Citation
Nesting Habitat Selection and Habitat Associations of Juvenile Jabiru Storks (Jabiru mycteria) in Belize, Central America: Implications for Conservation

Material Information

Title:
Nesting Habitat Selection and Habitat Associations of Juvenile Jabiru Storks (Jabiru mycteria) in Belize, Central America: Implications for Conservation
Creator:
FIGUEROA, OMAR ANTONIO ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Animal nesting ( jstor )
Bird nesting ( jstor )
Coniferous forests ( jstor )
Forest ecosystems ( jstor )
Forests ( jstor )
Lowland forests ( jstor )
Lowlands ( jstor )
Mangrove forests ( jstor )
Nesting sites ( jstor )
Storks ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Omar Antonio Figueroa. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
12/31/2006
Resource Identifier:
495638403 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

NESTING HABITAT SELECTION AND HABITAT ASSOCIATIONS OF JUVENILE JABIRU STORKS ( Jabiru mycteria ) IN BELIZE, CENTRAL AMERICA: IMPLICATIONS FOR CONSERVATION By OMAR ANTONIO FIGUEROA A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Omar Antonio Figueroa

PAGE 3

To Mary Elizabeth Figueroa.

PAGE 4

ACKNOWLEDGMENTS This research benefited tremendously from many individuals and organizations. First and foremost, my graduate advisor, Ken Meyer, provided encouragement, support, critical advice and consistent guidance. This project would probably still be conceptual if not for his partnership. My committee members, Peter Frederick and Scott Robinson, provided thoughtful comments. My study program at the University of Florida was funded by a Fulbright/Organization of American States Ecology Initiative Fellowship. Additional funding sources included a Compton Fellowship (Program for Studies in Tropical Conservation), a Jennings scholarship, the Protected Areas Conservation Trust, the Felburn Foundation, the Disney Wildlife Conservation Fund, the Wildlife Conservation Society, New Orleans Audubon Zoo, a Lewis A. Tyler Award and the Columbus Zoo Conservation Program. This generous support made my research possible. Several organizations in Belize provided crucial collaborative support. I would like to thank the Belize Forest Department for providing the necessary research permits, the Belize Zoo for allowing me to work on captive Jabiru storks and the Tropical Education Center for accommodations. The Belize Audubon Society, Programme for Belize and the Belize Forest Department allowed me to work on their managed lands. I thank Frank Plett and David Dueck for providing safe and reliable aerial surveys and Marcus Cucul for assistance with climbing and capture of juvenile Jabiru storks. iv

PAGE 5

I am indebted to many individuals for contribution too diverse and numerous to mention. I thank Osmany Salas, Sharon Matola, Tony Garel, Bruce and Carolyn Miller, Mark Myers, Marcelo Windsor, Hector Mai, John Pinelo, Valdemar Andrade, Richard and Carol Foster, Alejandro Paredes, Roberto Pott, Victor Alegria, Celso Pott, Miguel Choco, Mario Teul, Reynold Cal, Erneldo Bustamante, Norman Martinez, Joe Awe and Jonathan Urbina. I thank Wilber Martinez, an outstanding field companion, a genuine friend and a great field biologist who provided much needed assistance and encouragement during critical periods. His unwavering and selfless field assistance was greatly appreciated. I thank my wife Desi and daughter Rhiannon for their unrelenting support and understanding, especially during my prolonged absence in the past few months. I thank them for the lunch packs, warm meals and steadfast encouragement. I thank my brother and sisters, and their families for their support and for keeping in touch during challenging times. I am grateful to my parents, Arsenio and Jeanette, for things too numerous to mention. v

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES ...........................................................................................................viii LIST OF FIGURES .............................................................................................................x ABSTRACT ......................................................................................................................xii CHAPTER 1 INTRODUCTION........................................................................................................1 Regional Overview.......................................................................................................1 Habitat Selection...........................................................................................................2 The Study Species: The Jabiru Stork...........................................................................3 Distribution in Mesoamerica........................................................................................4 2 METHODS...................................................................................................................7 Study Area....................................................................................................................7 Nest Searching..............................................................................................................7 Microhabitat Vegetation Measurements.......................................................................9 Landscape-level Habitat Analysis..............................................................................10 Landscape-level Spatial Analysis...............................................................................10 Nest Site Selection......................................................................................................12 Movement...................................................................................................................15 Habitat Use Versus Available.....................................................................................16 Home-range Analysis.................................................................................................19 National and Regional Queries...................................................................................19 3 RESULTS...................................................................................................................25 Nest Site Selection......................................................................................................25 Microhabitat and Landscape Measurements..............................................................25 Nest Sites and Ecosystems..........................................................................................28 Habitat Use Versus Availability.................................................................................29 Nest Sites.............................................................................................................29 Post Fledging Movement.....................................................................................29 vi

PAGE 7

Habitat Associations of Juvenile Storks..............................................................30 Nest Sites by Protected Areas, Watershed Regions and Political Boundaries...........31 Home Range, Protected Areas and Watershed Regions.............................................32 National and Regional Queries...................................................................................32 4 DISCUSSION.............................................................................................................57 Movement...................................................................................................................63 Conservation Implications..........................................................................................64 Directions for Future Study........................................................................................66 5 CONCLUSION...........................................................................................................69 LIST OF REFERENCES...................................................................................................73 BIOGRAPHICAL SKETCH.............................................................................................78 vii

PAGE 8

LIST OF TABLES Table page 1 Broad scale ecosystem classes for Belize................................................................24 2 Comparisons of microhabitat variables measured for nest trees and randomly selected over story trees, and comparisons for landscape measurements at nest and random sites.......................................................................................................35 3 Species and relative percentage of eighty randomly selected unused (over story) trees within 180 meters of 22 active Jabiru stork nests............................................37 4 Principal component analysis (PCA) of variation along landscape variables measured for random sites (a) and nest sites (b)......................................................38 5 Principal component analysis (PCA) of variation along microhabitat variables for nest trees (a) and random trees (b)......................................................................38 6 Comparisons of ecosystems within which Jabiru storks built nests and the associated nearest ecotone with those of randomly selected points.........................42 7 Relative frequencies of ecotones associated with Jabiru stork nest sites and random sites..............................................................................................................43 8 Comparisons of habitat use (within 3000 m buffers of nest sites) to the proportions of available habitat within the 3000 m buffered MCP polygon for all nest sites...................................................................................................................45 9 Results of Chi square analyses showing that juvenile Jabiru storks used tropical lowland tall herbaceous swamps and short grass savanna with shrubs more than expected....................................................................................................................47 10 Results of the spatial cluster analysis procedure used to test the hypothesis that the activity pattern displayed by the tagged Jabiru storks could be the result of random chance..........................................................................................................47 11 Comparison of the cumulative proportion of habitat within 250 meter buffers of all Jabiru stork locations to the proportions of available habitat within the MCP home range and to the overall proportions available in Belize................................47 viii

PAGE 9

12 Comparison of the cumulative proportion of habitat along original Jabiru stork routes to the proportions of habitat available along randomly distributed routes within the MCP and along randomly distributed routes throughout Belize.............48 13 Distribution and relative percentages of critical Jabiru stork habitat within all seven countries of Central America and the proportions of all Central America Jabiru stork habitat represented by each country.....................................................51 ix

PAGE 10

LIST OF FIGURES Figure page 1 Location map of Belize and the wider Central American region...............................6 2 Map of northern Belize showing the general watershed regions surveyed for Jabiru stork nests......................................................................................................21 3 Map of central Belize showing the general watershed regions surveyed for Jabiru stork nests......................................................................................................22 4 Map of southern Belize showing the general watershed regions surveyed for Jabiru stork nests......................................................................................................23 5 Distribution of Jabiru stork nests by political division (district) in Belize, 2005....36 6 Dendrogram model based on CART analysis for predicting Jabiru stork nest sites (present)............................................................................................................39 7 The cross-validation relative error plot (cp plot) used to pick the tree size based on the 1 standard error rule and a smoothing parameter of 100...............................39 8 Distribution of correct classification rates (CCR) for 100 randomly generated trees versus the CCR of the model (triangle)...........................................................40 9 Dendrogram model based on CART analysis for comparing Jabiru stork nest sites (Present) with historical records (Historical)...................................................40 10 The cross-validation relative error plot (cp plot) used to pick the tree size based on the one standard error rule and a smoothing parameter if 100............................41 11 Distribution of correct classification rates (CCRs) for 100 randomly generated trees versus the CCR of the model (triangle)...........................................................41 12 Comparison of the proportions of vegetation composition within concentric buffers for nest (n=22) and random (n=22) sites......................................................44 13 Distribution of tracking locations for juvenile Jabiru storks from four nests in Belize. Point locations derived from GPS receivers incorporated in satellite transmitters carried by each bird..............................................................................46 x

PAGE 11

14 Cumulative vegetation proportions within 50, 75 and 95 percent fixed kernel probability polygons.................................................................................................49 15 Distribution of 22 Jabiru stork nest sites (2005) and 25 historical nest records relative to the protected areas network in Belize.....................................................50 16 Estimates of the activity ranges for juvenile Jabiru storks from four nests in Belize........................................................................................................................51 17 Distribution and abundance of critical Jabiru stork habitat throughout Central America....................................................................................................................52 18 Relative distribution of critical Jabiru stork habitat, 2005 nest sites (n = 22) and historical nest records (n = 25) in Belize..................................................................53 19 Map of Nicaragua showing the distribution of critical Jabiru stork habitat.............54 20 Map of Honduras showing the distribution of critical Jabiru stork habitat..............55 21 Map of Costa Rica showing the distribution of critical Jabiru stork habitat............56 22 Map of Belize showing the existing network of protected areas and the critical conservation areas for Jabiru storks.........................................................................68 xi

PAGE 12

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science NESTING HABITAT SELECTION AND HABITAT ASSOCIATIONS OF JUVENILE JABIRU STORKS (Jabiru mycteria) IN BELIZE, CENTRAL AMERICA: IMPLICATIONS FOR CONSERVATION By Omar Antonio Figueroa December 2005 Chair: Kenneth D. Meyer Major Department: Wildlife Ecology and Conservation Conservation measures often focus on emblematic or flagship species that are good indicators of the integrity and quality of habitat. Large species can often serve as good indicators because their large home ranges and spatial needs make them exceptionally vulnerable to habitat fragmentation and degradation. However, information on species-specific ecological needs is clearly lacking for many large species in Mesoamerica and this often limits effective conservation planning. The Jabiru stork is the largest bird in Mesoamerica and has been categorized as locally endangered by most countries in this region. Critical forest and wetland habitats are being lost or degraded at an alarming pace and very limited information is available on essential nesting and foraging habitats. The underlying goal of this study was to characterize habitat requirements and identify conservation measures for the local breeding population of Jabiru storks in Belize. xii

PAGE 13

I located 22 Jabiru stork nests in Belize from January 2004 to May 2004 and in April and May 2005. By comparing nest sites with randomly chosen unused sites, I determined which features of the nest tree, the surrounding vegetation, and the adjacent landscape influenced nest-site selection. Nests were constructed on either Ceiba pentandra (17) or Pinus caribbea (5). The nest trees had a larger dbh and their lowest branch was significantly higher than those of randomly selected over story trees. Compared with random sites, nest sites were at lower elevations, were nearer to other nests and were nearer to ecotones. Within a total available area of 4,292 km 2 , nesting Jabiru storks selected sites with significantly higher proportions of water, tropical lowland tall herbaceous swamps and short-grass savanna with needle leaf open forest. I used satellite transmitters equipped with GPS capabilities to explore post-fledging habitat associations and to estimate home range sizes of juvenile storks. I evaluated the spatial requirements of breeding territories and characterized habitat composition within breeding and foraging sites. One of the most important conservation needs for this species appears to be management and protection of critical breeding and foraging territories. A regional query of Central America revealed that these critical habitats are small in total area and concentrated in three distinct regions. Comparative ecological assessments in home range sizes, nesting habitat selection and foraging behavior for these three subpopulations are necessary. There is an urgent need for a comprehensive conservation strategy due to the very small population size and highly vulnerable status, especially in light of accelerating pressures to develop the Jabiru stork’s critical habitats, including a large portion of the nest sites. xiii

PAGE 14

CHAPTER 1 INTRODUCTION Regional Overview Belize is a relatively small country confined to the Caribbean coast of northern Central America (Figure 1). It is the second smallest country in the region (<23,000 square kilometers) and, with a population of 282,600 (CSO 2004), one of the least densely populated in Latin America. This relatively small population size has, until recently, partially confined anthropogenic disturbance to discrete sectors and allowed Belize to maintain relatively large and undisturbed areas of natural systems. In 1995, Belize ranked among the top four countries in the Neotropical realm in the proportion of total land mass (>.84) under some form of forest cover (FAO 1997). However, with an annual deforestation rate of 2.13 % from 1990-1995, the wider Mesoamerican region was third among the thirteen tropical-forest biodiversity hotspots (Brooks et al. 2002), equating to a 10% loss of forest cover in that period. Recent estimates place the total forest cover for Belize (including shrublands) at just over 69% (De Vries et al. 2003), with agriculture and urbanization growing at a combined rate of 93 square kilometers per annum (Ek 2004). With suitable wildlife habitat declining at an alarming pace throughout the region, Belize can play an increasingly important role in securing the viability of nationally, regionally and globally threatened flora and fauna. In order to reverse or abate the current trend, however, managers, policy makers and conservation planners must receive sound biological recommendations so that appropriate actions can be taken before the critical period between deforestation and extirpation/extinction 1

