sign in sign up
<<
Home , Mountain nyala

HABITAT QUALITY AND FORAGING ECOLOGY OF MOUNTAIN

( BUXTONI) IN THE MUNESSA FOREST AND THE BALE

MOUNTAINS NATIONAL PARK, SOUTH-EASTERN

Thesis submitted in partial fulfillment of the requirements for the degree of

"DOCTOR OF PHILOSOPHY"

By

Solomon Ayele Tadesse

Submitted to the Senate of Ben-Gurion University of the Negev

May 2012

BEER SHEVA

HABITAT QUALITY AND FORAGING ECOLOGY OF MOUNTAIN NYALA

(TRAGELAPHUS BUXTONI) IN THE MUNESSA FOREST AND THE BALE

MOUNTAINS NATIONAL PARK, SOUTH-EASTERN ETHIOPIA

Thesis submitted in partial fulfillment of the requirements for the degree of

"DOCTOR OF PHILOSOPHY"

By

Solomon Ayele Tadesse

Submitted to the Senate of Ben-Gurion University of the Negev

Signature of Supervisor: Prof. Burt P. Kotler ______

Signature of the Dean of the Kreitman School of Advanced Graduate Studies:

______

May 2012

BEER SHEVA

This work was carried out under the supervision of

Prof. Burt P. Kotler

In the Mitrani Department of Desert Ecology

The Jacob Blaustein Institutes for Desert Research

i

ABSTRACT Understanding how the quality and the characteristics of the habitat influence habitat use and foraging behaviors of an is of paramount importance to ecology and species management. Theories of habitat selection and patch use can be applied in many creative ways to form a set of relatively simple behavioral assays that provide leading indicators of habitat quality and habitat use. The present study utilized the application of different behavioral approaches including habitat suitability models based on activity densities, isodars based on activity densities, behavioral models based on time budgets, and patch use models based on natural giving-up-densities to assess the habitat quality and foraging ecology of the endangered Mountain Nyala (Tragelaphus buxtoni). I worked in the biodiversity rich Munessa Forest and National Park, Ethiopia, on populations of the endangered Mountain Nyala that live with considerable human and livestock pressures. The overall aim of this study was to examine the major environmental and anthropogenic factors affecting the habitat quality, habitat use, and foraging ecology of the charismatic Mountain Nyala. I conducted the fieldwork in the wet and the dry seasons in Munessa. I used different methods to acquire the field data. I conducted regular habitat inventory and population censusing along permanent transects stratified across major habitat types over the landscape. These included population censusing in both daylight and nighttime hours, with nighttime censusing being carried out only during the dry season and with the aid of a spotlight. Censusing yielded estimates of activity densities across habitats and age-sex classes. I measured important microhabitat variables in circular plots laid along each permanent transect and correlated these to local activity densities of Mountain Nyala to yield models of habitat suitability. I also carried out focal animal observations in order to evaluate how the behavioral responses of Mountain Nyala vary with group size, sex-age categories, and habitat types across seasons in Munessa. I assessed and quantified the impacts of human and livestock encroachments on the habitats of Mountain Nyala in Munessa. Accordingly, along permanent transects, I estimated the activity densities of livestock. I also inspected and quantified the extent of human and livestock encroachments on the habitats of Mountain Nyala in circular plots laid along each permanent transect. In addition, I developed pretested, open- and closed-ended interview questionnaires and then

ii

administered them to local people living in the adjacent three peasant associations and one village in Munessa. I also held focal group discussions with key local community members. Finally, I observed free-ranging Mountain Nyala in grasslands versus dense woodlands to quantify their time budgets, bite rates, and bite diameters while feeding on common natural forage plant species in the Bale Mountains National Park (BMNP). These yielded measures of foraging effort and natural giving-up-densities (GUDs: the amount or density of food resources left in a food patch when the most efficient forger leaves the resource patch), with greater bite diameters corresponding with lower costs and greater efficiencies. The browse species cropped by free-ranging Mountain Nyala were identified and the diameters of all browsed twigs were measured. Focal observations were carried out and bite rates were recorded for each sex and age category of target animal. Time budgets were also quantified for focal according to sex-age categories and habitat type. The numbers of individuals in the group were also recorded. Measurements of activity densities and environmental variables allowed me to construct models of habitat suitability and to estimate isodars describing density-dependent habitat selection. The habitat suitability model revealed that Mountain Nyala did not show significant habitat selection behavior during the wet season in Munessa. However, in the dry season, natural forest was the most selected habitat when only crown diameters of trees significantly affected the habitat suitability for Mountain Nyala. The slopes of the isodars also revealed that natural forest habitat is qualitatively, but not quantitatively, better than either the plantation or the clear cut habitat during the dry season. However, the result with spotlight censusing showed that Mountain Nyala selected the clear cut habitat during the night time when people and livestock are absent in Munessa. The behavioral study revealed that Mountain Nyala devoted much of their time to vigilance behavior during the wet season in Munessa; however, habitat type did not significantly affect vigilance. In addition, during the wet season, there was no significant difference in time spent in vigilance among the different sex-age classes. In the dry season, Mountain Nyala were significantly most vigilant in the clear cut habitat. Although more vulnerable animals, especially females with young, are expected to be more vigilant, adult males were more vigilant. This may reflect hunting pressure from humans that exclusively targets adult males.

iii

Livestock and human encroachments on the habitats of Mountain Nyala varied seasonally in Munessa. The activity density of livestock was significantly highest in the natural forest habitat during the wet season. In contrast, during the dry season, livestock did not show significant difference in their relative habitat use. Overall, livestock activity densities in all habitat types were higher in the wet season than in the dry season. In both wet and dry seasons, the extent of stem and crown damage by humans was significantly highest in the plantation habitat. In both seasons, the evidence of wood use and the number of stumps cut by humans were significantly highest in the natural forest. In both seasons, sign of habitat use by livestock did not differ among habitat types; rather it was dispersed throughout all habitats. However, in the wet season, the intensity of grazing / browsing by livestock was significantly heaviest in the natural forest habitat. Generally, the results revealed that the impacts of human and livestock encroachments were high and persistent throughout Munessa. As a result, Mountain Nyala avoided both human and livestock impacts by becoming active during periods (e.g. in the night time) when people and livestock are absent. The social study revealed that attitudes of local people toward Mountain Nyala and its population increase were significantly affected by several socio-economic variables such as livelihood strategy, land ownership, livestock ownership, and knowledge. In addition, through focal group discussions, key community members shared their abundant indigenous knowledge about the different behaviors of Mountain Nyala and their habitats. The results revealed that bite diameters, bite rates, vigilance rates, and proportion of time feeding, all differed between habitats. In particular, Mountain Nyala had greater bite diameters, higher bite rates, and spent a greater proportion of their time feeding, and less in vigilance in the grassland habitat. In addition, adult females had the highest bite rates, and the browse species Solanum marginatum had the greatest bite diameter. The results show that grasslands are a higher quality habitat than woodlands, offering lower foraging costs, greater safety, and more time for foraging. The results further show how behavioral indicators and natural giving-up densities can reveal habitat quality for endangered wildlife through the use of non-invasive techniques. The present studies revealed that Mountain Nyala have faced several human and livestock induced challenges which likely threaten their fitness in Munessa. Mutually

iv

supportive relationships between local people and the Munessa hunting block are crucial to the long-term success of Mountain Nyala population conservation efforts. Thus, the study suggested several management recommendations that should be put in place to conserve and sustainably utilize the endangered Mountain Nyala. Most importantly, introducing and promoting community-based conservation efforts that allow communities to derive economic benefits from ecotourism may promote conservation while at the same time providing a solution to resource use conflicts between the local people and the conservation of the Mountain Nyala population in Munessa. Introducing and advocating an economic benefit-sharing system with the full participation of the local community in the conservation and management processes is also equally important to plan and implement sustainable Mountain Nyala trophy sport hunting in the Munessa hunting block. To conclude, the combined results obtained from the different approaches used in the present studies can help local policy makers and wildlife managers to understand the ecology and the habitats of Mountain Nyala. Behavioral indicators based on foraging theory have many advantages. They are often fast, inexpensive, and simple to implement. More importantly, they provide answers from the forager’s perspective rather than ours, and they have the potential to provide leading indicators of change. Because behavior is adaptive, the resulting measures are leading indicators of habitat change and can form the basis for a more proactive management approach. Changes in the behaviors reveal changes in fitness and population wellbeing. The study thus improves our understanding of the adaptive habitat selection behaviors of Mountain Nyala. This is a basis for developing novel solutions to conserve and manage the endangered Mountain Nyala and its habitats in Ethiopia. The study also motivate local decision makers and wildlife managers to give due emphasis to the needs and the wants of the local people in the management processes. This will ultimately help establish ecologically sustainable, economically feasible, and socially acceptable conservation and management system for the endangered Mountain Nyala in Ethiopia.

Key words: Bale Mountains National Park, behavioral indicators, behavioral models, bite diameter, bite rates, density-dependent habitat selection, habitat quality, habitat suitability models, humans and livestock encroachments, Munessa, Tragelaphus buxtoni.

v

ACKNOWLEDGEMENTS Above all, I would like to thank the Almighty God for all his blessings, guidance, and mercy throughout my life, and for helping me to get to the point of accomplishing this work. I would like to extend my heartfelt gratitude to my supervisor, Prof. Burt P. Kotler, for his extraordinary willingness, unlimited assistance, and guidance in any matters concerning my work whenever and wherever I needed. I have really learnt a lot from the fruitful discussions that we had both in Israel and Ethiopia, and have benefited a lot from his experiences. In short, he is a wonderful supervisor and I need him to keep it up! I am also grateful to forward my sincere thanks to Prof. Arnon Lotem and Dr. Yaron Ziv for their time and patience in commenting on the proposal of my PhD studies and sharing with me their fruitful experiences. I would like to thank the three anonymous reviewers for their time and patience in commenting on the whole PhD dissertation and sharing with me their fruitful experiences. Dr. Paul Evangelista deserves to be acknowledged for sharing with me his critical ideas and practical knowledge on Mountain Nyala and their habitats in Ethiopia. And also his encouragement to carry out the field research in the Munessa Forest is much appreciable and motivating to me. Nigussie Chala is substantially thanked for providing accommodation and friendship on my trips to the Munessa Forest during the whole period of my fieldworks. Ekemo Ersedo and Tsegaye Ayalew deserve invaluable thanks for the many hours of their assistance with fieldwork, guidance, and companionship in the Munessa Forest. Anteneh Girma also needs to be much acknowledged for his many hours of field assistance in the Bale Mountains National Park. Alemayehu Abiso is very appreciated for his continuous driving service both in the day light hours and during the night time in the Munessa Forest and the Bale Mountains National Park. I am grateful to the Wondo Genet College of Forestry and Natural Resources (WGCF-NR) for the vehicle support to ease the fieldwork both in the Munessa Forest and the Bale Mountains National Park. It is a pleasure to thank the Albert Katz International School (AKIS) for providing me the financial support for my PhD studies in Israel. I am also indebted to thank the Rufford Small Grant Foundation (RSGF) and The Murelle Foundation (TMF) which helped covered the financial expenses for the fieldwork in Ethiopia.

vi

My heartfelt thanks go to Aramde Fetene for his help in transforming the GPS locations of my transect walks into the Ethiopian coordinate system and continuously guiding me with the Arc GIS 9.3 program to produce the maps of the study areas. Of course, many people have contributed much either directly or indirectly for the success of this research project. I would like to thank the following people for their unreserved assistance in sharing with me their critical ideas which added a lot of inputs to my research work. These are: Tadele Zewdie, Dr. Tsegaye Bekele, Dr. Demel Teketay, Dr. Solomon Zewdie, Dr. Melaku Bekele, Dr. Shomen Mukereji, Dr. Ruti Berger, Oded Berger, Abduraman Wario, Mohammednur Jemal, Sonny Solomon Bleicher, and Addisu Assefa. If I forget to mention the names of other people, I ask their forgiveness for the oversight, but I am very grateful for their invaluable ideas and assistance. I acknowledge my honest appreciation to the following institutions in Ethiopia: Ethiopian Wildlife Conservation Authority (EWCA), Oromiya Forest and Wildlife Enterprise (OFWE), Arsi Forest Branch, Munessa Forest District, and Bale Mountains National Park (BMNP) for their unreserved co-operation in giving me the permit to work on the endemic Mountain Nyala and its habitats in Ethiopia. I would like to thank the local people residing in different peasant associations around the Munessa Forest for sharing with me their fruitful ideas and abundant experiences about Mountain Nyala and its habitats during the focal group discussions and the interview questionnaires. I am also thankful to the leaders of the different peasant associations who coordinated and encouraged the local people to willingly involve themselves and participate in the valuable focal group discussions and interview questionnaires held around the Munessa Forest. Even though I don’t have a list of their names at hand, I would like to extend my appreciation to the eight newly employed forest professionals working at the Munessa Forest Distinct who helped me much to handle the interview questionnaires in the three peasant associations and one village around Munessa. I am indebted to forward my warm thanks to Keren Embar for her time and patience in facilitating the transfer of my monthly scholarship via Western Union every month while I was in Ethiopia for the field research. I also thank Keren Embar for all her moral supports during my hard times in the fieldwork in Munessa, Ethiopia. I am further indebted to thank Keren Embar for her help in Hebrew translations of many things in

vii

connection with my PhD studies including the title and the abstract of this dissertation. I would like to express my sincere appreciation to Ms Dorit Levin for her extraordinary assistance. I am really very grateful for all her understanding, friendly approach, and her usual readiness to help. My parents, my two sisters, my brother, and all my good friends in Ethiopia deserve special thanks for their unreserved moral supports during the whole periods of my study. My mother, Amelework Nigussie Mullatu, needs special appreciation for all her kind hearted maternity and supports in any matters since my childhood. I have really learnt strength and self confidence from my mother. My sister, Meseret Ayele Tadesse, needs special acknowledgements for all her moral support and kind hearted treatments during my hardship times in the fieldworks in the Munessa Forest and the Bale Mountains National Park. Most importantly, I would like to thank my wife, Ketsela Wuletaw Wubetu, for all her love, friendship, support, and the good times that I have shared with her during the past number of years both in Ethiopia and Israel. It is truly been awesome being with her and having her support during this research project. I couldn’t imagine getting to this point without her patience, continuous support, and encouragement. Our newly born daughter, Lamrot Solomon Ayele, is a gift for both of us and let God always bless her. Last but not least, great thanks to the populations of Mountain Nyala and their habitats in the Munessa Forest and the Bale Mountains National Park from which I have learnt much during the whole processes of this research work. The study was conducted in the Marco and Louise Mitriani Department of Desert Research, the Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev. This work was supported by the scholarships from the Albert Katz International School for Desert Studies, the Kreitman School of Advanced Graduate Studies, and the United States-Israel Bi-national Science Foundation. Solomon Ayele Tadesse

viii

LIST OF CONTENTS Page Abstract i Acknowledgements v List of Contents viii List of Tables x List of Figures xi List of Plates xiv Chapter 1: General Introduction 1 Chapter 2: Study Species and Study Areas 10 Study Species 10 Study Areas 19 Chapter 3: Habitat selection by Mountain Nyala assessed with habitat suitability model and isodar analysis in Munessa 28 Abstract 28 Introduction 29 Methods 32 Results 39 Discussion 48 Chapter 4: Habitat use by Mountain Nyala evaluated with behavioral indicators in Munessa 54 Abstract 54 Introduction 55 Methods 58 Results 63 Discussion 79 Chapter 5: Habitat use by Mountain Nyala determined using bite diameters, bite rates, and time budgets in the Bale Mountains National Park 86 Abstract 86 Introduction 87 Methods 89

ix

Results 95 Discussion 105 Chapter 6: Impacts of humans and livestock encroachments on the habitats of Mountain Nyala in Munessa 111 Abstract 111 Introduction 112 Methods 115 Results 117 Discussion 123 Chapter 7: Attitudes of local people toward Mountain Nyala and their habitats in Munessa 128 Abstract 128 Introduction 129 Methods 132 Results 139 Discussion 148 Chapter 8: Concluding Discussion and Recommendations 153 Concluding Discussion 153 Recommendations 163 References 169 Appendices 190 Appendix I 190 Appendix II 195 Abstract in Hebrew 197

x

LIST OF TABLES Page Table 3.1. A summary of the twelve transects used in the analysis. 36 Table 3.2. A summary of the activity densities of Mountain Nyala (MN) in different habitat types during the wet and the dry seasons field survey. 40 Table 3.3. A summary of night time activity densities of Mountain Nyala (MN) in different habitat types in the dry season with a spotlight censusing. 41 Table 3.4. A summary of multiple linear regression statistics and constituent variables of the habitat suitability index model for Mountain Nyala in Munessa during the wet season. 43 Table 3.5. A summary of multiple linear regression statistics and constituent variables of the habitat suitability index model for Mountain Nyala in Munessa during the dry season. 44 Table 3.6. The results of linear regression for isodars comparing the activity densities of Mountain Nyala in adjacent habitat types during the dry season 46 Table 4.1. The seasonal effects of habitat variables, group size, habitat type, sex-age categories, and the interaction of habitat type and sex-age categories on the activity time budgets of Mountain Nyala in Munessa. 65 Table 4.2. A summary of the behavioral models for Mountain Nyala in the wet and the dry seasons in Munessa. 69 Table 5.1. Habitat type-based mean bite diameters and standard deviations of twigs cropped and consumed by free-ranging Mountain Nyala in the Bale Mountains National Park. 96 Table 6.1. A summary of the activity density of livestock in different habitat types during the wet and the dry season field surveys. 117 Table 7.1. Sample characteristics and descriptive results for independent variables 141 Table 7.2. Descriptive results (percent of responses) of multiple items combined to measure “attitudes of local people toward Mountain Nyala, and their conservation and a management in Munessa” 144 Table 7.3. Descriptive results (percent of responses) of statements combined to measure “attitudes of respondents toward Mountain Nyala population increase in Munessa” a 145 Table 7.4. Multiple linear regression modela for attitudes toward “Mountain Nyala and c their habitats”b, and “Mountain Nyala population increase” 147

