University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Dissertations & Theses in Natural Resources Natural Resources, School of

5-2020

Charismatic Predators in Modern Africa: Spotted hyena (Crocuta crocuta) and Human coexistence in the Northern Tuli Game Reserve, Botswana

Jazmin Castillo University of Nebraska - Lincoln, [email protected]

Follow this and additional works at: https://digitalcommons.unl.edu/natresdiss

Part of the Hydrology Commons, Natural Resources and Conservation Commons, Natural Resources Management and Policy Commons, Other Environmental Sciences Commons, and the Water Resource Management Commons

Castillo, Jazmin, "Charismatic Predators in Modern Africa: Spotted hyena (Crocuta crocuta) and Human coexistence in the Northern Tuli Game Reserve, Botswana" (2020). Dissertations & Theses in Natural Resources. 312. https://digitalcommons.unl.edu/natresdiss/312

This Article is brought to you for free and open access by the Natural Resources, School of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Dissertations & Theses in Natural Resources by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.

CHARISMATIC PREDATORS IN MODERN AFRICA:

SPOTTED HYENA (CROCUTA CROCUTA) AND HUMAN COEXISTENCE IN THE

NORTHERN TULI GAME RESERVE, BOTSWANA

By

Jazmin Castillo

A THESIS

Presented to the Faculty of The Graduate College at the University of Nebraska In Partial Fulfillment of the Requirements For the Degree of Master of Science

Major: Natural Resource Sciences

Under the Supervision of Professors John P. Carroll and John P. DeLong

Lincoln, Nebraska May, 2020

CHARISMATIC PREDATORS IN MODERN AFRICA:

SPOTTED HYENA (CROCUTA CROCUTA) AND HUMAN COEXISTENCE IN THE

NORTHERN TULI GAME RESERVE, BOTSWANA

Jazmin Castillo, M.S. University of Nebraska, 2020

Advisors: John P. Carroll and John P. DeLong Increasing human populations has led researchers to investigate the impacts of high human population density and its impact on carnivore populations. Spotted hyenas

(Crocuta crocuta) hold a unique place in African ecosystems due to being a very successful top predator with an adaptable diet whereas other top predators, like the

African lion, are rapidly decreasing in abundance. We investigated past and current spotted hyena abundance within the Northern Tuli Game Reserve to better aid in wildlife management. Spotted hyenas showed no significant difference in the change in population abundance throughout the different years of the study (2008-2016). Spotted hyena populations also showed no preference in habitat types between river and upland sites whereas brown hyenas had greater detections in river sites. Spotted hyena distribution was negatively impacted by villages and proximity to the border and positively impacted by ecotourism camps and cattle posts. With the decline in riverine habitat and humans influencing wildlife space usage, it is important to further investigate whether certain areas within the reserve require different management zones for a better balance in a limited space surrounded by various human infrastructure. i

DEDICATION Les dedico este proyecto a ustedes quienes dejaron todo atrás para dar me una buena vida y oportunidad. Gracias Mami (Maria G. Martinez) y Papi (Alberto Castillo) por su apoyo y amor durante mi carrera.

To Maithem and Tata, my Omanican family. ii

ACKNOWLEDGMENTS

Firstly, thank you to the former and current presidents of the Republic of Botswana, his

Excellency Lieutenant General Dr. Seretse Khama Ian Khama and Dr. Mokgweetsi E.K.

Masisi for allowing me to visit their country and the opportunity to conduct this research

project. To the Department of Wildlife and National Parks, Botswana for permission to

conduct our research. Also, a special thanks to all of the landowners of Northern Tuli for

access to their properties, David Evans, Piet Van Rensburg, Redge Van Rensburg, Stuart

Quinn, Jannie Eager, and Anthony and Marcelle Robbins from Depotta Game Reserve.

Thank you to the Mashatu Game Reserve for providing support through the use of a Land Cruiser, equipment, local knowledge, and access to various parts of the reserve.

Thank you to EcoTraining, Inc. and Stuart Quinn from Tuli Wilderness Safari, for providing housing and vehicles during field seasons. To all of the UNL and UGA

Botswana Study Abroad groups that assisted in data collection. Each of you helped in different ways and had me laughing harder every trip.

Special thanks to all the programs that got me to where I am today. The McNair

Program, IANR, Cabela’s Internship, UNL Study Abroad Office, the NSF NRT Program, and UCARE. Each program allowed me to continue to travel to Botswana and conduct my research.

To my committee: John Carroll, John DeLong, and Liz VanWormer. Thank you for being patient and providing feedback to my project. I am glad to have each of you on my committee and helping me look at my project with a wider scope. iii

John, Papa Impala, thank you for everything. You have shaped me into the biologist I am today and I have grown so much by having you as my advisor. My family refers to you as my “Second Father” because you have taken me under your guidance from the moment I arrived at UNL. You have taught me to be unapologetic and be proud of who I am even for things as simple as properly pronouncing my name. I will treasure our bumbles out in the bush hiding geocaches and sneaking by Maria’s desk to ask you a quick question. Thank you for seeing in me the potential I had yet to see in myself. I look forward to many more adventures and new opportunities because the Dark Vortex of Los

Castillos continues.

Thank you to Andrei Snyman, for allowing me to conduct research with you. This opportunity is something many dream of conducting and I am grateful you allowed me to take on a project that showed me the coolness of spotted hyenas. Special thanks to Lauren

Satterfield and Katie McCollum for being the ones that collected the data for my third chapter. You are two badass women. You were the first women in graduate school that piqued my interest and showed me what was possible in the field of ecology. Thank you for answering my questions about graduate school when I was a freshman who had never heard about a Doctorate or a Master’s degree. Lauren thank you for taking the time out of your busy schedule to answer all of my data questions.

Thank you to Lyndsie Wszola and Gabby Palomo for putting up with my endless questions of coding and statistics. I look up to both of you and truly appreciate you both taking the time out of your schedules to help me with my project. Lyndsie thank you for always offering me advice, introducing me to deer meat, and talking me through my academic journey. You helped me out and I truly admire you. Thank you to Jessica iv

Burnett who also helped me with coding questions and being willing to zoom whenever I needed a question answered.

To all of my friends that kept me sane throughout this journey, to the Coffee Club

(Callie Seaton and Jessica McDaniel), Mason Williams & Tiger, Alex Cleveland,

Shannon Al-Badry, Carina Arriaga, Andrea Tostado, and Amy Hanson. You all helped me in some way throughout this journey and your friendships have meant the world to me. To the NRT Resilient Raccoons, our endless group chats, inside jokes, “collaboration days”, and crazy adventures. I am grateful the NRT provided me a lab of diverse colleagues.

Thank you to my entire family, the Castillos’, Pichardos’, Guerreros’, and

Martinez. Este logro es para todos ustedes. Especially los Primxs: Kathy, Javier Jr.,

Leonardo, Alondra, Ruby, Victor, and Gabriel.

Last but not least, Maithem. You are the only one who knows how much I truly suffered and sacrificed to get to this point. Your support and the constant push to get this degree finished helped keep me motivated. I am thankful our paths crossed and for helping me grow as a person and as a scientist. v

TABLE OF CONTENTS Page DEDICATION ...... i ACKNOWLEDGMENTS ...... ii LIST OF TABLES ...... vii LIST OF FIGURES ...... vii CHAPTER 1: INTRODUCTION ...... 1 REFERENCES ...... 6 CHAPTER 2: ABUNDANCE AND DISTRIBUTION OF SPOTTED HYENAS (CROCUTA CROCUTA) IN THE NORTHERN TULI GAME RESERVE, BOTSWANA ABSTRACT ...... 8 INTRODUCTION ...... 8 METHODS ...... 12 RESULTS ...... 16 DISCUSSION ...... 18 REFERENCES ...... 22 TABLES ...... 28 FIGURES ...... 34 CHAPTER 3: OCCUPANCY ANALYSIS OF SPOTTED HYENAS (CROCUTA CROCUTA) AND BROWN HYENA (HYAENA BRUNNEA) IN THE NORTHERN TULI GAME RESERVE, BOTSWANA ABSTRACT ...... 44 INTRODUCTION ...... 45 METHODS ...... 47 RESULTS ...... 51 DISCUSSION ...... 52 REFERENCES ...... 54 TABLES ...... 57 FIGURES ...... 60 vi

CHAPTER 4: DISCUSSION ...... 70 REFERENCES ...... 73

vii

LIST OF TABLES

Table 2.1. Spotted hyena abundance survey for all study years (2008-2009, 2015-2016)

on the Northern Tuli Game Reserve, Botswana ...... 33

Table 2.2. Spotted hyena linear fixed models AIC Table comparison with various

variables ...... 34

Table 2.3. Summary table on best fitting model ...... 35

Table 2.4. Candidate covariates and their respective definitions used in linear fixed

models ...... 35

Table 2.5. Simplified spotted hyena linear fixed effects models with AIC table

comparisons ...... 36

Table 2.6. The best fitting model with variables’ p-values from simplified models .....36

Table 3.1. Candidate covariates used in survey-specific occupancy models with their respective definitions ...... 58

Table 3.2. Survey-specific model selection results for spotted hyenas ...... 59

Table 3.3. Survey-specific model selection results for brown hyenas ...... 60

LIST OF FIGURES

Figure 2.1. Poorly maintained fencing along the western border and within the center of the Northern Tuli Game Reserve, Botswana ...... 37

Figure 2.2. Decision tree used to prevent double detection at calling station sites for abundance surveys of spotted hyenas, NTGR, Botswana ...... 38 viii

Figure 2.3. Spotted hyena calling station sites placed throughout the Northern Tuli Game

Reserve, Botswana in 2008 (n=14), 2009 (n=17), 2015 (n=19), and 2016 (n=19) ...... 39

Figure 2.4. The layout of various human activity throughout the Northern Tuli Game

Reserve, Botswana ...... 40

Figure 2.5. Spotted hyena calling station locations within NTGR, Botswana set as point counts with an effective calling radius of 3.2 km covering a total area of 32.2 km² .....41

Figure 2.6. Estimated spotted hyena density (hyenas/km²) distribution throughout NTGR

for the four years of data collection (2008, 2009, 2015, and 2016) ...... 42

Figure 2.7. Estimated spotted hyena abundance and density for the four years of data

collection (2008, 2009, 2015, 2016) within NTGR ...... 43

Figure 2.8. A linear relationship between various fixed effects and predicted hyena abundance ...... 43

Figure 2.9. A linear relationship between spotted hyena amount and different fixed effects ...... 44

Figure 2.10. Relationship of spotted hyena amount based on the interaction of cattle posts and their proximity to humans ...... 45