PAGE 15

2 expires (Brooks et al. 1999). While reversing the trend of development and population size may not be possible, it is imperative that sustainable decisions be made in the face of these growing pressures. Habitat Selection One of the fundamental endeavors of ecology is to describe how animals use habitat and the associated forces that determine how and why certain habitat types are preferred. Studies of habitat selection by birds span the twentieth century, and have evolved from correlative analyses of habitat characteristics and species abundance to models of density dependence, distribution, and to studies showing the importance of landscape structure and function (Jones 2001). Because of their specific role in guiding habitat management and conservation planning, habitat-selection studies are urgently needed in the Mesoamerican region and should be considered high priority. The main objective of this research was to characterize habitat requirements and identify conservation measures for Jabiru storks in Mesoamerica by studying the local breeding population in Belize. By examining one of the healthiest populations in the region, this study could then serve as a framework for research and conservation planning throughout Mesoamerica. The threatened status, unique physical features, and charismatic nature of the Jabiru stork make it an ideal species on which to focus research and protection efforts. My research had two principal components. First, despite the obvious significance of adequate nesting territories, there has not been a comprehensive attempt to characterize the Jabiru stork’s breeding habitat. I quantified breeding habitat preferences and determined some key habitat features that were associated with nest-sites. Because the nest-site selection process may result from factors operating at different spatial scales

PAGE 16

3 (Penteriani et al. 2001), I studied this behavior at the level of both nest microhabitat and landscape (within an area of 4,292 km2). Secondly, I studied post-fledging movements and analyzed the habitat features that were selected for by juveniles. Thus I present a quantitative description of nesting sites, and their surrounding landscape, and post-breeding habitat used by juveniles, and develop conservation recommendations that may guide protection of the critical habitats essential to this regionally endangered species. The Study Species: The Jabiru Stork The Jabiru stork (Jabiru mycteria, Lichtenstein 1819) is one of the largest storks in the world and, with the second longest wingspan of any new world species, one of the largest birds in the world (Hancock et al. 1992). Its geographic distribution lies within the Neotropic region, extending from southern Mexico through Central America and Northern South America to Northern Argentina and Uruguay (Sibley and Monroe 1990, Del Hoyo et al. 1992). Three distinct populations may occur within this relatively large expanse (Luthin 1987), with the smallest and perhaps most vulnerable scattered over disjunct areas of Mesoamerica (southern Mexico through Panama) (Luthin 1984). The Jabiru stork and the Wood Stork (Mycteria americana) are the only stork species found in Mesoamerica. Adult Jabiru storks stand at a height of 122-140 cm and have a wingspan of 230-260 cm (Del Hoyo et al. 1992). Adults are conspicuous and unmistakable with their white plumage, massive black bill (30.5-33 cm), and broad red band around the base of the neck (Howell and Webb 1995). Juveniles have a less recurved, shorter, and less pointed bill, a downy neck in the immediate area surrounding the red band, a downy head, and body feathers edged grayish-brown.

PAGE 17

4 Distribution in Mesoamerica Jabiru storks are widely scattered throughout Mesoamerica and nowhere dense. In Mexico, they have been observed as breeders in the Usumacinta drainage (Corea and Luthin 1988, Ogden et al. 1988, Hartasanchez 1992) and the eastern Yucatan peninsula (Lopez-Ornat et al. 1989). In Guatemala, a few solitary individuals have been reported for the northern third of the country (Luthin 1984), at Laguna del Tigre in April 2003, and in marshy savannas along the main highway about 10 km west of the Belize-Guatemala border (pers. obs.). Historic reports also indicate isolated occurrences along the pacific coast of Guatemala (Salvin and Goodman 1901, Deignan 1933, Tashian 1953). There were two historic reports for El Salvador at Laguna de Olomega (Dickey and Van Rossem 1938), but the species may have been extirpated from this country (Komar 1998). Jabiru storks have been reported along the pacific coast of Honduras at Laguna El Jicarito 75 km east of Laguna de Olomega, and along the pacific coastal wetlands of Nicaragua (Villalobos 1995). Villalobos (1995) also mentioned two separate sightings of one individual each and Monroe (1968) made other observations along the Atlantic coast of Honduras. Jabiru storks have also been documented along the Caribbean coastal wetlands of Nicaragua (Howell 1972, Camacho 1983). However, not until recently were the Honduran Mosquitia and the Nicaraguan Miskito coast confirmed as important to the regional population of Jabiru storks (Frederick et al. 1997). The presence of Jabiru storks in Costa Rica and the importance of the Tempisque basin have been well documented (Slud 1964, Stiles and Skutch 1991, Villareal 1995, 1997, 1998, Poveda 2003). Northern Costa Rica, in fact, may now be the southern limit of the Mesoamerican population. Two records exist for Panama, one confirmed at Bocas del Toro (1927) and another unconfirmed sighting on the La Jagua Marshes (Wetmore

PAGE 18

5 1965), but the species may no longer exist in this country (Ridgley and Gwynne 1989). In Belize, the Jabiru stork is known from several documentations (Russel 1964, Weyer in Scott and Carbonel 1986, Miller and Mill er 1995, 1998, Paredes 2004, Barnhill et al. 2005) and it has been suggeste d that Belize may hold one of the healthiest populations in the region (Luthin 1984, Luthin 1987). Additional notes on the occurrence and distribution of the Jabiru stork can also be found in several popular bird guides for the region and from Scott and Carbonell’s (1986) IWRB Directory of Neotropical Wetlands. Taken together, the literature on Jabiru storks in Mesoamerica is consistent in portraying a small, widely scattered, and probably cr itically endangered local population with major gaps in the understanding of basic biology, limiting factor s, and conservation needs.

PAGE 19

6 Figure 1. Location map of Belize and the wider Central American region. Adapted from ESRI: http://www.esri.com/data/download .

PAGE 20

CHAPTER 2 METHODS Study Area I considered all of Belize in my review of historic nest-site descriptions and recent sightings to determine where to focus my nest-searching efforts. Historic records consisted of archived reports in the Belize Biodiversity Information System (BBIS) and unpublished documents obtained from the Conservation Division, Belize Forest Department. I relied heavily on my own previous knowledge of Jabiru storks based on four years of coordinating national wetland surveys and two seasons of monitoring Jabiru stork nests. Sighting information was also solicited from tour guides, the Belize Birding Group (a local group with extensive knowledge of Belizean avifauna) and two requests were published in the newsletter of the Belize Audubon Society. I used sighting information to focus my efforts and to ensure a thorough sampling effort. Nest Searching I conducted low altitude aerial surveys from a Cessna 172 single-engine high-wing airplane from January through May 2004 and again in April and May 2005. Surveys were flown at 60 to 75 meters above ground level and at an air speed of 120 km/hr. Survey flights were conducted in three distinct regions of Belize: northern (Figure 2), central (Figure 3) and southern (Figure 4). Because preliminary work had indicated that specific watersheds were critical for Jabiru storks, I used these regions as a guide for planning the aerial surveys. The southern region consisted primarily of the Aguacaliente Wildlife Sanctuary, the Sarstoon-Temash wetlands, the Moho watershed, and the coastal 7

PAGE 21

8 wetlands south of Dangriga Town. The central region consisted of the Sibun and Manatee watersheds but also included the entire Belize River watershed east of Burrel Boom village. The northern region consisted of northern Belize north of Burrel Boom village and east of the Yalbac Mountains. For the northern surveys, particular attention was given to the Belize River watershed, the New River watershed, and the northern coastal wetlands and their associated broadleaf forests. In order to focus on the high probability areas within these large regions, I used historical nesting records, known breeding territories (pers. obs), recent and historical observations of storks during the breeding season, the occurrence of suitable vegetation, and my knowledge of the distribution and extent of suitable wetlands and nesting habitats. Additionally, I conducted a series of interviews in communities located near known and potential stork nesting habitats and with employees of four shrimp farms where Jabiru storks were known to forage. I sought information regarding numbers of storks frequenting the farms, temporal patterns in use, and any observation that would suggest a nearby breeding territory (e.g. storks carrying nesting material or performing courtship displays). Approximately 50 to 60 % of the northern and central survey regions were high probability areas. Within these areas, I generated north-south transects, two to four kilometers apart, and then used these as a guide for the flight routes. In southern Belize, 50 to 60 % of the Moho, Sarstoon and Temash watersheds were considered high probability areas. Again, I generated north-south transects separated by two to four kilometers and used these as a guide for the aerial surveys. For the remaining watershed regions in southern Belize (Figure 4), surveys were confined to within 10 km of the coastline resulting in about 70 % of these regions not being surveyed. I conducted

PAGE 22

9 surveys along a 500 m buffered the Belize, Sibun, Moho, Temash, Manatee and New Rivers, and again within a 500 m buffer of the New River, Northern, Southern and Crooked Tree lagoons. An additional 20 % of the northern and central watershed regions were surveyed using less intensive and more dispersed flight routes, and by opportunistic observations when traveling to and from target sites. The total survey time for this study was approximately 103 hours. As soon as a nest was identified, coordinates were recorded using a Garmin 12 XL handheld Global Positioning System (GPS) receiver and marked on a 1:50,000 topographic map. The species of nest tree was noted and the direction and distance to nearby landmarks (river or other body of water, village or other settlement) were mapped to aid in locating the nest from the ground. Microhabitat Vegetation Measurements Seven microhabitat variables were estimated on the ground at all nest sites. I used a Suunto clinometer and a fifty-meter open reel fiberglass tape to estimate nest height, tree height, height of lowest branch and height of nest branch to the nearest meter. The exact nest height was measured for four nests by lowering a rope from the bottom of the nest to the ground and then measuring the marked length of the rope. The fifty-meter tape was used to measure the horizontal distance from the tree trunk to the nest and the widest crown diameter. A metric diameter tape was used to measure the diameter at breast height (dbh) to the nearest cm. For each nest, the closest overstory tree in five random directions was selected and four microhabitat variables were measured for each of the five trees: tree height, height of lowest branch, crown diameter and dbh. A random direction was selected by sorting a number between 0 and 360, and then using this number as the angular bearing

PAGE 23

10 from the nest tree. All nest and random trees were identified to species. A leaf sample was collected from all unknown trees and submitted to the Belize National Herbarium for identification. Distance and direction from the nest tree to each random tree was measured using a measuring tape and a compass, respectively. Landscape-level Habitat Analysis I used the Central American Ecosystems Mapping Project (CAEMP) 1 , which characterizes 85 terrestrial ecosystem classes for Belize. These fine-scale classes were further grouped by the CAEMP into eleven broad scale ecosystem classes (including Land Use category) (Table 1). I used these two sets of ecosystem classes as the basis for my habitat analysis. The Central American Ecosystem Mapping Project was commissioned by the World Bank in cooperation with the Central American Commission for Environment and Development (CCAD) and all seven countries in the Central American isthmus. All ecosystem files (GIS compatible) are available from the World Bank (http://wbln0018.worldbank.org/MesoAm/UmbpubHP.nsf/) or the USGS-EROS web site (http://mitchnts1.cr.usgs.gov/data/otheragency.html). Use of this classification scheme as the basis for all habitat analyses ensured the regional compatibility of the resulting management and conservation recommendations. Landscape-level Spatial Analysis I used the Data Management Tools of ArcToolbox (ArcGIS 8.3) to define the spatial reference for all digital layers and then projected all files to the North American 1 The Central American Ecosystem Mapping project is part of larger project commissioned by the World Bank and the Government of Netherlands in cooperation with the Central American commission for Environment and Development (CCAD). The primary objective was to generate a uniform ecosystem map for the seven countries of the region.

