LANDSCAPE SCALE RESEARCH AS A TOOL FOR ENGAGING COMMUNITIES IN A SHARED LEARNING PROCESS FOR CONSERVATION AND MANAGEMENT IN THE RUPUNUNI, GUYANA By MATTHEW T. HALLETT 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 2017
2017 Matthew T. Hallett
To my family thank you for teaching me that I could achieve anything that I set my mind to and never letting me believe otherwise. Y ou may not have always understood what I was doing or why, but you always supported me anyway. To my friends near and far I could not have done this without you. Yo u provided me with a constant source of humor and love and formed the most loyal support system that anyone could ask for. To the love of my life you are the greatest gift that the University of Florida could have ever given me. Thank you for being my constant source of confidence, comfort, and inspiration, my statistics coach, and the person who supported me emotionally ( and financially ) through this process. To my advisor thank you for being the only person willing to allow me to pursue this crazy idea and for all of your support along the way To the city that raised me, thank you for making me the man I am today. Chicago, you are a living testament to the importance of diversity. You showed me what it means to care for my fellow man, to invest in people, to listen an d learn from the experience of others, and to serve my community
4 ACKNOWLEDGMENTS I would like to thank Fernando Li Ashley Holland Howard Barnabus Micah Davis, Joan Kenyon, Kenneth Butler, Gerard Pereira, Francisco Gomes, Ramlall Roberts, Cain Edwards, Samantha James, Conrad Peters Kayla DeFreitas, Andrew Albert Martin Carter, Justin DeFreitas Gerry Bell Milner Captain, Duane DeFreitas Jose George, Leroy Ignacio Nicolas Fredericks, Silverio Moses, Flavian Thomas, Jenkins Lawrence, Lisa Orella, Magnus David, Octavius Hendericks, Manuel Mandook, Jessica George Bernard Lawrence, Herman Phillips, Andrew Jackson, Daniel All icock, Gary Sway Rhomayne Li, Alyssa Melville, Judah Kenyon, Nicolas Mandook, Eion Grey, Don Melville Gomer Honorio, Aaron Holden, Courtney Peters, Asaph Wilson, Brian Duncan, Allen & Katie Harley, Meshach Pierre, Telford Roberts, Leon Baird, Martin Robe rts James Honorio, Hubert Gonzales, Marvin Francisco, Arrianne Harris, Maxie Ignace, Arthur King, Felix Holden Maya DeFreitas, Delene Lawrence, Faye Fredericks, Lakeram Haynes, Tommy Kenyon, Anthony Andries, Raquel Thomas Caesar, Thadaigh Baggallay, Anouska Kinahan, and Kyle Hallett for all their assistance, guidance, and support in the field To the Guyana Environmental Protection Agency, Ministry of Indigenous Protected Areas Commission, North Rupununi District Development Board and the Kanuku Mountains Community Representative Group, thank you for making my work in the Rupununi possible I am ever grateful to the leadership and residents of Yupukari, Quata ta, Markanata, Kwatamang, Wowet ta, Rupertee, Surama, Karasabai, Katoka, Shulinab, Meriwau, Quiko, and Rupunau villages, as well as the management and staff of Dadanawa, Karanambu, Saddle Mountain, Manari, and Waikin Ranch es and the Iwokrama River Lodge, Caiman House Research Station, S urama EcoLodge, Rewa EcoLodge and Farfan and Mendes Group for their overwhelming kindness, hospitality, generosity, cooperation and support. I am also thankful to Panthera Guyana the Iwokrama International Centre for Rainforest Conservation and
5 Developm Conservation Society Rupununi Learners Inc., and Karanambu Trust for their support of our work. I would like to give special thanks to John Blake, Andy Noss, Bette Loiselle, Karen Ka iner, Simone Athayde, and Brittany Bankovich, for their guidance and technical support along the way. Lastly, I am ever grateful for funding support provided by the Jacksonville Zoo & Gardens Cleveland Metroparks Zoo, the Ron Magill Conservation Award / Z oo Miami, the Frankfurt Zoological Society, Operation Wallacea, and the Tropical Conservation & Development Program, the School of Na tural Resources and Environment and the Department of Wildlife Ecology & Conservation at the University of Florida
6 TAB LE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ........... 4 LIST OF TABLES ................................ ................................ ................................ .................... 10 LIST OF FIGURES ................................ ................................ ................................ .................. 11 ABSTRACT ................................ ................................ ................................ ............................. 13 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ............. 15 Community Bas ed Conservation ................................ ................................ ........................ 16 Citizen Science ................................ ................................ ................................ ................... 18 Participatory Action Research ................................ ................................ ............................ 20 The Rupununi Wildlife Research Unit ................................ ................................ ................ 20 Project Design ................................ ................................ ................................ ............. 21 Project Scale ................................ ................................ ................................ ............... 23 Consultation and Compensation ................................ ................................ .................. 24 Immersive Experience ................................ ................................ ................................ 24 Indigenous Knowledge ................................ ................................ ................................ 25 Feedbacks and Two Way Learning ................................ ................................ ............. 27 Skills and Leadership Development ................................ ................................ ............. 29 Communication of Re sults ................................ ................................ .......................... 30 Collaboration and Consultation ................................ ................................ ................... 31 Conclusion ................................ ................................ ................................ ......................... 31 2 HABITAT USE, POPULATION DENSITY, AND ACTIVITY PATTERNS OF JAGUARS ( PANTHERA ONCA ) IN A HETEROGENEOUS ENVIRONMENT THE RUPUNUNI REGION OF GUYANA ................................ ................................ ................ 37 Introduction ................................ ................................ ................................ ........................ 37 Scaling Jaguar Ecology and Conservation ................................ ................................ ... 38 Jaguar Habitat Use ................................ ................................ ................................ ...... 39 Methods ................................ ................................ ................................ ............................. 41 Study area ................................ ................................ ................................ ................... 41 Camera trap Survey ................................ ................................ ................................ ..... 43 Statistical Analysis ................................ ................................ ................................ ...... 44 Habitat use ................................ ................................ ................................ ........... 44 Population density ................................ ................................ ................................ 45 Activity patterns ................................ ................................ ................................ ... 46 Results ................................ ................................ ................................ ............................... 46 Habitat Use ................................ ................................ ................................ ................. 47 Population Density ................................ ................................ ................................ ...... 48 Activity Patterns ................................ ................................ ................................ .......... 48
7 Discussion ................................ ................................ ................................ .......................... 49 Use of Savanna Habitat ................................ ................................ ............................... 49 Use of Bush Islands and Gallery Forest Habitat ................................ ........................... 50 Use of Lowland and Montane Forest ................................ ................................ ........... 52 Impact of Microhabitat on Camera trap Surveys ................................ .......................... 53 Presenc e of Hunters ................................ ................................ ................................ ..... 54 Population Density ................................ ................................ ................................ ...... 55 Activity Patterns ................................ ................................ ................................ .......... 57 Habitat as a driver of activity ................................ ................................ ................ 57 Prey as a driver of activity ................................ ................................ .................... 58 Proximity to humans as a driver of activity ................................ ........................... 59 Conclusion ................................ ................................ ................................ ......................... 59 3 OPPORTUNISTIC ENCOUNTERS BY RESOURCE USERS AS A TOOL FILLING KNOWLEDGE GAPS IN THE MANAGEMENT OF A RARE SPECIES: BUSH DOGS ( SPEOTHOS VENATICUS ) IN THE RUPUNUNI, GUYANA ............................... 78 I ntroduction ................................ ................................ ................................ ........................ 78 The Bush Dog A Rare and Elusive Neotropical Canid ................................ .............. 78 Research Tools for Surveying Rare Species ................................ ................................ 80 Local Ecological Knowledge and Research ................................ ................................ 82 Bush Dogs in the Rupununi Region of Guyana ................................ ............................ 84 Methods ................................ ................................ ................................ ............................. 85 Study area ................................ ................................ ................................ ................... 85 Camera trapping ................................ ................................ ................................ .......... 86 Interviews of Natural Historians in t he Rupununi ................................ ........................ 86 Results ................................ ................................ ................................ ............................... 88 Camera trapping ................................ ................................ ................................ .......... 88 Interviews of Natural Historians ................................ ................................ .................. 88 Discussion ................................ ................................ ................................ .......................... 90 Predatory Behavior ................................ ................................ ................................ ...... 90 Reproductive Behavior ................................ ................................ ................................ 91 Habitat Use ................................ ................................ ................................ ................. 92 Perceived and Observed Threats to Bush Dogs in the Rupununi ................................ .. 93 Pet trade ................................ ................................ ................................ ............... 93 Disease transfer from domestic canids and livestock ................................ ............. 95 Direct and indirect impacts of roads ................................ ................................ ..... 96 Loss of habitat ................................ ................................ ................................ ...... 97 Conclusion ................................ ................................ ................................ ......................... 99 4 DEVELOPING A ROBUST METHODOLOGY FOR IDENTIFYING INDIVIDUAL GIANT ANTEATERS ( MYRMECOPHAGA TRIDACTYLA ) WITH IMPLICATIONS FOR POPULATION ESTIMATION ................................ ................................ ............... 106 Introduction ................................ ................................ ................................ ...................... 106 The Giant Anteater ................................ ................................ ................................ .... 106 Feeding ecology ................................ ................................ ................................ 107
8 Habitat use ................................ ................................ ................................ ......... 108 Tra ditional Beliefs and Anteater ID in the Rupununi ................................ ................. 109 Visual Identification of Individual Animals ................................ ............................... 110 Methods ................................ ................................ ................................ ........................... 112 Study area ................................ ................................ ................................ ................. 112 Anteater Surveys ................................ ................................ ................................ ....... 113 Reference photos from AZA institutions ................................ ............................. 113 Opportunistic observations ................................ ................................ ................. 114 Camera trapping ................................ ................................ ................................ 115 Identification of individual anteaters ................................ ................................ .......... 116 Results ................................ ................................ ................................ ............................. 117 Reference Photos from AZA Institutions ................................ ................................ ... 117 Camera trap ima ges ................................ ................................ ................................ ... 118 Photos from opportunistic observations ................................ ................................ ..... 118 Discussion ................................ ................................ ................................ ........................ 119 Selecting the Region of Interest (ROI) ................................ ................................ ....... 119 Identifying Opposite Flanks ................................ ................................ ...................... 120 Pose, Lighting Conditions, and Image Quality ................................ ........................... 121 Data Collection Tools ................................ ................................ ................................ 122 Reference Images of Known Individuals ................................ ................................ ... 122 Camera Trap Surveys ................................ ................................ ................................ 123 Visual Encounter Surveys ................................ ................................ ......................... 125 Conclusion ................................ ................................ ................................ ....................... 127 5 CONCLUSIONS ................................ ................................ ................................ .............. 133 Implementation and results to date ................................ ................................ ................... 134 Applications for Co mmunity driven Management ................................ ............................ 136 APPENDIX A JAGUAR CAPTURE DATA ................................ ................................ ........................... 138 B R CODE FROM CHAPTER 1 STATISTICAL ANALYSIS ................................ ............ 155 C KEY CHARACTERISTICS OF BUSH DOG, CRAB EATING FOX, AND DOMESTIC DOG ................................ ................................ ................................ ............ 163 D SPECIES IDENTIFICATION SLIDES ................................ ................................ ............ 165 E BUSH DOG INTERVIEW DATA SHEET ................................ ................................ ...... 168 F BUSH DOG INTERVIEW DATA ................................ ................................ ................... 170 G IDENTIFICATION OF KNOWN INDIVIDUAL ANTEATERS FROM AZA INSTITUTIONS ................................ ................................ ................................ .............. 181 LIST OF REFERENCES ................................ ................................ ................................ ........ 183
9 BIOGRAP HICAL SKETCH ................................ ................................ ................................ ... 204
10 LIST OF TABLES Table page 2 1 Distribution of camera trap across habitat types ................................ ............................. 60 2 2 Distribution of camera trap sites acro ss trail type and anthropogenic activities ............... 61 2 3 Number of jaguar captures, sites with captures, capture per site, and relative abundance per habitat type ................................ ................................ ............................. 61 2 4 Results of Pearson's product moment correlation test between sampling effort and relative abunda nce ................................ ................................ ................................ ......... 62 2 5 Results of Kruskal Wallis test among habitat and trail types sampled by camera traps ... 62 2 6 Results of Dunn Test for multiple comparisons between habitat and trail types sampled by camera traps ................................ ................................ ................................ 62 2 7 Results of GLM and AIC of association between jaguar relative abundance and habitat type ................................ ................................ ................................ .................... 63 2 8 Number of occurrences and relative abundance of jaguars by sex overall and by sample site ................................ ................................ ................................ ..................... 63 2 9 Results of site level population estimation using spatially explicit capture recapture (SECR) methodology ................................ ................................ ................................ .... 64 2 10 Results of evaluation of overlap (%) in a variety of variables affecting jaguar activity patterns ................................ ................................ ................................ .......................... 64 3 1 Location, year, number of trap stations, number of trap nights, and number of bush dog encounters from previous camera trap studies conducted in Guyana, s tudies conducted by the author in bold ................................ ................................ ................... 100 3 2 Author, location, number of trap nights, number of occasions, and relative abundance (RAI) of bush dogs from previous camera trap studies ................................ ................ 101 3 3 Distribution of opportunistic encounters of bush dogs by key variabl e ......................... 102
11 LIST OF FIGURES Figure page 1 1 Results of priority setting workshop with Guyana Protected Areas Commission ................................ 33 1 2 Strategy for engagement integrating participation, capacity building, and inquiry in the facilitation of participatory research ................................ ................................ ......... 34 1 3 Examples of Rupununi Wildlife Research Unit social media outreach on Facebook, Instagram, YouTube, and Twitter ................................ ................................ .................. 35 1 4 Examples of Rupununi Wildlife Research Unit popular media outreach in tourism and culture magazines, as well as various newspaper outlets ................................ .......... 36 2 1 Map of distribution of camera trap sites across habitats in the Rupununi Region, Guyana ................................ ................................ ................................ .......................... 65 2 2 Examples of camera trap photos of jaguars in the Rupununi Region of Guyana across various habitat and trail types ................................ ................................ .............. 66 2 3 Plot of sampling effort and jaguar relative abundance ................................ .................... 68 2 4 Evaluation of overlap of male and female jaguar activity patterns ................................ .. 69 2 5 Comparison of jaguar activity patterns between Rupununi habitat types ........................ 70 2 6 Evalua tion of overlap of jaguar activity patterns in gallery and riverine forest ............... 71 2 7 Evaluation of overlap of jaguar activity pattern s in lowland and riverine forests ............ 72 2 8 Evaluation of overlap of jaguar activity patterns in lowland forests and bush islands ..... 73 2 9 Evaluation of overlap of jaguar activity patterns in lowland and montane forests ........... 74 2 10 Evaluation of overlap of jaguar activity patterns in in hunted and non hunted areas in lowland forests ................................ ................................ ................................ .............. 75 2 11 Evaluation of overlap of jaguar activity patterns on and off roads within lowland forests ................................ ................................ ................................ ............................ 76 2 12 Evaluation of overlap of jaguar activity patterns in dry creeks and other trail types within gallery forests ................................ ................................ ................................ ..... 77 3 1 Distribution of opportunistic encounters of bush dogs in the Rupununi ........................ 103 3 2 New records of bush dogs in the Rupununi Region from camera traps ......................... 104
12 3 3 Opport unistic photographic records from the Rupununi ................................ ............... 105 4 1 Map of camera trap locations and opportunistic encounters with giant anteaters at Karanambu Ranch, Guyana ................................ ................................ ......................... 128 4 2 Examples of images of giant anteaters from AZA accredited institutions (top), a came ra trap survey of Karanambu Ranch and Yupukari Village (middle), and opportunistic encounters during anteater tours at Karanambu Ranch (bottom) ............. 129 4 3 Example results from queries of known individuals from AZA accred ited institutions 130 4 4 Example results from queries of camera trap images obtained during a survey of Karnambu Ranch and Yupukari Village ................................ ................................ ....... 131 4 5 Example results from queries of photos from opportunistic encounters during an teater tours at Karanambu Ranch ................................ ................................ .............. 132
13 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 LANDSCAPE SCALE RESEARCH AS A TOOL FOR ENGAGING COMMUNITIES IN A SHARED LEARNING PROCESS FOR CONSERVATION AND MANAGEMENT IN THE RUPUNUNI, GUYANA By Matthew T. Hallett December 2017 Chair: John G. Blake Cochair: Bette A. Loiselle Major: Interdisciplinary Ecology T he Rupununi Region of Guyana serves as an ideal laboratory for testing research questions along natural and human use gradients It supports unique habitats that range from cerrado savanna to evergreen and deciduous tropical forest, mammalian f auna representative of the Guiana Shield and low density populations largely composed of indigenous people who maintain tra ditional subsistence lifestyles. New access to technology and global culture, national tural resources, and global climate change are driving rapid change in culture, economics, and climate, creating a need for communities to adapt their way of life. This study seeks to support this process by incorporating research tools and quantitative m ethods into existing traditional management using a participatory approach. Our primary objective is to increase understanding of the status of Rupununi wildlife and identify the most important drivers of their distribution and abundance; however, our pro cess was designed with the intent to equip indigenous communities to deal with resource issues in a changing world.
14 Moving beyond our initial ecological question, we engaged members of indigenous communities and private landowners in a learning process, as sisting in the development and testing of locally relevant research questions that address wildlife and natural resource issues. We provided training and leadership opportunities, and access to decision making in every critical aspect of the development a nd execution. We believe that e ngaging resource users as acti ve participants in research promotes trust, high quality data collection, and a broader under standing of ecological systems and the causal effects of resource exploitation behavior while provid ing the tools that communities need to manage their own resources in a sustainable way. We hope that this project serves as an example of how participatory research can serve as an effective strategy for achieving research and conservation outcomes in area s of high biodiversity.
15 CHAPTER 1 INTRODUCTION Every year, thousands of scientists head to the field seeking to answer a wide variety of questions related to biology, ecology, behavior, and conservation of spec ies, communities, and systems. Enduring images of wildlife scientists sketching undescribed species on the shores of an unexplored archipelago or sitting alone in a remote forest quietly scribbling observations no longer reflect reality. The modern resea rcher works alongside people within integrated systems, seeking to understand interactions between humans and nature and their contributions to mass extinctions and global environmental change. The daunting challenges looming on the horizon have many res earchers striving to make a conservation impact with their work This means pouring effort into not just making sure that their findings cycle back to the places where the data were collected, but facilitating processes to help make that information more accessible and ensure that it is applied to decision making or that it affects behavior in a positive way. At its base, research invests time, training and resources direc tly into communities where the o pportunities for development are rare. Research i s moving towards more participatory methods, creating the potential for scientific research to serve as a tool that promotes and even facilitates conservation ultimately help ing to stem the declin e of global biodiversity that researchers set out to study However, participatory methods require proper thought and careful planning with respect to how to engage communities and build capacity for research that is relevant and locally beneficial. In this chapter, we will explore trends in conservation and sc ience, and discuss how we applied lessons learned from the past to the design and implementation of a participatory research project in the Rupununi Region of Guyana.
16 Community Based Conservation Over the last 30 years, conservation has shifted its focus a nd scale from maximizing protected habitat within the range of wildlife species of interest to incentivizing human communities for the sustainable management wildlife habitat within their local environment. This shift represents the recognition of human b ehavior as the driving force behind environmental issues like habitat loss, climate change, and others (Schultz 2011) and serves as an attempt to stem the drivers of these issues instead of responding to the outcomes. Historically, the conservation and m anagement of wildlife and their habitats has depended largely on rules and enforcement. Top down conservation practices of the past sought to conserve by excluding people from utilizing wildlife and the resources that were set aside for them However, th e capacity of states to enforce laws in wilderness areas and effectively exclude people from resource use in favor of biodiversity conservation is limited, rendering these approaches largely ineffective (Agrawal & Gibson 1999). Centralized management crea tes a mismatch in scale for coupled human environmental systems (Folke et al. 2002), creating a need for bottom up approaches where solutions start at the lowest organizational level possible communities (Berkes 2004). More inclusive, community based ap proaches to conservation are a reaction to the failures of exclusionary approaches (Ghimire & Pimbert 1997) and have shown promise for bridging the gap between people who seek to use resources and those tasked with understanding, managing, and sustaining them Socio economic factors have been major drivers in the decentralization and shift towards increased local participation in conservation. The spread of democratic political structures and increased acknowledgment of human, particularly indigenous pe oples, rights have made centralized efforts that do not represent the needs of people undesirable (Agrawal & Gibson 1999). Community based conservation (CBC) efforts are perceived as more accountable, fair,
17 and legitimate than centralized initiatives (Per sha et al. 2011 ), and the prospect of preserving biodiversity while simultaneously reducing rural poverty has provided an economic incentive for engaging communities directly (Kiss 2004). With the shift to promote sustainable behaviors came the acknowledge ment that conservation is largely a l uxury exclusive to those who are able to meet their daily needs (Adams in the developing world, initiatives that seek to address poverty red uction and environmental protection in an integrated manner seem the most relevant and necessary approach (Adams 1990). As a result, community based conservation initiatives have largely opted for revenue generating activities such as trophy hunting (Chi l d 1996) and eco tourism (Manyar a & Jones 2007), that can increase the standard of living by providing economic incentives for co existing with wildlife rather than allowing communities to actively manage of their own resources within the traditional soci oeconomic paradigm (Gibson & Marks 1995). The challenge that remains with community based conservation based on economic incentives is that while exclusion from resources may have been decentralized, the wall between people and wildlife remains. In this case, interactions with resources are limited to those that meet their demand of foreign funders. This type of commercial natural resource management can be perilous because it establishes an understanding that resources are only useful for their monetary value, leaving them vulnerable to exploitation in the event that economic incentives from sustainable activities decrease or disappear. O verall the results of community based conservation initiatives have been mixed leading to a variety of debates in t he literature over the merits of the concept (Agrawal and Gibson 1999). We believe that failures in community based conservation are not due to the weakness or impracticality of the concept, but rather to its improper implementation (Murphree
18 2002; Berkes 2004). Past examples have shown a tendency for intended beneficiaries to be treated as passive recipients of project activities (Pimbert & Pretty 1995) and for projects to be too short term in nature and over reliant on expatriate expertise (Leach et al. 1999), causing projects to ultimately fall short of their goals. Lessons learned from community based initiatives have suggested that they should pay more attention to issues of equity and empowerment (Berkes 2004), requiring that concerted time and ener gy be invested in areas such as training and capacity building (Mutandwa & Gadzirayi 2007). Citizen S cience Historically, scientific investigation was a largely academic pursuit that was limited to highly trained and affluent members of society who could afford to undertake science as a career. Calls for the democratization of science have persisted for decades (Freire 1972; Kleinman 1998), but research continues to lag behind more applied fields in breaking down its ivory walls, despite the potential fo r participation to bene fit research itself. The field of ecology has also, more recently, undergone a number of important conceptual shifts, from a reductionist to a systems view of the world to include humans as part of natural systems, and from an expe rt based approach to participatory research, conservation and management (Levin 1999; Bradshaw & Beckoff 2001; Ludwig 2001; Berkes 2004) all of which further underline the need to engage people in science Citizen science was among the strategies deve loped to reconcile these shifts a method which engag es a dispersed network of volunteers to assist in professional research developed by or in collaboration with professional researchers (Trumbull et al. 2000) using methodologies for data collection whic h have been specifically designed or adapted to give amateurs a role (Silvertown 2009 ). This use of disperse d participants has been shown to complement and enhance convention al scientific studies in terms of efficiency, cost (Au et al. 2000, Pattengill
19 Se mmens & Semmens 2003) and scale (Cooper et al. 2007; Lee et al. 2006). Some studies have even shown that fieldwork is done better and more efficiently by members of local (especially rural) communities when compared to graduate stu dents or postdocs becau se of the benefit of experience w orking and living in the local environment (Simons 2011). Proliferation of citizen science programs was supported by the notion that involvement would also increase science literacy and environmental awareness among parti cipants (Conrad & Hilchey 2011). While the hope that citizen science harbors educational benefit s for those involved (Cooper et al. 2007), and can foster the type of environmental learning that lead s to behavior change and direct conservation action (Trum ball et al. 2000; Evans et al. 2005), the results of these initiatives have been variable. In fact, broad evaluations of citizen science programs have indicated that participation does not significantly improve understanding of scientific or natural proce sses, and that gains are typically limited to factual and/or procedural knowledge that does not translate well to shifting behavior (Jordan et al. 2011). Critics of citizen science often cite the rigid nature of their engagement as a limiting factor tha t merely employing volunteers to collect their data does not provide sufficient motivation to learn and grow Successful projects actively engage with participants, creating deeper involvement that resulted in increasingly robust learning outcomes (Dickin son et al. 2012). The best examples of citizen science are those that benefit both the participants and th e researchers equally (Silvertown 2009) by facilitating constructive dialogue between citizens and researchers (Brewer 2002). While participation in these programs is largely passive and restricted to prescribed standards for data collection, c itizen science represented the first real ngaging the public in science.
20 Participatory Action R esearch Participatory Action Research (PAR) is a systematic approach to co developing research programs with people, rather than for people, with the specific intent for the results to inform social action (Fals Borda & Rahman 1991; McIntyre 2007). In man y ways, PAR was developed to challenge the knowledge control and rigid roles that are inherent to citizen science, as a mechanism that puts the rhetoric of participation into action by empowering those involved to participate at a high level (Baum et al. 2 006). PAR breaks down structured roles by prohibiting any one participants from being the expert o r sole possessor of informa tion (Feire 1970), with interactions between researchers and participants specifically taking the form of multidirectional sharing of information an d perspectives (Brewer 2002). In PAR, different types of knowledge are recognized as valid, allowing participants to engag e as eq ual participants in a learning process where knowledge and experience can be shared and created in the name o f discovery and creative problem solving. The PAR process includes collaborative investigation and analysis, followed by interpretation, reflection, and action aimed at short term local solutions, as well as long term institutional or cultural change (Cr abtree & Sapp 2005). Recognizing both the active and relational aspects of research new knowledge is continually constructed through critical reflection on current awareness and specific actions (Parkes & Panelli 2001). In its application with community based conservation and management, PAR has been recognized for promoting organization and collaboration strengthening local institutions, and its ability to assist communities in building adaptive capacity (Mapfumo et al. 2013). The Rupununi Wildlife R esearch Unit The Rupununi is widely considered the most biodiverse region of Guyana a country that maintains a wealth of natural resources and >75 % forest cover (GFC 2013) The indigenous
21 Makushi and Wapichan people who inhabit the region have served as its stewar ds for millennia and the wealth of biodiversity and natural resources that persist there today are t he direct result of the ir traditional management systems and continued commitment to sustainable use However, the Rupununi is curren tl y facing dramatic social, e conomic, and env ironmental change, resulting from exposure to technology and global culture, demand from, and access to, global markets, and a changing global climate. T his overarching goal of this project is to equip these commun ities with too ls strategies and str uctures to deal with their rapidly changing world allowing them to continue the legacy of stewards hip set by previous generations Project Design Reaching a level of community development in which communities are independently managin g their resources in a sustainable manner has been elusive. Many indigenous communities had historically developed traditional management practices, which guided the sustainable use of resources at a time when populations were low and demand r arely exten d ed beyond limited regional trade. Rising to meet the challenges of sustainable resource management in a globalized world requires communities to increase their capacity in understanding and monitoring the impacts of contemporary resource use. To meet these new challenges and build the necessary capacity we developed a PAR project that takes a scaffolding ap proach which attempts to capitalize on expertise that may already exist in a community, while building skills and knowledge through collaboration Scaffolding is both a structure for a ctivities and a process by which to carry them out. A scaffolding process enables a learner to solve a problem, carry out a task, or achieve a goal which would be beyond their unassist ed efforts (Wood et al. 1976). The concept of scaffolding is working t owards the development of solutions to complex problems by starting with simpler cases and gradually building up to the full level of complexity (Norman and Spoh r er 1996). By
22 avoiding initial complexity, learners can develop knowledge and confidence while building towards an ultimate goal (Jackson et al. 1996) and has the potential to address some of the issues encountered in the past when attempting to prepare communities to manage resources in a sustainable manner. This project was designed to generate d ata that addressed important local and regional issues through a participatory process that builds capacity in participants and spawns unique inquiries as a result. Our general focus was informed by a previous survey which indicated that a lack of employment and t raining opportunities, conflict with jaguars and pumas, and overhunting of game mammals and fish were the primary issue s facing communities surrounding the Kanuku Mountains Protected Areas (Hallett et al. 2015 a ). A priority setting workshop facilitated by t he Guyana Protected Areas Commission r einforced these findings, a s community leaders identified conflict with jaguars / puma and overhunting of game fish and mammals as the primary natural resource issues that they would like to see addressed (Figure 1 1 ). Guided by an understanding of local issues, we designed a camera trap research project that would generate specific data that could help to address these concerns and inform man agement in the future In doing so, it was also our goal to design a process where Rupununi residents could g ain temporary, part time employment and training in research methods and tools that may be transferable to future projects and employmen t. Whil e chapter two of this dissertation was always intended to focus on the results of this camera trap re search we left the remainder of the project open with the intention of facilitating a process where communities could develop and answer their own locally relevant questions. We have adopted the ladder of citizen participation (Arnstein 1969, Connor 1988 ; Hart 2013) a s a visual representation of our strategy for engagement (Figure 1 2). The ladder
23 provides the idea l symbol for initiatives that emphasize building capacity, as participants are provided with experiences that in crementally increase their skills and knowledge. Our engagement strategy extends beyond the incremental progression of skills among participa nts to include the levels of participation associated with those skills, as well as the level s of complexity t hat are possible to explore when a given level of skills and participation are achieved Our goal of having participants design and implement their own research requires skills and confidenc e that are built through increased participation and while considering increasingly c omplex questions T his process began in chapter two by ensuring that we w ere transparent about the purpose and goals of our work, providing fair compensation to participants and creating opportunities for participants to actively contribute their knowledge to a project that was intentionally selected to capitalize on their expertise (the distribution of species of interest). Over time we provided opportunities for participants to contribute in different ways and engage at higher levels, w hile helping them build the skills to do so. Chapters three and four represent the culmination of this process as we collaborated directly with participants f rom our camera trap project to develop and impleme nt unique questions that were of inte rest to them. Project Scale We explicitly chose to operate this project at the landscape scale because we knew that a broad scale effort would allow for the accumulati on of data suitable for answering our initial ecological questions, while providing a structure where communities could work at the site level to identify and formulate locally relevant questions that they wanted to address or choose to work with other communities who were facing similar issues. The scale of this effort is important because it allowed us to work across scales in answering questions that are relevant to
24 conservation in the region. Creating these vertical and horizontal linkages ha s proven an important address ing the complexity of social and natural systems (Berkes 2004). Consultation and Compensation A basic tenant of any participatory effort should be to gain the informe d consent of the participants and communities that are engaged. In Guyana, informed consent is addressed before Environmental Protection Agency (EPA) require community lea ders to provide signed and dated letters that express their support for proposed research. However, our view is that consultation should go beyond the initial agreements to proceed. We keep communities and community leaders well informed of our actions, progress, and preliminary results through frequent and informal in person and written updates. This consistent consultation and information sharin g maintains transparency in our work and builds confidence in comm unities that we work with Participati on in conservation and research initiatives in the developing world often requires participants to forego subsistence activities that provide for themselves and their families. This is the reality for the subsistence farmers, fishermen, and hunters that s erve as the collaborators and natural historians in this project. Fair compensation has proven an important strategy for releasing participants of this burden (Child 1996), and, as such, we have included fair and prompt payment for services provided a bas ic tenant of our work. Immersive Experience One of the biggest challenges facing the scientific community is de mystifying the process of science and translating the process and results for non scientist citi zens (Brewer 2001). This is especially true in rural communities where understanding of formal scientific and larger scale ecological processes may be limited and represents an important step in community based efforts To develop ecological literacy, participants need to experience the wonder of
25 science in addition to factual information (Orr 1989). Information gained through experience is vital as it provides a requisite contextual base for assimilating informat ion collected through other indirect means (Tuss 1996), thus building skills that would allow communities to gather and apply information towards problem solving. In the experiential model, p articipants progress from action, to understanding the consequenc es of action in a particular context, to generalization in a broader context (Tuss 1996). We bring participants from Rupunun i communities into the scientific process by creating opport unities to set and check camera traps as part of our landscape scale study of drivers of the abundance, distribution, and activity patterns of jaguars across the Rupununi Region. Participants gai n experience through a project that informs pressing issues o f the region, but was designed by the author allowing them to focus on learning from the experience Indigenous Knowledge The support of l ocal field assistants has long been considered an asset to scientific research, particularly in remote rural areas where resources may be difficult to reach and previo us experience may be limited. Although, f or much of the history of science local people have been valued merely either as laborers or for providing insight s that may assist in locating research subjects or specimens. Despite their contribution to keeping scientists alive in rugged fie ld sites and contributing to successful data collection, credit has rarely been given and local knowledge rarely recognized. Recently, indigenous knowledge (I K) has received increasing academic and policy attention and has been applied to wider efforts i n biodiversity conservation, ecosystem assessments, and ecosystem management (Heyd 1995; Huntington 2000; Gadgil et al. 2003; Fabricius et al. 2006; Chalmers & Fabricius 2007; Anad n et al. 2009 ). IK direct interactions with their env ironment and is accumulated through trial an d error, learning
26 from feedback and interaction (Berkes et al. 2003). This type of practical and applied knowledge is accumulated over a lifetime resulting from the harvest or traditional use of species (Gilchri st et al. 2005). Indigenous knowledge contrasts with formal, scientific knowledge, which is the conventional source of information for formalized ecosystem management (Chalmers & Fabricius 2007). Scientific knowledge is precise and measured in an objective and repeatable manner (Moller et al. 2004) passing through strict and agreed upon set s of universally accepted rules (Fabricius et al. 2006). However, IK should not be considered merely the binary opposite of scientific knowledge (Battiste 2002), but rather an alternate subjective cultural bias (similar to the western construction of the scientific process), through which we can find common ground (Aikenhead & Ogawa 2007). The published literature is rich with debate over the merits of IK (Berkes et al. 2000, Huntington 2000), however, b ecause it is typically derived from people who lived with hunted, and trapped wildlife, IK is increasingly Holli ng 1990, Zabel et al. 2002). IK is of course not infallible and does have certain limitations. However, one of the insights from complexity (systems) thinking is that the multiplicity of scales prevents there from being one correct perspective in a complex system. I nstead, p henomena at each level hav e their own emergent properties and, as a result, a number of actors may hold different, but equally valid perspectives (Berkes 2004). In keeping with a conceptual shift towards a systems approach to applied ecology, adding local perspectives to scientific research presents t he opportunity to add depth and complexity to scientific data. Recognizing IK has been one of the primary objectives of this project and one of the first necessary steps to engaging indigenous communities.
