1 THE ECOLOGY AND CONSERVATION OF JAGUARS ( Panthera onca ) IN CENTRAL BELIZE: CONSERVATION STATUS, DIET, MOVEMENT PATTERNS AND HABITAT USE By OMAR ANTONIO FIGUEROA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Omar Antonio Figueroa
3 To D esi, Rhiannon, Mary Elizabeth, Kylie and Justin To Superman
4 ACKNOWLEDGMENTS A study such as this could never be a singular effort. I am extremely grateful and indebted to so many individuals and numerous organizations. At the risk of failing to mention or acknowledge individuals who have s omehow helped along the way I hereby venture to express my sincere gratitude to the many folks who have provided assistance and guidance along this long and many times difficult endeavor. At the university level, I would not have been able to reach this mi lestone were it not for the help and guidance of Susan Jacobson, my major professor. Simply put, without her help and support I would not have been able to finish this degree program. I am extremely grateful for her help, guidance and for the critical revi ew on my dissertation. I am also immensely grateful for the other member of my committee, Ken Meyer, Eric Keys, Leonard Pearlstine Peter Frederick and Howard Quigley for all the help and guidance provided during the development of the research proposal up until the final dissertation was completed. Thanks so much for assisting on the academic front. I thank Wilber Martinez, Reynold Cal, David Tzul and Gilford Hoare for their valuable field assistance in what was sometimes extremely difficult field condit ions. Their support and unwavering commitment to this research project was partly responsible for the eventual success. There is no other field team I would prefer to be a part of. To Gil Boese, Lillian Boese and David Gegg, thanks for allowing these guys to work with me and for the many financial contributions. To Sharon Matola, thanks so much for the help and support. I thank Ginny Fifield, John and Nancy Kennedy, the Protected Areas Conservation Trust, the Columbus Zoo, the Felidae Conservation Fund, the International foundation for Science, the Belize Zoo and Tropical Education
5 Telonics Northstar Science and Technology, the Foundation for Wildlife conservation, the Birds Without Borders Aves Sin Fronteras project, the Chester Zoo and the Belize Forest Department for all the assistance, financial and otherwise. I must especially thank Alan Rabinowitz and Panthera for believing in me after the first set of collars failed. I also especially thank Howard Quigley for help and guidance throughout the many phases of this project. His guidance was critical to the eventual success of this project. I thank my wife, Desi, without whose unwavering support and encouragement I woul d have surely faltered. You held forth while I pursued these magnificent cats. I cannot express how grateful I am for the great job you have done with the kids and how fortunate I am to have you. I am excited and optimistic about our future! We have a lot of catching up to do. I also thank Lloyd and Jessie Graniel for all the help and support over the past few years especially with the kids I thank my Parents for their unconditional support. I will never be able to thank them enough. I thank my brother, A rsenio and his wife, Martha, for their encouragement and support. Rhiannon will always be grateful for the time she spent at baby gator. To those I may have missed, and I know you are out there somewhere, I apologize for the oversight but rest assured I wi ll always be grateful for the help and support. In regard to the second chapter of this dissertation, w e are thankful to the Forest Department for making available copies of reports and research proposals. We appreciate the continued interest and substantive support provided by Alan Rabinowitz,
6 Howard Quigley and Panthera for jaguar research and conservatio n in Belize. This effort was directly or indirectly financed and supported by the Belize Forest Department, Tropical Conservation, Virginia Tech, The Philadelphia Zoo, an d AcornAlcinda Foundation. The Protected Areas Conservation Trust, Belize Audubon Society, Belize Zoo, Brevard Zoo, Columbus Zoo, International Foundation for Science, Chester Zoo, Foundation for Wildlife Conservation, Kaplan Scholarship award, Liz Claibor ne and Art Ortenberg Foundation, North of England Zoological Society, Runaway Creek Nature Reserve, Telonics, Felidae Conservation Fund, Wildlife Conservation Society, UK Darwin Initiative, UK Natural Environment Research Council, University of Belize, Uni versity of Idaho Laboratory for Ecological, Evolutionary and Conservation Genetics, University of Southampton, Virginia Fifield, Woodland Park Zoo have made important contributions to the research and conservation of jaguars in Belize.
7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 11 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 16 2 THE STATUS OF JAGUAR ( Panthera onca ) CONSERVATION IN BELIZE .......... 22 Synopsis ................................ ................................ ................................ ................. 22 Overview ................................ ................................ ................................ ................. 23 Study Area ................................ ................................ ................................ .............. 25 Methods ................................ ................................ ................................ .................. 26 Policy and Legal Framework ................................ ................................ ............ 26 Ecolog ical and Conservation Research ................................ ............................ 26 Threats to Jaguar Conservation ................................ ................................ ....... 27 Outcome ................................ ................................ ................................ ................. 27 Policy and Legal Framewo rk ................................ ................................ ............ 27 Protected Areas ................................ ................................ ................................ 28 Ecological and Conservation Research ................................ ............................ 31 Threats to Jaguars ................................ ................................ ........................... 36 Discussion ................................ ................................ ................................ .............. 40 General Recommendations and Propos ed Actions ................................ ................ 43 3 STUDY DESIGN AND METHODOLOGY ................................ ............................... 50 Study Area ................................ ................................ ................................ .............. 50 Diet ................................ ................................ ................................ ......................... 51 Scat Collection ................................ ................................ ................................ 51 Food Niche Breadth ................................ ................................ ......................... 53 Food Niche Ove rlap ................................ ................................ ......................... 54 Mean Weight of Vertebrate Prey ................................ ................................ ...... 54 Capture and Immobilization ................................ ................................ .................... 54 Home Range and Movement Patterns ................................ ................................ .... 56 Use of Habitat and Landscape Features ................................ ................................ 60 4 STUDY OUTCOMES AND DATA SUMMARIES ................................ .................... 66 Diet of Jaguars and Pumas ................................ ................................ ..................... 66
8 Capture and Telemetry ................................ ................................ ........................... 68 Home Range Sizes ................................ ................................ ................................ 69 Movement Patterns ................................ ................................ ................................ 71 Landscape and Habitat Use ................................ ................................ .................... 73 5 SYNTHESIS AND CONSERVATION IMPLICATIONS ................................ ......... 128 Diet ................................ ................................ ................................ ....................... 128 Home Range and Movement Patterns ................................ ................................ .. 135 Landscape and Habitat Use ................................ ................................ .................. 143 6 CONCLUSION ................................ ................................ ................................ ...... 148 APPENDIX A LETTER CODES USED IN THE CART ANALYSIS. CODES BASED ON THE CENTRAL AMERICAN ECOSYSTEM MAP: UNESCO CLASSIFICATION. ........ 15 1 B LETTER CODES USED IN THE CART ANALYSIS. CODES BASED ON THE CENTRAL AMERICAN ECOSYSTEM MAP: ECOSSYTEM CLASSIFICATION. 152 LIST OF REFERENCES ................................ ................................ ............................. 153 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 163
9 LIST OF TABLES Table page 2 1 Protected areas that have a role to play in protecting jaguar populations in Belize. ................................ ................................ ................................ ................. 46 2 2 Jaguar density estimates for different sites in Belize. ................................ ......... 46 2 3 Estimate for the number of jaguars in the two main conservation blocks in Belize. ................................ ................................ ................................ ................. 47 4 1 Total number of scats and prey items per species. ................................ ............ 79 4 2 Percent occurrence and relative biomass of prey consumed in 79 jaguar scats (n= 102 prey items). ................................ ................................ .................. 79 4 3 Pe rcent occurrence and relative biomass of prey consumed in 36 puma scats (n=42 prey items). ................................ ................................ ............................... 80 4 4 Food Niche Breadth (diet diversity) and Mean Weight of Vertebrate Prey (MWVP) of jaguars, pumas, male jaguars and female jaguars. .......................... 80 4 5 Food Niche Overlap (diet similarity) between jaguars and pumas, male jaguars and pumas, and female jaguars and male jaguars. ............................... 80 4 6 Percent occurrence and relative biomass of prey consumed in 65 Male jaguar scats (n= 82 prey items). ................................ ................................ ........ 81 4 7 Percent occurrence and relative biomass of prey consumed in 10 female jaguar scats (n= 15 prey items). ................................ ................................ ......... 81 4 8 Sex, age and body measurements of 16 jaguars ( Panthera onca ) captured in central Belize during 2008 2010. ................................ ................................ ........ 82 4 9 Sex, age and body measurements of 8 pumas ( Puma concolor ) captured in central Belize during 2008 2010, including two individuals* with functional collars th at provided data for this study. ................................ ............................. 83 4 10 Total trapping effort: number of captures, recaptures and trap nights per capture. ................................ ................................ ................................ .............. 83 4 11 Estimated size of annual home ranges (km 2 ) of jaguars ( Panthera onca ) and pumas ( Puma concolor ) in central Belize.. ................................ ......................... 84 4 12 Estimated size of the dry season home ranges (km 2 ) of seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in centra l Belize. .................. 84
10 4 13 Estimated size of the rain season home ranges (km 2 ) of seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. .................. 85 4 14 Mean daily distances moved (km/day), maximum daily distances moved (km/day) and total distance travelled for the duration of the study for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. ..... 85 4 15 M ean daily distances moved (km/day), maximum daily distances moved (km/day) and total distance travelled during the dry and rain seasons for seven jaguars and two pumas in central Belize. ................................ ................ 86 4 16 Results of spatial cluster analysis (average nearest neighbor, method = Euclidian) procedure used to test the hypothesis that the activity pattern displayed by collared cats could be a result of random chance. ......................... 87 4 17 Habitat selection using the habitat utilization method of Neu et al. (1974) with Bonferroni confidence intervals. ................................ ................................ ......... 88 4 18 Third order habitat selection using the Neu et al. (1974) habitat utilization method with Bonferroni confidence i ntervals. ................................ ..................... 89 4 19 Habitat utilization based on the Neu et al. (1974) method comparing proportions of habitat based on GPS location fixes versus proportions within the total study area (TSA). ................................ ................................ .................. 90 4 20 Compositional analyses testing second order habitat selection a nd ranking matrices for jaguars ................................ ................................ ............................ 91 4 21 Compositional analysis testing third order habitat selection and ranking matrices for jaguars ................................ ................................ ............................ 92 4 22 Correct classification rates for all nine classification trees.. ................................ 93
11 LIST OF FIGURES Figure page 2 1 Network of protected areas and the central jaguar corridor in relation to land cover in Belize. ................................ ................................ ................................ ... 48 2 2 main protected blocks. ................................ ................................ ........................ 49 3 1 Map of Central Belize showing the general study area with major roads, communities, land cover and protected areas. ................................ ................... 65 4 1 Spatial distribution of 136 felid scats collected in Central Belize. ....................... 94 4 2 Annual home range estimates (95% FK) for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ). ................................ ................................ ....... 95 4 3 Dry season home ranges (95% FK) for six jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. ................................ .......................... 96 4 4 Rain season home ranges (95% FK) of seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. ................................ .................... 97 4 5 Cumulative home range sizes (km 2 ) based on minimum convex polygon estimates calculated at two week intervals. ................................ ........................ 98 4 6 Plot of log10 net displacement (absolute distance from a given starting point) against log10 time for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ).. ................................ ................................ ................................ ........... 99 4 7 Smoothed model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies wit h time of day. ........... 100 4 8 Smoothed model showing how log10 transformed movement rates (distance moved per hour) for pumas ( Puma concolor ) varies with time of day. ........... 101 4 9 Smoothed model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies with days of the year. .... 102 4 10 Smoothed model showing how log10 transformed movement rates (di stance moved per hour) for pumas ( Puma concolor ) varies with days of the year. .... 103 4 11 Model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies with the ecosystem type the movement initiated in ................................ ................................ ........................ 104
12 4 12 Model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies depending on whether the movement crossed a road or did not cross a road. ................................ ........... 105 4 13 Variability of movement rates (kilometers per hour) for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ). ................................ .......... 106 4 14 Spatial cluster analyses for all nine cats generated z values which were less than 2.58, thereby indicating that the spatial pattern demonstrated by the movement of all cats was clustered. ................................ ................................ 107 4 15 Proportions of habitat based on fixed kernel probability polygons (95%) utilized by seven jaguars ( Panthera onca ) and two p umas ( Puma concolor ) in central Belize. ................................ ................................ ................................ ... 108 4 16 Proportions of habitat that was available to seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. ................................ ........... 109 4 17 Pruned classification tree model (Dendrogram) for Cat 01.. ............................. 110 4 18 Pruned classification tree model (Dendrogram) for Cat 02. .............................. 111 4 19 Pruned classification tree model (Dendrogram) for Cat 03.. ............................. 112 4 20 Pruned classification tree model (Dendrogram) for Cat 04. .............................. 113 4 21 Pruned classification tree model (Dendrogram) for Cat 05. .............................. 114 4 22 Pruned classification tre e model (Dendrogram) for Cat 06 ............................... 115 4 23 Pruned classification tree m o del (Dendrogram) for Cat 07 ............................... 116 4 24 Pruned classification tree model (Dendrogram) for Cat 08. .............................. 117 4 25 Pruned classification tree model (Dendrogram) for Cat 09. .............................. 118 4 26 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 01. ................. 119 4 27 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 02. ................. 120 4 28 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 03 ................. 121 4 29 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 04.. ................ 122
13 4 30 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 05.. ................ 123 4 31 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 06.. ................ 124 4 32 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 07.. ................ 125 4 33 Distribution of correct classification rates (CCR) for 1,000 randomly generat ed trees versus the CCR of the model (triangle) for Cat 08.. ................ 126 4 34 Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 09.. ................ 127
14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE ECOLOGY AND CONSERVATION OF JAGUARS ( Panthera onca ) IN CENTRAL BELIZE: CONSERVATION STATUS, DIET, MOVEMENT PATTERNS AND HABI TAT USE By Omar Antonio Figueroa May 2013 Chair: Susan K. Jacobson Cochair: Kenneth Meyer Major: Wildlife Ecology and Conservation Effective conservation strategies for large, long lived and wide ranging carnivores such as the jaguar ( Panthera onca ) depend on comprehensive understanding of spatial needs and habitat requirements. Given the accelerating patterns of human population growth and habitat fragmentation, assessments at large spatial scales are necessary We used GPS (Global Positioning System ) tracking technology to investigate movement patterns, habitat use and home range dynamics of Maya in the northwest and the Maya M ountain massif in the south. Accurate information on diet of elusive predators is also an important requisite for the development of effective management and conservation strategies, especially in human dominated landsca pes where anthropogenic forces, especially subsistence hunting, represent an immediate and direct source of competition for a dwindling prey base. The observed habitat used by jaguars was significantly different from expected for the total annual period, a nd both the dry and rain seasons. Jaguars avoided
15 agriculture and wetlands but preferred Lowland Broadleaf Forest and Shrublands (P<0.001). Jaguars showed strong diurnal and seasonal cycles in movement patterns: movement rate at mid afternoon was three tim es less than from midnight to mid morning, and movement rate during the dry season was on average two times the rate during the rain season. Pumas showed weaker diurnal and seasonal patterns. Diet of jaguars and pumas were constructed, and resource part itioning evaluated, from analyses of 136 scats containing 161 prey items. Scat origin, by genetic determination, showed that 79 scats (69 from males and 10 from females) were from jaguars and 36 were from pumas (with 21 of unknown origin). Jaguar diet was dominated armadillos ( Dasypus novemcinctus ) and puma diet was by pacas ( Agouti paca ). The collared peccary ( Tayassu tajacu ) was important to both diets. Female jaguars showed higher diet diversity and diet overlap was highest between female jaguars and pum as. The elevated paca content in the puma diet suggests that pumas may be at a higher relative risk since pacas are a preferred game meat.
16 CHAPTER 1 INTRODUCTION Conservation strategies often focus on emblematic or flagship species that are good indicators of the integrity of ecological systems. Large carnivores have potential to serve as good indicators because their large home ranges and spatial requirements make them vulnerable to processes such as habitat fragmentation or degradation that affect ecological integrity. The social and economic needs of expanding human populations, either directly or indirectly, alter entire landscapes and the myriad of species that are dependent on these systems. As landscapes become fragmented and contiguous habitats become smaller and more isolated, wildlife extirpations will increase across space and time. Wide ranging carnivores are particularly vulnerable to these landscape lev el processes and will be disproportionately represented in the patterns of local extinctions (Soule and Terborgh, 1999), particularly in areas where human populations are expanding (Woodroffe 2000). The primary focal species of this study was the jaguar ( Panthera onca ), with a secondary focus on the puma ( Puma concolor yet lower conservation concern. Historically, jaguars ranged from the southern USA through Central and South America all the way to Argentinean Patago nia. However, since the mid 1900s jaguars have been extirpated from broad portions of this range and now exist in populations that are fragmented (Swank and Teer, 1989, Nowell and Jackson, 1996) and with variable levels of connectivity. It is the largest f elid in the Americas (Sunquist and Sunquist, 2002) and of conservation concern throughout its geographic range have resulted in local extirpations (Seymour 1989, Swank and Teer,
17 1989, Sanderson et al. 2002). Swank and Teer (1989) reported that the distribution in Mexico and Central America shrunk to 33% of its original range and that 75% of remaining populations existed in reduced number s Due to significant reductions in geograp hic range and population size and a declining population trend, the jaguar is considered globally Near threatened (IUCN 2012). Patterns of shrinkage are evident, especially in Mesoamerica where habitat fragmentation, habitat loss and substantial human popu lation growth have exerted significant pressures on this species. A re evaluation of the conservation status of jaguars, across its entire range, is necessary and should be considered a priority. The geographic range of the puma is the most extensive of any terrestrial mammal in the Americas (Sunquist and Sunquist 2002), ranging from Canada and the western United States, through Central and South America, all the way to the southern tip of Chile. The puma is considered of Least Concern, but this is parti ally a reflection of its widespread geographic range (IUCN 2012). While numerous detailed studies have been conducted on the puma in western North America, very little is known about pumas in Central America (see Hornocker and Negri 2010, Sunquist and Sunq uist 2002). Pumas and jaguars are sympatric across the entire range of the jaguar but conservation efforts on the more charismatic jaguar have often overshadowed those for the puma. Habitat loss and fragmentation across the entire geographic range are majo r threats to the long term survival of this species. Pumas have been extirpated from large portions of its historic range in North America (Nowell and Jackson 1996) and are considered to be declining (IUCN 2012). The large home ranges, conservation needs a nd habitat requirements of jaguars and pumas present a myriad of challenges.
18 Research projects that advance our knowledge in these focal areas are necessary to design and promote effective management and conservation strategies. Conservation of large long lived, wide ranging carnivores such as the jaguar and the puma is contingent on comprehensive understanding of habitat requirements and spatial needs. Large scale assessments are necessary given the accelerating patterns of habitat fragmentation and human population growth rate. However, the cryptic behavior, low population densities and large home ranges of these large carnivores complicates sampling at required spatial and temporal scales. The advent of the global positioning system (GPS) tracking technology has created new opportunities to better understand the ecology and conservation needs of these elusive species. While GPS and satellite tracking in tropical forests still presents a multitude of challenges it does offer opportunities to answer i mportant questions that were previously difficult, if not impossible, to address. Central Belize presents a unique opportunity to study the ecology and conservation of jaguars and pumas within a human dominated landscape that is of immense ecological sig nificance. The diverse array of land uses includes protected areas (national parks, wildlife sanctuaries and private protected lands), rural communities, agriculture (citrus, cashew, fish, livestock), logging, oil exploration and eco tourism. These land us es are all nestled within an area that sustains these two top carnivores and the prey species upon which their survival depends. This area also represents the only remaining zone of connectivity between the Selva Maya forest in northwestern Belize (and the contiguous forests in northeastern Guatemala and southern Mexico) and the larger Maya Mountain massif in southern Belize ( C hapter 2).
19 This research project represents the first study of the ecology of jaguars and pumas in this important area. Ecosystem level conservation strategies focusing on these top predatory species can be effective conservation tools, especially in areas where human induced pressures are shaping the landscape at accelerated rates. Understanding factors that influence size and place range and spatial and temporal patterns of resource utilization within this range are vital components of species specific management and conservation strategies. Understanding the patterns of movement within defined home ranges ca n provide important information to guide management and conservation strategies of these wide ranging species. The distribution and abundance of prey is another important factor in the conservation of large carnivores but since prey base can vary from regi on to region, in terms of species diversity, abundance or availability, and these in turn can vary with anthropogenic factors (socio economics, patterns of land use / land cover changes, culture), any attempt to devise a conservation strategy must consider the unique set of factors that shape the focal area. These critical components which affect the ecology and conservation of these top specific conservation strategies are important for jaguars (Quigley and Crawshaw 1992) and pumas, and may offer the best chance of success. This study attempts to provide previously unavailable information on the ecology and conservation of jaguars in central Belize. Data generated from t his study has already been used to guide the establishment of the Laboring Creek Jaguar Corridor Wildlife Sanctuary and a nationally
20 important agreement on the importance of ecological connectivity in central Belize, discussed in Chapter 2. This disser tation has four general components. Chapter 2 serves as a general introduction and discusses the status of jaguar conservation in Belize, with particular focus on the central region where this study was conducted. This chapter will be published in the upco three focal areas: diet, use of habitat and landscape, and movement patterns. The overall methodology and results, partitioned according to these three core areas, are presented in C hapters 3 and 4 respectively. A detailed discussion of the results is presented in C hapter 5 and a general conclusion in C hapter 6 This study advances our understandi ng of the ecology and conservation of jaguars in a human dominated landscape. This is important because given the continent wide patterns of land use and land cover changes, the future of jaguar conservation will partly depend on the proper management of j aguar populations within these areas. First, we analyze and compare the food habits of jaguars and pumas by quantifying and characterizing the diet of both species. Accurate information about the diet of jaguars and pumas is an important component of any m eaningful conservation strategy, especially in areas like central Belize where important prey species such as the White lipped peccary ( Tayassu pecari ) ha ve been extirpated. We also evaluate the role of diet in resource partitioning between these two top p redators. Second, this study answers questions about We test the null hypothesis that the jaguars use of habitat is in proportion to availability. We
21 identify the dry season, rain season and annual home range characteristics (size and placement) for individual jaguars, and quantify habitat preferences at both second and third order selection. We landscape scale habitat variables such as terrain ruggedness, distance to nearest community, perpendicular distance to nearest road, perpendicular distance to western highway and perpendicular distance to nearest river. And third, we quantify diurnal and seasonal movement patterns of jaguars and pumas, a nd assess how movement rate of the jaguar is affected (inhibited or facilitated) by habitat type and road crossing.
22 CHAPTER 2 THE STATUS OF JAGUAR ( Panthera onca ) CONSERVATION IN BELIZE Sy nopsis Belize has a rich history of research and conservation of jaguars a With 63% of the country under forest cover and 35% under protected status, Belize has the potential to safeguard the viability of national jaguar conservation efforts. Given the reg ional patterns of land use and land cover changes over the last decade (e.g. land conversion, increased logging, expansion of human settlements), coupled with the associated pressures of a significant human population growth rate, Belize is strategically p oised to be integral to the stability of jaguar populations in northern Central America and southern Mexico. The national landscape is characterized by two relatively large blocks of protected forests with a collective capacity to sustain a viable populati on of jaguars and must therefore be monitored and managed as a major component of any long term jaguar conservation strategy. A series of imminent threats, including loss of connectivity in the central jaguar corridor, direct persecution, over harvest of w ild prey, and illegal encroachment and poaching in the Maya Mountain Massif must be evaluated, monitored, managed and mitigated. While a robust policy and legislative framework exists which is the focus of current improvement efforts, substantive gaps rema in in areas of law enforcement and monitoring. We provide a comprehensive review of documents relating to jaguar conservation and management in Belize, including peer reviewed publications, reports to governmental and non governmental organizations a Reprinted with permission from Figueroa, O. A., R. J. Foster, C. Wultsch, J. B. Mesa Cruz, M. J. Kelly, B. J. Harmsen, W. Sabido, S. Matola and S. K. Jacobson. 2013. Estatus de conservacion del Jaguar en Belice. En Medellin, R., C. Chavez, A. de la Torre, H. Zarza y G. C eballos (compiladores). El Jaguar en el Siglo XX1: La Perspectiva Continental. Fondo de Cultura Economica. Mexico D. F.