PAGE 24

11 Datum 1927 (NAD 27), Universal Transverse Mercator Zone 16 (UTM Zone 16). All Geographic Information System (GIS) analyses were conducted in this projection. Using ArcGIS 8.3 and the Central American Ecosystem layer, I estimated vegetation composition within concentric buffers with radii of 500, 1500, 3000 and 5000 meters around each nest. I used ArcGIS to estimate distance to nearest road, river, nest, and settlement, and the elevation (meters above sea level) at each nest. Using the Random Point Generator v. 1.3 extension in ArcView GIS 3.2 (Jenness 2005), I selected 22 random points within a polygon covering mainland Belize with the exclusion of all areas definitely not suitable for Jabiru storks, primarily the high altitude and densely forested areas of the southern Maya Mountains. All landscape measurements previously described were duplicated at these 22 random points. Records of 25 historical nest sites were obtained from the Belize Biodiversity Information System (BBIS) (Belize Audubon Society). Wildlife Conservation Society (via Bruce miller) obtained these approximate locations directly from Ford Young (deceased), who conducted country-wide aerial surveys in the late 1960s through mid 1980s. Young had marked these nest locations on a 1:250,000 topographic map, which I obtained. I digitized and geo-referenced this document using a scanner and ArcGIS 8.3, thus estimating approximate coordinates for all of the historical nest records. I also produced GIS-compatible files for these locations for use in subsequent spatial analyses. This file was then reviewed and two questionable records deleted. These two records were deleted because of their coordinates corresponded to high elevation values and densely forested montane habitats. The landscape analyses comparing recent nest and random sites were then duplicated for the approximate locations of the historical sites

PAGE 25

12 using the same set of randomly generated points. To minimize errors associated with the broad scale map used to document the historical records, only buffers of 3,000 and 5,000 m were used to analyze vegetation composition for the historical locations. All distance measurements were obtained by joining the two layers within ArcMAP. For example, to determine the distance between all nest sites and the nearest river, both layers were imported into ArcMAP and joined. The rules of the join were clearly defined to specify the distance from all points to the nearest polyline theme (in this case, river). The output file generated a unique table that specified the shortest distance between all nest locations and the nearest river. This method is consistent across nests and minimizes errors such as those associated with manual estimation using the ruler function of ArcMAP. All nest locations were encircled with concentric buffers with radii of 500, 1,500, 3,000 and 5,000 meters, overlaid on an ecosystem layer, and then clipped to determine the vegetation proportions within each buffer. To ensure that all resultant files were adjusted to include only the habitat composition within the specified buffers, four Personal Geo-databases (a relational database that stores spatial data) were created and used to store the clipped files for each of the four clipping parameters (i.e., one geodatabase for the 500 meter clips, one for the 1,500 meter clips etc.). The data were summarized and sorted in Microsoft Access to facilitate analysis. The same clipping procedure was used to determine vegetation proportions within concentric buffers for random and historical nesting sites. Nest Site Selection I obtained mean, range, minimum and maximum values, standard errors and standard deviations of measurements for all nest trees, random overstory trees and

PAGE 26

13 landscape measurements. Before performing univariate comparisons, habitat variables were tested for normality using the Kolmogorov-Smirnov test (Massey, 1951) and by visual inspection of normal quantile-quantile (Q-Q) plots (Rosenkrantz 2000). The Wilcoxson’s signed rank test was used to compare nest trees and random overstory trees for the four variables measured for both groups (Shiraki 1994, Dykstra et. al. 2000). I used Mann-Whitney U-tests (Siegel 1956) to evaluate landscape differences between nest and random sites (Shiraki 1994) and a Principal Component Analysis (PCA) to partition total variability of the data into orthogonal components and determine which habitat variables accounted for most of the variation. Principal components were then examined to determine whether these variables remained the same when nest and random overstory trees were compared at the microhabitat level and when nest and random points were compared at the landscape level. In order to determine the underlying habitat components that explained the differences between nest and random sites, I constructed classification trees using the Cartware/rpart software package in R, an open-source statistical package and programming language ( www.r-project.org ). Classification trees are constructed by a recursive partitioning of the data matrix such that the resulting groups (tree branches in the dendrogram) within each partition become increasingly more homogenous. The resulting tree presents easily understood and interpreted information regarding the underlying structure and predictive capability of the data. This method is especially vigorous since, in addition to providing a classification structure, it also provides an estimate of the misclassification probability for each group along each terminal branch. However, since a fullygrown tree has limited predictive capability (it is an exact

PAGE 27

14 representation of the data), the tree was simplified by a pruning (terminal branches removed) procedure which recombined subgroups when classification errors were not significantly increased. The decision to prune the tree at a given size was based on the V-fold cross-validation method, which provides error estimates for trees of a given size. In this procedure, the data set is divided into ten equal parts, the model derived with nine parts (learning set) and tested with one part (the validation set). The cross-validation is repeated ten times and the results are combined to develop the predictive accuracy and error rates for the tree. In this case, I used the smallest tree for which the estimated error was within one standard error of the minimum. The monte.cart function was used to obtain a P-value for the tree based on Monte Carlo re-sampling. One hundred trees were created by random permutations of the data. I then compared the correct classification rate (CCR) with the distribution of CCRs from the one hundred randomly generated trees. I conducted this analysis to evaluate the variables that best separate the randomly generated and actual trees. I then repeated the analysis for the historical nest records to evaluate whether the present condition of these historical sites were still potential nesting territories and to see what, if any, habitat variables best distinguished the two groups. Because many of the historical records were located in the same general area as some present nest sites, I expected strong similarities between these two groups and therefore a resulting model with lower predictive power (higher P value). Finally, I used discriminant analysis (Stevens 1986) to evaluate which characteristics contributed the most to differences between the two groups (Seamans and Gutierez 1995) and to cross-check how well the variables of the CART predictive model separated the two groups in this procedure. The SPSS 11.5 software package produces an

PAGE 28

15 output table that lists Wilks’ lambda values, an F statistic, and a significance value indicating how well the variables function as discriminators. Movement Six solar-powered satellite transmitters (PTTs) equipped with a GPS receiver function were attached to nestling Jabiru storks two to three weeks prior to fledging. Three juvenile storks of similar age had previously been fitted with satellite transmitters in South America (P. de Tarzo Zuquim Antas, pers. comm.). At this age, the birds have attained adult size but are not yet capable of flying. Three nests in living trees (four individuals) were accessed by a climber using mechanical ascenders on a rope secured over a strong limb. One nest (two individuals) in a dead tree was accessed by assembling a steel scaffolding tower adjacent to the nest tree. Once at the nest, a six foot pvc pole with an adjustable rope loop at one end was used to capture the nestling. The climber blind-folded the bird, secured the bill with a non-adhesive elastic tape, carefully placed it inside a cotton sack, and slowly lowered it to the ground with a rope. I measured tarsus, culmen (distal edge of nostril to tip) and weight and drew a small blood sample from the brachial vein. Two drops of blood were smeared on an identification card and used to determine sex. The remainder was stored in a lysis buffer (100mM TRIS-HCL pH=8; 100mM EDTA, pH=8; 10mM NaCl; 2% SDS [weight/volume]) for future genetic studies. The young storks were fitted with the transmitters using a backpack harness design and then returned to their nests by means of the same rope and cotton sack technique. The transmitters are constructed with tubes, front and back, through which the Teflon ribbon harness material passes. One piece of the ribbon is passed through the front tube, centered and knotted on either side of the PTT, to make two even length straps which will eventually pass over the shoulders.

PAGE 29

16 Another piece of ribbon likewise passes through the rear tube, centered and knotted to make the lower straps which will pass behind the wings and above the hips. The total handling time (i.e., from capture to replacement in the nest) took an average of 82 minutes (n = 6, range = 69-94 minutes) per bird. The PTT/GPSs were programmed to collect and store hourly GPS fixes taken from 05:00 to 19:00 hour local time. The units also collected and stored altitude, speed, and course data associated with each GPS location. Every two to three days the unit broadcasted the location data to orbiting satellites, from which the information was retrieved by Service Argos Inc., which processed and forwarded the results. I used the Perl application (ActivePerl 5.6.1.638, http://www.activestate.com) and the Argos/GPS PTT-100 Parsing Software ver. 6.3 (Microwave Telemetry, Inc.) to parse the data. This process separates the data by PTT and facilitates the conversion to a GIS-compatible format (shapefiles, feature classes, or feature datasets). All parsed files were reviewed to remove duplicate readings. Habitat Use Versus Available Shapefiles were created for all GPS location data obtained from the tagged storks. These individual activity ranges were then overlaid on ecosystem layers to compare used to available habitats for each stork. I used the Alternate Animal Movement Routes, v 2 extension in ArcView 3.2 (Jenness 2004) as an exploratory tool to compare habitat features associated with Jabiru stork locations with those that were available based on the mapped trajectory obtained from the detailed GPS data. I compared habitat characteristics of all route segments (generated by joining all GPS locations for each bird separately) by intersecting the segments with an ecosystem layer and calculating the distances and proportions of each route that traversed each unique habitat type. I

PAGE 30

17 generated 250 m buffers for all stork locations and then summed and sorted all habitat types to determine the cumulative proportion of each habitat. Since the transmitters utilized in this study provided several fixes per day, such high sampling effort more closely approximated the underlying trajectory and therefore provided better estimates of proportional habitat use, even though serial correlation may also increase (Aebischer et al. 1993). The total available area was defined in two different ways. First, the total available area was limited to the original country polygon, which excluded the southern Maya mountains. Second, the total available area was confined to the minimum convex polygons generated from the location data of all storks. I compared habitat features used by the Jabiru storks with those habitat features that were available along the path of a randomly-chosen route using the Alternate Animal Movement Routes v. 2 extension for ArcView 3.2. (Jenness, 2004). This method randomly distributes the original route segments, such that all alternative routes start at the same point and end at the same point and all have the same number of route segments. The set of random routes was developed by making a list of the distances and bearings traveled in each segment of the original route and then dividing each input route into a randomly rearranged set of vertex-to-vertex segments. Characteristics of the random routes could then be compared with those used by Jabiru storks by evaluating the attributes of polygon themes intersected by the routes. The polygon previously used to generate random points and the minimum convex polygon generated from the movement data for each stork were used as the boundary polygon in the randomization of route segments.

PAGE 31

18 I used the Animal Movement Extension for ArcView 3.2, v. 1.1 (Hooge and Eichenlaub, 1997) to explore site fidelity for each tagged stork. I used the first point as the starting point for the simulation. This method creates random angles and uses distances between sequential points to determine random walks. One hundred random walk simulations were generated based on the original data and a Monte Carlo simulation method was used to test whether the observed movement pattern displayed more site fidelity than should occur randomly, was a random pattern, or was overly dispersed. I also conducted a spatial cluster analysis (Average nearest neighbor, Distance = Euclidian) on the movement data to test whether the location points were randomly distributed or whether the data displayed a tendency to cluster. All movement data were analyzed relative to Belize’s protected-areas network, major watershed systems, and a digital elevation model to test for correlations with these features. A chi-square goodness-of-fit test (Siegel 1956) was used to test for differences in habitat use, thereby complimenting the use versus available comparisons. Additionally, I conducted habitat use versus available tests for the nest site data. For this analysis, total available area was defined by the area within a 3000 m buffer around the minimum convex polygon generated from the distribution of the 22 nest sites that were active during my study period. The total habitat proportions within this buffered polygon were calculated and sorted. Habitat used was defined as the proportions (summed and sorted) of habitat within 3,000 m buffers of each nest site. I used the one-sample t-test procedure to compare the proportions of habitat within the nest buffers (used) to the specified constant within the minimum convex polygon (available) (comparing sample means to hypothesized value: Sokal and Rohlf:169-175). This

PAGE 32

19 comparison was conducted using the SPSS 11.5 software. This procedure outputs the average difference between each data value and the hypothesized test value, a t test comparing this difference to zero, and a confidence interval. Home-range Analysis Using the Animal Movement Extension for ArcView 3.2 (Hooge and Eichenlaub 1997), I generated home range boundaries for the telemetered young birds based on the harmonic mean method (Dixon and Chapman 1980). I then removed the outlier locations that represented 2 % of the total for each of the four samples. This method removes points with the largest harmonic mean values and recalculates the range boundaries after each removal. Fixed kernel home range probability polygons (50%, 75% and 95%) were then generated for each distribution. I imported these probability polygons into ArcGIS 8.3 and used the boundaries to clip the ecosystem and watershed layers, thus determining the proportions of the vegetation types and watershed regions and the range of elevations within each polygon. I also overlaid the home range polygons on a map of Belize’s protected areas (i.e., public and private lands with various levels of conservation status) to determine the relative percentages of stork locations lying within protected and unprotected areas. For comparison with the probabilistic fixed kernel methods, I also produced minimum convex polygon (Hooge and Eichenlaub, 1997) estimates of the storks’ home ranges and performed the same analyses for vegetation types, watershed boundaries, and elevation. National and Regional Queries Finally, after determining the critical nesting and foraging habitats for the Belize population, I queried the GIS for Belize and then for the seven countries of the Central American isthmus to determine the distribution and relative abundance of these critical

PAGE 33

20 habitats. I used the ‘Select by Attributes’ function in ArcMap to develop a Structured Query Language (SQL) and to display these habitat types simultaneously for Belize and separately for the entire region. I reviewed the habitat classification scheme for each country (World Bank and CCAD 2000) to ensure that critical habitats were not excluded based on minor differences in names resulting from different geographic affinities. For example, SGSWS was one of the important habitats included in the SQL. However, limiting the SGSWS component of the query solely to SGSWS would have excluded the Pacific North and Central as well as the Valle del General and Central Valley variants that occur in Costa Rica. By verifying that all variants were included in the regional query, the results of this analysis does not underestimate the distribution and extent of these habitats throughout Central America.