27 Engaging community members as stakeholders in a process where they are contributing to a l earning process is critical for the appropriate application of IK Acknowledgment and validation ha ve shown to be vital in the development of a critical skill self directed learning (Garrison 1997). Self directed learners are those that feel confident in their abilities to make observations, ask questions and develop solutions to issues that may arrive in future natural resource management. As participants are exposed to the scientific process through our camera trapping effort we invite and support th e contribution of local and indigenous knowledge to the process. However, t he true benefit of incorporating IK into research is in the co production of new knowledge, in which parties recognize the strengths and weaknesses of different ways of understanding and work together to create a more complete pic ture of reality (Berkes 2009) Chapters three and four of this dissertation are outcomes of the co production of knowledge, but represent only a fraction of the potential that this project has created in the Rupununi reg ion. The contribution of indigenous knowledge to our researc h process is specifically expressed in the visual representation of our engagement strategy in the form of circular arrows that are pulling from sources of information outsid e (IK, western scientific tools and methods ) and within (refl ection) the process. Feedbacks and Two Way Learning Combining IK with science is, however, fraught with ethical, methodological and conceptual difficulties and merging them does not address the problem (Chalmers & Fabricius 2007). What western science has to offer that I K lacks is a broader appreciation of context beyond the local level (Becker & Ghimire 2003). I K offers access to information across scales of time and space that cannot be properly assessed through conventional methods (Wohling 2009), deep knowledge of a particular place resulting from longstanding inhabitation, as well as
28 unique worldviews that are developed through ways of understanding draw from spiritual and other realms outside of consideration by conventional research (Aikenhead et al. 2007). Consequently, I K is rarely integrated into wildlife management decision making, because it is a discipline that historically has relied on quantitative results derived through accepted scientific methods ( Mauro & Hardison 2000), when IK may actually be more relevant beca use specifically incorporates the human experiences, while quantitative studies view human influence as a bias that should be controlled or removed. Increasingly acknowledging that I K may yield biological information relevant to conservation efforts (Huntington 2000), but merely accessing and utilizing I K represents onl y a basic form of application. It is important to recognize that I K and scientific research of information have some common ground: they both rely on direct observation, experience, experimentation, and interpretation (Becker & Ghimire 2003). These commonalities represent a shared understanding that should serve as the starting point for facilitating a collaborative learning process that results in the co production of a knowledge system for sustainable management of ecosystems (Roux et al. 2006). Additionally, IK can serve as a tool that helps to engage local stakeholders by bringing them to the table as part of a team addressing shared conservation concern s an approach generally more productive than scientific studies alone (Za bed et al. 2002). One o f the recognized benefits of PAR is its ability to bring together the perspectives of people who depend on the landscape for their survival and those who make a career from studying it (Colfer et al. 2011). Actively engaging incorporating IK in to the PAR benefits creates the opportunity for knowledge to be co produced. The co production of knowledge through interpretation and reflection is e xplicitly express ed in the visual representation of our e ngagement strategy in the
29 form o f circular arrow on each step of the ladder (Figure 1 2) Participation capacity building, and complexity of the questions tha t we ask can only p rogress following reflection on experience and t he co production of new knowledge. Skills and Leadership Development There is little debate over whether or not incentives for conservation are important they are (Berkes 2004). In the case of community conservation, there has often been a mismatch between what conservationists have thought of as community benefits (i.e. sharing of financial benefits ) and what multiple stakeholders in communities may have thought of as benefits ( empowerment, equity, capacity; Brown 2002). While we have made a commitment to providing fair compensation to those who gi ve up their time to work with us, we have also invested heavily in building capacity (skills, knowledge, confidence) in our p articipants an investment that is already showing rew ards (as evidenced by chapters three and four). Our approach to skill and capacity building is explicitly outlined between the rungs of the ladder in the visual representation of our engagement strategy (Figure 1 2), however opportunities for capacity build ing were provided based on the needs of unique individuals and situations. Remaining flexib le in how we provide these opportunities a llows individuals to progress at their own pace. Skills and k nowledge a re unevenly spr ead within any group, and communities are no exception. Not everyone h as a holistic understanding of the environment and in rural communities; different groups and individuals use different landscapes for different purposes (Kaschula et al. 2005). This leads to the development of local experts or leaders with regard to k nowledge of local wildlife While i t is important to select and work with such local experts rather than arbitrarily selected individuals early on in a project when data collection is the goal (Donov an & Puri 2004), p roviding training and mentoring allows motivated individuals to learn,
30 grow, and emerge as leaders The development of leaders who can fulfill tasks independently was crit ical for our ability to manage the land scape scale research effort presented in ch apter two, but leaders also bec a me an invaluable asset as they developed their own questions in chapters three and fou r, and will continue to support t he application of the results presented here in the future Communication of Results Returning the re sults of research to the communities from which the data was collected is critical in building confidence and perceived va lue in the research process, while creating opportunities for results to be utilized and applied (Chavis et al. 1983). We consider the sharing of results with our project partners to be both part of the consultation and information sharing process that we discussed above, but also a critical part of the learning process P roviding (even preliminary) results of the p rojects that p articipants have contributed to c reates an important o pportun it y to reflect upon and interpret what these results mean for to them selves and their communities. The latter is actively facilitated through informal conversations about what results may mean or how they could be applied and these are the places where new knowledge is most often co produced Information sharing in the Rupununi is typically done through in person presentations conducted i n each community While this approach may be effective and allow for the highest level of engagement, mass communication is known to influence human thought and action by informing, motivating, and guiding people directly, as well as creating feedbacks thorough communities and networks that provide additional incentives and guidance for action (Bandura 2001). Social media has proven an effective and far reaching method for communicating conservation messaging with some audiences (Parsons et al. 2014), while using print and other popular media outlets may be more effective for others (Engels & Jacobson 2007). We share
31 highlights and preliminary results of our work alongside conservation messaging, promotion of sustainable livelihoods, and relevant environmentally themed new s stories on various social media platforms (Facebook, Instagram, YouTube, Twitter; Figure 1 3) and through the popular media in the form of newspapers and magazines (Figure 1 4). These outlets have allowed us to extend our audience beyond just residents of the communities that we directly engage with in the Rupununi, gaining recognition of the value of biodiversity and conservation in the broader Guyana. Collaboration and Consultation Collaboration and con sultation represents th e actions at the top ru ngs of the visual representation of our engagement strategy (Figure 1 2). Engaging with participants at this level requires researchers to relinquish much of the ownership of t he research questio n and data collection to local collaborators who have developed the skills and confidence necessary to design and execute projects that address local issues The research presented in chapters three and four clearly represent these higher level efforts, as t hese questions were developed by local collabo rators Our role in these projects h as become as consultants, helping to fill in specialized skills in fundraising, project design, data analysis, and writing. To be clear, local collaborators will be recognized as co authors of these efforts when the results of tho se chapters go to publication. Conclusion The project presented here represents a snapshot in time. We are currently benefitting from the development of relationships that have been more than seven years in the making, but our participatory research is still in its infancy. We will continue to work with our collaborators in the Rupununi to develop projects that address local issues, and look forward to shifting towards investigating way in which these results will be applied. This project will not conclude
32 w ith the submission of this dissertation, but this dissertation has served as a moment of reflection in what we foresee as a lifelong commitment to a very special place in this world.
33 Figure 1 1. Results of priority setting workshop with Guyana Protecte d Areas Commission ( p hotos courtesy of Matt Hallett )
34 Figure 1 2. Strategy for engagement integrating participation, capacity building, and inquiry in the facilitation of participatory research
35 Figure 1 3. Examples of Rupununi Wildlife Research Unit social media outreach on Facebook, Instagram, YouTube, and Twitter (p hotos courtesy of Matt Hallett )
36 Figure 1 4. Examples of Rupununi Wildlife Research Unit popular media outreach in tourism and culture magazines, as well as various newspaper outlets (p hotos courtesy of Matt Hallett )
37 CHAPTER 2 HABITAT USE POPULATION DENSITY, AND ACTIVITY PATTERNS OF JAGUARS ( PANTHERA ONCA ) IN A HETEROGENEOUS E NVIRONMENT THE RUPUNUNI REGION OF GUYANA Introduction The jaguar ( Panthera onca ) is the largest feli d in the Neotropics, occupying the role of top predator in terrestrial environments, and is also identified as a keystone, indicator flagship, and umbrella species (Terborgh et al. 1999) for a wide variety of habitats, from the southwestern United States to northern Argentina (Seymour 1989). Primary threats to jaguar populations are overhunting of prey, retal iatory killing following depredation of livestock (Caso et al. 2008), and continual loss and fragmentation of habitat that has reduced their range significantly in the last 100 years (Sanderson et al. 1999). the IUCN RedList of Threatened Species; are listed under Appendix I of CITES; receive full protection in Argentina, Brazil, Colombia, French Guiana, Guyana, Honduras, Nicaragua, Panama, Paraguay, Surinam, United States, and Venezuela; and are hunted under restricted conditions in Brazil, Costa Rica, Guatemala, Mexico, and Peru (Caso et al. 2008). Jaguars fare better than any other felid in the genus Panthera in terms of probability of survival, with 78% of their historic range still supporting dispersal (Zeller & Rabinowitz 2010). Blood and scat samples collected range wide showed high gene flow and dispersal ability evidence of landscape scale habitat con nectivity with few barriers (Eizirik et al. 2001; Wultsch et al. 2016). However, the biology, ecology and beh avior of jaguars are not well understood, despite the recent prolife ration of studies that have estimating population density using camera traps a nd capture recapture analysis (Maffei et al. 2011). Reliable data on jaguar population size and distributions are needed, as they form the basis of conservation and management strategies (Soi salo & Cavalcanti 2006).
38 Scaling Jaguar Ecology and Conservati on Historically, the main tool for biodiversity conservation in the Neotropics has been to restrict access and development through the creation of Protected Areas (PAs), most of which have been focused in the largest areas of remaining intact forest (Rodri gues et al. 2004). Certainly, large, well designed and well managed PAs can be effective tools for conservation (La Saout et al. 2013). However, the current distribution of PAs is not representative of the habitat diversity of the Neotropics (Chape et al 2005). Furthermore, their location, design, lack of funding, and poor management reduce their effectiveness (La Saout et al. 2013). For large carnivores, PAs are often not extensive enough to support viable populations (Parks & Harcourt 2002), and high mortality rates at their edges (Woodroffe & Ginsberg 1998) increase extinction probability over time (Brashares 2001). Jaguars range across vast areas at relatively low population densities, earning them the 2002). Individual jaguars roam widely, requiring large areas of habitat populated by sufficient prey for individual survival and mates for reproduction and population survival. As a result, jaguar conservation has moved towards range wide conservation ef forts focused at the landscape scale (Sanderson et al. 2002; Coppolillo et al. 2004; Rabinowitz and Zeller 2010). Jaguar conservationists have designated Jaguar Conservation Units (JCUs) as core areas with either sufficient prey and a self sustaining jagu ar population, or fewer jaguars but adequate habitat and a diverse prey base (Sanderson 2002). Connectivity is maintained by corridors, representing the shortest distance and least cost for dispersal between core jaguar populati on s (Zeller & Rainbowitz 2010). Corridors maintain movement between habitat patches, allowing for persistent exchange of genetic material between populations (Zeller & Rabinowitz 2010). Loss of genetic diversity
39 results in increased incidence of inbreed ing and genetic drift (Young & Clark 2000; Stockwell et al. 2003), as well as reductions in mating ability, male and female fecundity, offspring survival, fitness (Frankham et al. 2002), effective population size (Frankham 1996), and adaptive capacity (Leh man & Perrin 2006), thus threatening populations and species with extinction (Frankham 2005). Corridors may take the form of continuous habitat or h abitat patches which play a en core habitats (Sweanor et al. 2000), both of which provide cover and prey that may dramatically increase fitness and survival of dispersing individuals (So ndgerath & Schroder 2001). Connectivity between core habitat areas and the landscape scale needed for jaguar conservation makes the inclusion of unprotected, heterogeneous, or human dominated landscapes in jaguar conservation efforts critically important. Unprotected areas with low human population density, lack of infrastructure, poor soils (low pote ntial for conversion to agriculture), game species, and proximity to large protected areas have shown the greatest potential for conserving jaguars (Payan et al. 2013). Jaguar Habitat Use In many ways, jaguars are the most emblematic species representing N eotropical forests. However, tropical forests only cover ~50% of tropical South America (FAO 1997) and jaguars readily adapt to the savanna, wetland, scrub, and even desert habitats that cover the remaining area of the Neotropics (Eisenberg 1989). Resear chers have largely assumed that jaguars prefer moist lowland forests in close association with water (Schaller & Crawshaw 1980), however more recent analysis indicates that more heterogeneous environments like the Pantanal, Llanos, and Gran Chaco may actua lly be more suitable of sustaining jaguar populations than contiguous forests such as in Amazonia (Trres et al. 2007).
40 In heterogeneous landscapes, jaguars use a wide variety of habitats, including open savannas and closed forests (Scognamillo et al. 2013 ), pastures, and areas disturbed by human activity (Foster et al. 2010). Preference seems to vary based on site conditions, with individuals favoring forested areas in close proximity to water and prey (Sollman 2011; Davis et al. 2011). Distribution of r esources (forage, fruits, and browse ) for prey can be patchy and predators are forced to hunt in parts of the matrix where prey are most abundant (Rabinowitz & Nottingham 1986). Studies from the Venezuelan Llanos indicate that prey animals are most abundant a long the forest savanna ecotone, and jaguar foraging intensity follows suit (Scognamillo et al. 2003). Heterogeneous landscapes may allow for greater co existence and higher carnivore densities and diversity (Hanski 1994), as patchy distribution of resour ces allows potential competitors to avoid one another across space (Emmons 1987; Aranda & Snchez Cordero 1996) and t ime (Karanth & Sunquist 2000). capacity of heter ogeneous habitats, habitat mosaics, and habitats other than tropical forests to host sustainable jaguar populations is in need of further evaluation. Increased un derstanding of the abundance, distribution and activity patterns of jaguars in these habitat s will inform the prioritization of important jaguar conservation areas, as well as increase the opportunity to engage communities and private lan downers in jaguar conservation. Guyana is a country with an abundance of potential jaguar habitat, as >75% of its total human population densities (GFC 2013). Though Guyana does not contain any designated lity of supporting the long term
41 through central Guyana from the Pakaraima Mountains to the Iwokrama International Centre and into eastern and central Suriname (Z eller & Rabinowitz 2010). The Rupununi Region is tropical forest alongside cerrado savanna, savanna forests, and seasonally flooded wetlands. The Rupununi savannas are conside red part of the Gran Sabana / Rio Branco savanna system of neighboring Venezuela and Brazil an area of immediate conservation concern because of a low probability of long term survival of jaguar populations (Sanderson et al. 2002 ). The Rupununi Region o f Guyana hosts an annually flooded savanna wetland and is bordered by large tracts of lowland and upland tropical forest. Understanding the abundance and distribution of jaguar populations across this heterogeneous landscape represents a first step to ide ntifying key areas for conservation and management, corridors maintaining connectivity with forested areas in Brazil and Venezuela, and the contribution of SW Guyana to range wide jaguar conservation efforts. This study seeks to identify drivers of variat ion in the use, density, and activity of jaguars using camera traps set across a variety of h abitat types in the Rupununi. M ethods S tudy area Rupununi (Region 9), Guyana (Figure 2 1) is named for the river, savannas and wetlands that bear this name, but the region is actually an ancient rift valley, the Takutu Basin, that is bordered by the Iwokrama, Pakaraima and Kanuku mountains ( Crawford et al. 1985 ). The floor of Takutu Basin consists of cerrado savanna, gallery and savanna forests, rivers, creeks and seasonally flooded wetlands bordered by large tracts of lowland and montane tropical deciduous and evergreen forests. The Rupununi savannas Rio Branco savanna system (Montambault & Missa 2002) and tree: grass ratios indicate that the moist savannas of the North Rupununi are most analogous to the cerrado savannas of eastern
42 Brazil (Eden & McGregor 1992). The Rupununi savannas are cont iguous with the Iwokrama International Centre for Rainforest Conservation and Development and Guyana forests to the north, the Pakaraima Mountains and Gran Sabana to the west, and the Kanuku Mountains to the east, which join a vast expanse of intact Guiana Shield forest shared with Brazil, Surinam e and French Guiana ( Mittermeie r et al. 1998 ). Mountainous areas in Guyana have long been revered permanent residents (L. Haynes pers. comm.) As a result, these areas support a wealth of intact ha bitat and corresponding biodiversity. The primary habitat types found in the Rupununi, savanna and moist forest, are principally determined by soil conditions, with savanna habitat occurring where trees cannot take root because a hard underlying clay layer limits penetration of their roots (Montambault & Missa 2002). the reach of seasonal flood waters. Moist forest occurs on porous substrates along the slopes of hil ls and mountains, along rivers and in adjacent low lying areas which receive nutrient runoff (Clarke et al. 2001). The Rupununi savannas are found at elevation s of 120 150 m above sea level ( highest mountain peaks are at 1,067 m, with a num ber of minor p eaks above 900 m; Montambault & Missa 2002). Protected areas in the Rupununi are composed of roughly 99% forest and <1% savanna ( PAC 2015 ). This region of Guyana experiences a single rainy season (May to August), with the heaviest rainfall in May, and a longer dr y season (September to April). A verage annual rainfall is between 1,500 2,000 mm per year and average temperatures are between 25.9C 27.5C ( PAC 2015 ). During the height of rainy season, the main rivers rise by as much as 15 m, flooding low lyi ng forests and/or inundating adjacent savannas.
43 Human population density in the Rupununi is very low. Communities are typically small to medium in size, with few containing > 1,000 residents, for a total of ~10,000 people spread across 46 indigenous communities (Stone 2002). The notable exception is Lethem, a booming town at the Brazilian border. The region also contains a number of privately held cattle ranches and terrestri al mining operations. Rupununi communitie s are made up of predominantly i nine indigenous groups, as well as Chinese Brazilian Afro and Indo Guyanese (Stone 200 2). Makushi people are of Carib descent inhabit villages in the north and central Rupununi, and number ~7,750 in Guyana (NDS 1996). Wapichan people are of Arawak descent; inhabit villages in the south Rupununi, and number ~6,900 in Guyana (NDS 1996). R upununi villages typically range in size from ~120 615 people; subsistence fishing, farming and hunting are the primary means of livelihood (Stone 2002). Camera trap Survey Camera trap photos were obtained as part of a multi species camera trap study of the Rupununi Region following well established methods for camera trap research (Karanth & Nichols, 1998). Camera traps ( Bushnell Trophy Cam #119447C, #119734C, #119736C, and #119837C ; Bushnell KS, USA) were set 2 3 km apart with a single camera at ea ch site set 30 40 cm from the ground in proximity to observed signs of jaguars and their preferred prey. Cameras were active 24 h per day, with a 1 second delay between captures, recording the date and time with each 3 image sequence. Camera traps were set at 357 sample sites across 13 indigenous communities (Yupukari, Quatatta, Markanata, Kwatamang, Wowetta, Rupertee, Surama, Karasabai, Katoka, Shulinab, Meriwau, Quiko, and Rupunau), 5 private ranches (Dadanawa, Karanambu, Saddle Mountain, Manari, and W aikin), and two protected areas (Iwokrama International Centre for Conservation
44 and Development and Kanuku Mountains Protected Area) from May 2014 May 2017 for a total of 62,010 trap nights (Figure 1 1) Camera traps were spread across habitat types in a manner representative of the composition of the Rupununi habitat mosaic (Table 2 1 ). In an effort to avoid bias in the sample and increase understand ing of the impact of variables of interest, camera traps were set across a variety of trail types and a gradient of anthropogenic activity (Table 2 2 ). Images of the species of interest that occurred at the same trap site within a period of 30 minutes were excluded ensure that photo occasions were independent (Silver et al. 2004). Statistical Analysis S tatistical analysis was structured in three tiers to understand the impact of habitat heterogeneity on the (1) distribution, (2) abundance, and (3) circadian rhythms of jaguars in the Rupununi Region. Shifts in distribution were considered to be the most severe response to habitat variables (used rarely or completely absent) followed by abundance (persisting but decreasing), and activity patterns (present in numbers, but shifting behavior). The preceding trends are expressed as habitat use, population de nsity, and activity patterns below. Statistical analyses were performed in R (version 3.2.4; R Core Team 2013) unless otherwise indicated. Habitat u se A relative abundance index (RAI) was calculated by dividing the number of jaguar occurrences by the nu Due to seasonal flooding, theft, equipment malfunction, vegetation growth, and manipulation of cameras by human and animal subjects, trap effort varied among sites. A Coefficient (Pearson 1895) was calculated to measure the strength of the linear relationship between relative abundance and the number of trap nights. A strong linear relationship would
45 require standardization of trap effort at each site and, ultimately, t he loss of jaguar occurrences from the dataset T he relative abundance indices from each trap site were plotted to understand the distribution of the data. A Kruskal Wallis test, a non parametric one way analysis of variance (ANOVA ) that assesses the differences among three or more independently sampled groups and a non normally distributed continuous variable (Kruskal & Wallis 1952), was performed to identify any significant differences in jaguar relative abundance across habitat a nd trail types. hoc pair wise non parametric test of the strength of the relationships between the means of multiple pairs of variables (Dunn 1964; Zar 2010), was performed to identify frequency with which pair s of habitat and trail variables were used We converted the relative abundance data to presence and absence to remove the bias associated with the observed skewed distribution of captures at a few, heavily used sites. A generalized linear model (GLM), a flexible application of linear regression that relates to response variables via a link function, error distributions that are not normally distributed (Nelder & Bak er 1972), was applied to the nai ve occupancy data set to understand the relationship betwe Criterion (AIC; Akaike 1981), an estimator of the relative quality of statistical models, to rank the models that best predicted habitat use by jaguars in t he Rupununi Region of G uyana. Population d ensity Spatially explicit capture recapture (SECR) models utilize density as their population parameter, while detection is represented by a function of declining capture probability as the activity center moves further from each detec tor (Williams et al. 2002) or camera trap in this case By using the locations where each animal is detected to fit a spatial model of detection, SECR models are proven more effective at adjusting for edge effects and incomplete detection,
46 and accommodat ing individual variation in capture probability and home range size (Obbard et al. 2009). Individual jaguars w ere identified by the unique pattern of their rosettes (Silver 2004) and jaguar popul ation density was estimated using a maximum likelihood based spatially explicit capture re capture model in program DENSITY 18.104.22.168 (Efford et al. 2004). Population density at each sample site was compared based on habitat composition (% forest cover) to provide insights on habitat as a driver of jaguar abundance. Activity p atterns Data on the time and date of each encounter recorded during each occasion prov ides an important source of information with regard to the activity patterns of a species of interest. We chose to analyze the temporal overlap of jaguar act ivity patterns as a method for quantifying 3.2.4) to estimate the percentage of overlap in the activity patterns of jaguars at sites with variables of interes parametric measurement of the overlap between the probability distribution functions of these underlying of day as a circular random variable whose distribution may be bimodal an important difference from smoothing techniques like kernel density estimation. The overlap of activity patterns of jaguars were examined with respect to sex, habitat type, trail typ e, and presence of hunters. R esults After removing sites where equipment malfunction and theft led to low sampling effort, we achieved 442 ind ependent cap tures of jaguars from 264 trap sites (Figure 2 2; Appendix A). Jaguar records occurred in six main habitat types (in order of number of jaguar occurrences)
47 lowland forest, gallery forest, bush islands, riverine forest, montane forest, and cerrado savann a. Gallery and riverine forest sites showed the highest number of occurrences/site and RAI, followed by bush islands, lowland forests, montane forest, and cerrado savanna (Table 2 3). A plot of the RAIs at each site showed a non normal distribution of da ta (Figure 2 3) and a 4) indicated no significant correlation between sampling effort and the relative abundance of jaguars. Habitat Use The Kruskal Wallis test (Table 2 5) showed a significant relationsh ip between RAI and both habitat ( p = 0.002) and trail type ( p = 2.45e 07). A p ost comparisons (Table 2 6) showed significant differences between the relative abundance of jaguars at trap sites in savanna and bush islands ( p = 0.03), gallery forest ( p = 0.00), lowland forest ( p = 0.00), montane forest ( p = 0.00), and riverine forest ( p = 0.01). These results suggest that jaguars use savanna habitats differently than all forested habitats surveyed Within these habitat types the of jaguars along game trails compared to foot trails ( p = 0.01), dry creek beds ( p = 0.00), and roads ( p = 0.00), as well as foot trails when compared to roads ( p = 0.01). These results suggest that jaguars use dry creek beds, foot trails, and roads similarly (with a higher relative abundance on roads), while using game trails less frequently. A general linear model (Table 2 7) showed that the most important variables w ith regard to habitat use by jaguars are (in order of least to most important): montane forest, lowland forest, gallery forest, bush islands, elevation, presence of hunters, and cerrado savanna. Riverine forest sites were excluded because of insufficient sample size. The results of our AIC model evaluation (Table 2 7) suggest that Model 3 (AIC = 343.45, df = 6), Model 4 (AIC = 342.75, df = 5), and Model 5 (AIC = 343.09, df = 4) are equally well supported and that the presence of hunters (+
48 correlation), e levation (+), bush island habitat ( ), gallery forest habitat ( ), and cerrado savanna habitat ( ) are the most significant variables for predicting habitat use by jaguars in the Rupununi. R Code for statistical analysis is available in Appendix B. Popu lation Density We identified a total of 74 individual jaguars 39 male and 35 females (Table 2 8). The relative abundance of male jaguars (0.45) was almost double that of females (0.27), although sex ratios in the sample were nearly equal (Table 2 8). P opulation density varied between sites (Table 2 9), with the sites hosting the highest percentages of forest cover also hosting the highest density of jaguars. The Kanuku Mountains Protected Area (KMPA) showed the highest site level population density (5. 58 individuals / 100 km 2 ), followed by Surama Village (4.98), Rewa Village (4.76), Read Head (4.68), Saddle Mountain Ranch (3.56), Karanambu Ranch / Yupukari Village (2.91), Shulinab Village (2.84), and Dadanawa Ranch (2.54), and Karasabai Village (1.96). Activity Patterns Pair wise comparisons of overlap in jaguar activity patterns focused on variation driven by sex, habitat type, trail type, and human activity (Table 2 10). Male and female jaguar activity patterns showed a high degree of overlap (0.9; Figure 2 4). Activity pattern varied across habitat types (Figure 2 5), with the lowest overlap between gallery and riverine forests (0.72; Table 2 10; Figure 2 6), followed by lowland and riverine forests (0.79; Table 2 10; Figure 2 7), lowland forests a nd bush islands (0.83; Table 2 10; Figure 2 8), and lowland and montane forests (0.83; Table 2 10; Figure 2 9). Additional variation in activity patters was observed in hunted and non hunted sites in lowland forests (0.77; Table 2 10; Figure 2 10) on vs. off road sites in lowland forests (0.79; Table 2 10; Figure 2 11), and dry creek beds vs. other trail types in gallery forest (0.79; Table 2 10; Figure 2 12). Results demonstrate biologically significant shifts in behavior
49 that are likely driven by thermo regulation, activity patterns of preferred prey, and human activity. R Code for statistical analysis is available in Appendix B. D iscussion The Rupununi Region of Guyana hosts large areas of intact tropical forest alongside a mosaic of Neotropical savan na, wetlands, and gallery forest. Our research suggests that jaguars use each habitat type available in this matrix that also includes a variety of human activities (hunting, farming, livestock rearing), although they showed a clear prefe rence for foreste d habitats Understanding patterns of use has implications for jaguar conservation in the Rupununi, as the region is faced with a growing demand for its natural resources and rapid changes to its culture, economy, and climate Use of Savanna Habitat Our m odels indicate that savanna habitats supported the lowest relative abundance of any habitat type and a highly negative correlation with jaguar presence. Although the GLM model showed that other habitat types were negatively correlated with jaguar presenc e, the pair wise any of the forested habitats. We predicted that jaguars would prefer forested habitats to those in savanna because the lack of canopy and cove r and the close proximity to humans renders savanna habitat too hot, too exposed, and too dangerous for regular jaguar activity. While there were a low number of captures at savanna sites overall, RAIs were inflated by a high number of captures at a few sites. T he majority of t shared some common attributes in close proximity to water and forest edges in areas known for human jaguar conflict. At each of these sites, the cameras were set along existing foot trails, whic h jaguars were observed using repeatedly as they traveled between the forest and savanna. It is likely that these sites provide access to a single unique resource cattle ( Bos taurus )
50 Jaguars are known to regularly depredate cattle across their range ( Polisar et al. 2003), and it is believed that the access to large prey outweighs the risks, prompting jaguars to move out into open areas. that were free of cattle, suggest ing that jaguars may use savanna areas naturally with some frequency. Further research comparing the frequency of use in savannas with and without cattle would provide important insights into the importance of the presence of cattle in determining s use of savanna habitat. Use of Bush Islands and Gallery Forest Habitat Models indicate that jaguars use gallery forest and bush islands, both forested habitats found within the matrix, similarly. Both habitat types showed high RAIs, with gallery for est exhibiting the highest RAI of any site. However, the GLM model showed that both bush islands and gallery forest had a negative correlation with jaguar presence. These seemingly conflicting results indicate that while the overall percentage of use acr oss these habitat types may be low, there are some hotspots that are highly used and worth further consideration. Bush islands are forest fragments surrounded on all sides by savanna habitat. Generally formed on hills or small mountains, bush islands provide cover, but also serve as dry refuges during the seasonal inundation of the Rupununi savannas. Of the forested habitats represented here, bush islands showed the lowest percentage of trap sites used by jaguars. However, this does not mean that on e should conclude that bush islands are not important for jaguar conservation and management in the Rupununi. Forest fragments are known to play an unfavora an animal, examples of individuals in this study that were observed on several occasions in bush
51 islands before being re captured in highly forested sites indicates t hat bush islands in the Rupununi savannas may serve as stepping stones that facilitate dispersal between forested areas Additionally, while jaguars used bush islands of all sizes and proximities to large areas of intact forests, the majority of th e bush island captures came from the surveys of the two largest bush islands ( Kusad Mountain at Saddle Mountain Ranch and three mile bush at Karanambu Ranch ). These bush islands were orders of magnitude larger than the rest of the forest fragment s, allowing for multiple cameras to be set within a single bush island Comparison to a previous camera trap study at Karanambu Ranch suggests that these large bush islands support permanent, reproductive populations (E. Paemelaere pers. comm.) in additio n to dispersing individuals. Jaguars are known to be frequently associated with water (Schaller & Crawshaw 1980), which would reinforce their use of gallery forest habitat. We defined gallery forest as any narrow band of forest running along a river, cr eek, or oxbow lake, and bordered by savanna habitat on both sides. The percentage of forest cover was highly variable in gallery forest sites, with very dense, closed canopy forest bordering the larger rivers and sparse, open forests bordering creeks. Th e smaller r ivers and creeks that flow through the Rup ununi savannas are generally bordered by a narrow band of it palm ( Mauritia flexuosa ) and may stop flowing completely in the dry season, drying down to isolated pools in low lying areas. While our camera traps captured jaguars using gallery forests of all types, the denser gallery forests that run alongside large rivers showed more frequent use than open forests The latter may still have important con servation implications, as they still represent forested corridors running through undesirable habitat
52 While bush islands and gallery forest may not serve as primary jaguar habitat, they still play an important role in supporting healthy jaguar populations and sh ould be managed accordingly. These forests are increasingly targeted by logging and rotational farming in recent years because of their proximity and accessibility to villages, and have been subject to savanna intrusion following savanna fires (Kellman & Meave 1997 ; Biddulph & Kellman 1998 ). Further research into the impacts of bush island size gallery forest connectivity, and the distance of each from large forested areas may help identify the bush islands and gallery forests that are the most likely to serve as corridors and thus should be prioritized for conservation and management. Use of Lowland and Montane Forest Although l owland forest sites ha d a higher RAI and hosted more captures and more individuals t han montane sites p airwise models did not show a significant dif ference in habitat use bet ween lowland and montane forest s Both l owland and montane sites showed more even and consistent use than other forested habitats, with fewer captures at each site and longer interv als between occasions but a larger proportion of the total number of sites used Forested habitats are clearly important for sustaining jaguar populations, p opulation models showed a positive cor relation betw een percent of lowland and montane forest cover and population density. Although it appears that elevation can be discarded as a factor to consider when pr edicting jaguar habitat use, montane forests had fewer captures of fewer individuals with even longer intervals than lowland sites, and the distribution of the captures that did occur in montane forests w ere skewed towards lower elevations, near the foot of mountains. During the rainy season, many of the lowland habitats are inundated with the flood waters of rivers, creeks, and wetlands. This shift may force jaguars into upland areas during the rainy season, where they may continue to frequent water edges which have shifted to near the foot of mountains. The d ata
53 sho wed that jaguars do use areas up land areas >1000 m, most activi ty appears centered on low elevation areas near seasonal water edges and the occasions that did occur at high eleva tions appeared to be isolated events (only a single capture over the study period). All of the high elevation captures from this study involved seemingly you ng (jud ged by their relative size) male individuals, indicating that high elevations habitat may be useable, but not preferred h abitat. That said, prey species showed equal distribution across all elevation s, with jaguars using lower elevation mountains, particularly those with flat areas at their peaks, with some regularity. Further research into the use of elevation gradients by jaguars and their prey would provide additional insights on the importance of mountainous areas for supporting jaguar populations. A t lowland sites more captures occurred at sites that were closer to forests edges than those further away. This tendency to frequent areas near forest edges has been observed elsewhere, and is believed to be related to the distribution of prey (Scogn amillo et al. 2003). This preference for sites near edges may also have important conservation implications, as most Rupununi communities are also situated near forest edges, which increases the probability of jaguars coming into contact with people and domestic an imals. Edges are known to be among the highest mortality areas in protected areas (Woodroffe & Ginsberg 1998), an issue which may require special management to avoid the transformation o f edges into a population sink. Impact of Microhabitat on Camera trap Surveys Variations in microhabitat around a camera trap site can lead to inaccurate detection rates depending on the individual preferences of the species of interest (Weckel et al. 2006). Jaguars they are available (Harmsen et al. 2010). Setting camera tra ps along lightly trafficked (Blake et al. 2017) roads and trails is a proven method f or increasing capture rates in jaguar studies (Rabinowitz & Nottingham 1986; Maffei et al. 2004; Weckel et al. 2006; Harmsen et al. 2010;
54 Blake & Mosquera 2014) although a focusing trapping effort exclusively on roads is problematic bec ause f emale jaguars have shown to actively a void using roads (Conde et al. 2010). While s electing sites that boost capture rate has long been identified as an advantageous strategy for camera trap studies of large carnivores ( O Connell et al. 201 0 ) t his var iable should be seriously considered in studies seeking to understand preferences in habitat use as it creates the potential for erroneous results suggesting that areas with networks of established roads or trails support higher jaguar populations than road less areas simply because of the increased rate of cap tur e rate along roads (Harmsen et al. 2010). J aguars used human made roads and trails more frequently than off trail areas when available and dry creek beds may serve a similar role when human made infrastructure is not present. While the presence of roads and trails may not attract jaguars to an area on their own, it is clear that the presence of roads bias t he distribution of ( male ) jaguars in the survey area All cameras set along roads in th is study were in a single habitat type lowland forest. Although these sites did show a higher RAI than off road s ites (more captures at shorter intervals), patterns of use (measured by presence / absence) remained consistent. Presence of Hunters Unexpectedly, jaguar habitat use was positively correlated with the presence of hunters at camera trap sites in GLM model. Jaguars sensitivity to human activity seems to vary across their range, avoiding in time and space in some studies (Woodroffe 2000), while frequenting human dominated landscapes in others (Foster et al. 2010, this study). The apparent tolerance of hunting activity observed in this study may have social, cultural, or ecological interpretations. prey t o meet subsistence needs, doing so mainly with traditional bow and arrow. These socio cultural factors mean that overall hunting intensity and disturbance associated with that activity
55 are low. Most camera traps registered very low relative abundances of hunters, with only a few sites experiencing frequent hunting activity. Hunting intensity is seemingly low enough not to alter nat ural jaguar hunting patterns. Additionally, hunters frequent places where they are most likely to encounter prey. Jaguars also are known to structure their movement and activity patterns to increase the probability of encountering their primary prey species (Harmsen et al. 2011). With low intensity hunting in the region, we believe that the positive correlation with the pres ence of hunters may actually be an indication that jaguars prefer to use areas also fre quented by game species. Population Density Site level jaguar p opulation density in the Rupununi Region ranged from 1.96 5.58 individuals / 100 km 2 These density estimates fall within the known range of jaguar population densities from previous studies that employed camera traps and SECR methodology. The lowest population densities came from sites with the highest percentage of savann a habitat with in the study area (Karasabai Dadanawa) The 1.96 and 2.54 individuals / 100 km 2 were similar although slightly higher, than the 0.31 1.82 individuals / 100 km 2 estimated for comparable habitat in Bolivian Gran Ch aco (Noss et al. 2012). The highest population densities came from sites with the highest p ercentage of forest cover ( Rewa Head Re wa Surama KMPA ) T h e 4.68, 4.76, 4.98, and 5.58 individuals / 100 km 2 were similar although also slightly higher, t han the 4.4 0.7 individuals / 100 km2 estimated for Los Amigos Conservation Concession in the southwestern Amazon (Tobler et al. 2013). The sites whose population densities feel between the se highest and lowest densities ( Shulinab, Karanambu / Yupukari, Saddle Mountain ) were sites within the matrix that cont ained m ix of forest types (primarily gallery forest and bush islands) and savanna Population density at these sites ( 2.84, 2.91 and 3.56 individuals / 100 km 2 respectively) fell between the highest and
56 lowest densities observed, seemingly also correlated with the percent forest cover. Saddle Mountain and Karanambu / Yupukari were on the higher end of these middling densities, reiterating the importance of la rge bush islands, which may support both permanent residents and dispersing indivi duals. Interestingly, although the Kanuku Mountains Protected Area showed the highest jaguar population density in the region, land tenure does not appear to be a driving factor of population density, as two indigenous community owned sites had similar and only slightly lower population densities. Within the matrix, ranch sites appear to support higher densities than those managed by indigenous communities; however, these sites also contained greater pe rcent f orest cover than community lands Retaliatory killing of jaguars following depredation of livestock at ranch and community sites may play a role in reducing population density at some sites, as each of the sites with highest d ensity either l acked or had very few livestock Additionally, within each habitat type sites that supplemented their income with ecotourism (Surama, Rewa, Karanambu / Yupuka ri, Sa ddle Mountain, Dadanawa) rather than depending on cattle rearing and extractive activities, also showed higher densities than their counterparts with si milar habitat composition The impact of human jaguar conflict and land use patterns on the population density of jaguars are variable s that warrant further r esearch. Overall, population density seems to be positively correlated with forest cover and t he Rupununi showed relatively high population densities when compared with studies of similar habitat elsewhere in the jaguar s range E xtrapolating even a median population density observed in the matrix across the study area suggests that the Rupununi may support upwards of 600 jaguars. Although this extrapolation is presented with the recognition t hat this is not a n actual population estimate that captures all t he possible variation across the region however e ven
57 as a rough estimate the high number of individuals p ossible seems to j ustif y consideration of the Rupununi Region for a Jaguar Conservation Unit, with the potential for Jaguar Corridors to be identified with further study. Activity Patterns Jaguar activity patterns varied among habitat and trail types, which may provide insights into how jaguars deal with environmental conditions, pursuit of prey, and the presence of humans in the landscape. Male and female jaguars showed a high degree of overlap in activity, which differed from findings in the western Amazon ( Blake et al. 2012). We believe that the heterogeneity of the Rupununi landscape m ay serve as the pr imary dr iver of activity and, as a result, negates any sex related variation that has been observed in more homogenous landscapes Habitat as a driver of activity Among habitat types, those with similar composition a nd forest cover s howed t he high est overlap in activity patterns, while the lowest overlap s w ere observed b etween habitat s with the biggest differences in forest cover Savanna habitat was excluded from these comparisons, as savann a showed different usage patterns from any of the forested habitat types and the majority of the captures occurred at a low number of sites and included only a few individuals, thus increasing the probability of individual bias. However, the comparison of gall ery and riverine forests, which at first consideration would seem to be similar in that they are both forested habitats that border rivers and creeks, showed the lowest degree of overlap of any two habitat types (0.72). Jaguars are known to frequent areas nears water but jaguars in gallery forest habitat were most active before dawn and after dusk, while those in riverine forest s showed activ ity throughout the day with peaks in activity during the morning hours and before d usk. The composition and forest of gallery forests was highly variable, but included a number of sites along small, seasonal creeks that wit h open canopies. Shifts in activity patterns toward more
58 nocturnal activity at these sites may be driven by t he susce ptibility of this species to heat stress (McBri de & McBride 2007) resulting in the need to avoid the hot sun in areas of high exposure. Pr ey as a driver of activity Prey is known as a major factor driving the distribution of jaguars, but studies have shown that they may also shift their activity patterns to match their preferred prey (Harmsen et al. 2011) While the species composition may differ, functionally, lowland and riverine forests are very similar habitat types in terms of f orest structure and cover The primary difference between thes e habitat types is the presence of rivers or creeks that seasonally inundate the riverine forest understory Despite the similarities, a low degree of overlap ( 0.7 9) was observed between lowland and riverine forests which may be attributed to jaguars matching the ir behavior to their prefer red prey in each habit at Jaguars are known generalists, and dietary studies in lowl and forest indicate that a variety of large and medium sized mammals make up the majority of their diet with white lipped ( Tayassu pecari ) and collared ( P ecari tajacu ) peccaries and armadillos ( Dasypus novemcintus ) of ten representing the ir most common prey ( Garla et al. 2001; Gonz lez & Miller 2002 ; Weckel & Giuliano 2006 ; F oster et al. 2010 ) Data from this study showed consistent activity patterns across all times of the da y in lowland forest, providing furth er evidence for a generalist strategy that includes predation on a number of large and m edium sized mammals However, within r iverine forests, ja guars showed shifts in their behavior with dramatic peaks in activity during the morning hours and around dusk. Previous dietary studies conducted in flooded forests have show n th at riverine specialists such as spectacled caiman ( Caiman crocodilus ) and capybara ( Hydrochoerus hydrochaeris ) make up the largest proportion of their prey ( Schaller & Vasconcelos 1978; Da Silveira et al. 2010 ). The shift in activity pattern towards more diur nal activity indicates that jaguars may be matching the activity patterns of
59 caiman ( who bask during the daylig ht hours ) and capybara ( who graze along river ed ges around dusk and dawn ) Proximity to humans as a driver of activity Lastly, w hile jaguars showed a high level of overlap in their activity patterns across g ame trails foot trails and dry creek beds there was a low degree of overlap (0.79) in their activity on roads c ompared to all other trail types. Jaguar activity on game trails, foot trails, and dry creek beds was consistent across all times of the day, with some dips in activity just before dusk and dawn. However, on roads, jaguar activity decreased steadily thro ughout from peaks in the hours before dusk and after dawn, with the lowest activity during the middle of the d ay. The majority of roads in the sample were in lowland forest habitat which indicates that the presence of humans, and potentially the noise assoc iated with vehicles, may driv e these shifts in behavior t owards more nocturnal activity While this study focused exclusively on the effects of habitat on jaguar activity, this is the first indication of t he potential impact of hu man activity a variable which should be studied further. Conclusion While jaguars utilize all habitat types in navigating the mosaic that is the Rupununi Region of Guyana, they showed a clear preference for forested habitats over savanna. The importance of cover for hunting and thermoregulation low abundance of natural prey, and the threat of conflict with humans may limit us e of open savanna habitat The r egion s int act lowland and montane forest s s upport ed the h ighest use and highest jaguar densities i ndicating that thes e areas are the core habitats with the greatest probability of supporting long term survival. The high jaguar population density across region suggests that Rupununi forests warrant consideration as a Jaguar Conservation Unit while further research is needed to identify critical corridors with the broader habitat matrix of the Rupun uni savannas.
