments, which retards surface water loss, and the periodic development of solution features that drain surface water into the Floridan aquifer (Kushlan 1990). Due to these shifting processes, marshes and lakes in this region are unstable, and a particular site may convert from lake to marsh to dry land within a short time span. The vegetation association of a given marsh is determined largely by its hydrologic regime, fire frequency, and soils. The three principal soil materials of Florida s marshes-- peat, marl, and sand are also greatly influenced by hydrologic regime. Large highland marsh complexes used by eastern migratory Whooping Cranes include the Peace Creek marshes near Lake Wales, the Lake Apopka marsh near Zellwood, the Clermont marsh

18 between Lake Minneola and Lake Minehaha, the Emeralda marsh near Lake Griffin, Eustis 5 Meadows near Eustis, and Paynes Prairie, near Gainesville. Monitoring and Data Collection Radio-telemetry and Individual Observations.--All of the birds in the eastern migratory Whooping Crane population were outfitted with conventional VHF transmitters, which were attached to plastic leg bands (Melvin et al. 1983). A few birds were also tagged with satellite transmitters. The birds movements, however, were primarily monitored using conventional radio telemetry, and the satellite data were utilized when a bird s location was unknown or was geographically isolated from the population s core distribution. Trackers primarily homed on radio signals by using truck-mounted 7-element yagi antennae. Handheld antennae were also used when it was necessary to walk into an area to observe a bird or to determine whether a bird was in flight or standing. On migration or during periods of frequent movement, and when searching for a bird whose location was unknown, birds were often radio tracked from airplanes with wing strut-mounted antennae. During the study years, once a bird or group of birds settled on a winter home range (defined as an area occupied for 10 or more consecutive days), it was sampled as regularly as possible during varying time periods. A stay of 10 or more days was selected as the home range establishment criterion because of the observation that once birds had settled into one location for more than one week, they tended to stay in that location for either the entire winter or until completing a major mid-winter location shift after occupying the area for a significant amount of time (See Tables 1 and 2 in Appendix C for all Florida and South Carolina winter bird locations and periods of occupation during the study). The deviation

19 from this pattern of long-term home range occupation was a single bird, number 6-01, which associated closely with Sandhill Cranes and more readily moved between major Sandhill Crane wintering areas. Only those areas that 6-01 occupied for 10 or more days were included in the study. Because many pairs or groups exhibited coordinated movements and were almost never located more than a few meters apart from each other, the pair or group, rather than the individual bird, was identified as the independent entity in the analysis of habitat selection versus availability when birds associated in groups. The more traditional approach is to sample individual marked animals in a given population and to treat each individual separately. In this study, though, every individual is marked, and group composition can be verified during each observation. Other studies have also analyzed habitat selection by pairs or groups of animals rather than individuals (Morrison and Humphrey 2001, Palminteri and Peres 2012, and Cartwright et al. 2012). Eleven bird units consisting of 20 individual birds were studied in year 1, and 17 bird units consisting of 34 individuals were included in the second year of the study. In the first study year, there were 3 birds that wintered in Florida apart from any other birds of this population, and there were 2 such birds in the second year of the study. All other birds were in groups of 2-5 individuals. The day was separated into four observation periods: AM (sunrise-3hr post sunrise), Mid-day (3hr post sunrise-3hr pre-sunset), PM (3hr pre sunset-sunset) and Roost (sunsetsunrise), due to recognized daily behavioral patterns in cranes (Tacha et al. 1987). To the greatest extent possible, trackers obtained an even distribution of locations within the various time periods for each group of birds. 6

20 In each year, trackers included me and one tracking assistant whom I trained in the 7 field observation and recording techniques. Typically, two or three tracking routes were established according to the geographic locations of the birds, and these routes were run every two to three days, with the starting and ending points rotating. Due to the short duration of time that visual observations can be made of a bird on roost, private property access complications at dawn and dusk, and the relative stability of roosting locations for cranes (Hoffman 2000), fewer roost locations were obtained than daytime locations. During each observation, the following information was recorded: observation start and stop time, weather data (% cloud cover, temperature, precipitation, wind speed and direction), observer name, property name and ownership, county, general primary and secondary cover type (see below), habitat comments, general behavior, associations with Sandhill Cranes and sub-species of Sandhill Cranes when identifiable, associations with other Whooping Cranes (migratory or non-migratory), and general comments. Presence and use of a supplementary food source (such as cattle or horse feed) were recorded. When a health issue was apparent, it was noted. To obtain Whooping Crane locations, the observer s location was determined with a Global Positioning System (GPS) device and recorded on the data sheet. The bearing and distance from the observer to the bird were obtained using an electronic bearing and range finder (KVH Datascope) or a compass and a visual distance estimate. Multiple comparisons of measurements on the ground and estimates obtained with the rangefinder indicated a similar level of error associated with each method. Whenever possible, additional visual landscape cues such as large trees, roads, buildings, or habitat edges, were recorded to aid in plotting a location on an aerial photograph in ArcView. The locations were determined using

21 a distance and azimuth tool in ArcView 3.2 or using the Editor tools in ArcGIS 9. When 8 visual locations were unattainable due to visual barriers, lack of site access due to permission or terrain, or lack of light, radio triangulations were performed. These triangulations were also plotted in ArcView, and the center of the triangles were calculated and used to estimate the birds locations. When a resulting triangle was too large to determine with confidence the cover type that the bird or birds were occupying, the location was discarded from the study. When cover types noted in the field differed widely from those obtained using GIS data (the landuse/landcover data used are described below), the field-recorded types were generally used, particularly when the birds were observed roosting in a marsh or wet prairie, yet the GIS layer identified the area as an upland cover type. Such discrepancies were often due to the fluctuating demarcation between wetlands and surrounding uplands resulting from variations in water depth and coverage (Kushlan 1990). Habitat Selection versus Availability Habitat Classification and Home Range Calculation. I used landuse/landcover layers developed by Florida s Water Management Districts (FWMDs) to classify and quantify utilized and available cover types. The WMDs categorized land use/cover features according to the hierarchical Florida Land Use and Cover Classification System (FLUCCS), developed by the Florida Department of Transportation (Florida Department of Transportation, 1999). The features were photo-interpreted at 1:12,000 using 2005 or earlier 1-ft color infrared (CIR) digital aerial photographs (Mapping and GIS Section, Southwest Florida Water Management District 2006). For each WMD, the most recent available layers were used. Original data layers were GIS shapefiles containing coded polygons, and were

22 typically available for the U.S. Geological Survey quarter-quadrant. For location and home range analysis, I intersected the points or polygons with these original shapefile layers. To assess habitat use and availability at larger scales, I merged the appropriate layers and then transformed the shapefile into a 10m x 10m raster. In order to perform a meaningful habitat selection analysis for Whooping Cranes, I simplified the hundreds of land use/cover classifications in the FLUCCS system into 10 general cover types: Open pasture and other rural open lands; Scrub/shrub wetlands and uplands; Wooded or forested uplands; Forested wetlands; Wet prairies; Freshwater marshes; Saltwater marshes, bays and estuaries; Open water; Urban/residential; and Miscellaneous. Hawth s Analysis Tools were used to construct 95% and 50% (core) kernel home ranges using all daytime locations collected at a given winter home range of a bird or group of birds. The same was done for all roost locations associated with each home range. Multiple trials were run to determine an appropriate smoothing factor (h) for these data, and 175 was selected because the resulting kernels most closely reflected field observations of the birds movements between and among use sites. Hawth s Analysis Tools output polylines, which had to be converted to polygons in order to determine habitat composition. An ET Geowizards tool was used to perform this transformation. The resulting polygons were intersected with the land use/cover layer (see Figure 4 for an example of such an intersection), and the area of each cover type contained within the polygon was calculated using ArcView s X Tools. The percent composition of each cover type was then determined, and this was used to perform the selection analysis. When a bird or group of birds occupied more than one home range in a particular year, all home ranges were included in the study, provided that adequate location data were collected. For each year, the mean and standard 9

23 deviation of the number of observations at each home range were calculated. All home 10 ranges with fewer observations than one standard deviation below the mean were discarded from the analysis because too few observations were recorded at these sites to create home ranges boundaries with a high level of confidence. I used compositional analysis (Aebischer et al. 1993) to: (1) compare cover types in individual 95% home ranges to cover types available in the birds search areas, defined here as the areas over which the birds searched for suitable home ranges and explained in detail below; (2) compare cover types in the birds core home ranges to cover types available in the 95% home ranges, and (3) determine whether cranes used cover types in proportion to their availability within individual core home ranges. In compositional analysis, the log-ratio transformation results in linearly independent data (Aebischer et al. 1993). The unit of replication was the study year. For each year, the log-ratios of the utilized habitat composition were compared to the log-ratios of available habitat composition. If habitat selection was random, the difference between these two ratios would be close to zero. If selection was found to be significantly non-random (p<0.05), I built a ranking matrix. To determine which cover types were used proportionally more than others, the mean and standard error were calculated for the difference (e.g., use minus availability) for each cover type pair (e.g., freshwater marsh vs. forested wetland). Using all the home ranges in a study year, I ranked habitats by adding the positive difference values. To determine whether the paired use-availability difference showed a significant (p<0.05) departure from random use, a t-statistic was calculated as the ratio of mean over standard error (Aebischer et al. 1993). Search Area Approach.--Rather than comparing the habitat composition within each bird s home range to that of the entire study area or some arbitrarily defined area, I used

24 the information on the areas actually explored by each bird to construct each individual 11 search area. The essential comparison I used to assess habitat selection at the landscape level was between the habitat composition of the bird s search area and the composition of its home range. Using Hawthorne s Analysis Tools, a minimum convex polygon was drawn around all of the known landing locations of each individual bird in Florida. This polygon was then buffered by 5km, to account for the typical wide circling of the birds before selecting a landing site (based on several personal observations). See Figure 5 for an example of such a search area. When birds returned directly to a previously used home range at the beginning of the season, the search area remained the same as for previous years. When data were insufficient to create a search area for a bird, that bird was not included in the landscape-level analysis. I did, however, compare the composition of the home ranges of such birds with the mean composition of the combined search areas for all birds with known search areas. This analysis closely approximates the typical landscape-level habitat selection approach, where individual home ranges are compared to a general study area (Aebischer et al. 1993, Graham 2001). Typically, researchers do not have the opportunity to gather complete movement information for all of their study animals, and so cannot perform individual analyses at the landscape level. Because of the intensive monitoring program associated with this reintroduction and the complete location dataset, I was able to improve the accuracy of the landscape-level analysis by identifying search areas. Behavior within Cover Types Behavioral data collected during visual observations of Whooping Cranes on their winter home ranges in Florida (eleven bird units consisting of 20 total birds in year one, and