Figure 3.1. The largest landowner within the Northern Tuli Game Reserve (NTGR),

Botswana is known as the Mashatu Game Reserve and the location of our camera trap study site (2013-2014) ...... 61

Figure 3.2. Camera trap sites layout within the Mashatu Concession for 2013 and 2014 in eastern Botswana ...... 62 ix

Figure 3.3. Northern Tuli Game Reserve map layout of various human settlements within

the reserve ...... 63

Figure 3.4. Spotted and Brown hyena detections from camera trap study in 2013 within

the EcoTraining concession of the Mashatu Game Reserve ...... 64

Figure 3.5. Spotted and Brown hyena detections from the camera trap study in 2014 in

EcoTraining and Central concessions of the Mashatu Game Reserve ...... 65

Figure 3.6. Camera trap sites from 2013-2014 within the Mashatu Game Reserve,

Botswana indicating detection of both spotted and brown hyenas ...... 66

Figure 3.7. Camera trap sites from 2013 separated into detected or not detected sites

based on average distances from various covariates...... 67

Figure 3.8. Camera trap sites from 2014 separated into detected or not detected sites

based on average distances from various covariates...... 68

Figure 3.9. Camera trap sites from 2013-2014 based on the percentage of cameras

detecting both species of hyenas based on habitat type ...... 69

Figure 3.10: Spotted and brown hyena detections from both 2013 and 2014 based on

habitat type ...... 70 1

CHAPTER 1: INTRODUCTION Throughout several millennia, the relationship between humans and wildlife has evolved significantly with an accelerating increase in human population and per capita resource use. According to the United Nations’ World Urbanization prospects report, 55% of the

World’s population live in urban areas. Although it is estimated to increase to 68% by

2050, Africa remains predominantly rural with only 43% of its population living in urban areas (UN Dept for Economic and Social Affairs, 2019). Despite having the lowest population living in urban areas, the African continent currently has the greatest rate of population growth at 2.53%, with a current population of 1.2 billion people (UN Dept for

Economic and Social Affairs, 2019). It is expected that the human population on the

African continent will rise to 2.5 billion people in 2050. These rapid increases in human

population will lead to more of the land base diverted from natural or semi-natural habitats to human-dominated ones, increasing habitat fragmentation for wildlife. Habitat fragmentation consists of the decline in total habitat area and the redistribution of discontinuous habitat fragments; both of these elements are significant factors in species extinction (Wilcove et al., 1986).

Increasing human populations has led researchers to investigate the impacts of high human population density (HPD) and its impact on carnivore populations. HPD does not show signs of direct impacts on the extinction rate, but rather how an animal responds to these impacts determines the overall success in survival (Cardillo et al., 2004).

Implementing population prevention programs for species that are categorized as least concern saves more money than saving a species once already at risk. Understanding how resilient a species is within an ecosystem to ensure its survival given the current and 2

expected future impacts. The term ‘resilience’ is defined as the ability to absorb change

while still maintaining its relationship within the environment (Holling, 1973). Holling

also defines stability as ‘the ability of a system to return to an equilibrium state after a

temporary disturbance.’ A system can be resilient and not stable or stable with little

resilience. For better management of a system, consideration must be given to the

projected change in human populations and maintaining current environmental systems

within this framework.

It is important to manage ecosystems through a ‘learning by doing’ method also

known as adaptive management (Allen et al., 2011). Adaptive management takes into

account the lack of understanding of complex systems by integrating the need for the

constant learning of systems during the management implementation. This approach also

looks into all agents within the system and emphasizes the need for flexible management

in an uncertain future.

Botswana in 2000 had a population of 1.7 million and by 2018 increased to 2.3 million with an annual growth rate of 1.63% (United Nations, 2018). Although

Botswana’s growth rate is slowly increasing, the issues of human-wildlife coexistence are

not just the number of people, but also the location of people within the country. In the

2011 census, the Eastern District of Botswana contained 12.7% of the entire population

and had the highest human density in the country (Statistics Botswana, 2015). Farmers

and cattle owners have had a gradual increase since 2007, but a 2013 survey showed that

there has been a gradual decrease in active farmers and cattle holding (Annual

Agricultural Survey Report 2013, 2015). This decrease was due to harsh weather conditions and foot and mouth disease outbreaks. Only 7% of Botswana’s land qualifies 3

as arable land, making the land limited in usage (USAID, 2010). Finding a balance between large human concentrations, livestock, agriculture, and wildlife is important when managing for the future of the country.

Spotted hyenas hold a unique place in African ecosystems due to being a very successful predator with an adaptable diet while many of the other predators are rapidly decreasing in abundance and are categorized as Threatened (Cozzi et al., 2015;

Woodroffe, 1998; Woodroffe, 2000; Yirga et al., 2013). Currently ranked as Least

Concern by the IUCN Red List with a decreasing trend in population, spotted hyenas are one of the most abundant predators (Bohm & Honer, 2015). Perhaps because hyenas are

so successful and widespread they are seen as less important within the assemblage of

large carnivores. Previous predator management plans within eastern Botswana focused

less on both hyena species and more on African lions (Panthera leo), wild dogs (Lycaon

pictus), and cheetahs (Acinonyx jubatus) (Greater Mapungubwe Transfrontier

Conservation Area, 2010). However, where limited resources for research are normally

put to good use for large carnivores classified as endangered or vulnerable, such as wild

dogs and lions respectively, ensuring the stability of a population before reaching a

vulnerable state saves both money and efforts in the end (Cardillo et al., 2004).

My study area was in eastern Botswana within the Northern Tuli Game Reserve,

NTGR. This privately owned reserve holds a unique area of protected land within an

array of farms and villages. Its unique setting provides insight into private game

management in a country where 38% of Botswana’s land is dedicated to national parks

and game reserves (Winterbach et al., 2014). To improve wildlife management, the

Greater Mapungubwe Transfrontier Conservation Area (GMTFCA) was created in 2016. 4

Parts of South Africa, Zimbabwe, and Botswana form the GMTFCA, summing to a total of 487,200 ha with 28% being in Botswana, including NTGR (Greater Mapungubwe

Transfrontier Conservation Area, 2010).

The objective of my thesis is to understand the current distribution and abundance of spotted hyena populations within the Northern Tuli Game Reserve and determine what factors affect their overall distribution. This project was established to better manage the reserve and ensure stable populations in the various carnivore communities within the area.

In Chapter 2, I focused on determining the abundance and distribution of spotted hyenas throughout the Northern Tuli Game Reserve and how human settlements, habitat types, and the different years the study was conducted affect their population. Despite ongoing lion studies within the area, little was known about current spotted hyena populations in the area. Calling stations were conducted throughout the reserve in 2008,

2009, 2015, and 2016. Four years of surveys within the area were conducted to estimate the abundance of hyena populations and look into spatial impacts influencing the distribution of hyena populations. Understanding population dynamics on one of the most abundant predator species in the area is crucial for better management.

In Chapter 3, I investigated the relationship between the two more common hyena species within the area, spotted hyena (Crocuta crocuta) and brown hyenas (Hyaena brunnea). An occupancy study on spotted hyenas and brown hyenas within the Mashatu concession located within the Northern Tuli Game Reserve was conducted to determine how the presence of spotted hyenas impact the distribution of the lesser species, brown 5 hyenas. A look into what factors dictate the distribution of both species will aid in better management within the area.

6

REFERENCES Allen, C. R., Fontaine, J. J., Pope, K. L., & Garmestani, A. S. (2011). Adaptive

management for a turbulent future. Journal of Environmental Management, 92(5),

1339–1345.

Annual Agricultural Survey Report 2013. (2015).

http://www.statsbots.org.bw/sites/default/files/publications/Agriculture%20Main

%20Report%202013_Jul%2017%202015.pdf

Bohm, T., & Honer, O. R. (2015). Crocuta crocuta. International Union for Conservation

of Nature (IUCN). https://doi.org/10.2305/IUCN.UK.2015-2.RLTS.T5674A45194782.en

Cardillo, M., Purvis, A., Sechrest, W., Gittleman, J. L., Bielby, J., & Mace, G. M. (2004).

Human population density and extinction risk in the world’s carnivores. PLoS

Biology, 2(7), e197. https://doi.org/10.1371/journal.pbio.0020197

Cozzi, Gabriele, Börger, L., Hutter, P., Abegg, D., Beran, C., McNutt, J. W., & Ozgul, A.

(2015). Effects of trophy hunting leftovers on the ranging behaviour of large

carnivores: A case study on spotted hyenas. PLOS ONE, 10(3), e0121471.

https://doi.org/10.1371/journal.pone.0121471

Greater Mapungubwe Transfrontier Conservation Area. (2010).

https://www.notugre.com/home/gmtfca

Holling, C. S. (1973). Resilience and stability of ecological systems. Annual Review of

Ecology and Systematics, 4(1), 1–23.

Hyaenid. (2018). IUCN. https://www.iucn.org/commissions/ssc-

groups/mammals/specialist-groups-f-z/hyaenid 7

Statistics Botswana (Ed.). (2015). Population census atlas 2011: Botswana. Statistics

Botswana.

UN Dept for Economic and Social Affairs. (2019). World urbanization prospects: The

2018 revision. UNITED NATIONS.

USAID. (2010). Botswana | LandLinks. https://www.land-links.org/country-

profile/botswana/

Wilcove, D. S., McLellan, C. H., & Dobson, A. P. (1986). Habitat fragmentation in the

temperate zone. Conservation Biology, 6, 237–256.

Winterbach, H. E. K., Winterbach, C. W., & Somers, M. J. (2014). Landscape suitability

in Botswana for the conservation of its six large African carnivores. PLoS ONE,

9(6), 1–12.

Woodroffe, R. (1998). Edge effects and the extinction of populations inside protected

areas. Science, 280(5372), 2126–2128.

Woodroffe, Rosie. (2000). Predators and people: Using human densities to interpret

declines of large carnivores. Animal Conservation, 3, 165–173.

Yirga, G., Ersino, W., De longh, H. H., Leirs, H., Gebrehiwot, K., Deckers, J., & Bauer,

H. (2013). Spotted hyena (Crocuta crocuta) coexisting at high density with

people in Wukro district, northern Ethiopia. Mammalian Biology, 78, 193–197.

8

CHAPTER 2: ABUNDANCE AND DISTRIBUTION OF SPOTTED HYENAS IN THE NORTHERN TULI GAME RESERVE, BOTSWANA ABSTRACT Rapid human population growth on the African continent has led to an increase in human-wildlife conflicts with apex predators. Predator management is a major focus in conservation due to the high levels of persecution experienced by predators. Spotted

hyenas in eastern Botswana were historically extirpated from the area until the

establishment of the Northern Tuli Game Reserve. Calling station surveys were

conducted throughout the reserve in 2008, 2009, 2015, and 2016 to determine spotted

hyena abundance and distribution. Spotted hyena abundance and density did not vary

among the years. Concerning human populations, spotted hyena abundance increased

further from villages, the border of the reserve, the airfield, and the border post (p<

0.001). Spotted hyena abundance also increased closer to tourism camps (p <0.001).