60 Table 2 1. Distribution of camera trap across habitat types Site name Trap sites Trap nights # of photos Cerrado savanna Bush island Gallery forest Lowland forest Montane forest Riverine forest Rupunau 17 1,161 91,460 7 4 2 4 0 0 Dadanawa 21 3,095 34,389 3 16 10 2 0 0 KMPA 30 7,789 108,944 0 0 0 11 11 8 Mapari 15 3,477 62,746 0 0 0 6 7 1 Shulinab 22 3,902 72,934 7 1 8 3 5 0 Rewa Head 22 2,937 58,484 0 0 0 11 2 9 Rewa 20 4,352 44,334 4 0 1 7 4 10 Surama 24 5,233 91,701 1 0 0 18 2 3 BHI 22 1,934 41,226 1 3 5 9 4 0 Saddle Mtn. 20 4,478 88,496 6 8 6 0 0 0 Karasabai 20 3,876 61,205 8 2 5 4 1 0 Iwokrama 47 11,499 119,282 0 0 1 42 2 2 Yupukari 34 4,157 41,855 2 8 7 0 0 0 Karanambu 23 4,162 119,554 8 5 3 0 0 0 Manari 20 5 4 13 0 0 0 TOTAL 357 62,052 1,036,610 52 51 61 117 38 33
61 Table 2 2. Distribution of camera trap sites across trail type and anthropogenic activities Site name Trap sites Game trail Creek bed Foot trail Vehicle road Hunted sites Livestock present? Logged sites Near farm? Rupunau 17 9 2 6 0 12 8 0 8 Dadanawa 21 14 6 0 0 18 9 1 0 KMPA 30 30 0 0 0 0 0 0 0 Mapari 15 14 0 0 1 0 5 2 2 Shulinab 22 7 7 8 0 8 18 1 7 Rewa Head 22 16 4 1 0 0 0 0 0 Rewa 20 9 0 10 1 0 12 2 5 Surama 24 5 2 4 13 3 12 14 7 BHI 22 10 0 10 2 4 14 8 8 Saddle Mtn. 20 12 5 3 0 12 10 0 1 Karasabai 20 15 2 3 0 15 13 2 6 Iwokrama 47 23 1 9 14 0 7 27 0 Yupukari 34 29 1 4 0 9 17 10 9 Karanambu 23 14 1 5 3 8 12 4 3 Manari 20 TOTAL 357 207 31 63 34 89 137 71 56 Table 2 3. Number of jaguar captures, sites with captures, capture per site, and relative abundance per habitat type Habitat type Sites w/ captures Jaguar captures Captures/site RAI / 100 nights Savanna 6 19 3.17 0.35 Bush Island 17 82 4.82 0.92 Gallery forest 22 100 4.54 1.25 Riverine forest 13 57 4.38 1.20 Lowland forest 36 141 3.91 0.57 Montane forest 20 43 2.15 0.42
62 Table 2 4. Results of Pearson's product moment correlation test between samplin g effort and relative abundance Variables t value df p value LCI UCI corr Sampling Effort Rel. Abundance 0.84 261 0.4 0.07 0.17 0.05 Table 2 5. Results of Kruskal Wallis test among habitat and trail types sampled by camera traps Variable Chi squared df P value Habitat Type 19.04 5 0.002* Trail Type 33.56 3 2.45e 07* Table 2 6 Results of Dunn Test for multiple comparisons between habitat and trail types sampled by camera traps Comparison Z value Unadjusted p value Adjusted p value Bush Island Gallery 1.32 0.19 1 Bush Island Lowland 1.68 0.09 0.92 Gallery Lowland 0.27 0.78 1 Bush Island Montane 1.02 0.31 1 Gallery Montane 0.18 0.86 1 Lowland Montane 0.44 0.66 1 Bush Island Riverine 0.90 0.37 1 Gallery Riverine 0.21 0.83 1 Lowland Riverine 0.45 0.65 1 Montane Riverine 0.04 0.96 0.96 Bush Island Savanna 2.11 0.03* 0.38 Gallery Savanna 3.44 0.00* 0.01 Lowland Savanna 3.94 0.00* 0.001 Montane Savanna 2.96 0.00* 0.04 Riverine Savanna 2.72 0.01* 0.08 Dry Creek Bed Foot Trail 1.46 0.14 0.29 Dry Creek Bed Game Trail 3.81 0.00* 0.00 Foot Trail Game Trail 2.52 0.01* 0.04 Dry Creek Bed Road 1.21 0.23 0.23 Foot Trail Road 2.55 0.01* 0.04 Game Trail Road 4.55 0.00* 0.00
63 Table 2 7 Results of GLM and AIC of association between jaguar relative abundance and habitat type Model 1 Model 2 Model 3* Model 4* Model 5* Null Intercept 0.54 0.53 0.50 0.62 0.74 0.30 Elevation 0.003 0.003 0.003 0.003 0.003 NA Hunted 0.58 0.58 0.58 0.48 0.50 NA Bush Is. 0.6 0.61 0.64 0.53 NA NA Gallery 0.38 0.39 0.41 NA NA NA Lowland 0.05 0.044 NA NA NA NA Montane 0 NA NA NA NA NA Riverine NA NA NA NA NA NA Savanna 1.94 1.95 1.97 1.83 1.73 NA df 8 7 6 5 4 1 logLik 165.72 165.72 165.73 166.37 167.55 181.36 AIC 347.44 345.44 343.45 342.73 343.09 364.71 delta 4.71 2.71 0.73 0 0.36 21.99 Weight 0.03 0.09 0.23 0.34 0.28 0.00 Table 2 8 Number of occurrences and relative abundance of jaguars by sex overall and by sample s ite Male Female TOTAL Occurrences 276 169 445 Relative Abundance 0.45 0.27 0.72 Total Individuals 39 35 74 Karasabai 2 1 3 Dadanawa 2 2 4 Shulinab 4 4 8 Karanambu/Yupukari 4 4 8 Saddle Mountain 3 2 5 Rewa Head 5 4 9 Rewa 4 3 7 KMPA 4 4 8 Surama 5 4 9 BHI 3 2 5 Rupunau 3 1 4 Mapari 2 3 5
64 Table 2 9. Results of site level population estimation using spatially explicit capture recapture (SECR) methodology Site LL SE LCI UCI Density g0 sigma Ind. / 100 km 2 Karasabai 47.93 0.01 0.006 0.06 0.0196 0.07 2913.52 1.96 Dadanawa 59.41 0.01 0.01 0.07 0.0254 0.04 3831.26 2.54 Shulinab 77.18 0.01 0.01 0.07 0.0284 0.06 3637.11 2.84 154.37 0.01 0.01 0.06 0.0291 0.07 3516.32 2.91 Saddle Mtn. 81.49 0.02 0.01 0.09 0.0356 0.16 1804.61 3.56 Rewa Head 84.53 0.02 0.02 0.09 0.0468 0.06 4012.94 4.68 Rewa 74.73 0.02 0.02 0.1 0.0476 0.02 4954.56 4.76 Surama 175.83 0.02 0.03 0.13 0.0498 0.05 3510.29 4.98 KMPA 103.64 0.02 0.02 0.09 0.0558 0.06 2330.61 5.58 Table 2 10. Results of evaluation of overlap (%) in a variety of variables affecting jaguar activity patterns Comparison Overlap Lower CI Upper CI Male v. female 0.90 0.87 0.98 Lowland v. montane 0.83 0.77 0.96 Lowland v. riverine 0.79 0.74 0.96 Lowland v. bush island 0.83 0.76 0.94 Gallery v. riverine 0.72 0.59 0.84 Hunted v. non hunted 0.77 0.68 0.90 On v. off road 0.79 0.69 0.92 Creek bed v. other 0.79 0.70 0.95
65 Figure 2 1. Map of distribution of camera trap sites across habitats in the Rupununi Region, Guyana (ESRI 2016)
66 Figure 2 2. Examples of camera trap photos of jaguars in the Rupununi Region of Guyana across various habitat and trail types (photos courtesy of Matt Hallett)
67 Figure 2 2 Continued (photos courtesy of Matt Hallett)
68 Figure 2 3. Plot of sampling effort and jaguar relative abundance
69 Figure 2 4. Evaluation of overlap of male and female jaguar activity patterns
70 Figure 2 5. Comparison of jaguar activity patterns between Rupununi habitat types
71 Figure 2 6. Evaluation of overlap of jaguar activity patterns in gallery and riverine forest
72 Figure 2 7. Evaluation of overlap of jaguar activity patterns in lowland and riverine forests
73 Figure 2 8. Evaluation of overlap of jaguar activity patterns in lowland forests and bush islands
74 Figure 2 9. Evaluation of overl ap of jaguar activity patterns in lowland and montane forests
75 Figure 2 10. Evaluation of overlap of jaguar activity patterns in in hunted and non hunted areas in lowland forests
76 Figure 2 11. Evaluation of overlap of jaguar activity patterns on a nd off roads within lowland forests
77 Figure 2 12. Evaluation of overlap of jaguar activity patterns in dry creeks and other trail types within gallery forests
78 CHAPTER 3 OPPORTUNISTIC ENCOUNTERS BY RESOURCE USERS AS A TOOL FILLING KNOWLEDGE GAPS IN THE MANAGEMENT OF A RARE SPECIES: BUSH DOGS ( SPEOTHOS VENATICUS ) IN THE RUPUNUNI, GUYANA I ntroduction Historically, wildlife management planning and decision making has been built on population estimation and monitoring a model that relies heavily on time and resource intensive quantitative data collection through rigorous, structured scientific studies (Mauro and Hardison 2000). Well planned and executed research that takes place across broad spatial and temporal scales, inclu des precise measurements of population size, and comparisons to control sites for variables of interest is undoubtedly an effective method for accurately identifying the degree and drivers of population change (Moller et al. 2004). However, constraints to time and resources, as we ll as the ability of scientists to properly characterize complex human interactions with the environment limits our capacity to conduct such studies in the field a fact that becomes particularly apparent with regard to naturally rare, wide ranging species that inhabit remote or hard to access locations. The Bush Dog A Rare and Elusive Neotropical Canid The bush dog ( Speothos venaticus ) is a naturally rare and extremely elusive canid that ranges widely across South America. T his species is highly social, hunting in groups of 2 1 0 individuals (Aquino & Puertas 1997; Barnett et al. 2001; Beisiegel & Zuercher 2005). Little is known about the movements and habits of bush dogs but the y reportedly are mostly diurnal and semi nomad ic over extensive home ranges (>100 km 2 ). They also use a variety of habita ts including intact forest, well vegetated cerrado savanna, and disturbed agricultural and ranchland sites associated with forest fragments (de Souza Lima et al. 2012).
79 The wide ranging nature of bush dogs is associated with their predatory behavior (de Souza Lima et al. 2012). They are strongly carnivorous, favoring small mammals up to their own weight, including Cuniculus paca Dasyprocta spp Myoprocta spp., Dasypus spp an d various species of mice, rats, opossums, terrestrial reptil es, and ground birds (Eisenberg 1989; Emmons 1997; Hunter 2011). Bush dogs have also been documented pursuing larger prey, such as Hydrochaeris hydrochaeris Pecari tajacu and Mazama spp. (Hunt er 2011), and consuming terrestrial invertebrates and fruits in small amounts or inciden tally (Zuercher et al. 2005). Bush dog population s are estimated to have declined 20 25% from 1996 2008 (DeMatteo et al. 2011), but actual abundance and distribution are unknown. Major threats include habitat loss related to extractive activities, reduction of prey base by unsustainable hunting, and lethal diseases contracted from domestic canids (DeMatteo 2008). The degree to which each threat impacts bush dog popul ations is unknown because of the difficulty of studying this species in the under Appendix I of CITES (DeMatteo et al. 2011). Current knowledge of bush dogs comes f rom museum specimens (Engstrom & Lim, 2002; DeMatteo & Loiselle 2008; De Oliveira 2009), interviews with practitioners (DeMatteo 2008; DeMatteo & Loiselle 2008), indirect evidence in the form of tracks and scat (DeMatteo et al. 2004; Zuercher et al. 2005; Lima et al. 2009; DeMatteo et al. 2009; Pickles et al. 2011), opportunistic encounters (Peres 1991; Barnett e t al. 2001; DeMatteo & Loiselle 2008; Beisiegel 2009; De Oliveira 2009; Gil & Lobo 2012; Carreter o Pinzn 2013), camera trap captures (Beisiegel 20 09; Michalski 2010; Fusco Costa & Ingberman 2012), and studies of captive individuals (Kleiman 1972; Porto n et al. 1987; Montalli & Kelly 1989; MacDonald 199 6; DeMatteo et al. 2006). F ew field s tudies have been robust enough to detail bush dog diet, home
80 range, and behavior (Beisiegel & Ades 2002; Beisiegel & Zuercher 2005; de Souza Lima et al. 2012; DeMatteo et al. 2014). As a result, the abundance, distribution, and ecological requirements of this species are poorly understood, hampering conservation an d management efforts (Eisenberg 1989; Emmons 1997; Hunter 2011). Research Tools for Surveying Rare Species Field studies on bush dogs have thus far been limited by the ability of researchers to effectively locate packs in the wild. Both active and passi ve capture techniques have been suggested as potential solutions for filling knowledge gaps on this species, and although each has contributed some data, returns have been in sufficient to guide management. Physical capture and deployment of radio or GPS co llars provides the most comprehensive data on distribution, activity, and home range of species of interest. These tools are inherently expensive and bush dogs have proven exceedingly difficult to locate and physically capture, thus limiting application. Two studies that successfully captured and tracked bush dog packs produced estimates of home ranges that varied from 140 (Lima et al. 2012) to 709 km 2 (De Matteo et al. 2014). While authors hypothesized that differing amounts of natural habitat between study sites may account for the variation in home range size (de Matteo et al. 2014), more replicates are needed to increase the applicability of these results. The potential for a dramatic increase in the number of similar studies remains constrained by costs and the rare and el usive nature of this species. Domesticated detection dogs have been suggested as an effective technique for locating physical evidence (scat, hair) of bush dogs, with the genetic information in each sample providing important dat a on the structure and health of populations that goes beyond simply documenting their occurrence in a given area (De Matteo et al. 2009). The olfactory ability of detection dogs goes far beyond what human observers can sense in their environment and it ha s
81 been suggested that surveys with detection dogs are more cost and time effective than other met hods (De Matteo et al. 2009), but this method also requires access to specially trained canines, which limits its appl ication (Oliveira et al. 2016). Camera tra ps are considered an effective tool for surveying medium and large terrestrial rainforest mammals (Tobler et al. 2008), and a consistent and reliable means for estimating the relative abundance of cryptic species (Carbone et al. 2001) and of bush dogs in particular (de However, a number of robust camera trap studies conducted in areas with high quality habitat failed to detect bush dogs (when the expectation w as that they would be pre sent), including 14 of the previous 15 camera trap projects conducted in Guyana (Table 3 1). Some suggest that the lack of captures in these studies may have resulted from insufficient sample size or ineffective trap placement, because many studies in the Neotropics are focused on jaguars (Tobler et al. 2008). However, even studies that focused their study design specifically on bush dogs and/or those with high sampling effort have resulted in few captures (Table 3 2; de Oliveira et al. 2016). Camera tra p studies measuring occupancy or relative abundance assume a consistent detection probability across time and space and a positive linear relationship between capture rates and true abundance f ew to apply these analyses. Bush dogs are naturally rare and semi nomadic across large home ranges. They also tend to avoid roads and foot and game trails frequented by other species, making camera trapping alone too inconsistent for gaining a reasonab le understanding of the abundance and distribution of this species (Fusco Costa & Ingberman, 2012). We present the findings of a multi species landscape scale camera trap survey of the Rupununi Region of Guyana here and discuss the limitations of resultin g data for informing management.
82 Local Ecological Knowledge and Research Gaps in knowledge are recognized as a central factor limiting the effectiveness of bush dog conservation and management, and traditional qualitative research tools (i.e. transects, camera traps, collars) have shown limited success in closing these gaps. To address this issue local ecological knowledge, the dynamic and ever changing information and experiences embedded in the practices, institutions, relationships, and rituals of people and communities that are developed in a given place over time (Warburton & Martin 1999), should be integrated with limited data produced by western methods. R eaching a cross disciplines to broaden data sources and methods, may provid e the innovation needed to produce critical information that has, to date, been elusive but that is necessary for effective management of this species. The ecological LEK ) is used here as it most accurately characterizes those wh o contributed to our study a group that includes indigenous and non indigenous people whose knowledge may encompass generational or contemporary experience in our study area. LEK is derived from a lifetime of experience of those who depend on the local environment for their survival (hunting, fishing, and farming) and subsistence (wildlife trapping, commercial harvest). LEK differs from i ndigenous knowledge (IK) which is a body of knowledge that is aggregated over time by a given cultural group in thei r long term adaptation to a given biophysica l environment (Purcell 1998). Distinctions between IK and LEK may be important for incorporating IK into research, as observations and reflections made by indigenous people who maintain traditional belief system s may be based in a spiritual worldview that recognizes connections between organisms and the natural environment that are not recognized in the scientific literature (Pierotti & Wildcat 2000). Incorporating LEK into the scientific process has shown cons iderable potential for improving research and natural resource management (Baird et al. 2005), particularly when
83 focused on species and their d istribution (Moller et al. 2004; Kaschula et al. 2005). LEK may be particularly important to conservation and ma nagement when the species of interest is rare, elusive and /or have populations that occur in remote locations where extensive scientific studies may be impractical (Barsh 1997; F erguson et al. 1998). Although LEK may be imprecise and qualitative, the spa tial and temporal scales at which LEK is collected may be more consistent with the biology, ecology, and behavior of semi nomadic species than traditional scientific studies. The capability of rigorously vetted surveys of species observations by local nat ural historians to produce accurate and reliable data should result in LEK being considered analogous (Zabel et al. 2002 ; DeMatteo and Loiselle 2008). However, d espite its potential to contribute to research and management, LEK is often viewed with heavy skepticism or outright dismissed by academics and practitioners due to the lack of structure in data collection and lack of formal training among data collect ors (Gilchrist et al. 2005). By collaborating with local natural historians western scientists have boosted research productivity and efficiency, cut investigative c osts, and can produce large enough sample sizes for hypothesis testing, all while su pporting conservation and building capaci ty in local communities ( Moller et al. 2004; Simons 2011). Inviting the par ticipation of natural historians in to the investigative process es may provide the additional be nefit of addressing disconnect s between rese arch and implementation (Toomey 2016) Although this requires skillful facilitation, clear objectives from the outset, and a process that is underpinned by a philosophy that emphasizes empowerment, equity, trust and learning (Reed 2008), inviting resource users into the planning and investigative process only increases the chance that research findings will be applied locally in daily decision making, and build momentum towards shaping policy
84 Bush Dogs in the Rupununi Region of Guyana Bush dogs are an understudied canid of conservation concern. Their remote habitats, elusive and wide ranging nature and naturally low population densities ha ve made it difficult to understand and effectively manage this species. Guyana hosts an abundance of suitable bush dog habitat, but is one of only two countries (Suriname is the other) for which the status of this species has not been assessed (DeMatteo et al. 2011). Bush dogs were recently listed as gulations (EPA, 2009), but this status was adopted from an out of date IUCN RedList st atus and needs to be updated. Bush dogs are known in Guyana from opportunistic encounters at Kaieteur National Park, the Iwokrama International Centre for Rainforest Co nservation and Development, Timberhead (Barnett et al. 2001), Mabura Hill (ter Steege et al. 1996), and the Berbice an d Corentyne Districts (Quelch 1901). Indirect evidence was reported in the Rewa Head (Pickles et al. 2011), while local experts confirmed the presence of this species during rapid assessments of the Kanuku Mountains Protected Area (Parker et al. 1993; Montamba ult & Missa 2002) and Konashen Community owned Conservation Area (Alonso et al. 2008). Two wild caught specimens from the Rupununi Region are held as museum specimens (Barn ett et al. 2001; Engstrom & Lim 2002), and bush dogs were documented by camera traps in the Kanuku Mountains Protected Area (Hallett et al. 2017) In the Rupununi Region of Southwestern Guyana, bush dogs are known as Ai (Makushi), Wechawar (Wapishana), Savanna Dog or Short tail dog (Creole; NRDDB 2000). They are highly respected by indigenous people for their hunting skills and there is a long history of interaction s wi th the people of the region. They are not targeted for hunting or trapping (NRDDB 2000), nor are they considered a high conflict species (Hallett et al. 2015 b ), despite a reputation for depreda ting chickens elsewhere (Hunter 2011). With large tracts of intact fo rest
85 and savanna habitat, abundant prey, lack of direct hunting/trapping pressure, and little conflict with human communities, the Rupununi Region of Guyana should host healthy populations of this species. Lack of documentation of bush dogs indicates that either, (a) population densities are low; (b) research t ime and effort has been insufficient; or (c) new methods are needed to increase species detection and sample size While previous studies of bush dogs have incorporated surveys of conservation and la w enforcement pro fessionals (DeMatteo & Loiselle 2008; Fusco Costa & Ingberman 2012), reports of opportunistic encounters from natural historians represent a wealth of data that has yet to be employed in the study of this species This study documented 95 opportunistic encounters with bush dogs from across the Rupununi Region. These new records dramatically increase our understanding of bush dogs in Guyana, and represent an effective method that could be applied elsewhere for studying this, and other rare elusive, and wide ranging data deficient speci es. M ethods S tudy area The Rupununi (Region 9), Guyana (Figure 3 1), is an ancient rift valley (Crawford et al. 1985), that supports a habitat mosaic of cerrado savanna, gallery and savanna forests, rivers, creeks, and seasonally flooded wetlands, bordered by large, undeveloped tracts of lowland and montane tropical deciduous and everg reen forests. While the forests of northern Guyana host a relatively high density of logging and mining concessions and the ad jacent Rio Branco savanna to the west in Brazil has been largely degraded by large scale agriculture a nd urban development, Rupununi forests remain unbroken, in pristine condition, and extend virtually uninhabited to the east and south towards Suriname and Brazil. Human settlements in the Rupununi consist of small medium sized villages (100 700 people) which contain ~20,000 people spread over 40+ villages (Bureau of Statistics, 2012). Communities are composed of
86 predominantly indigenous Makushi and Wapish ana people, the majority of whom maintain traditional sub sistence lifestyles. Camera t rapping Camera trap photos were obtained as part of a multi species camera trap study of the Rupununi Region following well established methods for camera t rap research (Karanth & Nichols 1998). Camera traps ( Bushnell Trophy Cam #119447C, #119734C, #119736C, and #119837C ; Bushnell KS, USA) were set 2 3 km apart and 30 40 cm from the ground in proximity to observed signs of jaguars ( Panthera onca ) and their preferred prey. Cameras were active 24 h per day, with a 1 second delay between captures, recording the date and time with each 3 image sequence. We sampled 372 trap sites from May 2014 May 2017 for a total of 62,010 trap nights. Study design was informed by hab itat, proximity to resources, and trail type (Table 2 1 ; 2 2 ). Relative abundance index (RAI) was calculated by dividing the number of independent captures by the total number of traps night (then standardized for 100 trap nights). Interviews of Natural Historians in the Rupununi Rupununi residents who contributed observations of bush dogs were identified a cross disciplinary and multi scaled term describing those who observe species and describe their interactions with each oth er and the environment (Tewksbury et al. 2014). Our use of this term is meant to convey respect for local ecological knowledge and to those who have earned and knowingly contributed it to this project. The term also accurately represe nts our application focused objectives, as natural history has played a critical role in the conservation and responsible manage ment of species and places of interest, while serving as an important means for connecting science and soci ety (Tewksbury et al. 2014). Surve y sites and natural historians (NHs) were identified through a snowball samplin g approach (Biernacki & Waldorf 1981). We visited villages where bush dog sightings were
87 locally known and identified individuals who were respected for their s ubsistence hunting, fishing, or farming expertise. Those who confirmed first hand encounters were asked to provide informed consent to share their knowledge, and to share any other known locations of bush dog encounters and the names of potential contribut ors Interviews were conducted in each village until known leads were exhausted. Natural historians were screened based on their abilities to differentiate bush dogs from other canids found in the region (Appendix C) Each was asked to identify the ani mal in question by name and to describe its physical characteristics. Encounters were only considered valid if the animal(s) in question were described as having some variation of the following attributes: small canid with pale brown, tawny yellow or dar k brown coat, characteristic stocky build and combination of broad (bear like) head, small eyes, short muzzle, short rounded ears, short bushy tail, short legs, and/or relatively long body (Eisenberg 1989; Emmons 1997; Hunter 2011). The short tail is cons idered the key attribute for differentiating this species from other regional canids Canis familiaris and Cerdocyon thous Information was only included from NHs who were able to successfully differentiate bush dogs from biological sketches and photogra phs of similar species (Appendix D). Once accounts were verified, opportunistic encounters were documented through semi stru ctured interviews (SSI; Dickman 2008). NH s were asked to provide information on when (date, season, time of day) and where (geograp hic or relative location) the encounter occurred and what they observed (how many animal s, notable behaviors; Appendix E). They were encouraged to share anecdotal information and could stop the interview or refrain from answering questions at any time. L ocations were taken from the nearest point to the location described using Garmin eTrex 20 handheld GPS units. Sightings were independently verified by
88 other parties present where possible. Complete descriptions of opportunistic en counters were compiled ( Hallett et al. 2015 b ) and coded using key words in the data, with those containing insights on behavior, reproduction, habitat use, and t hreats prioritized (Appendix G ). R esults Camera trapping In total, 15 bush dogs (individual animals could not be identified) were captured on camera traps at six locations across the Rupununi (Figure 3 2) Bush dogs were photographed at cameras set near Surama, Karasabai, Shulinab, and Rewa Village s as well as t he Iwokrama International Centre for Rainforest Conservation and D evelopment (IIC) and the Kanuku Mountains Protected Area (KMPA). Across the entire study area bush dogs showed a relative abundance of 0. 01 with RAIs of 0.02, 0.03, 0.03, 0.02, 0.01, and 0.01 from Surama, Karasabai, Shul i nab, Rewa, IIC, and KMPA respectiv ely. Interviews of Natural Historians One hundred and three interviews with NH s claiming to have opportunistically encountered bush dogs were conducted, of which 95 were considered valid. Previously undocumented bush dog encounters were recorded from Pa ramakatoi, Surama, Rupertee, Toka, Aranaputa, Yakarinta, Apoteri, Rewa, Kaibaiku, Karasabai, Yupukari, Quatatta, Kwaimatta, Nappi, Katoka, Moco Moco, Quarrie, Parikwarnau, Shulinab, Meriwau, Potarinau, Sand Creek, Rupunau, Katoonarib, Maruranau and Awaruna u Villages, Dubulay, Manari and Dadanawa Ranches, Demerara Timbers Limited and Bai Shan Lin logging concession s state land along the B erbice, Rewa and Kwitaro Rivers and Kusad Mountain, the Iwokrama International Centre for Rainforest Conservation and Dev elopment and the Kanuku Mountains Protected Area (Figure 3 1; 3 3). The general consensus amon g NHs was that bush dogs exist in the Rupununi, potentially
89 even in relatively high numbers for this species, but that they are rarely seen even by those who have spent extensive amounts of time in forest and savanna. According to NHs bus h dogs are diurnal although a number of records also included descriptions of hearing their barks after dusk and before dawn as well. Mean observed pack size was 3.45 (mode = 4, range = 1 12 individuals). Several encounters described only a single animal or observations of only part of a pack. Behaviors observed included (from most to least often reported) running in single file, vocalizing (barking), pursuing prey, crossin g roads, searching / smelling along the ground, resting (laying down), interacting with other members of the pack, and interacting with offspring Interactions between members of packs included ng with tails wagging. Mean litt er size in encounters where pups were observed together was 3.3, but single pups were also obs erved. Bush dogs were observed hunting Cuniculus paca most frequently, followed by Dasyprocta leporina Mazama americana Mazama nemorivaga Pecari tajacu Tupinambis teguixin and Sus scofra Six successful kills were observed. Encounters o ccurred in a variety of habitat types with the majority occurring in lowland forest followed by upland forest savanna, bush islands, and gallery forest. With regard to land use, encounters occurred within (from most of least often reported) indigenous titled lands, protected areas, private ranches, state owned land, and logging and mining concessions. In terms of key resources, encounters occurred most often along riv ers, creeks and ponds, followed by along forest edges, within traditional hunting grounds, near traditional farming sites, along the foot of mountains, and within tourism areas. Observed threats to bush dog populations included wild caught individuals kep t locally as pets, legal possession by commercial pet traders, disappearances following human disturbance disease tra nsfers from domestic animals, road
90 strikes, and individuals killed for sport by hunters The number and frequency of observations of each behavior, prey species, habitat types, land use, key resources in proximity, and threats can be found in Table 3 3. D iscussion Bush dogs are notoriously difficult to study in the wild. Active and passive capture s of this species ha ve been sparse and firs t hand observations by university trained scientists in the field have resul ted in few published accounts While documentation of bush dogs from camera traps is important, the documentation of opportunistic encounters by local informants may serve as a cr itical tool for filling the knowledge gaps currently plaguing the management of th is species. Predatory Behavior Hunting behavior was described similarly in all accounts, with bush dogs following closely behind their potential prey in a single file, all in dividuals having their ears down, tails either straight up or wagging side to side, and barking intently. In the Rupununi, bush dogs are named ( Ai for their high pitched, yapping vocalizations and most reported hearing them long before the an i mals came into view. H unting packs were always described with the largest individual in front and smal ler animals towards the back. Bush dogs are revered in the Rupununi for their hunting prowess. They are believed to have extraordinarily high success rates, typically attributed to stamina and sheer determination. Hunts were described as bush dogs pursuing their target at a moderate pace for extended periods until their prey was eventually overtaken or collapsed due to extreme exhaustion. A powerful sense of smell allows bush dogs to follow their prey even when it is out of sight (Nilsson et al. 2014) and its high pitc hed barks keep prey constantly in motion (B. Phillips pers. comm.) Pack
91 hunting strategies based on stamina instead of speed are not unusual in the genus canid (Bailey et al. 2013) Descriptions of predatory events indicate d a preference for C. paca but a willingness to pursue a wide variety of prey. This supports existing literatu re on bush dog diets (Eisenberg 1989; Emmons 1997; Hunter 2011), but differs from a preference for D. novemcinctus documented in the Brazilian Pant anal (de Souza Lima et al. 2 009; 2013). Interviews revealed u nusual descriptions of predatory behavior such as bush dogs entering the underground den of C. paca and several accounts of overtaking prey in the water. Bush dogs are known to be comfortable underground and in the wat er, as they take up residence in abandoned burrows (Desbiez 2013) and are known to be strong swimmers (Hunter 2011) Cooperative hunting strategies were hypothesized by NHs as individuals were observed taking on different tasks in an effort to make a kill the l argest individuals were described as responsible for flushin g and initial attacks, while smaller individuals wait and support subduing prey. Feeding behavior was described as frenzied and lacking distinct order, with all individuals simultaneously te aring away pieces of fles h. On two occasions, NHs intervened and opportunistically seized potential prey from bush dogs an anecdote that may provide some insight on early inspirations for adopting wild and domesticated dogs to alleviate hunting effort Reproductive Behavior Mean litter size documented here was 3.3, although four encounters included only a single pup. NH abandoned or ventured too far away from den sites. Adul t females were never encountered in the vicinity of lone pups. Bush dog reproduction has been described as aseasonal, with birth peaks in the rainy season, average gestation of 67 days, and litter sizes typically ranging from 3 6 pups but reachi ng up to 1 0 (Hunter 2011).