PAGE 34

21 Figure 2. Map of northern Belize showing the general watershed regions surveyed for Jabiru stork nests.

PAGE 35

22 Figure 3. Map of central Belize showing the general watershed regions surveyed for Jabiru stork nests.

PAGE 36

23 Figure 4. Map of southern Belize showing the general watershed regions surveyed for Jabiru stork nests.

PAGE 37

24 Table 1. Broad scale ecosystem classes for Belize. Ecosystem Class / Land Cover Area (km 2 ) Percentage (%) Urban Coastal savanna Lowland pine forest Wetland Water Submontane pine forest Mangrove and littoral forest Lowland savanna Submontane broadleaf forest Agricultural uses Lowland broadleaf forest Total (approximate values) 110 234 304 410 457 473 938 1958 2210 3776 11291 22161 0.50 1.06 1.37 1.85 2.06 2.13 4.23 8.84 9.97 17.04 50.95 100 Calculated from the Belize Ecosystems Map (Meerman and Sabido, 2001)

PAGE 38

CHAPTER 3 RESULTS Nest Site Selection Visual inspection of the untransformed data comparing nest trees with random over-story trees revealed differences in the means of the four micro-habitat variables and, when comparing nest sites with random sites, differences in the means and variances of at least three landscape variables (elevation, distance to nearest nest and distance to nearest ecotone) (Table 2 ). However, since all linear variables were not normally distributed, I used log 10 to transform the data. Unless otherwise stated, results are presented as mean + SE (standard error). P-values less than 0.05 were considered significant. Microhabitat and Landscape Measurements Twenty-two breeding territories were identified as active during this study (Figure 5). Five nests (22.7%) were constructed on Pinus caribbea and seventeen (77.3%) on Ceiba pentandra trees (Table 3). Nest trees had a larger dbh (113.25 + 11.06 versus 59.71 + 3.47 cm) and their lowest branch was significantly higher (15.39 + 2.18 versus 8.05 + .530 m) than those of randomly selected nearest overstory trees (Wilcoxon Signed Ranks Test, Z = -2.635 and P < .008 for dbh and Z = -3.070 and P < .002 for height of lowest branch). Compared with random sites, nest sites were at lower elevations (24.7 + 5.76 versus 149.4 + 44.62, Mann-Whitney U-test, U = 99.00 and P < .006), were nearer to other nests (9,187.0 + 1,708.17 versus 16,611 + 1,510.1, Mann-Whitney U-test, U = 66 and P < .0001) and were nearer to ecotones (173.1 + 23.37 versus 1,078.1 + 220.4, Mann-Whitney U-test, U = 56.00 and P < 0.0001). 25

PAGE 39

26 Principal component analyses of the landscape variables for nest and random sites showed that the first three principal components explained the majority of total variation (87.4% for random sites and 79.9% for nest sites) (Table 4). For the random site landscape variables, PCA revealed gradients along elevation and the distances to nearest river, ecotone, and road. The PCA for the nest site landscape variables revealed a gradient along elevation and the distances to nearest river and road. For the microhabitat variables (Table 5), the first two principal components accounted for most of the total variation (95.6% for nest trees and 92.7% for random trees). Crown diameter and height of lowest branch loaded most significantly on the first and second components, respectively, for the nest microhabitat variables. Dbh and height of lowest branch loaded most significantly for the random overstory trees. Dbh represented the second most important loading along the first principal component in the nest microhabitat PCA. CART analysis for the nest sites produced a tree with four branches. Tropical lowland tall herbaceous swamp (TLTHS), tropical evergreen seasonal broadleaf lowland forest over lime rich alluvium (TESBLFOLRA), and short grass savanna with needle leaf open forest (SGSNLOF) were the three critical habitat variables that best separated the two groups (Figure 6). The cross-validation relative error plot suggested that a tree with four branches provided the best fit based on the one-standard-error rule (Figure 7). Simplifying this tree to four branches did not significantly increase the classification error. This tree had significant predictive power (CCR = 98% or 43/44, kappa [proportion improvement over chance] = 0.955 and P [occur] where predicted = 0.957). Based on these three habitat variables alone, the model correctly predicted 22 of 22 nests and 21 of 22 random sites.

PAGE 40

27 The model accurately predicted eleven nests based solely on the presence of at least 103.3 ha of TLTHS within a buffer radius of 3000 m. Six additional nests were accurately predicted with less than 103.3 hectares of TLTHS but more than 116.8 hectares of TESBLFOLRA. This TESBLFOLRA condition was present within the 5000 m buffers. Absence for 21 sites was predicted based on the co-occurrence of less than 103.3 hectares of TLTHS, 116.8 hectares of TESBLFOLRA, and 11.24 hectares of SGSNLOF. The SGSNLOF condition was identified within the 1500m buffer. Finally, presence was accurately predicted 83% of the time for the additional six sites. These six sites were defined by greater than 11.24 hectares of SGSNLOF within the 1500 m buffer, less than 103.3 hectares of TLTHS (3000 m buffer) and less than 116.8 hectares of TESBLFOLRA (5000 m buffer). One random site was misclassified. The Monte Carlo re-sampling showed that the predictive model represented by this specific four-branch tree could not be obtained by random chance alone (P<0.01) (Figure 8). One hundred permutations of the data separated the model from the plot of random trees. A three-leaf tree was obtained when CART was used to separate present from historical nest sites (Figure 9). Based on the one-standard-error rule, the cross validation relative error plot suggested a pruned tree with four branches as the best fit (Figure 10). The resulting model had a CCR of 77% (37/48), a Kappa value of 0.542 and a P (occur) where predicted of 0.75. Based on present day vegetation conditions of the historic sites, the principal habitat variables that best separated the two groups were Mixed mangrove scrub (MMS) and Agriculture. Eleven sites were classified as historic (91% CCR and 1 misclassification) when more than 47.89 hectares of MMS were present

PAGE 41

28 within the 5000 m buffer. With less MMS, Agriculture became significant. Thirteen sites were classified as historic (69% CCR and 4 misclassifications) when more than 45.91 hectares of agriculture were present within the 3000 m radius. With less than 45.91 hectares of Agriculture, 24 sites were classified as Present (75% CCR and six misclassifications). Six historical records and five present nest sites were misclassified. One hundred permutations of the data could not separate the model from the plot of random trees (Figure 11). Discriminant analysis (SPSS, 11.5) confirmed that TLTHS was highly significant in separating present nesting sites from the random sites (Wilks’ lambda = .819, F = 8.386, Sig. = .006). TESBLFOLRA was also significant (Wilks’ lambda = .848, F = 6.787, Sig. = .013) but SGSNLOF was not (Wilks’ lambda = .955, F = 1.774, Sig. = .191). TLTHS at the 1,500 buffer level was another strong discriminator (Wilks’ lambda = .882, F = 5.064, Sig. = .030). Nest Sites and Ecosystems Sixteen nests (73%) were built in Lowland Broadleaf Forests (LBF), four (18%) in Lowland Savanna and two (9%) in Agricultural areas (Chi-square = 15.636, P = .000) (Table 6). The most important ecosystem/ecotone combinations were LBF/Water bodies and LBF/Lowland Savanna with 8 (36%) and 7 (32%) sites respectively (Table 7). Collectively, these combinations represented 68% of nest and adjacent ecotone conditions. Sixteen (73%) random sites were also located in LBF; however, the ecotone combinations for random sites differed from those of nest sites (Table 7). Six (27%) random site/ecotone combinations were LBF/LBF, 5 (23%) were LBF/Agriculture, and 3 (14%) were LBF/Lowland Savanna.

PAGE 42

29 Cumulative vegetation composition within concentric buffers showed well-defined patterns for certain habitat variables (Figure 12). Relative to random sites, nest buffers averaged significantly higher proportions of TLTHS (0.90), Water (0.85), SGSWNLOF (0.84) and SGSWS (0.70), but lower proportions of Agriculture (0.33), Lowland pine forest (0.13) and Submontane broadleaf forest (0.00). Vegetation composition within random site buffers was devoid of Urban areas. However, small proportions of Urban areas were found within the 3000 m and 5000 m nest site buffers (Figure 12). Habitat Use Versus Availability Nest Sites Nesting Jabiru storks selected sites with significantly higher proportions of Water (t = 3.13 and P = .01), TLTHS (t = 2.46 and P = .02) and SGSWNLOF (t = 2.15 and P = .04), and selected against Lowland Pine Forest (LPF: t = -2.56 and P = .02) and Urban areas (t = -39.0 and P = .00) (Table 8). The proportions of LBF and MLF used by nesting Jabiru storks were similar to the proportion that was available (Table 8). Post Fledging Movement A total of 9,381 post-fledging location fixes were obtained for six tagged storks (Figure 13). Since two birds died within one month of leaving the nest, I used data from four individuals (9,286 location fixes) hereafter referred to as Maypen, HillBank, Manatee1 and Manatee2. Of the 9,286 fixes, Maypen, HillBank, Manatee 1 and Manatee 2 accounted for 4,052 (43.64 %), 1,614 (17.38 %), 2,381 (25.64 %) and 1,239 (13.34 %) respectively. Of the total fixes, 4,083 (44.0%) were in Lowland savanna and 3,277 (35.3%) were in Wetland. All Lowland savanna fixes were in either SGSWS (3,437) or SGSWNLOF (646). All wetland fixes were in TLTHS.

PAGE 43

30 The mean age at fledging (age at which nestlings first abandoned the nest) of tagged Jabiru storks was 96 days (n = 5). The earliest fledging was on April 19 and the latest on June 2. Mean distance traveled within the first 24 hours after tagging varied from only 35 meters for Manatee 1 to 2,700 meters for HillBank. The HillBank nest was located in a dense secondary broadleaf forest and adjacent to the 300 m wide New River lagoon. This fledging therefore had to cross this barrier at first flight. Adult storks remained with juveniles for 10 to 12 weeks after fledging. Fledgling stork movement during this time was confined to foraging habitat within 10 km of the nest site. For the duration of the telemetry study, stork movement was confined to both the Orange Walk and Belize districts, with one exception. On October 23, 2004, HillBank moved across northeastern Corozal district into the southern tip of Mexico, where it spent one day. It then moved into northern Ambergris Caye and back to mainland Belize. The entire movement lasted three days, covered over 180 km, and may be the first documented occurrence of this species in the southern tip of Mexico (near Xcalak), on Ambergris Caye, and in the Corozal district. Hillbank was 17 months old at the time of this movement. While Jabiru storks displayed the ability to cover large distances in short time periods, none migrated and none displayed any tendency for large-scale regional or seasonal movements. Habitat Associations of Juvenile Storks Maypen and Hillbank used TLTHS more than expected while Manatee1 and Manatee2 used SGSWS more than expected (Asymp. Sig. = .000) (Table 9). For each of the four birds, there was less than a 1% chance that the displayed movement pattern could have resulted from chance (Table 10). Hillbank, Maypen, Manatee1 and Manatee2 survived for 21, 11.5, 6.5 and 4 months respectively. One bird was confirmed dead

PAGE 44

31 (Manatee1) after the transmitter was recovered and the other three individuals were assumed to have perished after their transmitters stopped emitting signals. Juvenile Jabiru storks consistently selected for TLTHS and against LBF and Agriculture (Tables 11 & 12). The selection against agricultural areas was most evident when proportions of habitat along actual routes were compared to proportions along random routes within the total available area (Table 12). The selection for TLTHS was strongest when buffered locations were compared to the total available habitat throughout Belize (Table 11). Within the MCP areas, selection for TLTHS was strongest when habitat proportions along travel routes were compared to random routes (Table 12). Selection for LS was consistently demonstrated but was less-highly significant across all three comparisons. All location fixes were at least 250 meters from Urban areas (Table 11). This pattern of selection was repeated in the analyses of vegetation composition within fixed kernel probability polygons (Figure 14). Within the 50% probability polygon, TLTHS accounted for almost 50% of the total vegetation. As the polygon size increased, the total proportion of TLTHS decreased to 16% within the 95% polygon. SGSWNLOF showed a similar pattern, decreasing in proportion as the probability polygon increased, from 17% at the 50% polygon to 12% at the 95% polygon. Nest Sites by Protected Areas, Watershed Regions and Political Boundaries Four (18.18 %) of the 22 present nests and five (20 %) of the 25 historical sites were located within the national network of protected areas (Figure 15). Two of the four nests found inside protected areas were within one hundred meters of the periphery. Seven (31.81 %) additional nests were located outside but within 1,200 m of protected areas.

PAGE 45

32 Eleven nests were located within the Belize river watershed, three each within the New, Sibun, and Manatee river watersheds, and one each within the Rio Hondo and Moho watersheds. Twenty-one nests (95%) were located within two of the more populous districts (CSO 2004): 14 (64%) in the Belize district and seven (32%) in Orange Walk. Home Range, Protected Areas and Watershed Regions The 95 % fixed kernel probability polygon for fledgling birds encompassed an area of 48,821 ha. Of this total, 39,657 hectares (81.2%) fell outside the protected areas network (Figure 16). Elevations within these polygons ranged from 0 to 40 meters above sea level (12.95 + 7.56). A significant percentage of these polygons were located within three watershed regions: New River (33.1%), Belize (27.6%), and Manatee (33.1%). Similar results were obtained using the MCP home range estimates. Just over 18 % of the total MCP area fell within the protected areas network. Elevation values ranged from -3 to 59 meters above sea level (12.52 + 7.54). The Belize, New, Manatee, and Sibun river watersheds comprised 30.0, 15.0, 8.8 and 4.2 % respectively of the combined MCP polygons. National and Regional Queries Based on the results of this study, nine vegetation classes were queried for the entire Central American region. These included six lowland savanna habitats: waterlogged short-grass savanna without trees or shrubs, short-grass savanna with deciduous shrubs, short-grass waterlogged savanna (Mosquita variant), short-grass savanna with scattered needle-leaved trees, short-grass waterlogged savanna with broadleaved treed and three wetland habitats: tall herbs lowland swamp, tall sedge swamp and, herbs and grass swamp with shrubs and/or palms.