Spotted hyena distribution was negatively impacted by human interactions like villages

and the border of the reserve but positively impacted by human interactions like tourism

camps located throughout the reserve. Spotted hyenas within the NTGR showed

differences in detection depending on proximity to the type of human settlement the

calling station site was located.

INTRODUCTION Apex predators are by definition at the top of the trophic community, which often leads to

human-wildlife conflict (Stier et al., 2016; R. Woodroffe et al., 2005; Yodzis, 2001).

Sources of conflict can vary from predation on prey populations, competing for resources, and habitat fragmentation (Berger et al., 2001; Ritchie & Johnson, 2009). On the African continent, the rapid human population increase combined with higher per 9

capita resource demands fuel growing land use at the same time people are moving to

more urban areas (Davis, 1965). Therefore, conflict in some forms is likely to increase

but with a shift in more people moving to urban areas, this could lead to a positive impact

on carnivore populations (Karanth & Chellam, 2009).

Globally, human-induced large carnivore declines can be divided into five

different reasons: historical/cultural, valuational, ecological, management system, and

policy process (Clark et al., 1996). Understanding all the factors that may affect carnivore

populations is crucial for establishing more effective management practices. The creation

of policies for predator management can be beneficial by ensuring the involvement of all

stakeholders, but can also inhibit proper management when debates become polarized.

Site-specific carnivore management accounting for political and ecological conditions ensures a greater success with how humans interact with carnivores within their area

(Treves & Karanth, 2003). Predator management takes a major focus in society due to high levels of persecution and their perception as being a nuisance; however, the

management of predators must be done cautiously due to the complex relationship

between predator and prey populations (Mills, 1991). Studies have shown that the

removal of apex predators from a system can lead to trophic cascades that can alter

ecosystems in many ways (Finke & Denno, 2004; Johnson et al., 2007; Prugh et al.,

2009).

Spotted hyenas are one of the 31 largest mammalian carnivores on Earth. Of the

31 species, only six currently has a population trend of stable or increasing (Ripple et al.,

2014). As one of the most abundant predators in southern Africa, with an estimated

global population of 27,000 to 47,000, spotted hyenas are listed as ‘Least Concern’ by the 10

IUCN with an overall decreasing trend in population. Places like western, eastern, and

central Africa have noted a decline in spotted hyena populations while protected areas in

southern and some eastern parts of Africa have stable populations (Bohm & Honer,

2015). To conserve specific declining species, the IUCN established various Specialist

Groups based on taxonomic group i.e. the IUCN SSC Hyaena Specialist Group (HSG).

The HSG strives to ‘contribute and promote the understanding and conservation of the

species in the family Hyaenidae’ (Hyaenid, 2018).

Carnivore home range size is a better predictor of extinction risk than population size because of the risk of contact with humans on the edges of the reserve (R.

Woodroffe, 1998). For spotted hyenas, territory size in the Kalahari depended more on prey availability rather than the size of clans (Mills, 1984). Territories can overlap amongst different clans and lower-ranking hyenas are more likely to travel outside their territories for foraging due to their low status. Spotted hyenas will redistribute themselves if there are areas outside of their territories that are more profitable (Höner et al., 2005).

Woodroffe also noted a strong association between human density and a decline in carnivore populations (Rosie Woodroffe, 2000). However, in Ethiopia, spotted hyena density increased the closer to areas with higher human density (Yirga et al., 2017).

The highly adaptable behavior of hyenas makes its management a much more localized effort. In northern Ethiopia, an area with very low prey populations, spotted hyenas live near human-populated areas with a diet predominantly dependent on anthropogenic food (Yirga et al., 2013). Within this same area, spotted hyenas will change their diet during religious fasting periods and hunt livestock when human food waste is not available (Yirga et al., 2012). However, in the Northern Province of 11

Cameroon, spotted hyenas avoided villages near protected areas due to higher poaching

levels in the area (Croes et al., 2011). In Kenya, the Maasai Mara will kill spotted hyenas, lions, and leopards due to predation on livestock. Due to spotted hyenas only attacking livestock at night, immediately running away after attacks, and Maasai culture of killing lions, African lions were more persecuted than the more frequent livestock predator, spotted hyenas (Kissui, 2008). In South Africa, within Kruger National Park, the presence of human infrastructure has had an impact on the spatial distribution of food, which indirectly affected space-use of spotted hyenas (Belton et al., 2016). This variable

behavior with neighboring human settlements shows spotted hyenas will adapt their

behavior depending on their relationship with humans, whether good or bad.

Historical records within the Northern Tuli Game Reserve, NTGR, suggest

spotted hyena populations were initially very low, with most sightings in the northeastern

portion of the reserve noted in 1985 (McKenzie, 1990). The NTGR was originally intended for cattle grazing until established as a game reserve in the mid-1960s. Due to persecution from previous inhabitants of the area, various predators like spotted hyenas, leopards, and lions were low in abundance (McKenzie, 1990). It has been noted that spotted hyenas are much slower in recolonizing areas where they were previously persecuted than lions (Kruuk, 1972; Smuts, 1982).

In 2008, we began conducting an annual survey to understand the current spotted hyena population density and distribution within the area (Snyman, 2009). Before 2008, the majority of research in the NTGR focused on African lions, black-backed jackals, leopards, cheetahs, and African elephants. The survey was established to further aid management practices in NTGR and aid in understanding carnivore population dynamics. 12

In this chapter, we asked how spotted hyena abundance and distribution is affected by

human settlements, cattle posts, and other landmarks throughout the reserve, and to

assess variation across the years the study was conducted.

METHODS Field Site The Northern Tuli Game Reserve (NTGR) is owned by various private landowners. The

NTGR is located in the eastern tip of Botswana bordered by Zimbabwe to the east and

South Africa to the south. Eastern Botswana contains the highest concentration of

humans per square kilometer in the country (Statistics Botswana, 2015). A veterinary

fence line runs north to south on the western portion of the NTGR. This veterinary fence, along with poorly maintained fences along the Limpopo River to the south and a game fence along the western border of NTGR, is the only fencing within the reserve (Figure

2.1). Although the veterinary fence was implemented to stop the spread of foot and mouth disease, parts of the fence were seen to be damaged and poorly maintained throughout our study. Previous studies within the area have also documented movements of wild dogs (Lycaon pictus), lions (Panthera leo), and cheetahs (Acinonyx jubatus)

through the fences, suggesting the barrier has a minimal effect (Jackson et al., 2012).

The NTGR falls under the ecoregion category of ‘Afrotropical’ with mopane tree

(Colophospermum mopane) woodlands, umbrella thorn (Vachellia tortilis), and groups of

large fever berry trees (Croton megalobotrys) along the rivers (Estes & Greyling, n.d.).

The area is semi-arid with an average annual rainfall of 400 mm. This 72,000 ha property has a wet and hot summer season during November - April and the cool winter season during May - October (Mashatu Game Reserve, 2019). 13

Field Methods & Study Design Spotted hyena calling stations are point-count surveys where playback recordings are

broadcasted for 30 minutes with loudspeakers. For this study, we used the calls of a

buffalo calf (Syncerus caffer) in distress to generate a neutral response from animals in

the area. In contrast to some studies, we did not use recordings of interclan fighting or

spotted hyenas mobbing a lion (Mills et al., 2001; Ogutu & Dublin, 1998). The speakers

were rotated 90 degrees every 15 minutes to create an effective 360-degree calling radius

(Mills et al., 2001). On average, we conducted 3-4 calling stations a night to maximize traveling throughout the reserve and access on various properties.

We did not conduct calibration experiments to determine the likelihood and response distance for our study area because that had been done in similar landscapes.

Graf et al. (2009) determined a response radius of 2.8 km in KwaZulu-Natal Province,

South Africa. In the Okavango Delta, northern Botswana, Cozzi et al. (2013) found an average response radius of 2.93 km (G. Cozzi et al., 2013; Graf et al., 2009). For this study, we based our calling radius to a response probability of 3.2 km following Mills et al. (2001) from Kruger National Park (Mills et al., 2001) (Figure 2.5). This leads us to assume re-sampling the same individuals is unlikely at least for the calling stations done within the same night. We also assumed an equivalent response likelihood throughout the reserve with a probability of 0.61 (Mills et al., 2001).

Surveys were conducted between the winter months of May – June. Each year was treated as a closed population due to the completion of the yearly study within a few days. A study conducted over a short timeframe in which births, deaths, emigration, or immigration have no significant impact can be assumed as a ‘closed population’. Calling 14 stations were conducted only once each year to reduce habituation to calls and maximize response rates to stations (G. Cozzi et al., 2013). The sample sizes were 14 sites in 2008,

17 sites in 2009, and 19 sites in both 2015-2016 (Figure 2.3).

At each calling station, all predators who approached the site were counted but our more extensive data collection on hyenas included the time of start and finish of all sites, numbers observed, the arrival time of each animal, and landscape descriptions.

Spotted hyenas typically travel in clans, so a decision tree was used to avoid double counting throughout the study (Figure 2.2). Our vehicle was equipped with a spotlight that was shined periodically to scan for hyenas. A minimum of three observers were in the vehicle to aid in setup, hyena counting, and data collection. Cubs were not counted in the population estimates because of the low percentage of cub responses within our study and in the studies of Mills et. al. (2001) and Graf et al. (2009).

ArcMap 10.5.1 software was used for mapping and determining the distance between sites and distances to various human settlements. Distances to the nearest human settlement like negative human activity, border of the reserve, and tourism campsites were determined for each calling station site. Here we define negative human activity as anything that could be associated with negative interactions for hyenas, i.e. villages, the border post, and the Limpopo Valley Airfield. Since cattle posts were clustered toward the western portion of the reserve, the average distance to all the cattle posts for each site was determined (Figure 2.4).

Data Analysis 15

The estimates for detection probability ( =0.611) and the area covered by each calling

site (Area = 32.2 km²) are taken from Mills𝜋𝜋� et al., 2001. This study was done in Kruger

National Park, South Africa, which had habitat similar to those found in the NTGR.

We used a probability model to estimate the spotted hyena population size based

on the hyenas that responded to our calling stations(Mills et al., 2001). The objective is to

estimate the overall abundance of the spotted hyenas and the mean of the distribution N.