25 17 bird units consisting of 34 individuals in year two) were pooled and analyzed to determine whether specific cover types were complementary (readily replacing each other as suitable habitat) or supplementary (utilized in conjunction with other habitat types) for given activities and whether particular cover types fulfilled specific life history needs. Behavioral categories included: roosting, foraging, resting, loafing (including preening and resting), and walking locomotion. For each behavioral category, the proportion of observations of this behavior within each cover type was calculated. These proportions were examined visually, using bar graphs, but no statistical test was used in this study to further evaluate these data. Hypotheses were as follows: H1: Wetland habitats, including wet prairies, open water, forested wetlands, and freshwater marshes, are preferred for roosting. H2: Open habitats, including open pasture, wet prairie, open water, and freshwater marshes are preferred for walking locomotion. H3: Foraging needs are met by a variety of cover types; however, open pasture is the most important habitat type in Florida for this activity. H4: Wetland habitats are preferred for loafing activities. I hypothesized that the omnivorous diet and energy needs of Whooping Cranes would require them to utilize various habitat types in foraging for food. I presumed that the open quality of pastures, along with the ease of locating invertebrate prey in closely grazed, nutrient-enriched pastures would make these areas the most attractive to Whooping Cranes for foraging, particularly when dotted with freshwater marshes, wet prairies, and other shallow water sources. I hypothesized that Whooping Cranes seeking refuge from the heat of the afternoon sun or taking a break between foraging bouts would seek wetlands in which to 12

26 rest and preen. These areas provide drinking water, a place to bathe, and cooling water in 13 which to stand, all in an open context that allows for vigilance against predation. The use of wetlands for loafing activity was not expected, however, to be exclusive. Cranes will often loaf in pastures, agricultural fields, and other upland areas (Sparling and Krapu 1994). Philopatry Philopatry is defined here as the tendency of a migratory bird to return each year to the close vicinity of its initial home range, in this case the winter home range it occupied during its first year. The influence of philopatry on winter site selection by Whooping Cranes was evaluated in two ways. I assessed the return of each Whooping Crane to the Chassahowitzka NWR softrelease pen during the second and subsequent years. This analysis included the 60 freeflying birds in the population that were led to the saltmarshes of Florida s central Gulf Coast and spent part of their initial winter at the Chassahowitzka NWR soft-release pen. A return to or near Chassahowitzka NWR was indicative of an intentional return to the bird s initial winter home range (in this case the soft-release pen where they were held for long enough to meet our criteria for a home range). A return to Chassahowitzka NWR presents fairly pure evidence of a philopatric influence on Whooping Crane movements, as the site does not provide habitat that is suitable for long-term occupation by Whooping Cranes in the absence of supplemental feed and intensive habitat management. A return rate greater than 50% was considered to indicate a philopatric influence, and t-tests were performed to determine whether the percent of birds returning to the soft-release pen within two years of release was significantly (p<0.05) greater than the percent of non-

27 returning birds. Because two of the release cohorts encountered notably different 14 environmental or management influences at the soft-release pen or following release, additional analyses, which omitted each and then both of these cohorts, were performed. I also compared the return rate of females to that of males, again taking varying management and environmental influences into account. On the breeding grounds, female cranes tend to disperse further from their natal home range than do males (Nesbitt et al. 2002, Maguire 2008). I wished to determine whether this pattern occurred on the wintering grounds as well. Finally, I assessed whether the location of the soft-release pen influenced the location of self-selected winter home ranges by reintroduced Whooping Cranes. If self-selected winter home ranges were clustered, at a large geographic scale, around the Chassahowitzka release site, this would present further evidence for a philopatric influence. If the ultimate geographic distribution of self-selected winter home ranges was influenced by the location of the soft-release pen, this could significantly influence the distribution of wintering cranes in future years, particularly if the soft-release pen is moved to another geographic location. To maintain comparability between this study and that of Maguire (2008), I used the same statistical analyses in the assessment of underlying factors affecting site selection whenever possible. Hypotheses were as follows: H1: Reintroduced migratory Whooping Cranes will at least briefly revisit the Chassahowitzka soft-release pen at a rate of >50% during their first two winters (i.e., conforming to our definition of philopatry). H2: The proportion of birds returning to the Chassahowitzka soft-release pen will be significantly greater than the proportion of birds failing to return to the soft-release pen

28 during their second and subsequent winters. I used a t-test and a significance value of p<0.05 for this analysis. H3: Male Whooping Cranes will revisit the Chassahowitzka soft-release pen at a greater rate than will female Whooping Cranes. Again, 60 birds (36 males and 24 females) were included in this analysis, and a t-test and significance value of p<0.05 were used. H4: Self-selected winter home ranges will be clustered around the Chassahowitzka NWR soft-release pen. These data were analyzed using visual inspection of geographic distribution of home ranges on a map and graphically, using the percentage of home ranges within various distance categories from the Chassahowitzka NWR soft-release pen. 15 Site Tenacity Site tenacity is defined here as the return of an animal to a previously occupied winter home range, in this case excluding the soft-release pen. I assessed a total of 124 cases of winter home range selection by individuals (59 cases of female site selection and 65 cases of male site selection) over the first 7 years of the reintroduction. A total of 52 birds were included in this analysis. This sample size is lower than that of the philopatry analyses due to the mortality of some birds on the breeding grounds prior to their third winter. I assessed the influence of this site tenacity in various ways. First, I recorded whether or not individual Whooping Cranes established home ranges within the boundaries of their previously occupied self-selected home ranges. The previously occupied home range had to be utilized by the crane for a minimum of ten days for site tenacity to be confirmed. If a site was not utilized or was abandoned in favor of a different site later in the season, I attempted to

29 discover a potential explanation for this shift. The strength of the influence of site tenacity can be affected by such factors as age, social status, gender, and environmental conditions. Hypotheses were as follows: H1: Whooping Cranes will return to self-selected home ranges that they or a member of their social group used in the previous year at a greater rate than they will establish entirely new winter home ranges. A simple z test was used to determine the significance of the detected difference, and a significance level of p<0.05 was selected.. H2: Site tenacity will be stronger for males than for females. Again, z tests were used to statistically test the significance (p<0.05) of any detected difference in site tenacity expression between male and female Whooping Cranes. H3: Site tenacity will increase with age, in the absence of stochastic environmental influences. Descriptive statistics were used to assess the validity of this hypothesis. The sample size was too small for robust significance testing. 16 Habitat Imprinting In this study, habitat imprinting is defined as the selection by Whooping Cranes of home ranges having a similar habitat composition to that of their initial winter home range, in this case the soft-release pen site. While return to the vicinity of the soft-release pen may be confounded by philopatric influence, any use of surrounding saltmarsh or, even more so, of distant saltmarsh would indicate a degree of imprinting on the habitat composition of the soft-release pen. Hypotheses were as follows:

30 H1: Whooping Cranes, especially in second and subsequent winters, will establish winter home ranges in habitats similar to those in which they were initially held at the end of their first migration (the Chassahowitzka soft-release pen). This hypothesis was evaluated using the compositional analysis of habitat selection for self-selected home ranges. The selection of home ranges containing saltmarsh habitat would support this hypothesis. H2: Any effect of habitat imprinting on the wintering grounds is over-ridden by the unsuitability of the saltmarsh habitat of central Florida s Gulf Coast. H3: The general habitat composition on winter home ranges is similar to the general habitat composition of the Necedah NWR rearing sites and the surrounding areas. I used a chi-square goodness of fit test to test these hypotheses, and I used the rearing site habitat composition data compiled by Maguire (2008) in the comparisons. The Florida home ranges of eleven bird units for year one and 17 bird units for year two were included in this analysis, and a significance level of p<0.05 was selected. 17 Social Facilitation and Territoriality The influences of the presence of other cranes on winter site selection by Whooping Cranes in this population were examined to determine whether Whooping Cranes were either attracted to or repelled from an area due to the presence of other cranes. Interactions with other Whooping Cranes, both migratory and Florida non-migratory, as well as interactions with migratory and non-migratory Sandhill Cranes, were considered. During each visual observation of wintering migratory Whooping Cranes, proximity to and interactions with any cranes not in the static social group being observed were recorded. Associations with Sandhill Cranes and non-migratory Whooping Cranes were classified as either fully or non-

31 fully integrated. A Whooping Crane or static social group of Whooping Cranes was 18 considered fully integrated with another crane or group of cranes when its movements were coordinated in the same direction and speed as the other crane or group of cranes. Interactions among separate groups of migratory Whooping Cranes wintering in close proximity to each other were also monitored closely, and any aggressive or non-aggressive direct exchanges were recorded. Hypotheses were as follows: H1: Whooping Cranes winter home ranges tend to overlap with Sandhill Crane home ranges. All birds wintering in Florida for both study years were included in this analysis, including 11 bird units for year one and 17 bird units for year two. Home ranges were considered to be overlapping when Whooping Cranes were observed within close proximity (15m or less) of Sandhill Cranes and were tolerant of them. I considered this to be a possibility because the presence of Sandhill Cranes, which are much more abundant in Florida than are Whooping Cranes and are more gregarious than Whooping Cranes, might indicate habitat suitability for Whooping Cranes. Additionally, previous observations of wintering Whooping Cranes in this population suggested a relationship between Whooping Cranes and Sandhill Cranes. H2: Full association between Whooping Cranes and Sandhill Cranes will be more common among single Whooping Cranes than for paired birds or bachelor groups. Whooping Cranes were considered to be in full association with Sandhill Cranes when they were consistently observed within 5m of Sandhill Cranes and exhibited coordinated movements with those Sandhill Cranes. The behaviors of 6 lone Whooping Cranes during

32 the study as well as the two winter seasons prior to the study were evaluated to test this 19 hypothesis qualitatively. Whooping Cranes typically travel and associate in pairs or small family groups or bachelor flocks on the wintering grounds (Allen 1952). This contrasts with the behavior of Sandhill Cranes, which tend to winter in larger, gregarious groups or flocks (Johnsgard, 1983). I surmised, therefore, that the most integrated, consistent associations with groups of Sandhill Cranes would be by Whooping Cranes wintering singly. These Whooping Cranes would seek the safety afforded by associating with Sandhill Cranes and would more closely bond with them in the absence of close Whooping Crane associates. H3: Whooping Cranes home ranges overlap partially when they are found in close proximity, especially when they share roosting areas, where cranes tend to be more gregarious. This hypothesis was based upon previous observations of wintering Whooping Cranes in this population. Winter territoriality does not appear to be as strong for this population as for wild Whooping Cranes wintering in Aransas, which tend to occupy exclusive territories. Non-aggressive interactions or low-level aggressive interactions, along with overlapping winter home ranges would support this hypothesis of slightly different wintering behavior in this population of Whooping Cranes. H4: When a crane s search area overlaps with another Whooping Crane home range, that crane will tend to land in the area and will often settle nearby. This means that a bird flying over an established Whooping Crane home range will land in direct or close proximity to that home range and will often establish a home range nearby. Because we are not capable of ascertaining the detectability of Whooping Cranes by

33 each other, this can only be evaluated by observations of behavior that may be indicative of a general tendency. To assess this tendency, I compiled evidence of conspecific attraction on the wintering grounds, including twelve cases of prospecting Whooping Cranes visiting or settling in direct or close proximity to established Whooping Crane home ranges. H5: Site selection by Whooping Cranes will be influenced by the Whooping Cranes with which they are associating. This was assessed for two groups of birds: those associating with older birds while returning to Florida independently for the first time, and those with altered social associations from their previous winters (e.g., a social associate died or the association ended). The proportion of birds that established home ranges within the boundaries or in direct proximity (within 5km) to their social associates previously occupied home range was calculated. The home range selection of 14 groups of birds in new social associations during the study were evaluated for this analysis, and data from 6 new social associations in years following the study were also noted but not included in the analysis, as there was a possibility of incomplete information for those years. 20 Environmental Stochasticity The potential influences of stochastic environmental factors such as drought, storm events, unfavorable prevailing winds during migration, and supplemental feeding were examined when deviations from expected or established patterns of site selection occurred. On a case-by-case basis, the existing environmental conditions as well as the background of each bird displaying movements or behaviors that differed from the general pattern followed by other Whooping Cranes in this population or that differed from the bird s own established pattern were scrutinized to determine whether stochastic environmental factors might best