92 Of the three litters observed by NH s, each was guarded by a single female who growled and displayed her teeth when approached. Den sites of these observed litters included a hollow trunk of a fallen tree (D. DeFreitas pers. obs.) a cav ity formed between rocks along a creek (K. Davis pers. comm.) and an abandoned burrow most likely excavated by Priodontes maximus an observation docum ented previously (Desbiez 2013). In this case, the female was moving three very young pups from a burr ow that was beginning to flood following a heavy rain. Females physically carried small pups, holding them in their teeth by the scruff of the neck (L. Campion pers. comm.) Den sites and maternal care have rarely been observed in the wild in this species. Habitat Use Bush dogs are known to use a wide variety of habitats, as verified by camera trap captures and opp ortunistic encounters were recorded in savanna, bush islan ds, gallery forest, lowland forest, and montane forest. More of the opportunistic encounters were observed by NHs in some form of forested habitat (80%) than savanna (19%), and encounters that did occur in savanna were most often closely associated with f orest edges or large forest fragments D e Souza Lima et al. (2012) found that bush dogs used savanna habitats more than expected based on availability in similar habitat in the Pantanal. While bush dogs certainly use savanna habitat in the Rupununi, NH speculated that use is most likely limited to dispersal between forested are as or in pursuit of prey. Within habitat types, bush dogs were often associated with water (rivers, creeks and ponds), near habitat edges or along the foot of mountains T he bush dog s highly carnivorous nature likely results in movement patterns dictated by the presence of prey (de Souza Lima et al. 2012). Camera trap data from this study indicates that sites with these attributes showed higher diversity and relative abun dance of both predators and prey than other sites (pers. obs.).
93 Cuniculus paca are also known to favor areas near water (Michalski & Norris 2011), but our sample was also likely biased towards areas that people fav or for subsistence activities. In the fu ture, qu antifying sampling efforts of NH s (in the form of time spent in the forest) may allow for more detailed habitat use analyses Many observations occurred within or near to areas of relatively low or intermittent human activity (traditional hunting grounds, rotational farming areas, tourism areas, logging concessions employing selective logging), suggesting some tolerance for human activity. While bush dogs were frequently observed in proximity to roads or trails, they were never observed using the m to travel or hunt (they were always observed crossing perpendicular to the road ). Only individuals being kept in captivity were observed in close association with villages or commercial activity. Perceived and Observed Threats to Bush Dogs in the Rupu nuni Collecting data from natural historians provides unique access to information on site specific threats to populations of species of interest that are generally not present in the quantitative datasets of scientific resea rch. In this study, NH s identi fied local threats facing bush dog populations. While they closely match the range wide i ssues threatening bush dogs (De Matteo et al. 2011), the drivers and unique conditions described by NHs provide critical region specific details necessary for developi ng effective management inte rventions. Pet trade Seven encounters included descriptions of bush dogs that were captured and kept as pets locally (Figure 3 3). Before the arrival of the domestic dog with the Spanish and Portuguese, bush dogs were kept as hunting dogs by the ancestors of current residents (L. Haynes pers. comm.). Stories of their legendary hunting abilities are common and often connected to t he
94 mythical tamona forest spirits that are believed to live in the high forest and usually descri bed as the masters of animals responsible for their pr oliferation and movements (Daly 2015). Traditional beliefs dictate that not only are bush dogs very successful hunters, but also that temporarily capturing a bush dog, drawing blood from its nose, an d smearing the blood on the nose of a domestic dog will transform that dog into a skilled h unter (L. Haynes pers. comm.). For those who maintain this belief, the opportunistic encounter of bush dogs, especially pups, may present an opportunity to increase hunting success. Keeping wild animals as pets is a common practice in Makushi and Wapichan cultures, with animals living in loose association with a family homestead untethered and largely responsible for their own care (R. Roberts pers. comm.). Wild to raise into adulthood before the animals return to their natural habitat on their own (S. James pers. obs.). All bush dogs kept in local captivity were found as pups and rais ed alongside domestic dogs, with few surviving into adulthood (D. DeFreitas pers. obs.). Bush dogs are considered personal pets of the tamona and it is a widely held belief that tamona will punish those who harass or needlessly kill animals that they favor (A Jackman pers. comm.). Reverence or fear of tamona was cited by a number of NHs as a key deterrent to harassing, killing or capturing bush dogs, while others were able to overcome this belief. While the contemporary d esire to capture and keep b ush dogs may be partially driven by fascination created by traditional stories of their skill at hunting wild game, the erosion of traditional taboos associated with exposure to western religion (Luzar et al. 2012) and global popular culture (Freeland 1996 ) may also be reducing the perceived consequences that prevented these behaviors in the past. Keeping bush dogs as pets in the Rupununi is uncommon a practice that is
95 opportunistic in nature and may be driven by curiosity as well as the erosion of trad itional belief systems Three NHs also observed wild caught bush dogs in the custody of international pet traders. Bush dogs were only recently protected in Guyana under the Wildlife Management and Conservation Regulations (EPA, 2009), but international trade of this species has been regulated under Appendix I of CITES since 1977. Guyana is home to a thriving legal pet trade (M. Pierre pers. obs.) but a lack of resources and capacity for enforcement also allows for illegal trade to occur. It is unclea r where or to whom these individuals were destined to be sold or what drives a market for captive bush dogs, but any formal trade in this natural ly rare species is a concern. Disease transfer from domestic canids and l ivestock According to NHs, d isease trans fer from domestic canids was speculated to have resulted in the deaths of wild caught individuals in captivity. An individual kept in close association with domestic canids in Rupertee Village suffered from loss of hair (in patches) and body mass preceding death suggesting that it contracted mange (S. Andries pers. comm.) Another NH who cared for th r e e bu sh dog pups after they were given to Dadanawa Ranch by a local hunter believed that they died from the bacterial disease Leptosp ira sp. (D. DeFreitas pers. obs.). As a rancher, the NH is familiar with the signs and symptoms of this disease in cattle and domestic dogs (vomiting, lethargy, loss of appetite, jaundice). Bush dogs held in captivity around domestic dogs or livestock ar e especially at risk of contracting diseases, but free roaming domestic canids also form a potential vector for disease transfer to wild bush dogs Bush dogs are highly susceptible to diseases such as mange, leishmaniasis, parvovirus, rabies, and canine d istemper, and to parasites such as Dioctophyma renale Amphimerus interruptus Lagochilascaris spp., Spirocerca lupi Toxoplasma gondii and Echinococcus vogeli (DeMatteo et al. 2008, Jorge et al. 2011) many of which can be attributed to exposure to
96 domes tic canids. V iruses, diseases, and parasites spread quickly, killing entire packs of this highly social canid (Mann et al. 1980; DeMatteo, 2008). Domestic canids in the Rupununi live in loose association with people and are primarily used for hunting. L ocal dogs often form packs that hun t and scavenge i ndependently the vast majority of which are unvaccinated (F. Li pers. c omm.). The potential for interaction between feral domestic dogs and bush dogs outside of villages in adjacent forests and savannas (packs of feral domestic dogs were observed on camera traps in this study) extend s the threat of disease transfer to wild bush dog populations. Direct and indirect impacts of r oads Bush dogs were observed crossing roads in 20% of encounters in this study. Roa ds can serve as both an indirect (fragment habitat, provide access to hunters/trappers ) and direct (road strikes) threats to wildlife (Fahrig & Rytwinski 2009) Seven of the encounters on roads occurred along the Georgetown Lethem highway which serves as the primary means for oad is unpaved, but has been successively improved over time by annual grading permitting increasingly higher speeds and vehicle densities. A recent environmental impac t assessment of this road found a relatively low occur re nce of road strikes on medium and large mammals (E. Paemelaere pers. comm.); however, two NHs reported observing bush dog road kills (including a pup ). Current plans to pave the GT Lethem highway would dramatically increase speed and density of traffic, as well as noise, potentially leading to an increase in road strikes and further disrupting existing habitat along the roadway. Even infrequent direct mortality events from road strikes disproport ionately impacts naturally rare species. Increased noise and traffic could also functionally fragment habitat leading to decreased genetic diversity at the metapopulation level (Barber et al. 2010) This is especially a concern in the Rupununi, as the GT Lethem highway krama).
97 NHs related n ine additional descriptions of road crossings on two track roads from t he GT Lethem highway to villages and from villages to tradition al farming grounds. These roads are typically in very rough condition, support ing only low speed and vehicle density. These factors reduce the probabil ity of collisions with wildlife; h owever, several NHs indicated that encounters of this species decreas ed following the expansion and improvement of two track roads used by tractors to accommodate a wider range of vehicles and increase the efficiency of resource extraction (e.g. timber, agricultural products) NHs hypot hesized that these secondary effects (e.g. noise associated with road construction and resource extraction, increased traffic and access provided to hunters, farmers and fisherman ) were ultimately responsible for bush dogs either dispe rsing away from the area or shifting behavior to avoid enc ounters with people on these newly expanded roads Loss of h abitat Expansion of large scale agriculture and inappropriate fire management serve as the greatest threats to wildlife inhabiting the cerrado savanna habitat mosaic of the Rupununi lowlands. While savanna is not prime bush dog habitat, are known to use it more than expected (De Matteo et al. 2014), with NHs speculating that the Rupununi savannas occasionally facilitate dispersal and hunting activity Growing interest in the de velopment of the Rupununi as one of the largest intact stretches of cerrado savanna habitat ( the continent s most threatened and least protected habitat type ; Klink & Machado 2005; Durigan et al. 2007 ), as well as the forest frag ments and forested corridors that maintain connectivity between intact forested areas across Guyana, Suriname, Brazil, and Venezuela. NHs indicate that growing demand for construction materials from existing large scale agriculture operations has driven a n increase in small scale commercial logging in savanna
98 forests (H. Barnaba s pers. comm.) habitats that are naturally limited in diversity a nd biomass of timber trees. More frequent catastrophic savanna fires were described as threats to transitional forests between forest and savanna and forest fragments converting forest edges to savanna and primary to secondary forest. Indigenous people have been mimicking the natural regenerative properties of fire in the Rupununi savanna for centuries. Howeve r, loss of local knowledge and the interruption of predictable seasonal weather patterns by a changing climate have resulted in the increased frequency and intensity of savanna fires (N. Fredericks pers. comm.) NHs in Rupertee and Moco Moco village repor ted the disappearance of bush dogs following catastrophic fires that occurred in the 1980s (A. Primus pers. comm.) but more recent reports of opportunistic encounters at these locations indicate that they may have returned after forests recovered (L. Aldi e pers. comm.) Bush dogs are most frequently associated with large tracts of intact forest habitat, an asset which Guy ana currently possesses, with >7 5% of its total land area in forest (GFC 2013 ). The annual d eforestation rate in Guyana is relatively low, with a mean of 0.04% (range = 0.02% 0.08%; 1990 2013), compared to a global mean of 0.52% (GFC 2012). Relatively strict regulations of the Guyana Forestry Commission (GFC) and naturally slow growing Guiana Shield forests that support a limited numb er of commercially viable species h ave resulted in selective logging as the most common strategy for timber extraction in Guyana. The majority of current deforestation is driven by mining activities in the cou 2012). Howeve r, large, controversial logging concessions east of the Berbice and Rewa rivers indicate increasing pressure to exploit Rupununi resources ultimately threatening habitat for bush dogs and other wildlife
99 Conclusion Indigenous ecological knowledge is applicable to many aspects of scientific study, but is especially valuable in cases where species are rare or elusive. A lifetime of experience in the local environment makes local natural historians trained observers, and partnerships with communitie s and land owners in regions where bush dogs are known to exist may help researchers unlock the secrets of this species while creating incentives to support conservation. Camera trap captures presented here provide additional evidence of bush dogs in the wild in Guyana, but the rich dataset derived from semi structured interviews with local natural historians provides key insights into the biology, ecology, and conservation of bush dogs that may inform the manageme nt of this little known species.
100 Table 3 1 Location, year, number of trap stations, number of trap nights, and number of bush dog encounters from previous camera trap studies conducted in Guyana, studies conducted by the author in bold Site Author Year Trap stations Trap Nights BD occ. Eastern Kanuku Mountains, lower Kwitaro River J. Sanderson & L. Ignacio 2002 16 160 0 Konashen Community owned Conservation Area J. Sanderson, E. Alexander, V. Antone, C. Yukuma 2008 20 62 0 Rewa River R. Pickles, N. McCann, A. Holland 2011 12 218 0 Iwokrama International Centre for Rainforest Conservation & Development A. Roopsind, T, Caughlin, H. Sambhu, J. Fragoso, J. Putz 2011 52 1,613 0 Karanambu Ranch E. Paemelaere, E. Payan, D. McTurk 2012 67 1,972 0 Dadanawa Ranch E. Paemelaere, E. Payan, D. DeFreitas 2013 43 1,466 0 Variety Woods & Greenheart Ltd. Charabaru Concession E. Paemelaere 2013 36 1,593 0 Kanuku Mountains Region M. Hallett, A. Holland, A. Roberts, M. David, A. Jackman 2013 115 4,059 2 Upper Berbice River M. Pierre, L. Ignacio, D. Torres, E. Torres, E. Paemelaere 2014 32 1,411 0 Chenapau Village E. Paemelaere, N. Carter, F. Carter, R. Williams 2014 41 1,204 0 South Rupununi Savanna Kusad Mtn. & Parabara Village E. Paemelaere, D. Fernandes, L. Ignacio, A. Johnny 2014 39 1,295 0 Apoteri Village E. Paemelaere 2015 21 1,040 0 Bai Shan Lin Concession E. Paemelaere 2015 35 1,358 0 Mining District 2 Frenchman Site J. Liddell 2016 23 151 0
101 Table 3 1. Continued Demerera Timbers Limited Siparuni Concession M. Pierre, E. Paemelaere, E. Payan 2016 33 1,236 0 Landscape scale Study of the Rupununi Region M. Hallett et al. (this study) 2014 2017 372 62,010 6 Table 3 2 Author, location, number of trap nights, number of occasions, and relative abundance (RAI) of bush dogs from previous camera trap studies Author Study Location # of trap nights # of occasions RAI Beisiegel (2009) So Paulo, Brazil 4,818 1 0.021 Bianchi (2009) Mato Grosso do Sul, Brazil 2,238 1 0.045 Michalski (2010) Alta Floresta, Brazil 6,721 2 0.030 Negres et al. (2011) Tocantins, Brazil 7,929 1 0.013 Bergallo et al. (2012) Par, Brazil 3,572 3 0.084 Fusco Costa & Ingberman (2013) Paran, Brazil 4,112 3 0.073 Rocha et al. (2015) Amazonas, Brazil 4,894 3 0.061 Ferreira et al. (2015) Minas Gerais, Brazil 6,000 1 0.017 Meyer et al. (2015) Panama 31,755 8 0.025 De Oliveira et al. (2016) Brazilian Amazon 15,888 7 0.044 Hallett et al. (2017) Kanuku Mountains, Guyana 4,059 2 0.049
102 Table 3 3. D istribution of opportunistic encounters of bush dogs by key variable Variable Number of Accounts Percentage of Total Behavior Running in single file 61 64% Vocalizing 55 59% Hunting 27 28% Crossing road 20 21% Wandering/smelling 10 11% Resting 4 4% Sleeping in den 3 3% Playful interaction 2 2% Female with pups 2 2% Prey Species Cuniculus paca 18 67% Dasyprocta leporina 3 11% Mazama americana 2 7% Mazama nemorivaga 1 4% Pecari tajacu 1 4% Tupinambis teguixin 1 4% Sus scofra 1 4% Habitat Type Lowland forest 44 46% Upland/montane forest 21 22% Savanna 17 18% Bush island 8 8% Gallery forest 5 5% Land Use Indigenous Lands 65 68% Protected Area 16 17% Private Ranch 6 6% State owned Land 3 3% Logging Concession 3 3% Mining Concession 2 2% Proximity to Key Resources Near river / creek / pond 44 24% Near forest edge 43 23% Traditional hunting grounds 40 22% Near traditional farming area 31 17% Along mountain foot 17 9% Tourism area 8 4% Threats Pet trade 10 48% Human disturbance 6 29% Road strike 2 10% Disease 2 10% Individual killed by hunter 1 5%
103 Figur e 3 1. Distribution of opportunistic encounters of bush dogs in the Rupununi (ESRI 2016)
104 Figure 3 2. New records of bush dogs in the Rupununi Region from camera traps (photos courtesy of Matt Hallett)
105 Figure 3 3. Opportunistic photographic records from the Rupununi (photos clockwise from top right, courtesy of Meshach Pierre Duane DeFreitas, Meshach Pierre and Ronan McDermott) Meshach Pierre Duane DeFreitas Meshach Pierre Ronan McDermott
106 CHAPTER 4 DEVELOPING A ROBUST METHODOLOGY FOR IDENTIFYING INDIVIDUAL GIANT ANTEATERS ( MYRMECOPHAGA TRIDACTYLA ) WITH IMPLICATIONS FOR POPU LATION ESTIMATION I ntroduction The giant anteater ( Myrmecophaga tridactyla ) is a large, charismatic species whose unique behavioral and physical characteristics make it a readily recognizable flagship species for Neotropical forest and savanna habitats. Despite widespread interest in its biology, ecology, and conservation, relatively little is known about this species (Bertassoni et al. 2017). Most studies on giant anteaters have focused on feeding ecology and thermoregulation, with a few studies of radi o or GPS collared individuals providing insights into hom e range size and habitat use. Giant anteaters are a naturally rare species that ranges widely feeding on abundant and readily available prey traits that have limited our understanding of their pop ulation ecology and hindering the conservation and management of this species. The Giant Anteater The giant anteater ranges from Nicaragua to northern Argentina, inhabiting upland and lowland tropical moist and dry forest savann a and open grassland habitats (Eisenberg 1989). Although giant anteater populations are estimated to have decreased by >30% across their range (Miranda et al. 2014), reliable data are unavailable to determine true population densities and trends. Road st rikes, uncontrolled fires, overhunting, and loss of habitat are primary threats, and have resulted in reported extirpations from Belize, Guatemala, El Salvador, and Uruguay, as well as the states of Santa Catarina Rio de Janeiro, and Esprito Santo in Bra zil (Miranda et al. 2014). Naturally low population densities, l ow reproductive rates, and long parental care make anteater populations vulnerable to, and ill equipped to recover from, population declines (Rodrigues et al.
107 2008). As a result, they have b Species and under Appendix II of CITES (Miranda et al. 2014). Feeding ecology Giant anteaters are highly specialized predators that feed solely on ants and termites. The distribution, shape, a availability and the state of the specific conditions in which it resides (Di Blanco et al. 2016). They fulfill their energetic and nutritional needs by selectively feeding in short bu rsts spread across many foraging locations with prey preference determined by the type of defense and nutritional values of potential prey (Redford 1985). A large bodied, true mammalian (Redford & Dorea 1984) results in a relatively slow metabolism, and need for consistent access to an abundance of food resources (Medri et al. 2003). While ants and termites are certainly abundant in the Neotropics, an individual giant anteater may need t o eat >30,000 ants and termites each day to meet its needs (Dinerstein 2013). Securing this abundance of food resources requires individual anteaters to forage over larg e areas for their survival. The area that an individual animal occupies is determi ned by the availability of resources in that space individuals in resource poor areas must forage more widely than those inhabiting resource rich regions (Schoener 1974). Home ranges (minimum convex polygons) from studies of radio and GPS collared antea ters have varied from 0.8 13.5 km 2 (Miranda 2004; Di Blanco et al. 2015). Female anteaters maintain larger home ranges with more overlap (Silveira 1969; Medri & Mour o 2005; Bertassoni et al. 2017) because their distribution is thought to be determined solely by resource availability (Powell 2000), while male home ranges are largely determined by the distribution of available females (Shaw et al. 1987) because of high intraspecific competition (Braga et al. 2010). Remaining variation in home range size may be
108 explained by site variation in habitat quality and resource availability, as well as the distribution of key habitat features, such as distance to foraging and resting sites ( Bertassoni et al. 2017). Habitat use The largest anteater in the world, this species tolerates a wide range of habitats, but is found in greater association with dry, open habitats, which provide an abundance of ground level food resources easily accessible to this species (Quiroga et al. 2016). Giant anteaters are thought to exist at the highest densities in the savannas of Brazil, Guyana, Venezuela, and Colombia (Eisenberg 1989), regions that also support highest densities of termite and ant colonies on well drained soils (Redford 1985). Neotropical savanna and shrub habita ts are not homogenous, supporting a habitat mosaic with variation in their scrub and canopy layers driven by gradients of elevation, soil moisture and nutrients (Toby Pennington et al. 2000). These heterogeneous habitats provide giant anteaters with optim al foraging opportunities in proximity to forest edges or forested patches refuges with sufficient cover to hide from predators, rest, and find shelter from extreme temperatures ( Camilo Alves and Mouro 2006 ; Mouro and Medri 2007; Desbiez and Medri 2010 ; Bertassoni et al. 2017; Di Blanco et al. 2017). Forested areas may serve as particularly important refuges for this species, as their ability to thermoregulate is compromised if they are forced to rest in open habitats, making them increasingly vulne rable to predators (Camilo Alves & Mouro 2006; Mouro and Medri 2007 ; Di Blanco et al. 2017). Indeed, this species, a capable swimmer, will enter water to cool its body temperature ( Camilo Alves and Mouro 2006 ), but primarily seeks shade in which to res t. Jaguars ( Panthera onca ) are the primary predator of adult anteaters (Silveira 2004), although anteaters make up a very small percentage of their overall diet (Cavalcanti & Gese 2010). Jaguars inhabiting semi arid habitats like the Caatinga, Chaco an d Cerrado reportedly prey
109 on giant anteaters more frequently, seemingly because of lower densities of preferred prey species (Silveira, 2004 ; Astete et al., 2008; Rodrigues et al., 2008 ; Mc Bride et al., 2010 ). Cerrado savannas are the second largest eco region in the Neotropics (Mittermeier et al. 2005 ), but are also the most threatened by anthropogenic activities (Klink & Machado 2005) and the least protected (Durigan et al. 2007) Neotropical savanna habitats appear to be of particular importance t o anteater conservation and management because of the unique combination of resources that they support. However, a lack of data has hindered support for prioritizing savanna habitat to support anteater conservation. Anteater abundance and population tre nds have never been estimated from empirical data; in part because scientists have been unable to reliable identify individual animals. Traditional Beliefs and Anteater ID in the Rupununi Traditional beliefs in the Rupununi region of Guyana hold that dec eased acqua intances or family members who were wronged in some way before their death may take the form of the giant anteater in order to take revenge in the afterlife (L. Daly pers. comm.). Local fo lk lore also maintains that canaimas or evil spirits tha t may kill or physically harm people, may also take the form of the giant anteater ( as opposed to its typical form as a jaguar) in an effort to fool people into a sense of complacency before attacking (G. Pereira pers. comm.). Rupununi villagers generally keep a watchful eye out for repeat visits to the area around a village by the same individual anteater (Ra. Roberts pers. comm.). As result, anteaters in close proximity to people are periodically shot (J. George pers. comm.) because they are considered a bad omen and even a threat to human life and/or property. The identification of individual anteaters by their physical attributes is a skill that is developed informally and independently through a diverse set of strategies and approaches for differenti ating individuals.
110 Realizing that guests responded positively to stories about individual animals, guides at Karanambu Ranch applied this informal method to identifying and naming anteaters that they about where individuals are generally found, reproductive history of females, and any associations observed with other known individuals to help their guests connect to local wildlife and ultimately enrich their experi ence. Karanambu is renowned in Guyana for anteater sightings and as a result of the frequency of observation, their guides have become local anteater experts. The goal of this paper is the collaborative development of a formal method for anteater identif ication that brings together the process that Karanambu guides have developed for identifying individual anteaters over time with new technology for visual identification that allows this process to be standar dized and replicated elsewhere. Visual Identifi cation of Individual Animals Photographs represent a powerful source of data, in that they provide definitive confirmation of a species of interest at a given time and place. Increased access to inexpensive, high quality digital cameras has created the po tential for an enormous influx of data on species occurrences collected by a wide variety of actors with all levels of training and experience (Crall et al. 2013). Additionally, m any species have natural markings (stripes, spots, rosettes) that are indivi dual specific allowing for identification of individual animals ( Karanth 1995; Arzoumanian et al. 2005 ; Gamble et al. 2008 ). The ability to recognize individual animals is a powerful tool that allows researchers to estimate population size and density, f itness, vital rates, activity patterns, inter and intra species interactions, and life history (Bolger et al. 2012). However, v isually matching individuals is inefficient and impractical (Matth et al. 2017), making the direct application of informal ide ntification methods developed by practitioners inappropriate for dealing with large photo databases. Recent developments in
111 p hoto matching algorithms have dramatically increased the accuracy and efficiency of the photo matching process (Bolger et al. 2012 ). The visual identification software HotSpotter has successfully ide ntified unique spot patterns in zebra ( Equus grevyi ; E quagga ) giraf fe ( Giraffa camelopardalis reticulata ), and lionfish ( Pterois volitans ; Crall et al., 2013), Wyoming toad ( Anaxyrus baxteri ; Morrison et al. 2016), and Saimaa ringed seal ( Phoca hispida saimensis ; Zhelezniakov et al. 2015) with great er matching success than their visual ID programs using similar algorithms (Crall et al. 2013; Morrison et al. 2016) Giant anteaters are conspicuously marked they have a long, fan like tail, light colored legs that contrast with a dark colored body, and a bold black stripe that is bordered on either side by a contrasting white stripe and extends down the midline of their flanks from thro at to shoulder (Eisenberg 1989). Despite being one of the most charismatic species in the Neotropics, and one whose populations are declining and whose management has been hampered by a lack of data, individual identification using natural markings has ne ver been formally attempted with giant anteaters. Experienced professionals from both the zoological (P. Riger pers. comm.) and eco tourism (M. Roberts pers. comm.) fields have been informally recognizing individual anteaters for decades typically by id entifying a single characteristic trait of each individual that they repeatedly come into contact with within small populations. This paper recognizes and expands on this local expertise, to develop a formal, replicable method for identifying in dividuals of this species. While previous research has provided insights into some aspects of the biology and behavioral ecology of this species, their population ecology, demographics, and distribution are still unknown. Giant anteaters are a large, conspicuous ma mmal that can be observed regularly in a variety of Neotropical habitats. Whether first person observations or camera trap images,
112 anteater records are being accumulated across the tropics, but robust analysis of these data has not been possible to date b lacking individually identifiable markings. We present a robust method for identifying individual giant anteaters, with implications for capture recapture methodology for population estimation. M ethods S tudy area Karanambu Ranch is located in the North Rupununi savannas (Region 9 Upper Takutu Upper Essequibo) of southwestern Guyana (Figure 4 1). The 324 km 2 privately owned ranch, located on the western bank of the Rupununi River, was founded b y the McTurk family as an outpost for the balata (natural rubber) trade in the early 1920s before shifting to a combination of cattle rearing and eco tourism The ranch sits within an 800,000 ha wetland complex composed of freshwater floodplains, seasonal ly inundated savannas, and fragments of savanna woodland connecting the Rio Branco and Essequibo watersheds that are bordered by large, unbroken tracts of pristine tropical forest in the Pakaraima and Kanuku mountains and Iwokrama forest (Mistry & Roopsind 2004). Karanambu Ranch lands consist primarily of cerrado savanna habitat Savanna vegetation consists largely of perennial grasses from the genus Andropogon Mesosetum Paspalum and Trachypogon and a shrub layer dominated by the cayembe tree ( Curatella americana ; Shackley 1998). Shrub density varies based on soil moisture, nutrients, and history of fire, with hilltops covered by forest fragments, depressions with flooded savannas, and the boundaries of rivers, creeks, and ponds flanked by ripa rian forest. Approximately 100 people reside at Karanambu Ranch on a full or part time basis. Most reside in nearby Yupukari, Kwaimatta, and Masara villages, and are employed in eco tourism, scientific research, wildlife conservation / management, hospi tality, or livestock management
113 (McTurk & Spellman 2005). Karanambu maintains a herd of ~250 cattle (Dadanawa breed), and ~30 working hor ses (A. Holland pers. comm.). Anteater Surveys Images of giant anteaters were collected through three separate metho ds in order to test a formalized approach for identifying individual giant anteaters: (1) photos of individuals with known identifications housed in the collections of institutions accredited by the Association of Zoos & Aquariums; (2) photos from opportun istic encounters during anteater tours conducted at Karanambu Ranch, Guyana; and (3) camera trap photos from a survey of Karanambu Ranc h and neighboring Yupukari Village, Guyana Reference photos from AZA institutions Photographs of captive giant anteaters residing at Association of Zoos & Aquariums (AZA) accredited zoos were obtained through the Pangolin, Aardvark & Xenarthra Taxon Advisory Group (PAX TAG) and Giant Anteater Species Survival Plan (Anteater SSP). Progr am leader, Stacey Belhumeur of Reid Park Zoo (Tucson, AZ), contacted staff at institutions housing giant anteaters to request photographs and information on the sex breeding, and life history of animals in their collection. Institutions were notified tha t photographs provided would serve as known individuals in construction of a reference library that would aid in the development of a method for identifying in dividuals in a wild population. P rofile photographs ( both left and right flank) of 32 individual giant anteaters (18 male, 14 female) were provided by 19 AZA accredited institutions (Appendix G) Brevard Zoo (Melbourne, FL), Brookfield Zoo (Chicago, IL), Buffalo Zoo (NY), Cleveland Metroparks Zoo (OH), Dallas Zoo (TX), Fresno Chaffee Zoo (CA), Greens boro Science Center (SC), Houston Zoo (TX), Jacksonville Zoo (FL), Palm Beach Zoo (FL), Phoenix Zoo (AZ), Potawatomi Zoo (South Bend, IN), Reid Park Zoo (Tucson, AZ), Sacramento Zoo (CA), San Antonio Zoo (TX ),
114 Sedgewick County Zoo (Wichita, KS), Sunset Zo o (Manhattan, KS), Turtle Back Zoo (West Orange, NJ), and Zoo Boise (ID). These images served as the reference library for training the algorithm in program HotSpotter in preparation for identifying individual animals in the wild population at Karanambu Ranch (Figure 4 2). Opportunistic o bservations From 2011 2016 giant anteaters encountered during anteater safaris were documented by Gerard Pereira, manager of the Karanambu Trust the research and conservation arm of the ranch. Trust staff accompani es guides and tourists during this activity to provide additional information on natural history, indigenous culture, and resource management / conservation on the ranch As part of a collaboration with the lead author that began in 2011, trust staff bega n to collect data on giant anteater encounters that occurred at the ranch. Photos and videos were taken of each animal along with the date, time, and GPS lo cation (Garmin eTrex 20) of the encounter, as well as anecdotal data on behaviors observed, weather conditions, and descrip tion of sighting / locations. Multiple images from various angles were taken of each individual, with the goal of obtaining clear, unobstructed profile photos on each occasion. Anteater safaris are a primary attraction offered by Karanambu Lodge, a certified tour operator and eco depart from Karanambu Lodge at 05:30, with guides in the vehicles with guests searching for track roads, while ranch vaqueros scan the areas in between on horseback. Vehicles drive slowly, while spotters search both sides of roads. Vaqueros conduct ad hoc searches of the surrounding savanna, reducing their search area based on previous experience. Anteater tours typically last until ~09:00, when researchers and guests return for breakfast. The success rate of anteater sightings on safaris at Karanambu is ~95%. When giant
115 anteaters are spotted by lodge guides or ranch vaqueros, vehicles are maneuvered to put guests and researchers downwind in the best position for prolonged encounters in proximity to the guests In the case that the vehicles and vaq ueros are some distance apart, vaqueros remain upwind at a safe distance, holding their position or walking slowly behind anteaters to dire ct them towards the tourists. Care is taken by Karanambu staff to reduce stress on individual anteaters. Ideally, anteaters maintain natural behavior throughout an encounter, with only the scent and sound of human presence altering the direction of their movement. Vaqueros, guides, and guests remain quiet and chasing of anteaters is highly discouraged. Vaqueros and vehicles pull back immediately when a female with a pup is enco untered. Camera trapping Camera trap photos were obtained as part of a multi species camera trap study of large mammals in the Rupununi Region following well established methods for camera tra p research (Karanth & Nichols 1998 ; Silver 2004 ). Camera traps ( Bushnell Trophy Cam #119447C, #119734C, #119736C, and #119837C ; Bushnell KS, USA) were set 2 3 km apart and 30 40 cm from the ground with a single camera at each site set in proximity to ob served signs of wildlife Cameras were active 24 h per day, with a 1 second delay between captures, recording the date and time with each 3 image sequence. Due to the limited number of roads and trails, camera traps were distributed based on habitat, pro ximity to res ources, and trail type. In total, 372 trap sites were surveyed from May 2014 May 2017 for a total of 62,010 trap nights. Photographs utilized in this study were obtained from camera traps set across 40 trap sites at Karanambu Ranch and Yup ukari Village from June 2014 to March 2016 for a total of 7,747 trap nights (Figure 4 1).
116 Identification of individual anteaters We used the visual identification program HotSpotter (Crall et al. 2013) to identify individual giant anteaters based on their stripe pattern. HotSpotter employs two algorithms, a locate s key points and extract s associated 128 dimensional vector descriptors matching images based on a query of all descriptors available in an image database using an approximate near est neighbor search data structure (Crall et al. 2013) Descriptor matches are scored using Local Naive Bayes Nearest Neighbor methods an approach which simultaneously matches ea ch descriptor in a query to its k nearest n eighbors across all categories (McCann & Lowe 2011). HotSpotter queries provide ranked similarity scores, with higher scores implying more likely matches (Morrison et al. 2016) Images were uploaded to a new da tabase in HotSpotter for this project. Images of individuals with known identifications from AZA institutions were uploaded first to test HotSpotter HotSpotter requires that a rectangular region of inte rest (ROI) be identified for each image; we chose to focus on the 3 for each fo r matches to determine the accuracy of HotSpotter steps with multiple images taken from different angles, under various lighting schemes, with various backgrounds, and with images of the right flanks of AZA animals. Queries on both the original (left flank) and mirror image of the right flank were run to determine if anteaters have identical markings on both flanks. Once a satisfactory success rate was achieved, these steps were repeated with images from camera tra ps and opportunistic encounters at Kar a nambu Ranch and Yupukari Village.