Let the actual count for spotted hyena near a calling station be N. Let the observed count

for spotted hyenas near a calling station be y.

Since N is a count, N ~ Poisson ( ) and we know the mean: E (N) =μ=λ.

𝜆𝜆 As λ (average number of hyenas throughout the reserve) cannot be constant or

fixed throughout the reserve, λ is expected to have a gamma distribution with parameters

α and β,

1 𝜆𝜆 λ~ Gamma(α,β), ( ) = − , > 0 and for gamma distribution E(λ)=αβ= μ. 𝛽𝛽 𝛼𝛼 − 𝑒𝑒 𝜆𝜆 𝛼𝛼 𝑓𝑓 𝜆𝜆 𝛽𝛽 Γ 𝛼𝛼 𝜆𝜆 The conditional distribution of y|N ~ Binomial (π), where π is the response or

detection probability of the spotted hyenas near the calling sites. Then, y~Negative

Binomial (α,δ), where δ=βπ, E(y) = αδ = αβπ = μπ. Therefore, to get an overall estimate

of the population of hyenas we can equate the sample observed mean to the population

mean since the sample mean is an unbiased estimator of the population mean.

E(y) = μπ = => = 𝑦𝑦� 𝑦𝑦� 𝜇𝜇̂ 𝜋𝜋 We used the following equations to calculate abundance estimates and their

corresponding confidence limits: 16

: = ( ) 𝑥𝑥⁄𝜋𝜋 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 𝐷𝐷 𝑎𝑎 where ‘x’ is the number of individuals seen at a specific site, ‘ ’ is the response probability, and ‘a’ is the total area sampled per site. The proportion𝜋𝜋 of the reserve surveyed was also calculated by dividing the total amount of area surveyed with all calling station sites divided by the total area of NTGR.

The confidence limit was calculated as = ± . ( ) 2.

Statistical analyses were performed using R 3.6.1𝐶𝐶𝐶𝐶 𝐶𝐶𝐶𝐶𝐶𝐶software.𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 𝜇𝜇̂ 𝑆𝑆 𝐸𝐸 𝜇𝜇̂ ∗

Generalized linear mixed effects models were run with years as random effects and various distances to different types of human impacts as fixed effects. Our null model included no fixed or random effects with a Poisson distribution. Two human models were run. One that included calling sites’ distance to the border, the average distance to cattle posts, and distance to negative human activity, and the other combined border and negative human activity to only account for whichever was closest to each calling station site. The tourism model included distance to the nearest tourism campsite for each calling station site as a fixed effect and the last model included all possible fixed human effects.

A two-way interaction model was also run to look into cattle posts to proximity to human activity and how this impacts the number of hyenas detected.

RESULTS

We observed 42 spotted hyenas in 2008, 76 hyenas in 2009, 78 hyenas in 2015, and 54 hyenas in 2016. From these observations spotted hyena estimates varied from 108 to 169 across the NTGR (Table 2.1). Confidence limits indicate there was no significant 17

difference among the different years surveyed. The density of spotted hyenas within the reserve also showed similar results to abundance with confidence limits indicating no significant difference between the years (Figure 2.7).

The concentration of spotted hyena core pockets varied among the years. In 2008,

the majority of hyena density was located within the northern and central portion of the

reserve. The highest density of hyenas remained in the north and central portion of the

reserve in 2009, with more distribution found on the eastern portion. In 2015, there were

various core pockets of high-density areas throughout the reserve. In 2016, the hyena

distribution remained within the core of the reserve with low density on the eastern

portion of the reserve.

Our lowest AIC value came from our model with all possible fixed effects (Table

2.2), with factors like distance to tourism sites and human settlements having a strong

effect (p < 0.001). The model that looked at whichever came closer, human settlements or

the border (All humans), with respect to hyena numbers became one of our worst models

aside from the null model. This could be because some borders can be correlated to

human impact, i.e. those neighboring villages or farms, while other borders of NTGR are

far from direct human influence like game hunting concessions (Figure 2.4). The only

feature that was not related to hyena locations was the distance to cattle posts. This

indicates that the further from the border and negative human settlements the more likely

to have a spotted hyena detection, while the closer we are to tourism camps the more

likely we are to have a spotted hyena detection (Figure 2.8). The predicted amount of

spotted hyenas correlates with the actual amount of hyenas detected. However, the actual 18

amount of hyenas detected had larger limits than the predicted amount at a confidence interval of 95% (Figure 2.9).

To simplify models we compared year as a fixed effect to our null with a value of

less than 1 for the delta AIC further indicating that year has no significant impact and can

be removed from the models. The all humans variable strongly correlated with our

distance to border therefore both distances to border and distance to human settlements

were removed to keep All Humans as the replacement. Since cattle posts had a higher p-

value, a model was created to look at the interaction between cattle posts and all humans.

Cattle Posts and all humans became our best simplified model with an AIC value of

397.8 (Table 2.5). Individually, the model shows cattle posts and distance to humans

have a positive relationship with hyenas being closer. However, when comparing the

relationship between the two, hyenas show a negative relationship (Table 2.6). This

indicates cattle posts and human presence had the most significant impact on spotted

hyena presence in comparison to the other variables.

DISCUSSION Various hyena playback studies have been conducted throughout Africa to be used as a

low-cost estimation method due to their short-time frame in data collection. To continue to understand hyena population trends within the area we recommend to continue conducting playback surveys to monitor population trends as a low-cost method. In this process, we also recommend conducting response calibrations specific to the NTGR to determine if there are any significant differences within the area. Due to not conducting our detection sampling, the probability of detection of a hyena near a calling site might differ from what was found in Kruger National Park; therefore, our variance is inflated 19 with a lower estimation of abundance due to using Mills et al. (2001) response probability. However, although our abundance estimates may be lower than actual numbers, this study indicates a stable hyena population within the reserve that seems to yet be impacted by the increase in African lions within the area. Continuous monitoring of an abundant predator within the area will continue to provide crucial information to better manage the reserve and understand population dynamics.

Unfortunately, our study was only able to compare spotted hyena numbers to human influence within the reserve. Spotted hyena distribution shows core pockets of hyena density within the central portions of the reserve (Figure 2.6). Unfortunately, our data also consists of missing data in certain areas of the reserve, especially in 2016.

Although cattle posts showed no significant impact on spotted hyena amount individually, when combined based on cattle post proximity to humans there was a negative interaction of hyenas predicted. The closer to cattle posts that are close to humans, the less likely we will see hyenas, and the further the calling stations from humans the higher the predicted amount of hyenas (Figure 2.10). Humans do appear to influence the presence of spotted hyena populations, but due to a lack of comparison with

African lions within the area, it is difficult to conclude whether hyenas are inhabiting areas based on response to humans, lion presence, or both. We recommend future studies to look into local lion populations and overlay lion presence with the results from this study to determine which has a larger impact on spotted hyena distribution. In Namibia, it was noted hyenas generally avoided lions unless the lions got too close to their dens in which they would mob to ensure enough time for cubs to seek refuge (Trinkel &

Kastberger, 2005). 20

Spotted hyena clan sizes can also shed significant light on why certain areas show

higher numbers of hyenas. Spotted hyenas are constantly adapting to their environment,

and the number of hyenas within a clan can determine how they interact with their

competing predator, lions. In Etosha National Park in Namibia, spotted hyenas avoided

lions because of their lower population numbers than the lion populations (Trinkel &

Kastberger, 2005). Various studies have shown that the number of hyenas within a clan relative to the number of lions in the area determines whether spotted hyenas will steal lion kills or avoid lions to avoid being killed (Cooper, 1991; Honer et al., 2002; Mills,

1990). A more in-depth look into local clan territory sizes and interactions would provide an enhanced perspective on spotted hyena dynamics to this specific area.

Spotted hyena populations within the area are stable, with proximity to villages, the border post, the airfield, and the border of the reserve showing a negative relationship

with spotted hyena detection. While conducting fieldwork, we noticed a change in

behavior between sites near negative human interactions with spotted hyenas being easily

spooked from turning on our spotlight. These observations go along with our data of less

spotted hyenas are seen the closer sites are to negative human settlements and borders.

However, this does not suggest there are no hyenas in proximity to these areas, but rather

you are less likely to detect them. This is also why conducting calibrations within the

NTGR at different proximities to human settlements could also help determine how much

proximity could impact detection. Mills et al. only conducted calling stations based on

different habitat and other studies that also performed detection calibration did not

account for proximity to humans (G. Cozzi et al., 2013; Mills et al., 2001; Ogutu &

Dublin, 1998). 21

This study has provided the foundation to further understand the complex dynamics within a private game reserve among various types of human impacts. Spotted hyena studies have shown their adaptability and behavior can vary based on the number of competitors, available resources, and human influences. This adaptability to human activity will be crucial in how lions may influence spotted hyena distribution. As African lion populations continue to rise within the reserve, taking a closer look into how their populations overlap within the reserve concerning various types of human settlements in

NTGR will better aid in maintaining a balance between this interspecific competition within limited space.

22

REFERENCES

Belton, L. E., Cameron, E. Z., & Dalerum, F. (2016). Spotted hyaena space use in

relation to human infrastructure inside a protected area. PeerJ, 4, e2596.

https://doi.org/10.7717/peerj.2596

Berger, J., Stacey, P. B., Bellis, L., & Johnson, M. P. (2001). A mammalian predator-prey

imbalance: Grizzly bear and wolf extinction affect avian neotropical migrants.

Ecological Applications, 11(4), 947–960.

Bohm, T., & Honer, O. R. (2015). Crocuta crocuta. International Union for Conservation

of Nature (IUCN). https://doi.org/10.2305/IUCN.UK.2015-

2.RLTS.T5674A45194782.en

Clark, T. W., Curlee, A. P., & Reading, R. P. (1996). Crafting effective solutions to the

large carnivore conservation problem. Conservation Biology, 10(4), 940–948.

Cooper, S. M. (1991). Optimal hunting group size: The need for lions to defend their kills

against loss to spotted hyaenas. African Journal of Ecology, 29(2), 130–136.

Cozzi, G., Broekhuis, F., McNutt, J. W., & Schmid, B. (2013). Density and habitat use of

lions and spotted hyenas in northern Botswana and the influence of survey and

ecological variables on call-in survey estimation. Biodiversity and Conservation,

22(12), 2937–2956.

Croes, B. M., Funston, P. J., Rasmussen, G., Buij, R., Saleh, A., Tumenta, P. N., & de

Iongh, H. H. (2011). The impact of trophy hunting on lions (Panthera leo) and

other large carnivores in the Bénoué Complex, northern Cameroon. Biological

Conservation, 144(12), 3064–3072. 23

Davis, K. (1965). The Urbanization of the Human Population. Scientific American,

213(3), 40–53.