34 explain this aberrant behavior. Aberrant wintering behaviors fell into the following 21 categories: (1) wintering outside of the migration corridor and primary wintering area; (2) abandoning a winter home range in favor of another area mid-winter (3) failing to return to a well-established winter home range. Due to the increasing tendency of Sandhill Cranes to establish new wintering grounds north of their historical wintering range (USFWS 2012), it was expected that a small number of Whooping Cranes may also short-stop to winter along the migration corridor north of the target wintering area, and this was not considered an aberrant behavior unless it began to occur at a high frequency. I hypothesized that most aberrant movements by Whooping Cranes in this population that could not be explained by the influence of social factors could be attributed to environmental stochasticity. Innate Preferences When a crane s selection of a winter home range was not easily explained by any combination of the above factors, the existence of innate preferences, defined as instinctive tendencies that have a genetic, rather than learned, basis, was considered and examined for plausibility on a case-by-case basis. The combination of innate tendencies with one or more of the above factors was also considered when appropriate and when the influence of various factors was difficult to differentiate. RESULTS Habitat Selection versus Availability for Daytime Home Ranges Fourteen home ranges of eleven bird units or individuals (20 individual birds in total) were included in the habitat selection analysis for year one of the detailed study (winter

35 ), and eighteen home ranges of seventeen bird units or individuals (34 individual 22 birds in total) were included in the analysis for year two (winter ). The numbers of individuals and home ranges analyzed is uneven because birds sharing a search area and home ranges with other birds were included only once in the analysis, and some individuals or social groups had multiple home ranges in a given winter. 95% Home Range versus Search Area. -- On average, wooded or forested uplands comprised the greatest proportion of a search area, followed by forested wetlands, open pasture, urban and residential development, miscellaneous, saltwater marsh, freshwater marsh, scrub/shrub wetlands and uplands, open water, and finally wet prairie (See Table 1 for the composition of individual search areas by cover type). When I compared the habitat availability within selected home ranges to the mean search area composition for all known search areas, selection of 95% home ranges within search areas was non-random for year 1 (p<0.001). Daytime home range cover type preference ranks were as follows: (1) open pasture (2) freshwater marsh (3) wet prairie (4) open water (5) shrub/scrub (6) wooded upland (7) urban/residential (8) forested wetland (9) miscellaneous and (10) saltwater marsh. Using a critical p value of 0.05 with 13 degrees of freedom, significant preferences included freshwater marsh over miscellaneous and saltwater marsh, open pasture over miscellaneous, saltwater marsh, and urban/residential, and open water over saltwater marsh. The clear avoidance of saltwater marsh is not only statistically significant but also biologically significant, as the soft-release pen is located within saltwater marsh, and a majority of the birds visit this cover type initially, only to reject it in favor of open inland cover types. Next, I compared the habitat availability within qualifying home ranges (those for which the individual birds search areas were known) to the habitat availability within

36 individual search areas. In view of its absence from any self-selected home range, I 23 eliminated saltwater marsh from this analysis. Again, habitat selection at this scale was nonrandom (p <0.001). Preference ranks using this analysis were as follows: (1) freshwater marsh (2, tied) open pasture = open water (3) wet prairie (4) forested wetland (5) shrub/scrub (6) wooded upland (7) urban/residential and (8) miscellaneous. Significant preferences (p<0.05) were freshwater marsh over miscellaneous, open pasture over miscellaneous, and open water over miscellaneous. Using the average search area habitat composition, results for the second year were similar (p<0.001), with a few cover types moving up or down one place in the ranking. Ranks for year two were as follows: (1) open pasture (2) freshwater marsh (3 tied) wet prairie = open water (4) scrub/shrub (5) wooded upland (6) forested wetland (7) miscellaneous (8) urban/residential and (9) saltwater marsh. Significant preferences (p<0.05) included freshwater marsh over saltwater marsh, open pasture over saltwater marsh, urban/residential, and miscellaneous, wet prairie over saltwater marsh and open water over saltwater marsh. Using individual search areas in the compositional analysis for study year 2, habitat selection was again non-random (p<0.05). Cover type rankings were as follows: (1) wet prairie (2) freshwater marsh (3) open pasture (4) open water (5) shrub/scrub (6) wooded upland (7) forested wetland (8) urban/residential and (9) miscellaneous. Significant preferences (p<0.05) included freshwater marsh over miscellaneous, open pasture over miscellaneous, and open pasture over urban/residential. In year one, all home ranges contained open pasture, freshwater marsh, and open water. In year two, all home ranges contained open pasture, and all but one home range contained freshwater marsh and open water. Along with their high ranking for each year,

37 regardless of the method of analysis, this fact highlights the combined importance of these 24 three landscape elements as winter habitat of Whooping Cranes. Table 2 and 3 list the cover type composition of the 95% home ranges for migratory Whooping Cranes in Florida during the season one and two of data collection, and Table 4 provides a side-by-side comparison for 95% home range habitat rankings using the two methods for determining landscape-level habitat availability (mean vs. individual search areas). Core (50%) Home Range versus 95% Home Range.-- The comparison of core home range composition to 95% home range composition reveals those cover types most important to Whooping Cranes in their primary use areas within their larger home range. Compositional analysis again revealed selection to be non-random (p<0.05) at this spatial scale. Because no bird used saltmarsh within its home range, this habitat type was eliminated from this analysis for both years. In year one, similarly, no home ranges contained miscellaneous habitats; therefore, this type was also eliminated. Year one rankings were as follows: (1) open pasture (2) freshwater marsh (3) open water (4) forested wetland (5, tied) wet prairie = urban/residential and (6, tied) shrub/scrub = wooded upland. Significant differences (p<0.05) included open pasture over urban/residential and open pasture over wooded upland. Year two rankings were as follows: (1) open water (2) open pasture (3) freshwater marsh (4, tied) wet prairie = urban/residential (5) wooded upland (6, tied) forested wetland=miscellaneous and (7) shrub/scrub. Significant differences (p<0.05) include open water over miscellaneous and wet prairie over miscellaneous. Table 7 gives the average core home range compositions by habitat type for the two study years, and Tables 5 and 6 provide individual core home range compositions by habitat for years one and two, respectively.

38 Locations versus Core (50%) Home Range.--The finest scale of habitat selection analyzed in this study is the choice of specific cover types within the core home range. For both years, the proportional use of cover types differed significantly (p<0.05) from the availability of cover types within the core home range. For year one, rankings were as follows: (1) freshwater marsh (2, tied) open pasture=wet prairie (3) open water (4, tied) forested wetland=wooded upland (5) shrub/scrub and (6) urban/residential. Significant differences (p<0.05) included open pasture over forested wetland and shrub/scrub, freshwater marsh over forested wetland and shrub/scrub, forested wetland over shrub/scrub and urban/residential, open water over wooded upland and forested wetland, wet prairie over forested wetland and urban/residential, and wooded upland over forested wetland. Rankings for year two were as follows: (1) wet prairie (2) open pasture (3) freshwater marsh (4) open water (5) shrub/scrub (6) urban/residential (7) forested wetland and (8) wooded upland. Significant preferences (p<0.05) included freshwater marsh over urban/residential, open pasture over forested wetland and urban residential, shrub/scrub over forested wetland, wet prairie over forested wetland, urban/residential, open pasture, and open water, and urban residential over wooded upland. Tables 8 and 9 provide the proportions of locations in each cover type for individual bird units in years one and two, respectively. Table 10 and Figure 6, respectively, show overall proportions of locations within each cover type over the two study years. 25 Habitat Selection versus Availability on Roosting Sites Identical analyses were performed for roosting sites as for daytime home ranges at each of the selection scales. Due to the difficulty in attaining roost locations for multiple

39 groups of birds on a given evening and the relative stability of roost locations over time 26 during the study years, fewer roost locations were recorded for each group than were daytime locations. For the first year of the study, 15roost sites of 11 bird units were analyzed, and for the second year, 19 roost sites of 18 bird units were included in the analysis. 95% Roost Site versus Search Areas.--At the landscape scale of selection, proportional use of cover types at roost sites differed significantly from availability for both study years when comparing individual search areas to roost sites (p < for both years). Birds that shared roost sites but not individual search areas were considered separately for this analysis. No roost site contained salt marsh habitat, and miscellaneous habitats comprised such low proportions (< 1% on average) of roosting sites as to compromise the accuracy of the analysis, that these cover types were eliminated from the ranking matrix analysis. In , preference rankings were: (1) freshwater marsh (2) open pasture (3) open water (4) wet prairie (5) forested wetland (6) urban/residential (7) wooded upland (8) shrub/scrub. Significant preferences (p<0.05) included freshwater marsh over shrub/scrub and open pasture over shrub/scrub. Rankings for year two were as follows: (1) wet prairie (2) freshwater marsh (3) open water (4) open pasture (5) wooded upland (6) shrub/scrub and (7) forested wetland. Statistically significant preferences were freshwater marsh over shrub/scrub and wet prairie over wooded upland. Wet prairie was preferred over shrub/scrub and forested wetlands at a nearly significant level (the critical t-statistic was 1.71 and the t- statistics for these two comparisons were 1.65 and 1.64, respectively). Table 11 provides the composition of 95% roost sites by cover type for year one, and Table 12 provides the same information for year two of the study.