xi

LIST OF FIGURES Page Figure 2.1. Distribution map of Mountain Nyala throughout the highlands of Ethiopia. 13 Figure 2.2. Map of Munessa where most of the studies were conducted. 23 Figure 2.3. Map of Bale Mountains National Park as one of the study areas. 27 Figure 3.1. (a) A simple representation of the ideal free model of density-dependent habitat selection, (b) an isodar model generated from fitness-density curves depicted in (a). 38 Figure 3.2. Seasonal habitat use by Mountain Nyala in Munessa. 40 Figure 3.3. Dry season habitat use by Mountain Nyala in Munessa assessed with a spotlight censusing technique during the night time. 42 Figure 3.4. The environmental variables and their response curves included in the Mountain Nyala habitat suitability model for Munessa during the wet season. 43 Figure 3.5. The environmental variables and their response curves included in the Mountain Nyala habitat suitability model for Munessa during the dry season. 45 Figure 3.6. Isodars for the dry season between activity densities of Mountain Nyala in (a) natural forest and plantation habitats, (b) natural forest and clear cut habitats. 47 Figure 4.1. General activity time budget patterns of Mountain Nyala during the wet and the dry seasons in Munessa pooled over four categories, adult males, adult females, sub-adults, and juveniles. 64 Figure 4.2. The effect of habitat type on the vigilance level of Mountain Nyala during the wet and dry seasons in Munessa pooled over the four categories, adult males, adult females, sub-adults, and juveniles. 66 Figure 4.3. The effect of sex and age categories on the proportion of time vigilant during the wet and the dry seasons. 67 Figure 4.4. Seasonal variation of group sizes in Mountain Nyala across habitat type in Munessa. 68 Figure 4.5. A model for the vigilance behavior of Mountain Nyala in Munessa during the wet season. 70 Figure 4.6. A model for the feeding behavior of Mountain Nyala in Munessa during the wet season. 71 Figure 4.7: A model for the moving behavior of Mountain Nyala in Munessa during the

xii

wet season. 72 Figure 4.8. A model for the resting behavior of Mountain Nyala in Munessa during the wet season. 73 Figure 4.9. A model for the vigilance behavior of Mountain Nyala in Munessa during the dry season. 75 Figure 4.10. A model for the feeding behavior of Mountain Nyala in Munessa during the dry season. 76 Figure 4.11. A model for the moving behavior of Mountain Nyala in Munessa during the dry season. 77 Figure 4.12. A model for the resting behavior of Mountain Nyala in Munessa during the dry season. 78 Figure 5.1. The effect of habitat type on the diameters of twigs browsed by free-ranging Mountain Nyala in the Bale Mountains National Park (BMNP). 97 Figure 5. 2. The effect of natural forage plant species on the diameters of twigs browsed by free-ranging Mountain Nyala in the BMNP. 98 Figure 5.3. The effect of habitat type and natural browse plant species on the diameters of twigs browsed by free-ranging Mountain Nyala in BMNP. 99 Figure 5.4. The effect of habitat type on the bite rates of free-ranging Mountain Nyala in BMNP pooled over bite rates by four categories, adult males, adult females, sub-adults, and juveniles. 100 Figure 5.5. The effect of habitat type and sex-age categories on the bite rates of free- ranging Mountain Nyala in BMNP. 101 Figure 5.6. General activity time budget pattern of Mountain Nyala in BMNP pooled over four categories, adult males, adult females, sub-adults, and juveniles. 102 Figure 5.7. The effect of habitat type on the vigilance level of Mountain Nyala in the Bale Mountains National Park pooled over four categories, adult males, adult females, sub- adults, and juveniles. 103 Figure 5.8. The effect of habitat type and sex-age categories on the proportion of time vigilant by free-ranging Mountain Nyala in BMNP. 104 Figure 5.9. The effect of group size on the vigilance level of Mountain Nyala in the Bale Mountains National Park. 104

xiii

Figure 6.1. Seasonal habitat use by livestock in Munessa. 118 Figure 6.2. Seasonal stem damage on trees across habitat types in Munessa. 118 Figure 6.3. Seasonal crown damage on trees across habitat types in Munessa. 119 Figure 6.4. Seasonal wood use across habitat types in Munessa. 120 Figure 6.5. Cut stumps in different seasons and habitat types in Munessa. 121 Figure 6.6. Seasonal habitat use by livestock in Munessa. 121 Figure 6.7. Seasonal level of grazing / browsing by livestock in Munessa. 122

xiv

LIST OF PLATES Page Plate 2.1. A picture of Mountain Nyala taken at Dinsho Sanctuary, Bale Mountains National Park, Ethiopia. 11 Plate 2.2. An open area in the Munessa Forest which serves as a typical foraging ground for Mountain Nyala especially in the rainy season. 14 Plate 5.1. A picture showing how Verner caliper is used to measure the bite diameters of Carduus nyassanus plant species browsed by Mountain Nyala in the woodland habitat, Bale Mountains National Park, Ethiopia. 92

1

CHAPTER 1 GENERAL INTRODUCTION Species are more abundant in some habitats than others (Morris, 1987), with various environmental (i.e., biotic and abiotic) and anthropogenic factors determining the presence or absence of a species in a specific habitat. This suggests that there are intimate and complex relationships between a species and its habitats. The ability to effectively conserve and manage wildlife populations and their habitats depends largely on our capacity to understand and predict species-habitat relationships. Conservation activities are often prioritized to solve problems of species already at risk of extinction, or for which declining populations signal impending concern (Morris, 2003). In order to effectively manage conservation areas, conservation authorities need to develop optimal management strategies. Wildlife management often demands integrated knowledge on the biology of the species, its habitat, and the needs and wants of the local people living in the surrounding landscape (Anderson et al., 2002). So, wildlife managers should incorporate both ecological and societal issues into their research designs and management plans. Through their different mechanisms of adaptation, organisms may value habitats differently (Melton, 1987). For example, the habitat selection by individuals of a species is governed by the presence of quality habitat essentials (Stephens and Krebs, 1986), freedom of extreme competition from associated species (Tilman, 1982), and suitability of escape terrains for avoidance of predation risk (Stephens and Krebs, 1986; Kotler et al., 1994), human nuisance (Gill et al., 2001; Manor and Saltz, 2003), and livestock disturbances (Evangelista et al., 2007; Mamo, 2007). Previous studies demonstrated that habitat quality and characteristics influence temporal activity patterns, foraging behavior, anti-predator behavior, and other social organization of a species (e.g., Brown and Alkon, 1990; Rosenzweig, 1981; Cresswell, 1994; Kotler et al., 1994; Arenz and Leger, 1997; Druce, 2005). Understanding how wild animals use their habitats is of paramount importance to ecology and species management. Habitat quality is a measure of the contribution of an area to individual fitness and population persistence (Van Horne, 1983; Block et al., 1998; Morris, 1998). For example, decline in habitat quality may often represent transformation from optimal toward unsuitable environmental conditions for survival and reproduction

2

(Evangelista et al., 2007; Morris et al., 2009). Habitat quality is affected by costs such as risk of predation (e.g., Brown, 1988; Kotler et al., 1994) and benefits such as availability of the right kinds of food (e.g. Gross et al., 1995), and cover associated with residing in that habitat (e.g., Brown and Alkon, 1990; Arenz and Leger, 1997). Different approaches have been used to assess the habitat quality of a species. For example, the traditional approaches follow the population sizes of herbivores and associated vegetation structure to assess habitat quality (Boshoff and Kerley, 2001; Wilson and Kerley, 2003). However, such approaches can only note trouble after problems have become manifested in large declines of population size and/or vegetation. In contrast, behavioral indicators based on foraging theory provide leading indicators of habitat quality. The reason is that patch use and habitat selection behaviors are adaptive and reflect fitness consequences (Morris et al., 2009). Foraging theory predicts the optimal behavior of animals whilst foraging (Smith, 1993), and assumes that this behavior has been molded by natural selection (Sih, 1980). Organisms responding to environmental challenges should possess adaptive behaviors. Changes in these behaviors reveal changes in fitness and population wellbeing. Therefore, animal behaviors reveal much about habitat quality. However, the assumption that animals adaptively respond to their environment may not always be correct and therefore it should be carefully assessed and addressed. For example, animals may fail to respond adaptively to novel environmental factors to which they were not exposed during their evolutionary history, such as invasive species or technological innovations (e.g., artificial light, novel chemicals or food types). With the help of accurate assessment of habitat quality, wildlife managers can maintain the correct number of individuals within a natural system (Druce, 2005). At larger scales, accurate assessment of habitat quality allows conservation planning (Fish and Wildlife Services, 1980; Boshoff et al., 2002; Druce, 2005). Conservation managers can have access to a set of relatively simple behavioral assays based on foraging theory. These indices infer habitat quality and habitat use (Siegfried, 1980; Druce et al., 2006, 2009; Shrader et al., 2008a, 2008b). This is particularly relevant to conservation biology because behavioral indices based on adaptive behavior may provide leading indicators of habitat change (Morris et al., 2009). Unlike population density dynamics, behaviors can respond instantaneously to altered environmental conditions (Morris, 2003; Morris et al., 2009;

3

Berger-Tal et al., 2011). Such indices include habitat suitability models based on activity densities, isodars based on activity densities, behavioral models based on time budgets, and patch use models based on natural giving-up-densities. I discuss each in the respective Chapters (see Chapters 3, 4, and 5). A widely studied issue in conservation biology is the potential impact of human disturbances on wild animal populations and their habitats (Klein et al., 1995; Kellert et al., 1996; Primack, 2002). Wildlife populations often perceive humans as potential predators, and their behavioral responses to human disturbances can be assessed and quantified in a similar matter to risk of predation (e.g., Whittaker and Knight, 1998; Fernandez-Juricic and Telleria, 2000; Frid and Dill, 2002; Manor and Saltz, 2003; Gilroy and Sutherland, 2007). According to the risk-disturbance hypothesis, human disturbance is analogous to predation risk in that it can cause habitat shifts and other behavioral responses that later influence the fitness of individuals and the persistence of wildlife populations (Frid and Dill, 2002; Gilroy and Sutherland, 2007). However, it is important for conservationists to know whether avoidance of disturbance results in population change (Sutherland, 1996; Gill and Sutherland, 2000). Wild animals disturbed by humans are expected to increase their anti-predator and vigilance behavior at the expense of other activities. For example, Manor and Saltz, (2003) found that human disturbance is the overwhelming factor affecting vigilance. Wild animal species must then have to make trade-offs between selecting habitats that offer forage resources (Stephens and Krebs, 1986) or avoiding human activity (Gill et al., 1996; Gill and Sutherland, 2000; Mamo, 2007) especially where human activity is associated with increased mortality (Frid and Dill, 2002). A species with suitable habitat nearby may avoid disturbance simply because it has alternative sites to go to. In contrast, wild animals with no suitable habitat nearby will be forced to remain despite the disturbance, regardless of whether or not this will affect survival or reproductive fitness (Gill et al., 2001). So, from a conservation perspective, human disturbance on wildlife population is important if it affects survival or reproductive fitness and hence causes a population to decline (Gill et al., 2001). Several studies demonstrated that wild animals avoid areas where humans are present (e.g., Burger, 1981; Belanger and Bedard, 1990; Sutherland and Crock-ford, 1993;

4

Klein et al., 1995; Gill, 1996; Frid and Dill, 2002; Manor and Saltz, 2003; Mamo, 2007; Atickem et al., 2011). However, avoidance of areas because of human activity can often have demographically costly consequences (Johnson et al., 2004; McLoughlin et al., 2005). For example, human disturbances can affect the reproductive success by influencing pairing display or increasing the amount of parental care given to the young in wild animal species (e.g., Burger, 1981; Verhulst et al., 2001). Human disturbance reduces the quality and availabilities of habitats (e.g., Evangelista et al., 2007; Mamo, 2007; Atickem et al., 2011) and can also force wild animals to limit their access to foraging, nesting, and shelter areas (Gill et al., 1996; Manor and Saltz, 2003; Refera and Bekele, 2004). Even within protected areas, human disturbance has the potential to influence the distribution and abundance of wild animals (e.g., Stephen et al., 2001; Refera and Bekele, 2004; Mamo, 2007; Tadesse and Kotler, 2010; Atickem et al., 2011) including their spatial and temporal patterns of behaviors (e.g., Refera and Bekele, 2004; Fernandez-Juricic and Telleria, 2000; Mamo, 2007).

Conceptual Research Background The highlands of Ethiopia are rich in endemic fauna and flora species (Yalden and Largen, 1992; Hillman, 1993). Mountain Nyala (Tragelaphus buxtoni) is one of the endemic flagship species in the southeastern highlands of Ethiopia (Hillman and Hillman, 1987). However, compared with its closest relatives such as the Greater (Tragelaphus strepsiceros) and the Nyala of southeastern Africa (Tragelaphus angasii), the Mountain Nyala is not well studied (Hillman, 1985; Shuker, 1993; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007; Evangelista et al., 2008). For example, for many decades, field information on the habitat, population distribution, abundance, and dynamics of the elusive Mountain Nyala in Ethiopia was largely gathered and reported by trophy hunters (Evangelista et al., 2007). After its discovery in 1908, population estimates of Mountain Nyala remain unclear (Evangelista et al., 2007; Atickem et al., 2011). For example, the only reliable population estimates of Mountain Nyala were carried out in the Gaysay and the Dinsho areas in the northern end of the Bale Mountains National Park (Hillman, 1985; Hillman and Hillman, 1987; Woldegebriel, 1996; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo,

5

2007; Atickem et al., 2011). So, the Mountain Nyala in the Gaysay and Dinsho areas are reasonably well known by the general public, international community, conservation organizations, scientists, and government institutions (Malcolm and Evangelista, 2005; Evangelista et al., 2007). The largest numbers of Mountain Nyala are aggregated in the Dinsho Sanctuary, the Gaysay plain, and the fragmented forest patches adjacent to the Gaysay Valley in the Bale Mountains National Park, Ethiopia (Evangelista et al., 2007; Mamo, 2007; Atickem et al., 2011). Because of their accessibility, more scientific studies have been conducted on these Mountain Nyala than any others (e.g., Brown, 1969a, 1969b; Hillman, 1985; Hillman and Hillman, 1987; Woldegebriel, 1996; Stephens et al., 2001; Refera and Bekele, 2004; Mamo, 2007; Atickem et al., 2011). Even though much scientific knowledge regarding the Mountain Nyala has been documented by studying this population, these studies may not accurately represent the natural wild behaviors and other social organizations of the species (Evangelista et al., 2007). This is because the Dinsho Sanctuary and the Gaysay Valley have been enormously exposed to continuous impacts of humans and livestock (Stephens et al., 2001; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007; Atickem et al., 2011). As a result, those adverse impacts are likely to modify the natural behaviors as well as other social organizations of Mountain Nyala in the wild. In addition, those studies have not even addressed the full range of the species within the park’s boundaries (Evangelista et al., 2007; Atickem et al., 2011). Much is unknown about how Mountain Nyala are related to their habitats (Brown, 1969b; Evangelista et al., 2007; Mamo, 2007; Atickem et al., 2011) and their nutritional requirements (Malcolm and Evangelista, 2005; Evangelista et al., 2007), or how habitat characteristics and anthropogenic factors influence their foraging and patch use behaviors. Evangelista et al. (2007) also note that scientific studies overlooked several significant populations of Mountain Nyala in Ethiopia and simply rely on speculation rather than in- depth studies. However, since 2000, there have been several sizable populations identified that have yet to be acknowledged by the scientific or wildlife conservation communities in Ethiopia (Evangelista et al., 2007). This gap has resulted in a high degree of uncertainty in

6

regard to the status of the Mountain Nyala, and inhibits the effectiveness of management plans and conservation strategies (Evangelista et al., 2007). So, scientific research in regard to the ecology, behaviors, habitat availability, habitat selection and patch use of the different localized populations of Mountain Nyala in Ethiopia are crucial for the effectiveness of management plans and conservation strategies. I planned to work in the biodiversity rich Munessa Forest and Bale Mountains National Park, Ethiopia where populations of the critically endangered Mountain Nyala can be found. Mountain Nyala require a relatively large home range and diverse habitat resources (Brown, 1969b; Hillman, 1985; Evangelista et al., 2007; Atickem et al., 2011); therefore, they need management of landscapes rather than isolated patches of habitat. However, due to the ever-increasing humans and livestock pressures, there is growing interest in using the Munessa Forest and Bale Mountains National Park for livestock grazing and agriculture. Ensuring the long-term persistence of the populations of Mountain Nyala in the study areas may require accounting for the various ecological and anthropogenic factors. Stepping up efforts to protect and conserve Mountain Nyala and their habitats is, therefore, of greatest urgency. Mountain Nyala are important to the general public (Malcolm and Evangelista, 2005), whether as a game species (Lydekker, 1912; Sanford and Legendre, 1930; Mellon, 1975; Atickem et al., 2011), as the name of many prominent businesses (e.g., Nyala Insurance and Nyala Motors) (Malcolm and Evangelista, 2005), or as an aesthetic symbol of nature (Shuker, 1993). By improving management decisions, public support at the local level is crucial for the success of management plans. Assessing and quantifying the impacts of humans and livestock on the population of Mountain Nyala and its habitats is, therefore, another step. This is important to promote the formulation of policies which initiate community-based management plans and conservation strategies toward Mountain Nyala and their habitats.