117 R esults The ability of program HotSpotter to successfully identify individual giant anteaters was tested from images of animals in the collection of AZA accredited institution s, c amera trap photos, and images taken during anteater tours at Karanambu Ranch, Guyana ( Figure 4 2). Reference Photos from AZA Institutions Following the initial entry of 3 2 images of known individuals (one image of the left flank of each individual), HotS potter correctly matched 93.8% (30/32) of images of the left flank of the same individual taken on a different occasion (different date/time, location within exhibit; Figure 4 3). Failed matches were caused by low quality images (cell phone images with fe wer pixels) and images with drastically different lighting (low angle morning or afternoon light). These initial matches showed very high similarity scores (>50,000), with a high density of hot spots. Thirty two additional images of t he opposite flank o f the same 32 known individuals were entered into the database in an effort to understand whether or not the giant anteater has matching flanks. The initial queries resulted in zero successful matches. However, mirror images of the opposite flank of the 32 known individuals resul ted in successful matching of 84.4 % (27/32) of images. These matches showed lower similarity scores (>5,000) than matches of images from the same flank, but similarity scores on matches between the l eft and right flank of the same individual ranked higher than similarity scores of the left flanks of different individuals (Figure 4 3). Providing HotSpotter with additional images of the same individual only strengthened similarity scores and the abilit y of the program to successfully match images of the same individual, even increasing its ability to correctly identify images with different angles, lighting, and quality (Figure 4 3). Individual databases are classif
118 Camer a trap images Camera trap surveys of Karanambu Ranch and neighboring Yupukari Village resulted in 114 total photographs of giant anteaters, from 33 occasions, at 18 trap locations. Of the total sample of anteater photos, only 18 (16%) were useable for app lication to visual identification software. Hotspotter software successfully matched images of the same individual from a single sequence of images on 100% of queries, confirming the feasibility of the use of camera trap images with visual ID software (Fi gure 4 4). There was only a single instance of a successful match of camera trap photos taken of the same individual on two occasions (re captured at a d ifferent location and time). Variations in pose and image quality eliminated the majority of camera trap photos in the sample from application to visual identification software. All black and white images (66) produced by photos taken at night with infrared flash (IR) were unusable because the lack of color washed the contrasting aspect of the anteater s stripe. Photographs of individuals walking perpendicular to camera traps (30) were also unusable because the horizontal stripe on the he front or back of the animal. Photos from opportunistic o bservation s A total of 530 encounters with giant anteaters were documented at Karanambu Ranch from January 2011 July 2016, with Karanambu Trust staff collecting GPS data (GPS locations, photographs) on 306 encounters (Figure 4 1). Images were sorted by date and further separ ated into cases where multiple individuals were observed on a single tour. Usable profile photos of at least one flank were obtained from all 306 encounters and usable photos of both flanks were o btained from 199 encounters. Seventy two photos taken fr om encounters that occurred in 2015 were added to the database to test the ability of HotSpotter to successfully ID wild individuals. Following the
119 initial entry of 72 images (one image of the left flank from each occasion), HotSpotter correctly matched 9 4% (68/72) of additional images of the left flank of the same individual taken duri ng the same occasion (Figure 4 5). Unsuccessful matches were images with drastically different lighting (low angle morning or afternoon light) or that were taken at a great distance (>50 m) These initial matches showed very high similarity scores (>30,000), with a high density of hot spots. Entry of 7 2 mirror images of the opposite flank of the same individuals resulted in successful matching of 77.8 % (56/72) of images. D iscussion The results of this paper show, for the first time, a robust method for identifying a unique individually identifiable mark, which could allow re searchers to track captures and recaptures of individuals and thereby provide the opportunity to estimate population densities. Ideally, presenting this method will encourage the application of this method to existing images from studies already conducted as well as the proliferation of new capture recapture studies of giant anteaters, considering the lessons learned from methods e mployed here. Key aspects of employing this method successfully include determining the region of interest (ROI), comparing o pposite flanks, as well as managing issues related to image quality, image angle, and lighting. Selecting the Region of Interest (ROI) Local collaborators suggested three potential regions of interest which they have used to successfully identify individual anteaters in the past: (1) tail shape / size; (2) bands or markings attempting various derivations of the regions of interest, including full bod y, front half only, tail only, and stripe only, the most successful matches resulted from a ROI that focused on the horizontal stripe runnin
120 Our assumption that a combination of factors would be most useful was unfounded for application in HotSpotter a large area (although to additional identifying characteristics may stil l be useful for visual ID by practitioners in small samples). Tail only queries were also ruled out, as local collaborators suggested that this attribute may change several times over the life of an anteater. We also found that queries using both front h alf and full body ROIs were biased by hot spots identified in the vegetation or exposed rock faces in the background of the image s Issues with matching the background scenery were observed by Crall et al. (2013), but did not result in matching failures i n that case. We found that this was a more prevalent issue in the photos from AZA institutions where commonalities existed in background exhibit features (rocks, turf grass, palms) between institutions Tight cropping of the region of interest (ROI) arou the probability of failed matches in each of our samples. Identifying Opposite Flanks Issues arose with regard to matching the right and left flanks of the same animal in our sample. The HotSpotter algorithm ranked all of the photos of left flanks with higher similarity scores than any of the photos of right flanks, regardless of individual, suggesting that the patterns on the opposite flanks of the same individual did not match. This issue is well documented in large car arrays in an effort to document the distinct patterns on both sides of the an simultaneously. Adobe Photoshop we were able to dramatically increase the probability of successfully matching pattern was leading to failed mat ches, rather than the pattern of the stripe itself. Although none
121 of the matches between photos of the left flank and the mirror image of the right flank produced similarity scores as high as the matches of photos of left flank s from different occasions, queries of photos opposite flanks regularly awarded the highest similarity scores to the matching flanks from the same individual The consistency with which opposite flanks were matched was satisfactory and only improved as additional images were added t o the database. Pose, Lighting Conditions, and Image Quality Crall et al. (2013) cited issues with image angle or pose, lighting, and image quality, including f ocus, resolution and contrast as the primary causes of identification failures in their tests of the HotSpotter algorithm with various species. Images taken too far from the front, back, or top of giant anteaters could not be matched with photos of the same individual because the sed in some instances by adding correctly identified images of an individual from various angles to the database, but images with (i.e. broken tail, scar). Cl accurate identification of individuals. Lighting was the pr imary issue that caused matching failures in our dataset. The HotSpotter algorithm was unable to match black and white photos from the IR flashed used by camera traps in our survey. Additionally, stripe patterns were washed out in images where the direct, low angle morning and afternoon sun was behind the photographer. The contrast in coloration of the stripe is the key aspect that makes each stripe unique. Adding additional, correctly identified images of the same individual may help the algorithm overcome this issue, however the best approach would be to avoid setting camera traps or taking photos facing directly to the east or we st particularly around dawn and dusk
122 Other issues arose with images that were taken from a rel atively long distance or with cell phone cameras. The low pixel number in these images resulted in blurry ROIs when they were clipped from th e original image for analysis, and this l ack of detail in the ROIs ultimately resulted in a number of identification failure s These images may need to be removed from the most of the issues discussed in this section are manageable, but require additional human i nteraction to mitigate issues. Data Collection Tools The application of this method to existing techniques for surveying wildlife populations is a key step for esti mating giant anteater populations. We applied our method to reference images of known individuals, images from camera trap surveys, and photos from opportunistic encounters during anteater tours and will discuss issues and provide suggestions fo r applying this method to each. Reference Images of Known Individuals Our study benefitted greatly from a database of referenc e images graciously shared by 19 AZA accredited institutions. Gaining access to a reference dataset of known individ uals provided a critica l base for traini ng the HotSpotter algorithm, ultimately allowing for the identification of wild individuals at Karanambu Ranch. The creation of an accessible image database of known individual anteaters would be a valuable resource that would support of further application of this method, and is worth pursuing in the future. In general, zoo collections are a tremendous resource for researchers and practitioners seeking to inform management and conservation, and we hope that this effort will serve as an e xample of how collaborati ons can benefit both parties.
123 Camera Trap Surveys Camera traps have proven to be an effective tool for sampling a wide variety of medium and large tropical mammals, including cryptic and naturally rare species (Tobler et al. 2008 ), as well as in studies employing capture recapture methodology for estimating population densities of species with recognizable patterns (Karanath & Nichols 1998; Silver 2004). Unfortunately, images from our camera trap survey showed limited applicabili ty for the identification of giant anteaters using HotSpotter However, lack of success in identifying giant anteaters from camera trap photos in this study should not deter researchers from taking up this effort in the future. Individual anteaters were successfully identified from multiple images taken during the same occasion and re capture s of the same individual from different occasion s, however these successes occurred at an unacceptably low frequency. I dentification of individual anteaters in camer a trap studies is possible, but adjustments would need to be made to the tools and study design of future studies to improve performance of vi sual identification programs and achieve robust results With regar d to research tools, a switch from IR fl ash t o white flash cameras may address multipl e issues. No successful match es were made from the black and white photos taken at night by cameras employing (IR) flash even individuals in the same p hoto sequence. Employing cameras that use a white flash woul d provide color ph otos with standard lighting during night occasions. Even during the day in photos taken under the forest canopy, we routinely were forced to omit photos that were too dark or too bright for analysis. White flash cameras may hel p allevia te this issue as well by providing a consistent source of light that illuminates the subject of each image In terms of study design, the camera trap data utilized in this study came from a study primarily focused on jaguars (Hallett et al. 2017). Trap sp acing (2 3 km apart) and placement
124 (mixed placement with some cameras set on roads and trails) may have reduced the potential for captures and re captures of giant anteaters. The 2 3 km between cameras applied here is based on the minimum home range size observed in jaguars (10 km 2 ; Rabinowitz & Nottingham 1986) standard trap spacing in jaguar studies where home range size is unknown (Silver 2004). Although 10 km 2 is similar to the maximum home range (13.5 km 2 ; Di Blanco et al. 2015) observed in a colla red female anteater, home ranges as small as 0.8 km 2 (Miranda 200 4) have also been observed. As such, r educing the distance between cameras may be necessary to increase capture and re capture rate in studie s focused on giant anteaters. Trap placement that optimize s capture rate is also critical. Jaguars and other large carnivores have shown a preference for existing roads and trails (Harmsen et al. 2010), and as a result camera trap studies focused on these species may disproportionately select sites along roads and trails in an effort to increase overall capture rate Giant anteaters avoid the use of roads and trails, perhaps as a means for avoiding their primary predator (jaguars; Quiroga et al. 2016) or primary threat (road strikes; Miranda et al. 2014). We rarely observed giant anteaters on cameras set facing roads, and these few captures showed anteaters crossing perpendicular to roadways, as opposed to walking parallel to roads as observed with species known to actively use roads. Discrepancie s in space use and trap placement may make the results of camera trap studies focused on large carnivores inappropriate for characterizing the abundance or dis tribution of giant anteaters. Additionally, anteaters do not appear to create game trails by tr aveling regular routes, resulting in low overall capture rates even at off trail sites. These inconsistent movement patterns may have also resulted in m any photos of the photos of the front or rear of individuals where the stripe is not visible. The pote ntial for a scent or attractant placed in front of the
125 camera to lure and/or orient individuals in the optimal pose for obtaining profile images should be tested in future studies Although our matching of opposite flanks provided relatively consistent re sults, p air ed camera trap arrays would increase accuracy by providing simultaneous views of both fl anks of an individual from a single event Visual Encounter Surveys Visual surveys have showed limited success in the study of Neotropical mammals. Dens e, remote forests obstruct visibility and provide ample cover for cryptic species. However, giant anteaters inhabiting Neotropical scrub, savanna, and grassland habitats may present a rare opportunity to employ observational techniques. Researchers sugge st that giant anteaters may be most densely populated in the Llanos, Cerrado, Pantanal, and Chaco regions ( Eisenberg 1989) thanks, in part, to a higher density of their favored prey (ants, termites; Redford 1985). Although this species also inhabits lowla nd and upland forests where observational studies may not be feasible, vis ual surveys may provide valuable data on the abundance and distribution o f this species in these critical habitats habitats which are also among some of the most threatened and lea st protected on the continent. Our visual surveys employed vehicles with spotters slowly driving regular transects at standard speeds while vaqueros on horseback conducted ad hoc searches informed by previous experience. Applying this method at Karanambu R anch in the North Rupununi wetlands, anteaters were regularly observed during hours of normal activity (Shaw 1987) exhibiting natural behaviors. Although active search on horseback does flush hiding anteaters, vaqueros travelling the savannas on horseback is a common activity for a working cattle ranch which anteaters appear to have acclimated to over time. T his combination of on and off road searches provided a successful search method, allowing fo r regular visual observation.
126 Due to concerns about th e compromised ability of giant anteaters to thermoregulate (Camilo Alves & Mouro 2006), visual surveys should be conducted at ni ght or around dawn / dusk to minimize potential heat stress during opportunistic encounters and maximizing the potential of obs erving natural behaviors Surveys in this study began before sunrise (05:30) and refrained from prolonged pursuit or excessive manipulation of natural behavior. With the eeps individuals unaware the presence of researchers / tourists, allowing approach for photo document ation and detailed observation. Visual observation surveys may represent the ideal method for obtaining photographs that produce successful individual identification of anteaters using HotSpotter Among the biggest issues impairing the ability of HotSpotter variations in image quality, pose, and lighting conditions. These issues can be directly addressed during visual observations by purchasing a mid to high quality digital camera, taking multiple photos from various angles during each encounter, and putting photographers in the best position to obtain useable photographs for analysis. While every encounter may not present ideal conditions, experience d research teams and access to quality e quipment will help ensure that images obtained will useful for individual identification using HotSpotter Visual surveys during this study were conduct tourist attraction offered by Karanambu Lodge. Collaborations between researchers and practitioners have shown to promote multiple goals boosting productivity and cutting costs of research while supporting conservation and building capacity in local communities (Simons, 2011) The training provided to local collaborators who contributed to this study served to increase capacity among Karanambu staff, while providing data that may ultimately improve the management of anteater habitat. The revenue generating aspect of eco tourism in particular has shown to
127 increase the probability of related activities to contribute to conservation (Stem et al. 2003). As the probability that data provided by this study will be applied to conserve anteaters on ranch lands. Increasing c ollaborations between researchers, citizen scientists, and tour operators at ranches and eco lodges across the Chaco, Cerrado, Llanos, and Pantanal would create the potential to dramatically increase our understanding of giant anteaters in some of their most important habitats, while providing capacity building and income generating opportunities for the people that are directly responsible for thei r stewardship. It may be ex actly t his type of grassroots effort that is needed to fill the knowledge gaps that are constraining efforts to effectively manage giant anteater populations into the future. Conclusion Giant anteaters are a charismatic, flagship species of conservation concern that inhabits Neotropical forests, savannas, and wetlands. Knowledge gaps are hindering management efforts, as actual abundance and distribution of this species are unknown. In t his paper, we presented a method for identifying individual anteaters that, based on a reference dataset of captive individuals in the collections of AZA accredited institutions, showed promise when applied to a wild population at Karanambu Ranch, Guyana. The ability to successfully identify individual animals forms the necessary initial input for capture recapture analysis a proven method for populatio n estimation. P hotos accumulated from visual surveys, ideally in collaboration with an existing eco to urism or citizen science efforts already in place in open habitats, in other locations across this species range should be applied further test this method. O ptimizing the accuracy and efficiency o f individual identification of giant anteaters is critical for applying the method developed here to population estimation
128 Figure 4 1. Map of camera trap locations and opportunistic encounters with giant anteaters at Karanambu Ranch, Guyana (ESRI 2016)
129 Figure 4 2. Examples of images of giant anteaters from AZA institutions (top photos courtesy of Greensboro Science Center and Jacksonville Zoo & Gardens ), a camera trap survey of Karanambu Ranch and Yupukari Village (middle photos courtesy of Matt Hallett ), and opportunistic encounters during anteater tou rs at Karanambu Ranch (bottom photos courtesy of Gerard Pereira )
130 Figure 4 3. Example results from queries of known individuals from AZA accredited institutions. Shown here is an example of a successful identification of (a) the same individual fro m two separate images of the left flank (top left), (b) the same individual from an image of the left flank and mirror image of the right flank (top right), and (c) an individual to several images previously identified in the database (including images of bot h flanks) ; photos courtesy of Greensboro Science Center and Jacksonville Zoo & Gardens
131 Figure 4 4. Example results from queries of camera trap images obtained during a survey of Karnambu Ranch and Yupukari Village. Shown here is an example of a su ccessful identification of (a) the same individual from two images from the same sequence (left) and (b) a re capture of the same individual at a second trap location (right) ; photos courtesy of Matt Hallett
132 Figure 4 5. Example results from queries of photos from opportunistic encounters during anteater tours at Karanambu Ranch. Shown here is an example of a successful identification of (a) the same individual from two images taken during a single encounter (top left); (b) a re capture of the same individual on separate occasions at different locations (top right); and (c) an individual from two images previously identified in the database (i ncluding images of both flanks) ; photos courtesy of Gerard Pereira
133 CHAPTER 5 CONCLUSIONS While some policie s that exclude communities from natural resources in the name of have changed radically. Communities are now the locus of conservation thinking and driving bottom up approaches to conservation and management (Agrawal and Gib son 1999). However, community based conservation is rife also with it s own set o f uniq ue challenges to overcome Idealized images of coherent, long standing, localized sources of authority that are tied to what are assumed to be sustainable resource management regimes (Brosius et al. 1998) are just that idealized (Berkes 2004). The vision of community as the centerpiece of conservation and re source management is attractive. However, c ommunit ies are not merely static groups of i solated people. In reali ty, they are multidimensional, cross scale, social political units or networks that can be elusive and are constantly changin g over time (Carlsson 2000). Com munities do not a ct as simple, uniform agents that either conserve or despoil i nstead they should be viewed as congregations of unique individuals with their own interests and motivations that respond to pressure s and incentives and are embedded in larger systems (Berkes 2004). With the realization that communit ies a re difficult to define we set out to design a project that bui ld the confidence and capacity to allow communities to define thems elves. In previous efforts to enga ge communities in research, conservation, and management, t he focus on what or who community is are generally focused by forces from outside of the community that are attempting to define and organize it. Many such projects have fallen short of meeting their objectives b ecause c ommunity based initiatives should be based on local level solutions t o issues identified by the communities themselves (Leach et al. 1999). Very simply,
134 this means that communit ies should d erive their own solutions to their own problems. P revious projects failed to invest in capacity building, instead building systems which are developed, run and maintained by outside rs with members of communit ies r eceiving only residual benefits (Reed 2008) Passive b enefits also have a history of being unequally al located or not allocated at all (Child 1996) As a result, community compensation for participation in conservation has become a source of criticism of community based conservation as a concept and in practice (Kiss 2004). Studies have also shown tha t the conception of local incentives purely in terms of economic benefits is too narrow, too simplistic, and potentially counterproductive (Berkes 2004). Incentives are multidimensional. Equity and empowerment are often more important than mone tary incentives (Reed 2008) D ecision making processes s hou ld be legitimate, accountable, inclusive, and that take into account multiple stakeholders and interests are highly valued (Berkes 2004). We attempted to circumvent the shortcomings and capitalize on success of previous projects by designing a project based around locally relevant issues ( Figure 1 1) that provid es fair compensation, giv es careful consideration into how we engage with comm unities, is transparent about de cision ma kin g promot es participation and invest s h eavily in capacity building and leadership development (Figure 1 2) Implementation and results to date The goal of this dissertation was to engage Rupununi communit ies in a learning process for conservation and management. While the focus of chapter two was informed by issues that communities themselves defi ned as important, the design of the project was set before the project began t o provide a structure that would facilitate learning and growth through participation While the data generated by our camera trap proj ect is certainl y important open sour cing the remaining chapters of this dissertation made effective engagement and capacity building a
135 central tenant that would ultimately decide our fate. Certainly, the contribution of local collaborators ma kes la ndscape scale research feasible and efficie nt but it also gave community members a chance to contribute to researc h in a way that their knowledge was recognized and they were provided with opportunities t o participate at a high level Information gained through this type of experien tial learning provides the context that directs the formation of values and behavior with participants progressing from action, to understand the consequences of that action, to generalization of these consequen ces in a broader context (Tuss 1996 ). Participants often grow complacent with project over a period of time and request new a venues to maintain engagement ( Evans et al. 2005). This aspect creates both a challenge and an opportunity. Participatory action research provides an outlet for leaders that emerge from the process to take on greater responsibility and ownership over their own development. Effective P AR projects should respond to the experiences and needs of the community, foster collaboration between researchers and com munity research activities, and promote common knowledge and increases community awareness (Finn 1994). From the outset of this project we provided local collaborators with a clear path and incentives for leadership development, with increased responsibility resulting in increased autonomy, compensation, and recognition. A s participants gained interest and confide nce in working with the camera trapping project, we provided t hem with a dditional responsibil i ty autonomy, and opportunities to contribut e to decision making As participants progressed in skills and confidence, we provided training and leadership development opportunities to allow t hem to experience research while building towards independent action and eventually the development of t heir own questions The key to building this kind of capacity w as structuring activities in such a way that skills, knowledge and understanding progress ed from simpler cases to more complex. I ntegrating
136 research and training allow ed data collection and community development to occur simultaneously, as community members are allowed to assume more participation and ownership over research when they were ready. As p articipants b egan to mast er the tasks related to our camera trap project, they were encouraged t o start asking their own questions, based on their own interests or the interests and needs of their community Research has shown that when participa nts design and implement their own studies, they develop a strong vested interest in finding answers to their own questions and implementing the results (National Research Council 1997). While our initial engagement focused on increasing understanding of ecological processes (chapter two), chapters three and four represent community directed initiatives Research, conservation and development are not mutually exclusive. Though rarely documented in the literature, research is known to generate interest in bi odiversity that may lead to increased conservation and management. Much of this may depend on facilitation, but research strategies that emphasize participation are gaining traction in ecological research all over the world. Engaging stakeholders in the research process involves the development of a shared learning process that could change conservation behavio rs in participants and inspire conservation planning and action in communities. We believe that this dissertation provides an example of how research can be utilized as a tool to engage communities of resource users simultaneously increasing scienti fic understanding, contributing to conservation outcomes, and building capacity in communities who often have acc ess to few such opportunities. Application s for Community driven Management The s caffolded interaction and collaboration u tilized in this project was intended to provide a structure for build ing capacity t hrough participatory r esearch and towards community driven natural resource management When local communities are empowered to manage their
137 own resources, a new approach to science and management can be created (Berkes 2004) but transitioning from participat ion in research to resource management informed by data collection involves provi ding a centralized organizational infrastructure that is specifically designed to promote individual, community and regional science based management via an interactive feedback loop (Cooper et al. 2007). Hence, in the future this project will be entering a new phase in which we will begin developing the necessary structures that will facilita te the translation of the results presented here into action. Adaptive management is the application of scientifically informed natural resource management strategies whose recommendations are iteratively evaluated and re vised to improve outcomes and have shown to be an effective means for organizing citizens, research and resource management ac tivities to achieve cumulative, positive impacts on biological di versity (Cooper et al. 2007 ). Our emphasis on experience, interpretation, reflection, and the co production of knowledge were specifically designed to prepare community members for integration into adaptive management. Empowering a community to take ownership over itself and its own development could perhaps be the greatest achievemen t of participatory action research but only time will tell if the approach presented here will be successful in getting there
138 APPENDIX A JAGUAR CAPTURE DATA Site Image ID Habitat Date Time Sex Indiv. ID BHI13 01140129 Lowland 1/14/2017 12:35 Male 12 BHI14 01060113 Lowland 1/6/2017 20:30 Male 8 BHI14 03230712 Lowland 3/23/2017 11:32 Female (with cubs) 9 BHI5 08160563 Montane 8/16/2016 16:28 Male 11 BHI5 09060892 Montane 9/6/2016 16:52 Male 11 BHI5 09080932 Montane 9/8/2016 15:19 Male 8 BHI5 09240090 Montane 9/24/2016 6:20 Female 10 BHI5 11250820 Montane 11/25/2016 16:51 Male 12 BHI5 04070594 Montane 4/7/2017 19:04 Male 11 BHI6 06180377 Lowland 6/18/2016 8:51 Female 10 BHI6 07050083 Lowland 7/5/2016 4:13 Male 11 BHI6 07150232 Lowland 7/15/2016 15:41 Female 10 BHI6 07270467 Lowland 7/27/2016 23:25 Male 11 BHI6 08160899 Lowland 8/16/2016 19:00 Male 11 BHI6 04020828 Lowland 4/2/2017 22:05 Female 10 BHI6 05130739 Lowland 5/13/2017 20:57 Female 10 BHI8 09030162 Bush Island 9/3/2016 1:46 Female 9 DAD1 EK000764 Gallery 3/16/2016 16:33 Female 1 DAD10 EK001495 Gallery 12/1/2015 17:57 Female 1 DAD10 EK001498 Gallery 12/1/2015 19:04 Female 1 DAD10 EK002236 Gallery 1/10/2016 8:33 Female 3 DAD10 EK002487 Gallery 1/18/2016 4:46 Male 2 DAD10 EK002553 Gallery 1/20/2016 18:38 Female 3 DAD10 EK002742 Gallery 1/27/2016 6:50 Female 1
139 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID DAD10 EK002767 Gallery 1/27/2016 19:36 Female 1 DAD10 EK003463 Gallery 2/19/2016 4:05 Female 1 DAD10 EK003529 Gallery 2/21/2016 8:16 Female 1 DAD2 EK001766 Bush Island 4/15/2016 23:32 Male 50 DAD3 EK000271 Gallery 2/17/2016 21:37 Female 1 DAD7 EK000418 Gallery 11/22/2015 20:44 Female 1 DAD7 EK000470 Gallery 11/24/2015 10:20 Male 2 DAD7 EK001135 Gallery 12/13/2015 2:26 Female 1 DAD7 EK001289 Gallery 12/13/2015 17:30 Female 1 DAD8 PICT0133 Gallery 11/24/2015 23:47 Female 3 DAD8 PICT0659 Gallery 1/27/2016 2:00 Female 3 DAD8 PICT0674 Gallery 2/3/2016 6:50 Female 3 DAD9 EK000085 Gallery 12/5/2015 18:23 Male 2 KBAI1 03100280 Gallery 3/10/2016 5:12 Female 13 KBAI1 03110343 Gallery 3/11/2016 2:01 Female 13 KBAI1 03200728 Gallery 3/20/2016 19:24 Male 14 KBAI1 04030283 Gallery 4/3/2016 3:31 Female 13 KBAI1 04120860 Gallery 4/12/2016 18:29 Female 13 KBAI10 EK000130 Lowland 3/24/2016 8:07 Female 13 KBAI17 PICT0742 Lowland 12/14/2016 10:40 Male 14 KBAI17 PICT0751 Lowland 12/14/2016 15:06 Male 14 KBAI17 PICT0757 Lowland 12/15/2016 18:52 Male 14 KBAI17 PICT0768 Lowland 12/16/2016 18:23 Male 14 KBAI17 PICT0769 Lowland 12/17/2016 17:15 Male 14 KBAI17 PICT0775 Lowland 12/17/2016 22:21 Male 14 KBAI17 PICT0778 Lowland 12/17/2016 23:18 Male 14 KBAI2 09270737 Savanna 9/27/2016 2:56 n/a 15 KBAI2 10010800 Savanna 10/1/2016 21:14 Male 14