40 Core Roost Site versus 95% Roost Site.--Selection of a core roost site within a 27 95% roost site was found to be non-random at a significance level of p < for both study years. Year one rankings were as follows: (1) wet prairie (2) freshwater marsh (3) open pasture (4) open water (5) forested wetland (6) wooded upland and (7) shrub/scrub. Significant preferences included: freshwater marsh over forested wetland, shrub/scrub and wooded upland; forested wetland over wooded upland; open water over forested wetland and shrub/scrub; open pasture over forested wetland and shrub/scrub; wet prairie over forested wetland, shrub/scrub and wooded upland; and wooded upland over shrub/scrub. For year two, I eliminated shrub/scrub and other under-utilized, low availability cover types from the analysis in order to reduce the occurrence of falsely low standard errors and resultant high significance levels in the ranking matrix due to replacement methods for zeros in the dataset. Habitat selection at this scale was, again non-random (p<0.001). Cover types not included in the ranking, including shrub/scrub, miscellaneous, saltmarsh, and urban/residential were absent or nearly so from the core roosting areas of these birds and can therefore be considered to be selected against at this scale. Preference rankings were as follows: (1) open water (2) freshwater marsh (3) wet prairie (4) open pasture (5) forested wetland (6) wooded upland. There were no significant selection preferences in this analysis; however, the t- statistics for wet prairie over wooded upland and for wet prairie over forested wetland were close to the critical value (1.48 and 1.53 respectively, as compared to a critical value of 1.73 for p < 0.05). Core roost site compositions by cover type for years one and two of the study are found in Tables 13 and 14, respectively. Roost Locations vs. Core Roost Sites.--For both years, the selection of roost locations (the actual spot where the bird roosted) within core roosting areas was highly

41 significantly non-random (p < 0.001). In , preference rankings were: (1) 28 freshwater marsh (2) open water (3) wet prairie and (4) open pasture. During that winter, no birds were observed roosting in forested wetland although the cover type was available within some birds core roost areas. The type was not included in the analysis, however, due to low availability and non-utilization. The birds do, though, appear to be selecting against it at this level in comparison to the four listed cover types. Significant preferences were freshwater marsh over open pasture and open water over open pasture. For the second study year, the preference rankings were: (1) open water (2, tied) wet prairie=freshwater marsh (3) forested wetlands and (4) open pasture. Though birds were never observed roosting on dry land, they sometimes did roost in temporarily flooded pasture adjacent to wet prairie or freshwater marsh. In these cases, the computer-generated classification was reclassified to the nearest wetland type. Significant preferences included wet prairie over forested wetland, freshwater marsh over open pasture and open water over open pasture. Tables 15 and 16 present the proportion of roost locations in each cover type for individual bird units for years one and two of the study, respectively. Table 17 provides the ranking matrix for this particular scale of selection and illustrates the final step in the compositional analysis of habitat selection used in this study (after Aebischer et al. 1993). Behavior within Cover Types At the level of daytime home range selection, Whooping Cranes appeared to use cover types in a supplementary fashion (no 95% home ranges consisted of a single cover type), as all activities (foraging, walking locomotion, loafing (resting and preening/maintenance behaviors, and roosting) occurred in multiple cover types. See Figures

42 7 through 10 for graphical depictions of the utilization of cover types for these behaviors. 29 However, the proportion of observations of these behaviors in the different cover types did vary by behavioral category. Foraging was observed in 11 of the 12 cover types, with open pasture being the most utilized cover type for this behavior (57% of foraging observations were made in open pasture cover). Collectively, wet prairie, freshwater marsh, and open water accounted for 41% of foraging observations, in fairly equal proportions. These cover types, then, supplement each other to meet Whooping Cranes energetic requirements. Walking locomotion was observed, again, most often in open pasture cover, comprising 54% of walking locomotion observations, with open water, freshwater marsh, and wet prairie making up 42% of observations, collectively. The cover type used for walking locomotion is likely a function of both the ease of walking through that particular cover type and the preference for that cover type for meeting another requirement. Cover types were used in different proportions for loafing and roosting behaviors, with wetland cover types gaining importance for these behaviors. Loafing behaviors were most often observed in open water, comprising 43% of observations. Open pasture, freshwater marsh, and wet prairie made up 54% of loafing observations. Roosting was observed in only four cover types (forested wetland, open water, freshwater marsh, and wet prairie), suggesting that these types are critical for meeting this requirement and may complement or replace each other on the landscape in providing adequate roosting habitat for Whooping Cranes. The stability of roost sites over time supports this, as birds appear to require just a single suitable roost site within their winter home range. In short, Whooping Cranes require wetland cover for roosting, while various cover types are supplementary for meeting foraging requirements.

43 Philopatry 30 Though only one pair of birds established a home range near the Chassahowitzka soft-release pen, philopatry was indicated in this population by the high rate of initial revisits to the pen site during birds first two years. Over all years, 60.3% of all birds that survived to at least their first winter following release initially but temporarily revisited the Chassahowitzka soft-release pen during their first two winter seasons. Because some of the cohorts experienced different conditions or disruptive events on the wintering grounds or during their first independent northward migration, cohorts were analyzed separately to evaluate the potential impact of the differing conditions on the rate of return to the softrelease pen. Due to strong indications that stochastic environmental factors, discussed in detail under that heading, that precluded the influence of philopatry on the winter home range selection of the birds of the 2003 cohort, I wanted to examine the return rate for birds not influenced by these environmental events. If the 2003 cohort is removed from the analysis, the population return rate to Chassahowitzka NWR within two years of initial release increases from 60.3% to 66.7%. The 2003 cohort returned to Chassahowitzka at a rate of 40%. The 2005 cohort was managed differently on the wintering grounds from the previous cohorts, with a potential impact on their tendency to show a philopatric attraction to the Chassahowitzka NWR. This was the first year that a temporary holding pen site was selected at the Southwest Florida Water Management District s Hálpata Tastanaki Preserve, located in southern Marion County. The young birds arrived at this site on 13 December 2005, and 18 of the 19 birds were subsequently led behind ultra-light aircraft to the

44 Chassahowitzka NWR pen-site on 9-11 January. The nineteenth bird appeared flightless and failed to follow the ultra-light and so was transported via vehicle and then airboat to the softrelease pen. When the 2005 cohort is removed from the analysis, the population return rate for ultra-light led birds increases from 60.3% to 68.9%. The birds of the 2005 cohort returned briefly to Chassahowitzka NWR at a rate of 38.9%. Of the 18 birds in the cohort, five (27%) of them utilized the Hálpata-Tastanaki holding pen site for short periods during their second winter, and two of these five did not return to Chassahowitzka. The initial use of a holding pen at Hálpata Tastanaki Preserve, therefore, appeared to preclude the adoption of philopatry toward the saltmarshes of Florida s central Gulf Coast. The strongest indication of the influence of philopatry on this population, eliminating mitigating environmental and management influences, was the rate of return by all but the 2003 and 2005 cohorts. When just the 2001, 2002, and 2004 cohorts are examined, the overall return rate to Chassahowitzka NWR during the first two years was 83.3%. This was not quite as strong as the philopatric influence on the movement of birds on the breeding range, which returned to the immediate vicinity of Necedah NWR at a rate of 87% after their initial northward migration (Maguire 2008). However, an 83.3% return rate was sufficient to conclude that these birds are influenced by a tendency to return, at least temporarily, to the vicinity of the pen site from which they were soft-released following the conclusion of their southward migration and where they spent the majority of their first winter season. For the 2001, 2002, and 2004 cohorts, I also compared the differential rates of return by males and females. Overall, independent of their social associations, males from 31

45 these cohorts revisited Florida s Gulf Coast salt marshes at a rate of 87.5%, while females 32 returned at a rate of 78.6%. A simple two-proportion z test of these rates indicates that the male return rate was significantly greater than that of the females of this population (Z=4.777, p<0.0001). This preliminary analysis suggests a possible similar trend for weaker philopatry by females on the winter as well as the breeding grounds. In addition to calculating the rate of revisits to the soft-release pen, I examined the distance of the eventual self-selected home ranges of birds in this population to Chassahowitzka NWR to determine whether philopatry might be affecting the geographic distribution of ultimate winter home ranges. Figure 11 is a map of all winter home ranges, and Figure 12 graphically depicts the geographic distribution of winter home range distances, in 100km increments, from the Chassahowitzka NWR soft-release pen site. It appears that the large-scale distribution of home ranges is influenced by philopatry associated with Florida s Gulf Coast saltmarshes. A further analysis of home range locations revealed that, among those birds not settling near the central Gulf Coast saltmarshes, a majority of birds selected home ranges in specific geographic areas: the Paynes Prairie Wetland Complex in Alachua County, Florida, the Hixtown Swamp Wetland Complex in Madison County, Florida, the Ace Basin Wetland Complex in Colleton County, South Carolina, short stops along the migration route, and miscellaneous wintering areas in Florida or in Georgia or the Carolinas. One distant outlier wintered in Louisiana. Figure 13 reveals the distribution of home ranges by location category (I eliminated all Paynes Prairie locations from the first category, as some lie within 100 km and others lie just outside that distance).

46 Site Tenacity 33 In this study, I examined 124 cases of individual Whooping Cranes (both within social groups and singly) selecting wintering home ranges over the first seven years of the reintroduction. First, I determined the frequency with which all birds over all years returned to a previously occupied self-selected home range. On an individual level, the percentage of birds using a home range within the boundaries of the previously occupied area was 56%. This statistic indicates that site tenacity affected a majority of birds; however, there are additional considerations. When birds social associations changed, they were often influenced by their new social associates to adopt a home range within the boundaries of the associates former home range. Although they were not being tenacious to their previously occupied home range, the current home range selection was influenced by their social associates site tenacity. When these individuals were removed from the analysis, the overall proportion of home range selections influenced by site tenacity grew to 73%. This proportion is the best index of the general influence of site tenacity on individuals in this population, and it is significantly greater than the proportion of cases not influenced by site tenacity (z=7.112, p<0.001). An examination of the 27 cases in which the location of a bird s home range was known and in which site tenacity was not expressed reveals some interesting and noteworthy patterns. Of these 27 cases, 14 involved short-stops along the migration route, meaning that rather than returning to the expected wintering range, the birds selected a winter home range north of Florida along the migration route, 15 involved birds in new social associations, and five were directly attributable to drought severely diminishing the suitability of previous home ranges. These factors may also interact to increase the likelihood of birds settling in

47 new home ranges: For example, drought coupled with changed social associations may have increased the likelihood of home range shifts in 12 of the 14 short-stop cases. Further, 18 of the 27 cases of home range shifts occurred in the winters of and , years of drought on the southern end of the birds winter range. Overall, 23 of the 27 cases (85%) involved at least one of the above factors, indicating that while site tenacity strongly influenced home range selection for these birds, social and environmental factors affected its expression. I also examined the differential influence of site tenacity according to gender and age of birds. Females returned to their own previously occupied home range at a frequency of 59%, while males did so 72% of the time. This is a noteworthy but marginally significant difference (z=1.5259, p=0.063). Females accounted for 66% of the cases in which a bird settled in the previous home range of its social associate rather than its own previous home range, while males accounted for 33% of these cases, again a telling but marginally significant difference (z=1.8257, p=0.067). The influence of site tenacity appeared to strengthen with age, with the most considerable increase in the expression of site tenacity occurring between the first returning winter and the second winter. See Table 18 for a breakdown of site-tenacity expression by age. 34 Habitat Imprinting Results indicated weak, if any, influence of imprinting on the habitat at the softrelease pen. None of the birds adopted self-selected winter home ranges in habitat similar to the Chassahowitzka NWR soft-release pen. However, as noted above in the philopatry section, a majority of the birds did return briefly to the CNWR soft-release pen and

48 temporarily utilized the saltmarsh areas in and around the pen before eventually abandoning the area. This behavior appears to be more plausibly attributable to philopatry. One might discontinue the analysis at this point. However, cases in which birds utilized salt marsh habitat outside of the general vicinity of CNWR complicate the issue. Below are the instances of use of salt marsh habitat: In the winter of 2002, male 1-01 and female 2-01 settled for 15 days on the Greenleaf Bay area, 11km north of the CNWR soft-release pen, roosting in a saltwater bay and foraging during the day primarily in upland saltmarsh habitats such as an extensive area of saltgrass (Distichlis sp). The last night that they were observed roosting on the salt marsh was 13 December, which was also the first night of extreme high tides since their initial arrival and his then mate 2-02 also utilized salt marsh habitat for short periods of time in In the winter of 2004, male and females 3-03 and spent part of three days in salt marsh habitat about 15km north of the CNWR soft-release pen. Upon completing their southward migration to Florida, they had initially spent at least four days at an inland site in Marion County, and then returned to CNWR briefly before flying south to roost for one evening in the inland Green Swamp area in Polk Co. They then spent three days utilizing salt marsh both at CNWR and at the northern site before returning to Marion County for the remainder of the winter. The 2003 Cohort, as previously discussed, exhibited very different winter site selection from the other Whooping Cranes in this population, especially in the first year 35