Aims of the Study The overall aim of this study was to examine the major environmental and anthropogenic factors affecting the habitat quality, habitat use, and foraging ecology of

7

Mountain Nyala in the Munessa Forest and the Bale Mountains National Park, Ethiopia. The specific objectives of the study were to: (1) identify the preferred habitat types of Mountain Nyala by carrying out habitat suitability modeling and isodar analyses, (2) explore the proximate environmental variables to which Mountain Nyala are behaviorally responsive by developing behavioral indicators based on time budgets of focal individuals as functions of habitat types, environmental variables, group size, sex-age classes, and seasons, (3) evaluate the effect of group size and sex-age categories on the vigilance behavior of free-ranging Mountain Nyala across habitat types and seasons, (4) quantify the variation in group size as functions of habitat type and seasons, (5) quantify the bite diameters and bite rates of free-ranging Mountain Nyala foraging on common natural browse species across habitat types, (6) assess and quantify humans and livestock encroachments on the habitats used by Mountain Nyala across seasons, and (7) assess and quantify the knowledge and attitudes of local people toward Mountain Nyala and their habitats and thereby to look for community-based conservation strategies.

Relevance of the Study Behavioral indicators based on foraging theory have many advantages. They are often fast, inexpensive, and simple to implement. More importantly, they provide answers from the forager’s perspective rather than ours, and they have the potential to provide leading indicators of change (Kotler et al., 2001, Morris et al., 2009). The present study provides measures of habitat suitability and habitat selection that are based on behavioral variables. Because behavior is adaptive, the resulting measures are leading indicators of habitat change and can form the basis for a more proactive management approach. Habitat Suitability Index (HSI) models are crucial tools for helping managers and ecologists identify and evaluate habitat variables (Druce, 2005; Reid, 2005; Evangelista et al., 2008; Tadesse and Kotler, 2010), and thereby predict future conditions of the habitats for the survival and reproductive success of Mountain Nyala. The field information obtained through habitat sampling may help in designing future monitoring programs to track population size and status of Mountain Nyala. Isodars can be applied to identify the number of habitats and to measure scales of habitat selection (Morris, 2003) that the Mountain Nyala recognize across a heterogeneous

8

landscape. Isodar analyses reveal basic information on the underlying mechanisms of habitat selection by Mountain Nyala. Isodar analyses also complement the habitat suitability models through increasing our understanding on qualitative and quantitative differences (Morris, 2003) in habitat selection and use by Mountain Nyala. The study allowed the development of behavioral tools to quantify how Mountain Nyala respond to variation in habitat type, predation risk, and human nuisance and livestock disturbances. Changes in the behavior of individuals can indicate a change in environmental conditions. The study thus improves our understanding of the adaptive habitat selection behaviors of Mountain Nyala. This is a basis for developing novel solutions to conserve and manage the endangered Mountain Nyala and its habitats in Ethiopia. The intensity of resource patch use provides a measure of habitat quality (Olsson and Molokwu, 2007). However, it is often not feasible to train individuals of target species to use artificial food patches. In such cases, using natural measures of patch use would be preferable. In the current study, I developed techniques for measuring natural giving-up- densities (GUDs) using measurement of bite size based on twig diameter of branches browsed by free-ranging Mountain Nyala on natural forage species. The diameters of browsed twigs are in effect natural giving-up-densities, and can be used just like GUDs measured from artificial food patches to compare food preferences, compare habitat quality, look for impact of risk of predation and human disturbance, and track changes in range quality. Such measures are especially important in that they provide a measure of range quality perceived by the foraging Mountain Nyala. The field information may also direct habitat conservation priorities to the habitats that require greater levels of protection from human and livestock use (Evangelista et al., 2007; Mamo, 2007). Understanding how wild animals respond to anthropogenic activities is fundamental to resolving potential conflicts between humans and wild animals (Stephen et al., 2001; Beale, 2007; Mamo, 2007). The local communities residing around the study area were considered in this study. The motivation for such consideration is that a better understanding of the impacts of humans and livestock on the habitats of Mountain Nyala could help decision makers to formulate habitat management plans and conservation strategies. The study also assessed and quantified the knowledge and attitudes of local

9

people toward Mountain Nyala and their habitats. The outcomes of such study may help identify and promote community-based participatory wildlife conservation strategies.

Structure of the Thesis The dissertation is organized in eight Chapters. Chapter 1 presents general introduction, conceptual research background, aims, and relevance of the study. Chapter 2 provides an overview of the study species (i.e., Mountain Nyala) and the study areas (i.e., the Munessa Forest and the Bale Mountains National Park). Chapter 3 deals with habitat selection by Mountain Nyala evaluated by habitat suitability model and isodar analyses in Munessa, while Chapter 4 focuses on the seasonal habitat use of Mountain Nyala determined by behavioral indicators in Munessa. Chapter 5 describes how patch use of Mountain Nyala is determined by quantifying bite rates and measuring the diameters of twigs removed by foraging Mountain Nyala through quantifying their natural giving-up densities (GUDs) in Bale Mountain National Park. Chapter 6 is devoted to assess and quantify the impacts of humans and livestock encroachments on the habitats used by Mountain Nyala in Munessa. Chapter 7 deals with measuring the knowledge and attitudes of local people toward Mountain Nyala and their habitats in Munessa. Chapter 8 gives a synthesis of this work. It summarizes the research by synthesizing the outcomes from the different approaches used in the study, recommends possible conservation and management activities to be carried out, and suggests ideas for how research in this field should continue in the future.

10

CHAPTER 2 THE STUDY SPECIES AND THE STUDY AREAS

The Study Species Description The charismatic Mountain Nyala (Tragelaphus buxtoni) is an endemic flagship species which belongs to the family, the sub-family Tragelaphinae (Kingdon, 1997; Stuart, and Stuart, 2000). Mountain Nyala was discovered in 1908 as a spiral-horned (Kingdon, 1997). Both stripes and spots are regularly present on most individuals of Mountain Nyala (Kingdon, 1997; Evangelista et al., 2007). Mountain Nyala may have up to nine lateral white stripes on each side (Kingdon, 1997; Evangelista et al., 2007). Spots are often present on the flanks or in a linear pattern running horizontally across the side or back, with one to three spots commonly present on the face or cheek of an animal (Kingdon, 1997; Evangelista et al., 2007). However, markings are generally unique for individual animals with the spot and stripe patterns differing on each flank of an individual (Brown, 1969a; Hillman, 1985; Kingdon, 1997; Evangelista et al., 2007). Mountain Nyala is a sexually dimorphic animal. Adult males are easily distinguished by their large twisted horns and dark brown color. Adult females do not carry horns and have a reddish brown color. Horns of fully grown adult males usually grow up to 1 meter in length and may have three distinct ridges (Kingdon, 1997; Evangelista et al., 2007). Adult males are much larger than adult females. Adult male Mountain Nyala may weigh between 260 to 340 kg (varying among populations and habitat types) and stand approximately 135 cm at the shoulder (Kingdon, 1997; Stuart, and Stuart, 2000; Evangelista et al., 2007). However, adult females may weigh approximately 150 to 200 kg (Kingdon, 1997).

11

Plate 2.1: A picture of Mountain Nyala taken at Dinsho Sanctuary, Bale Mountains National Park, Ethiopia. The front individual is adult female whereas the rear one is adult male. (Photo by the author, 2011).

Breeding habits Little is known about the reproductive biology of Mountain Nyala (Evangelista et al., 2007). For example, Brown (1969b) suggests that Mountain Nyala give birth to calves at different times of the year with peaks during the months of April, May, June, October, November, and December. However, Hillman (1985) notes that Mountain Nyala are perennial breeders exhibiting a defined calving peak during the months of August and September. Mountain Nyala exhibit a polygynous reproductive strategy, which is common among non-territorial where potential mates may be widely distributed (Evangelista et al., 2007). Courtship begins when a female approaches a state of estrus and ends when copulation is permitted (Evangelista et al., 2007). Brown (1969b) suggests that

12

under optimal environmental conditions, reproductive success of Mountain Nyala may increase populations as much as 25% per year. However, Evangelista et al. (2007) argue that their reproductive potential is extremely sensitive and birth rates are quite liable to factors such as habitat quality, population densities, stress, climate, and negative human and livestock impacts.

Distribution and abundance The Mountain Nyala are confined in their distribution to the southeastern highlands of Ethiopia, bounded by the Chercher Mountains in the north and the Bale Mountains in the south (Refera and Bekele, 2004; Evangelista et al., 2007). Currently, the species is known to occur in six different locations in Ethiopia. From north to south, these sites are: Kuni Muktar, Dindin and Arba Gugu Forests, Munessa-Shashemene Forests, Mount Kaka, Galama-Chilalo Forest Priority Area (FPA), and the Bale Mountains (Malcolm and Evangelista, 2005; Evangelista et al., 2007) (see Fig. 2.1). The most prominent population of Mountain Nyala is found in the Bale Mountains (Brown, 1969a, 1969b; Evangelista et al., 2007; Mamo, 2007; Atickem et al., 2011). Population estimates for Mountain Nyala are difficult to assess due to the animal’s elusive nature (Evangelista et al., 2007; Atickem et al., 2011), preference for dense forest habitats (Atickem et al., 2011), and the uncertainty of their range (Evangelista et al., 2007). The first attempt to determine the total population estimates of Mountain Nyala was conducted by Leslie Brown. Brown (1969b) estimated that Mountain Nyala numbers ranged between 7,000 and 8,000 individuals in Ethiopia. In recent estimates, the total population size of Mountain Nyala throughout Ethiopia may not exceed 1,000 individuals (Refera and Bekele, 2004). However, having investigated several new ranges of Mountain Nyala in Ethiopia, Evangelista et al. (2007) argue that the population size of Mountain Nyala in Ethiopia should be greater than 4,000 individuals. On the other hand, using fecal pellet group counting technique, Atickem et al. (2011) recently estimated that the population size of Mountain Nyala in the Bale Mountains alone could be about 3,800 individuals. The differences in population estimates of Mountain Nyala are most likely due to low sampling efforts in the forested habitats by previous field studies where Atickem et al. (2011) reported the highest numbers.

13

Munessa Forest

Figure 2.1. Distribution map of Mountain Nyala throughout the highlands of Ethiopia. The Munessa Forest, the Bale Mountains National Park (BMNP), and the Controlled Hunting Areas (CHA) - where legal trophy sport hunting of Mountain Nyala have been practiced, are also shown. (Source: Evangelista et al., 2007).

Habitats The habitats of Mountain Nyala are heavily affected by precipitation and temperature, which are known to have direct influences on vegetation structure and diversity (Brown, 1969b; Evangelista et al., 2008). Brown (1969b) suggested that the Mountain Nyala is a habitat specialist occurring mainly above 3400 m and being particularly common in the heath zone. However, Mountain Nyala are capable of persisting in a wide range of habitats from 2500 to 4300 m (Hillman, 1985; Malcolm and

14

Evangelista, 2005). They are commonly reported to be found in a mosaic of high-altitude woodland, bush land, heathland, moorland, and valley-bottom grassland, ranging from 2700 to 4300 m above sea level, but mostly prefer mesic heathland and alpine habitats (Yalden and Largen, 1992; Kingdon, 1997). However, their range is much broader than previously reported, and they are distributed across a variety of montane forest habitats. For example, Evangelista et al. (2007) noted that the team observed Mountain Nyala at elevations as low as 1,600 m occupying dense forests. Mountain Nyala require two kinds of habitat, i.e. foraging grounds and cover in which to hide themselves from predators and shelter from extreme weather (Hillman and Hillman, 1987; Malcolm and Evangelista, 2005). For example, the small montane grassland around the Dinsho area in the Bale Mountains National Park provides an important feeding habitat for Mountain Nyala (Brown, 1969b; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). Mountain Nyala need dense cover for camouflage (Hillman and Hillman, 1987; Evangelista et al., 2007), and higher grounds and open sightlines to detect risk of predation and quickly escape (Evangelista et al., 2007). Generally, the habitat where the species is found ranges from closed canopy woodlands to grasslands and open alpine moorlands (Brown, 1969b; Hillman, 1985; Hillman and Hillman, 1987; Malcolm and Evangelista, 2005; Evangelista et al., 2007).

Plate 2.2. An open area in the Munessa Forest which serves as a typical foraging ground for Mountain Nyala especially in the rainy season. (Photo by the author, 2010).

15

Diet requirements There have been several insightful observations regarding the diet requirements of Mountain Nyala (e.g., Brown, 1969b; Hillman, 1985; Hillman and Hillman, 1987; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). Yet, it is believed that those observations are far from complete. Indeed, the forage use of Mountain Nyala vary with seasons, habitat types, and land-use activities (Refera and Bekele, 2004; Evangelista et al., 2007). For example, Mountain Nyala mostly forage on grass species during the early rainy season when young shoots have greater nutrition values and high digestibility (Brown, 1969b; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). As Mountain Nyala are mostly browsers, they are able to feed on a variety of vegetation. Herbs and shrubs are good sources of food for Mountain Nyala (Brown, 1969b; Hillman, 1985; Hillman and Hillman, 1987; Evangelista et al., 2007; Mamo, 2007). However, Mountain Nyala occasionally feed on , , fallen leaves, fruits, cultivars, and aquatic flora (Brown, 1969b; Hillman, 1985; Hillman and Hillman, 1987; Evangelista et al., 2007). Brown (1969b) suggests that Mountain Nyala rarely drink water. However, Evangelista et al. (2007) observed Mountain Nyala regularly drink water in the late afternoon.

Sociality and behaviors Mountain Nyala are considerably tolerant of each other so that they are often non- territorial social animals (Brown, 1969a; Hillman, 1985; Woldegebriel, 1996; Refera and Bekele, 2004; Evangelista et al., 2007). Mountain Nyala usually live by forming groups averaging from two to eleven individuals (Brown, 1969a; Hillman, 1985; Evangelista et al., 2007). However, their average group size may vary due to population density, habitat type, season, and availability of forage and water (Hillman, 1985; Refera and Bekele, 2004; Evangelista et al., 2007). Generally, average group sizes of Mountain Nyala are larger in open habitats than in forest habitats (e.g., in Bale Mountains National Park) (Hillman, 1985; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). Outside of the breeding season, Mountain Nyala segregate into single-sex herds of adult males or females (Brown, 1969a; Kingdon, 1997; Refera and Bekele, 2004; Evangelista et al., 2007) within which individual time budgets are similar and synchrony of activities is generally high as in other social ungulates (e.g., Gross et al., 1995; Muller et

16

al., 1995; Kohlmann et al., 1996; Conradt and Roper, 2000). Female groups consist of a matriarch, several mature females, juveniles and/or calves (Refera and Bekele, 2004; Evangelista et al., 2007). However, adult males usually form smaller bachelor groups of two to four individuals which are similar in age (Refera and Bekele, 2004; Evangelista et al., 2007). When males become very old, they are more likely to become solitary and avoid regular interaction with other individuals (Brown, 1969a; Evangelista et al., 2007). Mature male Mountain Nyala often display symbolic dominance through object- horning by forcefully rubbing their horns into woody vegetation or digging the ground (Brown, 1969a, 1969b; Hillman, 1985; Evangelista et al., 2007). Branches, foliages, and mud are usually lodged on horns, giving them a larger appearance that is readily shown off to other Mountain Nyala (Evangelista et al., 2007). Object-horning is primarily used to frighten other males or ascertain a symbolic dominance, and may also be used to signal fitness to females (Brown, 1969b; Hillman, 1985; Evangelista et al., 2007). Even though object-horning seems to be more frequent during the mating seasons, displays are observed throughout the year and can be common in solitary males and even among males belonging to the same bachelor group (Brown, 1969a; Hillman, 1985; Evangelista et al., 2007). The gregarious nature and social groupings of Mountain Nyala play a vital role in detecting potential predators by increasing the number of individuals that are attentive to risk of predation (Brown, 1969b; Evangelista et al., 2007). Since adult females are vulnerable to predation, their tendency to live in groups may be a survival advantage over less vigilant and solitary adult males (Brown, 1969b; Evangelista et al., 2007). During resting, a family group of Mountain Nyala arrays themselves in a defensive manner with each member having a different field of sight (Brown, 1969b; Evangelista et al., 2007). Although little is known about the sensory abilities of Mountain Nyala, the large ears are thought to reflect their sharp hearing, and the species is sensitive to humans moving upwind of them (Brown, 1969a, 1969b; Malcolm and Evangelista, 2005). Mountain Nyala swiftly move through thick natural forests and bamboo, up steep rocky slopes, and across open landscapes (Brown, 1969a; Evangelista et al., 2007). Even calves as young as two months old are extremely swift and physically capable of escaping predators in open and forested terrains (Evangelista et al., 2007). Mountain Nyala make

17

daily and to a lesser extent seasonal movements (Brown, 1969b; Hillman, 1985; Malcolm and Evangelista, 2005; Mamo, 2007).