23 and on going scientific research. We assess the status of jaguar conservation, review current threats, and propose recommendations that can serve to safeguard or enhance jaguar conservation in Belize. Overview Belize is the second smallest country in the Mesoamerican region (22,966 km 2 ), confined to the Caribbean coast of northern Central America. With a population of 312,698 inhabitants (CSO, 2010), it is one of the least densely populated nations in Latin Ame rica. However, with an average annual growth rate of 2.65% over the last decade (CSO, 2010), it has the highest rate of population increase in Mesoamerica, and one of the highest in Latin America (Saundry, 2009). The relatively small population size has, u ntil recently, partially confined anthropogenic pressures on wildlife to discrete geographic sectors and thus allowed the country to maintain relatively large, contiguous and undisturbed areas of natural vegetation. In 1995, Belize ranked among the top fou r countries in the Neotropical realm in the proportion of total land mass under some form of forest cover (>84%; FAO 1997). Since then the rural population has increased by 36.7% (CSO 2010) and, coupled with a consistent increase in development pressures, has contributed to a substantial decrease in the proportion of total land mass under forest cover (63% remaining under cover; Cherrington et al., 2010). Relative to Central America where the annual deforestation rate was 1.6% from 1990 to 2000 and 1.2% fro m 2000 to 2010 (FAO, 2010), Belize, with an average annual deforestation rate of 0.6% from 1980 to 2010 (FAO, 2010), stands out as a regionally important player in bio diversity. While agricultural expansion and urbanization in Belize results in an annual loss of approximately 100 km 2 of unprotected forest (Ek, 2004; Cherrington et al., 2010),
24 an otherwise robust network of protected areas suggests that Belize can play a n increasingly important role in securing the viability of nationally, regionally and globally imperiled biodiversity. In order to reverse or abate current trends, managers, policy makers and conservation planners must receive sound and timely ecological r ecommendations so that appropriate conservation and management actions can be taken before that critical period elapses between deforestation and extirpation/extinction (Brooks et al., 1999). While reversing the trend of development may not be possible, it is imperative that decisions regarding development take into account the maintenance of viable wildlife populations at a national level. Given the patterns of deforestation, habitat fragmentation and human population growth, large scale assessments are ur gently needed especially for the understudied, elusive and wide ranging species such as the jaguar ( Panthera onca ). Wide ranging species are particularly vulnerable to landscape level disturbances since viable populations can only be maintained across larg e connected areas of suitable habitat (Crooks, 2002; Noss et al., 1996). Fragmentation of such a landscape can cause local extinctions and finally a complete collapse of the meta populations (Soul and Terborgh, 1999), particularly in areas of high human p opulation density (Woodroffe, 2000). Jaguars are top predators and as such require large areas of contiguous habitat and abundant prey to sustain viable populations (Crooks et al., 2011). Their critical role in maintaining the structure and function of eco systems make them ideally suited to serve as umbrella species for the conservation of coexisting flora and fauna (Crooks and Sanjayan, 2006). In this regard, a comprehensive and large scale assessment of the status of jaguars, re evaluated over time, could serve as an important ecological
25 indicator of the functional integrity of the forested landscape. Belize has a rich conservation history and has actively promoted jaguar research and conservation efforts. It may support one of the largest jaguar populatio ns remaining in Central America (Sanderson et al., 2002) and, therefore, must be considered regionally important. However, within the last decade the accelerating patterns of land use and land cover changes coupled with an associated increase in human jagu ar conflicts represent threats to the viability of local jaguar populations. In this chapter we provide a comprehensive review of the status of jaguar conservation in Belize and make recommendations that can serve to safeguard and perhaps enhance the long term viability of jaguars in Belize. Study Area The entire country of Belize served as the focal area for this study (located between 15 north by Mexico, on the west and south by Gua temala and on the east by the Caribbean Sea. The north consists of flat, mostly swampy coastal plains with seasonally inundated marshy savannahs and heavily forested systems. In the south, low lying coastal areas rapidly give rise to the Maya Mountains rea ching a maximum elevation of 1,124 meters asl. The Maya Mountains comprise the main topographic feature and covers a characterized by a relatively rich and healthy array of f lora and fauna including over 3,750 species of vascular plants and 1,564 species of vertebrates (Meerman, 2006). There are 18 major river watersheds, each with an extensive network of tributaries. and human settlements
26 cover approximately 20%; the remainder is comprised of savannah, scrubland and wetlands (Land Information Centre, 2011). Methods Policy and Legal Framework We reviewed all available documentation of the network of protected forests, national conservation strategies and relevant legislation with the potential to impact jaguar conservation directly and indirectly through the forested systems upon which they depend. This included official reports on the status of the protected areas, lo cal and national conservation strategies, reports on land use and land cover, and all legislations pertaining to protected areas, forest resources and wildlife. Ecological and Conservation Research We conducted a comprehensive literature review on jaguar r esearch and conservation in Belize. This included the peer reviewed literature, university theses and dissertations, published and unpublished reports to local government and non governmental organizations, research proposals and reports submitted to the B elize Forest Department, and on going research. We summarized and sorted all published estimates of jaguar density for Belize. The entire country was partitioned according to land cover, habitat type, geography and protected status/land tenure. First we considered the two main blocks of protected areas, namely, the Maya Mountains in the south and the protected lands in the northwest. Considering the published jaguar density estimates for specific sites within these larger conservation areas, we extrapola ted to derive a population estimate for the two main blocks of protected areas. Considering contiguous land cover, similar habitat and protected status we extrapolated density estimates into adjoining areas.
27 Next, we obtained recent (2010) satellite imag ery and land use maps from the Land Information Centre and used these to determine the size and location of land within the unprotected landscape that was too fragmented, too developed or otherwise unlikely to support jaguars. We then considered the isolat ed array of protected lands and the unprotected matrix within which these are embedded. We summed and sorted all published estimates of jaguar density for sites within this area and extrapolated based on land cover, habitat type and geography. We then comp iled these figures to generate an estimate of the number of jaguars in Belize. Threats to Jaguar Conservation Integrating the reviews of policy and legal framework, and ecological and conservation research, we evaluated the influence and discuss the main t hreats to the long term viability of jaguars in Belize. Outcome Policy and Legal Framework Belize has a robust legislative foundation governing the protection and management of natural resources. The National Parks System Act (CAP 215) and the Forest Ac t (CAP 213) govern the administration and management of protected areas. The Ancient Monuments and Antiquities Act (CAP 330) govern the declaration and administration of Archaeological Reserves. Protected Areas are declared through Statutory Instruments ( SI), which also delineate the boundaries. The National Parks System Act enables the government to create, delineate boundaries, alter boundaries, administer, manage or de gazette National Parks, Wildlife Sanctuaries, Natural Monuments and Nature Reserves. The Wildlife Protection Act (CAP 220) was legislated
28 to afford Belize's wildlife protection by regulating hunting and the commercial dealing of wildlife. All Acts are available online at www.belizelaw.org. The principal tenet of the Wildlife Protection Act (CAP 220, Revised Edition 2000) is to protect the endangered and threatened wildlife of Belize and to regulate the hunting of those species that are traditionally hunted for domestic consumption. The Wildlife Protection Act affords the jaguar protection b y stipulating that it is illegal to actively hunt and kill jaguars for sport, leisure or consumption; however it has provisions that provide for the removal of jaguars when personal safety or damage to personal property is threatened. Frequently, these pro visions have been applied by farmers and others to justify hunting and killing of jaguars. The Forest Department is presently undertaking a revision of the Wildlife Protection Act to strengthen the provisions regarding protection of endangered and threaten ed species, including the jaguar, imposing stricter penalties, regulating the local trade in bush meat and wild caught pets. Additionally, Belize has developed a draft CITES Act which is currently in the last stages of revision to regulate the internation al trade of species listed in the Convention's Appendices thereby limiting even further the indiscriminate hunting and trade of endangered species. Protected Areas Belize has seventy (70) terrestrial protected areas covering approximately 35% of the mainla nd (Land Information Centre 2011). However, since 15 of these protected areas are either too small (< 10 km 2 ), too isolated, devoid of forest cover and/or too close to human settlements and because jaguars have very large spatial requirements with home ran ges up to 530 km 2 (Figueroa unpublished data), these 15 protected areas are not able to make any meaningful contribution to jaguar conservation. Therefore, at
29 least 55 protected areas have a role to play in jaguar conservation and they fall under seven dif ferent categories of protection (Table 2 1). Inclusive in this list are seven private protected areas covering a combined area of approximately 1,309 km 2 These private protected areas are officially recognized as part of the National Protected Areas Netwo rk but a legislative framework endorsing this partnership is yet to be defined. Area (RBCMA, 1,048 km 2 ), is an exception in that it is managed within a robust legal framewor k that secures the long term viability of this protected area. The RBCMA is owned and operated by the local non profit organization Programme for Belize (PfB). The PfB has a legally binding Memorandum of Understanding with the government of Belize which al lows certain fiscal incentives (for example land tax waivers and stamp duty exemption) for the PfB whilst ensuring that the PfB remains committed to conservation and proper development as stipulated in its Memorandum of Association. No other private protec ted area has a legal or otherwise binding commitment to remain under conservation management. Seven different categories of protection exist within the network of protected areas (Table 2 1). The underlying objectives and the management regimes of the dist inct categories differ, however the types of activities allowed within all protected areas are compatible with practices that promote jaguar conservation. The entire national network of protected areas is characterized by two relatively large but separated blocks, each comprised of several protected areas, and an additional array of protected areas scattered across the coastal lowlands (Figure 2 1). A contiguous network of eighteen protected areas covering approximately 5,324 km 2 extends across
30 most of the southwestern portion of the country and comprises one of the two aforementioned blocks. The core of this area is referred to as the Maya Mountain Massif and extends into the larger Maya Mountain range in Guatemala. A smaller but contiguous block of four pr otected areas covering approximately 1,145 km 2 extend from the Selva Maya forests of northeastern Guatemala (the Maya Biosphere Reserve) and southern Mexico (Calakmul Biosphere Reserve) into northwestern Belize and into the central lowland forests of the B elize River valley (Figure 2 1). This Selva Maya forest block forms the largest contiguous block of tropical forest north of the Amazon basin protected forested blocks (Figu re 2 2) increasing the size of the connected Selva Maya forests considerably. The array of isolated protected areas scattered across the central lowlands is approximately 1,170 km 2 While these protected lands are not contiguous with the two main protected blocks they still play an important role in jaguar conservation because most are comprised of forested systems that are contiguous with the unprotected but forested landscape. Belize has two protected areas specifically designated for the protection and conservation of jaguars: in 1986 the Cockscomb Basin Wildlife Sanctuary (CBWS) was designated for the protection of jaguars; and in April 2011 the Labouring Creek Jaguar Corridor Wildlife Sanctuary (LCJCWS) was designated to ensure connectivity and mainten ance of a viable population throughout Belize. In May 2011, the Mango Creek 1 Forest Reserve was extended and re aligned to connect with the CBWS thereby effectively enlarging the latter by approximately 56 km 2
31 Belize is implementing two policy based in itiatives with the potential to impact the viability of jaguar populations. The National Protected Areas Policy and Systems Plan and the National Land Use Policy focus on integrated landscape management approaches to ensure the proper functioning of the ex isting protected areas system and providing oversight of existing and potential land use activities, respectively. The former will ensure that the network of protected areas is managed as one system while the latter will guide the use and development of th e unprotected landscape via a national integrated planning framework. The National Integrated Water Resources Legislation (2010) should offer additional safeguards for jaguars and other wildlife by protecting waterways and watersheds that also serve as cor ridors and refuges. Ecological and Conservation Research The first detailed ecological study of jaguars in Central America was conducted in Belize in the early 1980s (Rabinowitz and Nottingham, 1986). In this landmark study, the authors used VHF telemetry as the primary tool to track and monitor the movement of jaguars in what is now known as the CBWS. Scats were also collected as part of this study and subsequently analyzed to determine jaguar diet. These jaguar scats were later pooled with a wider sample of scats from sympatric felids (ocelot ( Leopardus pardalis ), puma ( Puma concolor ), and jaguaroundi ( Puma yagouaroundi )), and analyzed to determine the prevalence of specific parasites (Patton et al. 1986). These studies affirmed the importance of this area for jaguar conservation. Later, in conjunction with the work of the Belize Audubon Society, these studies laid the groundwork for the establishment of the CBWS. A prolonged period of approximately fifteen years elapsed before jaguars were again the fo cus of ecological research.
32 and additionally research began in the Chiquibul Forest Reserve (CFR), also within the Maya Mountain Massif. Camera traps were the primary research tool used in simultaneous efforts to estimate abundance and density of jaguar populations (Silver et al., 2004). These camera trap studies continued from 2003 to 2008 within the CBWS (Harmsen, 2006; Harmsen et al., 2010a; Kelly et al., 2009) and were expan ded to include areas outside the sanctuary between 2004 and 2006 (Foster, 2008; Foster et al., 2010a). These studies included density estimation across the protected and unprotected landscape, but also investigated livestock predation and lethal control (F oster, 2008), habitat use (Foster et al., 2010a), diet (Weckel et al., 2006; Foster et al., 2010b), spatiotemporal interactions with competitors (Harmsen et al., 2009), marking behaviour (Harmsen et al., 2010b) and activity patterns in relation to their ma in prey (Harmsen et al., 2011); and also addressed methodological issues of studying jaguars and their prey (Foster et al., 2010c; Foster and Harmsen in press; Harmsen et al., 2010c; Harmsen et al., 2010d). Studies to estimate abundance and density of jagu ar populations in the CFR continued from 2002 2005 (Kelly unpublished data). Efforts in this area continued with mixed success until 2008, and were expanded to other areas of Belize (Table 2 2), thereby providing estimates of abundance and density for a wi de cross section of the entire country. Camera trap surveys for jaguars continue on an annual basis in the MPR, RBCMA, and CBWS, with the CFR being surveyed when possible. The ultimate goal is to estimate survival, recruitment, mortality and population gro wth rate from these long term studies.
33 Using published density estimates from twelve different studies (Table 2 2), the note that some of these studies indicated the highest jaguar densities in Mesoamerica and in some cases across the entire geographic range of the jaguar. It is also impor tant to note that these are approximate figures: extrapolation across the entire country has problems, especially in the Maya mountains where the more rugged areas are under b etween 276 to 531 jaguars spread across the two conservation blocks: 211 to 462 in the southern conservation block, and 65 to 69 in the northern conservation block (Table 2 3). These two conservation blocks are separated primarily by the Western Highway an d to a lesser extent by development pressures along the watersheds of the Sibun and Belize Rivers (Figure 2 1). The Coastal and Hummingbird Highways are also fast becoming important barriers. 2 ) fall outside t he two main blocks of protected areas. An isolated array of protected lands that is not contiguous with the two main protected blocks offers protection to some additional forests but significant portions remain without formal protection ( Figure 2 1). Withi n the unprotected landscape, there are at least six estimates of jaguar density; all derived using the same methodology as the studies within the protected areas. Given that the 1,500 km 2 forested block south of the RBCMA and extending into central Belize is similar and contiguous with that of the RBCMA, the maximum and minimum density estimate of 5.7 and 6.01 individuals per 100 km 2 (Waight et al., 2010) were used to derive a population estimate for this area; a minimum and maximum density estimate of 2.36 and 6.18
34 individuals per 100 km 2 (Bunyan and Kilshaw, 2008; Everatt et al., 2010) were used to obtain a population estimate for the 2,500 km 2 swath in northeastern Belize, extending south into the Belize River valley and the central jaguar corridor; and a minimum and maximum density estimate of 3.53 and 7.8 individuals per 100 km 2 (Kelly, 2003; M. Kelly unpublished data) for the 750 km 2 forest block that extends south of the Maya Mountains and into the Sarstoon Temash National Park. Collectively, this prov ides an abundance estimate of 170 to 223 jaguars for the area outside the two main conservation blocks. The remaining 2,100 km 2 of forested land is significantly fragmented and not included in our estimates. Connectivity is the central focus of two natio nally important studies with direct implications for jaguar conservation and management. In central Belize, a concerted, large scale and multi faceted effort is underway to identify and define a critical link between the Maya Mountain Massif in the south a nd the forests in northwestern Belize (RBCMA) which are contiguous with the Maya Biosphere Reserve in Guatemala and initiative is coordinated and implemented by the Environmenta l Research Institute (University of Belize), Panthera and the Belize Forest Department, in partnership with international collaborators (the University of Florida, the University of Southampton and the UK Darwin Initiative). Field research has focused on t he identification of a corridor using telemetry tracking and non invasive methods such as camera trapping, sign surveys and interviews to study site occupancy, abundance and movement patterns of jaguars, pumas and their prey species. This initiative has al so focused on important questions of land tenure, conservation incentives and government acceptance of
35 connectivity principals. It is executed within an adaptive management framework and has resulted in the establishment of a new protected area (LCJCWS) an d recognition by the Government of Belize of the importance of ecological connectivity and a jaguar corridor in central Belize. In southern Belize, site occupancy modeling techniques have been used in an attempt to delineate a corridor connecting the isola ted Sarstoon Temash National Park in the extreme south with the Maya Mountain Massif via the private protected areas of Golden Stream Private Reserve and the TIDE (Toledo Institute for Development and Environment) private reserve (Petracca, 2010). A conser vation and landscape genetics study assessing the status of jaguars, pumas, ocelots, margays ( Leopardus weidii ) and jaguarondis across multiple study sites (CFR and the Chiquibul National Park (CNP), CBWS, Fireburn/Balam Na Nature Reserve, Golden Stream Co rridor Preserve, MPR Forest Reserve, RBCMA, Sarstoon Temash National Park, Shipstern Nature Preserve) is currently underway (2007 present) by researchers from Virginia Tech University, US. This is the first genetic study of jaguars and sympatric felids in the Central American region. Non invasive genetic sampling techniques (detector dogs and molecular scatology) were optimized for use in a tropical environment. Protocols for collecting and storing scat samples were tested to increase the overall DNA amplif ication success from often degraded faecal DNA samples. A highly polymorphic set of genetic markers (14 microsatellite loci) was optimized to efficiently differentiate scat samples down to species and individual level (Wultsch et al., in prep). The genetic markers will provide the means to estimate population densities (through genetic mark recapture analysis), and examine directional gene flow, dispersal patterns and genetic variation and population structure, revealing
36 genetic health and degree of isolati on or fragmentation of wild felid populations. To date, this multi year project has successfully collected 1,190 felid scat samples in the wild. On going genetic analyses has identified 160 individual felids, including 65 individual jaguars, most of which were detected multiple times. Little is known about what factors impact the health of jaguars. Increased levels of human disturbance and conflict with jaguars could increase stress that negatively impacts jaguar health by increasing disease morbidity and mortality, or heightening animal aggression resulting in more conflict. An ongoing study conducted by a team from Virginia Tech in association with The Smithsonian Conservation Biology Institute is using non invasive analyses of faecal glucocorticoid metab olites to assess the impact of human disturbance on stress related hormones in free ranging jaguars. The study will provide comparative data on how the adrenal activity of jaguars in conflict with humans compares with those living in protected habitats. Th ese data will be used to identify sources of potentially detrimental stress for jaguars and help develop plans to reduce human jaguar conflict. Threats to Jaguars Competition with humans for food and space is probably the most important threat facing jagua r conservation. Human jaguar conflict has emerged as a very serious challenge (Brechin and Buff, 2005; Foster 2008) with potential to sabotage the long term viability of local jaguar conservation efforts. Foster (2008) quantified the rate of lethal control on lands neighbouring the CBWS and modelled jaguar population viability at the national level under off take rates observed around CBWS. However, more detailed information is required about reproductive rates, longevity and dispersal characteristics of ja guars to develop these models further. Depletion of the wild prey
37 base due to unregulated hunting pressures is an important threat that must be quantified and mitigated. It is important that levels of off take be evaluated, monitored and regulated. To star t this process, in 2010, a national survey was conducted to asses levels of game meat consumption across Belize (Collins, Foster and Harmsen, unpublished data); and a national survey is currently underway to assess hunting rates and awareness of wildlife l aws (Urbina, Foster and Harmsen unpublished data). Foster et al. (2010b) quantified differences in diet between jaguars ranging in protected and unprotected lands in southern Belize, and demonstrated that depletion of wild prey can force jaguars to make di etary shifts. These shifts may entail livestock predation, which often results in retaliatory persecution of jaguars (Foster, 2010). Within the last decade, jaguar populations in the Maya Mountains have been indirectly but severely impacted by activities a ssociated with the illegal extraction of xate palms ( Chamaedorea spp.). Xateros (palm harvesters) cross the porous border from Guatemala into the Maya Mountains of Belize to illegally harvest the xate palms. They may spend several weeks at a time in the Be lizean forests subsisting on wildlife resources, especially the critical prey base which inevitably becomes the main subsistence source. Failure to curtail this pressure will have serious indirect ramifications on the jaguar population in this important co nservation area. Maintaining connectivity for jaguars is an important challenge that deserves immediate attention. Habitat loss and fragmentation caused by deforestation and land conversion into agricultural areas are primary threats to the survival of thi s large, wide ranging carnivore species. The remaining patches of natural habitat may become too small and isolated to support a viable jaguar population. Small population size and
38 severe isolation of entire populations may decrease the genetic variation w ithin the population resulting in lower fitness and the loss of adaptive potential (Frankham and Ralls, 1998). Over the last decades, Belize has been experiencing widespread land conversion and development, as well as privatization of public lands. On goin g research efforts have been investigating various areas across Belize with the goal to identify core areas for jaguars and characterize potential habitat linkages and other conservation actions needed to mitigate impacts of habitat loss and fragmentation. For example, connect the two largest forest blocks, which are separated by human developments and the Western Highway in central Belize (Figure 2 2). This site is cur rently the focus of intense ecological research aimed at delineating a functional corridor. The Coastal and Hummingbird Highways also present an increasing challenge in the connectivity of these two main conservation blocks (Figure 2 1). While jaguars, pum as, tapirs and other mammals are still crossing these highway barriers (Figueroa unpublished data) the wide ranging white lipped peccary ( Tayassu pecari ) has been extirpated from the southeastern portion of the central corridor and individuals may no longe r cross the Western Highway (Foster and Harmsen unpublished data). Within the last decade, the Hummingbird Highway, which also bisects the network of protected areas, has attracted several development projects including cattle farming, citrus farming, ecot ourism ventures and community developments. The on going research aimed at delineating the central jaguar corridor must be expanded to consider the Hummingbird Highway since this is fast becoming a barrier to the movement of jaguars and other large mammals
39 The Northern Highway also presents a barrier separating the core jaguar population from jaguars in northeastern Belize (Figure 2 1). Development projections for this region coupled with the lack of any meaningful protected lands east of the Northern Hi ghway suggest that while jaguars in northeastern Belize may persist into the short term (10 to 20 years) they are unlikely to do so in the long term (> 50 years). The unprotected forests east of this highway will face increasing pressures and the movement barriers posed by the Northern Highway and the widespread development of intensive agricultural communities will become increasingly significant. In the extreme south, the impending development of the Southern Highway coupled with high levels of habitat de gradation and habitat loss due to agricultural practices will inevitably create movement barriers and will impact connectivity between the Maya Mountains and the forest habitats in the extreme south, including the Sarstoon Temash National Park. Within the last decade The Belize Zoo (TBZ) and its Tropical Education Centre has emerged as an important player in national efforts to advance jaguar conservation. Aggressive education campaigns implemented by TBZ, both onsite and outreach, has raised awareness and public recognition about the many challenges facing jaguar conservation. The multifaceted approach of TBZ involves interpretive displays at the zoo facility where visitors are educated about the diverse factors affecting jaguar conservation. With over 17,0 00 school children visiting annually, this important education opportunity has potential to impact the future of jaguar conservation. The proactive outreach program takes information about the central jaguar corridor to schools and villages located within the corridor area. TBZ also works closely with the Forest Department and Panthera in responding to cases of human jaguar conflict.