PAGE 46

33 Honduras and Nicaragua had the highest amounts of these vegetation classes but Belize had the highest relative percentage when compared to total national land area (Table 13). With the exception of Belize and the Caribbean coastal wetlands of Honduras and Nicaragua, these habitat types were widely scattered, very limited in area, and almost exclusively confined to lowland and coastal areas (Figure 17). In Belize, approximately 2,330 km 2 of these habitat types remain, disproportionately concentrated along the northern and central lowlands (Figure 18). Relatively small patches are scattered along the Pacific coast of Nicaragua, Honduras, El Salvador and Guatemala (Figure 17). A substantial percentage is concentrated along the coastal Caribbean wetlands of north-eastern Nicaragua (Figure 19) and along the eastern Caribbean coast of Honduras (Figure 20). In Costa Rica, small quantities are distributed along the Atlantic and Pacific slopes in the northern part of the country near the border with Nicaragua, and along the Pacific coast in the southwestern portion of the country (Figure 21). With approximately 205 km 2 present in Costa Rica, these habitats comprised 0.4 % of the total national land cover and 2 % of the total Jabiru stork habitat in the region. With 239 km 2 , 267 km 2 and 568 km 2 in Panama, El Salvador and Guatemala respectively, these habitats comprised 0.31, 1.29 and 0.52 % of total land cover for each country and 2, 3 and 5 % of the total Jabiru stork habitat in the region, respectively. These habitats can still be found covering 2.09 % of the entire Central American isthmus but are disproportionately distributed across (Table 13) and within countries (Figure 17). Approximately 66 % of the entire distribution is confined to the Caribbean coastal wetlands of eastern Honduras and northeastern Nicaragua. The 2,330 km 2 found in Belize represented 22 % of the total distribution in Central America. With a combined total of 9,401 km 2 , Belize, Honduras

PAGE 47

34 and Nicaragua accounted for 88 % of the total distribution of these critical Jabiru stork habitats in Central America. The remaini ng three countries, El Salvador, Guatemala and Panama accounted for a combined 1,074 km 2 , or 10 % of the total available habitat in the region. However, this 10 % was comprised of small and unconnected patches that were widely dispersed.

PAGE 48

35 Table 2. Comparisons of microhabitat variables measured for nest trees (2a) and randomly selected over story trees (2b), and comparisons for landscape measurements at nest and random sites (2c). 2a. Nest Trees n=22 Range Min. Max. Mean + SE SD Nest Height 45.7 9.2 54.9 18.77 2.31 9.81 Nest Branch 41.1 9.2 50.3 17.76 2.11 8.96 Lowest Branch* 43.3 4.8 48.9 15.39 2.18 9.24 Tree Height 56.9 10.9 67.8 28.6 3.34 14.15 DBH* 141 26 167 113.25 11.06 46.93 Crown Diameter 30.2 8.8 39 25.38 2.07 8.06 2b. Random Trees n=80 Range Min. Max. Mean + SE SD DBH* 140 20 160 59.71 3.47 31.01 Tree Height 28.5 7.6 36.1 20.63 .77 7.04 Lowest Branch* 20.2 1 21.2 8.05 .530 4.74 Crown Diameter 37.2 3 40.2 15.80 .833 7.45 2c. Range Mean + SE SD Variable (n = 22) Nest Random Nest Random Nest Random Nest Random Nearest Neighbor* 25600 25635 9187 16611.8 1708.2 1510.1 7639.2 6753.4 Elevation* 95 641 24.7 149.4 5.76 44.62 25.7 199.6 River 3264 4029 1049.8 1341.1 242.1 292.4 1082.7 1307.6 Road 5975 13515 2762.5 4547.9 396.1 959.1 1771.3 4289.1 Community 11667 8838 5717.3 4651.9 796.9 556.8 3563.9 2489.9 Ecotone* 395 3221 173.1 1078.1 23.37 220.4 104.5 985.5 *denotes significant difference for comparisons along this variable

PAGE 49

36 Figure 5. Distribution of Jabiru stork nests by political division (district) in Belize, 2005.

PAGE 50

37 Table 3. Species and relative percentage of eighty randomly selected unused (over story) trees within 180 meters of 22 active Jabiru stork nests. Frequency ( %) Species Nest trees Non-nest trees Guazuma ulmifolia Sabal mauritiiformis Acosmium panamense Vatairea lundellii Sabal yapa Brosimum alicastrum Cassia grandis Bucida buceras Pinus caribaea Ceiba pentandra Orbignya cohune Vitex gaumeri Ficus spp. Delonix regia Enterolobium cyclocarpum Spondias mombin Tabernamontana spp. Quercus oleoides Shizolobium parahybum Roystonea regia Calophyllum braziliense Manilkara (Achuras) zapota Ficus cotinifolia Unknown 1 Unknown 2 Unknown 3 *Wild Caimito Spondias purpurea 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 5 (22.7) 17(77.3) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 3 (3.75) 1 (1.25) 1 (1.25) 1 (1.25) 4 (5.0) 1 (1.25) 6 (7.5) 3 (3.75) 12 (15) 4 (5.0) 3 (3.75) 2 (2.5) 5 (6.25) 1 (1.25) 1 (1.25) 9 (11.25) 1 (1.25) 7 (8.75) 3 (3.75) 1 (1.25) 1 (1.25) 3 (3.75) 1 (1.25) 1 (1.25) 1 (1.25) 2 (2.5) 1 (1.25) 1 (1.25) *common name

PAGE 51

38 Table 4. Principal component analysis (PCA) of variation along landscape variables (log-transformed) measured for random sites (a) and nest sites (b). Only first three principal components reported because these explain the majority of the variation. a. Principal Component Cumulative % variance I 54.8 II 73.2 III 87.4 Variables Loadings Elevation 0.539 -0.370 0.156 Point to Point 0.109 River -0.754 -0.257 -0.347 Road 0.303 0.532 -0.758 Ecotone 0.216 -0.716 -0.515 Community b. Principal Component I II II Cumulative Variance 44.6 67.2 79.9 Variables Loadings Elevation -0.404 0.652 Nest to Nest -0.302 0.516 River -0.886 0.370 Road -0.401 -0.733 -0.400 Ecotone -0.203 0.137 0.236 Community -0.228 -0.295 Insignificant loadings not reported. Table 5. Principal component analysis (PCA) of variation along microhabitat variables for nest trees (a) and random trees (b). Only first three principal components reported because these explain the majority of the variation. a. Principal Component I II III Cumulative Variance 77.5 95.6 98.3 Nest Variables Loadings Tree height -0.372 0.241 Nest height -0.239 0.363 -0.527 Dbh -0.490 -0.110 0.134 Crown diameter -0.667 -0.549 Nest branch -0.222 0.394 -0.535 Lowest branch -0.266 0.584 0.639 b. Principal Component I II III Cumulative Variance 63.7 92.7 97.0 Random Variables Loadings Dbh -0.617 0.273 0.676 Tree height -0.431 Lowest branch -0.312 -0.942 Crown -0.580 0.194 -0.735 Insignificant loadings not reported

PAGE 52

39 Figure 6. Dendrogram model based on CART analysis for predicting Jabiru stork nest sites (present). The first split is along the TLTHS variable. If there is less than 103.3 ha follow the tree to the left (true conditions go left). With more than 103.3 ha follow the tree to the right to find that the prediction is Present and that this is correct 100% of the time for 11 cases. Kappa = 0.955 and Model = 98% (43/44). Figure 7. The cross-validation relative error plot (cp plot) used to pick the tree size based on the 1 standard error rule and a smoothing parameter of 100. The first point that is less than and within one standard error of the minimum corresponds to a tree with four leaves.

PAGE 53

40 Figure 8. Distribution of correct classification rates (CCR) for 100 randomly generated trees versus the CCR of the model (triangle). P < .01 indicates that the model was not produced by chance. P value based on Monte Carlo resampling. Figure 9. Dendrogram model based on CART analysis for comparing Jabiru stork nest sites (Present) with historical records (Historical). The first split is along Mangrove Scrub (MS). With MS greater than or equal to 47.89 ha follow tree to the left (true conditions go left) to find that the prediction is Historical and that this is correct for 91% of 11 cases. Kappa = 0.577 and Model = 77%.

PAGE 54

41 Figure 10. The cross-validation relative error plot (cp plot) used to pick the tree size based on the one standard error rule and a smoothing parameter if 100. The first point that is less than and within one standard error of the minimum corresponds to a tree with three leaves. Figure 11. Distribution of correct classification rates (CCRs) for 100 randomly generated trees versus the CCR of the model (triangle). P value of 0.2 based on Monte Carlo resampling does not separate the model for the plot of random trees.

PAGE 55

Table 6. Comparisons of ecosystems within which Jabiru storks built nests and the associated nearest ecotone with those of randomly selected points. All nests were constructed within an average distance of 173.1 + 23.37 meters from the nearest ecotone compared to 1078.1 + 220.4 meters for random sites. Nest Nearest Ecosystem Random Point Nearest Ecosystem Lowland Broadleaf Forest Water Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Water Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Wetland Agriculture Lowland Broadleaf Forest Lowland Broadleaf Forest Water Lowland Savanna Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Savanna Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Water Lowland Broadleaf Forest Lowland Savanna Lowland Broadleaf Forest Water Lowland Broadleaf Forest Agriculture Agriculture Lowland Broadleaf Forest Lowland Savanna Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Savanna Lowland Broadleaf Forest Basin Mangrove Lowland Broadleaf Forest Water Agriculture Agriculture Agriculture Lowland Savanna Submontane Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Agriculture Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Water Lowland Broadleaf Forest Lowland Pine Forest Lowland Broadleaf Forest Lowland Savanna Lowland Broadleaf Forest Submontane Broadleaf Forest Lowland Savanna Lowland Broadleaf Forest Submontane Broadleaf Forest Submontane Broadleaf Forest Lowland Savanna Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Pine Forest Lowland Savanna Lowland Broadleaf Forest Lowland Broadleaf Forest Agriculture Lowland Broadleaf Forest Water Lowland Broadleaf Forest Water Lowland Savanna Lowland Broadleaf Forest Lowland Broadleaf Forest Agriculture Lowland Broadleaf Forest Lowland Broadleaf Forest Lowland Broadleaf Forest Agriculture Lowland Broadleaf Forest Wetland Lowland Broadleaf Forest Lowland Broadleaf Forest 42

PAGE 56

43 Table 7. Relative frequencies of ecotones a ssociated with Jabiru stork nest sites and random sites. Frequency (%) Ecotone Nest Random Lowland Broadleaf Forest / Lowland Broadleaf Forest 2 (9) 6 (27) Lowland Broadleaf Forest / Water 8 (36) 1(4.5) Lowland Broadleaf Forest / Lo wland Savanna 7 (32) 3 (14) Lowland Broadleaf Forest / Agriculture 2 (9) 5 (23) Lowland Broadleaf Forest / Lowland Pine Forest 0 (0) 2 (9) Lowland Broadleaf Forest / Basin Mangrove 0 (0) 1 (4.5) Lowland Broadleaf Forest / Submontane Broadleaf Forest 0 (0) 2 (9) Lowland Broadleaf Forest / Wetland 2 (9) 0 (0) Lowland Savanna / Agriculture 1 (5) 0 (0) Submontane Broadleaf Forest / Submont ane Broadleaf Forest 0 (0) 1 (4.5) Agriculture /Agriculture 0 (0) 1 (4.5)

PAGE 57

00.10.20.30.40.50.60.70.80.9LBFAgricultureSGSWSWaterSGSNLOFTLTHSMLFSBFVegetationProportion Nest 500 Random 500 Nest 1500 Random 1500 Nest 3000 Random 3000 Nest 5000 Random 5000 44 Figure 12. Comparison of the proportions of vegetation composition within concentric buffers for nest (n=22) and random (n=22) sites.

PAGE 58

45 Table 8. Comparisons of habitat use (within 3000 m buffe rs of nest sites) to the proportions of available habitat within the 3000 m buffered MC P polygon for all nest sites. 95% Confidence Interval of the Difference Habitat Available MCP Jabiru Use t P (2-tailed) Mean Difference Min Max LBF 0.53 0.53 + 0.04 0.10 0.92 0.00 -0.088 0.097 Agriculture 0.13 0.09 + 0.04 -0.83 0.42 -0.03 -0.119 0.052 SGSWS 0.04 0.13 + 0.04 2.06 0.05 0.09 -0.002 0.181 SGSWNLOF * 0.04 0.12 + 0.04 2.15 0.04 0.08 0.002 0.153 TLTHS* 0.02 0.07 + 0.04 2.46 0.02 0.05 0.008 0.097 LPF* 0.07 0.00 + 0.00 -2.56 0.02 -0.01 -0.018 -0.002 Water * 0.01 0.04 + 0.01 3.13 0.01 0.03 0.011 0.053 Urban * 0.00 0.00 + 0.00 -39.00 0.00 -0.02 -0.003 -0.002 MLF 0.00 0.01 + 0.01 0.65 0.52 0.01 -0.014 0.028 *denotes significant difference

PAGE 59

46 Figure 13. Distribution of tracking locations for juvenile Jabiru storks from four nests in Belize. Point locations derived from GPS receivers incorporated in satellite transmitters carried by each bird (n= 1239 – 4052 fixes per bird).