Estes, L., & Greyling, L. (n.d.). Southeastern Africa: South Africa, Mozambique,

Botswana, Zambia, Zimbabwe, Swaziland, Namibia, and Malawi. World Wildlife

Fund. Retrieved November 6, 2019, from

https://www.worldwildlife.org/ecoregions/at0725

Finke, D. L., & Denno, R. F. (2004). Predator diversity dampens trophic cascades.

Nature, 429(6990), 407–410.

Graf, J. A., Somers, M. J., Gunther, M. S., & Slotow, R. (2009). Heterogeneity in the

density of spotted hyaenas in Hluhluwe-iMfolozi Park, South Africa. Acta

Theriologica, 54(4), 333–343.

Honer, O. P., Wachter, B., East, M. L., & Hofer, H. (2002). The response of spotted

hyaenas to long-term changes in prey populations: Functional response and

interspecific kleptoparasitism. Journal of Animal Ecology, 71(2), 236–246.

Höner, O. P., Wachter, B., East, M. L., Runyoro, V. A., & Hofer, H. (2005). The effect of

prey abundance and foraging tactics on the population dynamics of a social,

territorial carnivore, the spotted hyena. Oikos, 108(3), 544–554.

Hyaenid. (2018). IUCN. https://www.iucn.org/commissions/ssc-

groups/mammals/specialist-groups-f-z/hyaenid 24

Jackson, C. R., McNutt, J. W., & Apps, P. J. (2012). Managing the ranging behaviour of

African wild dogs (Lycaon pictus) using translocated scent marks. Wildlife

Research, 39(1), 31-34.

Johnson, C. N., Isaac, J. L., & Fisher, D. O. (2007). Rarity of a top predator triggers

continent-wide collapse of mammal prey: dingoes and marsupials in Australia.

Proceedings of the Royal Society B: Biological Sciences, 274(1608), 341–346.

Karanth, K. U., & Chellam, R. (2009). Carnivore conservation at the crossroads. Oryx,

43(1), 1-2.

Kissui, B. M. (2008). Livestock predation by lions, leopards, spotted hyenas, and their

vulnerability to retaliatory killing in the Maasai steppe, Tanzania. Animal

Conservation, 11(5), 422–432.

Kruuk, H. (1972). The Spotted Hyaena. University of Chicago Press.

Mashatu Game Reserve. (2019). Mashatu Game Reserve—Land of the Giants

(Botswana). https://mashatu.com/

McKenzie, A. A. (1990). Co-operative Hunting in the Black-backed Jackal (Canis

mesomelas). University of Pretoria.

Mills, M. G. L. (1984). The comparative behavioural ecology of the brown hyaena

(Hyaena brunnea) and the spotted hyaena (Crocuta crocuta) in the Southern

Kalahari. Koedoe, 27(2), 237–247.

Mills, M. G. L. (1990). Kalahari hyaenas: Comparative behavioural ecology of two

species. Unwin Hyman Ltd. 25

Mills, M. G. L. (1991). Conservation management of large carnivores in Africa. Koedoe,

34(1), 81–90.

Mills, M. G. L., Juritz, J. M., & Zucchini, W. (2001). Estimating the size of spotted

hyaena (Crocuta crocuta) populations through playback recordings allowing for

non-response. Animal Conservation, 4(4), 335–343.

Ogutu, J. O., & Dublin, H. T. (1998). The response of lions and spotted hyaenas to sound

playbacks as a technique for estimating population size. African Journal of

Ecology, 36(1), 83–95.

Prugh, L. R., Stoner, C. J., Epps, C. W., Bean, W. T., Ripple, W. J., Laliberte, A. S., &

Brashares, J. S. (2009). The rise of the mesopredator. BioScience, 59(9), 779–

791.

Ripple, W. J., Estes, J. A., Beschta, R. L., Wilmers, C. C., Ritchie, E. G., Hebblewhite,

M., Berger, J., Elmhagen, B., Letnic, M., Nelson, M. P., Schmitz, O. J., Smith, D.

W., Wallach, A. D., & Wirsing, A. J. (2014). Status and ecological effects of the

world’s largest carnivores. Science, 343(6167), 1241484–1241484.

Ritchie, E. G., & Johnson, C. N. (2009). Predator interactions, mesopredator release and

biodiversity conservation. Ecology Letters, 12(9), 982–998.

Smuts, G. L. (1982). Lion. MacMillan S.A.

Snyman, A. (2009). Hyena census report 2009. Unpublished.

Statistics Botswana (Ed.). (2015). Population census atlas 2011: Botswana. Statistics

Botswana. 26

Stier, A. C., Samhouri, J. F., Novak, M., Marshall, K. N., Ward, E. J., Holt, R. D., &

Levin, P. S. (2016). Ecosystem context and historical contingency in apex

predator recoveries. Science Advances, 2(5).

https://doi.org/10.1126/sciadv.1501769

Treves, A., & Karanth, K. U. (2003). Human-carnivore conflict and perspectives on

carnivore management worldwide. Conservation Biology, 17(6), 1491–1499.

Trinkel, M., & Kastberger, G. (2005). Competitive interactions between spotted hyenas

and lions in the Etosha National Park, Namibia. African Journal of Ecology,

43(3), 220–224.

Woodroffe, R. (1998). Edge effects and the extinction of populations inside protected

areas. Science, 280(5372), 2126–2128.

Woodroffe, R., Thirgood, S., & Rabinowitz, A. (Eds.). (2005). People and wildlife,

conflict or co-existence? In people and wildlife, conflict or co-existence? (pp. 1–

12). Cambridge University Press.

Woodroffe, Rosie. (2000). Predators and people: Using human densities to interpret

declines of large carnivores. Animal Conservation, 3, 165–173.

Yirga, G., De Iongh, H. H., Leirs, H., Gebrihiwot, K., Deckers, J., & Bauer, H. (2012).

Adaptability of large carnivores to changing anthropogenic food sources: Diet

change of spotted hyena (Crocuta crocuta) during Christian fasting period in

northern Ethiopia: Adaptability of large carnivores to changing anthropogenic

food sources. Journal of Animal Ecology, 81(5), 1052–1055. 27

Yirga, G., Ersino, W., De longh, H. H., Leirs, H., Gebrehiwot, K., Deckers, J., & Bauer,

H. (2013). Spotted hyena (Crocuta crocuta) coexisting at high density with

people in Wukro district, northern Ethiopia. Mammalian Biology, 78, 193–197.

Yirga, G., Leirs, H., De Iongh, H. H., Asmelash, T., Gebrehiwot, K., Vos, M., & Bauer,

H. (2017). Densities of spotted hyaena (Crocuta crocuta) and African golden wolf

(Canis anthus) increase with increasing anthropogenic influence. Mammalian

Biology, 85, 60–69.

Yodzis, P. (2001). Must top predators be culled for the sake of fisheries? Trends in

Ecology & Evolution, 16(2), 78–84.

TABLES

Table 2.1: Estimate of spotted hyenas per year and 95% confidence intervals for spotted hyena population density within the Northern

Tuli Game Reserve (NTGR).

Year Density LCI UCI Population Estimate LCI UCI % of reserve surveyed

2008 0.15 0.05 0.25 114 40 188 57%

2009 0.23 0.10 0.36 169 73 266 64%

2015 0.21 0.10 0.32 156 71 240 71%

2016 0.14 0.06 0.23 108 47 168 71% 28

Table 2.2: Generalized Linear Mixed-Effects Models for spotted hyena abundance as a function of distance to sites of human activity.

The year was included as a random effect in all models. Models were categorized between “positive” human interactions like Tourism

Camps or “negative” human interactions like proximity to the border, cattle posts, and human settlements i.e. villages, border post,

and the airfield.

Model AIC ∆AIC

Hyena Abundance ~ Tourism + Border + Cattle Posts + Human Settlements 385.13 0

Hyena Abundance ~ Border + Cattle Posts + Human Settlements 399.02 13.89

Hyena Abundance ~ Tourism Camps 399.87 14.74

Hyena Abundance ~ Nearest Border or Human Settlement 407.21 22.08

Hyena Abundance ~ Null 420.06 34.93

29

30

Table 2.3: Our best-fitted model was Hyena abundance with fixed effects of distance to tourism sites, border of the reserve, cattle posts, and negative human settlements. Cattle Posts had a p- value of 0.07 therefore, we cannot conclude that a significant difference exists between the parameter and the likelihood of detecting spotted hyenas.

Best Model

β (Parameter) Estimate P-value: Pr(|z|)

(Intercept) 0.71 0.02*

Tourism Sites -0.13 P < 0.001***

Border 0.09 0.003**

Cattle Posts 0.02 0.06

Human Settlements 0.05 P<0.001***

Significance codes: *** = 0, ** = 0.001, * = 0.01

Table 2.4: Candidate covariates used in linear fixed models with their respective definitions.

Covariate Definition

Humans Straight-line distance from nearest negative human interaction (i.e. Villages, Pont Drift Border Post, and

Limpopo Airfield) to calling station site

Border Straight-line distance from the nearest border of Northern Tuli Game Reserve to each calling station site.

Camps Straight-line distance from nearest ecotourism camp to calling station site.

Cattle Posts The average distance from all cattle posts to each calling station site.

All Humans Straight-line distance from the nearest Village, Border Post, Airfield, or Border of the Northern Tuli Game

Reserve to each calling station site.

31

Table 2.5: A run of simplified models were tested to identify which variables had a higher effect. Cattle posts had an insignificant effect but when added with all human impacts as a two-way interaction it became our best model. Year also showed no significant difference between our null model which backs up our results from our abundance estimates indicating year is not a significant effect on spotted hyena abundance.

Model AIC ∆AIC

Hyena Abundance ~ Humans * Cattle Posts 397.80 0

Hyena Abundance ~ Tourism Camps 399.87 2.07 Hyena Abundance ~ Null Model 420.06 22.26 Hyena Abundance ~ Year 421.03 23.23 32

Table 2.6: Our best-fitted model was hyena abundance with humans and cattle posts as two-way interactions. Individually, the closer to these effects the more hyenas but when looking at the interactions together you count more hyenas further away.

Best Model

β (Parameter) Estimate P-value: Pr(|z|)

(Intercept) 2.5 P<0.001***

Humans -0.27 0.03*

Cattle Posts -0.06 P<0.001***

Humans:Cattle 0.02 0.003**

Significance codes: *** = 0, ** = 0.001, * = 0.01 33

FIGURES

Figure 2.1: Location of the Northern Tuli Game Reserve, NTGR, located in the eastern tip of Botswana. There are no fences along the southern and eastern portion of the reserve. The southern border of NTGR is South Africa and the eastern border being Zimbabwe.