40 Discussion The multifaceted array of research projects that have taken place in Belize over the last few decades have provi ded the conservation community and policy makers with critical information necessary to help guide the management and conservation of jaguars. The various on going studies, also at the forefront of jaguar research, will provide much needed ecological infor mation necessary to secure the long term survival of this species. Given the alarming rate of land use and land cover changes sweeping the country, the future of jaguar conservation will partly depend on the timely availability of research based conservati on and management recommendations. The establishment of both the CBWS and the LCJCWS demonstrates that both the Government and non governmental organizations continue to work together for the advancement of jaguar conservation. However, the effort to maint ain and advance jaguar conservation is perpetual and the conservation community must be cognizant that any success or progress in this effort could easily be derailed given the multifaceted and increasing threats large carnivores such as jaguars are facing The population estimate of 446 to 754 jaguars for the entire country should be interpreted with caution. The camera trap based methodology used to derive the estimates upon which this is based is rapidly evolving (Foster and Harmsen, In press). In fact, several recent studies have shown that spatially explicit capture recapture (SECR) models are more realistic and that they result in lower density estimates using the same input data, than traditional methods based on the mean maximum distance moved (MMDM ) (Gerber et al., 2011). Therefore previous estimates from camera trapping may be biased towards overestimation. Given the extensive pressures facing jaguar conservation in Belize it is unlikely that the true estimate of jaguar population size
41 is toward th e high end of 754 individuals. The actual population estimate of jaguars in Belize is probably closer to the lower end of our projected estimate, that is, about 450 individuals. Population size is an important conservation parameter because it has a major impact on the dynamics of a population. Franklin (1980) first proposed the idea that 500 individuals was an effective population size necessary for long term survival. The concept of a minimum viable population inherent in the idea of 500 individuals has b een the subject of some critique (Caughley, 1994; Henriksen, 1997) but is still widely accepted and has potential to serve as an important management tool. Applying this concept to the national jaguar population in Belize underscores the collective import ance of ensuring connectivity between the Maya Mountain Massif and the forests in northwestern Belize. Maintaining adaptive genetic variation within the local jaguar population can only be accomplished by securing the ecological integrity of the two main protected blocks, and in securing their functional connectivity. In this regard, perhaps the single most pressing challenge threatening the long term viability of jaguars in Belize is that of securing and maintaining the ecological function of the central jaguar corridor. While the Maya Mountain Massif, together with the contiguous protected lands, may have enough territory to sustain a viable population in the long term, wild prey depletion associated with the illegal harvesting of xate and other natural resources must be eradicated or mitigated. Government and non governmental organizations now have a consistent presence in this area, however, stronger integration, enforcement, monitoring and evaluation is urgently needed. In 2010 levels of illegal xate a ctivity probably decreased
42 in the Maya Mountain Massif (G. Manzanero pers. comm.), but some reports suggest that the illegal operations are now becoming more sophisticated and have diversified into the exploitation of logwood, gold and other natural resour ces. These illegal patterns of exploitation are not sustainable and will have a detrimental impact on jaguars and other wildlife in the general area. Camera trap data has shown a decline in jaguar density in this area over time, but data should be re analy zed with spatially explicit models to determine if this is a true trend rather than an artifact of survey design. The Maya Mountain Massif is of regional importance as it represents a potentially viable and self contained population of jaguars that can ser ve as a source population for the wider region, namely the Maya Mountains of Guatemala, and the Selva Maya of Mexico and Guatemala. Maintaining healthy levels of genetic variation and connectivity for jaguars can avoid a decrease in fitness, and loss of ad aptive potential, which is particularly important in a changing environment. This can be accomplished by securing ecological integrity of protected areas, and functional connectivity between them. Prime habitat for jaguars is dwindling and rates of habitat fragmentation are rapidly increasing. Quick and efficient conservation and management actions are needed to enhance connectivity term survival in a landscape modified by humans. Extensive multi disciplinary research a ctivities are currently underway to identify population(s) (fragments). Applying the best available knowledge will aid in characterizing dispersal corridors between habitat patches existing in Belize and by doing so, successfully enhance levels of connecti vity. The creation of dispersal corridor(s) by establishing a new protected area (e.g. central jaguar corridor) or by
43 expanding currently existing protected areas is one of the most popular approaches in conservation to increase and/or maintain connectivit y. Increasing habitat quality and species management, creating sustainable use areas, decreasing human disturbance, or other threats in and around existing jaguar core areas also have potential to increase levels of connectivity. Within the unprotected lan dscape human jaguar conflicts and associated jaguar persecution levels are highest, perhaps a direct consequence of the patterns of land cover changes that disrupt the integrity of jaguar home ranges. Consequently, competition and confrontation with other jaguars and with humans will increase. The importance of the unprotected landscape must be characterized with urgency so that critical links and key conservation zones can be identified. General Recommendations and Proposed Actions Belize currently appears to have a healthy and viable population of jaguars but if the current threats continue unabated this will change. The network of protected areas potentially has enough resources to sustain a viable jaguar population, but further research and continued mon itoring efforts are needed. The window of opportunity to ensure long term survival is still open. The national jaguar population is most likely still contiguous (the current genetic study will confirm), but this may change unless steps are taken to curb ha bitat loss and fragmentation. An increase in habitat loss and fragmentation, jaguar persecution, over harvest of wild prey, and legislative and policy needs, underscore the urgency for continued actions. Despite the relatively small size of the Belizean te rritory, and given the relatively large tracks of intact forests coupled with a strong national conservation ethic, Belize is an important regional player in jaguar conservation. However, in order for Belize to
44 safeguard this critical role, it must first s ecure the long term conservation status of the means to enhance and secure a sustaina ble jaguar population in Belize: 1 The a ssess ment of connectivity using camera traps, telemetry and genetic analysis, and the identif ication of feasible conservation and management actions a. The c haracterization of a central jaguar corridor with enabling legislation. This is currently underway. b. The i dentif ication of core areas under threat and the appl ication of appropriate conservation actions (e.g. education programs, increase habitat quality and species management, creation of sustainable use areas). 2. Conduct a national systematic survey to quantify livestock p redation and jaguar persecution. Plans are being developed to run a national survey. This information would help government and non governmental agencies develop national strategies to curtail this serious threat. 3. The e ffective monitoring and prevention of illegal hunting and harvesting activities; and regulation of legal hunting activities. Species specific hunting regulations must be developed, implemented and enforced to avoid depleting wild populations beyond levels that compromise population viabili ty. These hunting regulations must be based on sound ecological data (density and demographic rates) to prevent sustained declines in breeding numbers. Regulations, such as closed seasons and hunting quotas, must be introduced and enforced over the entire country, inside and outside protected areas. 4. the i mplementation of a long term research and monitoring program within the Maya Mountain Massif, the central jaguar corridor and the Selva Maya block in northern Belize. These research and monitoring progra ms should be conducted within a national framework that allows for adaptive management of jaguars. 5. The c haracteriz ation of habitat use, movement behavior, demographics and genetic structure within the fragmented landscape so that Belize can efficiently conserve or restore functional connectivity. 6. The e stablishment of a national working group, led by the Belize Forest Department, to integrate and coordinate jaguar conservation efforts. This group must work synergistically and avoid duplication of efforts. The concerted efforts of government and non governmental organizations, researchers and communities is necessary to ensure that the country continues to have the capacity to conduct long term monitoring across multiple sites and adapt to emerging challenges in jaguar conservation.
45 7. The continued synergistic efforts of organizations such as the Belize Zoo and Tropical Education Centre, the University of Belize, Panthera and the Forest Department are crucial to the success of efforts aimed at mit igating human wildlife conflicts and ensuring longevity of jaguar populations. 8. Efforts must be made to activate the now defunct Jaguars Without Borders conservation initiative. The coordinated efforts of Belize, Guatemala and Mexico to advance conserv ation of the shared Selva Maya forests is absolutely critical to any local, regional or range wide jaguar conservation plan. Government and non government organizations, including academia must ensure the continued success of this tri national initiative.
46 Table 2 1. Protected areas that have a role to play in protecting jaguar populations in Belize. Belize Category IUCN Category Number of protected areas Total area (km 2 ) Forest Reserve Vl 15 3,748 National Park ll 15 1,628 Natural Monument la and lll 3 27 Nature Reserve la 3 437 Wildlife Sanctuary lV 6 763 Archaeological Reserve V 6 115 Private Protected Area lV 7 1,309 Total 55 8,027 Table 2 2. Jaguar density estimates for different sites in Belize. Study Site Reference Density (100 km 2 + SE) Habitat Chiquibul Forest Reserve Kelly, 2003 7.48 + 2.34 Tropical Moist Lowland Forest Cockscomb Basin Wildlife Sanctuary Silver et al., 2004 3.5 11 Tropical Moist Lowland Forest Gallon Jug Estate Miller, 2005 11.28 + 2.66 Tropical Moist Lowland Forest Gallon Jug Estate Miller, 2005 8.82 + 2.27 Tropical Moist Lowland Forest Fireburn Reserve (Northeastern Belize) Miller, 2006 5.3 + 1.76 Secondary Broadleaf Forest Cockscomb Harmsen, 2006 10.84 + 3.85 Tropical Moist Lowland Forest Chiquibul Forest Reserve M. Kelly unpublished data Mean 5.2 Tropical Moist Lowland Forest Northeastern Belize Bunyan and Kilshaw, 2008 4.42 6.18 Mosaic Mountain Pine Ridge M. Kelly unpubl. data in Maffei et al., 2011 2.32 5.35 Tropical Pine Forests RBCMA La Milpa Waight et al., 2010 5.70 + 2.10 Tropical Moist Lowland Forest RBCMA Hillbank Waight et al., 2010 6.01 + 2.10 Tropical Moist Lowland Forest Fireburn Reserve (Northeastern Belize) Everatt et al., 2010 2.36 Mosaic
47 Table 2 3. Estimate for the number of jaguars in the two main conservation blocks in Belize. Site Area ( km 2 ) Min and Max Density Estimates (# individuals per 100 km 2 ) Reference Population Estimate (# of I ndividuals) Southern Conservation Block CBWS plus adjoining PAs 1,111 4.82 and 11.45 Harmsen, 2006; Harmsen et al. 2010a 54 to 127 CFR plus adjoining PAs 3,780 3.53 and 7.48 Kelly 2003; M. Kelly unpubl. data 133 to 283 MPR 432 2.32 and 5.35 M. Kelly unpubl. data 10 to 23 MFR and adjoining PAs 396 3.53 and 7.48 Based on CFR estimates 14 to 29 Northern Conservation Block RBCMA plus adjoining PAs 1,145 5.7 and 6.01 Waight et al. 2010 65 to 69 Sub Total 276 to 531 (Protected Areas (PAs), Manatee Forest Reserve (MFR)).
48 Figure 2 1. Network of protected areas and the central jaguar corridor in relation to land cover in Belize. Land cover information based on Cherrington (2010) and obtained from the Land Information Center, Belmopan, Belize. Majo r highways are also shown.
49 Figure 2 main protected blocks.
50 CHAPTER 3 STUDY DESIGN AND METHODOLOGY Study Area The study area is located in central Belize and comprises an area of approximately 1,400 km 2 (Figure 3 central biological corridor. This corridor represents the only remaining source of biological connectivity between the Selva Maya forests in northern Belize and the protected areas of the Maya mountain massif in southern Belize. Anthropogenic forces are widespread and include established and growing communities, an expanding agriculture frontier (citrus, corn, sorghum, cashew, livestock), logging and increasing to urism activities. The Western highway bisects the study area and threatens to become an increasingly important barrier hindering biological connectivity within the corridor. A second major highway, the Coastal road, runs perpendicular to the Western highway. Two major river s run parallel to the Western highway, the Belize river on the north side and the Sibun river on the south side. Thirteen communities are located within the study area but localized along certain key landscape features: six along the Western highway, four along the Belize river valley, two along the Sibun river and one along the banks of Spanish creek. Terrain elevation ranges from sea level along the coastal plains to just over 200 meters asl. The land cover is a mosaic of different habitat types and land uses. Forest cover is dominated by Lowland Broadleaf Moist forests with smaller, interspersed amounts of Swamp forests and Shrublands (Iremonger and Brokaw 1995). The non forested landscape is dominated by Lowland savanna and to a lesser extent by Short Gr ass savannas, wetlands, livestock farming and citrus. The terrain ranges from low
51 and flat along the coastal areas to steep and rugged within the karst hills on the south side of the Western highway. The weather is characterized by two distinct seasons: a rain and a dry season. In 2010 and 2011, the annual rainfall in the central biological corridor averaged 1,840.5mm with 60% to 88% falling during the rain season (June to November) (Belize MET unpublished data). The temperature ranged from 12.9 C to 33.9 C with an annual mean of 30.03 C. The cooler months were November to January and the warmer, April to June. The relative temperature stability is due in part to the coastal proximity and the associated sea breeze. Diet Scat Collection Scats were analyzed t o determine the proportion of different prey species comprising the diet of jaguars and pumas. Scats were collected opportunistically between January 2008 and May 2011 throughout the entire study area by the primary author and three trained field assistant s. Additionally, researchers from two ongoing research projects in the study area collected and provided scats. The location coordinates (UTM) of all scats were recorded in the field using a Global Positioning System (GPS) unit. All scats were initially st ored in Ziploc bags for transportation from field to a centralized research base. Scats were later transferred to individually labeled paper bags and allowed to air dry in the shade. Dried scats were then stored in a water proof container until ready for g enetic and diet analysis. A portion of each scat containing bile was separated and placed in individually labeled paper bags for DNA analysis. Two table spoons of silica granules were added to each paper bag to ensure that all scat samples remained dry. Al l DNA analysis was conducted at the Global Felid Genetics Program, Sackler Institute of Comparative Genomics, American Museum of
52 Natural History, New York through a collaborative effort with this institute. Genomic DNA extraction of scat samples was conduc ted using the QIAamp DNA Stool Kit (Qiagen 2007), with modified protocols to improve amplification. To identify species, at least two different PCR (Polymerase Chain Reaction) were performed for each sample, using PuReTaq Ready To Go PCR beads in 25L rea ction volume (GE Health Care, Amersham Biosciences, 2004). Species were identified based on neighbor joining method for tree construction (1000 bootstraps) using the program Geneious Pro 5.4 (Drummond et al. 2011). To determine sex of each individual, the PCR product was done on a 2.5% agarose gel with SYBR Safe DNA gel stain in TBE buffer 1X (Wei et al. 2008). Sex determination was only conducted on jaguar scats. Following the removal of the portion for genetic analysis, the remainder was broken down, thor oughly washed and sifted to isolate undigested prey remains. Consumed prey was identified by separating and analyzing undigested prey remains which include bones, hooves, hairs and/or teeth. When necessary, a microscope was used to identify the prey remain s. Prey remains were compared with those from a reference collection compiled by Harmsen (2006) and made available for this study. A few additions were also made to this collection to represent species that were previously not included in the Harmsen (2006 ) collection. Contents of scats were calculated as percent occurrence (number of times a specific item was found as percentage of all items found) (Ackerman et al., 1984). The relative biomass of prey species consumed per scat was calculated using the cor rection factor (y = 1.98 + 0.035x) derived by Ackerman et al. (1984) for puma, where y is the weight of food consumed per scat and x is the mean prey weight. Mean prey weights
53 were adapted from Reid (1997), Novack et al. (2005) and local knowledge (O. Figu eroa unpublished data). Food niche breadth, food niche overlap and mean weight of vertebrate prey were estimated. Food Niche Breadth Food niche breadth (diet diversity) was computed for jaguars, pumas, male jaguars and female jaguars following the method o f Levins (1968) and later described by Jacksic and Braker (1983): where Pi is the relative occurrence of prey taxon i in the given species diet. This value ranges from one (minimum niche breadth, highest level of specialization) to n, where n is the total number of prey taxa and considers prey taken in relation to the available species. Colwell and Futuyma (1971) propose d a standardized version of food niche breadth to allow for comparisons among assemblages: w here B obs is the observed niche breadth, B min is the minimum niche breadth possible (=1), and B max is the maximum niche breadth pos sible (= n). Therefore, B sta will range between 0 and 1. Values approaching 1 indicate that all prey items were taken in equal proportion while values approaching 0 indicate that very few prey items were taken at a very high frequency and many prey items w ere taken at a very low frequency. The standardized value of the food niche breadth (B sta ) was calculated for jaguars, pumas, male jaguars and female jaguars.
54 Food Niche Overlap Food niche overlap (diet similarity) was calculated between jaguars and pumas, male jaguars and pumas, male jaguars and female jaguars, and female jaguars and pumas using the method described by Pianka (1973) and outlined by Jacksic and Braker (1983): where pi is the relative occurrence of taxon i in the die t of one species, and qi is the relative occurrence of the same taxon i in the diet of a second species. This index ranges between 0 and 1, where 0 signifies no overlap (completely different) and 1 signifies complete overlap (diets are completely similar). Mean Weight of Vertebrate Prey The mean weight of vertebrate prey was calculated for jaguars, pumas, male jaguars and female jaguars following the method described by Jacksic and Braker (1983). This method calculates the geometric mean by first summing th e products of individual prey times their log e transformed weight and then dividing this value by the total number of prey. Capture and Immobilization Jaguars and pumas were captured with foot snares. Snares were placed at strategic locations based on r eliable signs of jaguar activity (tracks, scats) or along known travel routes. Site placement depended foremost on safety requirements for a snared individual, that is, away from trees, snags, fences and/or large rocks. All vegetation within a 2.5m radius of a snare site was cleared and the snare assemblage was anchored to the ground. Constant and remote monitoring of snares was
55 accomplished by fitting a Telonics TBT 500 trapsite transmitter (Telonics Inc., Mesa, USA) to each snare to signal the triggering of a trap. These trap radios run continuously (confirming it is functioning) and change pulse rate when the trap is triggered. Jaguars and pumas were chemically immobilized using a remote drug delivery system (Pneu Dart, Inc., Model 179B Pistol; Williamsp ort, PA, USA). A telazol (tiletamine/zolazaline hydrochloride, 4.84 mg/kg) and xylazine hydrochloride (.99 mg/kg) mixture was used to anesthetize the captured animals (Kreeger, 1996). Immobilized animals were examined for general body condition. We monito red temperature, pulse rate and respiration rate of immobilized animals at 5 10 minute intervals, determined sex, fitted them with a radiocollar and conducted a series of morphometric measurements (body length, tail length, chest girth, neck girth, canin e length). Captured animals were aged using tooth eruption, tooth wear and pelage characteristics (Shaw 1987, Anderson and Lindzey, 2000). Two drops of blood were smeared on an identification card and saved for future genetic analysis. Immobilized cats wer e sometimes allowed to recover inside cages and released after recovery. Yohimbine (0.125mg/kg) was used as the reversal for xylazine (Kreeger, 1996). Capture and handling procedures adhered to guidelines established by the American Society of Mammalogists (Gannon et al. 2007) and the Belize Forest Department, and were approved by IFAS ARC at the University of Florida (Approval Number 014 09WEC). A total of 9 GPS satellite collars were deployed. We placed TGW 4580 (H) collars (Telonics Inc., Mesa, Arizona) on 6 jaguars and 1 puma, and North Star Globalstar collars (North Star Science and Technology, LLC. King George, Virginia) on1 jaguar and 1 puma. The TGW 4580 (H) units were programmed to colle ct and store a
56 location fix at seven hour intervals starting from the time of deployment. The units also collected and stored date and time of data acquisition, altitude, speed, course data, the number of satellites used for location estimates, search time and the horizontal dilution of precision. Every three days the units broadcasted the location data to orbiting satellites from which the information was retrieved by Service Argos Inc., which processed and made available the results. These collars were eq uipped with a programmable drop off mechanism that allowed the collars to drop off at a predetermined date and time. The Globalstar collars provided real time access to the acquired location fixes via a secure website. All processed files were reviewed to remove duplicate or erroneous readings. The collars were also equipped with VHF capabilities (148 152 MHZ) and both collar models were equipped with a store on board mechanism that allowed all acquired data to be archived on the unit for retrieval upon rec overy. Home Range and Movement Patterns The Central American Ecosystems Mapping Project (CAEMP) characterized 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). Updated v ersions (2011) of these habitat files were obtained from the Land Information Center (LIC), Ministry of Natural Resources, Belmopan City, Belize. This update included supervised classification of satellite imagery to account for recent developments and lan d conversion in central Belize. Updated files (2011) on roads and existing communities were also obtained from the LIC. The spatial reference for all digital layers were defined and if necessary, projected to the North American Datum 1927 (NAD 27), Univers al Transverse Mercator Zone 16 (UTM Zone 16). All
57 Geographic Information System (GIS) and spatial analyses were conducted in this projection. Outlier locations that represented two percent of the total for each of the nine cats were removed following the H armonic mean method which removes points by their largest harmonic mean value (White and Garrot 1990, Dixon and Chapman 1980). Values were recalculated after each removal (White and Garrot, 1990) leaving 98% of each data set for subsequent analyses. In thi s method points with the largest harmonic mean are removed and range boundaries recalculated after each removal. This analysis was conducted using the animal movement extension (Hooge and Eichenlaub, 1997) in ArcView 3.3. We defined home range areas as the area encompassed by 100% Minimum Convex Polygons (MCP) and 95% fixed kernel estimates. We calculated home range sizes separately for the rain season, the dry season and annual periods (combined seasons). Using the Geospatial Modeling Environment (Version 0.6.0.0) (Beyer, 2012) we calculated 50%, 75% and 95% Fixed Kernel (KP) probability polygons (Worton 1989, Kernohan et al. 2001) with least squares cross validation used to obtain the smoothing parameter (Worton 1995, Seaman et al. 1999) and Minimum Convex Polygons (MCP) (Kernohan et al. 2001). Fixed kernel performs better as a home range estimator and has a lower bias than adaptive kernel (Seaman and Powell 1996). Although the MCP has little ecological meaning and offers no information on the intensity of use, this home range estimator has been used in previous studies and thus allows for comparisons. These polygon contours (KP probability polygons and MCPs) defined the home range boundaries for subsequent analyses. We consider the autocorrelation issue irr elevant in this study because the
58 seven hour spacing of GPS locations provides independence of observations and the KP method is also robust to autocorrelation (De Solla et al. 1999, Kernohan et al. 2001). We dealt with the issue of sample size (number of locations needed to calculate a home range) for all KP estimates by conducting visual inspection of the area observation curves for evidence of an asymptote (Pyke 2006). This analys i s was conducted in GME (Beyer 2012) and ArcGIS 10.0. To assess the effect of sample size on the MCP estimates, we used the Animal Movement Extension (AME) in ArcView 3.3 (Hooge et al. 1999) to generate an area observation curve for each individual data set. AME uses a bootstrap function to produce these curves which then illustr ate the necessary sample size that produces a stable MCP (Gese et al. 1988). The MCP estimate was computed using 100 replicates of 10 randomly selected data points for each individual. We compared home range size and used paired t tests to test for seasona l differences. We used the Animal Movement Extension (Hooge and Eichenlaub, 1997) for ArcView 3.3 and the spatial analyst extension in ArcGIS 10.0, to explore site fidelity for each collared cat. A home range cannot exist if an animal does not exhibit site fidelity (Spencer 1990, Hooge 1995). The first point was used as the starting point for the simulation. This method creates random angles and uses distances between sequential points to determine random walks. One thousand random walk simulations were cre ated 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. This analysis tests the null h ypothesis of complete spatial randomness. The null is rejected when there is statistically significant clustering or dispersion of the data. The ratio of observed mean
59 distance and expected mean distance (OMD/EMD) is the index used to test the above hypoth esis. The EMD is the average distance between data points in a hypothetical random distribution. When this ratio is < 1 then the data has a tendency to cluster. When this ratio is > 1 then the data has a tendency toward dispersion. This test outputs a p value, a z value and a nearest neighbor ratio (OMD/EMD). We partitioned the data into discrete time intervals and calculated the distance moved (step lengths) during each interval. This estimate was used to determine the movement rate per step length per time period. We fitted models with a smooth term for each cats diurnal cycle, a smooth term for movement rate by Julian date, a term for the average movement rate of each cat, a term for the ecosystem the movement started in, and a term for whether or not the cat crossed a road in the step. We then used these models to evaluate how movement rate (log deviation from average movement) varies with time of day and with time of year (by Julian date). We evaluated the effect of roads on the movement rate by ca lculating the average movement rate for step lengths that crossed a road and comparing these with the average movement rate for step lengths that did not cross a road. We evaluated the effect of ecosystem on movement rate by calculating the average movemen t rate for step lengths when the movement started in a given ecosystem type and comparing these with the average movement rate for all step lengths in all ecosystem types. We generated models to explore how the average net displacement varied with time for each collared cat. Finally, we developed histogram models of the observed movement rates to evaluate the variability of movement speeds. These analyses were conducted in R version 2.14.2 (R Development Core Team 2012),
60 an open source statistical package and programming language (www.r project.org). Significance determined at the 0.05 level. Confidence intervals are 95%. Use of Habitat and Landscape Features Several analytical procedures were utilized to understand how jaguars and pumas use the availab le habitats and landscape features. After home range estimation and the delineation of the MCP, 50% KP, 75% KP and 95% KP contours, both second and third order habitat selection (Johnson 1980) were tested (Neu et al. 1974). This method uses a G test (Zar 1 984) and a Chi square goodness of fit test to test for differences in use and availability. If these tests indicated that differences occurred in the use versus available tests then Bonferroni simultaneous confidence intervals were calculated to determine which habitat types were selected or avoided. The Neu et al. (1974) method is robust in that it has acceptable Type 1 error rates (Bingham and Brennan 2004). This analysis was conducted using Resource Selection for Windows (Leban 1999). Habitat proportions were summed and sorted within the home range polygons. The 50% isopleths probabilities were used to delineate activity centers (areas of core use). The MCP polygons of all cats were merged and buffered by a distance of 11 km. This distance was chosen as it represents the average of the mean maximum distances moved within a 24 hour period. The merged and buffered area was used to represent the overall area that was available to all cats. Proportions of habitat within core areas (50% isopleths) and within home range polygons (95% isopleths) were compared to the proportions of habitat within the overall study area (merged and buffered MCP), with the core areas and the home range polygons representing habitat used and the overall study area representing the h abitats that were available. Additionally, use of habitat was
61 assessed at a finer scale by comparing the proportions of GPS location fixes within different habitat types (habitat used) with the proportions of habitat within individual MCP polygons (availab le habitat). Proportions of GPS locations were also compared with that available in the entire study area (merged and buffered MCP). To compliment these use versus available tests, we used the one sample t test procedure to further compare the proportions of habitat that were used (core areas, proportions of GPS locations and individual MCP polygons) to the specified constant within the total study area (comparing sample means to hypothesized value: Sokal and Rohlf: 169 175).This procedure outputs the avera ge difference between each data value and the hypothesized test value, a t test comparing this difference to zero, a confidence interval and the level of significance (2 tailed). In an effort to better understand how or if the use of habitat changes across the rain and dry seasons, these use versus available tests were duplicated across these two seasons. We also used Compositional Analysis (Aebischer et al. 1993) for contrasting habitats used versus those that were available. This method tests the Null hyp othesis that habitats are used in proportion of availability (no preference) and uses the individual animal as the experimental unit. Compositional Analysis is robust in that it accounts for the problem of apparent avoidance of certain habitat types which invariably occurs when a highly used habitat creates the perception of avoiding other, less utilized habitat types (Aebischer et al. 1993). This is achieved by using the differences in log ratios between use and availability for all habitat types. Composit ional Analysis provides ranks of selection (same as the Johnson Ranking Method, Johnson 1980) and the Null can be tested using a chi square test with an associated P value. This analysis was
62 conducted using Resource Selection for Windows (Leban 1999). P va lues less than 0.05 were considered significant. In order to better understand the underlying habitat components and landscape features that best explain presence (selection for specific habitat and landscape features) of jaguars and pumas we constructed c lassification trees using the Cartware/rpart software package in R version 2.14.2 (R Development Core Team 2012). The classification trees are constructed by means of a recursive partitioning of the data matrix such that the resulting branches in the dendr ogram (terminal groups) within each partition become increasingly homogenous. The resulting tree presents revealing information regarding the underlying structure and predictive capability of the data. The output structure of this data analysis is easily u nderstood and interpreted. This method is especially rigorous since, in addition to providing a classification structure (the dendrogram), it also provides an estimate of the misclassification probability for each group along each terminal branch. However since a fully grown tree has limited predictive capability (it is an exact representation of the data), the resulting trees were simplified by a pruning (removal of selected terminal branches) procedure which recombined subgroups without significantly in creasing the classification errors. The decision to prune a tree at any given size was based on the V fold cross validation method, which provides error estimates for trees of any given size. In this cross validation method, the data set is subdivided int o ten equal parts, the model derived with nine parts (learning set), and tested with one part (validation set). The cross validation is repeated ten times and the results are combined to develop the predictive accuracy and error rates for the dendrogram. I n this case, we retained the smallest tree
63 for which the estimated error was within one standard error of the minimum. Eight variables were used to construct the CART model: broad scale ecosystem classes, fine scale habitat categories, distance to nearest community, distance to the Western Highway, distance to nearest road, distance to nearest river (or permanent stream), terrain ruggedness and forest/non forest edge. For each location fix, ecosystem and habitat data were obtained by intersecting the locati on data with the relevant shapefiles representing the ecosystem and habitat data. All distance measurements (distance to community, to nearest road and to nearest river) were obtained by joining the two layers within ArcMAP (ArcInfo 10.0). For example to o btain the distance between the GPS location data and 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). Th e output file generated a unique table that specified the shortest distance between all nest locations and the nearest river. This method is consistent, efficient and minimizes errors such as those associated with manual estimation using the ruler function Terrain ruggedness was calculated as a statistical summary (standard deviation) of elevation values within a 250 meter radius of all location fixes. The length of edge in a 250 meter radius around each location fix was quantified as forest/non forest edg e. These analyses were conducted using ArcGIS 10.0, the Geospatial Modeling Environment (Beyer 2012) and R (R Development Core team 2012). Monte Carlo re sampling was used to obtain a P value for each dendrogram. In this method, one thousand trees were gen erated by random permutations of the data. The correct classification rate (CCR) for the original model was then compared with the
64 distribution of CCRs for the one thousand randomly generated trees. This analysis tests whether the model could in fact be o btained by random chance and evaluates the variables that best predict the presence of jaguars and pumas based on the detailed location data. A confusion matrix is generated for each model, where rows represent observed values and columns represent predict ed. The diagonals of the confusion matrix represent the correct classification while the off diagonals represent misclassification. A kappa value indicates the proportion improvement over chance.