PAGE 60

47 Table 9. Results of Chi square analyses showi ng that juvenile Jabiru storks used tropical lowland tall herbaceous swamps (TLTH S) and short grass savanna with shrubs (SGSWS) more than expected. Stork Habitat Chi square df Asymp. Sig. Maypen * TLTHS 6146.904 5 .000 HillBank * TLTHS 2035.653 5 .000 Manatee1 * SGSWS 6779.373 4 .000 Manatee2 * SGSWS 2446.102 4 .000 *significant result Table 10. Results of the spatial cluster an alysis (average near est neighbor, distance method = Euclidian) procedure used to test the hypothesis that the activity pattern displayed by the tagged Jabiru storks could be the result of random chance. Stork n OMD/EMD Z Critical value Sig. Manatee1 * 2381 0.12 67.2 -2.58 0.01 Manatee2 * 1239 0.14 80.3 -2.58 0.01 Maypen * 4052 0.20 53.3 -2.58 0.01 HillBank * 1614 0.13 105.9 -2.58 0.01 OMD = Observed mean distance *significant result EMD = Expected mean distance Table 11. Comparison of the cumulative propor tion of habitat within 250 meter buffers of all Jabiru stork locations to the proportions of available habitat within the MCP home range and to the overall proportions available in Belize. Habitat Stork Use MCP Belize LBF 0.14 0.46 0.58 SGSWS 0.21 0.09 0.06 SGSWNLOF 0.16 0.15 0.05 TLTHS 0.35 0.04 0.02 MLF 0.01 0.07 0.03 Coastal Savanna 0.003 0.04 0.009 Agriculture 0.02 0.07 0.2 Water 0.10 0.05 0.02 Seagrass 0.0004 0.03 0.002 LPF 0 0.004 0.01

PAGE 61

48 Table 12. Comparison of the cumulative proporti on of habitat along original Jabiru stork routes to the proportions of habitat ava ilable along randomly distributed routes within the MCP and along randomly dist ributed routes throughout Belize. (mean + SD). Habitat Original Routes Random Routes within MCP Random Routes within Belize Lowland Savanna .504 + 0.335 .460 + 0.413 .113 + 0.308 TLTHS .271 + 0.236 .050 + 0.199 .023 + 0.142 LBF .124 + 0.202 .286 + 0.414 .554 + 0.482 Water .050 + 0.134 .074 + 0.236 .021 + 0.132 Urban .015 + 0.042 .003 + 0.069 .004 + 0.062 MLF .014 + 0.144 .046 + 0.172 .033 + 0.173 Agriculture .003 + 0.037 .057 + 0.190 .204 + 0.389 Coastal Savanna .002 + 0.026 .027 + 0.212 .010 + 0.095 Seagrass .001 + 0.098 .011 + 0.138 .001 + 0.032 LPF .000 + 0.004 .004 + 0.063 .010 + 0.097 SBF .009 + 0.064 Other .000 + 0.001 .002 + 0.091 .011 + 0.108

PAGE 62

49 Figure 14. Cumulative vegetation proportions within 50, 75 and 95 percent fixed kernel probability polygons based on actual point locations derived from GPS receivers incorporated in satellite transmitters carried by each bird (n= 1239 – 4052 fixes per bird). 00.10.20.30.40.50.6WaterTLTHSProportion AgricultureLBFLPFSGSWSSGSWNLOFMLFUrbanLand Cover / Vegetation 95% 75% 50% LBF – Lowland broadleaf forest SGSWS – Short grass savanna with shrubs SGSWNLOF – Short grass savanna with needle leaf open forest MLF – Mangrove and littorial forest TLTHS – Tropical lowland tall herbaceous swamp

PAGE 63

50 Figure 15. Distribution of 22 Jabiru stork nest sites (2005) and 25 historical nest records relative to the protected areas network in Belize.

PAGE 64

51 Figure 16. Estimates of the activity ranges for juvenile Jabiru storks from four nests in Belize. Estimates are based on Adaptive Kernel Analysis. Colored ellipses indicate probability of expected occurrence based on actual point locations derived from GPS receivers incorporated in satellite transmitters carried by each bird (n = 1239 – 4052 fixes per bird). Table 13. Distribution and relative percentages of critical Jabiru stork habitat within all seven countries of Central America and the proportions of all Central America Jabiru stork habitat represented by each country. Based on a regional GIS query of the Central American Ecosystems map. Country Land Area (km 2 ) Jabiru Habitat (km 2 ) Percent of Country Proportion of total CA* Jabiru Habitat Belize 22,806 2,330 10.22 .22 Guatemala 108,430 568 .52 .05 El Salvador 20,720 267 1.29 .03 Honduras 111,890 3,012 2.69 .28 Nicaragua 120,254 4,059 3.38 .38 Costa Rica 50,660 205 .40 .02 Panama 75,990 239 .31 .02 Total 510,750 10,680 2.09 1 CA – Central America

PAGE 65

52 Figure 17. Distribution and abundance of critical Jabiru stork habitat throughout Central America. Based on a Geographic Information Systems query using the results of the Belize analyses and the Central American ecosystems map ( http://wbln0018.worldbank.org/MesoAm/UmbpubHP.nsf/ )

PAGE 66

53 Figure 18. Relative distribution of critical Jabiru stork habitat, 2005 nest sites (n = 22) and historical nest records (n = 25) in Belize. Critical habitats include tropical lowland tall herbaceous swamps, short grass savanna with shrubs and short grass savanna with needle leaf open forest.

PAGE 67

54 Figure 19. Map of Nicaragua showing the distribution of critical Jabiru stork habitat. Savanna systems include short grass savanna without trees or shrubs, waterlogged, with deciduous shrubs, with needle leafed trees and with broad leaved trees. Herbaceous swamps include tall herbs, tall sedge swamp and waterlogged tall-grass savannas. Agriculture systems including cropland/vegetation mosaics are shown for comparison.

PAGE 68

55 Figure 20. Map of Honduras showing the distribution of critical Jabiru stork habitat. Savanna systems include short grass savanna without trees or shrubs, waterlogged, with deciduous shrubs, with needle leafed trees and with broad leaved trees. Herbaceous swamps include tall herbs, tall sedge swamp and waterlogged tall-grass savannas. Agriculture systems including cropland/vegetation mosaics are shown for comparison.

PAGE 69

56 Figure 21. Map of Costa Rica showing the distribution of critical Jabiru stork habitat. Herbaceous swamps include tall herbs, tall sedge swamp and waterlogged tall-grass savannas. Agriculture systems including cropland/vegetation mosaics are shown for comparison.

PAGE 70

CHAPTER 4 DISCUSSION While some studies have provided information on nest and nest site characteristics (Kahl 1971, Spaans 1975, Thomas 1981, Gonzales 1996, Villareal 1997), this was the first attempt to quantify breeding habitat preferences and to determine the habitat features that influence nest site selection for Jabiru storks. While a satellite study of Jabiru storks was previously conducted on South American Jabiru storks (Antas and Nascimento 1997) my research represented the first attempt incorporating GPS technology to the study of movements and distribution of foraging habitats and the first (and only) movement study on the Mesoamerican population. My results suggest that features at both the microhabitat and landscape levels may influence nest-site selection by Jabiru storks. I found strong evidence indicating nonrandom use of habitat. The trees selected for nesting and the surrounding landscape composition were nonrandom features in the total available area. Within the stand, the Jabiru stork seems to choose one of the biggest over-story trees (significantly larger dbh and height of lowest branch) upon which to construct its nest. From an aerial perspective, nest trees stood out as emergent and appeared to be the largest in the stand, with the lowest branches above the adjacent forest canopy. These large trees provide the maneuvering space necessary for such large birds (2.4 m wingspan) and the required framework to support the massive nest structure (1.57 + .17 m diameter and .72 + .15 m depth, n = 4). While only four nests were physically measured, close observation of all 22 nests indicated that they were similar in size and shape, and all were wider than they 57

PAGE 71

58 were deep (contra Barnhill et al. 2005). Other essential features of the nest tree must accommodate flight in and out of the nest by adults and fledglings. This structural relationship with the adjacent vegetation provided the necessary space for nestlings to fledge and move in and out of their nests while their flight skills were developing. Jabirus nesting within a forested nest stand (17 of 22 nests) had to clear the adjacent forest trees during their first flights. Individuals that fail to clear the adjacent forest may not survive. Additionally, these structural features of the nest tree (emergent, isolated, height of lowest branch) may also decrease the vulnerability of these nests to potential mammalian predators such as the Margay (Leopardus weidi) and the Jaguarundi (Felinae yaguarondi). Nest sites were at low elevations (25.70 + 5.76 m above sea level) and the surrounding landscape had high proportions of water, tropical lowland tall herbaceous swamps, and short grass savanna (with needle leaf open forests). These distinctive landscape features (habitat composition within the general nest area) were highly nonrandom across the available habitat and the foraging range of juvenile storks was confined to within these habitats. Within this landscape, the closest spacing of simultaneously active nests was 5.3 km. The limited distribution of these critical habitats and the solitary and dispersed nesting strategies of Jabiru storks may limit the distribution of nests and simultaneously increase the vulnerability of this species to changes in the landscape. The selection of nesting habitat may also be driven by factors such as prey availability, interspecific competition, and predation, which may determine the success or failure of the nesting attempt. The availability of food and the intensity of predation may

PAGE 72

59 determine reproductive success in birds and these factors may influence the nest site selection strategies (Eberl and Picman 1993). If food is abundant in time and space Jabirus can establish all purpose territories. The scope of this study did not permit investigation of prey abundance or availability in the associated foraging habitats, but the design permitted me to test the prediction that nest sites were associated with certain key foraging habitats. Structural features of nest sites and certain behavioral adaptations of Jabiru stork’s may minimize the potential significance of predation during the nesting cycle. In the southern Llanos of Venezuela, Gonzales (1996) observed Crested Caracaras (Polyborus plancus) eating eggs and nestlings at the nest but commented that it was unclear whether this was predation or post-abandonment scavenging. At one nest in Belize I observed four Black Vultures (Coragyps atratus) perched within 2 m of an active Jabiru stork nest for prolonged periods during four separate observations in the first eight weeks of the nestling stage. At least one adult Jabiru stork was always present during this time. Nests were constructed high and above the canopy, thereby limiting access by potential terrestrial predators. Additionally, for several weeks after hatching and until nestling size could serve as a deterrent, adult Jabirus maintained an almost full-time presence at the nest rendering protection to the young storks. Nestlings also tended to retreat into nests when left alone and to restrict vocal outbursts to short periods of begging displays, thus reducing exposure and the associated risk of predation. Jabiru storks may mitigate predation pressures by favoring predator avoidance tactics such as the selection of a safe nesting site and a diligent defense of nestlings during the vulnerable stages (first 3 to 4 weeks). While the role of predation in nest site selection may be limited, it is probably

PAGE 73

60 one of the key factors driving post-fledging mortality. Adult Jabiru storks may spend 10 to 12 weeks with fledgling storks passing on necessary survival skills. This relatively short time period may be insufficient for the young storks to acquire crucial predator-avoidance skills. The large, conspicuous, and inexperienced fledglings may be at high risk from predation as they are gradually left to fend for themselves in habitats that are infested with Morelet’s crocodiles (Crocodylus moreletii). The Moorelet’s crocodile may be the only natural predator of post-fledging Jabiru storks but the significance of this pressure is significant because the distribution of these crocodiles may be closely associated with the critical Stork habitats. The twenty-two nests documented in this study represent the highest known concentration in Mesoamerica. While a thorough and comprehensive nest searching effort minimized the possibility of missing active nests, it is possible that a few were not recorded. I estimate this number to be no more than three, indicating that twenty-five breeding pairs may possibly exist in Belize. An additional 33 nests have been reported elsewhere in Mesoamerica (Correa and Luthin 1988, Hartasanchez 1992, Lopez-Ornat and Ramo 1992, Frederick et al. 1997, J. Villareal pers. com, O. Arrologo pers. com. and D. Medina pers. com.). While the availability of critical habitat may limit the total number of nests in Costa Rica to the reported 8 active nests, the same cannot be said for the Caribbean coastal wetlands of Honduras and Nicaragua. Frederick et al. (1997) first suggested the regional importance of this area after documenting 10 nests (6 in Nicaragua and 4 in Honduras) and 98 individuals (74 in Nicaragua and 24 in Honduras) during aerial surveys in March 1992 (Nicaragua) and February 1994 (Honduras). The result of my regional query supports this suggestion and indicates that this region may rival Belize

PAGE 74

61 in the importance for Jabiru storks. Based on this regional query it can be assumed that the Caribbean coastal wetlands of Honduras and Nicaragua may hold at least an equal number of nests as Belize. However, no further studies have been conducted in this region and except for the Frederick et al. (1997) study, the status and distribution of the waterbird population remains poorly known. I believe that a reasonable estimate for the number of Jabiru stork breeding pairs in Mesoamerica is approximately 70-80 pairs, with Belize and the coastal wetlands of Honduras and Nicaragua accounting for two thirds, or about 50 nests. This is the highest nest estimate for the region but the first based on a thorough sampling effort of one of the nesting subpopulations. Previous population estimates for the Mesoamerican population placed the total number of Jabirus at 150 to 250 individuals (Correa and Luthin 1988, Luthin 1987, Del Hoyo et al. 1992), but recently Paredes (2004) doubled these figures with an estimate of 400 individuals. Since Jabiru storks exhibit post breeding congregatory behavior, population estimates during these times will include juveniles, immatures, and adults. Because of high juvenile mortality, estimates based on counts conducted after the breeding season should be interpreted with caution. Based on my assumption of 80 breeding pairs in Mesoamerica, an average clutch size of 1-3 eggs per nest (n = 42), the relatively long-deferred reproductive maturity (3 to 4 years) and the high juvenile mortality of the tagged storks, I estimated the total regional population to be between 250 and 400 individuals. However, since it may take at least three years before the Storks attain reproductive maturity, the present recruitment levels may not be adequate to sustain the population and it is therefore possible that the Jabiru stork population in Mesoamerica may now be declining.