Predation Predation is one of the major selective forces in nature, shaping the distribution, abundance, habitat selection, population dynamics, behavior, and morphology of prey species (Lima and Dill, 1990). This implies that predation considerably affects the fitness of organisms in their natural habitats. Potential predators of Mountain Nyala include (Panthera pardus), Spotted Hyena (Crocuta crocuta), and African Lion (Panthera leo) (Brown, 1969b). As Leopard and Spotted Hyena share the majority of the Mountain Nyala’s range and habitat, they may pose a significant threat to the Mountain Nyala (Brown, 1969b; Malcolm and Evangelista, 2005). However, African Lions are rarely found at the altitudes frequented by Mountain Nyala (Brown, 1969b). Mountain Nyala can outrun most predators in a variety of landscapes (Evangelista et al., 2007). So, predation of mature Mountain Nyala by predators appears to be less common (Brown, 1969b; Evangelista et al., 2007). However, newly born calves of Mountain Nyala are susceptible to predation and may be killed even by jackals (Canis mesomelas) (Evangelista et al., 2007), (Phacochoerus africanus) (Evangelista et al., 2007), Anubis baboons (Papio anubis) (personal communication with the Munessa- Kuke Controlled Hunting Area scouts), and domestic dogs (Woldegebriel, 1996).

Threats Rapid human population growth followed by extensive cultivation and overgrazing threaten many of the endemic fauna and flora species in the highlands of Ethiopia (Hillman, 1993; Evangelista et al., 2007; Atickem et al., 2011). Habitat loss and fragmentation resulting from deforestation and uncontrolled burning is by far the largest threat to the persistence of Mountain Nyala in the southeastern highlands of Ethiopia (Miehe and Miehe 1994; Malcolm and Evangelista, 2005; Mamo, 2007; Atickem et al., 2011). For example, Miehe and Miehe (1994) and Evangelista et al. (2007) note that much of the habitat of Mountain Nyala in the northern and central mountains of its range has become highly altered and fragmented by human settlement and agriculture. When human

18

population pressure increases and agriculture expands, the physical environment of Mountain Nyala is changed. Natural habitats of Mountain Nyala have been destroyed or mostly converted to other unsuitable land use types throughout their ranges of distribution (IUCN, 2008). In addition, human activities create barriers that impede range expansion in Mountain Nyala and also modify their natural behaviors (Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). Another threat to Mountain Nyala is poaching (Hillman, 1985; Woldegebriel, 1996; Refera and Bekele, 2004; Malcolm and Evangelista, 2005; Evangelista et al., 2007; Mamo, 2007). Humans are thought to be the chronic enemies of Mountain Nyala in Ethiopia. In many parts of Ethiopia, considerable damage was done to Mountain Nyala populations at the time of government change in 1991 (Woldegebriel, 1996; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). For example, although the exact number of killed individuals was not known, many Mountain Nyala were shot in the Bale Mountains National Park (Woldegebriel, 1996). In addition, a number of Mountain Nyala were indiscriminately killed by the local people in different wildlife controlled hunting areas of Ethiopia (Evangelista et al., 2007; Ethiopian Wildlife Conservation Organization (EWCO), unpublished field reports). Potential competition with livestock has also negatively affected the habitat use by Mountain Nyala (Stephens et al., 2001; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007) because livestock outcompete and prevent the Mountain Nyala from using optimal habitats. In addition, through trampling and overgrazing, livestock reduce the availability and the quality of habitats for Mountain Nyala (Mamo, 2007). For example, Mountain Nyala have never been seen to graze in an area heavily grazed by livestock in the northern edge of the Bale Mountains National Park (Malcolm and Evangelista, 2005; Mamo, 2007). Furthermore, livestock induced disturbances may cause high levels of stress and increase diseases and parasite transmission in Mountain Nyala (Sillero-Zubiri et al., 1996; Stephens et al., 2001; Evangelista et al., 2007; Mamo, 2007). Mountain Nyala currently persist in Ethiopia largely because of their choice for inaccessible areas and highland ecosystems which are too steep for cultivation and settlement (Refera and Bekele, 2004; Malcolm and Evangelista, 2005; Atickem et al., 2011). As a result, Mountain Nyala are confined to “sky islands” on a handful of peaks and

19

ridges (Evangelista et al., 2007; Sillero-Zubiri, 2008). Those “sky islands” are mostly found in wildlife protected areas of Ethiopia, such as a wildlife national park, a wildlife sanctuary, and some controlled hunting areas. However, as Ethiopia’s human population and demand for cultivated lands expand, Mountain Nyala populations will become increasingly fragmented and isolated to smaller and smaller patches of habitat even at higher elevations (Evangelista et al., 2007; EWCO, unpublished field reports). This may prevent long distance movements and dispersal of Nyala and thereby limit gene flow among populations of the species and may ultimately cause deterioration of genetic viability in Mountain Nyala (Evangelista et al., 2007; Mamo, 2007; Sillero-Zubiri, 2008).

Conservation status Mountain Nyala are protected under the Ethiopian wildlife laws both at federal and regional levels; however, anthropogenic factors are increasingly causing problems for their conservation and management (Hillman, 1985; Hillman and Hillman, 1987; Woldegebriel, 1996; Stephens et al., 2001; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007; Sillero-Zubiri, 2008). The Ethiopian Wildlife Conservation Authority (EWCA) is unable to enforce laws intended to protect Mountain Nyala and their habitats in Ethiopia (Evangelista et al., 2007; Sillero-Zubiri, 2008). Since its discovery in 1908, Mountain Nyala has declined substantially in number and shrunk in range, mainly because of habitat destruction and uncontrolled hunting (Refera and Bekele, 2004; Evangelista et al., 2007). Because of high human and livestock encroachments, Mountain Nyala are known to have been extirpated from some parts of Ethiopia (e.g., Wondo Genet) (Malcolm and Evangelista, 2005). Due to the steady decline in numbers, Mountain Nyala is listed as an endangered C1 species by the IUCN (2008).

The Study Areas The Munessa Forest The Munessa Forest is situated in Oromiya administrative regional state some 240 km south of Addis Ababa, Ethiopia. Its geographical location is at 7°27′ N and 38°52′ E (See Fig. 2.2). The altitude range extends from 2100 to 2700 meters above sea level (Teketay and Granström, 1995). Depending upon altitude, the mean annual temperature

20

varies between 15°C and 20°C (Teketay and Granström, 1995). The mean annual rainfall is about 1250 mm. The soils are reddish, freely draining, and of medium to heavy texture (Teketay and Granström, 1995). The vegetation in Munessa is composed of natural and plantation forests where the main forest blocks are found on the escarpment and associated plateau lying between the rift valley lakes and the eastern edge of the rift valley (Teketay and Granström, 1995). Munessa has an area of 111 km2 of natural and plantation forests that could be potentially inhabited by a variety of wild animal species (Evangelista et al., 2007). The natural forests 2 2 approximately cover 85 km while the remaining 26 km are plantation forests (Evangelista et al., 2007). The natural forest in Munessa is heavily degraded and characterized by many gaps resulting from uncontrolled logging. For example, some forests have been totally cleared and converted into agricultural fields, while others suffer from different influences, such as heavy grazing and selective logging (Teketay, 1992). Some of the characteristic tree species in the upper story of the Munessa natural forest include Afrocarpus falcatus, Syzygium guineense, Prunus africana, Croton macrostachyus, Millettia ferruginea, hochstetteri, Schefflera abyssinica, abyssinica, Aningeria adolfi-friederici, and Bersama abyssinica. The natural forests are vital habitat for wild animal species including the Mountain Nyala because natural forests are sources of diverse food and important cover for wild animals. The plantations in the Munessa Forest are mainly composed of exotic tree species, such as Eucalyptus globulus, Eucalyptus grandis, Cupressus lusitanica, Pinus patula, Pinus radiata, and Grevillea robusta primarily planted for timber production. The Munessa Forest District was established to manage and harvest the plantation forests and process large volume of logs into commercial lumber at its four saw mills located at Degaga. Those plantation forests are potentially good habitat for Mountain Nyala and other wild animals existing in the area. For example, Evangelista et al. (2007) note that although the plantation trees offer sparse herbaceous understory, they are very important habitat for Mountain Nyala, and may provide escape refuge from risk of predation, valuable cover for thermal regulation (e.g., Black et al., 1976), sources of food, and travel corridors especially

21

in areas where natural forests are extremely disturbed and/or limited (Evangelista et al., 2007). Mountain Nyala were not noted in Munessa when Brown did his surveys in the 1960’s. However, Mountain Nyala are believed to have flourished in the plantations in Munessa from the Arsi Mountains and have been able to survive in the area where domestic herds and farming have been excluded (Malcolm and Evangelista, 2005; personal communication with the local elders in Munessa). Wild animals in Munessa were subjected to intensive poaching during the civil unrest following the fall of the Dergue government in 1991 (Hundessa, 1992, 1997; Evangelista et al., 2007; personal communication with local people in Munessa). Then, the impact from poaching on the population of Mountain Nyala was so intense that numbers of Mountain Nyala remained low in Munessa throughout the early 1990s (Evangelista et al., 2007). However, when the area was re-established as the Munessa-Kuke controlled hunting area in 1995, laws concerning wildlife conservation and illegal hunting were quickly enforced in the area (Evangelista et al., 2007; EWCO, unpublished field reports). Mountain Nyala is most commonly used by sport hunting operators to attract visiting hunters to Ethiopia (Lindsey, 2008). In recent years, the care management by professional outfitters has rescued the Mountain Nyala in the Munessa-Kuke controlled hunting area (Malcolm and Evangelista, 2005). For example, the Ethiopian Wildlife Conservation Organization (EWCO) carried out censusing of Mountain Nyala in Munessa using a transect method in 1995 and 1999 (EWCO, unpublished field report), estimating 95 and 81 individuals respectively. In 2007, the EWCO carried out another field census in Munessa using similar method and estimated 182 individuals (EWCO, unpublished field report). However, considering the available potential habitats where Nyala could travel and use in Munessa, Evangelista et al. (2007) argue that the population size should be greater than the EWCO’s estimates and could be closer to 200 individuals. Following hunting quotas proposed by EWCO, legal sport hunting has been practiced in Munessa-Kuke controlled hunting area. Some of the huntable wild game species in the Munessa hunting block include Mountain Nyala, Menelik’s Bushbuck (Tragelaphus scriptus meneliki), Anubis baboon, Black and white colobus monkey (Colobus guereza), and Spotted Hyena (EWCO, unpublished field report; personal communication with the Munessa-Kuke controlled hunting area scouts).

22

The use of licensed legal sport hunting as a means of conservation may allow the Mountain Nyala and other wild animals to make dramatic recoveries in Munessa (Malcolm and Evangelista, 2005; Lindsey, 2008; Sillero-Zubiri, 2008). This is because part of the money could be allocated to promote infrastructure developments which enhance wildlife and habitat conservation activities. The presence of trophy hunting operators could reduce illegal hunting. For example, in Zambia and Tanzania, wildlife sport hunting lease agreements require assistance with anti-poaching activities from hunting operators in hunting concessions (Lindsey, 2008). Even where anti-poaching is not a legal pre- requisite, operators often conduct anti-poaching patrols to protect the wildlife resource on which they depend (Lindsey, 2008). Trophy sport hunting is increasingly used as a tool for wildlife conservation in the sub-Saharan African countries including Ethiopia (e.g., Lindsey et al., 2007; Lindsey, 2008). However, it should be ecologically sustainable, economically feasible, and socially acceptable so as to be an efficient conservation tool (Sparrowe, 1990). Legal trophy sport hunting in the Munessa hunting block generates revenue which helps emphasize the economic value of wildlife to the local people through promoting a benefit-sharing scheme (Evangelista et al., 2007; EWCO, unpublished field report). For example, of the hunting revenue received from professional outfitters in Ethiopia, 20% of the revenue is allocated to local communities residing in the hunting concessions (Kubsa, 1999). Munessa is rich in biodiversity, mosaic of landscape, and human cultures that could be potential tourist attractions. Introducing community-based conservation efforts that allow communities to derive economic benefits from ecotourism may promote conservation while at the same time providing a solution to resource use conflicts in Munessa. Ecotourism activity can also improve and diversify the economic incomes of the local people through creating job opportunities (e.g., Fetene et al., 2012) such as tourist guiding services, souvenir selling, horse renting, etc., all of which can help make ecotourism economically viable in Munessa. Most importantly, Munessa is situated very close to the tourist destination, Lake Langano. A number of comfortable private safari lodges have been built around Lake Langano (e.g., Bishan Gari lodge, Wenney lodge, Sabbana beach resort, etc.). So, developing Munessa for wildlife-based tourism and ecotourism seems to be a promising business in the future.

23

Figure 2.2. Map of Munessa where most of the studies were conducted. The sample transect walks aligned in each habitat type (i.e., clear cut, plantation, and natural forest) are shown.

24

The Bale Mountains National Park Following the recommendation made by Leslie Brown, the Bale Mountains National Park (BMNP) was established in 1970. It is located between 60 29` and 70 10` North latitude, and 390 28` and 390 58` East longitude (Hillman, 1986; Stephens et al., 2001) (See also Fig. 2.3). The park was primarily established to protect Mountain Nyala, (Canis simensis), and other endemic species found in the Bale Mountains (Waltermire, 1975). The BMNP is situated in Oromiya administrative regional state some 400 km southeast of the capital Addis Ababa, Ethiopia. The park comprises an area of 2,200 km2 (Hillman, 1986). The altitude of the park extends from 1,500 meters at the southern edge to 4,377 meters above sea level at the summit of Mount Tulu Dimtu in the central peaks (Yalden and Largen, 1992). The annual and the daily temperatures in the park largely vary with altitude of the area. The highest peak in the park experiences the lowest temperature in the clear sky night in the dry season, but the highest temperatures in the day time during the same season (Yalden and Largen, 1992; Williams, 2002). The rainfall in the park is mainly characterized by one long rainy season from March through October with the greatest bulk of the rains falling in April and then August through October (Hillman, 1986; Williams, 2002). The mean annual rainfall differs with altitude of the area. For example, lower altitudes in the park receive between 600 mm and 1000 mm of annual rainfall; higher altitudes receive up to 1200 mm of annual rainfall (Williams, 2002). The Massif Mountains in the park play an essential role in climate control in the region by attracting large amounts of orographic rains (Hillman, 1986; Yalden and Largen, 1992; Williams, 2002). The park is the main source of water in the region. For example, over forty streams originate within the park boundary (Yalden and Largen, 1992; Williams, 2002). These streams join to form four major rivers which flow to the adjacent low-lying areas. It is known that the BMNP is the water tower in the country (Hillman, 1986; Williams, 2002). The BMNP has been intensively studied so that there are several detailed descriptions of the park’s flora and fauna (e.g., Hillman, 1986, 1988; Nigatu and Tadesse, 1989; Yalden and Largen, 1992; Miehe and Miehe, 1994; Gashaw and Fetene, 1996; Bussman, 1997). The BMNP is generally divided into five major vegetation zones

25

primarily classified based on altitude. Each of these zones has its own characteristic flora and fauna. The first vegetation zone in the park comprises the small mountain grassland in the Gaysay Valley which is situated in the northern edge of the park (Hillman, 1986, 1988; Williams, 2002). This grassland is rich in wildlife including Mountain Nyala (estimated to be approximately two-thirds of the global population) (Mamo, 2007), and other wild species such as Bohr (Redunca redunca), Menelik’s Bushbuck, and . The second vegetation zone lies northern slope of the park and is covered with woodlands dominated by tree species such as Hagenia abyssinica, Juniperus procera, and revolutum (Hillman, 1986; 1988; Yalden and Largen, 1992). The third zone is a belt of heather that is located above the second zone. The fourth zone is the central peaks and the plateau consisting of the Afro-alpine moorland (Williams, 2002). The Afro-alpine moorland is the largest area containing the Afro-alpine habitats (about 1,000 km2) in the African continent (Williams, 2002). The Afro-alpine is a suitable habitat for a number of rodent species, most of which are endemic to Ethiopia (Yalden and Largen, 1992; Williams, 2002). The afro-alpine area is also home to over half the global population of the Ethiopian wolves, which is one of the rarest and most in the world (Yalden and Largen, 1992). The fifth zone is situated in the southern escarpment and it is called the Harrena Forest. The Harrena Forest is mainly composed of dominant plant species such as Erica arborea, , Dombeya torrida, Rapanea simensis, Arundinaria alpina, Hagenia abyssinica, Schefflera abyssinica, Aningeria altissima, Syzygium guineense, Croton macrostachyus, Afrocarpus falcatus, and Cordia africana (Nigatu and Tadesse, 1989; Yalden and Largen, 1992; Bussman, 1997). The Harrena Forest is the second largest stand of moist tropical forest remaining in Ethiopia and includes the only cloud forest in the country (Nigatu and Tadesse, 1989; Bussman, 1997; Williams, 2002). The BMNP harbors unique and diverse fauna and flora resources in Ethiopia. For example, Yalden and Largen (1992) underlined the importance of the Bale Mountains as a center of endemism and reservoir of unknown genetic resources. Of the park’s recorded wild mammal species, about 26% of them are Ethiopian endemics, and there are also species that are only confined within the park itself (Yalden and Largen, 1992; Hillman,

26

1993). Some of the larger endemic wild mammal species in the park include Mountain Nyala, Ethiopian Wolf, Bale Monkey (Cercopithecus djamdjamensis), Giant Molerat (Tachyoryctes macrocephalus), and Stark’s Hare (Lepus starcki). Over 280 species of birds have been recorded in BMNP of which six species are endemic to Ethiopia (Yalden and Largen, 1992). The beautiful mountain scenery, sweeping valleys, and mosaic of human cultures coupled with the unique and diverse wild fauna and flora made the Bale Mountains National Park one of the most popular tourist destinations in the south-eastern highlands of Ethiopia. Despite its natural potentials, the BMNP has faced various problems. At the time of its establishment, there were few permanent inhabitants around the park (Waltermire, 1975; Evangelista et al., 2007). However, natural human population growth, immigration of people, and government relocation programs have all contributed to the high human population density now found in the area (Stephens et al., 2001; Mamo, 2007). As a result, deforestation due to fuel wood collection, clearing of forests for crop cultivation, and illegal commercial logging are major threats to the park (Hillman, 1993; Tedla, 1995; Stephens et al., 2001; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). Many of the local people living near the park rely on livestock rearing for their livelihood (Stephens et al., 2001; Evangelista et al., 2007; Mamo, 2007). As most of the traditional grazing lands outside the park have been converted to crop cultivation, the local people largely depend on the grazing lands within the park (Hillman, 1986; Stephens et al., 2001; Evangelista et al., 2007; Mamo, 2007). At the end of the dry season, the local people frequently burn the Erica heather to boost the growth of palatable forage for their livestock (Miehe and Miehe, 1994; Evangelista et al., 2007). All these are major threats to the conservation and management of wildlife and its habitats in the park. The boundary of the BMNP, particularly in the northern part, has been changed three times with changes in governments (Mamo, 2007). Consequently, the exact location of the boundary between the park and villages outside remains an issue of dispute (Mamo, 2007). In addition, the BMNP has never been formally gazetted (Hillman, 1986, 1993; Mamo, 2007); therefore, settlements and agricultural land expansion within the park’s boundary and illegal use of the park’s natural resources are difficult to legally control (Stephens et al., 2001; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007).