65 Figure 3 1. Map of Central Belize showing the general s tudy area with major roads, communities, land cover and protected areas.
66 CHAPTER 4 STUDY OUTCOMES AND DATA SUMMARIES Diet of Jaguars and Pumas A total of 136 felid scats were collected and a cumulative total of 161 prey items were identified in these scats (Table 4 1). The spatial distribution of all scats encompassed an area (minimum convex polygon) of approximately 366 km 2 (Figure 4 1). DNA analysis failed to identify the origin of 21 scats and these were removed from further anal ysis. Thirty six scats containing 42 prey items were from pumas while 79 scats containing 102 prey items were from jaguars. Jaguar scats contained an average of 1.29 identifiable prey per scat while puma scats contained 1.17 identifiable prey per s cat. Of the 79 jaguar scats, 65 were from males an d 10 from females, containing 82 and 15 prey items respectively. Scats belonging to male jaguars contained 1.26 identifiable prey per scat while scats belonging to female jaguars contained 1.5 identifiable prey pe r scat. Jaguar diet included at least 14 prey species: 11 wild mammals, 2 reptiles and at least one bird (Table 4 occurrence). The armadillo was the prey most often consumed by jaguars, comprising 35. 29% of the prey items and 34.38% of the total biomass consumed. Seven species (armadillos, collared peccaries, turtles, birds, pacas, northern raccoons and coatis) comprised 77.66% of the total biomass consumed by jaguars. Puma diet included at least 11 species: 9 wild mammals, 1 domestic mammal and at least one bird (Table 4 paca was the prey most often consumed by pumas, comprising 42.86% of the prey items and 40.15% of the total biomass consumed. Fiv e species (pacas, collared
67 peccaries, kinkajous, armadillos and tayras) comprised 84.21% of the total biomass consumed by pumas. The MWVP taken by jaguars was 4.62 kg (n=102 prey items) and for pumas was 7.62 kg (n=42) (Table 4 4). The MWVP taken by pumas was 65% heavier than the MWVP taken by jaguars. The jaguar diet had a higher species richness than that of the puma (14 species for jaguars versus 11species for pumas) and also showed a higher diet diversity (jaguar B sta = 0.37; puma B sta = 0.28) (Table 4 4). Diet overlap between jaguars and pumas was 0.40 (Pianka index) (Table 4 5). The diet of male jaguars was comprised of at least 13 species: 10 wild mammals, 2 reptiles and at least one bird (Table 4 5%. Th e armadillo was the prey most often consumed by male jaguars, comprising 37.80% of the prey items and 36.62% of the total biomass consumed. Seven species (armadillos, collared peccaries, turtles, northern raccoons, pacas, coatis and birds) comprised 85.28% of the total biomass consumed by male jaguars. The MWVP taken by male jaguars was 4.90 kg and the diet diversity (Bsta) was 0.35. The diet of female jaguars was comprised of at least 8 species: 6 wild mammals, one reptile and at least one bird (Table 4 7) All eight species had a relative occurrence comprising 33.33% of the prey items and 32.83% of the total biomass consumed. Four species (armadillos, collared peccaries, pacas and b irds) comprised 74.4% of the total biomass consumed by female jaguars. Four species (tamanduas, coatis, grisons and turtles) each represented 6.66% of the prey items and comprised 6.72%, 6.38%, 6.29% and 6.23% of the relative biomass consumed. The MWVP for female jaguars was 4.17
68 kg and the diet diversity (B sta ) was 0.64. Diet overlap (Pianka index) between female jaguars and male jaguars was 0.92 and between female jaguars and pumas was 0.51 (Table 4 5). With a Pianka index of 0.37, male jaguars and pumas showed the lowest overlap between any of the groups analyzed. Capture and Telemetry We captured sixteen jaguars (3 females and 13 males) (Table 4 8) and eight pumas (3 females and 5 males) (Table 4 9) during a total trapping effort of 2,919 trap nights (Ta ble 4 10). For ease of reference, the identification code of the nine cats that provided data for this study (Tables 4 8 and 4 9) were changed to Cat 01, Cat 02, Cat 03, Cat 04, Cat 05, Cat 06, Cat 07, Cat 08 and Cat 09 where cats 01 through 07 were jaguar s and cats 08 and 09 were pumas. We experienced almost complete failure with the initial deployment of seven GPS collars (Telemetry Solutions: model Quantum 5000) and were thus unable to obtain any meaningful data from these collars. Two jaguars and one puma were recaptured and fitted with functional collars. Six additional collars were deployed on five jaguars and one puma. Seven TGW 4580 (H) GPS collars (Telonics Inc., Mesa, Arizona) and two North Star Globalstar collars (North Star Science and Technolo gy, LLC, King George, Virginia) provided the data for this study. Overall, 45.6% of 8,456 location attempts by the Telonics TGW collars were successful. Of the successful attempts, 3.4% were 2 dimensional (3 satellites used) and 96.6% were 3 dimensionsal (at least 4 satellites used). Three dimensional fixes are more accurate than 2 dimensionsal fixes (Lewis et al. 2007, Moen et al. 1996). The average horizontal dilution of precision (HDOP) (SD) from the Telonics TGW collars was 5.94 .69 for the 3 dimen sional (3D) fixes and 4.4 .68 for the 2 dimensional (2D) fixes. DOP refers to the geometry of the satellite distribution in space and is known to
69 influence accuracy (Langley 1999, MELP 2001). Usually, a low DOP value indicates a higher GPS accuracy becau se satellites are more widely arranged in space (Langley 1999, MELP 2001). The average search time per acquired location fix (per collar) ranged from 87 40 seconds to 102 41 seconds. The North Star collars attempted a fix 912 times and had a success ra te of 37.83%. Of the successful fix attempts, 8.7% were 2D with HDOP ( SD) values of 7.8 .97 and 91.3% were 3D with HDOP ( SD) values of 9.3 .87. The average search time per acquired location fix (per collar) was 90.4 38 and 90.8 41 for the two N orth Star collars. None of the collars had fix attempts that were considered improbable (far from the general study area) and thus had to be removed. Home Range Sizes A total of 4,147 GPS location fixes (3,062 for male jaguars, 184 for a female jaguar and 910 for two male pumas) were used to estimate home range size for 9 collared cats (Table 4 11). Jaguars and pumas did not have exclusive home ranges (Figure 4 2). After partitioning of the data set to account for seasons, a total of 1,282 GPS location fixe s were used to estimate the size of dry season home ranges (Table 4 12) and 1,806 GPS location fixes were used to estimate the size of rain season home ranges (Table 4 13). Jaguars and pumas did not have exclusive home ranges during either the dry season o r the rain season (Figures 4 3 and 4 4). The log transformed home range size of jaguars differed across seasons. The annual home range size was significantly larger than the dry season home range size ( t = 4.01, d.f = 6, sig. (2 tailed) = .007 for MCP and t = 3.07, d.f = 6, sig. (2 tailed) = .03 for fixed kernel) However, the annual home range size was not significantly larger than the rain season home range size ( t = .980, d.f. = 6, sig. (2 tailed) = .37 for fixed kernel).
70 When home range was compared across seasons, the rain season home range size was significantly larger than the dry season home range size ( t = 4.838, d.f = 5, sig. (2 tailed) = .005 for MCP and t = 3.090, d.f = 5, sig. (2 tailed) = .03 for fixed kernel). Total home range size was si gnificantly larger than core areas when assessed for the entire annual range ( t = 7.857, d.f = 6, sig. (2 tailed) = .000 for MCP versus core areas, and t = 10.513, d.f = 6, sig. (2 tailed) = .000 for fixed kernel versus core areas) as well as for the dry season ( t = 6.324, d.f = 6, sig. (2 tailed) = .001 for MCP versus core areas, and t = 18.242, d.f = 6, sig. (2 tailed) = .000 for fixed kernel versus core areas) and the rain season ( t = 6.552, d.f = 5, sig. (2 tailed) = .001 for MCP versus core areas and t = 14.909, d.f = 5, sig. (2 tailed) = .000 for fixed kernel versus core areas). The core areas for the annual home ranges did not differ from either the dry season core areas ( t = .024, d.f = 6, sig. (2 tailed) = .982) nor the rain season core are as ( t = .956, d.f = 5, sig. (2 tailed) = .376). No significant difference was detected between the size of dry season core areas versus the size of rain season core areas ( t = .646, d.f = 5, sig. (2 tailed) = .542), however, the average size of the core area s (Mean SD) during the rain season was larger than the average size of the core area s during the dry season (64.33 43.83 for the rain season versus 47.67 21.46 for the dry season). A small sample size disallowed us from making similar comparisons for pumas. The cumulative home range sizes (km 2 ) based on MCP estimates calculated on a bi weekly basis, appeared to have reached an asymptote for all but one cat (Figure 4 5). Cat 02 displayed a very small increase toward the end of its monitoring but it is inconceivable that his range would have continued to increase since this individual had established himself as the dominant male in his territory. Moreover, this range estimate
71 was by far the largest reported in this study. The range of Cat 06 showed a sharp spike after week thirty and this trend was still rising when the collar was recovered Several cats (01, 02, 06, 08 and 09) showed a discernible increase in home range size at about the thirty week mark. The size of the home range of most cats stabilized after twenty weeks, for a period of about ten weeks, and then displayed an additional range expansion thereafter before reaching the final asymptote. The annual home range size for male jaguars averaged 257 km 2 (SD = 119) by minimum convex polygon and 264 km 2 (SD = 155) by fixed kernel. The female jaguar had an annual home range size of 1 11 km 2 (MCP) or 169 km 2 (FK). The annual home range of the male pumas averaged 205 km 2 (SD = 7) by MCP or 220 km 2 (SD = 26) by FK. During the dry season, the average home range size for male jaguars was 134 km 2 (SD = 34) by MCP and 188 km 2 (SD = 53) by FK. The female home range size during the dry season was 86 km 2 (MCP) and 129 km 2 (FK), while the average size for the male pumas was 109 km 2 (SD = 7) (MCP) versus 161 km 2 (SD = 18) (KP). During the rain season, the size of the home range for male jaguars ave raged 234 km 2 (SD = 91) (MCP) or 280 km 2 (SD = 144) (FK). Male pumas also averaged 159 km 2 (SD = 1.4) (MCP) or 230 km 2 (SD = 5) (FK) while the home range of the female jaguar was 99 km 2 (MCP) or 202 km 2 (FK) during the rain season. The rain season home ran ge of Cat 05 was excluded from these analyses because this individual was shot sometime toward the end of the dry season and his movements were severely impaired thereafter. Movement Patterns On average, jaguars moved 2.56 km/day for the duration of the st udy (Table 4 14). Mean daily distance travelled (km/day) for the duration of the study was significantly less than mean daily distance travelled during the dry season ( t = 2.736,
72 d.f = 6, sig. (2 tailed) = .03) but not significantly different from mean d aily distance travelled during the rain season ( t = 1.292, d.f = 6, sig. (2 tailed) = .24) (Table 4 15). On average, jaguars travelled similar distances daily during the rain season as they did during the entire study period (mean SD: 2.57 .75 for th e rain season versus 2.56 1.05 for the duration of the study). Mean daily distance travelled (km/day) during the dry season was significantly larger than mean daily distance travelled (km/day) during the rain season ( t = 2.624, d.f = 6, sig. (2 tailed) = .03). While the maximum daily distance travelled by jaguars was greater during the rain season than during the dry season (mean SD: 9.19 3.78 km versus 7.78 1.19 km) (Table 4 15) the difference was not significant ( t = .383, d.f = 6, sig. (2 ta iled) = .715). However, the maximum daily distance moved by jaguars during the dry season was significantly less than the maximum daily distance moved during the entire study ( t = 2.529, d.f = 6, sig. (2 tailed) = .04). The shape of the plot of net displa cement over time deviated strongly from the linear pattern expected of a purely random movement (Figure 4 6). Both jaguars and pumas displayed a bounded space use pattern (home range behavior) that stabilized after approximately 100 hours. The movement rat e during the day for jaguars were on average about one third (10 0.5 ) of their movement rates during the evening (Figure 4 7). Jaguars are more active at night, starting around 9PM, through the night, at dawn and the early morning up until around 9AM. Two small peaks occur during this time of activity, first between 11PM and 2AM, and again in the early morning between 7AM and 9AM. The movement rates at dusk were below the average. Pumas showed much less variability in diurnal movement rates (Figure 4 8) but this could be an artifact of a small
73 sample size (n=2 pumas). The movement rate for jaguars during the wet season (June through November), were on average about one half the movement rate during the drier months (Figure 4 9). Again, pumas showed much less variability in yearly movement rates (Figure 4 10) but again, this could be an artifact of the small sample size. In general, however, pumas showed substantially weaker yearly and diurnal cycles compared to jaguars. Movements for which the starting point was located in Lowland Broad leafed Moist Forest (LB LMF) were significantly longer than average while those movements that began in Shrub were significantly shorter than average (Figure 4 11). The effect of crossing a road was significant (Figure 4 12). Any move crossing a road was on average approximately 8 times faster than any move not crossing. The movement distribution of all cats (jaguars and pumas) were bimodal (Figure 4 13), that is, they all spent a lot of their time making relatively long range movements but occasionally made short range movements. For each of the nine collared cats, there was less than a 1% chance that the movement pattern could have resulted from chance (Table 4 16). In all cases, the z value was less than 2.58 thereby indicat ing that the movement pattern was clustered (Figure 4 14). Landscape and Habitat Use Habitat composition within jaguar home ranges differed significantly from what was available across the entire study area when fixed kernel estimates (95% probability co ntours, Figure 4 15) (habitat utilized) were compared with the total study area (available habitat TSA, Figure 4 16) during the dry season ( x 2 = 91.97, P < 0.001, G = 82.64, d.f = 6, P < 0.0001), during the rain season ( x2 = 180.03, P < 0.0001, G = 120.95, P < 0.0001), and during the entire annual period ( x 2 = 47.23, P < 0.0001, G =
74 40.23, P < 0.0001) (Table 4 17). After the removal of the injured individual, habitat composition within jaguar home ranges remained significantly different f rom what was available across the entire study area ( x 2 = 50.48, P < 0.0001 and G = 43.77, P < 0.0001 for annual FK versus TSA; x 2 = 180.28, P < 0.0001 and G = 120.94, P < 0.0001 for rain season FK versus TSA; x 2 = 279.14, P < 0.0001 and G = 199.15, P < 0 .001 for dry season FK versus TSA) (Table 4 17). Habitat composition based on proportions of jaguar locations differed significantly from what was available to each individual within its home range (MCP) during the dry season ( x 2 = 117.13, P < 0.0001, a nd G = 91.51, P < 0.0001), the rain season ( x 2 = 327.52, P < 0.0001 and G = 205.93, P < 0.0001) and for the entire year ( x 2 = 195.48, P < 0.0001 and G = 139.44, P < 0.0001) (Table 4 18). After the removal of the injured individual from this analysis, habit at composition based on proportions of jaguar locations remained significantly different from what was available within the individual home range polygons (MCP) during the entire year ( x 2 = 77.85, P < 0.0001 and G = 64.47, P < 0.0001), the dry season ( x 2 = 279, P < 0.0001 and G = 199.15, P < 0.0001) and the rain season ( x 2 = 70.25, P < 0.0001 and G = 62.98, P < 0.0001) (Table 4 18). Habitat composition represented by the proportion of GPS location fixes also differed significantly when compared with habitat proportions that were available across the total study area (TSA) during the entire year ( x 2 = 476.01, P < 0.0001 and G = 262.12, P < 0.0001), during the dry season ( x 2 = 362.58, P < 0.0001 and G = 220.65, P < 0.0001) and during the rain season ( x 2 = 782.61, P < 0.0001 and G = 368.07 and P < 0.0001) (Table 4 19). After removal of the injured individual from this analysis, habitat composition represented by the proportion of GPS location fixes remained significantly
75 different when compared with habit at that was available across the entire study site for the entire year ( x 2 = 187.86, P < 0.0001 and G = 40.23, P = 0.0001), during the dry season ( x 2 = 279.71, P < 0.0001 and G = 199.16, P < 0.0001) and the rain season ( x 2 = 145.36, P < 0.0001 and G = 111. 52, P < 0.0001). Compositional analysis comparing the proportional use of habitat was used to confirm the results of the previously described tests for second and third order habitat selection. The overall comparison of habitat use from the 95% FK home ran ges compared to habitat availability in the total study area, that is, second order selection (Johnson 1980), confirmed that jaguars did not establish annual home ranges at random ( x 2 (5 d.f .) = 14.25, P < 0.05) (Table 4 20). A ranking matrix ordered the habitat types in the sequence Shrub>Broadleaf>Water>Lowland Savanna>Agriculture>Wetland. Shrub was the highest ranking habitat and it was utilized significantly more than agriculture and wetland. There was no detectable difference in the use of the top two habitats. Similar to the pattern observed for habitat utilization during the entire annual period, the establishment of jaguar home ranges (95% FK) within the total study area during the dry season ( x (5 d.f .) = 26.2548, P < 0.0001) and the rain season ( x (5 d.f .) = 47.705, P < 0.0001) (Table 4 20) was not random. A ranking matrix ordered the habitat types in the sequence Broadleaf>Shrub>Agriculture>Lowland Savanna> Wetland>Water for the dry season and Shrub>Broadleaf>Agriculture>Lowland Savanna>Water>We tland for the rain season. Compositional analysis was subsequently used to test third order selection (Johnson 1980). Overall, jaguar use of habitat based on the proportional distribution of
76 GPS locations differed significantly from habitat composition w ithin the minimum convex polygons (x 2 (5 d.f .) = 19.58, P < 0.05) when considering the entire annual range (Table 4 21). A ranking matrix ordered the habitat types in the sequence Broadleaf>Shrub>Agriculture>Lowland Savanna>Water>Wetland. Broadleaf was th e highest ranking habitat and it was used significantly more than water and wetland. There was no detectable difference in the use of the top two habitats. Similar to the pattern observed for habitat utilization during the entire annual period, the propor tion of habitat locations based on GPS location fixes differed significantly from the proportions of habitat within minimum convex polygons, both during the dry season (x 2 (5 d.f .) = 23.72, P < 0.001) and the rain season (x 2 (5 d.f .) = 43.8, P < 0.001) (T able 4 21). A ranking matrix ordered the habitat types in the sequence Shrub>Broadleaf>Agriculture>Lowland Savanna>Wetland>Water for the dry season and Broadleaf>Shrub>Agriculture>Lowland Savanna>Water>Wetland for the rain season. In both cases, there was no detectable difference in the placement of Broadleaf and Shrub, suggesting that their relative placement was not important. Classification tree (dendrogram) models were generated for all collared cats. Both categorical (habitat categories) and numerica l (landscape measurements) data were used in these analyses, and individual models were developed for both jaguars and pumas. The number of branches for each tree ranged from four to ten for jaguars (two models with 4 branches, two models with ten branches and one each with six, eight and nine branches) and the two puma models contained six and eight branches (Figure 4 17 to 4 -25 ). The correct classification rate (CCR) for jaguar models (Cats 01 07) ranged from 87% to 96% while the CCR for the two puma mo dels (Cat 08 and 09)
77 were 90% and 92% (Table 4 22). All trees were pruned based on the cross validation relative error plot which selected a tree based on the one standard error rule. In this method, the number of leaves corresponding to the first point th at is within one standard error of the minimum is selected as the tree size. Monte Carlo re sampling showed that none of the nine predictive models could be obtained by random chance (P < 0.001 in all cases) (Figure 4 26 to 4 34 ). One thousand permutations of the data separated all nine models from the plot of random trees. Habitat was an important predictor for all models. Agriculture, Broadleaf forest and Shrubland were important predictors in all models, thereby validating the results of the compositiona l analysis and the Neu et al (1974) tests. Within the broad scale habitat category of broadleaf forest, jaguars selected for four different forest types, namely, Tropical evergreen seasonal broadleaf lowland forest (TESBLLF) over lime rich alluvium, TESBLL F on poor or sandy soils, TESBLLF on calcareous soils and TESBLL hill forest. Pumas, in addition to selecting for these broadleaf forest types, also selected for TESBLL swamp forest. Detailed descriptions of all habitat types, broad scale and fine scale, are available from the World Bank (http://wbln0018.worldbank.org/MesoAm/UmbpubHP.nsf/), the USGS EROS website (http://mitchnts1.cr.usgs.gov/data/otheragency.html) and the Central American Ecosystem Mapping Project (http://www.ecosystems.ws/ecosystems_map_c entral_america.html). These habitat variables, together with the perpendicular distance to the Western highway, were important predictors in all CART models (Figures 4 17 to 4 25 ). In addition, terrain ruggedness was an important predictor for four cats (t he female jaguar (Cat 04), a
78 young male jaguar (Cat 02), and the two pumas (Cat 08, Cat 09)) The presence of these four cats was characterized by terrain that was more rugged than the random locations. The perpendicular distance to roads was an important predictor for six cats, five jaguars (Cats 01, 03, 04, 05 and 06) and one puma (Cat 09). The presence of these cats was consistently closer to off roads than were random locations. Distance to nearest community was an important predictor for one jaguar (Ca t 03). In this case, jaguar presence was consistently further from communities than were random points.