49 following release, likely due to stochastic environmental events. All but 3-03, and (this group had migrated to Michigan in the spring of 2004 but then navigated around Lake Michigan and spent the remainder of the breeding season within the core reintroduction area in Wisconsin) wintered in either South Carolina or North Carolina during the winter of 2004, in five separate groups. I tracked four of these five groups closely as they arrived on their wintering grounds, and I observed an interesting pattern: all but one of these groups first utilized salt marsh habitats on the Atlantic Coast before settling further inland. Male (the one bird not following this pattern) also approached to within 15km of the coastal estuaries. However, he initially settled at a former inland rice plantation, consisting of several freshwater impoundments, which was being managed for waterfowl and included a flooded corn field, a favored foraging site for cranes in this population. Had he not encountered this site, his trajectory would soon have brought him to the salt marsh. Unlike the birds that settled in Florida, the South Carolina birds remained in close proximity to the salt marsh and utilized both brackish and freshwater wetlands. The brackish marshes in these locations differ greatly from those in Florida, however, due to their protection from extreme tidal fluctuations. These birds also all had access to flooded cornfields and freshwater marshes or ponds for foraging and drinking. The initial use of salt marsh habitat by these birds, though, is the most telling part of this story. Because they were so distant from CNWR and clearly were not lost (several of them returned to CNWR the following year), their use of salt marsh habitat can only be attributed either to innate preference or to habitat imprinting. The fact that reintroduced non-migratory Whooping Cranes in Florida, which share the same genetic lines as the eastern migratory birds, have never shown such an affinity for salt marsh (M. Folk, Florida Fish & Wildlife Conservation Commission, personal 36

50 communication), indicates that the more likely influence is indeed imprinting on the salt 37 marsh habitat at the CNWR soft-release pen. The second hypothesis regarding habitat imprinting was that any influence that habitat imprinting may have had on Whooping Cranes in this population was over-ridden by the unsuitability of the salt marsh habitat of coastal Florida. The failure of any birds to establish long-term winter home ranges on Florida s coastal salt marsh, despite the vast majority of birds returning briefly to this habitat type, strongly supports this hypothesis. As discussed extensively by Urbanek et al. (2010) the following biological characteristics of salt marsh habitat lead to the birds rejection of the area as long-term winter home ranges: Lack of availability of fresh water Lack of availability of stable roost conditions Domination of upland areas by black needlerush (Juncus roemerianus), which grows rank and tall in many areas, and inhibits the birds movements and reduces visibility Soft, unstable muck or jagged oyster shell substrates in the tidal pools and creeks In the absence of supplemental food and fresh water, these factors lead to the abandonment of salt marsh habitat in favor of the open pastures, freshwater marshes, and wet prairies of inland Florida. Next, I compared the habitat composition of the Necedah NWR rearing sites and surrounding areas to the general habitat composition of self-selected winter home ranges in Florida. I re-classified the habitat types used in this study and by Maguire (2008) into five general categories (see Table 19). A chi square goodness of fit test was performed for each

51 of these comparisons. The average composition of home ranges in Florida during the and winters differed significantly from that of the immediate rearing site vicinity in Wisconsin (x2=51.437, p<0.0001, df=5) and from that of the surrounding landscape in Wisconsin (x2=93.997, p<0.0001, df=5). It should be noted, however, that birds did utilize many similar habitat types on the breeding and wintering ranges even though the composition of these types within their home ranges was quite different. Table 18 provides the raw data used for the chi square analysis. Much like the habitat in and around the Necedah rearing facility, the Whooping Cranes in this population selected winter home ranges comprised of open areas with wetlands for roosting. Due to the differing structure and composition of habitats within Florida home ranges and the composition in and around the Necedah NWR rearing facility, however, habitat imprinting on the breeding range does not appear to adequately explain the patterns in winter habitat selection by Whooping Cranes in this population. At Necedah, the rearing wetlands are connected to large, managed impoundments, and the nearby fields primarily consist of grass strips managed as runways for the ultra-light aircrafts. In comparison, the wetlands selected by Whooping Cranes in Florida are primarily isolated natural freshwater marshes and wet prairies surrounded by improved cattle pastures. While the fields at the Necedah NWR rearing sites are not sources of abundant natural foods, the manure-enriched Florida pastures are a source of invertebrate prey and are primary foraging areas for wintering birds. So, not only are the structure and composition of these habitats different on the winter and summer ranges, the utilization of the cover types by the birds also differs. Finally, I wished to assess whether the use of the Hálpata Tastanaki holding pen site may have influenced home range selection by birds in this population. The pen 38

52 encompassed an ephemeral, marshy wetland and herbaceous upland. Surrounding landcover included floodplain swamp and oak scrub associated with the Withlacoochee River, and long-leaf pine (Pinus palustris)-turkey-oak (Quercus laevis) sandhills on the uplands (Southwest Florida Water Management District 2010). A heavily grazed cattle ranch with improved pastures bordered the preserve. In years following its construction, this pen was used to hold the juvenile birds until the majority of older birds had settled on self-selected winter home ranges, reducing the likelihood of their visiting the Chassahowitzka pen site. Though data are limited, the analysis below indicates the use of this holding pen site may be affecting the habitat selection patterns of those birds that are held there. I assessed the habitat composition of the area around the Hálpata Tastanaki holding pen and compared it to a typical self-selected Whooping Crane winter home range, and found that 32% of the area was composed of forested wetlands, as compared to an average of 6% in a typical Florida Whooping Crane home range during the two intensive study years. Twentyeight percent of the area was composed of wooded uplands, as compared to an average of 7% in analyzed home ranges. These are wooded habitat types in which crane mortality rates have been high. The areas around mortality locations contained more forested wetland and forested upland than did a typical crane home range during the two study years (Table 25). Two birds that were held at Hálpata Tastanaki Preserve were killed by predators during their first independent winter season. I performed a two-sample z test for comparison of means for percent composition of forested wetland and forested upland between the areas surrounding three winter mortality sites, including the two mentioned above, and typical winter home ranges. These areas contained a 39

53 significantly larger portion of forested wetlands (z=1.9275, p=0.026), and a nearly 40 significantly larger portion of forested uplands (z=1.4237, p=0.077). Social Facilitation and Territoriality Social interactions with other cranes played an important role in winter habitat selection by migratory Whooping Cranes. During the two years of the in-depth habitat study, all Whooping Crane home ranges in Florida overlapped with Sandhill Crane home ranges, and at some point, all Whooping Cranes interacted directly with Sandhill Cranes on their home ranges. The presence of Sandhill Cranes in an area may have acted as an indicator of habitat suitability to Whooping Cranes as they searched for wintering areas, but the density and dispersion of Sandhill Cranes in Florida meant that much of the suitable habitat was occupied by Sandhill Cranes, making overlap likely by chance alone. Whooping Cranes that wintered outside of the core wintering area, in states such as North Carolina and South Carolina, where Sandhill Crane densities were much lower to nearly absent, were less likely to settle in close proximity to Sandhill Cranes by chance alone. Number was eventually joined by a Sandhill Crane on his winter home range in South Carolina, with which he became fully associated, providing particularly strong evidence, given the low density of Sandhill Cranes in the region, of a tendency by these two species to interact and to choose sites based on the others presence. Figure 14 shows the distribution of nonmigratory Sandhill Cranes in Florida and the centers of all Whooping Crane home ranges prior to the winter of This figure provides further visual evidence of the spatial coincidence of Sandhill Cranes and Whooping Cranes in Florida. Areas of concentrated Whooping Crane home ranges overlap almost precisely with areas of high non-migratory

54 Sandhill Crane occupancy. The area of concentrated Whooping Crane home ranges in north central Florida (the Hixtown Swamp Wetland Complex and San Pedro Bay area) that was not occupied by Florida Sandhill Cranes was a major wintering site of migratory Sandhill Cranes. It was not possible to know with certainty whether the Whooping Cranes were using areas occupied by Sandhill Cranes primarily because those areas contained attractive habitat for both species or because the Whooping Cranes were attracted by the presence of Sandhill Cranes per se. I suspect, however, based on on-the-ground-observations and the extreme degree to which individual Whooping Crane home ranges overlapped with Sandhill Crane use areas, that Sandhill Cranes were indeed strongly influencing Whooping Cranes habitat selection at both the landscape and local scales. Sandhill Cranes also lead Whooping Cranes to sources of abundant or supplemental food. I often observed Whooping Cranes in Florida utilizing supplemental feed (in the form of both domesticated animal feed and wildlife/game feed) while in the company of resident Sandhill Cranes that most likely made them aware of the presence of the food source. Were single Whooping Cranes more likely to be fully associated with Sandhill Cranes than birds in larger social units? This held true for the four winter seasons that I examined ( through ). WCEP project data are incomplete on this subject for other years. The only birds that were consistently observed in full association with Sandhill Cranes were single birds. There were twelve such cases over the four seasons under consideration, involving six different Whooping Cranes. In all but one case, the bird was in full association with Sandhill Cranes prior to settling on its winter home range. This makes sense, as the tendency for single birds to seek the company of Sandhill Cranes could be 41

55 advantageous throughout the year; not just on the wintering grounds. Based on the behavior of all single birds in the population over this time period, whenever Sandhill Cranes were present within the home range of the Whooping Crane, the Whooping Crane was fully integrated with Sandhill Cranes. The single birds that did not winter in the company of Sandhill Cranes were in South Carolina or North Carolina, where Sandhill Crane densities were quite low. Did Whooping Cranes wintering in close proximity to one another have overlapping home ranges? Though I observed some aggression between Whooping Cranes on the wintering grounds, aggressive interactions were short-lived when a new group of cranes initially entered an area already occupied by other Whooping Cranes. If the newly arriving cranes endured that initial aggression, it eventually dissipated, and all of the cranes persisted at the site, sometimes in direct contact and other times occupying different portions of the general area. During the first four wintering seasons, there were six cases in which the home ranges of individuals, pairs, or social groups overlapped. These cases are summarized in Table 21. Figure 15 shows the spatial overlap of five winter home ranges. This pattern of overlapping home ranges on the wintering grounds has continued throughout the reintroduction years. Were Whooping Cranes either attracted to or repelled from one another in general when selecting winter home ranges? I hypothesized that Whooping Cranes would tend to land in the nearby vicinity when flying over an area already occupied by other Whooping Cranes. Table 22 summarizes the instances in which birds visited or settled in close proximity to established migratory or non-migratory Whooping Crane home ranges. These instances support the idea that these Whooping Cranes will tended to cluster together on the 42