27

Figure 2.3. Map of the Bale Mountains National Park as one of the study areas.

28

CHAPTER 3 HABITAT SELECTION BY MOUNTAIN NYALA ASSESSED WITH A HABITAT SUITABILITY MODEL AND ISODAR ANALYSIS ABSTRACT

Animals live in habitats that vary in quality, and the use of those habitats can influence an individual’s fitness. Habitat suitability index models and isodar analyses were applied to determine the seasonal habitat selection of Mountain Nyala in Munessa, Ethiopia. I hypothesized that Mountain Nyala select natural forest habitat compared with any other existing habitats in Munessa. I predicted that the abundance and distribution of Mountain Nyala is higher in the natural forest habitat than in other habitats. Activity densities of free-ranging Mountain Nyala were estimated through regular censusing following permanent transects aligned through three major habitat types. The fieldwork was carried out in the wet and the dry seasons. In addition, with the help of a spotlight, night time censusing of Mountain Nyala was carried out during the dry season in each of the three habitat types. Separate habitat suitability models were developed for the wet and dry season field data. The results revealed that Mountain Nyala did not show significant habitat selection behavior during the wet season. However, in the dry season, the natural forest was the most selected habitat by Mountain Nyala, when crown diameter of trees significantly affected the habitat suitability. The regression analyses of the dry season isodars revealed that the natural forest habitat was qualitatively, but not quantitatively, more suitable than both the plantation and the clear cut habitats. The slopes of the isodars suggest that the strength of density-dependence in the natural forest habitat is 2.4 and 2.7 times lower than in the plantation or the clear cut habitats, respectively. However, the result with spotlight censusing revealed that Mountain Nyala select the clear cut habitat during the night time at which they can more easily detect approaching potential predators (e.g., ) and escape in the open habitat than in the dense vegetation, or when people and livestock are absent in the area so that Mountain Nyala can access the forage in the open more freely. Key words: activity density, density-dependent habitat selection, habitat suitability models, isodars, Mountain Nyala, Munessa, season

29

INTRODUCTION The natural environment where an animal lives is composed of habitats that vary in quality (Druce, 2005), and the use of those habitats can influence an individual’s ability to survive and reproduce (Melton, 1987). Information on habitat use can be summarized in various types of habitat models that simulate the relationships between an animal population and its habitats (e.g., Verner et al., 1986; Druce, 2005; Mamo, 2007; Evangelista et al., 2008). These include models of habitat suitability that identify features of the environment that correlate with population density (Druce, 2005; Mamo, 2007) and models of habitat selection that identify salient features of the environment that determine how individuals choose to distribute themselves in space and time (e.g., Rosenzweig, 1974; 1981; Morris, 2003). Traditionally an animal’s habitat selection has been determined by identifying habitat features that affect habitat suitability (Fish and Wildlife Services, 1980; Conway and Martin, 1993). These could be features that directly promote reproduction and survival, as these are key fitness determining components (Conway and Martin, 1993; Block et al., 1998). This suggests that population distribution and habitat selection behavior of a particular species may be governed by factors such as resource availability and distribution (e.g., Boyce and McDonald, 1999; Atickem et al., 2011), season (e.g., Store and Jokimäki, 2003; Tadesse and Kotler, 2010), habitat characteristics (e.g., Druce, 2005; Reid, 2005), foraging behavior (e.g., Kotler et al., 1994), and availability of escape refuges (e.g., Brown and Alkon, 1990; Kotler et al., 1994). Relative habitat use can be inferred by directly observing and counting the number of individuals of the study species along transects aligned across major habitat types (e.g., Druce, 2005; Mamo, 2007; Tadesse and Kotler, 2010; Atickem et al., 2011). However, to maximize the usefulness of activity density estimates as an indicator of habitat quality of a particular species, the field data should be collected over a range of seasons and environmental conditions (e.g., Van Horne, 1983; Store and Jokimäki, 2003; Druce, 2005; Tadesse and Kotler, 2010). Habitat selection behavior provides useful information for management and conservation. Wildlife conservation problems can be solved with the proper application of habitat-selection theories (Morris, 2003). Theories of habitat selection can be applied in various ways to create a set of relatively simple behavioral assays that provide leading

30

indicators of habitat quality, habitat use, and habitat change (Morris, 2003; Morris et al., 2009). Many of the wildlife conservation and management solutions involve the use of habitat indices including habitat suitability index modeling (e.g., Schamberger and O’Neil, 1986; Verner et al., 1986; Fabricius and Mentis, 1991; Van Horne and Wiens, 1991; Evangelista et al., 2008) and habitat isodars (e.g., Morris, 2003; Morris et al., 2009). I briefly describe each in turn. Habitat suitability index (HSI) models are the simplest and perhaps the most frequently used form of ecological models (Schamberger and O’Neil, 1986). HSI models are intended to be general indicators of habitat suitability that are easily and reputably applied under field conditions (Uhmann et al., 2001). Habitat suitability modeling is a tool for predicting the suitability or quality of habitat for a particular species based on known affinities for environmental or habitat parameters (Van Horne and Wiens, 1991; Morrison et al., 1998). HSI models can be constructed based on the concepts of environmental variables and relative animal activity densities (Schamberger and O’Neil, 1986; Van Horne and Wiens, 1991; Tadesse and Kotler, 2010). The models assume that habitat is an important factor in determining the presence and relative abundance of the species in question (Farmer et al., 1982). The selection of appropriate environmental variables largely depends not only on the species studied, but also on the costs of collecting the variables and the purpose of use of the empirical data (Store and Jokimäki, 2003). Minimizing the number of variables in the HSI model serves two purposes: the model becomes more easily applied, and the likelihood of model over-fitting is reduced (Jeffers, 1982). However, to maximize the usefulness of the habitat suitability models, it is important that they should be constructed using as much prior information as possible (Johnson, 1980). The Habitat Suitability Index (HSI) is an equation that defines the contribution of the various environmental inputs in determining habitat suitability (Fish and Wildlife Services, 1980; Fabricius and Mentis, 1991; Van Horne and Wiens, 1991; Tadesse and Kotler, 2010). HSI is a unitless number scaled between 0 and 1 where 0 represents unsuitable habitat and 1 represents optimum habitat (Fish and Wildlife Services, 1980). HSI scores are used to compare among sites, over time, or between management scenarios (Fabricius and Mentis, 1991; Van Horne and Wiens, 1991). As animal density reveals information about the past conditions of the habitat in the system, it is a trailing indicator of

31

habitat quality (Morris et al. 2009). However, by using the behavior-based approach based on activity densities rather than population densities, one would allow for the development of habitat suitability models that provide a more leading indicator of habitat quality (e.g., Druce, 2005; Reid, 5005, Morris and Mukherjee, 2007; Tadesse and Kotler, 2010). Isodars are graphical lines in a state space of population densities in which fitness is equal across two adjacent habitats, reflecting an Ideal Free Distribution (Morris, 1987, 1988, 2003). Empirically, an isodar is obtained by regressing the density of the study species in one habitat type against the density in the other adjacent habitat. The slope of the isodar and its intercept reveal basic information on the underlying mechanisms of density- dependent habitat selection behaviors (Morris, 1987, 1988). If the slope is significantly different from 1, then the two adjacent habitats differ qualitatively (Morris, 1987, 1988, 2003). Qualitative differences reflect differences in the strength of density dependence. If the Y-intercept is significantly greater than zero, then the two habitats differ quantitatively (Morris, 1987, 1988). Quantitative differences usually refer to differences in resource availability or productivity. I examined habitat suitability and habitat selection behavior of Mountain Nyala in the Munessa Forest of Ethiopia. Natural forests provide good escape refuges (e.g., Mech, 1977; Ripple and Beschta, 2003, 2004; Evangelista et al., 2007), critical cover for thermal regulation (e.g., Black et al., 1976; Evangelista et al., 2007), and a wide opportunity for availability of quality forages (e.g., Evangelista et al., 2007). I hypothesized that Mountain Nyala prefer to select the natural forest habitat to any other existing habitats in Munessa. I predicted that the abundance and distribution of Mountain Nyala is higher in the natural forest habitat than in other habitats. Accordingly, I examined the habitat selection behaviors of Mountain Nyala to: (1) develop a habitat suitability index (HSI) model that would account for the variation in their activity densities across habitat types and seasons, (2) apply the isodar technique to look for density-dependent habitat selection behavior in Mountain Nyala, and (3) investigate the environmental factors that correlate to the relative activity densities of Mountain Nyala across seasons in Munessa.

32

METHODS Habitat Inventory First I conducted a reconnaissance survey to identify the habitats characterized by vegetation types. I identified three major types of habitat over my study area: natural forest, plantation, and clear cut habitats (see also Fig. 2.1). I briefly describe each of these three habitat types as follows. The natural forest habitat: The natural forest is characterized by heavily human disturbed upper canopy with a diversified understory. Characteristic indigenous tree species in the upper storey of the natural forest include Afrocarpus falcatus, Syzygium guineense, Prunus africana, Bersama abyssinica, Aningeria adolfi-friederici, Hagenia abyssinica, Celtis africana, Millettia ferruginea, and Croton macrostachyus. The natural forest habitat in Munessa is situated in undulating terrain so that it provides good escape and refuge for Mountain Nyala from risk of predation and human nuisance. Human nuisance is briefly defined as one form of anthropogenic disturbances that could induce behavioral responses similar to those associated with predation risk (Frid and Dill, 2002).The natural forest serves as valuable cover for thermal regulation and provides a good source of palatable forages for Mountain Nyala. I note that there are also several seasonal streams and permanent rivers (e.g., Mukenissa and Dalele rivers) originating from the natural forest habitat from which the Mountain Nyala and other wild animals get water. However, there are various human and livestock induced impacts on the natural forests in Munessa. For example, some natural forests have been totally cleared and converted into agricultural fields, while others suffered from different influences, such as heavy grazing and selective logging (Teketay, 1992). Illegal tree cutting for fuel wood and construction materials as well as free-range livestock grazing and browsing are the most prominent threats to the natural forest habitat especially in the crop growing rainy season (personal observations). The plantation habitat: The plantation forest is mainly composed of exotic tree species, such as Eucalyptus globulus, Eucalyptus grandis, Cupressus lusitanica, Pinus patula, Pinus radiata, and Grevillea robusta. The plantations in Munessa are potentially good habitat for Mountain Nyala. Although the plantations offer a sparse herbaceous understory in the dry season, they provide a very important cover and other resources

33

during the wet season following the good rains (Evangelista et al., 2007). For example, plantation forests provide escape refuge from risk of predation and human disturbances, valuable cover for thermal regulation, sources of food, and travel corridors especially during the rainy season in areas where natural forests are extremely disturbed and /or limited (e.g., Evangelista et al., 2007; Ethiopian Wildlife Conservation Organization, unpublished field report). Moreover, young plantations (e.g., stand of Cupressus lusitanica) with unprunned branches also provide the Mountain Nyala with good cover for concealment from risk of predation in Munessa (personal observation). Illegal tree cutting activities by the local people for fuel wood and construction materials as well as free-range livestock grazing and browsing are also the common threats to the plantation habitat in Munessa especially in the crop growing rainy season. Moreover, every year, the Munessa Forest District harvests a large volume of standing trees from the plantation forest for commercial lumber production. The clear cut habitat: The clear cut habitat is characterized by relatively freely draining areas that are rich in grass and other palatable herbs for Mountain Nyala. So, it serves as good feeding habitat especially in the rainy season. In the clear cut habitat, there are also some salt licks which attract a number of Mountain Nyala during the night time when people and livestock are not around. The salt licks are likely locations for seeing groups of Mountain Nyala in the rainy season. As most of the clear cut habitat is surrounded by natural forests and plantations, it is fairly easy for Mountain Nyala to escape from predators and human nuisance including illegal hunting. The clear cut habitat is repeatedly replanted with seedlings by the Munessa Forest District, but these often failed to survive due to over grazing and browsing by livestock and lack of proper forest management (personal communication from Munessa Forest District workers at Degaga). In order to collect Mountain Nyala habitat use data from all the existing habitats across the landscape, I set out permanent walking transects with the aid of a GARMIN 75 GPS device, with each transect sampling a major existing habitat type within the study site. I established a total of 12 permanent transects i.e. four transects in each habitat type (see also Fig. 2.1). Following Druce (2005) and Tadesse and Kotler (2010), along each transect, I quantified the viewable area of each habitat by walking perpendicularly from a given line transect until the unevenness of the topography or the thickness of the vegetation cover no

34

longer allowed me to view that transect. The GPS locations on both sides of all the viewable parts of each habitat type were then taken. This activity is important in determining the sample area of each habitat type along each transect walk. Following the minimum sample area proposed by Patton (1992), the sampling protocols covered 2-5% of the total area of the study site. Transects varied in length from 0.8 km to 2.3 km, i.e. the length of each transect varied with the size of each habitat patch. The total area of my study site was about 111 km2. So, I took transect samples whose total area is 3.83 km2, which is 3.45% of the study site.

Population Censusing Day time Mountain Nyala censusing: To measure activity density of Mountain Nyala along transects, I conducted regular population censusing by walking along each transect. During each survey, I recorded the habitat type of each observed Mountain Nyala along with its gender, age class, and group size. The season at the time of field observation was also recorded. Counts were carried out early in the morning from 6:00am to 9:00am local time when Mountain Nyala are most active. In addition, humans and livestock are absent in the area during the early morning hours. In order to avoid any likely disturbance caused by the observers on each Mountain Nyala during the transect walks, the field observation and censusing was carried out carefully and calmly. Moreover, binoculars were used to detect the Mountain Nyala at a far distance because approaching the Mountain Nyala in close distance may disturb and urge them to flee earlier before the counting was done. I conducted the field surveys from May through August 2010 for the wet season and from December 2010 through April 2011 for the dry season. Each transect was assessed six times for the wet season, but seven times for the dry season. Night time Mountain Nyala censusing using a spotlight: Due to the intensive human and livestock disturbances in the late dry season, the Mountain Nyala in Munessa become active during the night time when people and livestock are absent in the area. I found that spotlighting complemented the daylight field data on the activity density of Mountain Nyala in Munessa. In this study, the spotlighting was carried out from 8:30pm to 12:00pm local time from December 2010 through April 2011. Each of the three habitats

35

(i.e. clear cut, plantation, and natural forest) in the study site was surveyed ten times either by car or on foot for times when the area is inaccessible for driving.

Measurement of Habitat Variables I quantified environmental variables in each habitat patch along each permanent transect used for estimating the activity densities of Mountain Nyala as outlined in the previous section. This then allowed me to correlate activity densities and time budgets with the various environmental variables to construct habitat suitability indices and behavioral models (see also Chapter 4) for Mountain Nyala in Munessa. A systematic sampling design was employed to collect important microhabitat and environmental data. The first plot within each habitat patch on each transect was randomly located; then successive plots were added at 100 meter intervals along each transect. The total number and distribution of sample plots for each habitat type varied with the total size of each habitat patch. A total of 109 plots were assessed (i.e. 31 plots in the clear cut habitat, 41 plots in the plantation habitat, and 37 plots in the natural forest habitat). I laid out a circular sample plot with a radius of 5 meters on each line transect. All trees within the 5 meters radius circular plot were identified and counted, and sample crown diameter for each tree species was measured with a meter tape. A circular nested plot with a radius of 2 meters was laid out within the larger circular plot and then all shrubs within this nested plot were identified and counted. Within each nested plot, another sub-nested plot with a radius 0.5 meters was laid, and percent plant cover (grasses and herbs) and percent cover bare soil were estimated with a square grid. In addition, at the center of each circular sample plot, the slope condition (using declinometer) and elevation (altitude) (using altimeter) were measured. The field data were collected from May through August 2010 for the wet season, while the dry season data were collected from December 2010 through April 2011.