79 Table 4 1.Total number of scats and prey items per species. Species Number of scats Number of prey items Jaguar ( Panthera onca ) 79 102 Puma ( Puma concolor ) 36 42 Unknown (failed to ID) 21 17 Canid sp. 1 Total 137 161 Table 4 2. Percent occurrence and relative biomass of prey consumed in 79 jaguar scats (n= 102 prey items). Prey Percent Occurrence (%) Prey weight (kg) Correction factor a Relative Biomass Consumed (%) Collared peccary ( Tayassu tajacu ) 14.70 17.53 2.59 17.33 Red brocket deer ( Mazama a mericana ) 2.94 16.70 2.57 3.44 Paca ( Agouti paca ) 5.88 6.04 2.19 5.86 Northern tamandua ( Tamandua mexicana ) 2.94 6.15 2.20 2.95 Northern raccoon ( Procyon lotor ) 5.88 5.20 2.16 5.78 Nine banded armadillo ( Dasypus novemcinctus ) 35.29 4.69 2.14 34.38 White nosed coati ( Nasua narica ) 5.88 3.18 2.09 5.59 Kinkajou ( Potos flavus ) 1.96 3.30 2.10 1.87 Central American agouti ( Dasyprocta punctata ) .98 2.81 2.08 .93 Hog nosed skunk ( Conepatu s sp ) 1.96 2.32 2.06 1.84 Greater grison ( Galictis vittata ) 1.96 2.40 2.06 1.84 Bird 6.86 1.00 2.02 6.31 Iguana ( Ctenosaura sp ) 3.92 2.50 2.07 3.69 Turtle 8.82 1.75 2.04 8.19 a based on Ackerman et al (1984).
80 Table 4 3. Percent occurrence and relative biomass of prey consumed in 36 puma scats (n=42 prey items). Prey Percent Occurrence (%) Prey weight (kg) Correctio n factor Relative Biomass Consumed (%) White tailed deer ( Odocioleus virginianus ) 2.38 35.28 3.21 3.27 Collared peccary ( Tayassu tajacu ) 26.19 17.53 2.59 29.02 Red brocket deer ( Mazama americana ) 2.38 16.70 2.57 2.62 Paca ( Agouti paca ) 42.86 6.04 2.19 40.15 Nine banded armadillo ( Dasypus novemcinctus ) 4.76 4.69 2.14 4.36 White nosed coati ( Nasua narica ) 2.38 3.18 2.09 2.13 Tayra ( Eira barbara ) 4.76 3.28 2.09 4.26 Kinkajou ( Potos flavus ) 7.14 3.30 2.10 6.42 Grey fox ( Urocyon cineargenteus ) 2.38 2.65 2.07 2.11 Bird 2.38 1.00 2.02 2.06 Cow (calf) 2.38 45 3.55 3.62 Table 4 4. Food Niche Breadth (diet diversity) and Mean Weight of Vertebrate Prey (MWVP) of jaguars, pumas, male jaguars and female jaguars. Specie (group) Niche Breadth a Standardized (B sta ) b MWVP (kg) Pumas 3.77 0.28 7.62 Jaguars 5.76 0.37 4.62 Male jaguars 5.15 0.35 4.90 Female jaguars 5.49 0.64 4.17 a Niche breadth following Levins (1968); b niche breadth following Colwell and Futuyma (1971). Table 4 5. Food Niche Overlap (diet similarity) between jaguars and pumas, male jaguars and pumas, and female jaguars and male jaguars. Groups Similarity index (Pianka) jaguars and pumas 0.40 male jaguars and pumas 0.37 female jaguars and pumas 0.51 female jaguars and male jaguars 0.92 Mean overlap 0.55 0.25
81 Table 4 6. Percent occurrence and relative biomass of prey consumed in 65 Male jaguar scats (n= 82 prey items). Prey Percent Occurrence (%) Prey weight (kg) Correction factor Relative Biomass Consumed (%) Collared peccary ( Tayassu tajacu ) 15.85 17.53 2.59 18.59 Red brocket deer ( Mazama americana ) 3.66 16.70 2.57 4.26 Paca ( Agouti paca ) 4.88 6.04 2.19 4.84 Northern tamandua ( Tamandua mexicana ) 2.44 6.15 2.20 2.43 Northern raccoon ( Procyon lotor ) 7.32 5.20 2.16 7.16 Nine banded armadillo ( Dasypus novemcinctus ) 37.80 4.69 2.14 36.62 White nosed coati ( Nasua narica ) 4.88 3.18 2.09 4.62 Kinkajou ( Potos flavus ) 2.44 3.30 2.10 2.32 Hog nosed skunk ( Conepatus sp ) 1.22 2.32 2.06 1.14 Greater grison ( Galictis vittata ) 1.22 2.40 2.06 1.14 Bird 4.88 1.00 2.02 4.44 Iguana ( Ctenosaura sp ) 3.66 2.50 2.07 3.43 Turtle 9.75 1.75 2.04 9.01 Table 4 7. Percent occurrence and relative biomass of prey consumed in 10 female jaguar scats (n= 15 prey items). Prey Percent Occurrence (%) Prey weight (kg) Correction factor Relative Biomass Consumed (%) Collared peccary ( Tayassu tajacu ) 13.33 17.53 2.59 15.82 Paca ( Agouti paca ) 13.33 6.04 2.19 13.38 Northern tamandua ( Tamandua mexicana ) 6.66 6.15 2.20 6.72 Nine banded armadillo ( Dasypus novemcinctus ) 33.33 4.69 2.14 32.83 White nosed coati ( Nasua narica ) 6.66 3.18 2.09 6.38 Greater grison ( Galictis vittata ) 6.66 2.40 2.06 6.29 Bird 13.33 1.00 2.02 12.34 Turtle 6.66 1.75 2.04 6.23
82 Table 4 8. Sex, age and body measurements of 16 jaguars ( Panthera onca ) captured in central Belize during 2008 2010. Animal ID Sex Age (yrs) Weight (kg) Body Length (cm) Tail Length (cm) Total Length (cm) Chest (cm) Neck (cm) JM01 M 6 8 58.2 128 54 182 72.5 48 JM02 M 2 3 52.2 123 59 182 72 47 JF03 F 4 5 43.5 116 56 172 63 44.5 JM04* M 6 8 56.2 125 59 184 78 48 JM05 M 4 5 53.5 134 65 199 76 49.5 JM06 M 3 4 52.6 126 66 192 75 50 JF07* F 7 8 39.9 118 59.5 175.5 75 42 JM08 M 5 6 55.3 123 57 180 64 50 JM09 M 5 6 56.2 123 52 175 81 53 JM10* M 2 3 53.5 125 55 180 85 51 JM11* M 3 4 66.2 133 66 199 81 51 JM12* M 9 11 44.5 126 58 184 75 49 JM13* M 4 5 54.4 126 62 188 74 48 JM14* M 3 4 59.9 127 64 191 75 50 JM15 M <2 47.2 116 64 180 69 47 JF16 F 4 6 45.35 119 64.5 183.5 67 45 Totals Jaguar Male (mean SD) 54.61 5.40 125.77 4.53 60.08 4.77 185.85 7.48 75.19 5.44 49.35 1.75 Jaguar Female (mean SD) 42.92 2.77 117.67 1.53 60.01 4.27 177.01 5.89 68.33 6.11 43.83 1.61 *represents individual with functional collars that provided data for this study.
83 Table 4 9. Sex, age and body measurements of 8 pumas ( Puma concolor ) captured in central Belize during 2008 2010, including two individuals with functional collars that provided data for this study. Animal ID Sex Age (yrs) Weight (kg) Body Length (cm) Tail Length (cm) Total Length (cm) Chest (cm) Neck (cm) PM01* M 3 4 35.4 125 76 201 62.5 35 PM02 M 3 4 39.9 110 72 182 63.5 39.5 PM03 M 2 3 30.8 130 72 202 63 37.5 PF04 F 6 8 21.8 104 65 169 54 28.5 PF05 F 1 2 16.3 102 66 168 48 27 PF06 F 1 2 20.8 96 67 163 56 29 PM07* M 2 3 39.9 118 70 188 69 39 PM08 M 2 3 38.5 111 68 179 Totals Puma Male (mean SD) 36.90 3.87 118.80 8.70 71.60 2.97 190.40 10.64 64.50 3.03 37.75 2.02 Puma Female (mean SD) 19.63 2.93 100.67 4.16 66.00 1.00 166.67 3.21 52.67 4.16 28.17 1.04 *represents individual with functional collars that provided data for this study. Table 4 10. Total trapping effort: number of captures, recaptures and trap nights per capture. Species Captures (recaptures) Trap nights per capture Jaguar 16 (5) 139 Puma 8 (7) 195 Tapir 7 (0) 417 Totals 31 (12) 68
84 Table 4 11. Estimated size of annual home ranges (km 2 ) of jaguars ( Panthera onca ) and pumas ( Puma concolor ) in central Belize. Range estimates based on minimum convex polygon and fixed kernel (95%, 75% and 50%). Total number of GPS locations used to calculate the estimate is also shown. Animal ID Species Home Range (km 2 ) Total # of GPS Locations MCP Kernel Probability Polygon 95% 75% 50% Cat 01 Jaguar 168 154 68 32 461 Cat 02 Jaguar 461 531 276 138 373 Cat 03 Jaguar 216 196 95 44 843 Cat 04 Jaguar 111 169 87 44 184 Cat 05 Jaguar 136 127 43 10 304 Cat 06 Jaguar 325 369 183 81 362 Cat 07 Jaguar 240 208 90 36 719 Cat 08 Puma 200 202 97 46 345 Cat 09 Puma 210 239 124 54 556 Table 4 12. Estimated size of the dry season home ranges (km 2 ) of seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. Range estimates based on minimum convex polygon (MCP) and fixed kernel (95%, 75% and 50%). Total number of GPS locations used to calculate the estimate is also shown. Animal ID Species Home Range (km 2 ) Total # of GPS Locations MCP Kernel Probability Polygon 95% 75% 50% Cat 01 Jaguar 103 147 79 39 157 Cat 02 Jaguar 194 231 105 55 119 Cat 03 Jaguar 114 139 65 30 211 Cat 04 Jaguar 86 129 71 39 105 Cat 05 Jaguar 136 169 81 31 202 Cat 06 Jaguar 110 274 165 88 43 Cat 07 Jaguar 144 169 74 35 207 Cat 08 Puma 114 174 85 44 80 Cat 09 Puma 104 149 61 30 158
85 Table 4 13. Estimated size of the rain season home ranges (km 2 ) of seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. Range estimates based on minimum convex polygon (MCP) and fixed kernel (95%, 75% and 50%). Total number of GPS locations used to calculate the estimate is also shown. Animal ID Species Home Range (km 2 ) Total # of GPS Locations MCP Kernel Probability Polygon 95% 75% 50% Cat 01 Jaguar 155 153 61 29 230 Cat 02 Jaguar 386 550 295 148 155 Cat 03 Jaguar 179 241 130 65 256 Cat 04 Jaguar 99 202 107 51 75 Cat 05 Jaguar 15 4.42 .98 .43 133* Cat 06 Jaguar 245 327 155 63 280 Cat 07 Jaguar 203 208 88 30 410 Cat 08 Puma 160 227 101 52 105 Cat 09 Puma 158 234 123 61 162 *data for a period of 2.5 months Table 4 14. Mean daily distances moved (km/day), maximum daily distances moved (km/day) and total distance travelled for the duration of the study for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. Animal ID n Distance travelled (km/day) Total distance travelled (km) Duration of Study (days) Min. Max. Mean Cat 01 461 0 8.04 2.28 878 385 Cat 02 376 0 8.72 2.02 819 409 Cat 03 843 0 11.94 3.09 1,551 502 Cat 04 184 0 6.48 1.95 312 160 Cat 05 339 0 8.06 1.46 224 154 Cat 06 365 0 16.21 2.49 1,005 403 Cat 07 719 0 18.36 4.64 1,424 307 Cat 08 345 0 10.13 2.43 924 380 Cat 09 569 0 10.18 2.36 987 419
86 Table 4 15. Mean daily distances moved (km/day), maximum daily distances moved (km/day) and total distance travelled during the dry and rain seasons for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. Animal ID Dry Season / Rain Season n distance travelled (km/day) total distance travelled (km) duration of study (days) minimum maximum mean Cat 01 159 / 232 0 / 0 6.35 / 8.04 3.02 / 2.56 260 / 488 86 / 191 Cat 02 121 / 157 0 / 0 8.72 / 8.04 2.71 / 2.08 231 / 397 85 / 191 Cat 03 213 / 257 0 / 0 7.84 / 7.23 2.86 / 2.08 291 / 217 102 / 104 Cat 03* / 201 / 0 --/ 11.94 --/ 3.49 --/ 625 --/ 179 Cat 04 107 / 77 0 / 0 6.07 / 6.48 2.19 / 1.70 168 / 141 77 / 83 Cat 05 44 / 135 0 / 0 9.36 / 4.25 4.47 / .49 103 / 33 23 / 68 Cat 06 204 / 282 0 / 0 8.06 / 16.21 2.24 / 3.71 191 / 691 85 / 186 Cat 07 207 / 410 0 / 0 8.09 / 11.36 4.81 / 2.43 408 / 746 85 / 175 Cat 08 80 / 105 0 / 0 10.13 / 9.47 2.80 / 2.10 224 / 359 80 / 171 Cat 09 158 / 162 0 / 0 9.51 / 7.77 1.81 / 1.91 152 / 376 84 / 197 *Individual sampled for two rain seasons and one dry season.
87 Table 4 16. Results of spatial cluster analysis (average nearest neighbor, method = Euclidian) procedure used to test the hypothesis that the activity pattern displayed by collared cats could be a result of random chance. Animal ID Season n OMD/EMD z value p value Cat 01 Rain 232 0.56 12.59 0.000 Dry 159 0.67 7.86 0.000 Annual 461 0.62 15.53 0.000 Cat 02 Rain 157 0.52 11.50 0.000 Dry 121 0.56 9.20 0.000 Annual 376 0.53 17.33 0.000 Cat 03 Rain 257 0.56 11.80 0.000 Rain 210 0.49 15.33 0.000 Dry 213 0.49 14.18 0.000 Annual 843 0.47 29.41 0.000 Cat 04 Rain 77 0.70 5.01 0.000 Dry 107 0.86 2.80 0.005 Annual 184 0.69 7.81 0.000 Cat 05 Rain 135 0.38 13.64 0.000 Dry 204 0.43 15.64 0.000 Annual 339 0.36 22.59 0.000 Cat 06 Rain 282 0.49 16.07 0.000 Dry 44 0.73 3.44 0.001 Annual 365 0.46 19.81 0.000 Cat 07 Rain 410 0.42 22.34 0.000 Dry 207 0.58 11.66 0.000 Annual 719 0.48 26.45 0.000 Cat 08 Rain 105 0.54 8.97 0.000 Dry 80 0.63 6.38 0.000 Annual 345 0.55 15.85 0.000 Cat 09 Rain 162 0.54 11.02 0.000 Dry 158 0.39 14.46 0.000 Annual 556 0.49 22.63 0.000 OMD = observed mean distance EMD = expected mean distance
88 Table 4 17. Habitat selection using the habitat utilization method of Neu et al. (1974) with Bonferroni confidence intervals. Second order habitat selection tested for the annual, dry season and rain season home ranges. Only habitats that were preferred or avoided are shown. No P value indicates a preference or avoidance that was not statistically significant. Habitat utilized within 95% Fixed kernels (KP) were compared to what was available across the total study area (TSA). Comparison Habitat Confidence Interval Proportion Available d.f. P FK vs TSA Agriculture 0.0738 0.1361 0.1362 Avoid Annual Broadleaf 0.5552 0.6546 0.5165 Prefer Lowland Savanna 0.1233 0.1979 0.2348 Avoid 6 <0.05 Shrub 0.0711 0.1326 0.0298 Prefer 6 <0.0001 Wetland 0.0009 0.0231 0.0415 Avoid 6 <0.05 Other 0.000 0.0177 0.0270 Avoid FK vs TSA Broadleaf 0.5907 0.6867 0.5308 Prefer 6 <0.05 Dry Season Lowland Savanna 0.1213 0.1942 0.2413 Avoid 6 <0.05 Shrub 0.0432 0.0937 0.0306 Prefer 6 <0.0001 Wetland 0.000 0.0159 0.0427 Avoid 6 <0.05 FK vs TSA Agriculture 0 .0738 0.1361 0.1362 Avoid Rain Broadleaf 0.5552 0.6546 0.5165 Prefer Season Lowland Savanna 0.1233 0.1979 0.2348 Avoid 6 <0.05 Shrub 0.0711 0.1326 0.0298 Prefer 6 <0.0001 Wetland 0.0009 0.0231 0.0415 Avoid 6 <0.05 Other 0.000 0.0177 0.0270 Avoid FK vs TSA* Agriculture 0.0412 0.1036 0.1362 Avoid 6 <0.05 Annual Broadleaf 0.5629 0.6796 0.5165 Prefer Wetland 0.0006 0.0302 0.0415 Avoid Other 0.000 0.0248 0.0270 Avoid FK vs TSA* Agriculture 0.0163 0.0634 0.1362 Avoid 6 <0.0001 Dry Season Broadleaf 0.6443 0.7546 0.5165 Prefer 6 <0.0001 Lowland Savanna 0.0761 0.1526 0.2348 Avoid 6 <0.0001 Shrub 0.0853 0.1649 0.0298 Prefer 6 <0.0001 Wetland 0.000 0.0133 0.0415 Avoid 6 <0.05 Other 0.000 0.000 0.0270 Avoid 6 <0.05 FK vs TSA* Agriculture 0.0412 0.1036 0.1362 Avoid 6 <0.05 Rain Broadleaf 0.5629 0.6796 0.5165 Prefer Season Wetland 0.0006 0.0302 0.0415 Avoid Other 0.000 0.0248 0.0270 Avoid *Data on injured jaguar removed from analysis.
89 Table 4 18.Third order habitat selection using the Neu et al. (1974) habitat utilization method with Bonferroni confidence intervals. The ranges of habitat proportions based on the actual location fixes are compared with the proportions available within individual home ranges. Third order selection tested for the dry season, rain season and th e entire annual period. Habitat utilized based on proportions of GPS locations were compared with habitats available within individual minimum convex polygons (MCPs). No P value indicates a preference or avoidance that was not statistically significant. Comparison Habitat Confidence Interval Proportion Available d.f. P GPS vs Individual MCPs Lowland Savanna 0.0845 0.1499 0.2192 Avoid 6 <0.0001 Annual Shrub 0.1215 0.1958 0.0557 Prefer 6 <0.0001 Wetland 0.000 0.0067 0.0227 Avoid 6 <0.05 Other 0.000 0.0137 0.0148 Avoid GPS vs Individual MCPs Lowland Savanna 0.0779 0.401 0.2225 Avoid 6 <0.0001 Dry Season Shrub 0.1066 0.1761 0.0565 Prefer 6 <0.0001 Wetland 0.000 0.0118 0.0231 Avoid GPS vs Individual MCPs Lowland Savanna 0.0786 0.1423 0.2192 Avoid 6 <0.0001 Rain Season Shrub 0.1600 0.2414 0.0557 Prefer 6 <0.0001 Wetland 0.000 0.000 0.0227 Avoid 6 <0.05 GPS vs Individual MCPs* Broadleaf 0 .6088 .7223 0.5939 Prefer Annual Lowland Savanna 0.0975 0.1808 0.2339 Avoid 6 <0.05 Shrub 0.0690 0.1430 0.0472 Prefer 6 <0.0001 Wetland 0.000 0.0072 0.0307 Avoid 6 <0.05 GPS vs Individual MCPs* Agriculture 0.0168 0.0629 0.0632 Avoid Dry Season Broadleaf 0.6453 0.7535 0.6046 Prefer 6 <0.05 Lowland Savanna 0.0768 0.1519 0.2382 Avoid 6 < 0.0001 Shrub 0.0861 0.1642 0.0480 Prefer 6 <0.0001 Wetland 0.000 0.0131 0.0313 Avoid 6 <0.05 GPS vs Individual MCPs* Broadleaf 0.6088 0.7223 0.5939 Prefer Rain Season* Lowland Savanna 0.1085 0.1948 0.2339 Avoid 6 <0.05 Shrub 0.0641 0.1363 0.0472 Prefer 6 <0.0001 Wetland 0.000 0.000 0.0307 Avoid 6 <0.05 *Data on injured jaguar removed from this analysis.
90 Table 4 19. Habitat utilization based on the Neu et al. (1974) method comparing proportions of habitat based on GPS location fixes versus proportions within the total study area (TSA). The ranges of habitat proportions based on the actual location fixes are compared with the proportions available within the TSA No P value indicates a preference or avoidance that was not statistically significant. Comparison Habitat Confidence Interval Proportion Available d.f. P GPS vs TSA Broadleaf 0.5255 0.6260 0.5165 Prefer 6 <0.05 Annual Lowland Savanna 0.0845 0.1499 0.2348 Avoid 6 <0.0001 Shrub 0.1215 0.1958 0.0298 Prefer 6 <0.0001 Wetland 0.000 0.0067 0.0415 Avoid 6 <0.001 Other 0.000 0.0137 0.0270 Avoid GPS vs TSA Broadleaf 0.5457 0.6437 0.5308 Prefer Dry Season Lowland Savanna 0.0779 0.1401 0.2413 Avoid 6 <0.0001 Shrub 0.1066 0.1761 0.0306 Prefer 6 <0.0001 Wetland 0.000 0.0118 0.0427 Avoid 6 <0.0001 GPS vs TSA Agriculture 0.0732 0.1354 0.1362 Avoid Rain Season Lowland Savanna 0.0786 0.1432 0.2348 Avoid 6 <0.0001 Shrub 0.1600 0.2414 0.0298 Prefer 6 <0.0001 Wetland 0.000 0.000 0.0415 Avoid 6 <0.0001 Other 0.000 0.0149 0.0270 Avoid GPS vs TSA* Agriculture 0.0299 0.0862 0.1362 Avoid 6 <0.05 Annual Broadleaf 0.6088 0.7223 0.5165 Prefer 6 <0.05 Lowland Savanna 0.0975 0.1808 0.2348 Avoid 6 <0.05 Shrub 0.0690 0.1430 0.0298 Prefer 6 <0.0001 Wetland 0.000 0.0072 0.0415 Avoid 6 <0.001 Other 0.000 0.0191 0.0270 Avoid GPS vs TSA* Agriculture 0.0163 0.0634 0.1362 Avoid 6 <0.0001 Dry Season Broadleaf 0.6443 0.7546 0.5165 Prefer 6 <0.0001 Lowland Savanna 0.0761 0.1526 0.2348 Avoid 6 <0.0001 Shrub 0.0853 0.1649 0.0298 Prefer 6 <0.0001 Wetland 0.000 0.0133 0.0415 Avoid 6 <0.05 Other 0.000 0.000 0.0270 Avoid 6 <0.05 GPS vs TSA* Agriculture 0.0530 0.1208 0.1362 Avoid Rain Season Broadleaf 0.5705 0.6868 0.5165 Prefer 6 <0.05 Lowland Savanna 0.1085 0.1948 0.2348 Avoid 6 <0.05 Shrub 0.0641 0.1363 0.0298 Prefer 6 <0.0001 Wetland 0.000 0.000 0.0415 Avoid 6 <0.05 Other 0.000 0.0208 0.0270 Avoid *Data on injured jaguar removed from this analysis.