PAGE 75

62 All of the Belize nests and 42 percent of the total post-fledging location fixes were associated with either tropical lowland tall herbaceous swamps or short grass savanna with needle leaf open forest. These two habitats represented the larger and more permanent foraging habitats available to Jabiru storks. They are less vulnerable to extreme weather conditions and are capable of sustaining prey species during critical periods, such as prolonged droughts. An additional 37 percent of the post fledging location data was associated with short grass savanna with shrubs. This latter type may represent the principal foraging habitat for the six pairs nesting within the Sibun and Manatee river watersheds (Central Belize). These habitats can be severely affected by drought conditions. During the dry season (March through May/June), the prey base is significantly reduced and the foraging patches become smaller, less productive and widely dispersed. As the breeding season progresses, these nesting storks may be forced to travel further distances in search of food. As a result, nesting pairs need to extend their foraging ranges to the larger wetlands north of the Sibun watershed and may be forced to suspend defense of their breeding territories and the associated foraging habitats. This imposes an additional cost in time and energy and can affect survival and reproductive success. The selection of suboptimal nesting habitats, such as those further from TLTHS and SGSWNLOF or with reduced concentrations, may provide additional challenges for nesting pairs, especially during late in the nesting cycle and during less productive years when foraging opportunities may decrease earlier in the breeding season. In Belize, the breeding season for Jabiru storks runs from mid-November through early June. The period from nest building/repair to fledging takes about 24 weeks (n = 5, pers. obs.). Since adult Jabiru storks invest an additional 10 to 12 weeks with post

PAGE 76

63 fledging juveniles, the total time investment in breeding and reproduction may surpass 36 weeks. Such a protracted breeding season underscores the significance of adequately selecting breeding territories that are within reasonable flight distances of productive foraging habitats. Moreover, during the first few weeks after hatching, the nestling growth rate and energy requirements are high. During the first two weeks, adults may make 9 to 13 foraging trips daily (pers. obs), returning to the nest with high-protein food items, primarily toads (Bufo spp.) and swamp eels (Symbranchus marmoratus) necessary to meet the energy demands of the growing nestlings. As the season progresses, swamp eels and fish dominate the nestling diet, but turtles and crustaceans may also be delivered (pers. obs.). Storks that select breeding territories with higher proportions of these critical habitats in the surrounding landscape will have an adequate food supply for their lengthy nesting season. Nest-site selection may, therefore, strongly influence success for breeding Jabiru storks. Because of the high demands imposed by a lengthy breeding season, Jabirus almost certainly do not attempt to re-nest after a failure. Breeding pairs that select and defend territories near the larger wetlands such as those associated with Crooked Tree Wildlife Sanctuary and the New River Lagoon (higher proportions of TLTHS and SGSWNLOF) may be able to forage within a few kilometers for the entire breeding season. Movement It has been suggested that the Belize population of Jabiru storks is migratory (Corea and Luthin 1988, Hancock et al. 1992, Howell and Webb 1995, Miller and Miller 1998), leaving the country from June and returning in November (Luthin 1984). I found no evidence of any seasonal movements despite year-round observations over a six-year period. This suggestion of seasonal movements may have been based on anecdotal

PAGE 77

64 reports of significant concentrations followed by prolonged periods of few sightings from July to October, but this also could be explained by local post breeding dispersals into more remote, less visible habitats such as tall grass swamps, as evidenced by the storks tracked by satellite. Jabiru storks displayed the ability to soar at high altitudes (> 1000 m) and to cover large distances rapidly. It is possible that the Belize population undertakes regional movements north into the Usumacinta drainage of Mexico or west into Guatemala. However, such movements most likely would be in response to extreme weather conditions, such as the excessive rains associated with hurricanes, or severe, localized droughts, and may not represent true migration. For example, two countrywide aerial surveys after hurricane Iris struck Belize in October of 2001 failed to document the presence of Jabiru storks (pers. obs.). It is possible that the entire Belize population dispersed northward to avoid the devastating effects of the storm. I consider the Belize population to be resident and non-migratory, with any regional movements occurring sporadically in response to extreme weather conditions and not as part of a defined, consistently displayed seasonal migration. In Costa Rica, where foraging habitat is more limited, it is possible that a regular seasonal migration may occur. If so, then I would hypothesize that these birds move north into the larger wetlands of the Miskito coast of Nicaragua and La Mosquitia in Honduras. Conservation Implications The existing protected areas network of Belize falls short of providing adequate protection for the country’s Jabiru stork population. The results of this study illuminate clear opportunities to address the conservation needs of this vulnerable population. Since Jabirus use the same nest site in successive seasons and because significant portions of the associated foraging habitats are found within a 3,000 m radius, a 3000 m buffer

PAGE 78

65 would protect both the critical nesting sites and the associated foraging habitats. This buffer would also protect nesting storks from logging and other invasive threats. At a minimum, the security of the core-use areas of all tagged storks should be considered a very high priority for conservation. Ideally, however, the present amount and distribution of the critical habitats identified in this study, namely the Tropical lowland tall herbaceous swamps, the Short grass savanna with needle leafed open forest and the Short grass savanna with shrubs should be protected and maintained across the landscape. Probably the most important protected area for Jabiru storks in Mesoamerica is the Crooked Tree Wildlife Sanctuary in northern Belize. However, flaws in the delineation of this sanctuary severely limit its potential to offer long-term and sustainable protection for Jabiru storks. Expanding this protected area to include the TLTHS on the southern end and the adjacent SGSWNLOF (short grass savanna with needle open forest) along the existing eastern and western perimeters would significantly enhance the effectiveness of this reserve. Incorporating these lands would secure the long term viability of the local population and would allow the Crooked Tree Wildlife Sanctuary to offer sustained protection for Jabirus and a host of sympatric species. A substantial proportion of the core-use areas and active nest sites derive little or no protection from the present protected areas network in Belize (Figure 22). Local conservation biologists, national and international non-governmental organizations, and relevant government agencies must undertake a concerted effort to develop a regional conservation action plan that would ensure the long term viability of the regional population. Threats facing the Belize population include inbreeding depression, small and vulnerable population size, loss of foraging and nesting habitat, and

PAGE 79

66 high juvenile mortality/low recruitment. Any regional conservation plan must consider the ecological processes that maintain these critical Jabiru stork habitats, in particular, the processes that maintain the tropical lowland tall herbaceous swamps. Indiscriminate logging (or other destructive forces) must be avoided since it may take 60 to 100 years for the establishment of large, isolated and emergent Ceiba trees within a forested landscape. The role of fire in maintaining the integrity and productivity of Short grass savanna (with needle leaf open forest) must be clearly understood. While logging or intensive agriculture within buffer zones may not seem compatible, the feasibility of selective logging practices should be carefully studied. The entire hydrology of the foraging areas, including the integrity of all water bodies, water courses, ground water and watersheds must be maintained and pollution of any surface water avoided. Directions for Future Study Despite the Jabiru stork’s natural potential to serve as an ideal flagship species and the associated opportunities to advance regional conservation efforts throughout Mesoamerica, the species remains poorly studied, with critical gaps in our knowledge of its basic natural history, population ecology, and conservation needs. If the species is resident and non migratory, then the survival of the local breeding population in Belize may be at risk from inbreeding depression and the associated reduction in fitness. A study to determine the extent of this risk is necessary. Low post-fledging survival and the influence of predation in guiding this stage of the stork’s life cycle must be assessed. Nest site preference should be studied further in relation to nesting biology to determine how the microhabitat and landscape features influence reproductive success and the population’s intrinsic rate of growth.

PAGE 80

67 A rapid assessment of the Caribbean coastal wetlands of Honduras and Nicaragua must be conducted to estimate population size, to map nesting territories, and to determine the number of breeding pairs. The genetic study previously described should be duplicated on both the Costa Rican and Honduran/Nicaraguan populations. The Costa Rican population, because of its smaller size, may be at greater risk. Expanding the telemetry study to incorporate other regions in Mesoamerica will define movement parameters for the regional population and confirm the distribution and extent of key foraging habitats. A robust telemetry sample can answer key biological questions regarding age at first breeding, examine juvenile mortality and the associated causes, and map travel routes and core-use areas. A comprehensive satellite telemetry study of the regional population offers the only feasible method of obtaining critical answers on a timely basis. The elaboration of a comprehensive and effective regional conservation plan may hinge on this approach. Finally, since the Mesoamerican population is disjunct from the larger one in South America, with no records in southern Costa Rica nor in Panama, a genetic-based study of the species’ phylogeography to identify differences between these populations should be pursued. Describing the extent of genetic isolation for the Mesoamerican population would establish the relative urgency for conservation planning and management action across the region.

PAGE 81

68 22. Map of Belize showing the existing network of protected areas and the critical carried by each bird (n = 1239 – 4052 fixes per bird). Figure conservation areas for Jabiru storks. Critical areas determined by (1) buffered nest locations and (2) estimates of activity ranges for juvenile storks from four nests. Estimates are based on Adaptive Kernel Analysis from actual point locations derived from GPS receivers incorporated in satellite transmitters

PAGE 82

CHAPTER 5 CONCLUSION The extensive geographic rangeabiru storks on existing distribution maps is inaccurate and very misleading. Except for perhaps a few isolated individuals, the Jabiru stork may no longer exist along the pacific coast from Guatemala to southern Nicaragua, and the viable populations elsewhere in Central America may now be limited to three distinct groups – Belize, the Atlantic coast of Honduras (Mosquitia) and Nicaragua (Miskito), and Costa Rica. The window of opportunity for defining and addressing the conservation needs of the Belize population may still be open. However, aggressive deforestation and other destructive land-use trends, exacerbated by the natural vulnerability of the Jabiru stork’s preferred habitat due to location, access, and the potential for agricultural development, represent the most obvious and pressing threats. Such conspicuous threats underscore the urgent need for immediate action. The high natal philopatry displayed by juvenile storks and the non-migratory tendency of the general population presents a challenge for conservation. The very limited distribution of these critical habitat types coupled with the very restricted ranges of the marked juveniles within these critical habitats suggests that these habitats are strongly preferred and essential to the storks. The importance of these habitats and the urgent need for their protection cannot be overemphasized, particularly in the light of national long-term planning projections and the fact that most of these habitats fall outside the existing network of protected areas. The low availability of these critical habitats may strictly limit the Jabiru storks present opportunities for dispersing and may portrayed for J 69

PAGE 83

70 of irreves these llowed n. The y sive explain why this population is very sedentary albeit highly mobile. While this may aggravate any deleterious effects of genetic isolation and small population size, it also presents an opportunity for conservation action. Creating and maintaining a networksuitable habitat consisting of widely dispersed but sufficiently large patches of critical habitats could be very effective given the species high mobility, long life span, and the strong natural tendency to prospect over the wider landscape for favorable conditions. While it may be possible to use artificial platforms and plant or husband nest treesto boost nest site availability and perhaps productivity, the loss of foraging habitats is rsible. These critical foraging habitats are characterized by a unique set of landscape features. Within the landscape, the core-use areas are ephemeral and productivity may differ at varying spatial and temporal scales. For several decadelands have remained relatively intact and undisturbed in Belize. This may have athe Jabiru stork population to remain localized, non-migratory and relatively stable despite its small size. Drastic changes at local and regional scales are now fragmenting the landscape and threatening the ecological integrity of these critical habitats. Alterations to these systems will disrupt ecological processes at the landscape level and reduce the natural potential of these systems to support the Jabiru stork populatiopotential destructive and large-scale impacts of aquaculture, rice farming and ecotourismmust be carefully mitigated. In early 2005, significant portions of critical Jabiru stork habitat were cleared in the immediate area surrounding the Crooked Tree Wildlife Sanctuary for what appeared to be agriculture and ecotourism related projects. Not onldo these actions immediately destroy critical habitat, but the ramifications are pervaacross the entire landscape and pose an immediate threat to the entire Belize population.