Poorly maintained fences run along the western border of the reserve and a veterinary fence within the center implemented to prevent the spread of foot and mouth disease. 34

35

Figure 2.2: Decision tree used to prevent double detection at calling station sites for abundance surveys of spotted hyenas, NTGR, Botswana.

Figure 2.3: Spotted hyena calling station sites placed throughout NTGR, Botswana in 2008 (n=14), 2009(n=17), 2015(n=19), and

2016(n=19). 36

Figure 2.4: Various human activity throughout NTGR, Botswana providing both positive and negative human-wildlife interactions. 37

Figure 2.5: Spotted hyena calling station locations within NTGR, Botswana. These were set out as point counts with an effective radius of 3.2 km covering a total area of 32.2 km² (Mills et al., 2001). 38

Figure 2.6: Estimated spotted hyena density (hyenas/km²) distribution throughout the Northern Tuli Game Reserve (NTGR) for the 39

four years of data collection (2008, 2009, 2015, and 2016).

40

Figure 2.7: Chart showing spotted hyena abundance and density in their respective years with confidence intervals. All confidence intervals overlap indicating the difference between the years is not statistically significant.

41

Figure 2.8: Linear relationship between predicted hyena amount and (A) distance to the border of the reserve, (B) distance to negative human settlements like villages, border post, and airfield

(C) distance to cattle posts (D) distance to tourism camps.

42

Figure 2.9: Linear relationship between spotted hyena amount and (A) distance to the border of the reserve, (B) distance to negative human settlements like villages, border post, and airfield (C) distance to cattle posts (D) distance to tourism camps.

43

Figure 2.10: An increase in hyena amounts is predicted if calling stations are located further from cattle posts near humans than if stations were conducted closer to the cattle posts near humans.

44

CHAPTER 3: OCCUPANCY ANALYSIS OF SPOTTED HYENAS (CROCUTA

CROCUTA) AND BROWN HYENAS (HYAENA BRUNNEA) IN THE NORTHERN

TULI GAME RESERVE, BOTSWANA

ABSTRACT

Since the 1960s, the property in what is now the Northern Tuli Game Reserve (NTGR) has shifted away from cattle ranching and towards increasing ecotourism. As a result, much of the

predator populations have slowly begun increasing. The reappearance of apex predators like

spotted hyenas (Crocuta crocuta) and African lions (Panthera leo) has led to trophic cascades between the apex predators and mesocarnivores within the area. For further understanding of the dynamic between spotted hyenas and brown hyenas (Hyaena brunnea) in the area, we conducted an occupancy analysis within the Mashatu Game Reserve situated within NTGR. The occupancy analysis used the bycatch data from camera trap studies conducted during 2013 and 2014 which focused on leopards (Candelario, 2015). We set out 9 camera trap sites during a 3 week period in

2013 and 23 sites in 2014 for 9 weeks. Site usage for both brown and spotted hyenas was

analyzed independently with covariates including several types of human activity and terrain.

There were more brown hyena detections within the EcoTraining concession located in the

western portion of the reserve than in the Central concession. Spotted hyenas were detected in all

sites except for one site in 2014 and 5 sites in 2013. Overall, spotted hyenas had a lower average

distance to human settlements than brown hyenas. Brown hyenas showed a preference for river

sites over upland habitat, possibly due to better coverage for hunting and protection from predators. Riverine habitat is declining within the reserve, and with the low abundance of brown

hyenas, future work is needed to ensure which areas are crucial for the continuation of

conservation for this near-threatened species.

45

INTRODUCTION

Brown hyenas (Hyaena brunnea) are a smaller and less colonial relative of the spotted hyena

(Crocuta crocuta). Unlike their larger relative, they are predominantly scavengers. Brown hyenas have a wide-ranging diet in which they depend on various items from large mammal remains to wild fruits and insects. Most food items eaten by brown hyenas are typically sufficient to cover the energetic needs of one hyena, this is why brown hyenas will travel alone rather than in clans like spotted hyenas (Mills, 1990).

Brown and spotted hyenas will often compete if they are scavenging within the same area however, spotted hyenas are the more dominant of the two species and will take the kills (Mills,

1984). Although brown and spotted hyena diets do not entirely overlap, brown hyenas do benefit from consuming the remains of carcasses from spotted hyena and lion kills (Mills, 1990).

Although their distribution overlaps, few areas contain significant populations of both species.

Within the Kalahari National Park in western Botswana, brown hyena populations were thriving due to low numbers of spotted hyena populations in the area (Mills, 1990). These interactions suggest the two species may have difficulty coexisting unless brown hyenas are scavenging from spotted hyena kills. However because spotted hyenas can respond aggressively towards brown hyenas even if food is not present, it was noted brown hyenas will avoid areas where spotted hyenas frequently visit (Mills, 1984).

Brown hyenas are currently ranked as “near threatened” by the IUCN red list, but still showing a stable population (Wiesel, 2015). The reintroduction of brown hyenas within fenced reserves showed a preference for proximity to roads (Welch et al., 2016). This preference, if shown outside of protected areas, could lead to negative impacts on brown hyena populations.

46

Edwards et al. (2019) identified one of the largest populations of brown hyenas within an enclosed reserve in Namibia.

In Botswana, the government will compensate for the loss of livestock only if it was killed by predators under the “Big Five” category. Therefore, farmers who lose livestock to brown hyenas are not compensated which can lead to underreporting the extent of brown hyena- livestock conflicts (Schiess-Meier et al., 2007). Although it is illegal to kill wildlife in Botswana, farmers are allowed to kill predators if their livestock or wellbeing is in danger. Farmers typically view brown hyenas as a threat and persecute brown hyenas who predominantly scavenge on old carcasses (Maude, 2005). This conflict between pastoralists and brown hyenas has led to local extinctions further declining populations (Wiesel, 2015).

We studied hyenas on the Mashatu Game Reserve which is located within a much larger game reserve, the Northern Tuli Game Reserve (NTGR). NTGR is a privately-owned game reserve established in 1964 where various landowners came together to retire their cattle operations and converted to wildlife conservation. Although likely historically found in the region, brown hyenas did not occur in the NTGR in 1964 with only scarce sightings of spotted hyenas (McKenzie, 1990). As the reserve gradually changed towards ecotourism, predator populations slowly began increasing. Human persecution of large predators during the cattle grazing years left the area with few large predators, leading to greater populations of older impala and a booming population of black-backed jackals (Canis mesomelas) (McKenzie, 1990).

The reserve contains minimal fencing but has a veterinary fence running along the center of the reserve and a Game fence along the western border of the reserve. Both fences are poorly maintained with some areas trampled by wildlife. Sightings of wildlife physically crossing these fences indicate they are not effective barriers to wildlife movement (Jackson et al., 2012).

47

The NTGR is located in eastern Botswana with the Shashe River on the eastern edge of

the reserve and the Limpopo River along the South. Previous wildlife research conducted in the

area has focused on various animals like African lions (Panthera leo), Black-backed jackals,

African elephants (Loxodonta africana), Cheetahs (Acinonyx jubatus), and Leopards (Panthera

pardus). Minimal research has been done looking at spotted hyena populations within the area,

and no research has been done on brown hyenas here. With brown hyenas sighted within the

reserve in the last few years, it is crucial to begin studying the population dynamics of the very

reclusive brown hyena. Little is known about current brown hyena presence or distribution in the

reserve. In this chapter, we asked how spotted hyena presence and various habitat types affect

brown hyena occupancy within the Mashatu Game Reserve.

METHODS

Camera trapping is a non-invasive method that allows scientists to document wildlife without

disturbance and little cost. To document the interactions between both hyena species, camera

traps allow us to capture even the more reclusive species, brown hyenas.

Study Sites

The largest parcel in the NTGR is the Mashatu concession. Mashatu consists of 29,000 ha, about

40% of the entire reserve. There are two sections within the concession, the central portion

located within the core of the reserve and a western portion of about 3,300 ha adjacent to the

private farm in the area, known as EcoTraining. Various private farms to the south surround

NTGR and Talana Farms is located west of EcoTraining camp.

Data Collection

48

In 2013, a camera trapping study was conducted within the EcoTraining portion of Mashatu that looked at the effect of baited and non-baited camera trapping sites (Satterfield et al., 2017).

Moultrie M80 Game Spy Cameras were placed along roads or game trails at approximately 0.5m off the ground. To allow full-body shots of animals, the cameras were placed 3-6m away from the paths and at an angle to allow ample time to capture the individual within multiple shots.

Cameras were housed within a small metal container, specific to the camera model, to ensure no blockage of sight and tied in place with cable ties. Although there were multiple occurrences of elephants tampering with cameras, the camera-box-cable system allowed for at least the retrieval of memory cards if the cameras were destroyed. To minimize the likelihood of elephants tampering with cameras, cameras were not placed in areas that showed significant elephant activity.

In 2014, a second camera trapping study was conducted both in the EcoTraining and

Central concessions of Mashatu. Candelario (2015) looked at leopard site usage within the

Mashatu Game Reserve. Both the 2014 and 2013 studies were combined to look into spotted and brown hyena presence within the areas surveyed in the Mashatu Game Reserve. Camera traps were placed similarly to the 2013 control sites, with an even distribution on a grid layout of the area (Figure 3.2). The 2014 study had cameras in both the EcoTraining portion of Mashatu,

3,300 ha and the Central portion of Mashatu, 10,000 ha. Each location was placed in a 2.5 km x

2.5km grid with one camera within each grid and at least 1 km apart from any other camera.

Twenty-five camera-trapping sites were used in 2014. Two of the 25 cameras malfunctioned within the Central portion, which left 8 cameras within the EcoTraining Concession and 15 cameras within the Central Mashatu Concession.

49

For the purpose of this study, only the control camera sites from 2013 were used to

compare to the non-baited camera sites from 2014. In 2013, 10 un-baited sites were conducted

but only nine were used for this study due to a camera malfunction. All sites were spaced at

approximately 1 km apart (Figure 3.2). The 2013 cameras were set out for approximately three

weeks between 12 June – 2 July and 2014 cameras were set out for approximately 9 weeks from

26 May – 28 July. Both sampling periods were conducted in a short period to assume each year

are individual, closed populations (Karanth & Nichols, 1998). Closed populations are defined as

a sampling effort conducted over a period short enough to minimize births, deaths, emigration, or

immigration events.