91 Table 4 20. Compositional analyses testing second order habitat selection and ranking matrices for jaguars based on comparing proportional habitat utilized within fixed kernel (95% contours) home ranges with proportions of available habitat across the enti re study area: (a) during the annual period, (b) during the dry season and (c) during the rain season. A triple sign represents significant deviation from random at P < 0.05. Ho: D=0; Resources are used in proportion of availability. (a) Annual Resource Agric Broad LS Shrub Water Wetland Rank Agric ----+ (1) Broad + + + + (4) LS + + (2) Shrub +++ + + + +++ (5) Water +++ + + (3) Wetland --(0) (b) Dry Season Resource Agric Broad LS Shrub Water Wetland Rank Agric + + + (3) Broad + +++ + +++ + (5) LS --+ + (2) Shrub + + +++ + (4) Water ----(0) Wetland + (1) (c) Rain Season Resource Agric Broad LS Shrub Water Wetland Rank Agric + +++ +++ (3) Broad + + +++ +++ (4) LS + +++ (2) Shrub + + + + +++ (5) Water ----+ (1) Wetland --------(0) *+ and signs indicate preference or avoidance, while a triple sign denotes statistical significance.
92 Table 4 21. Compositional analysis testing third order habitat selection and ranking matrices for jaguars based on comparing the proportions of GPS location fixes for each animal in each habitat type with the proportion of each habitat type in the animals MCP range, (a) during the entire annual period, (b) during the dry season and (c) during the rain season. A triple sign represents significant deviation from random at P < 0.05. (a) Annual Resource Agric Broad LS Shrub Water Wetland Rank Agric + +++ +++ (3) Broad + + + +++ +++ (5) LS + + (2) Shrub + + +++ +++ (4) Water ------+ (1) Wetland ------(0) (b) Dry Season Resource Agric Broad LS Shrub Water Wetland Rank Agric + --+++ + (3) Broad + + --+++ +++ (4) LS --+ + (2) Shrub +++ + + +++ +++ (5) Water ------(0) Wetland ----+ (1) (c) Rain season Resource Agric Broad LS Shrub Water Wetland Rank Agric + + (3) Broad + + + +++ +++ (5) LS + + (2) Shrub + + +++ +++ (4) Water ----+ (1) Wetland ----(0) *+ and signs indicate preference or avoidance, while a triple sign denotes statistical significance.
93 Table 4 22. Correct classification rates for all nine classification trees. Rows of the matrix represent observed values and columns represent predicted. For Cat 01, there were 36 + 425 presence observations and the model correctly predicted 425 of these. Therefore, for Cat 01, 213 absences and 425 presences were correctly predi cted out of 711 observations (CCR = 89.73%). Cat 01 absence presence Cat 02 absence presence absence 213 37 absence 232 18 presence 36 425 presence 57 316 Cat 03 absence presence Cat 04 absence presence absence 233 17 absence 201 49 presence 33 487 presence 9 175 Cat 05 absence presence Cat 06 absence presence absence 223 27 absence 193 57 presence 22 282 presence 15 347 Cat 07 absence presence Cat 08 absence presence absence 173 77 absence 214 36 presence 25 694 presence 25 320 Cat 09 absence presence absence 240 10 presence 54 515
94 Figure 4 1. Spatial distribution of 136 felid scats collected in Central Belize.
95 Figure 4 2. Annual home range estimates (95% FK) for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ). Puma home ranges displayed in broken lines.
96 Figure 4 3. Dry season home ranges (95% FK) for six jaguars ( Panthera onca ) and two pumas ( Puma con color ) in central Belize. Puma ranges shown in broken lines.
97 Figure 4 4. Rain season home ranges (95% FK) of seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. Puma ranges shown in broken lines.
98 Figure 4 5. Cumulative home range sizes (km 2 ) based on minimum convex polygon estimates calculated at two week intervals.
99 Figure 4 6. Plot of log10 net displacement (absolute distance from a given starting point) against log10 time for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ). A purely random movement (no correlations between previous steps, relatively short steps, no home range) would have a slope of 0.5 (represented by red line). All movements start off with a slope higher than 0.5 and the n asymptote at around 100 hours.
100 Figure 4 7. Smoothed model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies with time of day. Movement rate at midnight and at 9AM was on average 2.5 times higher than at mid afternoon (~3PM).
101 Figure 4 8. Smoothed model showing how log10 transformed movement rates (distance moved per hour) for pumas ( Puma concolor ) varies with time of day. Movement rate at midnight was on average 1.5 times higher than at midday.
102 Figure 4 9. Smoothed model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies with days of the year. Movement rate during the dry season was on average 2 times the movement rate during the wet season.
103 Figure 4 10. Smoothed model showing how log10 transformed movement rates (distance moved per hour) for pumas ( Puma concolor ) varies with days of the year. Pumas displayed very little variation in movement rates across the two seasons.
104 Figure 4 11. Model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies with the ecosystem type the movement initiated in. Intercept represents the overall average movement rat e.
105 Figure 4 12. Model showing how log10 transformed movement rates (distance moved per hour) for jaguars ( Panthera onca ) varies depending on whether the movement crossed a road or did not cross a road. Any move crossing a road is on average 8 times faster (10.9) than any move not crossing a road.
106 Figure 4 13. Variability of movement rates (kilometers per hour) for seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ).
107 Figure 4 14. Spatial cluster analyses for all nine cats generated z values which were less than 2.58, thereby indicating that the spatial pattern demonstrated by the movement of all cats was clustered.
108 Figure 4 15. Proportions of habitat based on fixed kernel probability polygons (95%) utilized by seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Cat 01 Total Dry Rain Cat 02 Total Dry Rain Cat 03 Total Dry Rain Cat 04 Total Dry Rain Cat 05 Total Dry Rain Cat 06 Total Dry Rain Cat 07 Total Dry Rain Cat 08 Total Dry Rain Cat 09 Total Dry Rain Agriculture Broadleaf Lowland Savanna Shrub Water Wetland Other
109 Figure 4 16. Proportions of habitat that was available to seven jaguars ( Panthera onca ) and two pumas ( Puma concolor ) in central Belize. 0.14 0.52 0.23 0.03 0.01 0.04 0.03
110 Figure 4 17. Pruned classification tree model (Dendrogram) for Cat 01. Model based on CART analyses for predicting jaguar ( Panthera onca ) locations based on habitat and landscape measurements associated with 461 GPS lo cations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each terminal node is labeled with the proportion of observations within the specified category. The total numbe r of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left. The first split is along the categorical variable habitat category. If habitat is any of the thirteen variables bcfgijkmopqrs (see Appendix 1 for specific habitat categories represented by these letter codes) then follow the tree to the left (true conditions go left) to find that the prediction is a random point (0) 89% of the times for 162 cases. Follow the tree to the right to find that if habitat is any of the six habitat types represented by the letter codes adehln, and if the distance to the Western highway is greater than 6.75 km, and the distance to off roads is less than 3.72 km then the prediction is presence and this predict ion is correct 92% of the time for 462 cases. Correct classification rate: Null = 65%, Model = 90% (638/711), Kappa = .78. Unpruned tree has 24 terminal branches.
111 Figure 4 18. Pruned classification tree model (Dendrogram) for Cat 02. Model based on CART analyses for predicting jaguar ( Panthera onca ) locations based on habitat and landscape measurements associated with 376 GPS locations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each terminal node is labeled with the proportion of observations within the specified category. The total number of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left The first split is along the categorical variable habitat. If the habitat is any of the variables bfmnpq (see Appendix 1 for specific habitat variables represented by these letter codes) then follow the tree to the right (true conditions go left). Additi onally, if the distance to the Western highway is greater than 2.3km but less than 8.9km, and if the distance to nearest river is less than 2.2km then the prediction is presence and this prediction is correct 97% of the times for 251 cases. Correct classif ication rate: Null = 60%, Model = 88% (548/623). Kappa = .823. Unpruned tree has 32 terminal branches.
112 Figure 4 19. Pruned classification tree model (Dendrogram) for Cat 03. Model based on CART analyses for predicting jaguar ( Panthera onca ) locations based on habitat and landscape measurements associated with 843 GPS locations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each terminal node is labeled wi th the proportion of observations within the specified category. The total number of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left. The first split is along the categorical variable habitat. If the habitat category is represented by the letter codes adghln or r (see Appendix 1 for a list of the fine scale habitat classes and the corresponding letter codes) then follow the tree to the right. If distance to the Western highway is grea ter than 4.677km but less than 15.06 km, and the distance to community is less than 4.939 then the prediction is presence and this prediction is correct 98% of the time for 361 observations. Distance to off roads and distance to nearest river were also im portant predictors. Correct classification rate: Null = 68%, Model = 94% (720/770). Kappa = 0.87. Unpruned tree has 21 terminal branches.
113 Figure 4 20. Pruned classification tree model (Dendrogram) for Cat 04. Model based on CART analyses for predictin g jaguar ( Panthera onca ) locations based on habitat and landscape measurements associated with 184 GPS locations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each te rminal node is labeled with the proportion of observations within the specified category. The total number of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left. The first split is along the categorical variable habitat. If the habitat category is represented by any of the letter codes a or b (see Appendix 2 for a list of the broad scale habitat classes and the specific corresponding letter code) then follow the tree to the right. Subsequ ently, if the distance to off roads is less then 2.7km, and the distance to the Western highway is between 3.1km and 15.9km, and terrain ruggedness is greater than 2.0, then the prediction is presence and this prediction is correct 78% of the time for 224 observations. Correct classification rate: Null = 58%, Model = 87% (376/434). Kappa = .733. Unpruned tree has 21 terminal branches.
114 Figure 4 21. Pruned classification tree model (Dendrogram) for Cat 05. Model based on CART analyses for predicting ja guar ( Panthera onca ) locations based on habitat and landscape measurements associated with 339 GPS locations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each termin al node is labeled with the proportion of observations within the specified category. The total number of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left. The first split is along the categorical variable habitat. If the habitat category is represented by any of the letter codes aeh or n (see Appendix 1 for a list of the fine scale habitat classes and the specific corresponding letter code) then follow the tree to the right (true cases go left). Subsequently, if the distance to the Western highway is less than 8.3 but greater than 2.2 then the prediction is presence and this prediction is correct 97% of the time for 221 cases. Distance to off roads, habitat edge and distance to communit ies are also important variables affecting 75% of 72 predictions. Correct classification rate: Null = 55%, Model = 91% (505/554). Kappa = .823. Unpruned tree has 24 terminal branches.
115 Figure 4 22. Pruned classification tree model (Dendrogram) for Cat 06. Model based on CART analyses for predicting jaguar ( Panthera onca ) locations based on habitat and landscape measurements associated with 365 GPS locations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each terminal node is labeled with the proportion of observations within the specified category. The total number of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left. The first split is along the categorical variable habitat. If the habitat category is represented by any of the letter codes aegijlmo or p (see Appendix 1 for a list of the fine scale habitat classes and the specific corresponding let ter code) then follow the tree to the right (true cases go left). Next, if the distance to nearest off road is less than 3.35km and the distance to the Western highway is less than 8.966km, then the prediction will be presence for 97% of 194 cases given th at the habitat is neither agilo or p. Distance to nearest river was also an important predictor. Correct classification rate: Null = 59%, Model = 88% (540/612 or .8812). Kappa = .906. Unpruned tree has 24 terminal branches.
116 Figure 4 23. Pruned class ification tree model (Dendrogram) for Cat 07. Model based on CART analyses for predicting jaguar ( Panthera onca ) locations based on habitat and landscape measurements associated with 719 GPS locations (presence =1) and 250 randomly distributed points (rand om = 0). Each partition in the dendrogram is labeled with the splitting rule and each terminal node is labeled with the proportion of observations within the specified category. The total number of observations corresponding to a given terminal group is pr esented in parenthesis. In this model true conditions go left. The first split is along the quantitative variable distance to Western highway. If this distance is less than 8.34 km, and the habitat category is bd or h then the prediction is presence and t his prediction is correct for 90% of 751 cases. See Appendix 2 for a list of the broad scale habitat classes corresponding to the letter codes. Forest/non Forest edge is also an important predictor. Correct classification rate: Null = 74%, Model = 89% ( 867/969)Kappa = .775. Unpruned tree has 43 terminal branches.
117 Figure 4 24. Pruned classification tree model (Dendrogram) for Cat 08. Model based on CART analyses for predicting puma ( P uma concolor ) locations based on habitat and landscape measuremen ts associated with 345 GPS locations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each terminal node is labeled with the proportion of observations within the specif ied category. The total number of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left. The first split is on the quantitative variable distance to off roads. If the distance to nearest off road is less than 2.377 follow the tree to the right. Subsequently, if the distance to the Western highway is less than 13.85km but greater than 4.613, and if the habitat is either bd or i, and if the terrain ruggedness is greater than 2.563 then the pred iction is presence and this prediction is correct 90% of the time for 356 cases. Correct classification rate: Null = 58%, Model = 90% (534/595) Kappa = .795. Unpruned tree has 25 terminal branches.
118 Figure 4 25. Pruned classification tree model (Dendrogram) for Cat 09. Model based on CART analyses for predicting puma (Puma concolor) locations based on habitat and landscape measurements associated with 569 GPS locations (presence =1) and 250 randomly distributed points (random = 0). Each partition in the dendrogram is labeled with the splitting rule and each terminal node is labeled with the proportion of observations within the specified category. The total number of observations corresponding to a given terminal group is presented in parenthesis. In this model true conditions go left. The first split is along the categorical variable habitat. If the habitat category is represented by any of the letter codes adelmo or p (see Appendix 1 for a list of the fine scale habitat classes and the specific corresponding letter code) then follow the tree to the right (true cases go left). Subsequently, if distance to off roads is less than 2.256, and terrain ruggedness is greater than 2.7, and distance to the Western highway is greater than 2.304 km but less than 13.54km, and if the perpendicular distance to nearest river is less than 1.112 then the prediction is presence and this is prediction is correct for 98% of 452 cases. Null = 69%, Model = 92% (755/819), Kappa = 0.844. Unpruned tree = 20 leaves.
119 Figure 4 26. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 01. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model cou ld not have been generated by random chance.
120 Figure 4 27. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 02. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model could not have been generated by random chance.
121 Figure 4 28. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 03. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model cou ld not have been generated by random chance.
122 Figure 4 29. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 04. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model could not have been generated by random chance.
123 Figure 4 30. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (t riangle) for Cat 05. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model could not have been generated by random chance.
124 Figure 4 31. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 06. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model cou ld not have been generated by random chance.
125 Figure 4 32. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 07. P value based on Monte Carlo resampling. Permut ation test result: P < 0.001 indicating that this model could not have been generated by random chance.
126 Figure 4 33. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 08. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model cou ld not have been generated by random chance.
127 Figure 4 34. Distribution of correct classification rates (CCR) for 1,000 randomly generated trees versus the CCR of the model (triangle) for Cat 09. P value based on Monte Carlo resampling. Permutation test result: P < 0.001 indicating that this model could not have been generated by random chance.
128 CHAPTER 5 SYNTHESIS AND CONSERVATION IMPLICATIONS Since the first study of jaguars in the wild in Belize (Rabinowtiz and Nottingham 1986) numerous studies on different aspects of jaguar ecology have been conducted across the entire country (see chapter 2). In the midst of widespread land cover changes that have swept the wider Central American region since th ese first studies, Belize has managed to maintain relatively large trac t s of contiguous forests. Large carnivores, such as jaguars and pumas, that require extensive trac t s of forested habitats with abundant prey have been able to survive and maintain viabl e numbers. In fact, as suggested by Sanderson et al. (2002), Belize may still hold one of the healthier jaguar populations in Central America. However, the long term survival of the jaguar and the puma will ultimately depend on their ability to survive and reproduce within fragmented landscapes. In this regard, studies that advance our understanding of the ecology and conservation needs of these top carnivores within these altered landscapes will become increasingly important, especially in Central America where no single protected area is large enough to sustain a viable jaguar population. Proper management of the human influenced landscape that invariably connects networks of protected lands, such as the landscape of central Belize, is necessary for the lo ng term survival of both jaguars and pumas. Diet The jaguar is sympatric with the puma throughout most of its geographic range. Together these two top predators help maintain the structure and function of entire ecosystems within which they co exist (T erborgh et al. 2002). One of the primary mechanisms through which these felids are able to impact these systems is via their
129 predatory behavior. The jaguar, for example, is an opportunistic hunter (Seymour 1989) and is known to prey on more than 85 species across its range (Sunquist and Sunquist 2002). The puma, however, has been overshadowed by the jaguar and much less is known about the Central and South American populations (Laundr and Hernndez 2010). Understanding the dietary needs of these two sympat ric felids and how sympatry is facilitated is an important component of any meaningful management and conservation strategy for these predators. The diet of sympatric jaguars and pumas ha ve been studied in Mexico (Nunez et al. 2000), in Venezuela (Scognam illo et al. 2003), in Guatemala (Novack et al. 2005) and in southern Belize (Foster et al. 2010). Across these study sites, diet composition has varied in terms of species richness, diet diversity and the mean weight of vertebrate prey. D iet overlap betwee n sympatric jaguars and pumas also differed across these studies. Th ese differences are perhaps partly a reflection of the differences in the availability of prey. As the availability and abundance of prey is compromised due to, for example, habitat fragm entation and an associated increase in subsistence hunting, jaguars and pumas will be forced to increase their diet diversity due to lower levels of selected prey species. It also follows that, in well protected landscapes, where the prey base is healthy, jaguars and pumas will have lower diet diversity due to the increased availability of preferred prey. Wide ranging species such as jaguars and pumas must demonstrate a certain level of diet flexibility in order to withstand the landscape level changes that are altering the abundance and availability of the prey base. The long term viability of these predator s will partly depend on their ability to make these adaptations
130 T he MWVP differed for jaguars and pumas, with the puma having a MWVP that was substantially (165%) heavier than that for jaguars. This is interesting because jaguars are, in general, larger than pumas. However, significant size and weight differences exist between the sexes. At the extreme of such comparisons (in central Belize), on average, male jaguars can weigh 270% more than female pumas. Certainly, this can have implications on prey selection. For example, it is unlikely that a female puma (average weig ht of 19.63kg) would prey on heavily armored species such as the armadillo or turtle because of the limited physical ability to penetrate the protective carapace associated with these prey. Male jaguars are also substantially larger than both female jaguar s and male pumas, weighing on average, 150% more than male pumas and 130% more than female jaguars. On average, pumas also took heavier prey than male (155%) and female (174%) jaguars. The size and relative strength of the jaguar (Biknevicius et al. 1996) gives it a competitive advantage by allowing it to select and consume prey species such as armadillos and turtles that require less hunting effort. While the puma still has the ability to take armadillos, the cost benefit of consuming this armored prey (No vack et al. 2005) may serve to discourage pumas from targeting this species, as evidenced by the relatively low proportion of relative biomass armadillos contribute to the diet of pumas. The most important contribution to the diet of jaguars and pumas, bot h in terms of percent occurrence and of relative biomass, was the armadillo and the paca, each contributing 35.29% (34.38%) and 42.86% (40.15%) respectively, similar to studies in southern Belize (Foster et al. 2010) and the Guatemalan rainforest (Novack e t al. 2005). The differences in MWVP between jaguars and pumas can be attributed to the fact that pumas take relatively more pacas while
131 jaguars take more armadillos. Additionally, the white collared peccary is taken more frequently by pumas. The diet of jaguars in central Belize was more species rich (14 prey species for jaguars versus 11 prey species for pumas) and more diverse (Food Niche Breadth of .37 for jaguars versus .28 for pumas) than the diet of pumas. This relationship remains the same when ma le jaguars are compared with pumas but not when female jaguars are compared. The diet of female jaguars was less species rich (8 prey species for female jaguars versus 11 prey species for pumas) than that of pumas but more than twice as diverse (Pianka ind ex of .64 for female jaguars versus .28 for pumas). Female jaguars have smaller home ranges and must therefore exploit a wider resource base if they are to meet the requirements to survive and reproduce. While armadillos still represent a significant porti on of the diet of female jaguars (32.83% relative biomass), collared peccaries, pacas and birds were also relatively more important. The degree of dietary overlap between jaguars and pumas was higher in central Belize than in the Guatemalan rainforests (No vack et al 2005) and in southern Belize (Foster et al 2010). This is probably a reflection of a prey base that is less diverse due to the many anthropogenic forces that shape the central Belize landscape. The dietary overlap was more pronounced between pu mas and female jaguars. Assuming that the pumas were predominantly males (concurrent camera trap data suggests this was the case) then these two groups are very similar in size and body weight. This would help to explain the greater overlap in diet (Pianka index of 0.51). The relatively smaller home range of the female jaguar, coupled with the fact that females tend to live a more secretive lifestyle, especially during cub rearing, suggests that female jaguars cannot be as
132 selective as male jaguars and must expand their niche breadth. A wider niche breadth makes female jaguars more versatile in that it reduces their dependence on any one particular species and allows them to exploit a wider array of available prey. This in turn may enhance survival and the a bility to raise and care for cubs. Inevitably, the diet of jaguars and pumas will have implications on the condition of dentition. Tooth wear and general condition, as well as gum line recession are criteria that have been used to age the puma ( Currier 1 979, Ashman et al. 1983, Shaw 198 7 and Laundre et al. 2000) and should also be applicable to jaguars The diet of these two felids will have direct implications on tooth wear, gum line recession and general condition of the dentition and these characteri stics will inevitably deteriorate with time There are advantages and disadvantages, for both jaguars and pumas, in the selection of their preferred prey. While the puma may lack the physical ability to hunt and consume armadillos and turtles on a consiste nt basis, it probably benefits from specializing on the paca because of the limited adverse effects the consumption of this prey can have on dentition (due to challenges associated with penetrating the tough carapace). Given the absence of the tough carap ace associated with armadillos and turtles, one of the benefits the pumas may derive by concentrating on the paca s probably the maintenance of intact dentition that deteriorates at a much slower rate thereby prolonging their hunting ability (and efficienc y) However, the dependence of the puma on the paca has implications for puma conservation since it makes this species exceptionally vulnerable to fluctuations in the availability of this particular species. The paca is also the preferred game meat of many local residents in central Belize (O. Figueroa pers. obs.) and it is possible that subsistence hunting and the game meat
133 market may already be impacting the paca population and, by extension, the puma population in this area. The jaguar, being larger and more powerful, has the ability to exploit the resource that is most available. This confers an immediate competitive advantage for the jaguar. Armadillos and turtles are less agile than pacas and therefore, presumably, more readily available. The jaguar has the physical ability to easily penetrate the defenses of these armored prey species, thereby expending less energy than if it were to pursue the more agile and elusive paca (Novack et al. 2005) However, consistent feeding on these species means that j aguars must penetrate the tough protective carapace and this can have profound and long term impacts on the dentition, in particular, intact canines. The overall dentition, but in particular the canines of jaguars tend to deteriorate at a much faster pace than that of pumas (O. Figueroa pers. obs.). Given the high frequency with which jaguars consume armadillos and turtles in central Belize, and the absolute reliance of these cats on intact dentition, the length of time that jaguars remain in peak physical condition is probably shorter than that for pumas. Therefore, while pumas must expend more energy in pursuing the paca it is possible that there may be long term benefits associated with this strategy. Further studies on how diet affects dentition and the subsequent implications for jaguar and puma health would provide important information with important conservation implications for these two sympatric carnivores. If female jaguars are in more direct competition with male pumas than male jaguars ar e, as suggested by the food niche overlap, then female jaguars must develop adaptive strategies to overcome this disadvantage. One strategy seems to be that of developing a wider niche breadth (diet diversity) than either male pumas or male
134 jaguars. This w ider niche breadth also makes the female jaguar more resistant to any fluctuations in the availability of any particular prey species. This is important, especially in central Belize where hunting pressures are becoming increasingly important. Moreover, b ecause of the exorbitant numbers of armadillos needed to sustain female jaguars if the diet was comprised solely of armadillos (Nunez et al. 2002, Foster et al 2010), then increasing diet diversity would be one strategy to limit dependence on one species. For example, it has been estimated that a single adult female jaguar would need to kill 500 to 630 armadillos per year if her diet was comprised solely of this species (Foster 2008, Foster et al. 2010). Given the large home ranges of jaguars and pumas, a nd the relatively small sizes of protected areas within the study site it was not possible to compare the diet of jaguars and pumas within protected and unprotected lands. Further complicating any such comparisons is the fact that male jaguars and male pum as can move relatively large distances within short time periods, and have the potential to traverse any given protected area in Belize within 24 hours. Studies that aim to compare diet of jaguars and pumas between habitats, or between protected and unprot ected lands (e.g. Foster et al. 2010), or between hunted and non hunted areas (e.g. Novack et al. 2005), or even within defined areas (e.g. Nunez et al. 2000) must create meaningful buffers to ensure that cats hunting in one area of comparison are not depo siting scats in another. The results of this study can be used to delineate such buffers for studies in Mesoamerica. Conversely, the advances in genetic analysis of scats can be used to identify these remains to individual cats and then this information c an be used to guide comparisons of diet.