56 landscape when they are within detectable range (conservatively presumed to be within 43 15km) of one another. Were Whooping Cranes influenced by their social associates when selecting winter home ranges? The importance of social associations was already touched upon in the site tenacity section of this study, wherein I found that fluctuating social associations affected the winter site fidelity of individuals in this population. To determine the importance of the influence of other birds on the site selection of Whooping Cranes in this population, I examined the home range selections of all birds in novel social associations (those differing from previous winters) that included at least one bird with a previously established home range. Most (85%) birds in new social associations returned to within 5km of one of the associates previously established home ranges. The data are summarized in Table 22. Additional data from two years following the study are included in Table 23. Winter site selection by individual Whooping Cranes was strongly influenced by the birds social associates. The most common situation involved birds adopting the home range of new social associates and abandoning their own home range from the previous year. Except when drought conditions made a previously occupied home range unsuitable, it was exceedingly rare for birds in social groups not to return to one of the birds previous home ranges, particularly as the birds got older. During the first four years of the project, in fact, only two out of the 13 novel social groups failed to do so and shortstopped near Winchester, Tennessee in and returned to that location the following year. The home range selection of 1-02 and 6-03 was the second case of failure to return to a previously established home range. Male 6-03 appeared to skew female 1-02 s home range selection to the NE. As discussed in the Environmental Stochasticity section, 6-03 was one

57 of the birds that encountered an extreme storm cell on its initial northward migration wintered closer to the primary wintering area than did the other birds that encountered the storm cell, possibly indicating an influence by 1-02 (see Figure 16 for a visual depiction of this apparent interaction). 64% of all birds that did not return to their previous home ranges experienced a change in social association from the previous year, indicating an interaction between philopatry and social facilitation. 44 Environmental Stochasticity The first two release cohorts of Whooping Cranes in this population (the 2001 and 2002 cohorts) followed a similar pattern upon returning to the wintering grounds on their own. A vast majority of the birds returned briefly to the Chassahowitzka pen site in the first and second year following release, and those birds not returning to the pen site short-stopped somewhere within the migration corridor from Necedah NWR in Wisconsin to Chassahowitzka NWR in Florida. After briefly visiting the soft-release pen, all returning birds eventually abandoned the saltmarsh in favor of inland wintering areas in north-central Florida. This, then, became the expected pattern under normal rearing and climatic conditions. The three primary deviations from this pattern are summarized below and in Table 24. Detailed narratives of the first two cases are included in Appendix B. Case 1: 2003 Cohort: Group I. The first group of eight birds to leave the CNWR pen site in the spring of 2004 encountered strong westerly winds that pushed them significantly east of the migration route used the previous fall. They were additionally flushed from a roost location in the mountains at night. They migrated to lower Michigan rather than to Wisconsin, where they split into two groups. While one group of birds settled

58 in Michigan, the other eventually returned to Wisconsin that summer. The group of birds 45 that summered in Wisconsin showed a typical wintering pattern for this population, while the birds that summered in Michigan wintered in North Carolina and persisted in these atypical summering and wintering patterns, prompting manipulation by project staff to correct their patterns and bring them into contact with the rest of the population. Case 2: 2003 Cohort Group II.--The second group of eight birds to leave the CNWR pen site encountered a violent storm cell just north of the Georgia-Florida border. After the storm had passed, these birds all continued north in multiple groups, and migrated as expected to the Necedah NWR breeding grounds. The following fall, in five separate groups, with the exception of 6-03, these birds all migrated to southeastern South Carolina (See Figure 16). The only plausible explanation for the patterns exhibited by the 2003 release cohort birds was their unique exposure to stochastic environmental conditions on their first northward migrations. No other bird or group of birds in this population has exhibited such tendencies to migrate to the Atlantic Coast rather than remaining in the primary migration corridor. One other bird, female 8-05, wintered outside this corridor, and her behavior did not seem to have an environmental explanation. I will examine her case in the innate preferences section of this study. Case 3: Mid-Winter Home Range Shifts.--Early in this study, when environmental conditions in Florida were normal and conducive to wintering Whooping Cranes, as many wet prairies and marshes contained enough water for safe roosting, it was fairly rare for birds to shift wintering locations mid-winter. Such shifts were often inexplicable, as in the case of 6-01, who often shifted locations multiple times over the course of a winter. His behavior

59 may have been due to his close association with migratory Sandhill Cranes, which are known to shift wintering locations more readily (Pogson and Lindstedt 1991, Wenner and Nesbitt 1987). Sometimes, though, a specific environmental variable may explain this mid-winter movement. In the winters of and , the departure of wintering birds from Tilobe Farms Ranch in Pasco County coincided with the discontinuation of the seasonal feeding of wild turkeys on that ranch. The area may not produce enough natural foods to support the large number of Whooping Cranes and the multiple pairs of Florida Sandhill Cranes utilizing the area. During the winters of and , birds shifted winter home ranges with greater frequency than had been observed previously (S. Kerley, International Crane Foundation, personal communication). They also short-stopped along the migration route at a greater rate than ever before and showed a weakened tendency to return to their previous use areas (see the Site Tenacity section of this study). The long-term drought affecting the primary wintering range doubtless was at the root of these changes in the wintering ecology of this population of Whooping Cranes. 46 Innate Preferences In the observed wintering ecology of this reintroduced migratory population of Whooping Cranes, there has been only one instance of winter site selection that did not appear to be influenced by the factors of philopatry, site tenacity, social facilitation, habitat imprinting or environmental stochasticity. During the winter of , female 8-05 settled in southeastern Louisiana. Her winter home range was significantly closer to the wintering range of the wild Whooping Crane population to which she is genetically related than to all other known winter home ranges of reintroduced eastern migratory Whooping Cranes,

60 though it is still over 7,000km from the Aransas NWR. It is plausible, therefore, that she was drawn further west by some innate affinity for the migration pattern of her genetic ancestors. This conclusion is very tentative, as lone young females uninfluenced by geographic barriers or influenced by stochastic environmental conditions have tended to disperse more widely than other birds in this population experiencing normal climatic and geographic influences (Maguire 2008, Urbanek 2010). This temporary deviation (the bird returned to Florida with a male Whooping Crane the following winter), therefore, is more likely an extreme example of this tendency by lone females to separate from the central geographic range of the population, particularly when they are young. I suspect that innate preferences for open habitats containing shallow bodies of water play a role in the winter site selection of many of the birds in this population. These birds require relatively fresh water and appear to instinctively prefer a large open view of their surroundings (Armbruster 1990, Lewis 1995). As discussed in the habitat imprinting section, these preferences can be influenced and altered by the choice of a rearing site; however, a bird cannot long ignore its basic energetic and other survival needs. It is worth noting here that the habitat choices made by reintroduced migratory Whooping Cranes and non-migratory Whooping Cranes in Florida are essentially identical, with birds from both populations selecting the open habitats found primarily on cattle ranches (M. Folk, personal communication). The parallel preferences of these two populations of birds in Florida, which were raised under different circumstances and in different environments, appears to indicate that they may simply be choosing those habitats on the Florida landscape that are inherently most suitable for cranes. In fact, due to poor survival rates following releases on publicly owned state wildlife management area, 47

61 managers of the Florida non-migratory flock began to consider release sites on privately held cattle ranches in 1994 (Nesbitt et al. 1997). It is interesting to note here as well that when a small group of birds from the Florida non-migratory flock departed Florida and headed to South Carolina, they utilized the same general areas and habitats as the 2003 eastern migratory cohort birds that winter in that state (M. Folk, personal communication). Most Sandhill Cranes wintering in South Carolina also use the Ace Basin fields and marshes. Here again, then, there appears to be some inherent preference for these communities as suitable crane habitats. 48 DISCUSSION The Whooping Crane reintroduction procedure and the intensive monitoring of individual birds presented a unique opportunity to study the process of habitat selection. I was able to record details of the habitat selection process that have rarely been possible in populations that were not so intensely monitored, and several aspects of the way birds were managed allowed me to test hypotheses that would be challenging in wild populations. Habitat Selection versus Availability Habitat selection by Whooping Cranes in this population was significantly nonrandom at the landscape, home range, and individual location levels for both daytime home ranges and roosting sites. Freshwater marsh, open pasture, wet prairie and open water were among the top four selected habitats for each of the ranking matrices. Saltwater marsh was strongly selected against despite the birds briefly visiting the Chassahowitzka NWR release site or its vicinity under the influence of philopatry. For the two study years, all home ranges

62 contained open pasture, and all but one also contained freshwater marsh and open water. In many cases, the juxtaposition of these habitat types was likely as important as their presence on the landscape. For example, wetlands within an open context were important for safe roosting, as heavier cover surrounding a wetland provided concealment for predators. The unevenness of bird units and home ranges included in the compositional analyses is a weakness in this study, as it creates a pseudo-replication error. This error will be addressed in any future publication by replacing multiple home ranges of a given group or individual with the average composition of these home ranges. This minor correction is unlikely to significantly alter the results of these analyses. This study provides a snapshot of how the birds selected habitat in the initial years of the reintroduction. Future research may compare the patterns established by these birds early in the reintroduction with those emerging since the study years to determine whether their habitat preferences persist with changing reintroduction techniques and locations or whether they adjust significantly under the influence of changing circumstances and environmental factors. 49 Behavior within Cover Types I found that Whooping Cranes in this population utilized various cover types in a supplemental fashion to meet each of their daytime behavioral requirements, particularly in the case of foraging. Though open pasture was used most often for this activity, the birds utilized all 11 cover types in pursuit of food, suggesting that, in the study area, their nutritional requirements cannot be met in a single cover type. Being omnivores, it is not surprising that Whooping Cranes would search for food in varying environs; however, future

63 research might examine food availability across the preferred cover types in Florida to 50 determine whether these birds are adequately meeting their nutritional requirements on their wintering grounds, as this may affect reproductive success upon return to their breeding grounds in the spring. Assessing the body condition of wintering Whooping Cranes prior to spring departure may also help to answer this question. Philopatry In this study, I discovered ample evidence for a philopatric influence on winter home range selection by this population of reintroduced Whooping Cranes. A significant majority of birds revisited their initial winter home range (the CNWR soft-release pen), particularly when exposed to uniform rearing techniques and non-extreme weather conditions. This philopatric tendency, however, was overcome by the consistent rejection of the Chassahowitzka salt marsh as suitable habitat by these Whooping Cranes. A majority of eventual self-selected winter home ranges were located within 100km of Chassahowitzka, however, indicating a persistent philopatric influence on the geographic distribution of Whooping Cranes on the wintering grounds. Based on these findings, I surmise that if the habitat composition of the initial winter home range (the CNWR soft-release pen) had been suitable and extensive we might have expected a pattern of philopatry more similar to that shown by this population on their summer range (Maguire 2008) and by the Wood Buffalo- Aransas flock on the winter range (Stehn 2010), wherein birds first establish territories in the empty spaces between established territories until an area is saturated and territories have shrunk to their smallest viable size, and then establish new territories in the closest suitable available habitat adjacent to the already established territories. This type of selection pattern,

64 though it was the pattern originally anticipated by managers of the reintroduced cranes, poses challenges to the reintroduction and would likely necessitate changes in the winter management and release techniques. In the years following this study, additional winter release sites as well as summer rearing sites have been utilized for this population. Future research should concentrate on whether the patterns seen in the first years of the reintroduction have persisted, and how the selection of new release sites has affected the wintering ecology and management of this population. 51 Site Tenacity As has been found on the breeding grounds (Maguire 2008), site tenacity has a clear influence on the winter home range selection of Whooping Cranes in this population, with varying strength according to gender, age and social status. The tendency of birds to return to areas that they or their social associates occupied in previous years has important implications for habitat conservation and management strategies with regard to the long-term viability of a migratory population of Whooping Cranes wintering in Florida and other southeastern states. As human population growth and associated development pressure continues and Florida s natural and agricultural areas are converted into gated communities and golf courses, the consistency in site tenacity by these birds allows us to identify the areas that may be most important to the population over at least the next several years. Conservation partners such as the Southwest Florida Water Management District and the Natural Resources Conservation Service may want to establish habitat conservation and improvement goals in these strategic areas. An awareness of development plans in areas that