140 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID KBAI5 03150057 Lowland 3/15/2016 9:35 Female 13 KBAI5 08010170 Lowland 8/1/2016 10:19 Female 13 KBAI9 EK000307 Savanna 8/3/2016 21:58 Male 14 KBU1 EK000637 Bush Island 3/5/2015 11:42 Female 27 KBU11 EK000001 Bush Island 4/6/2015 1:21 Male 26 KBU11 EK000008 Bush Island 4/16/2015 4:15 Male & Female 26 & 27 KBU16 PICT0065 Gallery 12/28/2014 18:40 Male 26 KBU16 EK000028 Gallery 3/30/2015 17:37 Male 26 KBU17 PICT0144 Gallery 10/13/2014 1:00 Male 28 KBU17 PICT0148 Gallery 10/13/2014 8:50 Male 28 KBU17 PICT0205 Gallery 10/22/2014 8:08 Female 27 KBU17 PICT0211 Gallery 10/23/2014 10:20 Male 28 KBU17 PICT0214 Gallery 10/23/2014 14:59 Male 28 KBU17 PICT0221 Gallery 10/24/2014 2:14 Male 28 KBU17 PICT0405 Gallery 11/29/2014 20:06 Male 28 KBU17 PICT0406 Gallery 12/1/2014 6:27 Male 28 KBU17 PICT0485 Gallery 12/18/2014 13:37 Female 27 KBU17 PICT0633 Gallery 1/28/2015 18:33 Male 28 KBU17 PICT0641 Gallery 2/3/2015 19:05 Male 28 KBU17 PICT0679 Gallery 2/8/2015 20:53 Male 28 KBU17 PICT0682 Gallery 2/12/2015 2:20 Male 28 KBU17 PICT0844 Gallery 2/25/2015 21:22 Male 28 KBU19 PICT0122 Bush Island 5/1/2015 1:37 Female 27 KBU2 PICT0389 Bush Island 12/29/2015 22:45 Male 26 KBU2 PICT0457 Bush Island 11/15/2015 20:26 Male 26 KBU2 PICT0468 Bush Island 11/17/2015 21:29 Female 25 KBU2 PICT0476 Bush Island 11/19/2015 23:11 Female 25
141 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID KBU2 PICT0561 Bush Island 12/1/2015 2:42 Male 26 KBU20 PICT0004 Gallery 3/1/2015 5:48 Male 28 KBU20 PICT0067 Gallery 3/30/2015 18:18 Male 28 KBU20 PICT0568 Gallery 2/4/2016 21:38 Male 28 KBU5 PICT0174 Bush Island 7/12/2014 18:14 Female 25 KBU5 PICT0114 Bush Island 1/15/2015 13:06 Female 27 KBU5 PICT0100 Bush Island 6/22/2015 22:07 Male 26 KBU5 PICT0187 Bush Island 7/25/2015 19:11 Female 25 KBU5 PICT0193 Bush Island 7/29/2015 18:24 Male 26 KBU5 PICT0220 Bush Island 8/14/2015 6:54 Male 26 KBU5 PICT0303 Bush Island 9/17/2015 1:58 Male 26 KBU5 PICT0010 Bush Island 10/31/2015 23:29 Male 26 KBU5 PICT0142 Bush Island 12/10/2015 0:28 Male 26 KBU5 PICT0683 Bush Island 1/13/2016 23:21 Male 26 KBU6 EK000093 Gallery 4/27/2015 2:30 Male 29 KBU6 PICT0151 Gallery 2/23/2016 4:57 Male 28 KBU7 PICT0049 Gallery 11/13/2015 22:17 Male 28 KBU7 PICT0052 Gallery 11/17/2015 20:38 Female 27 KBU7 PICT0065 Gallery 12/4/2015 18:44 Male 26 KBU7 PICT0067 Gallery 12/15/2015 17:54 Male 26 KBU7 PICT0073 Gallery 12/27/2015 20:09 Male 26 KBU7 PICT0078 Gallery 1/6/2016 15:32 Male 26 KBU7 PICT0086 Gallery 1/17/2016 0:13 Male 26 KBU7 PICT0100 Gallery 2/1/2016 23:18 Male 26 KBU7 PICT0149 Gallery 3/2/2016 21:39 Male 29 KBU7 PICT0154 Gallery 3/3/2016 5:40 Male 29 KBU9 EK000034 Gallery 11/10/2014 20:36 Male 26 KBU9 EK000060 Gallery 11/13/2015 18:39 Female 24
142 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID KBU9 EK000119 Gallery 12/5/2015 6:00 Male 28 KBU9 EK000163 Gallery 1/14/2016 5:34 Male 28 KBU9 EK000174 Gallery 1/21/2016 15:57 Male 24 KBU9 EK000195 Gallery 1/25/2016 5:03 Female 25 KBU9 EK000205 Gallery 2/2/2016 0:13 Female 25 KMPA12 11050887 Lowland 11/5/2016 0:25 Female 37 KMPA12 11230104 Lowland 11/23/2016 5:58 Female 37 KMPA13 02240040 Riverine 2/24/2017 16:11 Male 40 KMPA14 09210463 Riverine 9/21/2016 18:07 Male 40 KMPA14 09270538 Riverine 9/27/2016 14:19 Male 40 KMPA14 10220751 Riverine 10/22/2016 21:46 Male 40 KMPA14 10250786 Riverine 10/25/2016 11:46 Male 40 KMPA14 10300809 Riverine 10/30/2016 0:26 Male 39 KMPA14 11240019 Riverine 11/24/2016 7:33 Male 40 KMPA14 11260029 Riverine 11/26/2016 0:08 Male 40 KMPA14 12130161 Riverine 12/13/2016 3:25 Male 39 KMPA14 12250270 Riverine 12/25/2016 9:25 Male 39 KMPA14 02060647 Riverine 2/6/2017 20:44 Male 39 KMPA14 02100689 Riverine 2/10/2017 14:41 Male 39 KMPA14 02210922 Riverine 2/21/2017 7:22 Male 40 KMPA14 03090232 Riverine 3/9/2017 2:58 Male 40 KMPA14 03170374 Riverine 3/17/2017 13:01 Male 40 KMPA14 03200431 Riverine 3/20/2017 17:34 Male 40 KMPA14 Riverine 3/20/2017 18:10 Female 41 KMPA14 03210437 Riverine 3/21/2017 6:18 Female 41 KMPA14 03260524 Riverine 3/26/2017 20:26 Male 40 KMPA14 04030646 Riverine 4/3/2017 19:27 Female 41 KMPA14 04210863 Riverine 4/21/2017 7:20 Female 41
143 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID KMPA14 04250905 Riverine 4/25/2017 18:38 Male 40 KMPA14 05180133 Riverine 5/18/2017 10:36 Female 41 KMPA14 06040287 Riverine 6/4/2017 9:52 Male 40 KMPA14 06050294 Riverine 6/5/2017 10:55 Male & Female 40 & 41 KMPA14 06050299 Riverine 6/5/2017 17:34 Female 41 KMPA14 06060308 Riverine 6/6/2017 6:49 Male 40 KMPA16 04230812 Montane 4/23/2017 19:41 Male 39 KMPA17 01250854 Montane 1/25/2017 14:33 Female 38 KMPA17 05130448 Montane 5/13/2017 3:40 Female 38 KMPA17 06100711 Montane 6/10/2017 3:38 Male 39 KMPA18 09010205 Lowland 9/1/2016 9:15 Female 38 KMPA18 01260953 Lowland 1/26/2017 0:20 Female 38 KMPA2 EK000350 Riverine 7/11/2014 14:56 Male 43 KMPA2 11130422 Riverine 11/13/2016 3:45 Female 42 KMPA22 08210076 Montane 8/21/2016 8:33 Male 44 KMPA22 09040177 Montane 9/4/2016 5:41 Female 41 KMPA22 11250519 Montane 11/25/2016 14:32 Male 44 KMPA22 02100581 Montane 2/10/2017 11:33 Male 44 KMPA23 09280986 Lowland 9/28/2016 12:45 Male 44 KMPA23 09280987 Lowland 9/28/2016 14:58 Male 44 KMPA23 09300004 Lowland 9/30/2016 6:20 Male 44 KMPA23 10110199 Lowland 10/11/2016 5:01 Female 37 KMPA23 11200097 Lowland 11/20/2016 0:25 Male 44 KMPA23 05020058 Lowland 5/2/2017 4:24 Female 37 KMPA23 05060084 Lowland 5/6/2017 4:00 Male 44 KMPA23 05140131 Lowland 5/14/2017 23:32 Male 44 KMPA29 11010356 Montane 11/1/2016 14:23 Female 42
144 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID KMPA30 05190566 Riverine 5/19/2017 10:42 Male 43 KMPA5 05120487 Montane 5/12/2017 9:03 Male 43 KMPA7 01300748 Montane 1/30/2017 5:33 Female 42 KMPA8 04290608 Lowland 4/29/2017 6:22 Female 42 MAP10 EK000194 Lowland 7/9/2014 7:45 Male 47 MAP10 EK000204 Lowland 7/10/2014 12:44 Male 47 MAP10 EK000580 Lowland 8/20/2014 8:29 Male 47 MAP10 EK002413 Lowland 12/21/2014 15:32 Male 49 MAP10 EK000253 Lowland 1/26/2015 15:25 Female 48 MAP12 EK000433 Montane 7/10/2015 6:14 Female 48 MAP3 EK001130 Montane 8/11/2015 19:49 Male 47 MAP3 EK001722 Montane 9/26/2015 2:02 Female 46 MAP4 PICT0680 Montane 7/26/2015 21:12 Female 45 MAP5 EK000981 Lowland 1/5/2016 8:38 Male 47 MAP6 EK004453 Riverine 1/8/2016 18:41 Female 45 REW10 03300924 Lowland 3/30/2016 13:14 Female 35 REW10 03310005 Lowland 3/31/2016 13:55 Male 34 REW14 EK000547 Montane 8/11/2016 5:39 Female 36 REW15 EK001063 Montane 8/6/2016 9:27 Female 36 REW16 EK000909 Montane 6/21/2016 13:05 Male 33 REW16 EK001771 Montane 8/5/2016 15:27 Female 36 REW16 EK002264 Montane 9/4/2016 10:04 Male 33 REW2 02270055 Riverine 2/27/2016 17:05 Male 30 REW2 07180373 Riverine 7/18/2016 2:08 Male 30 REW2 11230729 Riverine 11/23/2016 23:19 Female 31 REW2 02280762 Riverine 2/28/2017 2:53 Female 31 REW3 EK000340 Riverine 3/22/2016 8:38 Female 32 REW5 EK000426 Lowland 4/19/2016 11:51 Male 33
145 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID REW5 EK000555 Lowland 9/8/2016 12:44 Male 30 REW5 EK000647 Lowland 9/16/2016 6:14 Male 30 REW5 EK000539 Lowland 2/2/2017 0:40 Male 30 REW6 05280329 Montane 5/28/2016 3:18 Female 35 REW6 02220335 Montane 2/22/2017 16:27 Male 34 REW7 03240307 Lowland 3/24/2016 2:54 Male 33 REW8 EK000069 Gallery 3/1/2016 16:01 Male 33 REW8 EK006924 Gallery 3/24/2016 6:32 Male 33 REW8 EK006953 Gallery 4/5/2016 22:30 Male 33 REW8 EK006955 Gallery 4/7/2016 9:05 Male 33 REW8 EK006965 Gallery 4/20/2016 6:04 Male 35 REW9 EK004156 Lowland 9/17/2016 11:12 Male 33 REW9 EK004552 Lowland 11/21/2016 18:46 Male 33 REW9 EK000177 Lowland 1/7/2017 17:28 Male 33 REW9 EK001170 Lowland 3/2/2017 9:04 Male 33 REW9 EK001345 Lowland 3/24/2017 0:32 Female 31 RH10 EK000169 Riverine 12/12/2015 10:13 Female 21 RH11 EK000791 Lowland 2/25/2016 17:44 Female 21 RH12 EK006918 Montane 3/2/2016 7:07 Male 20 RH14 EK000309 Lowland 12/23/2015 3:20 Male 20 RH16 EK000056 Riverine 12/6/2015 22:22 Male 16 RH16 EK000203 Riverine 2/7/2016 22:19 Male 16 RH16 EK000221 Riverine 2/23/2016 4:58 Male 16 RH16 EK000233 Riverine 3/1/2016 8:38 Male 16 RH16 EK000243 Riverine 3/5/2016 17:57 Male 16 RH16 EK000251 Riverine 3/5/2016 18:40 Male 16 RH16 EK000267 Riverine 3/6/2016 11:43 Male 16 RH16 EK000275 Riverine 3/6/2016 17:12 Male 16
146 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID RH16 EK000281 Riverine 3/7/2016 7:41 Male 16 RH16 EK000291 Riverine 3/8/2016 7:15 Male 16 RH16 EK000296 Riverine 3/8/2016 8:34 Male 16 RH16 EK000314 Riverine 3/8/2016 17:41 Male 16 RH16 EK000373 Riverine 3/25/2016 10:35 Male 17 RH18 EK000995 Lowland 3/30/2016 3:50 Female 18 RH19 EK001556 Riverine 1/31/2016 18:38 Female 18 RH2 EK002846 Riverine 4/8/2016 4:50 Female 19 RH20 EK008395 Lowland 2/11/2016 17:31 Female 18 RH5 EK000388 Lowland 12/17/2015 21:51 Male 22 RH6 EK000875 Lowland 3/6/2016 10:37 Female 21 RH6 EK000923 Lowland 3/9/2016 17:52 Female 21 RH8 EK000124 Montane 12/4/2015 7:55 Female 73 RH9 EK001604 Riverine 2/1/2016 16:41 Male 20 RUP11 04230433 Bush Island 4/23/2017 4:36 Female 6 RUP11 04230441 Bush Island 4/23/2017 15:04 Female 6 RUP11 04240492 Bush Island 4/24/2017 0:53 Female 6 RUP12 04130789 Bush Island 4/13/2017 20:01 Male 7 RUP12 04290850 Bush Island 4/29/2017 5:10 Male 7 RUP2 12290240 Lowland 12/29/2016 7:57 Male 4 RUP2 01150642 Lowland 2/3/2017 12:01 Male 4 RUP2 01280813 Lowland 2/16/2017 5:06 Male 5 RUP2 02070883 Lowland 2/26/2017 1:51 Male 5 RUP2 02170228 Lowland 3/8/2017 6:36 Male 5 SH1 EK001228 Gallery 3/14/2016 14:23 Male 65 SH10 EK000103 Gallery 11/7/2015 23:31 Female 68 SH10 EK000313 Gallery 11/29/2015 19:45 Female 67 SH10 EK000433 Gallery 12/16/2015 6:07 Female 68
147 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID SH10 EK001264 Gallery 1/29/2016 0:26 Female 68 SH10 EK001283 Gallery 1/31/2016 19:06 Female 68 SH10 EK001439 Gallery 2/20/2016 5:36 Female 67 SH12 PICT0053 Savanna 11/13/2015 5:46 Male 66 SH12 PICT0098 Savanna 12/15/2015 3:08 Male 69 SH15 EK000985 Montane 8/18/2016 11:55 Male 70 SH15 EK002354 Montane 11/23/2016 22:12 Male 70 SH15 EK003029 Montane 2/18/2017 1:17 Male 70 SH17 EK001545 Lowland 10/23/2016 22:57 Female 71 SH18 EK003425 Montane 1/30/2017 22:46 Male 70 SH18 EK003893 Montane 3/13/2017 19:57 Male 70 SH19 EK000226 Savanna 10/5/2016 19:29 Female 71 SH20 EK000140 Savanna 6/25/2016 0:17 Female 71 SH21 EK000371 Bush Island 7/23/2016 22:17 Female 71 SH21 EK000416 Bush Island 8/14/2016 20:15 Female 71 SH21 EK000608 Bush Island 2/6/2017 5:06 Male 70 SH22 11240968 Montane 11/24/2016 0:10 Male 70 SH22 11240990 Montane 11/24/2016 20:14 Male 70 SH22 01150419 Montane 1/15/2017 19:07 Male 70 SH22 03200332 Montane 3/20/2017 8:21 Female 72 SH22 04030606 Montane 4/3/2017 21:54 Male 70 SH22 04040611 Montane 4/4/2017 4:33 Male 70 SH4 EK000100 Gallery 11/17/2015 6:32 Male 65 SH8 EK000213 Savanna 11/11/2015 19:07 Female 67 SH8 EK000372 Savanna 11/30/2015 9:02 Male 66 SH8 EK000379 Savanna 12/2/2015 4:54 Male 66 SH8 EK000591 Savanna 1/20/2016 10:03 Female 67
148 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID SH8 EK000651 Savanna 1/22/2016 23:19 Female (with cub) 67 SH8 EK001224 Savanna 2/3/2016 23:51 Male 65 SH8 EK001263 Savanna 2/14/2016 6:38 Female 68 SH8 EK001428 Savanna 2/17/2016 19:14 Female 67 SH8 EK001431 Savanna 2/17/2016 21:12 Female 67 SH8 EK001434 Savanna 2/18/2016 0:03 Female 68 SH8 EK001491 Savanna 2/20/2016 7:51 Male 66 SH8 EK004173 Savanna 4/11/2016 19:15 Female 68 SH9 EK000111 Gallery 11/14/2015 5:18 Male 66 SH9 EK000329 Gallery 12/29/2016 18:57 Female 68 SH9 EK001269 Gallery 2/29/2016 5:22 Male 66 SM11 EK000245 Gallery 6/23/2016 21:29 Male 51 SM11 EK000739 Gallery 7/26/2016 19:28 Male 51 SM11 EK000859 Gallery 7/29/2016 19:36 Male 51 SM11 EK001021 Gallery 8/4/2016 22:33 Male 51 SM11 EK001051 Gallery 8/11/2016 18:00 Female 53 SM11 EK001093 Gallery 8/19/2016 4:42 Male 51 SM11 EK000172 Gallery 2/25/2017 20:40 Male 52 SM11 EK000278 Gallery 3/20/2017 13:58 Male 52 SM11 EK000282 Gallery 3/20/2017 16:51 Male 52 SM11 EK000286 Gallery 3/21/2017 1:01 Male 52 SM11 EK000289 Gallery 3/21/2017 6:11 Male 52 SM14 EK000085 Bush Island 1/25/2017 18:30 Male 55 SM15 06130050 Bush Island 6/13/2016 23:47 Male 55 SM15 09050832 Bush Island 9/5/2016 17:49 Male 55 SM15 09060844 Bush Island 9/6/2016 16:47 Female 55 SM15 09070852 Bush Island 9/7/2016 3:47 Male 52
149 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID SM15 09140893 Bush Island 9/14/2016 15:28 Male 55 SM15 09160905 Bush Island 9/16/2016 19:53 Male 52 SM15 09230964 Bush Island 9/23/2016 17:31 Male 55 SM15 09290027 Bush Island 9/29/2016 23:25 Male 55 SM15 09290043 Bush Island 9/29/2016 23:56 Male 55 SM15 10060047 Bush Island 10/6/2016 16:33 Male 55 SM15 10130079 Bush Island 10/13/2016 22:21 Male 55 SM15 10190131 Bush Island 10/19/2016 19:57 Female 54 SM15 10190151 Bush Island 10/19/2016 22:45 Male 55 SM15 10240174 Bush Island 10/24/2016 2:04 Male 55 SM15 10280499 Bush Island 10/28/2016 16:15 Male 55 SM15 11090547 Bush Island 11/9/2016 18:15 Male 55 SM15 11150601 Bush Island 11/15/2016 7:52 Male 55 SM15 12070811 Bush Island 12/7/2016 7:36 Male 55 SM15 12150848 Bush Island 12/15/2016 16:51 Female 53 SM15 01280308 Bush Island 1/28/2017 7:17 Male 55 SM15 02070375 Bush Island 2/7/2017 17:17 Male 55 SM15 02240508 Bush Island 2/24/2017 3:58 Male 55 SM15 02260616 Bush Island 2/26/2017 3:34 Male 55 SM15 03110704 Bush Island 3/11/2017 22:24 Male 55 SM15 03170800 Bush Island 3/17/2017 4:48 Female 54 SM15 04170033 Bush Island 4/17/2017 18:51 Male 55 SM15 05080767 Bush Island 5/8/2017 20:37 Male 55 SM4 EK000197 Gallery 12/16/2015 5:13 Male 51 SM4 EK000280 Gallery 12/17/2015 1:56 Male 51 SM5 EK000200 Bush Island 11/19/2015 4:40 Female 53 SM5 EK000684 Bush Island 12/5/2015 5:12 Female 53 SM5 EK000703 Bush Island 12/5/2015 14:38 Female 53
150 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID SM5 EK000715 Bush Island 12/6/2015 17:36 Female 53 SM5 EK000787 Bush Island 12/11/2015 8:58 Female 53 SM5 EK000796 Bush Island 12/11/2015 9:34 Female 53 SM5 EK001842 Bush Island 2/13/2016 9:02 Female 53 SM5 EK002330 Bush Island 2/24/2016 5:23 Female 54 SM5 EK000202 Bush Island 6/11/2016 6:57 Female 54 SM5 EK000362 Bush Island 10/16/2016 4:00 Female 54 SM5 EK000655 Bush Island 12/11/2016 7:41 Female 53 SM6 EK002089 Bush Island 1/3/2016 12:47 Female 53 SM6 EK002099 Bush Island 1/3/2016 21:54 Female (with cub) 53 SM6 EK002105 Bush Island 1/4/2016 19:32 Female (with cub) 53 SM6 EK002139 Bush Island 1/8/2016 19:48 Female 53 SM6 EK002473 Bush Island 2/10/2016 17:23 Male 52 SM6 EK003274 Bush Island 3/20/2016 9:00 Male 52 SM6 EK003346 Bush Island 3/26/2016 18:27 Female 53 SM6 EK003364 Bush Island 4/5/2016 5:23 Male 52 SM7 EK000297 Gallery 1/18/2016 17:51 Female 53 SM7 EK000369 Gallery 1/30/2016 9:07 Male 52 SM7 EK000373 Gallery 1/30/2016 22:51 Male 52 SM8 EK000056 Bush Island 7/19/2016 3:46 Male 51 SM8 EK000067 Bush Island 8/5/2016 10:18 Male 52 SM8 EK000238 Bush Island 10/9/2016 3:06 Male 52 SUR10 EK001525 Lowland 10/16/2015 10:38 Female 62 SUR10 EK002453 Lowland 12/24/2015 2:36 Female 62 SUR10 EK002714 Lowland 1/2/2016 22:15 Female 61
151 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID SUR10 EK002759 Lowland 1/5/2016 7:38 Female (with cubs) 61 SUR10 EK002923 Lowland 1/11/2016 14:31 Female (with cubs) 61 SUR10 EK002945 Lowland 1/12/2016 9:18 Female 63 SUR10 EK002997 Lowland 1/15/2016 9:12 Female 63 SUR10 EK003173 Lowland 1/21/2016 22:35 Female 61 SUR10 EK003177 Lowland 1/22/2016 7:36 Female 61 SUR11 EK002256 Lowland 9/15/2015 15:11 Male 57 SUR11 EK000389 Lowland 3/3/2016 8:38 Male 57 SUR11 EK005218 Lowland 5/17/2016 11:00 Female 61 SUR11 EK005334 Lowland 5/20/2016 16:58 Male 59 SUR11 EK007321 Lowland 7/11/2016 6:11 Female 61 SUR12 EK000703 Riverine 1/21/2016 8:17 Male 57 SUR12 EK000706 Riverine 1/21/2016 9:03 Male 57 SUR12 EK000712 Riverine 1/21/2016 10:51 Male 57 SUR12 EK000715 Riverine 1/21/2016 16:42 Male 57 SUR15 EK000623 Lowland 8/1/2016 6:31 Male 57 SUR15 EK000752 Lowland 8/13/2016 3:03 Male 60 SUR15 EK000932 Lowland 8/23/2016 22:10 Male 57 SUR15 EK000938 Lowland 8/24/2016 1:35 Male 60 SUR15 EK000942 Lowland 8/24/2016 16:55 Male 60 SUR15 EK001634 Lowland 9/25/2016 15:44 Female 61 SUR15 EK002064 Lowland 10/26/2016 2:27 Male 57 SUR15 EK002243 Lowland 11/13/2016 18:53 Male 57 SUR15 EK002318 Lowland 11/23/2016 22:35 Male 60 SUR15 EK002460 Lowland 12/2/2016 18:03 Female 62 SUR15 EK002465 Lowland 12/2/2016 22:35 Male 60
152 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID SUR16 EK000440 Lowland 8/26/2016 15:52 Male 57 SUR16 EK000618 Lowland 9/28/2016 5:33 Male 64 SUR16 EK000913 Lowland 10/26/2016 1:54 Male 57 SUR16 EK000930 Lowland 10/28/2016 9:45 Male 64 SUR16 EK000935 Lowland 10/28/2016 14:08 Male 64 SUR16 EK000939 Lowland 10/29/2016 2:58 Male 64 SUR16 EK000944 Lowland 10/29/2016 3:46 Male 57 SUR16 EK000947 Lowland 10/29/2016 18:13 Male 57 SUR17 EK000253 Lowland 8/12/2016 22:59 Female 63 SUR18 EK001043 Lowland 9/4/2016 21:07 Female 62 SUR18 EK001053 Lowland 9/5/2016 15:23 Female 63 SUR18 EK002036 Lowland 11/25/2016 15:20 Female 62 SUR20 EK000089 Lowland 7/22/2016 1:52 Male 60 SUR24 EK002156 Lowland 10/11/2016 0:57 Male 60 SUR24 EK002328 Lowland 10/23/2016 10:56 Female 63 SUR24 EK002399 Lowland 10/26/2016 23:56 Female 63 SUR24 EK002442 Lowland 10/31/2016 20:47 Male 60 SUR24 EK003225 Lowland 12/3/2016 13:28 Female 62 SUR3 EK000480 Lowland 8/9/2015 7:10 Male 56 SUR3 EK000760 Lowland 8/24/2015 19:50 Male 57 SUR3 EK002205 Lowland 11/24/2015 1:41 Female 58 SUR3 EK002222 Lowland 11/25/2015 3:57 Male 57 SUR3 EK002348 Lowland 12/7/2015 17:25 Male 57 SUR3 EK003494 Lowland 2/19/2016 23:46 Female 58 SUR3 EK000149 Lowland 2/21/2016 7:28 Female 58 SUR3 EK000285 Lowland 2/24/2016 12:59 Male 56 SUR3 EK000932 Lowland 3/26/2016 21:41 Female 58 SUR3 EK001041 Lowland 4/2/2016 15:13 Male 57
153 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID SUR3 EK001166 Lowland 4/8/2016 4:43 Female 58 SUR3 EK001292 Lowland 4/10/2016 22:10 Male 57 SUR3 EK002446 Lowland 5/28/2016 23:38 Male 56 SUR4 EK001445 Lowland 8/5/2015 4:01 Female 58 SUR4 EK004370 Lowland 5/23/2016 3:35 Male 57 SUR7 EK000413 Montane 8/17/2015 14:54 Male 56 SUR7 EK000619 Montane 9/23/2015 14:05 Male 59 SUR8 EK001407 Lowland 8/29/2015 20:10 Male 56 SUR8 EK002639 Lowland 11/25/2015 19:10 Female 58 SUR8 EK003189 Lowland 1/28/2016 2:36 Female 58 SUR9 EK000166 Lowland 7/27/2015 0:32 Male 57 SUR9 EK000722 Lowland 8/19/2015 21:24 Male 57 SUR9 EK000932 Lowland 8/28/2015 17:39 Mlae 57 SUR9 EK000974 Lowland 8/29/2015 21:04 Male 57 SUR9 EK001139 Lowland 9/10/2015 15:44 Male 56 SUR9 EK001280 Lowland 9/23/2015 19:10 Male 60 SUR9 EK001301 Lowland 9/24/2015 17:14 Female 58 SUR9 EK001336 Lowland 9/27/2015 22:30 Male 60 SUR9 EK002223 Lowland 11/10/2015 15:19 Male 59 SUR9 EK002424 Lowland 11/25/2015 1:46 Female 58 SUR9 EK002456 Lowland 11/29/2015 9:48 Female 58 SUR9 EK002572 Lowland 12/11/2015 4:03 Female 58 SUR9 EK000285 Lowland 3/8/2016 1:29 Male 60 SUR9 EK000355 Lowland 3/14/2016 10:55 Male 57 SUR9 EK000430 Lowland 3/19/2016 7:28 Male 57 SUR9 EK001311 Lowland 5/13/2016 8:01 Male 59 YUP1 PICT0305 Bush Island 8/31/2014 15:58 Female 25 YUP1 PICT0049 Bush Island 7/28/2015 18:14 Female 25
154 APPENDIX A Continued Site Image ID Habitat Date Time Sex Indiv. ID YUP2 EK000485 Bush Island 11/4/2014 1:20 Male 24 YUP9 EK000532 Gallery 11/16/2015 17:25 Male 24 YUP9 EK000890 Gallery 1/22/2016 22:50 Female 23 YUP9 EK001011 Gallery 2/4/2016 5:47 Male 24 YUP9 EK001103 Gallery 2/19/2016 1:14 Male 24 YUP9 EK001128 Gallery 2/21/2016 1:25 Male & Female 23 & 24 YUP9 EK001146 Gallery 2/25/2016 23:37 Male 24
155 APPENDIX B R CODE FROM CHAPTER 1 STATISTICAL ANALYSIS jag=read.csv("Jag_trap2.csv", header=T) head(jag) cor.test(jag$Trap.Nights, jag$Rel..Abund) hist(jag$Rel..Abund) cor.test( jag$Elevation,jag$Habitat) plot(jag$Elevation,jag$Habitat, label=T) head(jag) kruskal.test(Rel..Abund~Habitat, data=jag) warnings() boxplot(Rel..Abund~Habitat, data=jag) install.packages("FSA", dep=T) library(FSA) test=dunnTest(Rel..Abund~Habitat, data= jag) test ###test for differences in RAI between trail types jag=read.csv("Jag_trap2.csv", header=T) head(jag) head(jag) kruskal.test(Rel..Abund~Trail.Type, data=jag) boxplot(Rel..Abund~Trail.Type, data=jag) install.packages("FSA", dep=T) library(FSA ) test=dunnTest(Rel..Abund~Trail.Type, data=jag) test ## Use linear modeling (logistic regression) corjag=jag[c(14,16:21)] head(corjag) corjag2=as.matrix(corjag) cor(corjag2)
156 APPENDIX B continued ?corr # Null model null_mod=glm(Presence~1, family="binomial", data=jag) summary(null_mod) # Full model full_mod=glm(Presence~Elevation+Bush.Island+Gallery+Lowland+Montane+Savanna+Riverin e+Hunter_pa, family="binomial", data=jag) summary(full_mod) # Remove riverine(not enough presences to include) mo d_vr1=glm(Presence~Elevation+Bush.Island+Gallery+Lowland+Montane+Savanna+Hunter_ pa, family="binomial", data=jag) summary(mod_vr1) # Remove montane mod_vr2=glm(Presence~Elevation+Bush.Island+Gallery+Lowland+Savanna+Hunter_pa, family="binomial", data=jag) su mmary(mod_vr2) #Remove lowland mod_vr3=glm(Presence~Elevation+Bush.Island+Gallery+Savanna+Hunter_pa, family="binomial", data=jag) summary(mod_vr3) #Remove gallery mod_vr4=glm(Presence~Elevation+Bush.Island+Savanna+Hunter_pa, family="binomial", data=jag) summary(mod_vr4) exp( 0.621378)/(1+exp( 0.621378)) #Remove bush island mod_vr5=glm(Presence~Elevation+Savanna+Hunter_pa, family="binomial", data=jag) summary(mod_vr5) AIC(mod_vr5) stepAIC(full_mod) model.avg(mod_vr5,mod_vr4,mod_vr3) library(AICcmodavg) selection=model.sel(null_mod,full_mod,mod_vr1,mod_vr2,mod_vr3, mod_vr4, mod_vr5, rank="AIC") as.data.frame(selection) ??model.sel ## Jaguar Activity Patterns library(overlap)
157 APPENDIX B continued data("kerinci") head(kerinci) class(kerinci$Zone) #conve rt decimal time to radians timeRad < kerinci$Time 2 pi tig2 < timeRad[kerinci$Zone == 2 & kerinci$Sps == 'tiger'] densityPlot(tig2, rug=TRUE) #Test with jaguar data jags=read.csv("jag_caps_hab.csv", header=T) head(jags) jtime_rad=jags$Time_dec 2 pi #Activity patterns on roads vs. other trail types in lowland #Subset data, calculate time in radians for data subset jag_lowland=subset(jags, jags$Habitat=="Lowland") jag_low_time_rad=jag_lowland$Time_dec 2 pi #Roads jag_roads < jag_low_time_ rad[jag_lowland$Trail_type=="Road" & jag_lowland$Sp_ID == 'JAG'] densityPlot(jag_roads, rug=T,adjust=.5, main="Lowland Roads", ylim=c(0,0.1)) #Other trail types jag_other< jag_low_time_rad[jag_lowland$Trail_type!="Road" & jag_lowland$Sp_ID == 'JAG'] den sityPlot(jag_other, rug=T,adjust=.5, main="Lowland Other", ylim=c(0,0.1)) #compare roads versus other trail types -the overlap par(mfrow=c(1,1)) min(length(jag_roads),length(jag_other)) lowland_trail_ovr=overlapEst(jag_roads,jag_other, type="Dhat1") over lapPlot(jag_roads,jag_other,adjust=.5, main="", ylim=c(0,.08)) legend('topleft', c("Roads", "Other"), lty=c(1,2), col=c(1,4), bty='n') #Bootstrap #Resample roads roads_boot < resample(jag_roads, 10000) #Resample females other_boot < resample(jag_other, 10000) #Bootstrap jag_roads_other_boot < bootEst(roads_boot, other_boot, type="Dhat1") # takes a few seconds #Calculate mean overlap ( BSmean < mean(jag_roads_other_boot) )
158 APPENDIX B continued #generate CIs (use basic0) bootCI(lowland_trail_ovr, jag_r oads_other_boot) #Activity patterns on creek beds vs. other trail types in gallery #Roads jag_gallery=subset(jags, jags$Habitat=="Gallery") jag_gal_time_rad=jag_gallery$Time_dec 2 pi #Creek beds jag_creek < jag_gal_time_rad[jag_gallery$Trail_type==" Dry Creek Bed" & jag_gallery$Sp_ID == 'JAG'] densityPlot(jag_creek, rug=T,adjust=.5, main="Gallery Creek Beds", ylim=c(0,0.1)) #Other trail types jag_gal_other< jag_gal_time_rad[jag_gallery$Trail_type!="Dry Creek Bed" & jag_gallery$Sp_ID == 'JAG'] densi tyPlot(jag_gal_other, rug=T,adjust=.5, main="Gallery Other", ylim=c(0,0.1)) #compare creek beds versus other trail types -the overlap par(mfrow=c(1,1)) min(length(jag_creek),length(jag_gal_other)) gallery_trail_ovr=overlapEst(jag_creek,jag_gal_other, typ e="Dhat1") overlapPlot(jag_creek,jag_gal_other,adjust=.5, main="", ylim=c(0,.11)) legend('topleft', c("Creek Beds", "Other"), lty=c(1,2), col=c(1,4), bty='n') #Bootstrap #Resample creek beds creek_boot < resample(jag_creek, 10000) #Resample other gal_other_boot < resample(jag_gal_other, 10000) #Bootstrap jag_creek_other_boot < bootEst(creek_boot, gal_other_boot, type="Dhat1") # takes a few seconds #Calculate mean overlap ( BSmean < mean(jag_creek_other_boot) ) #generate CIs (use basic0) bootCI( gallery_trail_ovr, jag_creek_other_boot) #Hunting pressure in lowland and gallery habitats #Subset data low_hab=subset(jags,jags$Habitat=="Lowland") low_time_rad=low_hab$Time_dec 2 pi #density plot for captures in combined lowland and gallery habitats where hunted jag_low_hunted < low_time_rad[low_hab$Hunted == "Yes" & low_hab$Sp_ID == 'JAG'] densityPlot(jag_low_hunted, rug=T,adjust=.5, main="Lowland hunted", ylim=c(0,0.1)) #density plot for captures in combined lowland and gallery habitats where NOT hunted jag_low_no_hunt < low_time_rad[low_hab$Hunted == "No" & low_hab$Sp_ID == 'JAG']
159 APPENDIX B continued densityPlot(jag_low_no_hunt, rug=T,adjust=.5, main="Lowland not hunted", ylim=c(0,0.1)) #compare hunted versus not hunted in lowland habitat par( mfrow=c(1,1)) min(length(jag_low_hunted),length(jag_low_no_hunt)) low_hunt_no_hunt=overlapEst(jag_low_hunted,jag_low_no_hunt, type="Dhat1") overlapPlot(jag_low_hunted, jag_low_no_hunt,adjust=.5, main="") legend('topright', c("Hunted", "Not Hunted"), lty=c( 1,2), col=c(1,4), bty='n') #Bootstrap #Resample gallery low_hunt_boot < resample(jag_low_hunted, 10000) #Resample females low_nohunt_boot < resample(jag_low_no_hunt, 10000) #Bootstrap jag_lowhunting_boot < bootEst(low_hunt_boot, low_nohunt_boot, type="Dhat1") # takes a few seconds #Calculate mean overlap ( BSmean < mean(jag_lowhunting_boot) ) #generate CIs (use basic0) bootCI(low_hunt_no_hunt, jag_lowhunting_boot) par(mfrow=c(3,2)) #density plot for captures in Bush Island habitat jag_BI < jt ime_rad[jags$Habitat == "Bush Island" & jags$Sp_ID == 'JAG'] densityPlot(jag_BI, rug=T,adjust=.5, main="Bush Island", ylim=c(0,0.1)) #Lowland jag_Low < jtime_rad[jags$Habitat == "Lowland" & jags$Sp_ID == 'JAG'] densityPlot(jag_Low, rug=T,adjust=.5, main=" Lowland", ylim=c(0,0.1)) #Montane jag_Mon < jtime_rad[jags$Habitat == "Montane" & jags$Sp_ID == 'JAG'] densityPlot(jag_Mon, rug=T,adjust=.5, main="Montane ", ylim=c(0,0.1)) #Gallery jag_Gal < jtime_rad[jags$Habitat == "Gallery" & jags$Sp_ID == 'JAG'] den sityPlot(jag_Gal, rug=T,adjust=.5, main="Gallery", ylim=c(0,0.1)) #Riverine jag_Riv < jtime_rad[jags$Habitat == "Riverine" & jags$Sp_ID == 'JAG'] densityPlot(jag_Riv, rug=T,adjust=.5, main="Riverine", ylim=c(0,0.1)) #Savanna jag_Sav < jtime_rad[ jags$Habitat == "Savanna" & jags$Sp_ID == 'JAG'] densityPlot(jag_Sav, rug=T,adjust=.5, main="Savanna", ylim=c(0,0.1)) #compare gallery and riverine
160 APPENDIX B continued par(mfrow=c(1,1)) min(length(jag_Gal),length(jag_Riv)) galriv=overlapEst(jag_Gal,jag _Riv, type="Dhat4") overlapPlot(jag_Gal, jag_Riv,adjust=.5, main="") legend('topleft', c("Gallery", "Riverine"), lty=c(1,2), col=c(1,4), bty='n') #Bootstrap #Resample gallery Gal_boot < resample(jag_Gal, 10000) #Resample females Riv_boot < resample(jag_ Riv, 10000) #Bootstrap jag_galriv < bootEst(Gal_boot, Riv_boot, type="Dhat4") # takes a few seconds #Calculate mean overlap ( BSmean < mean(jag_galriv) ) #generate CIs (use basic0) bootCI(galriv, jag_galriv) #compare lowland and montane par(mfrow=c(1,1 )) min(length(jag_Low),length(jag_Mon)) ##Less than 50 samples, use Dhat1 lowmon=overlapEst(jag_Low,jag_Mon, type="Dhat1") overlapPlot(jag_Low, jag_Mon,adjust=.5, main="") legend('topleft', c("Lowland", "Montane"), lty=c(1,2), col=c(1,4), bty='n') #Bootstrap #Resample gallery Low_boot < resample(jag_Low, 10000) #Resample females Mon_boot < resample(jag_Mon, 10000) #Bootstrap jag_lowmon < bootEst(Low_boot, Mon_boot, type="Dhat1") # takes a few seconds #Calculate mean overlap ( BSmean < mean(jag _lowmon) ) #generate CIs (use basic0) bootCI(lowmon, jag_lowmon) #compare lowland and riverine par(mfrow=c(1,1)) min(length(jag_Low),length(jag_Riv)) ##Less than 50 samples, use Dhat1 lowmon=overlapEst(jag_Low,jag_Riv, type="Dhat1") overlapPlot(jag_Low, jag_Riv,adjust=.5, main="") legend('topleft', c("Lowland", "Riverine"), lty=c(1,2), col=c(1,4), bty='n')
161 APPENDIX B continued #Bootstrap #Resample gallery Low_boot < resample(jag_Low, 10000) #Resample females Riv_boot < resample(jag_Riv, 10000) #Bootstrap jag_lowriv < bootEst(Low_boot, Riv_boot, type="Dhat1") # takes a few seconds #Calculate mean overlap ( BSmean < mean(jag_lowriv) ) #generate CIs (use basic0) bootCI(lowmon, jag_lowriv) #compare lowland and bush island par(mfrow=c(1,1)) min(l ength(jag_Low),length(jag_BI)) ##Less than 50 samples, use Dhat1 lowmon=overlapEst(jag_Low,jag_BI, type="Dhat4") overlapPlot(jag_Low, jag_BI,adjust=.5, main="") legend('topleft', c("Lowland", "Bush Island"), lty=c(1,2), col=c(1,4), bty='n') #Bootstrap #R esample gallery Low_boot < resample(jag_Low, 10000) #Resample females BI_boot < resample(jag_BI, 10000) #Bootstrap jag_lowbi < bootEst(Low_boot, BI_boot, type="Dhat4") # takes a few seconds #Calculate mean overlap ( BSmean < mean(jag_lowbi) ) #generat e CIs (use basic0) bootCI(lowmon, jag_lowbi) #### Looking at between sex differences par(mfrow=c(2,1)) #Males only -all habitats (general activity pattern) jag_male < jtime_rad[jags$Sex == "Male" & jags$Sp_ID == 'JAG'] densityPlot(jag_male, rug=T,adjust=.2, main="Male Jaguar Captures", ylim=c(0,0.08)) #Females jag_female < jtime_rad[jags$Sex == "Female" & jags$Sp_ID == 'JAG'] densityPlot(jag_female, rug=T,adjust=.2, main="Female Jaguar Captures") ##Overlap test min(length(jag_male),length(j ag_female)) ## Over 50, use type=Dhat4
162 APPENDIX B continued par(mfrow=c(1,1)) mf2est=overlapEst(jag_male,jag_female, type="Dhat4") overlapPlot(jag_male, jag_female,adjust=.2, main="") legend('topleft', c("Males", "Females"), lty=c(1,2), col=c(1,4), bty='n ') #Bootstrap males maleboot < resample(jag_male, 10000) dim(maleboot) #Bootstrap females femaleboot < resample(jag_female, 10000) dim(femaleboot) jag_mac < bootEst(maleboot, femaleboot, type="Dhat4") # takes a few seconds ( BSmean < mean(jag_mac) ) #generate CIs bootCI(mf2est, jag_mac) bootCIlogit(mf2est, jag_mac)
163 APPENDIX C KEY CHARACTERISTICS OF BUSH DOG, CRAB EATING FOX, AND DOMESTIC DOG Bush Dog Crab eating Fox Domestic Dog Size: HB: 610 750 mm; T: 110 130 mm; HF: 110 120 mm; E E : 40 51 mm; WT: 5 7 kg HB: 590 765 mm; T: 22 41 mm; HF: 132 165 mm; EE: 66 80 mm; WT: 4.5 8.5 kg HT: 380 570 mm; WT: 10 18 kg, ( variable based on breed ) Legs / Feet: Legs very short, black or dark brown in coloration, webbed feet Feet and lower legs usually brown, darker than body coloration, feet not webbed Long legs and relatively large feet, in proportion with body length, feet not webbed Head: Broad, bear like head, short muzzle, small brown eyes Pointed, fox like head, muzzle darker brown than cheeks Head relatively small compared to body, long and rounded snout, coloration same as body Ears: Short, rounded ears Medium sized, thinly haired, without thick pale hair on inner rim, pointed, upright Ear shape and position variable, used to communicate emotional state, often pointed at tip, sometimes drooping Body: Small stocky dog unlike any other canid with a long cylindrical body, very short legs Medium sized fox, elongated body, relatively short legs, full tail Body length in proportion with head and legs, variable based on breed Coat: Long and soft fur, coloration varies from blonde to dark brown, usually pale brown to tawny yellow around the head and neck, darkening gradually to dark brown hindquarters and tail, sometimes white spot on chest Course fur, upper body dark grizzled grey with light to heavy black streaks, mid back with streak from head to tail, neck tawny orange, underparts cream to buff dark bristly appearance Single coat o f course guard hair, often with counter shading darker coloration on back and lighter coloration on underparts
164 APPENDIX C c ontinued Tail: Short, stumpy tail, black in coloration and thickly furred Moderately bushy, shorter than hind leg, black at the tip Thin, straight tails, generally used to communicate emotional state Call: Groups communicate with high pitched whining, yap while chasing prey Short, yapping barks Key Attribute: Short bushy tail, webbed feet, broad head, short ears, cylindrical body Long, bushy tail, relatively short legs, small feet, pointy snout Long, thin tail, relatively long legs, large feet
165 APPENDIX D SPECIES IDENTIFICATION SLIDES [ a dapted from Hunter, L., & Barrett, P. (2011). Carnivores of the world Princeton University Press New Jersey ] Priscilla Barrett Priscilla Barrett Priscilla Barrett Priscilla Barrett Priscilla Barrett Priscilla Barrett Priscilla Barrett Briton Rivire
166 APPENDIX D c ontinued [adapated from www.arkive.og]
167 APPENDIX D c ontinued [photos courtest of Matt Hallett]
168 APPENDIX E BUSH DOG INTERVIEW DATA SHEET Witness Name: ______________________________ Village: _______________________ Verbal Consent? Y / N Signed Consent: I, _______________________(name) hereby acknowledge that I have provided a truthful account of a firsthand encounter with bush dogs ( Speothos venaticus ) in the Rupununi, Guyana. I understand that this account will be reproduced in a scientific publication with proper recognition given to me by name in the form of ( Name pers. comm.). _____________________________________ __________ Date: _________________________ Date of Encounter (or nearest estimate): _______________________________________ Season: Rainy / Dry (circle one) Description of weather conditions: _______________________________________ ______________________________________________________________________________ _______ Time of day: Day / Night / Dawn / Dusk / Morning / Midday / Afternoon Anyone else that can confirm sighting? Y / N Name: __________________________________ GPS Location: _________________________________ Sighting In: Forest / Savanna Location Near To: River / Creek / Road / Trail / Village / Farm / Hunting Spot / Other Description of Location: _____________________________________________________________________________