135 The dietary habits of jaguars and pumas reduce s competition and thereby facilitate coexistence. The jaguars preference for armadillos as well as the pumas preference for pacas have been demonstrated in the Cocksco mb basin of southern Belize (Rabinowitz and Nottingham 1986, Foster et al. 2010) and in Guatemala (Novack et al. 2005). The results of this study were similar when jaguars (males and females combined) and pumas were compared. However, differences exist whe n sex is considered in these comparisons. While many studies on jaguar diet have been conducted within the last decade none have considered differences between the sexes. This is important because very limited information exists on the ecology of female ja guars and the longevity of any jaguar population will undoubtedly depend on the success of females and their ability to raise and care for young. Home Range and Movement Patterns There is a wide variation in published reports on the size of jaguar home ranges (Schaller and Crawshaw 1980, Rabinowitz and Nottingham 1986, Crawshaw and Quigley 1991, Nunez et al. 2002, Scognamillo et al. 2003, Crawshaw et al. 2004, Cavalcanti and Gese 2 009). The size and placement of home ranges are influenced by several factors. It has been suggested that female jaguars will select home ranges based on the abundance and availability of prey whereas male jaguars will select home ranges based on the prese nce of established female home ranges (Sandell 1989), as well as the abundance and availability of prey. Females have smaller ranges than males. The range of a female jaguar overlaps, to varying degrees, with that of several males. Therefore, any given fem ale within an established territory will have access to several males. The territories of male jaguars are not exclusive. However, there are important differences in the size and placement of the home ranges of male jaguars.
136 Within any given area that is occupied by jaguars, there will exist several overlapping territories and a well defined hierarchical system (social structure). Some jaguars are predisposed to becoming dominant males whereas others are not. In central Belize, two dominant male jaguars we ighed 112% and 120% more than the other resident males. Presumably, because newly established dominant male s will not be completely familiar with a newly established home range, the size of these newly established ranges will probably be larger than that of other resident males. Initially, a newly established dominant male will traverse a much larger area and then, with time, as they become more familiar with their territory, and learn the spatial locations of important areas (female territories, productiv e hunting areas etc), the size of these ranges will probably shrink in an effort to maximize efficiency During this initial exploratory phase the young male will map the available space to determine where the critical areas are (high prey availability, fe male territories). Eventually, the dominant male will adjust the size of his home range so as not to expend time and energy patrolling unsuitable areas. The other resident (non dominant) males will have home ranges that are, on average similar in size to t hat of an established dominant male. However, through an elaborate communication system (olfactory, visual and vocal) jaguars minimize any potential encounters. While confrontations between male jaguars are rare, they do occur, especially during the recept ive period of female jaguars. We found direct evidence to suggest that a confrontation occurred between a newly established dominant male with the former (and older) dominant male. On February 13 th and 14 th of 2011 two male jaguars and one female jaguar we re located in very close proximity of each other. The two male jaguars carried functional GPS collars. A combination of both GPS tracking
137 and camera data documented an apparent encounter between the two males. First, the older male was documented moving ne xt to the female. Subsequently, the two males were documented in the same area, with the younger male following closely behind the older male along an important travel route. The older male was observed with what appears to be an injury below his left shou lder. Immediately thereafter, the younger male was documented in the company of the previously described female. The older male was subsequently marginalized to less productive habitats but appeared to move back into his former range within five days after the untimely death of the dominant male. We found no evidence to suggest that female jaguars had exclusive territories. Three females were captured in the study area and while only one was monitored for an entire season, their territories overlapped for t he brief monitoring periods (3 and 4 weeks). The large size of the territories of male and female jaguars makes it physically impossible to defend as exclusive. However, roaming of large areas may actually increase reproductive success by increasing the ch ances of encountering females (Davies 1978, Lott 1984). Within an established social structure there will be an established dominant male, several resident males and possibly a few older males. The resident males compete for resources and can potentially c ompete for females as well. The older males are marginalized and forced into less productive habitats where food resources are limited and access to females is greatly reduced. In the Brazilian Pantanal, Cavalcanti and Gese (2009) found evidence that the movement of female jaguars were not completely confined to within the territories of males. However, other studies (Rabinowitz and Nottingham 1986, Schaller and Crawshaw 1980) have
138 suggested that resident females were restricted to within the range of resi dent males. Data from this study suggests that the range of several males can overlap with the range of a single female. However, while several males may have access to females in any given area the established hierarchical system provides competitive adv antages for the established dominant male. Therefore, while the mating system may be polygynous and promiscuous (Cavalcanti and Gese 2009) and whereas several males may have access to female territories, not all cats have equal opportunities (probability) of mating. It has been suggested that jaguars may exhibit some degree of sociality (Schaller and Crawshaw 1980, Cavacanti and Gese 2009) but this behavior may only occur in areas with abundant prey. In central Belize, where the prey base cannot be conside red abundant and may actually be scarce at certain times of the year (dry season) we found no evidence suggesting any degree of sociality and our data indicates that jaguars in this area were solitary. Studies in the Panatanal of Brazil have shown that the range of jaguars during the wet season was smaller than the range during the dry season (Crawshaw and Quigley 1991, Cavalcanti and Gese 2009). It was suggested that during the wet season prey became concentrated in smaller areas and this was reflected in the smaller ranges during this time (Crawshaw and Quigley 1991). In central Belize, the wet season home range size was significantly larger than the home range size during the dry season, the opposite of the pattern observed in the Brazilian Pantanal. Howe ver, the geographic range of the jaguar is very large, and within this extensive range jaguars exist in many different habitat types and across many different regions (Oliveira 1994, Sanderson et al. 2002). Undoubtedly, the ecology of jaguars will vary acr oss regions and habitats.
139 The dry seasons in central Belize can be intense with very little (and sometimes none) precipitation during the months of April and May. During this time water can become extremely scarce with only major rivers providing a reliab le source of this important resource. Since all prey species depend on this resource, the prey species will also become less widespread and it is possible that the availability will also be lowered during this time Jaguars will respond to shifts in the av ailability of prey by shifting the core areas of seasonal ranges to correspond with these shifts in prey availability. The lower availability of prey during the dry season also means that within this smaller range jaguars will be more active since they wil l need to increase their search effort to locate available prey. During the wet season food resources become more abundant and presumably more available This means that jaguars can expand their ranges and no longer need to search more intensively in small er areas. Jaguars tend to move less during the wet season since food is not as difficult to locate. While they tend to move less during the wet season the total available area increases with increasing resource availability. Jaguars now make more long dist ance movements to reach other productive areas and this is reflected in a larger home range during the wet season. Similar to studies in the Brazilian Pantanal (Cavalcanti and Gese 2009), the size of core areas during the dry season did not differ from the size of core areas during the wet season. In addition, the size of the core areas during the annual period did not differ from those during the two seasons. The similar size of the core areas across the wet and dry seasons, as well as for the annual perio d, regardless of significant variations in the size of the home ranges during the respective periods is interesting. There seems to be a minimum area that must be available before the establishment of a viable home
140 range. Outside this core zone jaguars can tolerate human induced pressures (for example citrus farms, cattle farms) but these core areas will provide a high proportion of the resources necessary for jaguar survival. Our study suggests that the average size of these core areas for jaguars are appr oximately 50 to 65 km 2 and are characterized by low levels of human disturbance and preferred jaguar habitats. Variable patterns of activity have been reported for jaguars across their range (Emmons 1987, Schaller and Crawshaw 1980, Rabinowitz and Notti ngham 1986, Crawshaw and Quigley 1991, Crawshaw 1995, Farrell 1998, Scognamillo et al. 2003, Cavalcanti and Gese 2009). Based on these reports, jaguars in Peru and Brazil were active during the day and during the night, whereas, in Mexico, Belize and Venez uela they were mostly active at night. Crawshaw and Quigley (1991) reported jaguars to be more diurnal with activity peaks at dawn, noon and dusk. Cavalcanti and Gese (2009) reported that jaguars were active at dawn and dusk, and that they travelled less d uring the day. Camera trap data in southern Belize also suggested that jaguars had activity peaks at dawn and dusk and that they were less active during the day (Harmsen et al. 2009). The only other radio telemetry study conducted in Belize (Rabinowitz and Nottingham, 1986) found that jaguars were predominantly nocturnal but that there was individual variation in activity. In our study we found that jaguars were most active at night and early to mid morning. Jaguar activity was greatly reduced from mid day to dusk and was lowest in the mid afternoon (~3PM). Rabinowitz and Nottingham (1986) and Harmsen et al. (2009) found that jaguars exhibited a peak in activity in the early evening (6PM to 9PM). Our study found jaguars to be relatively less active during t his time period but had a clear peak in activity in the early morning, up until around 9AM. In
141 accordance with Rabinowitz and Nottingham (1986) we also found that jaguars exhibited individual variation in activity, but in central Belize this variation was characterized by different short peaks in the early to mid morning time period. In general, while most studies suggest that jaguars are much less active during the day our study shows that in central Belize jaguars are active during the early to mid mornin g period. In central Belize, jaguars are active at night, starting around 9 or 10PM, up until the early to mid morning (around 9AM). In the mid afternoon (around 3PM) jaguars were, on average, about 3 times less active than they were in the late night to e arly morning times. Cavalcanti and Gese (2009) reported no seasonal differences in the movement rates of jaguars at either dawn, dusk or during the day. Crawshaw and Quigley (1991) also reported that there were no seasonal differences in activity. How ever, Rabinowitz and Nottingham (1986) reported that one male jaguar was significantly more active during the dry season. Our study also showed substantial seasonal differences in movement rates and that jaguars were least active during the wet season. The seasonal patterns of activity probably reflect the availability of prey (Sunquist and Sunquist 2002, Emmons 1989), with an increase in activity or movement rates reflecting a decrease in the availability of prey during the drier months. Jaguars probably i ncrease their hunting effort to search for prey species that are less abundant during this time. The diurnal peaks probably coincide with time periods when the jaguars prey are most active and the chances of encountering these are highest. In general, p umas in central Belize showed much weaker diurnal and seasonal cycles in movement rates than did jaguars. The movement rate of pumas did not
142 deviate significantly from the average movement rate, either within the 24 hour cycle or across seasons. It could b e that pumas, in co existing with the more dominant jaguar, cannot afford to scale down their patterns of activity to any meaningful extent during the day when jaguars are most inactive. Scaling down activity at night is also not an option since the primar y puma prey is the nocturnal paca. Instead, pumas are forced to remain active across diurnal and seasonal cycles to better exploit the prey base and thus enhance their chances of survival. The size of the puma sample was very small in terms of the number of pumas, however, the movement rates used in these comparisons were based on over 600 steplengths (distance between consecutive location fix). Central Belize is a mosaic landscape where major fluctuations i n resource availability can occur from season t o season, and from year to year. It has been suggested that the territories of males are determined by the spatial distribution of females, whereas the territories of females are determined by the abundance and distribution of food resources (Sandell 1989) Solitary felids like the jaguar, must communicate with conspecifics to regulate their system of land tenure, advertise the social status of males and females, and defend mating opportunities (Gorman and Trowbridge 1989, Sunquist and Sunquist 2002). This communication strategy can take on different forms, such as marking (Kichener 1991, Sunquist and Sunquist 2002) or vocalizing (Peters and Wozencraft 1989, Kitchener 1991, Sunquist and Sunquist 2002). Invariably, however, the effectiveness of these communi cation strategies will depend on frequent travels to all critical areas within any given home range. Markings can deteriorate rapidly, especially in the tropics, and will need to be replenished on a regular
143 basis. Given the large home ranges of jaguars, th ey must utilize a navigational strategy that will allow them to traverse these large areas on a regular basis. In central Belize, the navigational strategy utilized by jaguars can be characterized as many relatively long range movements interspersed with o ccasional short range movements. This strategy is almost the opposite of what is considered optimal for animals foraging in patchy habitats (Schlesinger and Klafter 1986,Viswanathan et al. 1996, 1999) but it is an effective strategy that allow jaguars to t raverse and mark their massive home ranges on a timely and meaningful manner. However, while it allows jaguars to make long range patrols that are necessary to maintain delineation of range boundaries and demarcation of critical areas, it limits the jaguar hunters and suggests a hunting strategy that is probably opportunistic as suggested by Seymour (1989) Therefore, jaguars do not spend a lot of time pursuing any particular prey species but probably take prey opport unistically, as it becomes available along the many long range patrols. Landscape and Habitat Use Jaguars are known to show a preference for forested habitats (Quigley 1987, Crawshaw and Quigley 1991, Farrel 1998, Sunquist and Sunquist 2002), but a lso inhabit a wide range of other habitat types. The high degree of similarity between the results of both the compositional rank orders and the results of the Neu et al. (1974) habitat utilization method for habitat types at second and third order selecti on implies that the results of our habitat preference analyses are robust. Factors operating at one scale (order of selection) are not influencing selection at another scale. At both the second and third order level of selection, and across both the rain a nd dry seasons, jaguars displayed a consistent pattern of preference and avoidance of certain habitat
144 types. While the degree of preference or avoidance was not constant the general pattern of avoidance of agriculture, lowland savanna, and wetland, and pre ference for broadleaf forest and shrub remained constant at both scales of selection and across both the rain and dry seasons. Therefore, in central Belize jaguars showed a measurable preference for forested habitats and avoided open areas. All of the jagu ars in this study avoided the only paved road in the study area. Only one of the home ranges included a small portion of the Western highway while none of the core areas overlapped with this major highway. However, in areas where the preferred forested ha bitat was located adjacent to the paved road, jaguars utilized these habitats as part of established territories. Off roads (unpaved) were fairly common across the entire study area and was one of the main factors contributing to fragmentation in central B elize. For jaguars to exist in central Belize they must adapt to the presence of off roads since it is not possible to place these huge home ranges in this area without intersecting the road network. Conde et al. (2010) found that male jaguars in the Maya n forest of the Yucatan Peninsula did not avoid roads but females did. Dickson et al (2005) suggested that puma movement was not inhibited by dirt roads and that these roads may even promote movement. In our study, jaguar movement rate was 10 times faster when the movement (step length) crossed a road. This would seem to suggest that the increased rate of movement indicates avoidance (jaguars are less comfortable and therefore moving at increased speeds). However, since jaguars must make constant long range movements to patrol and delineate home range boundaries and critical zones, they probably take advantage of the existing road network for these travel routes. While jaguars tolerate this level of fragmentation, it is also clear that when
145 it comes to paved roads with increased vehicular traffic, there is a strong avoidance. Movement rates within different habitats also showed that jaguars moved at rates significantly higher than the average when in broadleaf forest and that they moved at rates significantl y slower than average when in shrub habitat. Dickson et al. (2005) suggested that pumas travel fastest in habitats that they perceive as less profitable and that travel speeds could be used to complement more complex habitat assessments such as composition al analyses. While this suggestion seems logical it does not seem to apply to jaguars in central Belize. Jaguars selected for shrub and for broadleaf forests but while the movement rates in shrub was lower than average, the movement rate in broadleaf fores t was significantly faster than average. The higher movement rate in the broadleaf forest could be explained by the fact that since this is the most abundant habitat type within any given home range then it follows that significant portions of the long ran ge movements will inevitably pass through this habitat type. It is also likely that jaguars prefer moving through this habitat type because of the relative security provided by the forest cover as opposed to moving through more open habitats that would mak e them more v isible Also, a significant portion of the off roads are located within this repeatedly to move around their home range. These travel routes are used repeatedly and they probably make up some kind of mental map (spatial memory). Additionally, using these routes probably minimizes effort and allows jaguars to traverse the established territories most efficiently. Therefore the higher movement rate in broadleaf fore st should not be misinterpreted to suggest any kind of avoidance of this habitat type. Additionally, the significantly higher movement rate when crossing roads, is probably,
146 partially due to jaguars taking advantage of the existing road infrastructure and using this to facilitate the long range patrols. Female jaguars and male pumas favored relatively more rugged terrain. Female jaguars probably select for this feature since it helps them stay away from resident males, especially during times of cub reari ng. While for male pumas, this may be part of a larger strategy to facilitate sympatry with the male jaguars. Perpendicular distance to nearest off roads and perpendicular distance to nearest community were also important variables in the models predicting presence of both male and female jaguars. It is important to note that within fragmented landscapes where the presence of roads is widespread, wide ranging carnivores like jaguars and pumas will be forced to utilize these components of the landscape struc ture if they are to establish home ranges within these human dominated landscapes. Because jaguars and pumas have such large home ranges, and given that the extensive road network permeates the important habitat s then both jaguars and pumas will be forced to utilize the road network. As long as the fragmentation is such that there still remains relatively large patches of suitable habitat, and that these habitat patches are still capable of sustaining viable prey populations, then these large felids will b e able to occupy these landscapes. Their use of these extensive road networks should not be seen as a selection for these features but instead as ecological flexibility and an ability to adapt to these anthropogenic forces. The ability of jaguars and puma s to use these habitats, and to establish home ranges therein, does not also suggest that the individuals within this area are viable and able to reproduce and raise young.
147 Jaguars showed a measurable preference for both forested habitats and the more dist urbed shrublands in central Belize. Maintaining connectivity between the two main conservation blocks will depend on ensuring that sufficient amounts of these two land cover types remain within the central corridor. Management and conservation strategies f or this area must consider the importance of these two habitat types. While jaguars and pumas utilized the same habitats, pumas and female jaguars showed a preference for more rugged landscape features wh en compared with male jaguars. Th e selection of more rugged terrain by pumas and female jaguars appears to be another method used by these two felids to facilitate sympatry in central Belize.
148 CHAPTER 6 CONCLUSION The human dominated landscape in central Belize posits a major challenge, as well as a major opportunity, for the conservation of jaguars and pumas. The habitat and landscape mosaic in this region currently provide the requirements for movement and survival but failure to curtail existing pressures, developmental and otherwise, can chan ge this. Jaguars and pumas have the ecological flexibility to withstand and absorb substantial changes to their natural and preferred habitat, however, it is imperative that conservation efforts identify and delineate habitat requirements and conservation needs for these two wide ranging species in this area. While jaguars and pumas are still able to move and survive in central Belize, it remains unknown how much fragmentation or other human induced changes these felids will be able to withstand. While the window of opportunity for advancing jaguar conservation in this important conservation area may still be open, direct persecution, aggressive deforestation and other land use rred habitat and the natural vulnerability of wide ranging species represent the most obvious and pressing threats. There is an urgent need for immediate and focused action in central Belize. We agree with Quigley and Crawshaw (1992) that the best hope of ensuring the longevity of jaguar (and puma) populations is through the development of site specific conservation strategies. The Western Highway (recently renamed the George Price highway), dissects this central region and is fast becoming a major movement barrier, especially for the highly mobile species like the jaguar and the puma. Two of the nine jaguars collared with functional GPS collars were killed by collisions with vehicles while attempting to
149 cross this highway. Additionally, one male puma and on e male jaguar were shot and killed by hunters along the Sibun river valley. This exceptionally high, human induced mortality is a cause for serious concern. While jaguars and pumas will be able to absorb a certain level of landscape alterations, direct per secution can decimate the local population. Direct persecution and mortality due to vehicular collisions has the potential to prevent this central region from serving the intended purpose of a biological corridor, connecting the larger Selva Maya forest in the north with the Maya Mountain massif in the south. A robust educational campaign on the importance of this region for securing ecological connectivity is needed. Additionally, legislative amendments could potentially help to discourage the targeted and direct persecution of jaguars and pumas. Government and non government organizations must work in unison to develop a strategic, long term plan for this area. The establishment of the Laboring Creek Jaguar Corridor Wildlife Sanctuary, agreements regardin g connectivity between the government and non government organizations and the ongoing research and education efforts in central Belize demonstrates that in spite of the numerous challenges facing the integrity of the Central Belize Biological corridor str ategic gains have been accomplished in the recent past. However, many of the challenges and imminent threats (for example agriculture, community development, road construction among others) are irreversible while any effort to secure and maintain the ecolo gical integrity of this region has to be in perpetuity. The conservation community must continue to make strategic gains in this endeavor and develop long term plans to help guide current and future developments.
150 Studies to further our understanding on the diet, movement ecology, conservation needs and habitat requirements in the human dominated landscape are important. In isolation, very few protected areas in Central America are large enough to make any meaningful contribution to jaguar conservation. Know ledge on how jaguars and pumas move through the unprotected landscape and the dietary flexibility outside protected lands are vital components of any meaningful long term jaguar conservation plan. The importance of the forested habitats in central Belize a nd the urgent need for their proper management cannot be overemphasized, particularly in the light of national long term development projections for this area. Efforts must be made to develop a regional strategy for the larger Selva Maya forests. The now d efunct Jaguars Without Borders (Jaguares Sin Fronteras), a tri national initiative to advance jaguar conservation among the shared forests of Belize, Guatemala and Mexico must be revisited and revived. Belize has the potential to make substantial contr ibutions to the long term viability of jaguars in northern Central America. Curtailing the major threats described herein and ensuring the ecological integrity of the two main conservation blocks will secure the viability of jaguars on a national scale. Tr i national efforts will be necessary to secure jaguars in the larger Selva Maya forests (the largest forested block in Central America), and this in turn will constitute a major component of any effort to secure jaguars across the entire geographic range.
151 APPENDIX A LETTER CODES USED IN THE CART ANALYSIS. CODES BASED ON THE CENTRAL AMERICAN ECOSYSTEM MAP: UNESCO CLASSIFICATION. Cat 01 Cat 03 Cat 05 Cat 06 Cat 09 Habitat Letter codes a a a a a Agriculture b b b b b Brackish/Saline lake c c c c c Caribbean mangrove forest d d d d d Deciduous broad Leaved lowland disturbed shrubland e e e e e Deciduous broad leaved lowland riparian shrubland of the plains f f f f f Evergreen broad leaved lowland shrubland dominated by leguminous shrubs g g g g g Evergreen broad leaved lowland shrubland, Miconia variant h h i h h River i i j i i Seagrass beds j j k j j Short grass savanna with shrubs k k l k k Shart grass savanna with scattered needle leaved trees l l m l l Tropical evergreen seasonal broadleaf lowland forest over lime rich alluvium m m n m m Tropical evergreen seasonal broad leaved lowland forest on poor or sandy soils n n o n n Tropical evergreen seasonal broad leaved lowland forest on calcareous soils o o p o o Tropical evergreen seasonal broad leaved lowland hill forest p p q p p Tropical evergreen seasonal broad leaved lowland swamp forest q q r q q Tropical evergreen seasonal needle leaved lowland forest r s r r Tropical lowland tall herbaceous swamp s t s s Urban r Tropical freshwater reed swamp s Tropical lowland tall herbaceous swamp t Urban h Fish ponds and shrimp farms
152 APPENDIX B LETTER CODES USED IN THE CART ANALYSIS. CODES BASED ON THE CENTRAL AMERICAN ECOSYSTEM MAP: ECOSSYTEM CLASSIFICATION. Cat 02 Cat 04 Cat 07 Cat 08 Habitat (Broad scale classification) Letter codes a a a a Agriculture b b b b Lowland broad leaved moist forest c c c c Lowland pine forest d d d d Lowland savanna e e e e Mangrove and littoral forest f f f f Seagrass g g g g Shrubland h h h h Urban i i i i Water j j j j Wetland
153 LIST OF REFERENCES Ackerman, B. B., F. G. Lindzey, and T. P. Hemker. 1984. Cougar food habits in southern Utah. Journal of Wildlife Management 48:147 155. Aebischer, N. J., P. A. Roberstson and R. E. Kenward. 1993. Compositional analysis of habitat use from animal radio tracking data. Ecology 74 (5):1313 1325. Anderson, C. R., Jr., and F. G. Lindzey. 2000. A guide t o estimating mountain lion age classes. Wyoming Cooperative Fish and Wildlife Research Unit: Laramie, Wyoming, USA. Ashman, D. L., G. C. Christensen, M. L. Hess, G. K. Tsukamoto and M. S. Wichersham. 1983. The mou ntain lion in Nevada. Nevada Department of Wildlife Report W 48 15, Reno, USA. Beyer, H. L. 2012. Geospatial Modelling Environment (Version 0.6.0.0). (software). URL: http://www.spatialecology.com/gme. Bicknevicius, A. R., B. Van Valkenburgh, and J. Walk er. 1996. Incisor size and shape: implications for feeding behaviours in saber Paleontology 16:510 521. Bingham, R. L., and L. A. Brennan. 2004. Comparison of Type 1 error rates for statistical analyses of resource sel ection. Journal of Wildlife Management 68:206 212. Brechin, S. R., and J. Buff. 2005. Exploring human jagu ar conflicts in Belize, Central America, part ll: a nation wide study. Twelve month final progress report to Wildlife Conservation Society, 2005. Broo ks, T. M., S. L. Pimm, and J. O. Oyugi. 1999. Tim e lag between deforestation and bird extinction in tropical forest fra gments. Conservation Biology 13 (5):1140 1 150. Bunyan, J., and K. Kilshaw. 2008. Assessment of popu lation density and structure of mammal assemblages in NE Belize based on cam era trapping studies. School of Biological Sciences. University of Bristol. A report to Belize Forest Department. Caughley, G. 1994. Directions in conservation biology. Journal of Animal Ecology 63:215 244. Cavalcanti S. M. C., and E. M. Gese. 2009. Spatial ecology and social interactions of jaguars (Panthera onca) in the southern Pantanal, Brazil. Journal of Mammalogy 90:935 945.