PAGE 84

71 d may imate in eithero iverep). ow ution of key foraging habitat and the If these pressures continue at the present rate and the important habitats are significantly altered or reduced, then the already fragile local Jabiru stork population will not withstand these pressures and will succumb. Similar pressures elsewhere in the region may have been responsible for the extirpation of Jabiru storks from some areas anhave resulted in the small, patchy and localized distribution in other areas. In country-wide aerial surveys on July 2 nd and 3 rd 2004 I counted 157 Jabiru storks.This figure, together with the 22 known nest sites, represents the highest est category for any country in Mesoamerica. Based on five years of conducting intensive country-wide aerial surveys and my increased knowledge of this species, I dnot believe that these figures represent an increase in population size. Instead, this apparent increase can be attributed to intensive surveys by skilled observers, increased interest in the species, the confounding effects of shrinking habitat and a sustained fyear effort. Therefore, the suggestion that the Belize Jabiru stork population may have increased in recent decades (Barnhill et al. 2005) may be erroneous and can potentially mislead conservation efforts in Mesoamerica. Ongoing genetic studies comparing the Belizean birds with those from the Brazilian Pantanal indicates significant genetic differentiation (based on microsatellite data) between the two populations and very lowhaplotype diversity (genetic variability) for the Belize population (Lopes et. al. in pEven without the genetic results, and given the clear geographic separation of the Mesoamerican population it can be argued that these Jabiru storks are unique and represent an evolutionary significant unit. The small population size, low genetic variability, high level of inbreeding, ljuvenile survival/recruitment, limited distrib

PAGE 85

72 aggrered. rts. ssive deforestation and development pressures on these critical habitats clearly indicate that the Jabiru stork population in Mesoamerica is now critically endangeThe results of this study offer a real opportunity to efficiently focus conservation effoThe critical areas documented (Figure 22) should be considered immediate priority for local conservation because the viability of the local Jabiru stork population cannot be secured without these habitats.

PAGE 86

LIST OF REFERENCES Aebischer, N. J., P. A. Robertson and R. E. Kenward. 1993. Compositional analysis of habitat use from animal radio-tracking data. Ecology. 74(5):1313-1325. Antas, P. T. Z. and I. L. S. Nascimento. 1997. Tuiui: Under the skies of the Pantanal: biology and conservation of the Tuiui, Jabiru mycteria. Empresa das Artes, So Paulo, Brazil. Barnhill, R. A., D. Weyer, W. F. Young, K. G. Smith and D. A. James. 2005. Breeding biology of Jabirus (Jabiru mycteria) in Belize. Wilson Bulletin. 117(2):142-153. Brooks, T. M., S. L. Pimm and J. O. Oyugi. 1999. Time lag between deforestation and bird extinction in tropical forest fragments. Conservation Biology. 13(5):11401150. Brooks, T. M., R. A. Mittermeier, C.G. Mittermeier, G.A.B. Da Fonseca, A.B. Rylands, W. R. Konstant, P. Flick, J. Pilgrim, S. Oldfield, G. Magin and C. Hilton-Taylor. 2002. Habitat loss and extinction in the hotspots of biodiversity. Conservation Biology. 16(4):909-923. Camacho, M. G. 1983. Notes on aquatic birds of Nicaragua, Jabiru mycteria. Inst. Nicarag. Rec. Nat. Ambient. 14 pp. Central Statistical Office (CSO). (2004). http://www.cso.gov.bz . (accessed March 2005). Correa, J. and C. Luthin. 1988. Proposal for the conservation of Jabiru storks in Southeast Mexico. Pp 607-615, In Memoria del simposio sobre la ecologa y conservacin del delta de los ros Usumacinta y Grijalva. Instituto nacional de investigacin sobre recursos biticos, divisin regional, Tabasco, Mxico. De Vries, G. W., M.F. Haines, S.B. Hufnagel, A.K. Laird, K.D. Rearick and O.E. Salas. 2003. Enhancing collaboration for conservation and development in southern Belize. Master’s Thesis. University of Michigan. Del Hoyo, J., A. Elliot and J. Sargattal. Eds. 1992. Handbook of the birds of the world. Vol 1. Lynx edicions, Barcelona. Deignan , H. G. 1933. The Jabiru (Jabiru mycteria) in western Guatemala. Auk. 50(4):429 73

PAGE 87

74 Dickey, D. R. and A. J. Van Rossem. 1928. The birds of El Salvador. Field Museum of natural history zoological series no. 23. 609 pp. Dixon, K. R. and J. A. Chapman. 1980. Harminic mean measure of animal activity areas. Ecology. 61:1040-1044. Dykstra, C. R., J. L. Hays, F. B. Daniel and M.M. Simon. 2000. Nest site selection and productivity of suburban Red-shouldered hawks in southern Ohio. Condor 102:401-408. Eberl, C. and J. Picman. 1993. Effect of nest-site location on reproductive success of Red-throated loons (Gavia stellata). Auk. 110(3):436-444. Ek, E. 2004. Monitoring land use and land cover changes in Belize, 1993-2003: A digital change detection approach. Master’s Thesis. Ohio University. Food and Agricultural Organization (FAO). 1997. State of the world’s forests 1997. FAO, Rome, Italy. Frederick, P. C. J. C. Sandoval, C. Luthin and M. Spalding. 1997. The importance of the Caribbean coastal wetlands of Nicaragua and Honduras to Central American populations of waterbirds and Jabiru Storks. Journal of Field Ornithology. 68(2):287-295. Gonzales, J. A. 1996. Breeding biology of the Jabirus in the southern llanos of Venezuela. Wilson Bulletin. 108(3):524-534. Hancock, J. A., J. A. Kushlan, and M. P. Kahl. 1992. Storks, ibises and spoonbills of the world. Academic Press, San Diego, California. Hartasanchez, I. E. 1992. Aspectos ecolgicos de los humedales alrededor de la laguna de trminos con nfasis en especies ciconiformes Delta Usumacinta-Trijalva, Campeche, Mxico. Hooge, P. N. and B.Eichenlaub. 1997. Animal movement extension to ArcView. ver. 1.1. Alaska Science Center–Biological Science Office, U.S. Geological Survey, Anchorage, AK, USA. Howell, T. R. 1972. Birds of the lowland pine savanna of Northeastern Nicaragua. Condor. 74(3):316-340. Howell, S. N. and S. Webb. 1995. A guide to the birds of Mexico and northern Central America. Oxford University Press. Oxford. 849 pp. Jenness, J. 2004. Alternate animal movement routes (altroutes.avx) extension for ArcView 3.x, v.2. Jenness Enterprises. Available at: http://www.jennessent.com/arcview/alternate_routes.htm .

PAGE 88

75 ——. 2005. Random point generator (randpts.avx) extension for ArcView 3.x, v. 1.3. Jenness Enterprises. Available at: http://www.jennessent.com/arcview/random_points.htm . Jones, J. 2001. Habitat selection studies in avian ecology. Auk. 118(2):557-562. Kahl, M. P. 1971. Observations on the Jabiru and Maguari storks in Argentina, 1969. Condor. 73:220-229. Komar, O. 1998. Avian diversity in El Salvador. Wilson Bulletin. 110(4):511-533. Lopez-Ornat, A., J. F. Lynch, B. M. De Montes. 1989. New and noteworthy records of birds from the eastern Yucatn Peninsula. Wilson Bulletin. 101(3):390-409. Lopez-Ornat, A. and C. Ramo. 1992. Colonial waterbird populations in the Sian Ka’an biosphere reserve (Quintana Roo, Mexico). Wilson Bulletin. 104(3):501-515. Luthin, C. 1984. World working group on storks, ibises and spoonbills. Report 2, 1984. ICPB-Brehm Fonds. Walsrode. ——. 1987. Status of and conservation priorities for the world’s storks species. Colonial Waterbirds. 10:181-202. Massey, F. J. 1951. The Kolmogorov-Smirnov test of goodness of fit. Journal of the American Statistical Association. 46 (253): 68-71. Meerman, J. C., and W.Sabido. 2001. Central American Ecosystems Map – Belize Volume 1. Programme for Belize, Belize City, Belize. Miller, B. W. and C. M. Miller. 1995. National protected areas management plan and zoological report: faunal and site analysis. National protected areas systems plan for Belize, 3. Zoological report. Belize City: Ministry of Natural Resources and USAID. ——. 1998. Ornithology in Belize since 1960. Wilson Bulletin. 110(4):554-558. Monroe, B. L. 1968. A distributional survey of the birds of Honduras. Ornithological Monographs No. 7. 458 pp. Ogden, J. C., C. E. Knoder and A. Sprunt Jr. 1988. Colonial waterbird populations in the Usumacinta Delta, Mxico, p 565-605, in Ecologa de los ros Usumacinta y Grijalva. Instituto nacional de investigaciones sobre los recursos biticos. Tabasco, Mxico. Paredes, A. J. 2004. Status, distribution, habitat requirements, and foraging ecology of the Jabiru stork (Jabiru mycteria) in northern and central Belize, Central America. Master’s thesis. University of Florida.

PAGE 89

76 Penterianni, V., B. Faivre and B. Frochot. 2001. An approach to identify factors and levels of nesting habitat selection: a cross-scale analysis of Goshawk preferences. Ornis Fennica. 78:159-167. Poveda, M. G. 2003. El comportamiento de Jabiru mycteria durante la poca reproductiva en humedales de la zona norte de Costa Rica. Boletn Zeledonia 7(1), Junio, San Jos Costa Rica. www.zeledonia.org (accessed March 2005). Ridgley, R. S. and J. A. Gwynne Jr. 1989. A guide to the birds of Panama with Costa Rica, Nicaragua and Honduras. Princeton University Press. Princeton. 534p. Russell, S. M. 1964. A distributional study of the birds of British Honduras. Ornithological Monographs No. 1. AOU. 458 pp. Salvin, O. and F. D. Goodman. 1897-1904. Biologia Centrali-Americana: AVES. Vol III (Text). Taylor and Francis. Scott, D. A. and M. Carbonell. 1986. A directory of neotropical wetlands. International Waterfowl Research Bureau, Slimbridge, United Kingdom. Seamans, M. E. and R. J. Gutierez. 1995. Breeding habitat of the Mexican Spotted Owl in the Tularosa mountains, New Mexico. Condor. 97:944-952. Seigel, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill Book Co., New York. Shiraki, S., 1994. Characteristics of White-tailed Sea Eagle nest sites in Hokkaido, Japan. Condor. 96:1003-1008. Sibley, C. G. and B. L. Monroe. 1990. Distribution and taxonomy of birds of the world. New Haven, Connecticut, Yale University Press. Slud, P. 1964. The birds of Costa Rica: Distribution and ecology. Bulletin of the American museum of natural history. Vol. 128. New York, New York, USA 430 pp. Sokal, R. R. and F. J. Rohlf. 1995. Biometry: The principle and practice of statistics in biological research, 3 rd ed. W. H. Freeman, New York. Spaans, A. L. 1975. The status of the Wood Stork, Jabiru, and Maguari Stork along the Surinam coast, South America. Ardea. 63:116-130. Stevens, J. 1986. Applied multivariate statistics for the social sciences. Lawrence Erlbaum Associates, Hillsdale, NJ. Stiles, F. G. and F. Skutch. 1991. A guide to the birds of Costa Rica. Cornell University Press, Ithaca, New York, USA. 511pp.

PAGE 90

77 Tashian, R. E. 1953. The birds of southeastern Guatemala. Condor. 55(4):198-210. Thomas, B. T. 1981. Jabiru nest, nest building and quintuplets. Condor. 83:84-85. ——. 1985. Coexistence and behavior differences among the three western hemisphere storks, pg. 921-931, In P. A. Buckley, M. S. Foster, E. S. Morton, R. S. Ridgley and F. G. Buckley. Eds. Neotropical ornithology. Ornithological monograph No. 36. American Ornithologist’s Union, Washington, D. C. Villalobos, L. 1995. Reporte de Jabiru mycteria (Ciconiiformis:ciconiidae) en la costa Pacifica de Honduras. Brenesia. 43-44:91-92. Villareal-Orias, J. 1995. Estado actual del Galan sin Ventura Jabiru mycteria en Guanacaste, Costa Rica. Cotinga. 3:54-55. ——. 1997. Estado actual, presas y uso de habitat del Jabiru (Jabiru mycteria) en la cuenca del Rio Tempisque, Costa Rica. Universidad Nacional. Thesis. Heredia. 104pp. ——. 1998. A new nesting record for the Jabiru in Costa Rica. Colonial Waterbirds. 21(2):256-257. Wetmore, A. 1965. The birds of the republic of Panama-Part 1: TINAMIDAE (Tinamous) to RYNCHOPODIDAE (Skimmers). Smithsonian Institution. Washington. 483 p. World Bank and CCAD. 2000. Ecosystems of Central America (Arcview map files at 1:250,000). World Bank, Comisin Centroamerica de Ambiente y Desarrollo (CCAD), World Institute for Conservation and Environment (WICE), and the Centro Agronmico Tropical de Investigacin y Enseanza (CIAT), Washington, D.C. ( http://www.worldbank.org/ca-env ) (accessed January 2005).

PAGE 91

BIOGRAPHICAL SKETCH Omar Antonio Figueroa was born in 1969 in San Ignacio Town, Belize, Central America. In December of 1996, he obtained his bachelor’s degree (cum laude) from the University of North Florida. Immediately thereafter, Omar returned home and for the next five years coordinated an avian research and conservation project. During this five-year period his volunteer initiatives included one year on the board of the Mesoamerican Society for Biology and Conservation, three years as the national representative for the regional Partners in Flight Initiative and four years as the national coordinator for national wetland surveys. In August of 2003, Omar was awarded a Fulbright/OAS Ecology Initiative Fellowship to pursue his Master of Science degree. In 2005, he was awarded a Dexter Fellowship and now intends to pursue a doctoral degree in wildlife ecology and conservation. Omar intends to conduct all research related to his graduate degree in his native country of Belize and is anxious to return home upon the completion of his study program. 78