169 APPENDI X E c ontinued How many animals: ____________________________ Length of encounter: ___________________ _______ Interactions: Vocalizations / Sniffing / Feeding / Tail wagging / Biting / Growling / Hunting Noticeable Behaviors: ______________________________________________________________________________ ______________________________________________________________________________ Description of animal: ______________________________________________________________________ ________ ______________________________________________________ ________________________ ____
170 APPENDIX F BUSH DOG INTERVIEW DATA Date Site Lat Long Location Habitat # of ind. Behavior Observer Jul 12 Iwokrama ICRCD 4.6529 58.6848 Screaming Piha trail near Iwokrama Field Station Lowland Forest 3 Running through forest away from forest edge/station clearing B. Lim / T. Horsley Aug. 2013 Iwokrama ICRCD 4.6238 58.7140 GT Lethem Road near 3 mile check point Lowland Forest 1 Crossing road M. Davis N/A Iwokrama ICRCD 4.4266 58.8055 GT Lethem road at Bamboo Creek Lowland Forest 4 Crossing road B. Allicock N/A Iwokrama ICRCD 4.3834 58.8378 GT Lethem Road near Big Turu Creek Lowland Forest 4 Running alongside road K. Singh N/A Iwokrama ICRCD 4.2590 58.8620 Iwokrama forest along mtn. foot near Atta Lodge Lowland Forest 2 Pursuing Ciniculus paca P. Allicock N/A Iwokrama ICRCD 4.2583 58.8349 Iwokrama forest along mtn. foot near canopy walkway Lowland Forest 6 Pursuing Ciniculus paca P. Allicock N/A Iwokrama ICRCD 4.3876 58.8679 Iwokrama forest in Big Turu Creek, not far from GT Lethem road Lowland Forest 2 Pursuing Dasyprocta leporina K. Singh 2016 Iwokrama ICRCD 4.2490 58.9190 GT Lethem Road just after turn off for canopy walkway, photographed near bridge Lowland Forest 4 Protographed crossing GT Lethem Highway R. McDermott
171 APPENDIX F c ontinued Date Site Lat Long Location Habitat # of ind. Behavior Observer 2017 Iwokrama ICRCD 4.2460 58.9270 GT Lethem Road just after turn off for canopy walkway Lowland Forest 6 Observed crossing GT Lethem Highway G. Sway 2010 Iwokrama ICRCD 4.1832 59.0396 GT Lethem Road near south gate Lowland Forest 2 Two pups found dead alongside road, suspected road kill A. Roopsind N/A Surama Village 4.1547 58.0623 GT Lethem road near Surama Gate Lowland Forest 3 Crossing road D. Allicock 2012 Surama Village 4.1279 59.0646 GT Lethem road near Surama Junction Lowland Forest 2 Crossing road S. James / L. Haynes N/A Surama Village 4.1554 59.0855 Traditional farming areas near Surama Pond Lowland Forest 5 Pursuing Ciniculus paca M. Captain 2004 Surama Village 4.1419 59.0902 Tractor road near traditional farming grounds Lowland Forest 5 Pursuing Ciniculus paca F. Milton Sept. 2013 Surama Village 4.1526 59.1096 Tractor road near traditional farming grounds Lowland Forest 4 Crossing road K. Singh Aug. 2013 Surama Village 4.1397 59.1443 Surama forests in Pakaraima Mountains near Sabba Creek Montane Forest 6 Pursuing Ciniculus paca B. Allicock Jan. 2014 Surama Village 4.1073 59.1354 Crossing savanna from mtn. forest through Camodi Bash (swamp) Savanna 4 6 Crossing road F. Buckley N/A Surama Village 4.1250 59.0938 Running through the forest near Surama Creek Lowland Forest 3 5 Running in single file line through forest E. Allicock 2004 Surama Village 4.1308 59.1025 Road/culvert leading into Surama, over creek near forest/savanna edge Lowland Forest 4 Crossing road leading into village K. Singh
17 2 APPENDIX F c ontinued Date Site Lat Long Location Habitat # of ind. Behavior Observer 1977 Rupertee Village 3.9754 59.1233 Tractor road near traditional farming grounds Savanna 1 Crossing road E. Vanlong 1979 Rupertee Village 4.0095 59.1449 Tractor road near traditional farming grounds Montane Forest 12 Crossing road S. Andries N/A Rupertee Village 3.9969 59.1278 In savanna crossing between mountain forests Savanna 1 Crossing road M. Vanlong 1999 Rupertee Village 4.0076 59.1323 Tractor road near traditional farming grounds Montane Forest 3 6 Crossing road; Barking G. Duarte 2004 Rupertee Village 4.0095 59.1449 Tractor road near traditional farming grounds Montane Forest 4 Crossing road J. Sciptio 1994 Rupertee Village 4.0105 59.1588 Creek bed in forested area near traditional farming grounds Montane Forest 2 Crossing dry creek bed C. Vanlong 2012 Rupertee Village 4.0284 59.1369 Traditional hunting grounds in high forest far from village (back dam) Montane Forest 10 Drinking water from a small and mostly dry creek bed J. Vanlong N/A Rupertee Village 4.0239 59.1377 Forested area near bush mouth near village and traditional farming area Lowland Forest 8 Crossing road; Pursuing Ciniculus paca J. Vanlong N/A Rupertee Village 3.9754 59.1233 Found near small pond and traditional farming grounds Montane Forest 1 Found pup raised as pet for 1 year; died in captivity in Lethem J. Vanlong Apoteri Village 4.0152 58.5764 Captured along traditional balata harvesting route Lowland Forest 1 Found pup raised as pet; died in captivity S. Andries
173 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer N/A Rewa Village 3.9680 58.7990 Lowland forest area near traditional farming grounds Lowland Forest 3 6 Running through forest C. Haynes ~2010 Rewa Village 3.9730 58.8020 Lowland forest area near traditional farming grounds Lowland Forest 4 Running through the forest in a single file line, barking as they ran C. Edwards N/A Berbice River 3.9195 58.3381 Animals captured near Berbice River, held in East Bank Demerara Lowland Forest 8 Animals captured and transferred to exporter on the coast W. Melville Nov. 2013 Yakarinta Village 3.9575 59.2518 GT Lethem road near Yakarinta Junction Savanna 3 Running alongside road V. Edwards N/A Aranaputa Village 3.9044 59.4195 Near village crossing an open area between mountain forests Savanna 4 7 Pursuing Dasyprocta leporina V. Hamilton 2000 Toka Village 3.9160 59.3816 Rocky area near village, forest edge, mountain foot and small stream Savanna 5 Sleeping with pups in hole btw. rocks; Growled when disturbed K. Davis 2007 Karasabai Village (Kaibaiku) 4.1355 59.5318 Crossing a swampy area in savanna btw. montane forest and bush island Savanna 4 6 Pursuing Mazama americana K. Davis ~2010 Karasabai Village 4.1130 59.5880 Traditional hunting grounds in Pakaraima Mountains above farming grounds Montane Forest 3 Running through the forest, Sniffing along the ground A. Albert
174 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer Mar. 2012 Kwaimatta Village 3.8131 59.3006 Along Rupununi River near traditional farming grounds Gallery Forest 4 Running along Rupununi River; Some biting/growling noises H. Ambrose N/A Yupukari Village 3.6564 59.3611 Along Awarikuru Lake near traditional farming grounds Gallery Forest 3 Running in single file line along hill foot H. Ambrose 2013 Quattata Village 3.6595 59.4760 In bush island along a creek on hill top near Old Dutch Fort Bush Island 5 Wandering Sniffing along the ground; Playful interactions M. Mandook N/A Quattata Village 3.6419 59.4913 Near large bush island on hill top near Old Dutch Fort Savanna 2 Walking in open savanna along forest edge M. Mandook N/A Markanata Village 3.6420 59.5030 Marakanata bush island near bush edge and close to a pond Bush Island 3 Pursuing Mazama americana M. Mandook N/A Markanata Village 3.6370 59.4950 Middle of Markanata bush island near small creek Bush Island 1 Running through forest; Sniffing along the ground M. Mandook 2017 Markanata Village 3.6490 59.5310 Running through long grass along Pirara Creek Savanna 5 Pursuing Sus scofra sow with piglets, kill observed, individuals chased off by observer on horseback J. Fidel N/A Markanata Village 3.6620 59.5310 Savanna area between several large bush islands, ~2 mi from main road Savanna 5 Crossing road F. Li 2008 Nappi Village 3.3696 59.5110 Traditional hunting area in mountains near Maipaima Lodge Montane Forest 6 Pursuing Ciniculus paca D. Aldie
175 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer N/A Nappi Village 3.3981 59.5645 Along forest edge, bush mouth area near traditional farming grounds Savanna 4 Running in savanna along the forest edge G. Fredericks 2013 Nappi Village 3.3239 59.5175 Running along a flat hill top behind Nappi Mountain Montane Forest Running through the forest, Sniffing along the ground M. David Feb. 1983 Manari Ranch 3.4479 59.8202 Traditional hunting area along the Takutu River and Manari Creek Gallery Forest 4 Pelt of an individual shot by hunter who encountered group along creek C. Melville N/A Katoka Village 3.5270 59.2925 Along Simoni Creek in traditional hunting grounds Lowland Forest 5 Pursuing Mazama nemorivaga K. Edwards N/A Kanuku Mountains Protected Area 3.3820 59.3060 Along the Rupununi River near the mouth of Hiari Creek Gallery Forest 4 Pursuing Ciniculus paca B. Lawrence N/A Kanuku Mountains Protected Area 3.4408 59.1687 Along Simoni Creek far from any use areas Lowland Forest 3 Running along riverbank on Simoni Creek K. Mandook N/A Kanuku Mountains Protected Area 3.3358 59.2614 Along Mapari Creek in tourism area near traditional farming grounds Montane Forest 3 6 Sitting in understory vegetation along the riverbank D. DeFreitas 1976 1978 Kanuku Mountains Protected Area 3.2717 58.9512 Simoni Creek head near traditional balata, fishing and hunting grounds Lowland Forest 3 5 Pursuing Ciniculus paca R. Merriman
176 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer 2008 Moco Moco Village 3.3037 59.6358 High forest traditional hunting area near Kumaka Falls Montane Forest 1 Wandering Sniffing along the ground with head down L. Piash 2009 Moco Moco Village 3.2898 59.6816 Near traditional farming grounds south of village at Raguaga Falls Montane Forest 4 Pursuing Tupinambis teguixin K. Ambrose 2011 Moco Moco Village 3.2989 59.6714 Traditional farming grounds at Raguaga Falls Lowland Forest 1 Pursuing Ciniculus paca K. Ambrose Mar. 2013 Moco Moco Village 3.2754 59.7027 Along mountain foot near Moco Moco creek Montane Forest 3 Wandering Sniffing along the ground and in the air C.S. Williams 2012 Moco Moco Village 3.2835 59.6969 Traditional framing grounds near Moco Moco creek Lowland Forest 4 Female moving 3 pups from den site L. Campion N/A Moco Moco Village 3.3274 59.7306 Open savanna between Moco Moco and Lethem Savanna 2 Walking across savanna towards forest, recently pursued prey L. Campion 2008 Moco Moco Village 3.2661 59.7123 Along Cruza Creek in traditional hunting grounds Montane Forest 2 Wandering Sniffing along the ground and in the air L. Aldie Dec. 2013 Moco Moco Village 3.2732 59.7126 Along Cruza Creek in traditional hunting grounds Montane Forest 3 4 Vocalizations and tracks identified leading to den site L. Aldie Jan. 1985 Quarrie Village 3.2369 59.7518 Along creek near traditional farming grounds and forest/savanna border Lowland Forest 5 Pursuing Ciniculus paca F. Raymundo
177 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer N/A Quarrie Village 3.2203 59.7644 Traditional hunting area near small creek and farming grounds Lowland Forest 4 Pursuing Ciniculus paca K. Peter N/A Quarrie Village 3.2191 59.7499 Traditional hunting area near small creek and farming grounds Lowland Forest 3 5 Pursuing Ciniculus paca P. Morari N/A Parikwarnau Village (Imprenza) 3.1008 59.8033 Traditional hunting area near creek and isolated mountain Gallery Forest 2 Pursuing Pecari tajacu C. Melville N/A Parikwarnau Village (Imprenza) 3.0563 59.7980 Traditional hunting area near creek and isolated mountain Savanna 4 Crossing savanna between mountain forest and creek C. Melville N/A Shulinab Village 3.0552 59.7184 Savanna near village and bush edge, animal run down and hand caught Savanna 1 Animal captured and kept in captivity shortly and released M. Malcom ~2005 Shulinab Village 3.0810 59.5810 Traditional hunting area near farming grounds Montane Forest 8 Pursuing Ciniculus paca G. Bell ~2010 Shulinab Village 3.0860 59.5780 Traditional hunting area near farming grounds Montane Forest 4 Running through the forest in a single file line G. Bell N/A Meriwau Village 3.0675 59.6704 Traditional hunting grounds near forest edge of Kanuku Mountains Savanna 1 Pursuing Ciniculus paca G. Peter N/A Sand Creek Village 3.2017 59.4113 Along Rupunun i River and mtn. foot near traditional farming grounds Montane Forest 3 4 Pursuing Ciniculus paca A. Jackman
178 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer N/A Sand Creek Village 3.1829 59.3877 Along Rupunun i River and mtn. foot near traditional farming grounds Montane Forest 6 8 Pursuing Dasyprocta leporina near a foot trail along the Rup. R. B. Phillips 1980 Potarinau Village 3.0360 59.7663 Traditional hunting ground near farming area Lowland Forest 2 Pair mating/fighting; Young adult male captured, held in captivity J. George Kusad Mountain 2.7948 59.8421 Traditional farming areas at the foot of Kusad Mountain Lowland Forest 4 Pursuing Ciniculus paca J. George Dog mountain 2.7247 59.9401 Traditional grazing area near small mountain named for bush dogs Bush Island 5 Running along mountain foot J. George Late 1990's Katoonarib Village 2.736 59.584 Several bush dogs kept in the village as pets Bush Island 3 Several individuals held in local captivity A. Wilson 1960's Dadanawa Ranch 2.824741 59.524431 Pups found along trail to Kwitaro River; raised to adults at the ranch Lowland Forest 4 Four pups kept as pets; survived into adulthood; deposited as specimens in Natural History Museum in London ~2005 Dadanawa Ranch 2.824741 59.524431 Pups found in hollow log near Kwitaro River; kept a t ranch as pets alongside dom. d ogs Lowland Forest 3 Three pups kept as pets; did not survive until adulthood; believed to have contract Leptospirosis sp.
179 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer N/A Rupunau Village 2.9415 59.3646 Traditional hunting grounds in bush islands on hill tops near the village Bush Island 4 Pursuing Ciniculus paca L. Ignacio N/A Rupunau Village 2.9313 59.3367 Traditional hunting grounds in bush islands on hill tops near the village Bush Island 3 Wandering Sniffing along the ground and in the air L. Ignacio 1960's Shea Village 2.8280 59.0380 Pups encountered along a trail from village to traditional fishing grounds along Kwitaro R. Lowland Forest 4 Found pups raised as pets; died in captivity as adults at Dadanawa Ranch; deposited as specimens S. Brock N/A Awarunau Village 2.6654 59.1947 Animal caught in bush island on a hill near the village Bush Island 1 Found pup raised as pet; died in captivity T. Griffith N/A Maruranau Village 2.7662 59.2021 Animal found near village Savanna 1 Animal found near village, kept in local captivity for some time G. Pereira N/A Kwitaro River 2.8343 59.5156 Animals found in a hollow log near the Kwitaro River Lowland Forest 3 Found pups raised as pets; died in captivity at Dadanawa Ranch D. DeFreitas N/A Bamboo Creek (Rewa River) 3.1512 58.6263 Lowland forest along river bank ~70 miles up Rewa River Lowland Forest 3 4 Standing on high riverbank along the edge of Bamboo Creek D. Laurentino 2005 Rewa River 2.7825 58.6186 Lowland forest along river bank near the head of the Rewa River Lowland Forest 8 Standing near Rewa River on a small sand bank A. Holland N/A Rewa River 2.7220 58.5990 Lowland forest along river bank near the head of the Rewa River Lowland Forest 3 5 Running along riverbank on Rewa River D. DeFreitas
180 APPENDIX F continued Date Site Lat Long Location Habitat # of ind. Behavior Observer N/A DTL Logging Concession 4.5757 58.5581 Logging road in privately held concession near Essequibo River Lowland Forest 3 6 Crossing logging road H. James 2008 Mango Landing (Region 8) 5.3080 58.9360 Found along mining access road as pups, given to wildlife trader with facility on the East bank of the Demerara Lowland Forest 4 Bush dogs discovered as pups, adult not observed nearby; Observed being held by an international wildlife trader M. Pierre N/A Paramakatoi Village (Region 8) 4.7777 59.6213 Animal captured in Pakaraima Mtns., held in Paramakatoi Village Montane Forest 1 Found pup raised as pet; died in captivity E. Edwin 2015 Bai Shan Lin Concession (Region 10) 4.1539 58.1773 Lowland forest within Bai Shan Lin logging concession Lowland Forest 3 Crossing logging road during biodiversity survey L. Ignacio N/A Dubulay Ranch (Region 10) 5.6680 57.8880 Open intermediate savanna in Dubulay Ranch Savanna 2 Running through open savanna in the middle of the day A. Mendes N/A Berbice Road (Region 10) 5.6370 58.2830 Lowland forest along Berbice Road Lowland Forest 2 Crossing the Berbice Road A. Mendes N/A Berbice Road (Region 10) 5.6060 58.1780 Berbice Road near Dubulay Ranch Lowland Forest 1 Adult bush dog found dead along road, suspected road kill A. Mendes
181 APPENDIX G IDENTIFICATION OF KNOWN INDIVIDUAL ANTEATERS FROM AZA INSTITUTIONS Institution Anteater Name Sex # of photos Positive ID (Y/N) Reason for incorrect ID Photo Credits Brevard Zoo Abner Male 3 Yes n/a Kerry Sweeney Brevard Zoo Boo Female 4 Yes n/a Kerry Sweeney Brookfield Zoo Lupito Male 2 Yes n/a n/a Buffalo Zoo Delilah Female 2 Yes n/a n/a Buffalo Zoo Haji Male 2 Yes n/a n/a Cleveland Metroparks Zoo Amendi Male 2 Yes n/a n/a Cleveland Metroparks Zoo Kutter Male 2 Yes n/a n/a Cleveland Metroparks Zoo Pica Female 2 Yes n/a n/a Dallas Zoo Jimi Male 2 Yes n/a n/a Dallas Zoo Tullah Female 2 Yes n/a n/a Fresno Chaffee Zoo Caliente Male 2 Yes n/a Meghan Kelly Greensboro Science Center Eury Male 5 Yes n/a n/a Houston Zoo Olive Female 3 Yes n/a n/a Houston Zoo Pablo Male 3 Yes n/a n/a Houston Zoo Rio Female 4 Yes n/a n/a Jacksonville Zoo & Gardens Killroy Male 4 Yes n/a John Reed Jacksonville Zoo & Gardens Stella Female 7 Yes n/a John Reed Palm Beach Zoo Cruz Male 3 Yes n/a n/a Palm Beach Zoo Odelia Female 3 Yes n/a Nancy Nill Phoenix Zoo Beaker Male 11 Yes n/a Carrie Flood
182 APPENDIX G continued Institution Anteater Name Sex # of photos Positive ID (Y/N) Reason for incorrect ID Photo Credits Potawatomi Zoo Jo Hei Male 7 Yes n/a n/a Reid Park Zoo Curl Tail Male 3 Yes n/a n/a Reid Park Zoo Sophia Female 3 Yes n/a n/a Sacramento Zoo Amber Female 2 Yes n/a n/a San Antonio Zoo Sprout Female 5 No Low pixel images n/a San Antonio Zoo Humphrey Male 7 Yes n/a n/a Sedgewick County Zoo Ibini Female 5 Yes n/a Steve Jones Sedgewick County Zoo Matteo Male 4 Yes n/a n/a Sunset Zoo Angelina Female 5 Yes n/a n/a Turtle Back Zoo Barques Male 15 No Low pixel images n/a Zoo Boise McCauley Male 2 Yes n/a n/a Zoo Boise Gloria Female 2 Yes n/a n/a TOTALS 32 18 M 14 F 128 n/a
183 LIST OF REFERENCES Adams, W. M., Aveling, R., Brockington, D., Dickson, B., El liott, J., Hutton, J., Roe, D., Vira, B., & Wolmer, W. (2004). Biodiversity conservation and the eradication of poverty. S cience 306 (5699), 1146 1149. Adams, W. M. 1990. Green development: environment and sustainability in the Third World Routledge, London, UK. Agrawal, A., & Gibson, C.C. ( 1999 ) E nchantment and Disenchantment: The Role of Community in Natural Resource Conserv ation. World Development 27(4), 629 649. Aikenhead, G. S., & Ogawa, M. (2007). Indigenous knowledge and science revisited. Cultural Studies of Science Education 2 (3), 539 620. Akaike, H. (1981). Likelihood of a model and information criteria. Journal of E conometrics 16 (1), 3 14. Alonso, L.E., McCullough, J., Naskrecki, P., Alexander, E., Wright, H.E. (2008) A rapid biological assessment of the Konashen Community owned Conservation Area, Southern Guyana. Conservation International Rapid Assessment Program (RAP) Working Papers 51, 1 91. Anad n, J. D., Gimnez, A., Ballestar, R., & Prez, I. (2009). Evaluation of local ecological knowledge as a method for collecting extensive data on animal abundance. Conservation Biology 23 (3), 617 625. Aquino, R. & Puertas, P. (1997). Observations of Speot hos venaticus (Canidae: Carnivora) in its natural habitat in Peruvian Amazonia. Zeitschrift fur Saugetierkunde 62 117 118. Aranda, M., & Snchez Cordero, V. (1996). Prey spectra of jaguar ( Panthera onca ) and puma ( Puma concolor ) in tropical forests of Mexico. Studies on Neotropical Fauna and Environment 31 (2), 65 67. Arnstein, S. R. (1969). A ladder of citizen participation. Journal of the American Institute of planners 35 (4), 216 224. Arzoumanian, Z., Holmberg, J., & Norman, B. (2005). An astronomi cal pattern matching algorithm for computer aided identification of whale sharks Rhincodon typus Journal of Applied Ecology 42 (6), 999 1011. Astete, S., Sollmann, R., & Silveira, L. (2008). Comparative ecology of jaguars in Brazil. Cat News 42 9 14.
184 Au, J., Bagchi, B., Chen, B., Martinez, R., Dudley, S. A., & Sorger, G. J. (2000). Methodology for public monitoring of total coliforms, Escheri coli and toxicity in waterways by Canadian high school students. Journal of Environmental Management 58, 213 2 30. Bailey, I., Myatt, J. P., & Wilson, A. M. (2013). Group hunting within the Carnivora: physiological, cognitive and environmental influences on strategy and cooperation. Behavioral Ecology and Sociobiology 67 (1), 1 17. Baird, I.G., Flaherty, M.S. & B aird, G. (2005) Mekong River fish conservation zones in Southern Laos: Assessing effectiveness using local ecological knowledge. Environmental Management 36 (3), 439 454. Bandura, A. (2001). Social cognitive theory of mass communication. Media psych ology 3 (3), 265 299. Barber, J. R., Crooks, K. R., & Fristrup, K. M. (2010). The costs of chronic noise exposure for terrestrial organisms. Trends in ecology & evolution 25 (3), 180 189. Barnett, A., Shapley, R. & Engstrom, M. (2001) Records of the bus hdog, Speothos venaticus (Lund, 1842), from Guyana. Mammalia 65, 232 237. Barsh, R. L. ( 1997 ) Forests, indigenous peoples, and biodiversity. Global Biodiversity (Canadian Museum of Nature) 7 (2) 20 24. Baum, F., MacDougall, C., & Smith, D. (2006). Participatory action research. Journal of Epidemiology & Community Health 60 (10), 854 857. Becker, C. D., & Ghimire, K. (2003). Synergy between traditional ecological knowledge and conservation science supports forest preservation in Ecuador. Conservatio n E cology 8 (1). Beisiegel, B.M. (2009) First camera trap record of bush dogs in the state of Sao Paulo, Brazil. Canid News 12 (5), 1 5. Beisiegel, B.M. & Ades, C. (2002) The behavior of the bush dog ( Speothos venaticus Lund 1842) in the field: a review. Revista de Etologia 4, 17 24. Beisiegel, B. M. & Zuercher, G.L. (2005) Speothos venaticus Mammalian Species, 783, 1 6. Berkes, F., Colding, J., Folke C. ( 2000 ). Rediscovery of Traditional Ecological Kno wledge as Adaptive Management. Ecological Applications 10(5), 1251 1262. Berkes, F. Colding, J., Folke, C. ( 2003 ). Navigating social ecological systems: building resilience for complexity and change Cambridge University Press, Cambridge, U.K.
185 Berkes, F. ( 2004 ). Rethinking Community based Conservation. Conservation Biology 18 (3), 621 630. Berkes, F. (2009). Indigenous ways of knowing and the study of environmental change. Journal of the Royal Society of New Zealand, 39:4, 151 156. DOI: 10.1080/030142209095105 68 Bertassoni, A., Mouro, G., Ribeiro, R. C., Cesrio, C. S., Oliveira, J. P. D., & Bianchi, R. D. C. (2017). Movement patterns and space use of the first giant anteater ( Myrmecophaga tridactyla ) monitored in So Paulo State, Brazil. Studies on Neotropical Fauna and Environment 52 (1), 68 74. Biddulph, J., & Kellman, M. (1998). Fuels and fire at savanna gallery forest boundaries in southeastern Venezuela. Journal of Tropical Ecology 14 (4), 445 461. Biernacki, P., & Waldorf, D. (1981) Snowball sampling: Problems and techniques of chain referral sampling. Sociological Methods & Research 10 (2), 141 163. Blake, J. G., & Mosquera, D. (2014). Camera trapping on and off trails in lowland forest of eastern Ecuador: does location matter?. Mastozoolo ga N eotropical 21 (1), 17 26. Blake, J. G., Mosquera, D., Loiselle, B. A., Romo, D., & Swing, K. (2017). Effects of human traffic on use of trails by mammals in lowland forest of eastern Ecuador. Neotropical Biodiversity 3 (1), 57 64. Bolger, D.T., Morr ison, T.A., Bennet, V., Lee, D., Farid, H. (2012). A computer assisted system for photographic mark recapture analysis. Methods in Ecology and Evolution, 3, 813 822. Bradshaw, G.A. & Bekoff M. ( 2001 ) Ecology and social responsibility: the re embodiment of science. Trends in Ecology and Evolution 16, 460 465. Braga, F. G., Santos, R. E., & Batista, A. C. (2010). Marking behavior of the giant anteater Myrmecophaga tridactyla (Mammalia: Myrmecophagidae ) in Southern Brazil. Zoologia (Cur itiba) 27 (1), 07 12. Brashares, J. S., Arcese, P., & Sam, M. K. (2001). Human demography and reserve size predict wildlife extinction in West Africa. Proceedings of the Royal Society of London B: Biological Sciences 268 (1484), 2473 2478. Brewer, C. ( 2001 ). Cultivating conservation liter acy: down education is not enough. Conservation Biology 15(1), 1203 1205. Brewer, C. ( 2002 ). Outreach and partnership programs for conservation education where endangered species c onservation and research occur. Conservation Biology 16 4 6.
186 Brosius, J. P., Tsing, A. L., & Zerner, C. (1998). Representing communities: Histories and politics of community based natural resource management. Society & Natural Resources 11(2). http://www.tandfonline.com/doi/abs/10.1080/08941929809381069 Brossard, D., Lewenstein, B., & Bonney, R. (2005). Scientific knowledge and attitude change: The impact of a citizen science project. International Journal of Science Education 27 (9), 1099 1121 Brown, W. A. (2002). Inclusive governance practices in nonprofit organizations and implications for practice. Nonprofit Management and Leadership 12 (4), 369 385. Bureau of Statistics. (2012) Population and housing census. Census 2012 preliminary report http://www.statisticsguyana.gov.gy/pubs.html Downloaded on 2 February 2016 Camilo Aves, C. de S.P., Mouro, G. D. M. (2006). Responses of a specialized insec tivorous mammal ( Myrmecophaga tridactyla ) to variation in ambient temperature. Biotropica 38 (1), 52 56. Carbone, C., Christie, S., Conforti, K., Coulson, T., Franklin, N., Ginsberg, J.R., Griffiths, M., Holden, J., Kawanishi, K., Kinnaird, M., Laidlaw, R ., Lynam, A., Macdonald, D.W., Sunquist, M., Tilson, R., Wan Shahruddin, W.N. (2001). The use of photographic rates to estimate densities of tigers and other cryptic mammals. Animal Conservation, 4, 75 79. Carlsson, L. ( 2000 ) Policy networks as collective action. Policy Studies Journal 28, 502 520. Carretero Pinzn, X. (2013). Bush dog sighting in a large forest fragment in the Colombian Llanos. Canid Biology and Conservation 16 (5), 16 17. Caso, A., Lopez Gonzalez, C., Payan, E., Eizirik, E., de Oliveira, T., Leite Pitman, R., Kelly, M. & Valderrama, C. ( 2008 ) Panthera onca The IUCN Red L ist of Threatened Species 2008: http://dx.doi.org/10.2305/IUCN.UK.2008.R LTS.T15953A5327466.en Downloaded on 25 September 2017 Cavalcanti, S. M., & Gese, E. M. (2010). Kill rates and predation patterns of jaguars ( Panthera onca ) in the southern Pantanal, Brazil. Journal of Mammalogy 91 (3), 722 736. Chambers, R. E., & McBeth, M. K. (1992). Community encouragement: Returning to the basis for community development. Community Development 23 (1), 20 38. Chalmers, N., & Fabricius, C. (2007). Expert and generalist local knowledge about land cover d Coast: can local ecological knowledge add value to science? Ecology and Society 12 (1).
187 Chape, S., Harrison, J., Spalding, M., & Lysenko, I. (2005). Measuring the extent and effectiveness of protected areas as an indicator for meeting global biodiversity targets. Philosophical Transactions of the Royal Society of London B: Biological Sciences 360 (1454), 443 455. Chavis, D. M., Stucky, P. E., & Wandersman, A. (1983). Returning basic research to the community: A relationship between scientist and citizen. American Psychologist 38 (4), 424. Child, B. (1996). The practice and principles of community based wildlife management in Zimbabwe: the CAMPFIRE programme. Biodiversity & Conservation 5 (3), 369 398. Colfer, C. J. P., Andriamampandry, E., A saha, S., Lyimo, E., Martini, E., Pfund, J. L., & Watts, J. (2011). Participatory Action Research for Catalyzing Adaptive Management: Analysis of a" Fits and Starts" Process. Journal of Environmental Science and Engineering 5 (1). Conde, D. A., Colchero, F., Zarza, H., Christensen, N. L., S exton, J. O., Manterola, C., Ch vez, C., Rivera, A., Azuara, D. & Ceballos, G. (2010). Sex matters: Modeling male and female habitat differences for jaguar conservation. Biological Conservation 143 (9), 1980 1988. Connor, D. M. (1988). A new ladder of citizen participation. National Civic Review 77 (3), 249 257. Conrad, C. C., & Hilchey, K. G. (2011). A review of citizen science and community based environmental monitoring: issues and opportunities. Environmental monitoring and assessment 176 (1), 273 291. Cooper, C., Dickinson, J., Phillips, T., & Bonney, R. (2007). Citizen science as a tool for conservation in residential ecosystems. Ecology and Society 12 (2). Crabtree, R. D., & Sapp, D. A. (2005). Technical Communication, Participato ry Action Research, and Global Civic Engagement: A Teaching, Research, and Social Action Collaboration in Kenya. REFLECTIONS: A journal of rhetoric, civic writing and service learning 4 (2). Crawford, F.D., Szelewski C.E., Alvey, G.D. (1985) Geology and e xploration in the Takutu Graben of Guyana Brazil. Journal of Petroleum Geology 8 (1), 5 36. Da Silveira, R., Ramalho, E. E., Thorbjarnarson, J. B., & Magnusson, W. E. (2010). Depredation by jaguars on caimans and importance of reptiles in the diet of jaguar. Journal of Herpetology 44 (3), 418 424. Daly, L. (2015) The symbiosis of people and plants: Ecological engagements among the Makushi Amerindians of Amazonian Guyana D octoral thesis, University of Oxford.
188 Davis, M. L., Kelly, M. J., & Stauffer, D. F. (2011). Carnivore co existence and habitat use in the Mountain Pine Ridge Forest Reserve, Belize. Animal Conservation 14 ( 1), 56 65. DeMatteo, K.E. (2008) Using a survey of carnivore conserva tionists to gain new insight into the ecology and conservation status of the bush dog. Canid News 11.3. DeMatteo, K.E. & Kochanny, C. (2004) Determining an effective and safe radio tracking collar for bush dogs ( Speothos venaticus ). Canid News 7.7. DeMatteo, K. E. & Loiselle, B.A. (2008) New data on the status and distribution of the bush dog ( Speothos venaticus ): evaluating its quality of protection and directing research efforts. Biological Conservation, 141 2494 2505. DeMatteo, K.E., Michalski F. & Leite Pitman, M.R.P. (2011) Speothos venaticus IUCN Red List of Threatened Species v. 2013.2 Http://www.iucnredlist.org [accessed 17 February 2014] DeMatteo, K. E., Porton, I.J., Kleiman, D.G. & Asa, C.S. (2006) The effect of the male bush dog ( Speothos venaticus ) on the female reproductive cycle. Journal of Mammalogy 87 723 732. DeMatteo, K. E., Rinas, M.A., Sede, M.M., Davenport, B., Argelles, C., Lovett, K. & Parker, P.G. (2009) Detector dogs: an effective technique for bush dogs ( Speot hos venaticus ) surveys. The Journal of Wildlife Management 73 1436 1440. DeMatteo, K.E., Rinas, M.A., Argelles, C.F., Zurano, J.P., Selleski, N., Di Bittetti, M.S., Eggert, L.S. (2014) Noninvasive techniques provide novel insights for the elusive bush dog (Speothos venaticus). Wildlife Society Bulletin 38.4: 862 873. Desbiez, A.L.J., & Medri, . M. (2010). Density and habitat use by giant anteaters ( Myrmecophaga tridactyla ) and souther n tamanduas ( Tamandua tetradactyla ) in the Pantanal wetland, Brazil. Edentata 11 (1), 4 10. Desbiez, A.L.J. & Kluyber, D. (2013) The Role of giant armadillos ( Priodontes maximus ) as physical ecosystem engineers. Biotropica 45, 537 540. de Oliveira T. G. (2009) Distribution, habitat utilization and conservation of the Vulnerable bush dog Speothos venaticus in Northern Brazil. Oryx 43 (2), 247 253. de Oliveira, T.G. (2011 ) Ecologia e conservao de pequenos felinos no Brasil e suas implicaes para o manejo. PhD thesis. Universidade Federal de Minas Gerais, Belo Horizonte, Brazil de Oliveira T. G., Michalski, F., Botelho, A. L., Michalski, L. J., Calouro, A. M., & Desbiez, A. L. (2016). How rare is rare? Quantifying and assessing the rarity of the bush dog Speothos venaticus across the Amazon and other biomes. Oryx 1 10.
189 de Souza Lima, E., DeMatteo, K.E., Jorge, R.S.P., Cesar Dalponte, J., Souza Lima, H. & Klorfine, S.A. (2012) First telemetry study of bush dogs: home range, activity and habitat selection. Wildlife Research 39. 512 519. Daz, W. (2002). The v egetation a long the Lower Kwitaro River on the e astern e dge of the Kanuku Mountain Region, Guyana. Conservation International Guyana, Georgetown, Guyana. Di Blanco, Y. E., Jimnez Prez, I., & Di Bitetti, M. S. (2015). Habitat selection in reintroduced giant anteaters: the critical role of conservation areas. Journal of Mammalogy 96 (5), 1024 1035. Di Blanco, Y. E., Desbiez, A. L., Jimnez Prez, I., Kluyber, D., Massocato, G. F., & Di Bi tetti, M. S. (2017). Habitat selection and home range use by resident and reintroduced giant anteaters in 2 South American wetlands. Journal of Mammalogy 98(4) 1118 1128 Dickinson, J. L., Shirk, J., Bonter, D., Bonney, R., Crain, R. L., Martin, J., ... & Purcell, K. (2012). The current state of citizen science as a tool for ecological research and public engagement. Frontiers in Ecology and the Environment 10 (6), 291 297. Dickman, A.J. (2008) Key determinants of conflict between people and wildlife, particularly large carnivores, around Ruaha National Park, Tanzania PhD Thesis. University College of London and Institute of Zoology, Zoological Society of London, London, U.K. Dinerstein, E. (2013). Scent of an Anteater. The Kingdom of Rarities, Island Press/Center for Resource Economics, Washington D.C., pp. 151 178). Donovan, D., & Puri, R. (2004). Learning from trad itional knowledge of non timber forest products: Penan Benalui and the autecology of Aquilaria in Indonesian Borneo. Ecology and Society 9 (3). Dunn, O.J ( 1964 ) Multiple comparisons using rank sums. Technometrics 6: 241 252. Durigan, G., Siqueira, M. F. D., & Franco, G. A. D. C. (2007). Threats to the Cerrado remnants of the state of So Paulo, Brazil. Scientia Agricola 64 (4), 355 363. Eden, M.J., McGregor, D.F.M. (1992). Dynamics of the forest savanna boundary in the Rio Branco Rupununi region of northern Amazonia. In: Furley P A Proctor J., Ratter J.A., (Eds.). Nature and dynamics of forest savanna boundaries. London (UK): Chapman and Hall pp 77 88. Efford, M. G., Dawson, D. K., & Robbins, C. S. (2004). DENSITY: software for analyzing capture recapture data from passive detector arrays. Animal Biodiversity and Conserv ation 27 (1), 217 228.