154 Cherrington, E. A., E. Eck, P. Cho, B. F. Howell, B. E. H ernandez, E. R. Anderson, A. I. Flores, B. C. Garcia, E. Sempris, and D. E. Erwin. 2010. Forest cover and deforestation in Belize: 1980 2010. SERVIR, Cathalac. 42pp. Colwell, R. R., and D. J. Futuyma. 1971. On the measurement of niche breadth and overlap. Ecology 52:567 572. Conde, D. A., F. Colchero, H. Zarza, N. L. Christensen, Jr., J. Sexton, C. Manterola, A. Rivera, C. Chavez, D. Azuara and G. Ceballos. 2010. Sex matters: modelling male and female jaguar habitat for conservation. Biological Conservation 143:1980 1988. Crawshaw, P. G. 1995. Comparative ecology of ocelot ( Felis pardalis ) and jaguar ( Panthera onca ) in a protected subtropical forest in Brazil and Argentina. PhD Dissertation, University of Florida, Gainesville, Fl. Crawshaw, P. G., Jr., J. K. Mahler, C. Indrusiak, S. M. C. Cavalcanti, M. R. P. Leite Pitman, and K. M. Silvius. 2004. Ecology and conservation of the jaguar ( Panthera onca ) in Iguacu National Park, Brazil. Pp. 286 296 in People in nature: wildlife conservation in South and Central America (K. M. Silvius, R. E. Bodmer, and J. M. V. Fragoso, eds.). Colombia University Press, New York. Crawshaw, P. G., and H. B. Quigley. 1991. Jaguar spacing, activity and habitat use in a seasonally flooded environment in Brazil. Journal of Zoology (Lo ndon) 223:357 370. Crooks, K. R. 2002. Relative sensitivities of mammalian carnivores to habitat fragmentation. Conservation Biology 16(2):488 502. Crooks, K. R., C. L. Burdett, D. M. Theobald, C. Rondinin i, and L. Boitani. 2011. Global patterns of fragme ntation and connectivity of mammalian carnivore habitat. Philosophical Transactions of the Royal Society B 366:2642 2651. Crooks, K. R., and M. Sanjayan. 2006. Connectivity Conservation. Cambridge, UK: Cambridge University Press. Currier, M. J. P. 1979. An age estimation technique and some normal blood values for mountain lions ( Felis concolor ). Dissertation, Colorado State University, Fort Collins, USA. Davies, N. B. 1978. Ecological questions about territorial behaviour. Pp. 217 350 in Behavioral ecology : an evolutionary approach (J. R. Krebs and N. B. Davies, eds.). Sinauer Associates, Inc., Publishers, Sunderland, Massachusetts. Davis, M. 2009. Densities, habitat use and mesopredator r elease of the ocelot in Belize. sburg.
155 Davis, M. L., M. J. Kelly, and D. F. Stauffer. 2011. Carnivore co existence and habitat use in the Mountain Pine Ridge Forest Reserve, Belize. Animal Conservation 14(1): 56 65. Dickson, B. G., J. S. Jenness, and P. Beier. 2005. Influence of vegetatio n, topography and roads on cougar movement in southern California. Journal of Wildlife Management 69:264 276. Dixon, K. R., and J. A. Chapman. 1980. Harmonic mean measure of animal activity areas. Ecology 61:1040 1044. De Solla, S. R., R. Boundariansky, a nd R. J. Brooks. 19 99. Eliminating autocorrelation reduces biological relevance of home rang e estimates. Journal of Animal Ecology 68:221 234. Drummond, A. J., B. Ashton, S. Buxton, M. Cheung, A. Cooper, C. Duran, M. Field, J. Heled, M. Kearse, S. Markowit z, R. Moir, S. Stones Havas, S. Sturrock, T. Thierer, and A. Wilson. 2011. Geneious v5.4, Available from http://www.geneious.com/ Eck, E. 2004. Monitoring land use and land cover changes in Belize, 1993 2003: A digital change detection approach. Master The sis. Ohio University. Emmons, L. H. 1987. Comparative feeding ecology of felids in a neotropical rainforest. Behavioral Ecology and Sociobiology 20:271 283. Everatt, K., L. Andresen, and M. Kelly. 2010. Jagua r ( Panthera onca ) abundance and density for the Fireburn Reserve and Balam Na J ungle Estate in northern Belize using photographic capture recapture sampling Report for Forest Department, 23/25 Unity Boulevard, Belmopan. Farrell, L. E. 1999. The ecology of the puma and the jaguar in the Venezuelan llan os. Master Thesis. University of Florida, Gainesville. Figueroa, O. A., R. J. Foster, C. Wultsch, J. B. Mesa Cruz, M. J. Kelly, B. J. Harmsen, W. Sabido, S. Matola and S. K. Jacobson. 2013. Estatus de conservacion del Jaguar en Belice. En Medellin, R., C. Chavez, A. de la Torre, H. Zarza y G. Ceballos (compiladores). El Jaguar en el Siglo XX1: La Perspectiva Continental. Fondo de Cultura Economica. Mexico D. F. Food and Agricultural Organization (FAO). 1997. Stat F AO, Rome, I taly. Food and Agricultural Organization (FAO). 2010. Glob al Forest Resources Assessment 2010. Forestry Paper 163.FAO, Rome, Italy. Food and Agricultural Organization (FAO). 2011. State of Rome, Italy. (available at http://ww w.fao.org/docrep/013/i2000e/i2000e.pdf).
156 Foster, R. 2008. The ecology of jaguars in a h uman influenced landscape. PhD Dissertation, University of Southampton, Southampton, UK. Foster, R. J., and B. J. Harmsen. 2012. A critique of density estimation from camera trap data. Journal of Wildlife Management 76(2):224 236. Foster, R. J., B. J. Harmsen, and C. P. Doncaster. 2010a. Habitat use of sympatric jaguars and pumas across a gradient of human disturbance in Belize. Biotropica 42(6): 724 731. Foster, R. J., B. J. Harmsen, and C. P. Doncaster. 2010 b Sample size effects on diet analysis from scats of jaguars and pumas. Mammalia 74(3):317 321. Foster, R. J., B. J. Harmsen, B. Valdes, C. Pomilla, and C. P. Doncaster. 2010 c Food habits of sympatric jaguars and pumas across a gradient of hu man disturbance. J ournal of Zoology (London) 280: 309 318. Frankham, R., and K. Ralls. 1998. Conservation biology Inbreeding leads to extinction. Nature 392:441 442. Franklin, I. R. 1980. Evolutionary changes in small popul ations. Pp. 135 150 in M. Soule and B. A. Wilcox (eds). Conservation biology: an evolutionary ecological perspective. Sinauer Associates, Sunderland, Massachussetts. Gannon, W. L., R. S. Sikes, and the Animal Care and Use Committee of the American Society of Mammalogists. 2007. Guideli nes of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 88:809 823. Gerber, B. D., S. M. Karpanty, and M. J. Kelly. 2011. Evaluating the potential biases in carnivore capture recapture studies associated w ith the use of lure and varying density estimation techniques using photographi c sampling data of the Malagasy civet. Population Ecology 54:43 54. Gese, E. M., O. J. Rongstad, and W. R. Mytton. 1988. Home range and habitat us e of coyotes in southeastern Colorado. Journal of Wildlife Management 52:640 646. Gorman, M. L., and B. J. Trowbridge. 1989. The role of odor in the social lives of carnivores. Pp. 57 88 in Carnivore behaviour, ecology and evolution (J. L. Gittleman, ed.) Cornell University Press, Ithaca, New York. Harmsen, B. J. 2006. The use of camera traps for estimating abundance and study the ecology of jaguars ( Panthera onca ). PhD Dissertation. University of Southampton, Southampton, UK. Harmsen, B. J., R. J. Fost er, and C. P. Doncaster. 2010 a Heterogeneous capture rates in low density populations and consequences for capture recapture analysis of camera trap data. Population ecology 53(1): 253 259.
157 Harmsen, B. J., R. J. Foster, S. Gutierrez, S. Marin, and C. P. D oncaster. 2010b. Scrape marking behavior of jaguars and puma. Journal of Mammalogy 91(5): 1225 1234. Harmsen, B. J., R. J. Foster, S. C. Silver, L. E. T. Ostro, and C. P. Doncaster. 2011. Jaguar and puma activity patterns in relation to their main prey. Ma mmalian Biology 76 (3):320 324. Harmsen, B. J., R. J. Foster, S. C. Silver, L. E. T. Ostro, and C. P. Doncaster. 2010 c The ecology of jaguars in the Cockscomb Basin Wi ldlife Sanctuary, Belize. In D. MacDonald, editor. The biology and conservation of wild felids. Oxford University Press, Oxford. Harmsen, B. J., R. J. Foster, S. Silver, L. Ostro, and C. P. Doncaster. 2010 d Differential use of trails by forest mammals and the implications for camera trap studies, a case st udy from Belize. Biotropica 42: 126 1 33. Harmsen, B. J., R. J. Foster, S. C. Silver, L. E. T. Ostro, and C. P. Doncaster. 2009. Spatial and temporal interactions of sympatric jaguars (Panthera onca) and pumas (Puma concolor) in a Neotropical forest. Journal of Mammalogy 90:612 620. Henriksen, G. 1997. A scientific examination and critique of minimum viable population size. Faunanorv 18: 33 41. Hooge, P. N., and B. Eichenlaub. 1997. Animal movement extension to ArcView, ver. 1.1. Alaska Biological Science Center, U.S. Geological Survey, Anchor age, AK, USA. Hornocker, M. G., and S. Negri. 2010. Cougar ecology and conservation. The University of Chicago Press, Chicago, Illinois. Iremonger, S., and N. V. L. Brokaw. 1995. Vegetation classification for Belize. In R. Wilson (ed.). Toward a National P rotected Area System Plan for Belize, Synthesis Report. Programme for Belize, Belize City, Belize. 114pp. IUCN 2012. IUCN Red List of Threatened Species. Version 2012.2. . Downloaded on 09 January 2013. Jacksic, F. M., and H. E. Braker 1983. Food niche relationships and guild structur e of diurnal birds of prey: competition versus op portunism. Canadian Journal of Zoology 61:2230 2241. Johnson, D. H. 1980. The comparison of usage versus availability measurements for evaluating resource preference. Ecology 61:65 71. Kelly, M. J. 2003. Jaguar monitoring in the Chiquibul Forest, Belize. Caribbean Geography 13 (1):19 32.
158 Kernohan, B. J., R. A. Gitzen, and J. J. Millspaugh. 2001. Analysis of Animal Space Use and Movement. Pp. 125 166, In: Mill spaugh, J. J., and J. M. Marzluff (eds). 2001. Radio Tracking and Animal Populations. Academic Press, USA, 474 pp. Kitchener, A. 1991. The natural history of the wild cats. Mammal series (E. Neal, ed.). Christopher Helm, London, United Kingdom. Kreeger, T J., 1996. Handbook of wildlife chemical immobilization. International Wildlife Veterinary Services. Laramie, Wyoming, USA. Laundre, J. W., and L. Hernandez. 2010. What we know about Pumas in Latin America. In Hornocker, M. and S. Negri. (eds). Cougar Ecology and Conservation. The University of Chicago Press, Chicago, Illinois. Laundre, J. W., L. Hernandez, D. Streubel, K. Altendorf, and C. L. Gonzalez. 2000. Aging mountain lions using gum line recession. Wildlife S ociety Bulleti n 28:963 966. Leban, F. 1999. Resource Selection for Windows. Version 1.00 (Beta 8.4 May 28, 1999). Levins, R. 1968. Evolution in changing environments. Princeton, NJ. Princeton University Press. Lott, D. F. 1984. Intraspecific variation in the social sy stems of wild vertebrates. Behavior 88:266 325. Maffei, L., A. J. Noss, S. C. Silver, and M. J. Kelly. 2010. Abundance/Density Case Karanth (eds). Camera Traps in Animal Ecology m ethods and analyses. Springer. Meerman, J. C. 2006. Compilation of information on biodiversity in Belize. Final Report to INBio and Belize Forest Department. 116pp. Miller, C. M. 2005.Jaguar density in Gallon Jug Estate, Belize. Unpublished report. Wildli fe Conservation Society, Gallon Jug, Belize. Miller, C. M. 2006.Jaguar density in Fireburn, Belize. Unpublished report. Wildlife Conservation Society, Gallon Jug, Belize. Miller, C. M., and Miller, B. 2005.Jaguar density in La Selva Maya. Unpublished repor t. Wildlife Conservation Society, Gallon Jug, Belize. Ministry of Natural Resources, Local Government and the Environment, Forest Department. Strategic Pan 2005 2010. 44pp.
159 Neu, C. W., C. R. Byers, and J. M. Peek. 1974. A technique for analyses of utilizat ion availability data. Journal of Wildlife Management 38:541 545. Noss, R. F., H. B. Quigley, M. G. Hornocker, T. Me rrill, and P. C. Paquet. 1996. Conservation biology and carnivore conservation in the Rocky Mountains. Conservation Biology 10:949 963. Nov ack, A. J., M. B. Main, M. E. Sunquist, and R. F. Labisky. 2005. Foraging ecology of jaguar ( Panthera onca ) and puma ( Puma concolor ) in hunted and non hunted sites within the Maya Biosphere Reserve, Guatemala. Journal of Zoology (London) 267:167 178. Nowel l, K., and P. Jackson. 1996. Wild Cats: Status Survey and Conservation Action Plan. IUCN/SSC Cat Specialist Group, IUCN, Gland, Switzerland. Nunez, R., B. Miller, and F. Lindzey. 2000. Food habits of jaguars and pumas in Jalisco, Mexico. Journal of Zoology (London) 252:373 379. Nunez, R., B. Miller, and F. Lindzey. 2002. Ecologia del jaguar en la reserva de la Biosfera Chamela Cuixmala, Jalisco, Mexico. In: Medelln, R. A., C. Equihua, C. L. B. Chetkiewicz, P. G. Crawshaw, A. Rabinowitz, K. H. Redford, J. G Robinson, E. W. Sanderson and A. B. Taber (Eds.), El Jaguar en el Nuevo Milenio. Ediciones Cientificas Universitarias, Universidad Nacional Autonoma de Mexico/Wildlife Conservation Society, New York, pp. 107 126. Oliveira, T. G. 1994. Neotropical cats ecology and conservation. Editora da Universidade Federal do Maranhao, Sao Luis. Maranhao, Brasil. Patton, S., A. Rabinowitz, S. Randolph, and S. S. Johnson. 1986. A Coprological survey of parasites of wild Neotropical Felid ae. Journal of Parasitology 72: 517 520. Pianka, E. R. 1973. The structure of lizard communities. Annual Review of Ecology and Systematics 4:53 74. Pike, D. A. 2006. Movement patterns, habitat use, and growth of hatchling tortoises, Gopherus polyphemus. Copeia 2006:68 76. Peters, G. and W. C. Wozencraft. 1989. Acoustic communication by fissiped carnivores. Pp14 56 in Carnivore behaviour, ecology and evolution (J. L. Gittleman, ed.). Cornell University Press, Ithaca, New York. Petracca, C. 2010. Use of site occupancy modeling to deli neate a jaguar corridor in southern Belize. Master Thesis. Duke University. Quigley, H. 1987. Ecology and Conservation of the jaguar in the Pantanal region, Mato Grosso do Sul, Brasil. PhD Dissertation University of Idaho, Moscow, Idaho.
160 Quigley, H. B., a nd P. G. Crawshaw. 1992. A conservation plan for the jaguar Panthera onca in the Pantanal region of Brazil. Biological Conservation 61:149 157. R Development Core Team (2012). R: A language and environment for statistical computing. R foundation for Stati stical Computing, Vienna, Austria. ISBN 3 900051 07 0, URL http://www.R project.org/. Rabinowitz, A. R., and B. G. Nottingham. 1986. Ecology and behavior of the jaguar (Panthera onca) in Belize, Central America. Journal of Zoology 210:149 159. Reid, F. A. 1997. A Field Guide to the mammals of Centra l America and southeast Mexico. Oxford. Oxford University Press. Sandell, M. 1989. The mating tactics and spacing patterns of solitary carnivores. Pp. 164 182 in Carnivore behaviour, ecology and evolution (J. L Gittleman, ed.) Cornell University Press, Ithaca, New York. Sanderson, E. W., K. H. Redford, C. L. B. Chetkiewicz, R. A. Medellin, A. R. Rabinowitz, J. G. Robinson, and A. B. Taber. 2002. Planning to save a species: the jaguar as a mode l. Conservation Biology 16(1): 58 72. Saundry Peter (Contributing Author); CIA (Content Source); Mark McGinley (Topic Editor) "Latin America and Caribbean Population Growth Rates". In: Encyclopedia of Earth. Eds. Cutler J. Cleveland (Washington, D.C.: Environmental Information Coalition, National Council for Science and the Environment). Schaller, G. B., and P. G. Crawshaw. 1980. Movement patterns of jaguar. Biotropica 12:161 168. S cognamillo, D., I. E. Maxit, M. Sunquist, and J. Polisar. 2003. Coexistence of jaguar ( Panthera onca ) and puma ( Puma concolor ) in a mosaic landscape in the Venezuelan Llanos. Journal of Zoology (London) 259:269 279. Seaman, D. E., J. J. Millspaugh, B. J. K ernohan, G. C. Brundige, K. J. Raedeke, and R. A. Gitzen. 1999. Effects of sample size on kernel home range estimators. Journal of Wildlife Management 63:739 747. Seaman, D. E., and R. A. Powell. 1996. An evaluation of the accuracy of kernel density estima tors for home range analysis. Ecology 77:2075 2085. Seymour, K. L. 1989. Panthera onca Mammalian species 340:1 9. Shaw, H. 1987. Mountain lion field guide. 3rd Edition. Special Report Number 9. Arizona Game and Fish Department.
161 Shleinger, M. F., and J. Kl after. 1986. Levy walk versus Levy Flight. In: Stanley, H. E., and N. Ostrowsky. (eds). On growth and form: fractal and non fractal patterns in physics. Martinus Nijhoff, Boston pp 279 283. Silver, S. C., L. E. T. Ostro, L. K. Marsh, L. Maffei, A. J. Noss, M. J. Kelly, R. B. Wallace, H. Gomez, and G. Ayala. 2004. The use of camera traps for estimating jaguar Panthera onca abundance and density using captu re/recapture analysis. Oryx 38: 148 154. Snook, L. K. 2003. Regeneration, growth and sustainability of Ma Yucatan forests. In Lugo, A. E., J. C. Figueroa Colon and M. Alayon. (eds). Big leaf Mahogany: genetics, ecology and management. Springer Verlag, Heidelberg. Sokal, R. R., and F. J. Rohlf. 1995. Biometry: the principle and practice of st atistics in biological research, 3rd ed. W. H. Freeman, New York. Soul, M. E., and J. Terborgh. 1999. Conserving Nature at Regional and Continental Scales: A Scientific Program for Nor th America. BioScience 49(10): 809 817. Spencer, A. R., G. N. Cameron, a nd R. K. Swihart. 1990. Operationally defining home range: temporal independence exhibited by hispid cotton rats. Ecology 71:1817 1822. Sunquist, M., and F. Sunquist. 2002. Wild cats of the wo rld. The University of Chicago Press, London, UK. Swank, W. G. and J. G. Teer. 1989. Status of the Jaguar 1987. Oryx 23:14 21. Terborgh, J., L. Lopez, P. Nunez, M. Rao, G. Shahabudin, G. Orihuela, M. Riveros, R. Ascanio, G. H. Adler, T. D. Lambert and L. Balbas. 2002. Ecological meltdown in predator free forest fr agments. Science 294:1923. Viswanathan, G. M., V. Afanasyev, S. V. Buldyrev, E. J. Murphy, P. A. Prince, and H. E. Stanley. 1996. Levy flight search patterns of wandering albatrosses. Nature 381:413 415. Viswanathan, G. M., S. V. Buldyrev, S. Havlin, M. G. E. da Luz, E. P. Raposo, and H. E. Stanley. 1999. Optimizing the success of random searches. Nature 401:911 914. Waight, I., L. Andresen, and M. J. Kelly. 2010. The impact of sustainable logging on jaguar ( Panthera onca ) densities and trap success rates o f other mid to large wildlife species in the Rio Bravo Conservation and Management Area (RBCMA), Hillbank and La Milpa in northern Belize. Report for Forest Department, 23/25 Unity Boulevard, Belmopan, Belize. Weber, W., and A. Rabinowitz. 1996. A Global Perspective on Large Carnivore Conservati on. Conservation Biology 10(4): 1046 1054.
162 Weckel, M., W. Giuliano, and S. Silver. 2006. Jaguar ( Panthera onca ) feeding ecology: distribution of predator and prey through time and space. Journal of Zoology (London) 2 70: 25 30. Wei, K., Z. Zhang, W. Zhang, X. Xu, X. Liang, G. He, F. Shen, L. Zhang, R. Hou, and B. Yue. 2008. PCR CTPP: a rapid and reliable genotyping technique based on ZFX/ZFY alleles for sex identification of tiger ( Panthera tigris ) and four other endang ered felids. Conservation Genetics 9:225 228. White, G. C., and R. A. Garrot. 1990. Analysis of Wildlife Radio Tracking Data. Academic Press, San Diego, CA, USA. Woodroffe, R. 2000. Predators and people: using human densities to interpret declines of large car nivores. Animal Conservation 3: 165 173. Worton, B. J. 1989. Kernel methods for estimating the utilization distribution in home range studies. Ecology 70:164 168. Worton, B. J. 1995. Using Monte Carlo simulation to evaluate kernel based home range esti mators. Journal of Wildlife Management 59:794 800. Zar, J. H. 1984. Biostatistical Analysis, 2nd ed. Prentice Hall, Englewood cliffs, NJ.
163 BIOGRAPHICAL SKETCH Omar Antonio Figueroa was born in 1969 in San Ignacio Town, Belize, Central North Florida. Immediately thereafter, Omar returned home and for the next five years coordin ated 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 Par tners in Flight Initiative and four years as the national coordinator for national wetland surveys. In August of 2003, Omar was awarded a Fulbright/Organization of American States Ecology Initiative Fellowship to pursue a Master of Science degree. The cent erpiece of his MS research focused on using a Global Positioning System to study habitat associations and movement patterns of the regionally imperiled Jabiru Stork in Mesoamerica. This study represents the first and only attempt to use GPS tracking techno logy to study Jabiru Storks in Mesoamerica. In 2005, he was awarded a Dexter fellowship to pursue a PhD in wildlife ecology and conservation. The centerpiece of his PhD research focused on using GPS to study habitat use and movement patterns of one of the most iconic species in Latin America, the jaguar. In 2009 Omar was nominated by the Prime Minister of Belize to serve as a Senator. His senate term ended in March of 2012. In 2010, Omar was awarded the globally competitive research and Rabinowitz Kaplan Prize for the Next Generation in Wild Cat He received his PhD from the University of Florida in the spring of 2013. Omar is married and has three kids.