In 2014, cameras were replaced with a more updated version, Moultrie M80 XT active

PIR cameras. The newer cameras provided clearer images with shorter exposure times. However,

the 2014 cameras were set to similar features as 2013 to ensure similar quality. Cameras were set

at 5-second delay time with the maximum shot sequence. This meant in 2013 the cameras were a

3-shot model and 2014 a 4-shot model. This extra image, however, did not have a significant impact because our study focuses on presence-absence and the animal was always captured within the first three shots. Cameras were active 24 hours/day in both years and visited every four days after placement to ensure there was sufficient battery life, memory storage, and no disturbance to cameras from animals or people.

Data Analysis

Site usage for both brown and spotted hyenas was analyzed independently for 2013 and 2014.

Images of brown and spotted hyenas were separated by site. The days were split into night and

day within 24 hours between the sunset of one day to the sunset of the following day. Individuals spotted between 6 am to 6 pm were labeled as found in the day and those found outside of this

50

period were categorized as found at night. Our study excluded images from the last day of the

study due to the removal of cameras at various times. Independent events of detection were

separated by one hour at a site. This reduces the probability of counting the same individual

more than once on the same day (Candelario, 2015; Jiménez et al., 2010; Kawanishi & Sunquist,

2004). However, due to proximity between camera sites, there is still a possibility of detecting

the same individual at multiple sites, so our study will solely focus on site usage of both brown

and spotted hyenas.

All maps were created using ArcMap 10.5.1. Distances were determined by calculating a straight distance between each site of each year and various human settlements. Human settlements consisted of the border of the NTGR, tourism campsites, cattle posts, and negative

human activity for wildlife. A table describing the definition for each covariate is listed in Table

3.1. Each distance covariate was used in our site-specific occupancy models.

A binomial presence/absence occupancy model was created to determine site usage

throughout the two areas of the reserve. Spotted and brown hyena occupancy models were run

separately. Each occupancy model had various covariates determining occupancy, and detection

was set to one. The spotted hyena null model had individual sites as our occupancy while brown

hyenas had both occupancy and detection set to one due to few detections not allowing the model

to run (Table 3.2 and 3.3). Statistical analyses were performed using R 3.6.1 software. AICc was

not calculated due to the uncertainty of the effective sample size for this study (MacKenzie et al.,

2017).

Detections were plotted based on the different years and areas within the reserve to look at the ratio of spotted versus brown hyena detections at each site (Figures 3.4 and 3.5). All the sites that detected spotted hyenas were averaged together based on each covariate and compared

51 to the average distance from the sites that did not detect spotted hyenas. The same was done for brown hyenas for both years (Figures 3.7 and 3.8). Camera trap sites were labeled as located in either river or upland habitats. The percentage of river and upland sites that had detections were graphed out for each hyena species for 2013, 2014, and EcoTraining or Central areas (Figure

3.9).

RESULTS

In 2013, spotted hyenas were detected in four of the nine sites and brown hyenas detected at only two sites (Figure 3.4). For 2014, spotted hyenas were detected in 19 of the 23 sites and brown hyenas detected at only six sites (Figure 3.5). Brown hyenas were only detected in one site located in the central portion of the reserve. This same site also had no spotted hyena detections and was near an old abandoned farm classified as a river site.

Our occupancy models indicated our null model to be our best fitting for brown hyenas and our human model best fitting for spotted hyenas (Tables 3.2 and 3.3). However, ∆AIC among all models were very minimal of 0.000046 for spotted hyenas and 2.00 for brown hyena models. Due to the low levels of detections, especially for brown hyenas, there is not enough data to conclude which covariate has the most impact on occupancy.

Our surveys detected a higher number of brown hyenas within the EcoTraining concession located in the western portion of the reserve than in the central portion (Figure 3.5).

Sites that had large amounts of spotted hyena detections had zero to few detections of brown hyenas and those with many brown hyena detections had minimal to no spotted hyena detections

(Figure 3.10).

52

Distances to various covariates like the border, tourism camps, humans, and cattle posts

with respect to sites that either detected or did not detect either species of hyena showed spotted hyenas with closer distance than brown hyenas (Figure 3.8). This could be that brown hyenas, being a lesser species, are even less likely to be detected near humans than spotted hyenas.

DISCUSSION

Spatial occupancy differences between brown and spotted hyenas could be due to EcoTraining

having various bordering rivers, the Motloutse, Shashe and Limpopo River, providing crucial

riverine habitat for brown hyena refugia. Two sites, which had similar amounts of detections of

both spotted and brown hyenas, were both river sites. Spotted hyenas are still more common than

brown hyenas in the reserve. The large fever berry trees (Croton megalobotrys) create a

protective riverine habitat for various carnivores within the area, especially for brown hyenas.

However, O’Connor (2010) has noted changes in forested areas along the Limpopo due to

drought stress and elephant impacts. These changes in habitat could lead to a decline of the

already near-threatened brown hyenas. Unfortunately, mesocarnivores are not studied to the

extent of top predators, with only black-backed jackals as the only mesocarnivore studied within

NTGR (McKenzie, 1990).

Although brown hyenas were detected predominantly within the EcoTraining concession,

EcoTraining also holds various on-foot excursions within the western edge of NTGR where most

brown hyenas were detected. Williams (2018) noted on foot activity from humans led to a

negative interaction with brown hyenas while vehicle activity had the opposite influence.

Although our occupancy models could not determine what impacted occupancy, further studies

looking into the influence of ecotourism sites and villages on brown hyenas would be crucial for

future management.

53

We would recommend if this study were to be conducted again, to have cameras taking

images in a continuous setting. Therefore the chance of an animal not setting off the camera sensor is minimized. For this study, camera trap sites were not baited, which could have influenced such low numbers of detections. Satterfield et al. (2017) noted baited sites improved detectability within the area for both spotted and brown hyenas. Although the camera traps were initially deployed in 2013 for a baited camera trap study and 2014 for a leopard occupancy study, this study has shown that camera trapping data can be useful even for analysis that were not originally intended (Edwards et al., 2019).

It is especially crucial with the decline of riverine habitat within the area for further

conservation. Spotted hyenas did not show major preferences in habitat within the study with

detections throughout the reserve. However, the areas where brown hyenas were present were

near humans and riverine habitat, which provide refugia from predators like spotted hyenas and

lions. Steps towards conserving these riverine areas within the NTGR will further preserve

brown hyena presence and aid in the decline of their population. Having abundant apex predators

like spotted hyenas and lions help maintain mesopredator control and avoid possible

overpopulation of mesocarnivores (Prugh et al., 2009). A balance in the diversity of species both

apex and mesocarnivores prevents over predation on certain prey communities lessening the

impact on prey community dynamics (Finke & Denno, 2004). As human populations continue to

change landscapes toward agriculture it is important to understand the dynamics of the various

trophic levels within the area and how to properly manage the survival of all species, both apex predators and mesocarnivores.

54

REFERENCES

Candelario, L. R. (2015). Assessing the causes, impacts, and future implications of increased use

of preserve boundaries by leopards on research, tourism, and human-wildlife interactions

in Southern Africa. University of Georgia.

Edwards, S., Noack, J., Heyns, L., & Rodenwoldt, D. (2019). Evidence of a high-density brown

hyena population within an enclosed reserve: The role of fenced systems in conservation.

Mammal Research, 64(4), 519–527.

Finke, D. L., & Denno, R. F. (2004). Predator diversity dampens trophic cascades. Nature,

429(6990), 407–410.

Jackson, C. R., McNutt, J. W., & Apps, P. J. (2012). Managing the ranging behaviour of African

wild dogs (Lycaon pictus) using translocated scent marks. Wildlife Research, 39(1), 31–

34.

Jiménez, C. F., Quintana, H., Pacheco, V., Melton, D., & Tello, G. (2010). Camera trap survey of

medium and large mammals in a montane rainforest of northern Peru. Rev. peru. biol.,

17(2), 191–196.

Karanth, K. U., & Nichols, J. D. (1998). Estimation of tiger densities in India using photographic

captures and recaptures. Ecology, 79(8), 2852–2862.

Kawanishi, K., & Sunquist, M. E. (2004). Conservation status of tigers in a primary rainforest of

Peninsular Malaysia. Biological Conservation, 120(3), 329–344.

55

MacKenzie, D. I., Nichols, J. D., Royle, J. A., Pollock, K. H., Bailey, L., & Hines, J. E. (2017).

Occupancy estimation and modeling: Inferring patterns and dynamics of species

occurrence. Elsevier.

Maude, G. (2005). National park and a neighbouring community cattle area in Botswana.

University of Pretoria.

McKenzie, A. A. (1990). Co-operative hunting in the black-backed jackal (Canis mesomelas).

University of Pretoria.

Mills, M. G. L. (1984). The comparative behavioural ecology of the brown hyaena (Hyaena

brunnea) and the spotted hyaena (Crocuta crocuta) in the southern Kalahari. Koedoe,

27(2), 237–247.

Mills, M. G. L. (1990). Kalahari hyaenas: Comparative behavioural ecology of two species.

Unwin Hyman Ltd.

O’Connor, T. (2010). Transformation of riparian forest to woodland in Mapungubwe National

Park, South Africa, between 1990 and 2007. Austral Ecology, 35(7), 778–786.

Prugh, L. R., Stoner, C. J., Epps, C. W., Bean, W. T., Ripple, W. J., Laliberte, A. S., &

Brashares, J. S. (2009). The rise of the mesopredator. BioScience, 59(9), 779–791.

Satterfield, L. C., Thompson, J. J., Snyman, A., Candelario, L., Rode, B., & Carroll, J. P. (2017).

Estimating occurrence and detectability of a carnivore community in eastern Botswana

using baited camera traps. African Journal of Wildlife Research, 47(1), 32–46.

56

Schiess-Meier, M., Ramsauer, S., Gabanapelo, T., & König, B. (2007). Livestock predation—

insights from problem animal control registers in Botswana. Journal of Wildlife

Management, 71(4), 1267–1274.

Welch, R. J., Tambling, C. J., Bissett, C., Gaylard, A., Müller, K., Slater, K., Strauss, W. M., &

Parker, D. M. (2016). Brown hyena habitat selection varies among sites in a semi-arid

region of southern Africa. Journal of Mammalogy, 97(2), 473–482.

Wiesel, I. (2015). The IUCN Red List of Threatened Species [IUCN Red List of Threatened

Species]. Parahyaena brunnea. https://dx.doi.org/10.2305/IUCN.UK.2015-

4.RLTS.T10276A82344448.en

Williams, K. (2018). A unique symbiosis between farmers and brown hyenas: Feature - the

brown hyena. Farmer’s Weekly, 2018(18026), 34–35.

TABLES Table 3.1: Candidate covariates used in survey-specific occupancy models. Each covariate calculated the nearest distance to each camera trapping site. Negative human interactions for hyenas included Humans, Border, and Cattle posts while positive interactions included camps. Habitat and Location were not categorized as negative or positive because they were not considered human interactions.