190 Eisenberg, J. F. (1989). Mammals of the N eotropics: Panama, Colombia, Venezuela, Guyana, Suriname, French Guiana 1 Chicago: University of Chicago Press. Eizirik, E., Kim, J. H., Menotti rien, S. J., & Johnson, W. E. (2001). Phylogeography, population history and conservation genetics of jaguars ( Panthera onca Mammalia, Felidae). Molecular Ecology 10 (1), 65 79. Emmons, L. H. (1987). Comparative feeding ecology of felids in a neotropical rainforest. Behavioral E cology and S ociobiology 20 (4), 271 283. Eisenberg, J.F. (1989) Mammals of the Neotropics The Northern Neotropics, Vol. 1: Panama, Colombia, Venezuela, Guyana, Suriname, French Guiana University of Chicago Press, Chicago, US A. Engels, C. A., & Jacobson, S. K. (2007). Evaluating long term effects of the golden lion tamarin environmental education program in Brazil. The Journal of Environmental Education 38 (3), 3 14. Engstrom, M.D. & Lim, B.K. (2002) Diversity and conservat ion of mammals of Guyana. In: Cabellos, G.Y., Simonetti, J. (Eds.). Diversidad y conservacion de los mamiferos neotropicales CONABIO & UNAM, Mexico. Engstrom, M.D. & Lim B. K. ( 2008 ) Checklist of the Mammals of Guyana. http://www.mnh.si.edu/biodiversity/bdg/guymammals.html Downloaded on 25 September 2017. Environmental Protection Agency, Guyana (EPA). (2009) Environmental Protection (Wildlife Management and Conservation) Regulations Georgetown, Guyana. ESRI. (2016). ArcGIS Desktop: Release 10. 4.1 Environme ntal Systems Research Institute Redlands, CA (USA) Evans, C., Abrams, E., Reitsma, R., Roux, K., Salmonsen, L., & Marra, P. P. (2005). The Neighborhood Nestwatch Program: Participant outcomes of a citizen science ecological research project. Conservation Biology 19 (3), 589 594. Fabricius, C., Scholes, R.J., Cundil l G. ( 2006 ). Mobilizing knowledge for ecosystem management. In : Reid, W.V., Berkes, F., Wilbanks, T.J., Capistrano D. (Eds.). Bridging Scales and Knowledge Systems: Concepts and Applications in Ecosystems Management Island Press, Washington D.C. FAO. ( 1997 ) S tate of the World's Forests. Rome Italy.
191 Fahrig, L., & Rytwinski, T. (2009). Effects of roads on animal abundance: an empirical review and synthesis. Ecology and society 14 (1). http://www.ecologyandsociety.org/vol14/iss1/art21/ Fals Borda, O., & Rahman, M. A. (1991). Action and knowledge: breaking the monopoly with participatory action research. Rowman & Littlefield Publishers, USA. Ferguson, M., A. D., Williamson R. G., and Messier F ( 1998 ) Inuit knowledge of long term changes in a po pulation of arctic tundra caribou. Arctic 51, Finn, J. L. (1994). The promise of participatory research. Journal of Progressive Human Services 5 (2), 25 42. Folke, C., Carpenter, S., Elmqvist, T., Gunderson, L., Holling, C. S., & Walker, B. (20 02). Resilience and sustainable development: building adaptive capacity in a world of transformations. AMBIO: A journal of the human environment 31(5), 437 440. Foster, R. J., Harmsen, B. J., Valdes, B., Pomilla, C., & Doncaster, C. P. (2010). Food habits of sympatric jaguars and pumas across a gradient of human disturbance. Journal of Zoology 280 (3), 309 318. Frankham, R. (1996). Relationship of genetic variation to population size in wildlife. Conser vation Biology 10 (6), 1500 1508. Frankham, R. ( 2005 ) Genetics and extinction. Biological Conservation 126, 131 140. Frankham, R., Briscoe, D. A., Ballou, J. D. (2002). Introduction to conservation genetics Cambridge University P ress England Freeland, J. (1996). The Global, the National and the Local: forces in the development of education for indigenous peoples the case of Peru. Compare 26 (2), 167 195. Freire, P. (1972). Pedagogy of the Oppressed. 1968. Trans. Myra Bergman Ramos. New York: Herder Fusco Costa, R. & Ingberman, B. (2012) Record of bush dogs Speothos venaticus in a continuous remnant of coastal Atlantic Forest in southern Brazil. Oryx 47 (1), 105 108. Gadgil, M., Olsson, P., Berkes, F., Folke C. ( 2003 ) Exploring the rol e of local ecological knowledge in ecosystem ma nagement: three case studies. In : Berkes, F., Colding J., Folke C. (E ds.). Navigating socio ecological systems: building resilience for complexity and change University Press, Cambridge, pp. 189 209. Ga mble, L., Ravela, S., & McGarigal, K. (2008). Multi scale features for identifying individuals in large biological databases: an application of pattern recognition technology to the marbled salamander Ambystoma opacum Journal of Applied Ecology 45 (1), 170 180.
192 Garla, R. C., Setz, E. Z., & Gobbi, N. (2001). Jaguar ( Panthera onca ) food habits in Atlantic rain forest of southeastern Brazil. Biotropica 33 (4), 691 696. Garrison, D. R. (1997). Self directed learning: Toward a comprehensive model. Adult education quarterly 48 (1), 18 33. Gh imire, K.B., & Pimbert, M.P. ( 1997 ). Social change and conservation Earthscan, London, United Kingdom. Gibson, C. C., & M arks, S. A. (1995). Transforming rural hunters into conservationists: an assessment of community based wildlife management programs in Africa. World Development 23 (6), 941 957. Gil, G. E. &Lobo, J.M. (2012) Situacion del zorro vinagre (Speothos venatic us) en el extreme sur de su distribucion (Argentina). Interciencia 37: 21 28. [In Spanish]. Gilchrist, G., Mallory, M., & Merkel, F. (2005). Can local ecological knowledge contribute to wildlife management? Case studies of migratory birds. Ecology and S ociety 10 (1). Gilchrist, G., & Mallory, M. (2007). Comparing expert based science with local ecological knowledge: What are we afraid of?. Ecology and Society 12 (1). Gonzlez, C. A. L., & Miller, B. J. (2002). Do jaguars ( Panth era onca ) depend on large prey? Western North American Naturalist 218 222. Guyana Forestry Commission (GFC) (2012) MRV Interim Measures Report October 2010 31 December 2011. Guyana Forestry Commission Guyana REDD + Monitoring Reporting & Verification System (MRVS ) Version 3. Guyana Forestry Commission. (2013). Independent Forest Monitoring First Assessment Report 2013. Georgetown, Guyana, pp. 1 43. Hallett, M. (2015 a ) Survey of human wildlife conflict, resource use and attitudes towards conservation and sustainable development in the Kanuku Mountains, Rupununi, Guyana. figshare. http://dx.doi.org/10.6084 /m9.figshare.1287792 Hallett, M., James, S., Li, F., DeFreitas, D., Holden, F., Pereira, G., Holland, A.P. (2015 b ) Local Accounts of Opportunistic Encounters with Bush Dogs ( Speothos venaticus ) in the Rupununi Region of Guyana. figshare. http://dx.do i.org/10.6084/m9.figshare.1334305 Hallett, M.T., Roberts, A.R., Holland, A.P., Jackman, A. (2017). First photographic records of bush dogs ( Speothos venaticus ) in the wild in Guyana. Canid Biology and Conservation In Review. Hanski, I. (1994). Patch occupancy dynamics in fragmented landscapes. Trends in Ecology & Evolution 9 (4), 131 135.
193 Harmsen, B. J., Foster, R. J., Silver, S., Ostro, L., & Doncaster, C. P. (2010). Differential use of trails by forest mammals and the implications for camera trap s tudies: a case study from Belize. Biotropica 42 (1), 126 133. Harmsen, B. J., Foster, R. J., Silver, S. C., Ostro, L. E., & Doncaster, C. P. (2011). Jaguar and puma activity patterns in relation to their main prey. Mammalian Biology Zeitschrift fr Suget ierkunde 76 (3), 320 324. Hart, R. A. (2013). Children's participation: The theory and practice of involving young citizens in community development and environmental care Routledge, Abingdon, UK. Heyd, T. (1995). Indigenous knowledge, emancipation and alienation. Knowledge and Policy 8 (1), 63 73. Hunter, L. (2011). Carnivores of the World Princeton University Press, Princeton and Oxford, U.K. Huntington, H. P. (2000). Using traditional ecological knowledge in science: methods and applications. Ecolo gical applications 10 (5), 1270 1274. Jackson, S. L., Stratford, S. J., Krajcik, J. S., & Soloway, E. (1996). Model It: A Case Study of Learner Centered Software Design for Supporting Model Building National Science Foundation, Arlington, VA and Ann Arbor, MI. Jordan, R. C., Gray, S. A., Howe, D. V., Brooks, W. R., & Ehrenfeld, J. G. (2011). Knowledge gain and behavioral change in citizen science programs. Conservation Biology 25 (6), 1148 1154. Jorge, R.S.P., Ferreira, F., Ferreira Neto, J.S., Vasconcellos, S.D.A., Lima, E.D.S., Morais, Z.M.D. & Souza, G.O.D. (2011) Exposure of free ranging wild carnivores, horses and domestic dogs to Leptospira spp in the northern Pantanal, Brazil. Memrias do Instituto Oswaldo Cruz 106 (4), 441 444. Karanth, K. U. (1995). Estimating tiger Panthera tigris populations from camera trap data using capture recapture models. Biological conservation 71 (3), 333 338. Karanth, K.U., & Nichols, J.D. (1998) Estimation of tiger densities in India using photographic captures and recaptures. Ecology 79, 2852 2862. Karanth, K. U., & Sunquist, M. E. (2000). Behavioural correlates of predation by tiger ( Panthera tigris ), leopard ( Panthera pardus ) and dhole ( Cuon alpinus ) in Nagarahole, India. Journal of Zoology 250 (2), 255 265.
194 Kaschula, S.A., Twine, W.E. & Scholes, M.C. (2005) Coppice harvesting of fuel wood species on South African common: utilizing scientific and indigenous knowledge in community based natu ral resource management. Human Ecology 33, 387 418. Kellman, M., & Meave, J. (1997). Fire in the tropical gallery forests of Belize. Journal of Biogeography 24 (1), 23 34. Kiss, A. (2004). Is community based ecotourism a good use of b iodiversity conserv ation funds? Trends in ecology & evolution 19 (5), 232 237. Kleiman, D. G. (1972) Social behavior of the maned wolf ( Chrysocyon brachyurus ) and bush dog ( Speothos venaticus ): a study in contrast. Journal of Mammalogy 53 791 806. Kleinman, D. L. (1998). Beyond the science wars: Contemplating the democratization of science. Politics and the Life Sciences 17(2), 133 145. Klink, C. A., & Machado, R. B. (2005). Conservation of the Brazilian cerrado. Conservation Biology 19 (3), 707 713. Kruskal, W. H., & Wallis, W. A. (1952). Use of ranks in one criterion variance analysis. Journal of the American S tatistical Association 47 (260), 583 621. Le Saout, S., Hoffmann, M., Shi, Y., Hughes, A., Bernard, C., Brooks, T.M., Bertzky, B., Butchart, S.H.M, Stua rt, S.N., Badman T. Rodrigues, A.S.L. (2013). Protected areas and effective biodiversity conservation. Science 342 (6160), 803 805. Leach, M., Mearns, R., & Scoones, I. (1999). Environmental entitlements: dynamics and institutions in community based natural resource management. World development 27 (2), 225 247. Lee, T., Quinn, M., & Duke, D. (2006). Citizen, science, highways, and wildlife: using a web based GIS to engage citizens in collecting wildlife information. Ecology and Society 11 (1). http ://www.ecologyandsociety.org/vol11/iss1/art11/ Lehmann, L., Perrin, N. (2006). On metapopulation resistance to drift and extinction. Ecology 87 (7), 1844 1855. Levin, S.A. ( 1999 ). Fragile dominion: complexity and the commons Perseus, New York. Lima, E. S., Jorge, R.S.P., & Dalponte, J.C. (2009) Habitat use and diet of bush dogs, Speothos venaticus in the northern Pantanal, Mato Grosso, Brazil. Mammalia 73 13 19. Ludwig, D. ( 2001 ). The era of management is over Ecosystems 4, 758 764. Luzar, J. B., Silvius, K. M., & Fragoso, J. M. (2012). Church affiliation and meat taboo s in indigenous communities of G uyanese Amazonia. Human ecology 40 (6), 833 845.
195 Macdonald, D. W. (1996) Social behavior of captive bush dogs ( Speothos venaticus ). Journal of Z oology 239 (3), 525 543. Maffei, L., Cullar, E., & Noss, A. (2004). One thousand jaguars ( Panthera onca ) in Bolivia's Chaco? camera trapping in the Kaa Iya National Park. Journal of Zoology 262 (3), 295 304. Maffei, L., Noss, A. J., Silver, S. C., & Ke lly, M. J. (2011). Abundance/density case study: Jaguars in the Americas. In : O'Connell, A.F., Nichols, J.D., Karanth, K.U. ( eds. ). Camera traps in animal ecology: methods and analyses Springer Science & Business Media, Japan, pp. 119 144). Mallory, M. L ., Gilchrist, H. G., Fontaine, A. J., & Akearok, J. A. (2003). Local ecological knowledge of ivory gull declines in Arctic Canada. Arctic 293 298. Manyara, G., & Jones, E. (2007). Community based tourism enterprises development in Kenya: An exploration of their potential as avenues of poverty reduction. Journal of sustainable tourism 15 (6), 628 644. Mapfumo, P., Adjei Nsiah, S., Mtambanengwe, F., Chikowo, R., & Giller, K. E. (2013). Participatory action research (PAR) as an entry point for supporting climate change adaptation by smallholder farmers in Africa. Environmental Development 5 6 22. Matth, M., Sannolo, M., Winiarski, K., Spitzen van der Sluijs, A., Goedbloed, D., Steinfartz, S., & Stachow, U. (2017). Comparison of photo matching algorithms commonly used for photographic capture recapture studies. Ecology and Evolution 7 (15), 5861 5872. Mauro, F., & Hardison, P. D. (2000). Traditional knowledge of indigenous and local communities: international debate and policy initiatives. Ecological applications 10 (5), 1263 1269. McBride Jr, R. T., & McBride, R. T. (2007). Safe and selective capture technique for jaguars in the Paraguayan Chaco. The Southwestern Naturalist 52 (4), 570 577. McBride, R., Giordano, A., & Ballard, W. B. (2010). Note on the winter diet of jaguars Panthera onca in the P araguayan transitional chaco. Fauna Par aguay. Publications Bellbird Boletin MHNP Books McCann S., Lowe D.G. (2011). Local naive bayes nearest neighbor for image classification. CVPR, 2011. 2, 3, 4 McIntyre, A. (2007). Participatory action research (Vol. 52). Sage Publications, Thousand Oaks, CA.
196 McTurk, D., & Spelman, L. (2005). Hand rearing and rehabilitation of orphaned wild giant otters, Pteronura brasiliensis on the Rupununi river, Guyana, South America. Zoo Biology 24 (2), 153 167. Medri I.M., Mour o G M Harada AY (2003) Dieta de tamandu bandeira ( Myrmecophaga tridactyla ) no Pantanal da Nhecol ndia, Brasil. Edentata 5:29 34 Messmore, A. (1996). Measuring the impact of grassroots outreach. Science Communication 17(4), 430 442. Michalski, F. (2010) The bush dog Speothos ve naticus and short eared dog Atelocynus microtis in a fragmented landscape in Southern Amazonia. Oryx 44 (2), 300 303. Miranda, F., Bertassoni, A. & Abba, A.M. ( 2014 ) Myrmecophaga tridactyla The IUCN Red List of Threatened Species 2014. http://dx.doi.org/10.2305/IUCN.UK.2014 1.RLTS.T14224A47441961.en Downloaded on 18 September 2017 Mistry J, Roopsind M. (2004). The North Rupununi Adaptive Management Process (NRAMP) Darwin Initiative G uyana Partnership. https://repository.royalholloway .ac.uk/file/10bd80ea 3090 dba8 9013 eea3bab0004d/7/NRAMP_2008_Full_Version.pdf Downloaded on 18 September 2017. Mittermeier, R.A., Myers, N., Thomsen, J.B., da Fonseca, G.A.B., Silvio O ( 1998 ) Biodiversity h otspots and m ajor t ropical w ilderness a reas: Approaches to s etting c onservation p riorities. Conservation Biology 12 (3): 516 520. Mittermeier, R. A., Myers, N., Mittermeier, C. G., & Robles, G. (1999). Hotspots: Earth's biologically richest and most endangered terrestrial ecoregions CEMEX, SA, Agrupacin Sierra Madre, SC. Combining Science and Traditional Ecological Knowledge: Monitoring Populations for Co Management. Ecology and Society 9 (3), 2. Montalli, R. J.,& Kelly, K. (1989) Pathologic survey and review of diseases of captive maned wolves ( Chrysocyon brachyurus ) and bush dogs ( Speothos venaticus ). Erkrankungen der Zootiere 61 67 79. Montambault, J.R., Missa, O. (2002) A Biodiversity Assessement of the Eastern Kanuku Mountains, Lower Kwitaro River, Guyana. Conservation International Rapid Assessment Program (RAP) Working Papers 26, 1 86. Montgomery, G. G. & Lubin, Y. D. (1977). Prey influences on movements of neotrop ical anteaters. In : Phillips, R. L. & Jonkel, C. (Eds .). Proceedings of the 1975 predator symposium University of Montana, Missoula: Montana Forest and Conservation Experiment Station, pp. 103 131.
197 Morrison, T. A., Keinath, D., Estes Zumpf, W., Crall, J. P., & Stewart, C. V. (2016). Individual identification of the endangered Wyoming toad Anaxyrus baxteri and implications for monitoring species recovery. Journal of Herpetology 50 (1), 44 49. Murphree, M. W. (2002). Protected areas and the commons. Common Property Resource Digest 60 1 3. Mutandwa, E., & Gadzirayi, C. T. (2007). Impact of community based approaches to wildlife management: case study of the CAMPFIRE programme in Zimbabwe. The International Journal of Sustainable Development & World Ecolog y 14 (4), 336 344. National Development Strategy (NDS). (1996). Guyana National Development Strategy Ministry of Finance. Georgetown, Guyana. National Research Council (NRC). ( 1997 ). Science for all children: a guide to improving elementary science education for your district National Aca demy Press, Washington D.C Nelder, J. A., & Baker, R. J. (1972). Generalized linear models John Wiley & Sons, Inc., New York. Nilsson, S., Sjberg, J., Amundin, M., Hartmann, C., Buettner, A., & Laska, M. (2014). Behavioral responses to mammalian blood odor and a blood odor component in four species of large carnivores. PloS one 9 (11). https://doi.org/10.1371/journal.pone.0112694 Norman, D. A.,& Spohrer, J. C. (1996) Lea rner ce nt e red education. Communica tions of the ACM 39(4), 24 27. North Rupununi District Development Board (NRDDB). (2000) Local Ecological Knowledge of Some Large Mammals in the North Rupununi and Iwokrama Forest NRDDB, Guyana. Noss, A. J., Gardner, B., Maffei, L., Cullar, E., Mont ao, R., Romero Muoz, Sollman, R., & O'Connell, A. F. (2012). Comparison of density estimation methods for mammal populations with camera traps in the Kaa Iya del Gran Chaco landscape. Animal Conservation 15 (5), 527 535. Obbard, M. E., Howe, E. J., & Kyle, C. J. (2010). Empirical comparison of density estimators for large carnivores. Journal of Applied Ecology 47 (1), 76 84. : ranth, K.U. (Eds.). Camera traps in animal ecology Springer Japan, pp. 71 96
198 O'Connell, A. F., Nichols, J. D., & Karanth, K. U. (Eds.). (2010). Camera traps in animal ecology: methods and analyses Springer Science & Business Media. O rr, D. ( 1989 ). Ec ological literacy. Conservation Biology 3, 334 335. Parker, T.A. III, Foster, R.B., Emmons, L.B., Freed, P., Forsyth, A.B., Hoffman, B. & Gill, B.D. (1993) A biological assessment of the Kanuku Mountains region of southwestern Guyana. Conservation Inter national Rapid Assessment Program (RAP) Working Papers 5, 1 70. Parkes, M., & Panelli, R. (2001). Integrating catchment ecosystems and community health: the value of participatory action research. Ecosystem Health 7 (2), 85 106. Parks, S. A., & Harcourt, A. H. (2002). Reserve size, local human density, and mammalian extinctions in US protected areas. Conservation Biology 16 (3), 800 808. Parsons, E. C. M., Shiffman, D. S., Darling, E. S., Spillman, N., & Wright, A. J. (2014). How twitter literacy can benefit conservation scientists. Conservation Biology 28 (2), 299 301. Pattengill Semmens, C. V., & Semmens, B. X. (2003). Conservation and management applications of the reef volunt eer fish monitoring program. In: Melzian B.D., Engle, V., M cAlister, M., Sandhu S., & Eads L.K. Coastal Monitoring through Partnerships Springer Netherlands, pp. 43 50 Payan, E., Carbone, C., Homewood, K., Paemelaere, E., Quigley, H., Durant, S.M. (2013). Where will jaguar roam? The importance of survival in unprotected lands. In: Ruiz Garcia, M., Shostell, J.M. (Eds.). Molecular Population Genetics, Evolutionary Biology and Biological Conservation of Neotropical Carnivores Nova Science Pub Inc ., New York, p p 603 627. Pearson, K. (1895). Note on regressio n and inheritance in the case of two parents. Proceedings of the Royal Society of London 58 240 242. Peres, C. A. (1991) Observations on hunting by small eared ( Atelocynus microtis ) and bush dogs ( Speothos venaticus ) in central western Amazonia. Mammalia 55 635 639. Persha, L., Agrawal, A. Chhatre A. ( 2011 ). Social an d Ecological Synergy: Local Rulemaking, Forest Livelihoods and Biodiversity Conservation. Science 331, 1606 1608. Pickles, R.S.A., McCann, N.P. & Holland, A.P. (2011) Mammalia n and avian diversity of the Rewa Head, Rupununi, Southern Guyana. Biota Neotropica 11 (3), 237 251. Pimbert, M., & Pretty, J. N. (1995). Parks, people and professionals United Nations Research Institute for Social Development Geneva, Switzerland.
199 Polisar, J., Maxit, I., Scognamillo, D., Farrell, L., Sunquist, M. E., & Eisenberg, J. F. (2003). Jaguars, pumas, their prey base, and cattle ranching: ecological interpretations of a management problem. Biological C onservation 109 (2), 297 310. Porton, I .J., Kleiman, D.G. & Rodden, M. (1987) Aseasonality of bush dog reproduction and the influence of social factors on the estrous cycle. Journal of Mammology 68 (4), 867 871. Powell, R. A. (2000). Animal home ranges and territories and home range estimato rs. In: Boitani, L., Fuller, T. (Eds .). Research techniques in animal ecology: controversies and consequences Colombia University Press, New York, 1 476 Protected Areas Commission (PAC), Guyana. ( 2015 ) Kanuku Mountains Protected Area Management Plan 2015 2019 Georgetown: PAC, Guyana Purcell, T. (1998). Indigenous knowledge and applied anthropology: Questions of definition and direction. Human organization 57 (3), 258 272. Quelch, J.J. (1901) Animal life in British Guiana Argosy Publishers, Ltd., Georgetown, Guyana. Quiroga, V. A., Noss, A. J., Boaglio, G. I., & Di Bitetti, M. S. (2016). Local and continental determinants of giant anteater ( Myrmecophaga tridactyla ) abundance: Biome, human and jaguar roles in population regulation. Mammalian Biology Zeitschrift fr Sugetierkunde 81 (3), 274 280. R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R project.org/. Rabinowitz, A. R., & Nottingham, B. G. (1986). Ecology and behaviour of the Jaguar ( Panthera onca ) in Belize, Central America. Journal of Zoology 210 (1), 149 159. Rabinowitz, A., & Zeller, K. A. (2010). A range wide model of landscape connectivity and conservation for the jaguar, Panthera onca Biological conservation 143 (4), 939 945. Redford, K. H. (1985). Feeding and food preference in captive and wild giant anteaters ( Myrmecophaga tridactyla ). Journal of Zoology 205 (4), 559 572. Reed, M. S. (2008). Stakeholder participation for environmental ma nagement: a literature review. Biological conservation 141 (10), 2417 2431. Ridout M.S., & Linkie M. (2009) Estimating overlap of daily activity patterns from camera trap data. Journal of Agricultural, Biological, and Environmental Statistics 14, 322 337
200 Rodrigues, F. H. G., de Miranda, G. B., Medri, . M., dos Santos, F. V., Mouro, G. D. M., Hass, A., ... & Rocha, F. L. (2003). Fitting radio transmitters to giant anteaters ( Myrmecophaga tridactyla ). Edentata 5 37 40. Rodrigues, F.H.G., Medri, I.M., De Miranda, G.H.B ., Camilo Alves, C., Mourao, G. ( 2008 ) Anteater behavior and ecology. In: Vizcaino, S.F., Loughry, W.J. (Eds.), The Biology of the Xenarthra University of Florida, pp. 257 268. Roopsind, A., Caughlin, T. T., Sambhu, H., Fragoso, J., & Putz, F. E. (2017). Logging and indigenous hunting impacts on persistence of large Neotropical animals. Biotr opica 49 (4), 565 575. Roux, D., Rogers, K., Biggs, H., Ashton, P., & Sergeant, A. (2006). Bridging the science management divide: moving from unidirectional knowledge transfer to knowledge interfacing and sharing. Ecology and Society 11 (1). https://www .ecologyandsociety.org/vol11/iss1/art4/main.html Sanderson, E.W., Chetkiewicz, C. L.B., Rabinowitz, A., Redford, K.H., Ro binson, J.G., Taber, A.B. (1999). Un anlisis geogrfico del estado de conservacin y distribuicin de los jaguares a travs de su rea de distribuicin. In: Medell n, R.A., Equihua, C., Chetkiewicz, C. L.B., Crawshaw, P.G., Jr. Rabinowitz, A., Redford, K.H., Robinson, J.G., Sanderson, E.W., Taber, A.B. (Eds.), El Jaguar en el Nuevo Milenio Ediciones Cient ficas Universitrias, Univ ersidad Nacional Autnoma de M xico / Wildlife Conservation Society, New York, pp. 551 600 Sanderson, E. W., Redford, K. H., Chetkiewicz, C. L. B., Medellin, R. A., Rabinowitz, A. R., Robinson, J. G., & Taber, A. B. (2002). Planning to save a species: the jaguar as a model. Conservation Biology 16 (1), 58 72. Schaller, G. B., & Vasconcelos, J. M. C. (1978). Jaguar predation on capybara. Z. Sugetierk 43 296 301. Schaller, G. B., & Crawshaw Jr, P. G. (1980). Movement patterns of jaguar. Biotropica 161 168. Schoener, T. W. (1974). The compression hypothesis and temporal resource partitioning. Proceedings of the National Academy of Sciences 71 (10), 4169 4172. Schultz, P. (2011). Conservation means behavior. Conservation Biology 25 (6), 1080 1083. Scognamillo, D., Maxit, I. E., Sunquist, M., & Polisar, J. (2003). Coexistence of jaguar ( Panthera onca ) and puma ( Puma concolor ) in a mosaic landscape in the Venezuelan llanos. Journal of Zoology 259 (3), 269 279. Seymour, K. L. (1989). Panthera onca Mammalian S pecies (340), 1 9.
201 Shackley, M. (1998). Designating a protected area at Karanambu ranch, Rupununi savannah, Guyana: resource management and indigenous communities. Ambio 27 (3), 207 210. Shaw, J. H., Machado Neto, J., & Carter, T. S. (1987). Behavior of free living giant anteaters ( Myrmecophaga tridactyla ). Biotropica 255 259. Silveira, L. (2004). E cologia comparada e conservao da ona pintada ( Panthera onca ) e ona parda ( Puma concolor ), no Cerrado e Pantanal. PhD Dissertation University of Brazilia Silveira, E. K. P. (1969). Historia Natural do Tamandu bandeira, Myrmecophaga tridactyla Linn. 1 758, Myrmecophagidae. Velozia, 7, 34 43. Silver, S. (2004). Assessing jaguar abundance using remotely triggered cameras. Wildlife Conservation Society 1 27. Silvertown, J. (2009). A new dawn for citizen science. Trends in ecology & evolution 24 (9), 467 471. Simons, C. (2011). Uncertain future of tropical ecology. Science 332, 298 299. Soisalo, M. K., & Cavalcanti, S. M. (2006). Estimating the density of a jaguar population in the Brazilian Pantanal using camera traps and capture recapture sampling in combination with GPS radio telemetry. Biological Conservation 129 (4), 487 496. Sollmann, R. (2011). Ecology and conservation of the jaguar (Panthera onca) in the Cerrado grasslands of central Brazil Doctoral dissertation, Freie Universi tt Berlin Stem, C. J., Lassoie, J. P., Lee, D. R., Deshler, D. D., & Schelhas, J. W. (2003). Community participation in ecotourism benefits: The link to conservation practices and perspectives. Society &Natural Resources 16 (5), 387 413. Stockwell, C. A., Hendry, A. P., & Kinnison, M. T. (2003). Contemporary evolution meets conservation biology. Trends in Ecology & Evolution 18 (2), 94 101. Stone, S. (2002). Community Resource Evaluation Master Report. Conservation International Guyana, Georgetown, Guyana. Swean or, L. L., Logan, K. A., & Hornocker, M. G. (2000). Cougar dispersal patterns, metapopulation dynamics, and conservation. Conservation Biology 14 (3), 798 808. Terborgh, J., Estes, J.A., Paquet, P., Ralls, K., Boyd Heger, D., Miller, B.J. & Noss, R.F. (19 99) The role of top carnivores in regulating terrestrial ecosystems. Continental Conservation: Scientific Foundations of Regional Reserve Networks (eds M.E.Soul & J.Terborgh), Island Press, Covelo, CA, pp. 39 64.
202 ter Steege, H., Boot, R.G.A., Brouwer, L .C., Caesar, J.C., Ek, R.C., Hammond, D.S., Haripersaud, P.P., van der Hout, P., Jetten, V.G., van Kekem, A.J., Kellman, M.A., Khan, Z., Polak, A.M., Pons, T.L., Pulles, J., Raaimakers, D., Rose, S.A., van der Sanden, J.J. & Zagt, R.J. (1996) Ecology and logging in tropical rain forest in Guyana, with recommendations for forest management: Tropenbos Series, No. 14 The Tropenbos Foundation, Wageningen, The Netherlands. Tewksbury, J. J., Anderson, J. G., Bakker, J. D., Billo, T. J., Dunwiddie, P. W., Groom M. J., ... & Del Rio, C. M. (2014). Natural history's place in science and society. BioScience 64(4), 300 310. Tobler, M. W., Carrillo Percastegui, S. E., Leite Pitman, R., Mares, R., & Powell, G. (2008). An evaluation of camera traps for inventorying large and medium sized terrestrial rainforest mammals. Animal Conservation 11 (3), 169 178. Tobler, M. W., Carrillo Percastegui, S. E., Hartley, A. Z., & Powell, G. V. (2013). High jaguar densities and large population sizes in the core habitat of the so uthwestern Amazon. Biological Conservation 159 375 381. Toby Pennington, R., Prado, D. E., & Pendry, C. A. (2000). Neotropical seasonally dry forests and Quaternary vegetation changes. Journal of Biogeography 27 (2), 261 273. Toomey, A. H. ( 2016 ) What happens at the gap between knowledge and practice? Spaces of encounter and misencounter between environmental scientists and local people. Ecology and Society 21 (2), 28. http://dx.doi.org/10.5751/ES 08409 210228 T rres N M Filho J A F D De Marco Jr ., P Jacomo A T A Silveira L (2007) Jaguar distribution and conservation status in Brazil. In: Hughes J Mercer R (eds) Felid Biology and Conservation Conference 17 20 September: abstracts WildCRU, Oxford, p p 109 Trumbull, D. J., Bonney, R., Bascom, D., & Cabral, A. (2000). Thinking scientifically during participation in a citizen science project. Science education 84 (2), 265 275. Tuss, P. (1996). From student to scientist: An experiential approach to science education. Science Communica tion 17 (4), 443 481. Walters, C. J., & Holling, C. S. (1990). Large scale management experiments and learning by doing. Ecology 71 (6), 2060 2068. Warburton, H., & Martin, A. (1999). Local people's knowledge in natural resources research. Natural Resources Institute, University of Greenwich, UK. Weckel, M., Giuliano, W., Silver, S. (2006). Jaguar ( Panthera onca ) feeding ecology: distribution of predator and prey through time and space. Journal of zoology 270 (1), 25 30.
203 Williams, B. K., Nichols, J. D., & Conroy, M. J. (2002). Analysis and management of animal populations Academic Press, Cambridge, Massachusetts. Wohling, M. (2009). The problem of scale in indigenous knowledge: a perspective from northern Australia. Ecology and Society 14 (1). Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving. Journal of child psychology and psychiat ry 17 (2), 89 100. Woodroffe, R. (2000 ). Predators and people: using human densities to interpret declines of large carnivores. Animal Conservation forum 3( 2 ):165 173 Woodroffe, R., & Ginsberg, J. R. (1998). Edge effects and the extinction of populatio ns inside protected areas. Science 280 (5372), 2126 2128. Wultsch, C., Waits, L. P., & Kelly, M. J. (2016). A comparative analysis of genetic diversity and structure in jaguars ( Panthera onca ), pumas ( Puma concolor ), and ocelots ( Leopardus pardalis ) in fragmented landscapes of a critical Mesoamerican linkage zone. PloS one 11 (3), e0151043. Young, A.G., Clarke, G.M. ( 2000 ) Conclusions and future directions: what do we know about the genetic and demographic effects of habitat fragmentation and where do we go from here? In: Y oung, A.G., Clarke, G.M. (Eds.). Genetics, Demography and Viability of Fragmented Populations Cambridge University Press, Cambridge, England, pp. 361 366. Zabel, C. J., Roberts L. M., Mulder B. S., Stauffer H. B., Dunk J. R., Wo lcott K., Solis D., Gertsch M., Woodbridge B., Wright A., Goldsmith G., and Keckler C. ( 2002 ) A collaborative approach in adaptive manage ment at a largelandscape scale. In: J. M. Stott, P. J. Heglund, M. L. Morrison, J. B. Haufler, M. G. Raphael, W. A Wall, and F. B. Samson (Ed s. ). Predicting species occurrence: issues of accuracy and scale Isla nd Press, Washington, D.C., pp. Zar, J.H. ( 2010 ) Biostatistical Analysis 5 th ed. Pearson Prentice Hall: Upper Saddle River, NJ. Zhelezniakov, A., Eerola, T., Koivuniemi, M., Auttila, M., Levnen, R., Niemi, M., Kunnasranta, M., & Klviinen, H. (2015). Segmentation of saimaa ringed seals for identification purposes. In International Symposium on Visual Computing (pp. 227 236). Spri nger International Publishing. Zuercher, G.L., Gipson, P.S. & Carrillo, O. (2005) Diet and habitat associations of bush dogs Speothos venaticus in the Interior Atlantic Forest of eastern Paraguay. Oryx 39 (1), 86 89.
204 BIOGRAPHICAL SKETCH Matthew Thomas Hallett was born and raised in Chicago, Illinois. He received a Bachelor of Arts in biology with minors in e nvironmental studies and political s cience from the College of Charleston i n 2005 and a Master of Arts in zoology with a concentration in c ommunity based c onservation from Miami University in 2009. He previously worked in public education at the zoos and aquariums in the U.S. and in research and community based conservation in Kenya, South Africa, Malaysian Borneo and Guyana. Experiences in the field have inspired Matt to consistently pursue work that falls at the intersection of humans and nature with the challenge of developing creative solutions that maximize the benefits to both.