Data Analysis Activity densities In order to estimate the activity density of Mountain Nyala in each habitat type, I first created a visibility map for each transect by which I determined the sampled area. The

36

GPS locations which were taken on both sides of the visible points for each transect were transformed into the Ethiopian coordinate system and imported into Arc GIS 9.3 program, and overlaid on the map of the Munessa Forest. I then digitized the GPS coordinates of all the visible points for each transect walk using lines to make a polygon along all the transect walks. This gave me the areas of all the habitats sampled in each transect as shown in Table 3.1.

Table 3.1. A summary of the twelve transects used in the activity density analysis. The visible area of each habitat covered by transects is included.

Total Transect No visible Habitat type area in all transects 1 2 3 4 5 6 7 8 9 10 11 12 (ha) Clear cut 120.9 36.1 0 0 37.3 27.6 19.9 0 0 0 0 0 0 Plantation 142.3 0 37.8 38.6 0 0 0 0 0 0 38.2 0 27.7 Natural forest 119.9 0 0 0 0 0 0 29.8 45.7 18.8 0 25.6 0 Total 383.1 36.1 37.8 38.6 37.3 27.6 19.9 29.8 45.7 18.8 38.2 25.6 27.7

I incorporated the above generated sample area information (Table 3.1) with population census data obtained through transect walk counting during the wet and the dry seasons. I divided the number of Mountain Nyala counted along each transect during each census by the visible area of the respective transect for each habitat type. That enabled me to estimate the activity density of Mountain Nyala in each habitat for the wet and the dry seasons as shown in Table 3.2. As the data were composed of both categorical and continuous variables, I used ANOVA to check whether Mountain Nyala were exhibiting a choice in their habitat use (Snedecor and Cochran, 1989) during the wet and the dry seasons. To do so, I included habitat type as a predictor, with activity density of Mountain Nyala as the dependent variable. I defined alpha value of 0.05. I did the analysis with STATSTICA version 10.

Habitat suitability models I developed habitat suitability models with the assumption that all sex and age classes of Mountain Nyala have equal access to all prevailing habitat types in Munessa. I

37

developed distinct habitat suitability index models for the wet and the dry seasons. I correlated environmental variables to the activity densities of Mountain Nyala in order to obtain models of habitat suitability. For environmental parameters, I incorporated abundance of trees, crown diameter, abundance of shrubs, percent cover of herbs and grass, percent cover of soil substrates, percent of slope condition, and elevation (altitude). I then combined them with the activity density estimates generated in the previous section. Given the number of environmental parameters under consideration, I used only the simplified response curves for those environmental parameters included in this study. Accordingly, I converted the range of activity densities of Mountain Nyala into a 0 to 1 scale, with 0 representing poor habitats and 1 representing optimal habitats for Mountain Nyala (Fish and Wildlife Services, 1980). In that way, I transformed the samples into a habitat suitability index (HSI) scaled from 0 to 1. Taking the typically non-linear nature of the relationships between environmental variables and the relative activity densities of Mountain Nyala into account, I used a polynomial regression by which separate HSI response curves were generated for the wet and the dry seasons (Gutzwiller and Anderson, 1986; Ter-Braak and Looman, 1987). Multiple linear regression analysis is normally applied when activity density data are available (Jokimäki and Huhta, 1996). All the environmental data included in this study were composed of continuous and discrete variables. I checked for multicollinearity among all the independent variables, but none of the independent environmental variables considered in the habitat suitability model showed multicollinearity. Accordingly, I used multiple linear regressions to determine the coefficients of the HSI variables entered into the models for the wet and dry seasons. The multiple regressions also help to determine the statistical contribution of each environmental variable to the habitat suitability model during the wet and the dry season. I defined alpha value of 0.05. I did the analysis with STATSTICA version 10.

Isodars Many ecological theories that deal with the different factors of habitat selection are developed based on the assumption of IFD (Ideal Free Distribution) (Fretwell and Lucas, 1969). The IFD provides a basis for understanding how individuals should distribute

38

themselves among habitats in response to habitat quality and population density. It emerges when habitat selection is density-dependent and animals choose their habitats based on the highest available fitness returns and can enter their chosen habitat on the same basis as those already present (Fretwell and Lucas, 1969). Such a distribution of individuals can be revealed with habitat isodars by plotting the densities of individuals in adjacent habitats (e.g., Morris 1987, 1988, 2003). Morris (1995, 2003) demonstrated that isodars reflect quantitative (differences in productivity) and qualitative (differences in density dependencies) differences in habitat quality, and specify the conditions when population density will, or will not, match the abundance of resources. Below is an example of an isodar and how it is derived from the densities of individuals that are ideally distributed in two adjacent habitats. The intercepts and the slopes of the isodar reveal basic information about density-dependent habitat selection (Morris, 1995, 2003).

a b

Figure 3.1. (a) A simple representation of the ideal free model of density-dependent habitat selection. Two habitats are shown, each with a characteristic shape and decline in reproductive success with increasing density. At low density, individuals should choose habitat A because their expected fitness is greater than in habitat B. The expected fitness in habitat A will be reduced with increases in density. Individuals should begin to occupy habitat B when the average fitness there is equivalent to that in A. The densities should be adjusted by movement between habitats such that the average fitness is equal in both (horizontal lines, the pairs of points represented by symbols are replotted in (b)). The pair of habitat depicted here are perceived to differ from one another qualitatively (different slopes - differences in the intensity of density-dependence) and quantitatively (different intercepts - often from productivity differences). (Source: Morris, 1995). (b) An isodar generated from fitness-density curves depicted in (a). The isodar plots the set of densities in habitat A versus those in B such that the expected fitness of an individual is the same in both (intersections of all possible horizontal lines with the fitness-density curves). The fitness-density curves, in this case, diverge from one another, yielding an isodar with slope >1.0. (Source: Morris, 1995).

Several studies (e.g., Druce, 2005, Reid, 2005, Morris and Mukherjee, 2007;

39

Tadesse and Kotler, 2010, and the references there in) confirmed the possibility of activity densities of animals (behavioral assays) as input variables for isodar analyses in order to reveal basic information on the underlying mechanisms of density-dependent habitat selection behaviors by individuals of a species. Accordingly, individuals should have higher activity density (spend much more time) in a habitat where their fitness is higher compared with the adjacent habitat (Fretwell and Lucas, 1969; Morris, 1987, 1988, 1995, 2003). Because behavior is adaptive, the resulting measures are leading indicators for a more proactive management approach (Kotler et al., 2001; Morris et al., 2009). In the present study, to analyze the isodars, I used the daily activity densities of Mountain Nyala, which showed significance differences among the alternative adjacent habitats. The isodar analyses were restricted to those conditions where both habitat types were occupied and their activity densities were greater than zero in both habitat types since inclusion of zero activity densities can bias the isodar analyses, and hence may not yield a uniquely determined activity density in the alternative habitat (Morris, 1987; 1988; 2003). Adjacent habitats which met the aforementioned criteria were natural forest versus plantation habitats and natural forest versus clear cut habitats in the dry season. I defined alpha value of 0.05. I performed the analysis with STATSTICA version 10.

RESULTS Activity Densities Mountain Nyala activity density estimated from day time censusing Mountain Nyala did not show any significant habitat selection behavior in the wet season (F (2, 69) = 2.58; p = 0.829) (Fig. 3.2). In contrast, during the dry season, the activity density of Nyala in the natural forest habitat differed significantly from all other habitats (F

(2, 81) = 24.19; p < 0.001). Mountain Nyala selected the natural forest habitat in the dry season, being seen there about three quarters of their time (≈75.00%) and achieved their greatest maximum density (0.94 Mountain Nyala/ha; Table 3.2). The maximum densities varied between 0.28 and 0.94 Mountain Nyala/ha in the dry season (Table 3.2). In addition, the dry season field survey showed that the activity density of Mountain Nyala was the highest in the natural forest habitat (mean = 0.31 Mountain Nyala /ha; stdev = 0.25), followed by the plantation habitat (mean = 0.06 Mountain Nyala /ha; stdev = 0.09), and

40

was the least in the clear cut habitat (mean = 0.04 Mountain Nyala /ha; stdev = 0.08), (Table 3.2). During the dry season, a post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in activity density between natural forest versus clear cut habitats and natural forest versus plantation habitats (Fig. 3.2).

Table 3.2. A summary of the activity densities of Mountain Nyala (MN) in different habitat types during the wet and the dry seasons field survey: mean and maximum activity densities of MN were shown as the number of times MN were seen in each habitat type and as a percent of total number of times that habitat was sampled. N = number of transects sampled that habitat type. The minimum activity densities of MN in all habitat types both in the wet and the dry seasons were zero.

Maximum Sightings Season Habitat type N Mean Density Number of per transect Density (MN/ MN sightings walk (MN/ ha) ha) (%) Clear cut 24 0.24 1.02 16 66.67 Plantation 24 0.12 0.31 16 66.67 Wet Natural forest 24 0.15 0.51 17 70.83 Total 72 0.51 1.84 49 68.06 Clear cut 28 0.04 0.28 7 25.00 Plantation 28 0.06 0.28 13 46.43 Dry Natural forest 28 0.31 0.94 21 75.00 Total 84 0.46 1.50 45 53.57

0.6 Clear cut Plantation 0.5 Natural forest

0.4

0.3

0.2 Density (MN/ha)

0.1

0 Wet Dry Season

Figure 3.2. Seasonal habitat use by Mountain Nyala in Munessa. The error bars represent +1 stdev.

41

Mountain Nyala activity density estimated from night time censusing using a spotlight The area of each transect for each habitat type was adapted from the forest compartment map of the Munessa Forest District. The area of each transect for each habitat type was as follows: clear cut = 212.38 ha, plantation = 216.24 ha, and natural forest = 204.36 ha. The sample area of each transect was combined with the spotlight census data to analyze the night time activity densities of Mountain Nyala in each habitat type as shown in Table 3.3. In contrast to the day time dry season census data (see also the above section), the Mountain Nyala in Munessa changed their habitat selection during the night time. Spotlighting revealed that the activity density and habitat selection behavior of Mountain Nyala in the dry season was significantly biased to the clear cut habitat during the night time (F (2, 27) = 29.12; p < 0.001) (Fig. 3.3). The result showed that the activity density of Mountain Nyala was the highest in the clear cut habitat (mean = 0.13 Mountain Nyala /ha; stdev = 0.04), followed by the plantation habitat (mean = 0.06 Mountain Nyala /ha; stdev = 0.01), and was the least in the natural forest habitat (mean = 0.05 Mountain Nyala /ha; stdev = 0.02), (Table 3.3). A post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in activity density between clear cut versus plantation habitats and clear cut versus natural forest habitats (Fig. 3.3).

Table 3.3. A summary of night time activity densities of Mountain Nyala (MN) in different habitat types in the dry season with a spotlight censusing: minimum, mean, and maximum activity densities of MN were shown as the number of times MN were seen in each habitat type and as a percent of total number of times that habitat was sampled. N = number of transects sampled that habitat type.

Minimum Mean Maximum Number of Sightings per Habitat type N Density Density Density MN transect walk (MN/ ha ) (MN/ ha) (MN/ ha) sightings (%) Clear cut 10 0.09 0.13 0.21 10 100.00 Plantation 10 0.04 0.06 0.07 10 100.00 Natural forest 10 0.02 0.05 0.11 10 100.00

42

0.18 0.16 0.14 0.12 0.1 0.08 0.06

Density (MN/ha) Density 0.04 0.02 0 Clear cut Plantation Natural forest Habitat type

Figure 3.3. Dry season habitat use by Mountain Nyala in Munessa assessed with a spotlight censusing technique during the night time. The error bars represent +1 stdev.

Habitat Suitability Models Wet season habitat suitability models I developed the habitat suitability models by correlating the activity densities of Mountain Nyala with measurements of the habitat variables. The habitat suitability index for Mountain Nyala slightly increased with an increase in the abundance of trees; slightly increased up to a certain point and then steadily decreased with an increase in crown diameter of trees; was independent of the abundance of shrubs; slightly decreased and then steadily increased with an increase in the percent cover of grass and herbs; steadily decreased with an increase in the percent cover of soil substrates; was constant up to a certain altitude and then slightly decreased with an increase in altitude of the landscape; and slightly increased with an increase in the percent slope condition over the landscape (Fig. 3.4). None of the habitat variables entered into the habitat suitability model had a beta coefficient that is significant at 0.05 alpha value (Table 3.4). Overall, the wet season multiple linear regression models revealed that the seven habitat variables altogether explained 15.1% of the variance for the habitat suitability (Table 3.4).

43

Table 3.4. A summary of multiple linear regression statistics and constituent variables of the habitat suitability index model for Mountain Nyala in Munessa during the wet season. The incorporated variables are listed in the order in which they were entered into the model.

2 F(6,59) P Multiple R Habitat variables Coefficients t(1, 59) p 1.743 0.127 0.151 Intercept (constant) -0.969 -1.333 0.188 Abundance of trees -0.221 -1.596 0.116 Crown diameter of trees (m) 0.018 0.121 0.904 Abundance of shrubs -0.038 -0.294 0.769 Percent cover of grass and herbs 0.157 0.104 0.072 Percent of soil substrates -0.157 -0.104 0.072 Elevation /Altitude (m) -0.221 -1.673 0.099 Slope condition (%) 0.134 1.448 0.153

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

Suitability index Suitability 0

Suitability index Suitability 0 0 2 4 6 8 10 12 0 5 10 15 20 25 30 35 Crown diameter of trees (m) Abundance of trees

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 index Suitability 0

Suitability index Suitability 0 0 10 20 30 40 50 60 70 80 90 100 0 2 4 6 8 10 12 Grass and herbs (%) Abundance of shrubs

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2

Suitability index Suitability 0 0.2 Suitability Suitability index 2100 2150 2200 2250 2300 2350 0 Altitude (m) 0 10 20 30 40 50 60 70 80 90 100 Bare Soils (%)

1 0.8 0.6 0.4 0.2

index Suitability 0 0 10 20 30 40 50 Slope condition (%) Figure 3.4. The seven environmental variables and their response curves included in the Mountain Nyala habitat suitability model for Munessa during the wet season.

44

Dry season habitat suitability models The habitat suitability index for Mountain Nyala steadily decreased with an increase in the abundance of trees; significantly increased with an increase in tree crown diameters; steadily increased with an increase in the abundance of shrubs; was constant over low percentages of cover of grass and herbs, but decreased with high percent cover of grass and herbs; slightly increased and then became constant with further increases in the percent cover of soil substrates, but decreased with high percent cover of soil substrates; slightly decreased with an increase in altitude; but slightly decreased with an increase in the percent slope condition over the landscape (Fig. 3.5). Overall, the dry season multiple linear regression models revealed that the seven habitat variables altogether explained 12.2% of the variance for the habitat suitability (Table 3.5).

Table 3.5. A summary of multiple linear regression statistics and constituent variables of the habitat suitability index model for Mountain Nyala in Munessa during the dry season. The incorporated variables are listed in the order in which they were entered into the model. Significant coefficients at 95% confidence level are marked with an asterisk (*).

2 F(6,66) P Multiple R Habitat variables Coefficients t(1, 66) p 1.528 0.183 0.122 Intercept (constant) 0.088 0.108 0.914 Abundance of trees -0.011 -1.077 0.939 Crown diameter of trees (m) 0.334* 5.683 0.020 Abundance of shrubs 0.269 1.234 0.066 Percent cover of grass and herbs -0.073 -0.522 0.603 Percent cover of soil substrates 0.073 0.522 0.603 Elevation /Altitude (m) -0.003 -0.024 0.981 Slope condition (%) -0.146 -1.127 0.264

Overall, the habitat suitability index for Mountain Nyala over the landscape was significantly affected by habitat type (F (2, 150) = 9.49; p < 0.001) and the interaction of habitat type and season (F (2, 150) = 14.53; p < 0.001). However, season alone did not have any significant effect (F (1, 150) = 0.50; p = 0.480) on the habitat suitability for Mountain Nyala in Munessa. During the dry season, the result suggested that the natural forest habitat is the most suitable habitat for Mountain Nyala in Munessa.

45

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

Suitability index Suitability 0 0 0 2 4 6 8 10 12 index Suitability 0 5 10 15 20 25 30 35 Abundance of trees Crown diameter of trees (m)

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 Suitability index Suitability 0

Suitability index Suitability 0 0 2 4 6 8 10 12 0 20 40 60 80 100 Abundance of shrubs Grass and herbs (%)

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

Suitability index Suitability 0 Suitability index Suitability 0 0 10 20 30 40 50 60 70 80 90 100 2100 2150 2200 2250 2300 2350 Bare soils (%) Altitude (m)

1 0.8 0.6 0.4 0.2

Suitability index Suitability 0 0 10 20 30 40 50 Slope condition (%)

Figure 3.5. The seven environmental variables and their response curves included in the Mountain Nyala habitat suitability model for Munessa during the dry season.