Covariate Definition

Humans Straight-line distance from nearest negative human interaction (i.e. Villages, Pont Drift Border Post, and

Limpopo Airfield) to the camera trap site

Habitat Sites were categorized as “River” if located within 100m of a river or “Upland” if further than 100m.

Location Sites categorized as “EcoTraining” or “Central” depending on location.

Border Straight-line distance from the nearest border of Northern Tuli Game Reserve to each camera trap site.

Camps Straight-line distance from the nearest ecotourism camp to each camera trap site.

Cattle Posts The average distance from all cattle posts to each camera trap site.

57

Table 3.2: Survey-specific model selection results for spotted hyenas. Each model had detection as fixed with occupancy changing based on the different covariates stated in Table 3.1.

Model AIC ∆AIC AIC weight Cumulative weight Number of Parameters

Ψ(HUM) p(.) 1354.26 0.00 2.5e-01 0.25 3

Ψ(HAB) p(.) 1354.26 0.000046 2.5e-01 0.50 3

Ψ(LOC) p(.) 1354.26 0.000046 2.5e-01 0.75 3

Ψ(BOR) p(.) 1354.26 0.000050 2.5e-01 1.00 3

Ψ(Site) p(.) 1396.26 42.000035 1.9e-10 1.00 24

HUM = Human, HAB = Habitat, LOC = Location, and BOR = Border. 58

Table 3.3: Survey-specific model selection results for brown hyenas. Each model had detection as fixed with occupancy changing based on the different covariates stated in Table 3.1.

Model AIC ∆AIC AIC weight Cumulative weight Number of Parameters

Ψ(.) p(.) 506.35 0.00 0.40 0.40 2

Ψ(HUM) p(.) 508.35 2.00 0.15 0.55 3

Ψ(LOC) p(.) 508.35 2.00 0.15 0.70 3

Ψ(HAB) p(.) 508.35 2.00 0.15 0.85 3

Ψ(BOR) p(.) 508.36 2.00 0.15 1.00 3

HUM = Human, HAB = Habitat, LOC = Location, and BOR = Border. 59

FIGURES

Figure 3.1: Northern Tuli Game Reserve (NTGR) in eastern Botswana. NTGR is a game reserve owned by various individuals and/or corporations. These are known as “Freeholds” in Botswana and represent a very small portion of land area. Commercial activities are referred to as concessions. One of the largest concession of land is managed under Mashatu Game Reserve. The Mashatu game reserve contains two properties, the southwestern part known as the EcoTraining concession and the central portion known as the 60

Central Concession.

Figure 3.2: Outline of the camera trapping site layout within the Mashatu Concession for 2013 and 2014. Ten sites were laid out in

2013 and 23 sites in 2014. In 2014, sites were placed both in the EcoTraining area, done in 2013, and the central portion of the 61

reserve.

Figure 3.3: NTGR map outlining various human activities throughout the reserve. Tourism sites were categorized as neutral interactions with wildlife as they are ecotourism camps and the negative interactions consisted of cattle posts, the border of the reserve, villages, and government sites. 62

Figure 3.4: A boxplot layout of both spotted and brown hyena detections in 2013 within the EcoTraining concession. Only two sites detected brown hyenas, one in the river and the other in upland habitat. Site CT5 was the only site that had both species detected in

2013. 63

Figure 3.5: Boxplot layout of all 2014 camera trapping sites separated by location, i.e. EcoTraining and Central concessions of the

Mashatu Game Reserve. Brown hyenas were predominately detected in EcoTraining, the western portion of the reserve, than in the

Central concession. Spotted hyenas were detected in almost all sites except three sites in the EcoTraining area. 64

Figure 3.6: Camera trap sites within the Mashatu Game reserve with both 2013 and 2014 camera trap detections. Sites varied with only one species detected both, or none. Overall, brown hyenas had the lowest number of detections in both years. 65

Figure 3.7: The 2013 Camera trap sites whether detected or not detected with an average distance to various covariates, i.e.: A)

Average distance of camera traps to the border of NTGR, B) Average distance to nearest ecotourism campsites, C) Average distance to nearest negative human settlements, D) Average distance to all cattle posts for each site. 66

Figure 3.8: The 2014 Camera trap sites whether detected or not detected with an average distance to various covariates, i.e.: A)

Average distance of camera traps to the border of NTGR, B) Average distance to nearest ecotourism campsites, C) Average distance to nearest negative human settlements, D) Average distance to all cattle posts for each site. 67

Figure 3.9: A) Percentage of camera trap sites that detected both species of hyenas in 2013. B) Percentage of camera trap sites that detected both species of hyenas in 2014. C) Percentage of camera trap sites only in the EcoTraining concession that detected both species of hyenas. D) Percentage of camera trap sites only in the Central concession that detected both species of hyenas. All figures 68

show a percentage of detections within their respective habitat types, i.e. Upland and River.

Figure 3.10: Spotted and brown hyena detections from both 2013 and 2014 based on habitat type, i.e. River and Upland. Sites with large numbers of spotted hyena detections had no brown hyena detections and high brown hyena detections had few or no spotted hyena detections. 69

70

CHAPTER 4: DISCUSSION The Northern Tuli Game Reserve (NTGR) provides a unique look into private game

management in a country where the government has reserved approximately 38% of its land to

wildlife conservation like national parks and game reserves (Winterbach et al., 2014). Being a

privately-owned ecotourism site and surrounded by a matrix of villages, grazing land, and

agricultural farms, the wildlife within NTGR is impacted by various types of human

infrastructures. Home to various wide-ranging large carnivores like African lions (Panthera leo),

spotted hyenas (Crocuta crocuta), leopards (Panthera pardus), and cheetahs (Acinonyx jubatus) the NTGR requires proper management to prevent human-wildlife conflicts with local and neighboring human settlements.

Currently, research projects have focused on more charismatic species within the area with minimal projects done on spotted and brown hyenas. Initially, with the lack of large predators within NTGR, lack of trophic cascades allowed black-backed jackals (Canis

mesomelas) to congregate into small packs to take down larger prey (McKenzie, 1990).

However, the conservation of only one predator does not guarantee the removal of these cascading effects. Studies have shown the importance of maintaining predator diversity will dampen trophic cascades due to competition among predators lessening the impact on herbivore communities (Finke & Denno, 2004). Each African predator holds a unique niche within the ecosystem with lions predominantly eating larger prey and spotted hyenas being able to eat medium to smaller prey (Mills, 1990; Winterbach et al., 2014). Even though spotted hyenas have a wide-ranging diet, Mills (1990) noted lions still account for about 70% of kills with similar prey species. Competition between lions and hyenas can vary depending on prey availability, the

71 number of each species within the area, and the presence of male lions (Honer et al., 2002;

Kruuk, 1972; Mills, 1990; Watts & Holekamp, 2009).

Our calling station surveys have shown spotted hyena populations have remained stable from 2008-2016. Although their population density has not had a large change, there have been changes in spatial usage. Spotted hyena distribution showed a negative relationship between villages and the border of the reserve and a positive relationship towards cattle posts and ecotourism camps. In Kruger National Park, South Africa, it was also noted human infrastructures influenced spotted hyena space use (Belton et al., 2016). Šálek et al. (2014) noted an increase in population but a decline in home ranges for different carnivore species in more urban habitats. Although the NTGR is not an urban environment, there are still different types of human interactions that need to be looked at as to how it influences the different predators within the area. There is also a need for maintaining a balance in the abundance of species. In Namibia, they noticed when hyenas outnumbered lions they were more likely to mob lions even if food is not readily available (Trinkel & Kastberger, 2005).

Not only is there a demand for further research on spotted hyena and lion population management, but also how top predators are impacting mesocarnivores like brown hyenas. The bycatch data from camera trap studies within the Mashatu Game Reserve have shown a clear distinction in areas where spotted and brown hyenas were detected. The only areas that detected both species at equal levels were river sites, habitats that are highly preferable for brown hyenas due to better areas for refugia. However, these riverine habitats in NTGR are slowly transforming into woodland habitats, which could lead to the further decline of animals who depend on riverine habitats like brown hyenas (O’Connor, 2010).

72

This project aims to serve as a building block for further research to assist in the conservation of all wildlife within the area. Our results have shown core pockets of spotted hyena population distribution as well as areas brown hyenas were detected most than other areas.

Further research is needed to better understand clan dynamics throughout NTGR but a focus could be on the depleting riverine habitat along the Limpopo River and a comparison on the small number of detections within the central portion of the reserve. A more in-depth project is also needed to understand clan dynamics among spotted hyenas and how the NTGR compares to other reserves both private and public.

73

REFERENCES Belton, L. E., Cameron, E. Z., & Dalerum, F. (2016). Spotted hyaena space use in relation to

human infrastructure inside a protected area. PeerJ, 4, e2596.

https://doi.org/10.7717/peerj.2596

Finke, D. L., & Denno, R. F. (2004). Predator diversity dampens trophic cascades. Nature,

429(6990), 407–410.

Honer, O. P., Wachter, B., East, M. L., & Hofer, H. (2002). The response of spotted hyaenas to

long-term changes in prey populations: Functional response and interspecific

kleptoparasitism. Journal of Animal Ecology, 71(2), 236–246.

Kruuk, H. (1972). The spotted hyaena. University of Chicago Press.

McKenzie, A. A. (1990). Co-operative hunting in the black-backed jackal (Canis mesomelas).

University of Pretoria.

Mills, M. G. L. (1990). Kalahari hyaenas: Comparative behavioural ecology of two species.

Unwin Hyman Ltd.

O’Connor, T. (2010). Transformation of riparian forest to woodland in Mapungubwe National

Park, South Africa, between 1990 and 2007. Austral Ecology, 35(7), 778–786.

Šálek, M., Drahníková, L., & Tkadlec, E. (2014). Changes in home range sizes and population

densities of carnivore species along the natural to urban habitat gradient: Carnivores

along the natural-urban habitat gradient. Mammal Review, 45, 1–14.

Trinkel, M., & Kastberger, G. (2005). Competitive interactions between spotted hyenas and lions

in the Etosha National Park, Namibia. African Journal of Ecology, 43(3), 220–224.

Watts, H. E., & Holekamp, K. E. (2009). Ecological determinants of survival and reproduction in

the spotted hyena. Journal of Mammalogy, 90(2), 461–471.

74

Winterbach, H. E. K., Winterbach, C. W., & Somers, M. J. (2014). Landscape suitability in

Botswana for the conservation of its six large African carnivores. PLoS ONE, 9(6), 1–12.