65 are currently occupied by eastern migratory Whooping Cranes also affords project biologists with opportunities for targeted public education to minimize conflicts between humans and Whooping Cranes, including undesirable habituation to human activity and supplemental feeding. 52 Habitat Imprinting In this study, I found weak evidence of habitat imprinting on the winter site selection patterns of Whooping Cranes in this population. Though birds consistently abandoned the Florida salt marsh in which they were soft-released at the end of their first migration, in favor of inland habitats, their use of salt marsh areas outside the direct vicinity of CNWR, especially those in distant South Carolina and North Carolina, indicated an affinity for a habitat type that can most easily be attributed to imprinting on the salt marsh habitat surrounding the CNWR release site. This tendency to briefly visit areas that are similar in structure and composition to the soft-release site has important implications for future management of reintroduced birds. I propose that, if birds are led directly to and softreleased into habitats that meet their biological needs (unlike the salt marshes of Florida s central Gulf Coast), they will imprint on habitat of the type and composition of that area. I would add an important cautionary note here. If birds are deliberately imprinted on areas where predation risk is high (i.e., with high predator densities or ample predator cover, including wooded uplands, forested wetlands or shrublands) but that meet their dietary and fresh water needs, we may be unwittingly exposing them to increased mortality potential, as may have been the case at the Hálpata Tastanaki holding pen site.

66 Social Facilitation and Territoriality 53 At the scale of home range selection within the winter range as well as the use of particular sites within a home range, social factors surfaced as important influences on winter site selection in this population of Whooping Cranes. When they encountered other cranes, Whooping Cranes appear to be influenced by Sandhill Cranes of both the migratory and nonmigratory subspecies as well as migratory and non-migratory Whooping Cranes in choosing where to settle and how to utilize the local resources in a given area. All Whooping Crane home ranges in Florida overlapped with Sandhill Crane home ranges, and Whooping Cranes often sought the company of other Whooping Cranes. In contrast to the Whooping Cranes in the AWBP, winter home ranges of these birds sometimes overlapped, and aggression among social groups in close proximity to one another was low to moderate. It is unclear what the ecological implications of this may be, since the birds are generally widely dispersed across the landscape and saturation of available winter habitat does not seem likely in the near future. However, it is certainly noteworthy that these birds are behaving so differently from wild Whooping Cranes. Perhaps birds will defend winter territories when they are wintering with offspring. The only birds to successfully breed thus far wintered separately from any other Whooping Cranes but did interact with Sandhill Cranes. They were observed unison calling (an indication of defensiveness) when Sandhill Cranes approached them at food sources, but also often tolerated Sandhill Cranes in close proximity to them. There may be a strong advantage in releasing Whooping Cranes in close proximity to other cranes, from which they may learn appropriate predator avoidance behaviors as well as the best places to forage and safely roost. These advantages need to be weighed against any perceived disadvantages of Whooping Cranes associating with local cranes. Florida Sandhill

67 Cranes, for example, are often habituated to human activity and may influence Whooping 54 Cranes to utilize areas in close proximity to human dwellings and to utilize supplemental feed available in those areas. There have been a few instances of this in Florida thus far, and they have been handled on a case-by-case basis to minimize negative impacts on the Whooping Cranes. It is likely unrealistic to believe that, in this day and age, reintroduced Whooping Cranes won t come into conflict with humans or become exposed from time to time to undesirably high levels of human activity. In fact, some level of tolerance to human activity will likely be an advantage to this population, as over-sensitivity to humans is implicated in the original decline in the species, and human encroachment into previously uninhabited areas is likely to continue apace in Florida and beyond (Wenner and Nesbitt 1987, Morrison and Humphrey 2001). This underscores the importance of continued education and partnership with local residents and stakeholders in protecting reintroduced Whooping Cranes and their habitats. I found that birds were commonly influenced by their social associates when selecting wintering home ranges. Because social associations often change several times within the first few years of a crane s life, the choices initially made by one bird can have far-reaching effects in the wintering ecology of a reintroduced population of Whooping Cranes. Therefore, the selection of an alternate winter release site, even for just one year, will likely affect birds from multiple cohorts for many years into the future. This makes the selection of new wintering locations a serious decision for project personnel. The influence of social associates on winter site selection by Whooping Cranes in this population also affords us with an ability to identify key areas for potential habitat protection or improvement as well as environmental education initiatives to benefit Whooping Cranes and other species inhabiting

68 those areas. Barring excessive urban development or climatic shifts, birds will likely 55 continue to influence each other to return to areas of concentrated Whooping Crane activity year after year. As the winter release sites and management techniques are refined over time, it will be critically important to continue to closely monitor the movements and social associations of these birds to determine how the alteration in techniques may be affecting birds abilities to locate suitable winter areas that promote the long-term viability of the population. The potential use of St. Mark s NWR, in northwestern Florida, for example, will need to be evaluated based on the choices made by the birds soft-released there when they return to the wintering grounds in subsequent years. The impact, for example, that the distance of this location from the primary greater Sandhill Crane migration corridor and from the use areas of Florida Sandhill Cranes, as well as previously released migratory Whooping Cranes, will need to be assessed. Environmental Stochasticity When birds deviated from typical or established patterns of wintering behavior, I found that stochastic environmental variables can often provide a plausible explanation for these deviations. In some cases, such as birds that summered in Michigan and wintered in North Carolina after being blown off course on their first northward migration, the consequences of the environmental event can be far-reaching and negatively affect the birds potential contributions to the reintroduced population. In other cases, the effect is benign and simply an interesting departure from expected behavior, as in the case of the 2003 birds that encountered a storm cell on their first northward migration and wintered in South Carolina or eastern Florida.

69 In yet other cases, the birds ability to adjust their movements and behaviors 56 according to environmental conditions is encouraging from the standpoint of long-term population viability. Though there was a significant increase in winter mortalities during the first drought year in Florida ( ), the tally may have been higher had birds not abandoned their previous home ranges in favor of more permanently inundated areas. One possible consequence of these movements is that the more permanent water bodies are inhabited by larger populations of alligators, and one of the mortalities was attributed to an alligator. The following winter, many birds did not complete migration to Florida and instead wintered further north along the migration route. This plasticity in the birds wintering ecology, though it may not promote pair formation on the wintering grounds due to potentially greater dispersal of birds across the landscape, may prove advantageous as weather patterns shift and urban encroachment affects their initially preferred wintering areas. Over the past few years, some birds have wintered at Hiwassee Refuge in Tennessee, and it will be important for project personnel to work with local biologists to handle growing political pressures in the area to alter management on the refuge to decrease its attractiveness to cranes, which are perceived to negatively impact waterfowl hunting opportunities (Aborn, 2010). Innate Preferences In this study, I discovered one case of winter site selection at the scale of geographic distribution that may have been influenced by genetic tendencies toward the wintering range of the wild flock of Whooping Cranes from which all of the genetic material for this population is derived. However, I could not eliminate the possibility that this was simply an

70 extreme case of the greater tendency toward wide dispersal exhibited by young birds in crane populations, which is exacerbated in females. The geographic distribution of cranes in this reintroduced migratory population, therefore, appears to be influenced to the greatest extent by philopatry, site tenacity, and environmental stochasticity. At the scale of home range selection within a given search area, though, innate preferences may play a role. Because the influence of social facilitation is impossible to remove from the equation, it is not possible to determine to what extent innate preferences for open habitats with wetland components may be determining site selection. Habitat imprinting at Necedah NWR may also be influencing choices made at this scale; however, the habitat types in Necedah versus central Florida are different enough that this alone does not appear strong enough to determine where birds will winter in Florida. The Florida nonmigratory flock of Whooping Cranes provides a very useful comparison in this case, because the birds in that flock are not raised in a natural environment in Wisconsin. Despite their very different circumstances in rearing and release, birds from the two populations show strong preferences for the same habitat types in Florida (open pastures, wet prairies, and freshwater marshes located in close proximity to one another). Similarly, birds from both populations that dispersed to South Carolina, where the influence of Sandhill Cranes is essentially removed, utilized the same general areas and habitat types. It appears, then, that though humans can significantly influence the choices made by reintroduced populations of Whooping Cranes, the birds may begin to rely, at least in part, on innate preferences under various conditions. In our case, managers chose a salt marsh release site that does not appear to meet the basic needs of Whooping Cranes. The birds, therefore, are forced to abandon the area despite briefly returning to it under the influence of philopatry. 57

71 Similarly, birds influenced by environmental stochasticity may settle in locations that do not support similar habitat types to those in which they are raised or released. Under these scenarios, birds in this population often made several movements before finally settling on a home range. Therefore, it is possible that they were in search of an area that would meet their needs without possessing any pre-conceived precise search image of what that area might look like, other than it being a fairly open area containing a water source (the case for most observed short-term sites). Site selection, then, becomes a trial-and-error process eventually resulting in a long-term selection of an area that will meet the cranes needs for an extended period of time. The areas that the cranes select under this model, then, should be the areas that best meet cranes needs. 58 ACKNOWLEDGEMENTS I would like to sincerely thank those individuals and organizations that provided support and guidance throughout this project. The International Crane Foundation, the U.S. Fish and Wildlife Service Partners for Fish and Wildlife Program, the Natural Resources Foundation of Wisconsin, and the University of Wisconsin Madison generously provided funding and employment that made this project possible. I am grateful for the support of the Whooping Crane Eastern Partnership and all of the original partner organizations: the Whooping Crane Recovery Team, U.S. Fish and Wildlife Service (USFWS), Wisconsin Department of Natural Resources, International Crane Foundation, U.S. Geological Survey (USGS) Patuxent Wildlife Research Center, Operation Migration, Natural Resources Foundation of Wisconsin, National Fish and Wildlife Foundation, and the USGS National

72 Wildlife Health Center. I am particularly grateful to John Christian and Richard Urbanek of the USFWS for their unwavering support of this project. I would like to also thank Windway Capital Corporation for the use of their aircraft and staff when tracking the cranes by air. I would like to especially thank pilot Michael Voechting for his tirelessness and enthusiasm during long hours tracking cranes on migration. I thank the staff of the Chassahowitzka National Wildlife Refuge for the provision of housing and support during the initial years of the reintroduction. I thank the staff of the Necedah National Wildlife Refuge for use of the tracking vehicle, binoculars, and spotting scope essential for conducting this research. I thank Marty Folk and Marilyn Spalding of the Florida non-migratory Whooping Crane reintroduction team for their advice, assistance, and support. A special thanks to tracking team interns Julia Martinson and Christopher Malachowski, who skillfully gathered much of the data for this research. Mark Nipper (Operation Migration), Sara Zimorski (ICF) Richard Urbanek (USFWS), and Marianne Wellington (ICF) all of the winter monitoring team, also provided important assistance and support throughout the data-gathering phase of this research. I am particularly indebted to the private landowners and managers in Florida who generously provided access to their properties for the monitoring of these Whooping Cranes. Without their cooperation, this research would not have been possible. I also thank the staff of the Lake Woodruff National Wildlife Refuge for providing access to the refuge for monitoring and for sharing their biological expertise. I sincerely thank Dr. Stanley Temple; without his guidance and steadfast support, this project could not have come to fruition. I thank Dr. David Drake, Dr. Anna Pigeon, and Dr. 59