Isodars Significant isodars were only obtained from the dry season comparisons, for natural forest versus plantation habitats, and for natural forest versus clear cut habitats (Table 3.6; Fig. 3.6). Morris (1987, 1988, 2003) suggested that if the slope is significantly different from 1, then the two adjacent habitats differ qualitatively, i.e. differ in the intensity of density dependence. In this study, the slope of the isodar between the natural forest and the plantation habitats was significantly different from 1 (Table 3.6) indicating that the natural forest is qualitatively better than the plantation. Furthermore, based on the intercept and the slope of the isodar for natural forest versus plantation habitats (Table 3.6), I calculated and compared the strength of density-dependence in the natural forest versus in the plantation

46

habitat. Thus, the slope of the isodar revealed that the strength of density-dependence in the natural forest habitat is 2.4 lower than in the plantation habitat. Morris (1987, 1988, 2003) suggested that intercepts which significantly differ from 0 reveal quantitative differences between adjacent habitats. However, in the case of the natural forest versus the plantation habitats, the Y-intercept is not significantly different from 0 (Table 3.6), which illustrates that there is no significant quantitative difference between these two habitats. I calculated the CI for the slope of the isodar of natural forest versus plantation habitats (Fig. 3.6a). I found a CI of 0.472. Therefore, the 99% confidence interval is -0.107 to 0.837. That is, I am 99% confident that the true slope of the regression line for the isodar of natural forest versus plantation habitats is in the range defined by 0.365 + 0.472. The slope of the isodar between the natural forest and the clear cut habitats is significantly different from 1, suggesting that the natural forest is qualitatively better than the clear cut habitat (Table 3.6). In addition, based on the intercept and the slope of the isodar for natural forest versus clear cut habitats (Table 3.6), I calculated and determined the strength of density-dependence in the natural forest versus in the clear cut habitat. Accordingly, the slope of the isodar showed that the strength of density-dependence in the natural forest habitat is 2.7 times lower than in the clear cut habitat. However, there is no significant quantitative difference between the natural forest and the clear cut habitats (Table 3.6). I calculated the CI for the slope of the isodar of natural forest versus clear cut habitats (Fig. 3.6b). I found a CI of 0.542. Therefore, the 99% confidence interval is -0.185 to 0.899. That is, I am 99% confident that the true slope of the regression line for the isodar of natural forest versus clear cut habitats is in the range defined by 0.357 + 0.542.

Table 3.6. The results of linear regression for isodars comparing the activity densities of Mountain Nyala in adjacent habitat types during the dry season

Overall p Isodars Intercept P value Slope P r2 F value N value ratio Natural forest  Plantation 0.015 0.754 0.365 0.04 0.906 23.88 < 0.001 4 Natural forest  Clear cut 0.129 0.627 0.357 0.03 0.903 28.92 < 0.001 3

47

a.

0.15

0.10

0.05

in Density the plantation habitat

0.00 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 Density in the natural forest habitat b. 0.5

0.4

0.3

0.2

0.1

Density in theDensity cut clear habitat

0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Density in the natural forest habitat

Figure 3.6. Isodars for the dry season between activity densities of Mountain Nyala in (a) natural forest versus plantation habitats, (b) natural forest versus clear cut habitats.

48

DISCUSSION Activity Density and Distribution of Mountain Nyala Habitat choice affects the abundance and distribution of a species (e.g., Fretwell and Lucas, 1969; Rosenzweig, 1974, 1981; Morris, 1988). Mountain Nyala require open sightlines to detect approaching predators including humans (Evangelista et al., 2007), so open habitats provide Mountain Nyala with good opportunity to detect and escape from risk of predation. In the wet season, the result revealed that the clear cut habitat is a suitable feeding ground for Mountain Nyala through providing highly nutritive and digestible grass and herbs (e.g., Evangelista et al., 2007). However, due to the selective foraging behavior of Mountain Nyala, natural foraging grounds which are overgrazed / browsed by livestock were not grazed / browsed by Mountain Nyala. Previous studies also noted that Mountain Nyala avoid habitat patches which are overgrazed by livestock in the Bale Mountains (e.g., Malcolm and Evangelista, 2005; Mamo, 2007; Atickem et al., 2011). When Mountain Nyala caught sight of observers in the clear cut habitat, they typically moved straight away into the adjacent plantation and natural forests. This suggests that the clear cut habitat is risky, but both plantation and natural forest serve as escape refuges (Mech, 1977, Ripple and Beschta, 2003, 2004) for Mountain Nyala from risk of predation and human nuisance. The plantation habitat which is relatively rich in palatable undergrowth plant species could serve as a foraging ground for Mountain Nyala during the wet season. In addition, the plantation habitat serves as a corridor and shelter from extreme weather for Mountain Nyala (Evangelista et al., 2007). The seasonal streams flowing across the plantation and the natural forest habitats serve as source of water for Mountain Nyala during the rainy season. In the wet season, a group of Mountain Nyala is commonly seen around the salt licks in Munessa when people and livestock are not around. This suggests that salt is important nutrient for Mountain Nyala that influences their habitat selection and distribution. Previous studies also noted that habitat selection and distribution of Mountain Nyala in the Bale Mountains National Park is influenced by the availability of salt licks (e.g., Mamo, 2007; Ethiopian Wildlife Conservation Authority, unpublished field reports). The result revealed that the diurnal activity density of Mountain Nyala was biased to the natural forest in the dry season. This may be because the natural forest provides the

49

Mountain Nyala with diversified food (e.g., Evangelista et al., 2007), escape refuges from risk of predation (e.g., Mech, 1977, Ripple and Beschta, 2003, 2004), and critical shelter from environmental extremes especially during the dry season (e.g., Black et al., 1976; Evangelista et al., 2007). However, during the extended dry season, following the high livestock grazing pressures, palatable forages are greatly depleted in all habitat types, but most seriously in the clear cut and in the plantation habitats (Appendix I - Table 3.1). Even the undergrowth status in the natural forest is poor during the extended dry season compared with the wet season (Appendix I - Table 3.1). This inevitably results in a shortage of palatable forage for Mountain Nyala. Mountain Nyala were observed to adjust their habitat selection and distribution even on a daily basis in Munessa. For example, during the extended dry season, the presence of foot prints of Mountain Nyala in the plantation and clear cut habitats suggest that the Mountain Nyala become mostly active during the night time to search for forage in Munessa. The result of the spotlight censusing carried out during the night time supports this scenario. Unlike the activity density in the early morning hours (i.e., 06:00 – 09:00 local time), the night time activity density of Mountain Nyala during the dry season was biased to the clear cut habitat. This shows that Mountain Nyala alter their habitat selection even on a daily basis. The possible explanation for this is that the Mountain Nyala can more easily detect approaching potential predators and escape in the open habitat than in the dense vegetation during the night time. Evangelista et al. (2007) also note that Mountain Nyala are good runners to escape from their potential predators. In some occasions during my spotlight censusing, I encountered Leopard both in the plantation and the natural forest. Leopards are one of the potential predators of Mountain Nyala in the study area (Evangelista et al., 2007). This could be why Mountain Nyala select the clear cut habitat during the night time as Leopards are actively hunting then. Furthermore, humans and their livestock are absent at night, and Mountain Nyala can access the forage in the open more freely. As a result, the percentage of Mountain Nyala sightings during the transect walk was relatively higher for the spotlight census than for the daylight hours census in the dry season (see also Table 3.2 & Table 3.3).

50

Habitat Suitability Models Habitat suitability models are important tools for evaluating habitat quality based on critical environmental factors and relative activity density of animals (Schamberger and O’Neil, 1986; Morrison et al., 1998; Tadesse and Kotler, 2010). To maximize the usefulness of habitat suitability models, it is important that they should be constructed using habitat variables chosen for a priori reasons (Johnson, 1980; Henley, 2001), and selected using a method that consistently selects models of an appropriate level of complexity (Burnham and Anderson, 2002). Generally, there are two common approaches to develop habitat suitability modeling: logistic regression and multiple linear regression (Block et al., 1998). Logistic regression analysis is applied in many cases when only presence-absence data are collected (e.g., Store and Jokimäki, 2003; Evangelista et al., 2008; Atickem et al., 2011), whereas multiple linear regression analysis is normally applied when activity density data are available (e.g., Jokimäki and Huhta, 1996) as in the present study. The habitat suitability models developed in the present study incorporated seven habitat variables considered critical for Mountain Nyala. None of the habitat variables significantly correlated with the suitability index in the wet season (Table 3.4). In contrast, during the dry season, the HSI revealed that only crown diameter of trees had a significant beta coefficient in the habitat suitability model (Table 3.5), suggesting that tree crowns provide important shade and cover to Mountain Nyala from hot weather in the dry season (Evangelista et al., 2007). Overall, the habitat suitability models revealed that the suitable habitat for Mountain Nyala in Munessa did not significantly differ during the wet season. However, during the dry season, the results showed that the suitable habitat for Mountain Nyala significantly varied with habitat type. Accordingly, the habitat suitability model suggested that the natural forest is the most suitable habitat for Mountain Nyala. This is inline with the hypothesis of the present study. For example, from the habitat suitability models, the component of the dry season habitat suitability for Mountain Nyala that importantly reflects those significant differences between the aforementioned habitat types is the mean crown diameter of trees (see also Appendix - Table 3.1), which is the highest for the natural forest habitat. This implies that the natural forest provides the Mountain Nyala with

51

valuable cover for thermal regulation from hot weather during the day times in the dry season (e.g., Black et al., 1976; Evangelista et al., 2007). In addition, the natural forest habitat provides the Mountain Nyala with safe refuge from risk of predation (e.g., Mech, 1977; Ripple and Beschta, 2003, 2004), and humans and livestock disturbances (e.g., Evangelista et al., 2007, 2008). Availability of palatable forages was much better in the natural forest than either in the clear cut or in the plantation habitat during the dry season. The present study revealed that the abundance and diversity of shrubs was the highest in the natural forest habitat (see also Appendix I - Table 3.1, 3.2). This suggests that shrubs may serve as good sources of browse species for Mountain Nyala especially in the dry season when grass and herbs become less abundant and low in their nutritional quality in Munessa. Previous studies also noted that Mountain Nyala rely on shrubs as their typical sources of food in the Bale Mountains National Park during the dry season when grass and herbs are less abundant and indigestible (e.g., Brown, 1969; Woldegebriel, 1996; Hillman and Hillman, 1987; Refera and Bekele, 2004; Mamo, 2007; Chapter 5). Other authors suggested that food appears to be critically the most limiting factor determining the distribution, abundance, and habitat selection behaviors of herbivores during the dry season (e.g., Muller et al., 1995; Mengesha and Bekele, 2008; Tadesse and Kotler, 2010). In general, the present habitat suitability model could be employed to monitor and evaluate the habitat resource availability and, later, to determine and monitor the carrying capacity (Morris, 2003; Tadesse and Kotler, 2010) of the habitats for Mountain Nyala population.

Isodars In the present study, I analyzed isodars by regressing the activity densities of Mountain Nyala in adjacent habitat types seasonally. Significant isodars were only obtained for the dry season comparisons as shown in the result. Following the procedures outlined by Morris (1987, 1988, 1995, 2003), the regression analyses of the dry season isodars revealed that the natural forest habitat was qualitatively, but not quantitatively, more suitable than both the plantation and the clear cut habitats. Accordingly, the slopes of the isodars suggest that the strength of density-dependence in the natural forest habitat is 2.4 and 2.7 times lower than in the plantation or the clear cut habitats, respectively.

52

Qualitative differences often arise from differences in risk of predation (Morris, 2003), suggesting that the natural forest is safer than the other habitats, especially as population density rises. This is because the natural forest provides the Mountain Nyala with critical cover for concealment (e.g., Black et al. 1976, Evangelista et al. 2007) and escape refuge from risk of predation (e.g., Mech 1977, Ripple and Beschta 2003, 2004). In addition, resource identity and habitat structure may result in qualitative differences between habitats (Rosenzweig, 1987; Morris, 1988), which could be supported by the availability of more diversified browse species in the natural forest than either in the plantation or the clear cut habitat (see also Appendix I – Table 3.2). Species differ in the number of habitats that they occupy, and the differences in activity density also reflect preferential occupation of habitats that yield the greatest fitness

(Morris, 2005; Morris and Mukherjee, 2007). The present isodars suggest that the fitness of Mountain Nyala in the natural forest habitat is higher than either in the plantation or in the clear cut habitat during the dry season. Thus, future management plans for Mountain Nyala should give due emphasis toward the conservation and protection of the remnant natural forests in Munessa. Isodars can be further linked with theories of optimal foraging and patch use (e.g., Rosenzweig, 1987; Morris and Kingston, 2002; Morris, 2003; Druce, 2005; Morris, 2005; Morris and Mukherjee, 2007) to assess habitat characteristics and underlying mechanisms of habitat selection behavior in Mountain Nyala. For example, Morris (2005) applied isodars and patch use techniques through measurement of giving-up densities to assess differences in predation risk between habitats and density-dependent habitat selection behavior in Snowshoe hare. Optimal foraging theory may also provide a unique avenue with which to explore the adaptive behaviors of Mountain Nyala related to resource densities over different habitats (Morris and Mukherjee, 2007). Isodars can be applied to measure scales of habitat selection and to identify the number of habitats (Morris, 2003; Morris, 2005) that the Mountain Nyala recognize across the heterogeneous landscape. Moreover, isodars can be used to detect human occupation and modification of habitats as an indicator of future threats to the population of Mountain Nyala. For example, theoretical and empirical studies with isodar analyses suggest that the proportions of threatened taxa (e.g., and birds) are positively correlated with the human isodar (i.e., isodar that predicts the human habitat selection behavior based on urban versus rural habitats) where threats to biodiversity belong in the 'high threat' group (Morris and Kingston, 2002; Morris, 2003).

53

To summarize, an understanding of the adaptive habitat selection behaviors of Mountain Nyala enhances conservation and management activities. Habitat suitability models and isodar analyses were applied to evaluate the seasonal habitat quality and habitat selection behaviors of Mountain Nyala. The habitat suitability models, the isodar analyses, and the spotlight censusing helped to understand how the habitat selection behaviors of Mountain Nyala vary with the seasonal availability of habitat resources (e.g., food, cover), predation risk characteristics of the habitats, and impacts of humans and livestock disturbances. For example, the habitat suitability model showed that the natural forest is the most suitable habitat for Mountain Nyala in the dry season during the day time. Isodar analyses also complemented the habitat suitability model through revealing information on how the natural forest is qualitative better than either the plantation or the clear cut habitat for Mountain Nyala. However, the spotlight censusing revealed that Mountain Nyala selected the clear cut habitat during the night time when they can more easily detect approaching potential predators (e.g., Leopards) and escape in the open habitat than in the dense vegetation, or when people and livestock are absent in the area so that the Mountain Nyala can access the forage in the open more freely. The field information obtained through habitat sampling may help in designing future monitoring programs through tracking the population of Mountain Nyala in the Munessa Forest. The isodar analyses also support the habitat suitability models by increasing our understanding on the qualitative and quantitative differences in habitat selection and patch use behaviors in Mountain Nyala.