73 John Harrington for serving on my committee. A special thanks to John Cary and Nick 60 Kueler for their technical support. LITERATURE CITED Aborn, D. A Possible Competition between Waterfowl and Sandhill Cranes at Hiwassee Wildlife Refuge, Tennessee Proceedings North American Crane Workshop 11: Aebischer, N. J., P.A. Robertson, and R.E. Kenward Compositional Analysis of Habitat Use From Animal Radio-Tracking Data. Ecology 74: Armbruster, M. J Characterization of Habitat Used by Whooping Cranes During Migration. U.S. Fish & Wildlife Service, Biological Report 90(4). 16 pp. Allen, R. P A Report on Whooping Cranes Northern Breeding Grounds. New Nork, NY: National Audubon Society. Cartwright, R., B. Gillespie, K. LaBonte, T. Mangold, A. Venema, K. Eden, and M. Sullivan Between a Rock and a Hard Place: Habitat selection in female-calf humpback whale (Megaptera novaeangliae) pairs on the Hawaiian breeding grounds. Plos ONE, 7: doi: /journal.pone Chavez-Ramirez, F. and R.D. Slack Differential Use of Coastal Marsh Habitats by Nonbreeding Wading Birds. Colonial Waterbirds 18: Cody, M. L An Introduction to Habitat Selection in Birds. Pages 4-46 in M.L Cody, editor. Habitat Selection in Birds. Academic Press, Inc. Orlando, FL.

74 Dreitz, V.J. and F.L. Knopf Mountain Plovers and the Politics of Research on Private Lands. Bioscience 57: Dunning, J. B., B. J. Danielson, and H. R. Pulliam Ecological Processes that Affect Populations in complex landscapes. Oikos 65: Ellis, D.H., G.F. Gee, S.G. Hereford, G.H. Olsen, T.D. Chisolm, J.M. Nicolich, K.A. Sullivan, N.J. Thomas, M. Nagendran and J.S.Hatfield Post-Release Survival of Hand-Reared and Parent-Reared Mississippi Sandhill Cranes. The Condor 102: Ewel, K.C Swamps. Pages in R.L. Myers and J.J. Ewel, editors. Ecosystems of Florida. University of Central Florida Press, Orlando. Florida Department of Environmental Protection Learning from the Drought: Annual status report on regional water supply planning. 20 pp. Florida Department of Environmental Protection Hixtown Swamp, pages in Florida Forever Annual Report. Retrieved from: Florida Department of Transportation, Surveying and Mapping Office Geographic Mapping Section Florida Landuse Cover and Forms Classification System Handbook. 95 pp. Graham, C Habitat Selection and Activity Budgets of Keel-Billed Toucans at the Landscape Level. The Condor 103: Hildén, O Habitat selection in birds. Annales Zoologici Fennici 2: Hoffman, T Greater Sandhill Crane Central Valley Population Survey Results, Fall North American Crane Working Group. 61

75 Johnsgard, P. A Cranes of the World. Indiana University Press, Bloominton, Indiana. 257 pp. Johnson, D. H The Comparison of Usage and Availability Measures for Evaluating Resource Preference. Ecology 61: Kruse, K. L., J. A. Dubovsky, and T.R. Cooper Status and harvests of Sandhill Cranes: Mid-Continent, Rocky Mountain, Lower Colorado River Valley and Eastern Populations. Administrative Report, U.S. Fish and Wildlife Service, Denver, Colorado. 14pp. Kushlan, J. A Freshwater marshes. Pages in R.L. Myers and J. J. Ewel, editors. Ecosystems of Florida. University of Central Florida Press, Orlando. Lewis, J. C Whooping Crane (Grus americana), in A. Poole, editor. The Birds of North America Online. Cornell Lab of Ornithology; Ithaca, New York. Retrieved from: doi: /bna.153 Maguire, K. J Habitat Selection of Reintroduced Whooping Cranes Grus americana on their breeding range. Unpublished M.Sc.Thesis, University of Wisconsin, Madison, USA. Mapping and GIS Section, Southwest Florida Water Management District Southwest Florida Water Management District Land Use and Cover Retrieved from: Meine, C. D., and G. W. Archibald (Eds) The Cranes: Status survey and conservation action plan. IUCN, Gland, Switzerland, and Cambridge, U.K. 294pp. Melvin, S. M., R. C. Drewien, S. A. Temple, and E. G. Bizeau Leg-Band Attachment of Radio Transmitters for Large Birds. Wildlife Society Bulletin 11:

76 Morrison, J. L., and S. R. Humphrey Conservation Value of Private Lands for Crested Caracaras in Florida. Conservation Biology 15: Nesbitt, S. A., M. J Folk, M. G. Spalding, J. A. Schmidt, S. T. Schwikert, J. M. Nicholich, M. Wellington, J. C. Lewis, and T. H. Logan An Experimental Release of Whooping Cranes in Florida-The first three years. Proceedings North American Crane Workshop 7: Nesbitt, S. A., S. T. Schwikert, and M. J. Folk Natal Dispersal in Florida Sandhill Cranes. Journal of Wildlife Management 66: Palminteri, S., and Peres, C. A Habitat Selection and Use of Space by Bald-Faced Sakis (Pithecia irrorata) in Southwestern Amazonia: Lessons from a multiyear, multigroup study. International Journal of Primatology 33: Pogson, T. H., and S. M. Lindstedt Distribution and Abundance of Large Sandhill Cranes, Grus canadensis, Wintering in California's Central Valley. The Condor 93: Southwest Florida Water Management District Hálpata Tastanaki Preserve. Pages in Recreation Guide to Southwest Florida Water Management District Lands. Sparling, D. W., and G. L. Krapu Communal Roosting and Foraging Behavior of Staging Sandhill Cranes. Wilson Bulletin 106(1): Stehn, T. V., and F. Prieto Changes in Winter Whooping Crane Territories and Range Proceedings North American Crane Workshop 11: Tacha, T.C. P.A.Vohs, G.C. Iverson Time and Energy Budgets of Sandhill Cranes from Mid-Continental North America. The Journal of Wildlife Management 51:

77 Wenner, A. S., and S. A. Nesbitt Wintering of Greater Sandhill Cranes in Florida. 64 Proceedings North American Crane Workshop 4: Whooping Crane Recovery Team, Whooping Crane Recovery Plan. U.S. Fish and Wildlife Service. Region 2, Albuquerque, New Mexico. United States Department of Agriculture, Natural Resources Conservation Service Land Resource Regions and Major Land Resource Areas of the United States, the Caribbean, and the Pacific Basin. U.S. Department of Agriculture Handbook 296. Urbanek, R. P., J. W. Duff, S. R. Swengel, and L. E. A. Fondow Reintroduction Techniques: Post-release performance of Sandhill Cranes (1) released into wild flocks and (2) led on migration by ultralight aircraft. Proceedings North American Crane Workshop 9: Urbanek, R. P., L. E. A. Fondow, S. E. Zimorski, M. A. Wellington, and M. A. Nipper Winter Release and Management of Reintroduced Migratory Whooping Cranes (Grus Americana). Bird Conservation International 20:43 54.

78 Figures and Tables 65 Figure 1. Minimum convex polygon drawn around all home ranges in Florida during the 2 study years.

79 Figure 2. Whooping Crane home ranges and the major land resource areas in Florida. 66

80 67 67 Figure 3. Florida statewide average annual precipitation from 1895 through Source: Florida DEP, 2008.

81 68 68 Figure 4. Example of core (red line) and 95% (purple line) home ranges intersected with Florida landuse/landcover types.

82 69 69 Figure 5. Example of buffered minimum convex polygon drawn around locations where a bird landed within a search area.

83 70 Wet Prairies Open Water Freshwater Marshes Wooded Upland Open Pasture Forested Wetland Agricultural Crops Shrub/Scrub Streams, Waterways, Springs & Sloughs Other Saltwater Marshes Urban/Residential Figure 6. Percentage of observations of Whooping Cranes in each cover type.

84 71 71 Foraging Locations by Cover Type 60.00% 50.00% 40.00% % Locations 30.00% 20.00% 10.00% 0.00% Agricultural Crops Freshwater Marshes Open Pasture and Other Rural Open Lands Open Water Other Saltwater Marshes, Bays or Estuaries Shrub/Brush or Mixed Rangeland Stream, Waterways, Springs & Sloughs Cover Type Urban/Residential Wet Prairies Wooded or Forested Upland Figure 7. Foraging locations by cover type.

85 72 72 Walking Locomotion Locations by Cover Type 60.00% 50.00% 40.00% % Locations 30.00% 20.00% 10.00% 0.00% Forested Wetland Freshwater Marshes Open Pasture and Other Rural Open Lands Open Water Wet Prairies Wooded or Forested Upland Cover Type Figure 8. Walking locomotion locations by cover type.

86 73 73 Loafing Locations by Cover Type 45.00% 40.00% 35.00% 30.00% % Locations 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% Agricultural Crops Forested Wetland Freshwater Marshes Open Pasture and Other Rural Open Lands Open Water Cover Type Shrub/Brush or Mixed Rangeland Stream, Waterways, Springs & Sloughs Wet Prairies Wooded or Forested Upland Figure 9. Loafing locations by cover type.

87 74 74 Roost Locations By Cover Type 35.00% 30.00% 25.00% % Locations 20.00% 15.00% 10.00% 5.00% 0.00% Forested Wetland Freshwater Marshes Open Water Wet Prairies Cover Type Figure 10. Roosting locations by cover type.

88 Figure 11. Map of all winter home ranges, Filled pink stars denote Chassahowitzka NWR soft release pen site in Florida; and Necedah NWR in Wisconsin. Outlined stars are Hiwassee Refuge in Tennessee and Jasper-Pulaski Wildlife Management Area in Indiana. 75

89 76 76 Distribution of Home Range Distances from CNWR 60% 50% 40% % of Home Ranges 30% 20% 10% 0% Distance Interval (km) Figure 12. Distribution of winter home range distances from Chassahowitzka NWR soft-release pen site.

90 77 77 Home Range Distribution % of Home Ranges 60% 50% 40% 30% 20% 10% 0% 1 Category Within 100km (excepting PP) Misc FL Short Stop Hixtown Misc Carolina/Georgia Paynes Prairie Ace Basin LA Figure 13. Winter home range distribution by special category. Hixtown Swamp and Paynes Prairie are major greater Sandhill Crane wintering areas in Florida.

91 Figure 14. Map showing overlap in Whooping Crane winter home ranges (yellow dots) and Florida Sandhill Crane distribution (green polygons). 78

92 Figure 15. Overlapping minimum convex polygon home ranges on the Pruitt/Barthle Ranch area in Pasco County, Florida in winter (study year 2). 79