54

CHAPTER 4 HABITAT USE BY MOUNTAIN NYALA EVALUATED WITH BEHAVIORAL INDICATORS IN MUNESSA ABSTRACT Little is known how the different behaviors of Mountain Nyala are related to and influenced by environmental conditions, social factors (e.g., group size), and individual characteristics (e.g., sex-age classes). I studied the habitat selection and patch use behavior Mountain Nyala with behavioral indicators using time budgets of focal individuals in Munessa, Ethiopia. I hypothesized that the behavioral responses of Mountain Nyala vary with habitat type, season, group size, and sex-age categories. I predicted that adult female Mountain Nyala more strongly select habitats with fewer risks of predation and human nuisance disturbances than do adult males. I also predicted that Mountain Nyala spend more time on vigilance or moving behaviors in habitats with higher risk of predation and human nuisance disturbances. I applied the group scan method. The focal animal observations were carried out along transects aligned through three major habitat types. Crucial environmental variables were collected in plots laid along transects in each habitat type. The fieldwork was carried out in the wet and the dry seasons. The results showed that several environmental variables influenced the habitat use of Mountain Nyala. In the wet season, Mountain Nyala devoted much of their time being vigilant, but vigilance did not differ across habitat types and sex-age classes. In the dry season, vigilance differed across habitats and was the highest in the clear cut habitat. Contrary to expectations, adult males were the most vigilant compared with the other sex-age classes in the dry season. Adult male biased legal and illegal hunting may be the cause of high male vigilance. However, group size of Mountain Nyala was significantly greatest in the natural forest habitat during the dry season. To conserve and manage the Mountain Nyala, wildlife managers should promote novel behaviors that enhance survivorship in response to novel hazards. The study thus improves our understanding on the adaptive habitat selection behaviors of Mountain Nyala. This is a basis for developing novel solutions to conserve and manage the endangered Mountain Nyala and its habitats in Ethiopia. Key words: behavioral indicators, behavioral models, environmental variables, focal individuals, group size, time budgets, vigilance

55

INTRODUCTION An understanding of animal time budgets is crucial to behavioral ecology. For example, behavioral modeling based on time budgets of focal individuals help understand how animals perceive their environment (Stephens and Krebs, 1986; Manor and Saltz, 2003; Hochman and Kotler, 2006b), and how they trade-off food acquisition and safety (Stephens and Krebs, 1986; Lima and Dill, 1990; Brown et al., 2001; Brown and Kotler, 2004). Behavioral models allow inferences about the quality of the environment where an organism lives using “behavioral indices of habitat quality” (Siegfried, 1980; Druce, 2005). This can be accomplished in practice by correlating behavioral time budgets with influential environmental variables (e.g., Tadesse and Kotler, 2011). Activity patterns in animals are influenced by and changed in response to a number of factors, including disturbances caused by predators (Brown, 1999) and humans (Kitchen et al., 1999; Tadesse and Kotler, 2011). Activity time budget patterns of animals are generally regarded at least in part as the outcome of two conflicting demands: the activity required to maximize nutritional, social, and reproductive objectives, and the need to minimize the costs and risks imposed by the environment (Rosenzweig, 1974; Sih, 1980). Risk of predation, escape terrain, vigilance, disturbance, and habitat characteristics are some of the main factors that help determine the behaviors and time budget of foragers. Risk of predation has an overwhelming effect on the habitat selection, habitat use, and behaviors of prey species (e.g., Lima and Dill, 1990; Brown, 1999). Spatial variation in predation risk produces a landscape of fear for the prey species (Brown et al., 2001; Laundre et al., 2001; Shrader et al., 2008a). For example, mountain ungulates prefer steep terrain to reduce risk of predation (Kotler et al., 1994; Muller et al., 1995; Hochman and Kotler, 2006b). The ability of an animal to escape from predators across local terrain is important in determining habitat selection and patch use behavior of a species. Escape substrate surrounding a resource patch affects the ability of a forager to flee approaching predators or reach a refuge (Kotler et al., 2001). Escape terrain and vegetation cover can be important prey refugia where the potential to encounter predators greatly decreases or the ability to escape encounters greatly increases. For example, in a wolf-ungulate system, ungulates may seek refuge by migrating to areas outside the core territories of wolves

56

where their survival is higher (Mech, 1977; Ripple and Beschta, 2003, 2004). Similarly, Kotler et al. (1994) and Hochman and Kotler (2006b) noted that Nubian Ibex ( nubiana) minimize their predation risk by foraging most often near escape terrain of extremely steep slopes or cliffs or rugged terrains that are difficult or impossible for predators to access. Animals often use vigilance to reduce risk of predation while using their habitats (e.g., Brown, 1999; Manor and Saltz, 2003; Hochman and Kotler; 2006b). A reduction in individual vigilance with increasing group size is frequently reported in the field of animal behavior. Many social wild animal species show a reduction in individual vigilance with increased group size due to the “detection” or “dilution” effects (e.g., Pulliam, 1973; Siegfried, 1980; Lagory, 1986; Elgar, 1989; Lima and Dill, 1990; Quenette, 1990; Lima, 1995; Roberts, 1996; Hunter and Skinner, 1998; Brown, 1999; Treves, 2000; Manor and Saltz, 2003). Decreased vigilance in larger groups may be due to the fact that the more highly vigilant individuals that occur on the edge of the group comprise a smaller proportion of the group as its size increases (Inglis and Lazarus, 1981; Brown, 1999; Beauchamp, 2003). However, to remain in the group, individuals must synchronize their activities with other members of the group (e.g., Gross et al., 1995; Muller et al., 1995; Kohlmann et al., 1996; Conradt and Roper, 2000). Disturbances affect the habitat selection, distribution, behaviors, and physiology of wild animals. Disturbance amplifies the problems caused by habitat fragmentation because it decreases the available habitat (Moenting and Morris, 2006; Atickem et al., 2011). However, disturbance seems to be more intense when activities are dispersed within the habitats, or not practiced on a predictable basis (Burger, 1981). Consequently, disturbances may force animals to devote more time to safety related behaviors, such as increased levels of vigilance that come at the expense of foraging, nesting, and resting activities (Gill et al., 1996; Third World Conference on Mountain Ungulates, 2002; Manor and Saltz, 2003). For example, Done White et al. (1999) have quantified the effect of disturbances on Grizzly Bears (Ursus arctos horribilis) by alpinists in Montana, USA. This led to a 53% reduction in feeding time by bears, a 52% increase in movements, and a 23% increase in aggressive behaviors. Human disturbances can also affect reproductive success by altering pairing displays or increasing the amount of parental care given to the young (e.g., Verhulst et al., 2001; Lindsay et al., 2007).

57

Human disturbance is analogous to predation risk in that it can cause habitat shifts and other behavioral responses that later influence the fitness of individuals (Done White et al., 1999; Frid and Dill, 2002; Moenting and Morris, 2006). Depending on its frequency and severity, human disturbance can result in short-term or long-lasting effects on wildlife species. Short-term effects are easier to observe and quantify, but the long-lasting effects have strong impacts on fitness (Bell and Owen, 1990). So, a strategy to better manage the effects of disturbance on wildlife is to understand the proximate mechanisms underlying the response of animals to humans. Habitat characteristics affect the trade-offs between food acquisition and safety (e.g., Lima and Dill, 1990; Gill et al., 1996; Gill and Sutherland, 2000; Brown et al., 2001; Brown and Kotler, 2004; Druce, 2005; Beale, 2007) in Mountain Nyala. Mountain Nyala require two kinds of habitat (Malcolm and Evangelista, 2005; Mamo, 2007). Mountain Nyala use a foraging habitat where they acquire sufficient food both in quantity and quality (e.g. Brown, 1969b; Evangelista et al., 2007; Mamo, 2007). Mountain Nyala also need a habitat which provides them with cover in which to hide from predators, disturbances, and shelter from extreme weather (Refera and Bekele, 2004; Malcolm and Evangelista, 2005). For example, Mountain Nyala use dense vegetation for camouflage, and higher grounds and open sightlines to detect risk of predation and quickly escape (Evangelista et al., 2007). The gregarious nature and social groupings of Mountain Nyala are believed to play a vital role in detecting potential predators by increasing the number of individuals that are attentive to risk of predation (e.g., Brown 1969b; Evangelista et al. 2007). Previous studies demonstrated that the behaviors of wild animals are affected by habitat type (e.g., Druce, 2005; Reid, 2005; Hochman and Kotler, 2006; Tadesse and Kotler, 2011), season (e.g., Shettleworth et al., 1995; Provenza, 1997; Tadesse and Kotler, 2011), group size (e.g., Pulliam, 1973; Siegfried, 1980; Lagory, 1986; Elgar, 1989; Lima and Dill, 1990; Quenette, 1990; Lima, 1995; Roberts, 1996; Hunter and Skinner, 1998; Brown, 1999; Treves, 2000; Manor and Saltz, 2003), and sex-age classes (e.g., Laundre et al., 2001; Ruckstuhl and Kokko, 2002; Reid, 2005; Rieucau, and Martin, 2008; Tadesse and Kotler, 2011). I hypothesized that the behavioral responses of Mountain Nyala vary with habitat type, season, group size, and sex-age categories. Using time budget patterns of free-ranging focal Mountain Nyala, the variations could be explained by behavioral models

58

of activity time budget hypothesis. I predicted that adult female Mountain Nyala more strongly select habitats with fewer risks of predation and human nuisance disturbances than do adult males. This should be true especially for females with young because the need to bear and raise young may increase the predation cost to females and thereby limit their use of space (e.g., Laundre et al., 2001; Ruckstuhl and Kokko, 2002; Rieucau, and Martin, 2008; Tadesse and Kotler, 2011). In addition, unlike adult males, females are physically less capable of escaping and/or defending themselves from predators attack so that they are vulnerable to risk of predation. I also predicted that Mountain Nyala spend more time on vigilance or moving behaviors in habitats with higher risk of predation and human nuisance disturbances. However, Mountain Nyala should spend more time on feeding or resting behaviors in habitats with low risk of predation and human nuisance disturbances. I predicted that the vigilance level of Mountain Nyala reduces with the increase in group size due to the “detection” or “dilution” effects. In this study, I investigated the habitat use and activity time budgets of free-ranging Mountain Nyala during the wet and the dry seasons in Munessa. The objectives of this study were as follows. (1) Analyze the time budgets allocated by free-ranging Mountain Nyala in different activities (i.e., proportions of time vigilant, feeding, moving, and resting) in the wet versus dry season. (2) Explore the environmental conditions that affect the behaviors of Mountain Nyala. To do so, I developed behavioral models based on time budgets of focal Mountain Nyala as functions of environmental variables, group size, sex- age categories, and season. (3) Analyze the effect of group size and sex-age categories on the vigilance level of free-ranging Mountain Nyala as functions of habitat types and seasons. (4) Determine whether group size of Mountain Nyala is a function of habitat types and season.

METHODS Focal Animal Observations I applied the group scan sampling method (Altmann, 1974; Martin and Bateson; 1993) through focal animal observation to quantify how the activity time budget patterns of free-ranging Mountain Nyala are related to habitat types, environmental variables, seasons, group size, and sex-age categories. I applied focal observations to animals

59

encountered along 12 permanent transects i.e. four transects in each habitat type which were used to estimate the activity densities of free-ranging Mountain Nyala in Munessa (see also Chapter 3). Following Altman (1974) and Martin and Bateson (1993), I classified the behavioral activities of each focal Mountain Nyala into four types: vigilance, feeding, moving, and resting. The activity types were defined as follows. Vigilance referred when the animal had its head up, neck erect, ears oriented forward and eyes wide open for watching, and appeared alert. Feeding was defined as the entire process of food searching, locating, biting, and ingestion as characterized by slight movements with the head in the down position or foraging with head in a bush and/or shrub. Moving was defined as when the animal showed steady movement with the head held horizontally or running. Resting referred to when the animal was lying down or sleeping, standing under the shade, standing still with head held horizontally and ears drooping, but not appearing alert. I took field observations on focal individual Mountain Nyala encountered while walking along each transect in each habitat. In order to minimize any effect caused by the presence of the observer, I waited for the groups of Mountain Nyala to calm down for a minimum of 10 minutes before starting the focal observations. The total number of individuals in a group was recorded. Individuals were considered to be in the same group if the separation distance was approximately less than 50 meters (Hillman and Hillman, 1987; Refera and Bekele, 2004). If the focal animal was with a group of Mountain Nyala then sex-age classes of its group (numbers of adult males, adult females, sub-adults, and juveniles) were recorded. Due to the high humans and livestock induced disturbances in the study area including illegal and legal hunting, the sensitive individuals or groups of Mountain Nyala usually run when they encountered with the observers during the transect walks. However, I tried my best to use sight blinds not to be seen by those individuals or groups of Mountain Nyala. In addition, I was using binoculars so that I did not need to closely approach the Mountain Nyala during the transect walks. I identified the sex and age classes of the focal individuals as follows. Adult males are easily distinguished by their long twisted horns, large body size, and dark brown / grey color. Sub-adult males are identified by their straight spike horns (but not twisted), medium body size, and brownish color. Adult females are differentiated by their absent horns, red brown appearance, and smaller body size than adult males. Sub-adult females are differentiated from adult females by their smaller body size and reddish color.

60

However, previous studies revealed that sub-adult males and sub-adult females more or less exhibit similar activity behaviors (e.g., Brown 1969b; Refera and Bekele, 2004; Evangelista et al., 2007, Mamo, 2007). Thus, for simplicity purpose, I merged sub-adult males and sub-adult females together for this study. Juveniles are defined by their small body size and absence of horns. Following Brown (1969a) and Refera and Bekele (2004), I divided focal observations according to sex and age classes of the focal individual. If the group contained a mix of ages and genders, stratified random sampling was employed to select focal individuals. The stratification was based on age and sex categories while random sampling was achieved through use of random numbers in selection of the focal animal. Then I took focal observations of animals representing each category of sex and age. I carried out focal observations for 10 minutes for each focal animal with the help of binoculars. For each focal observation, I noted the type of activity in which the focal Mountain Nyala was engaged at the start of the observation period and recorded the length of time spent in different activities as mentioned above. Focal animal observations were carried out early in the morning from 6:00 am to 9:00 am local time (ideal times to sample group size and activity time budgets) when Mountain Nyala were active. In addition, humans and livestock were relatively absent at this time in Munessa so that Mountain Nyala safely used their habitats. Focal observations were conducted at a mean sighting distance of approximately 50 meters. I recorded the habitat type, transect number, date, season, and time for each focal observation. My main focus here was on feeding and vigilance because these behavioral activities best reflect aspects of food and safety. However, I noted that my behavioral model did not strictly model resting behavior per second because Mountain Nyala mostly spend time resting during the mid-day (Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). That was not my aim in this study. But I included time resting during the most important hours of Mountain Nyala’s activity. I took care to sample Mountain Nyala on different days, in different locations throughout their range, and in different habitats. This necessarily reduced the chances of sampling the same individual twice. In addition, I took care to collect focal observations from only a single individual of each age and sex class from a group each day and to collect data in different locations each day. These spread observations out across sex and age classes, habitats, time, and location as much as was possible for these populations. The

61

interspersion of data collection over days, locations, sex-age classes, and habitats increased my confidence in the independence of observations and helped minimize artifacts of interpretation. Using the time for the focal animal observation and the time spent in each activity, I determined the proportion of time that Mountain Nyala spent in each activity. Mountain Nyala population may exhibit seasonal fluctuations in the patterns of their time budgets according to dry versus wet seasons. I carried out the field study from May through August, 2010 for the wet season data. I observed a total of 119 focal individual Mountain Nyala in Munessa. For the dry season data, I took a total of 116 focal individual Mountain Nyala from December 2010 through April 2011 in Munessa. I analyzed seasons separately. The behavioral responses of wild animals may be influenced by habitat types and/or environmental variables (Senft et al., 1985; Henley, 2001). It is important that the behavioral models should be constructed using habitat variables chosen a priori (Johnson, 1980), and selected using a method that consistently selects models of an appropriate level of complexity (Burnham and Anderson, 2002). I took the environmental variables data from the habitat suitability models (see also Chapter 3) that could have greatest influence on the behavioral responses of Mountain Nyala in Munessa. Thus even though I was not able to collect environmental data at each exact location for each focal observation, I included representative data for the habitat type and transect from which each focal observation was taken. I developed behavioral models based on the consideration for the behavioral dynamics (patterns of activity time budgets of focal animals) of Mountain Nyala in accordance with the habitat variables, group size, sex-age categories, and seasons.

Data Analysis Activity time budget analysis: To analyze the overall activity time budget patterns of free-ranging Mountain Nyala, I pooled together data for time allocation of each activity type from focal animal observations for all the four age-sex categories, i.e. adult males, adult females, sub-adults, and juveniles. I square root transformed the data to ensure normality. Activity type (i.e., feeding, vigilance, moving, and resting) was a predictor, a proportion of time for each activity type was a dependent variable. I analyzed the wet and

62

the dry season data separately using ANOVA. I used a post-hoc Tukey's HSD test for the multiple comparisons across activity types in each season. Multivariate Analysis of Variance (MANOVA): the purpose of MANOVA is to explore how the independent variables influence the patterning of responses on the dependent variables (i.e., time budgets of focal animals) (Carey, 1998). The MANOVA approach helps analyze the effects of categorical and continuous independent variables, and the interactions between categorical independent variables on the dependent variables at the same time in the same analysis (Scheiner, 2001). Then, univariate follows MANOVA with univariate ANOVAs on each variable and allows for better interpretation of the multivariate results (Carey, 1998; Scheiner, 2001). Once the habitat type at the location of the focal animal was determined from the field survey, I included the data on the environmental variables from the habitat suitability models. I first ran tests to check for collinearity between potential explanatory environmental variables; however, I did not find collinearity between any of the explanatory environmental variables. Since the percent cover of bare soils and percent cover of herbs and grass in each plot are not independent (i.e., they share a degree of freedom), this might have consequences for the MANOVA analysis. Herbs and grass are valuable sources of food for Mountain Nyala (e.g., Evangelista et al., 2007; Mamo, 2007), thus I excluded the percent cover of bare soils from the MANOVA analysis. So, the independent variables entered into the MANOVA analyses included the following: abundance of trees, crown diameter of trees, abundance of shrubs, percent cover of herbs and grass, altitude, percent of slope condition, group size, sex-age categories, and habitat type. Then I entered the data collected from individual focal animal observations (i.e., proportions of time vigilant, feeding, and moving) after square root transforming them to help ensure normality. Since time is limited, proportions of time vigilant, feeding, moving, and resting are not independent (i.e., they share a degree of freedom). This might have consequences for the MANOVA analysis. Thus, I did not include the proportion of time resting in the MANOVA analyses (see also the MANOVA results presented in Table 1). Linear regression: I used linear regression to generate the response curves that defined the relationship between each independent variable (i.e., abundance of trees, crown diameter of trees, abundance of shrubs, percent cover of herbs and grass, percent cover of

63

bare soils, altitude, percent of slope condition, group size, and sex age category) on each dependent variable (i.e., proportion of time vigilant, feeding, moving, and resting by focal Mountain Nyala). For the categorical independent variable (i.e., sex-age category of the focal animal), I used a dummy variable to correlate the categorical independent variable with each dependent variable in the linear regression by which I was able to determine the correlation coefficients. A dummy variable can be defined as a qualitative representative variable incorporated into a regression, such that it assumes the value 1 whenever the category it represents occurs, but 0 when it is absent. Dummy variables are used as devices to sort data into mutually exclusive categories (such as adult males/adult females, sub- adults/juveniles) in the present study (Gujarati, 2003). I produced separate behavioral models for the wet and the dry seasons as shown in Table 4.2. Finally, I used ANOVA to analyze the variations of the group size in Mountain Nyala across habitat type and season. When there was significant difference in group size across habitat types, I used a post-hoc Tukey's HSD test for the multiple comparisons. For all analyses, I defined the alpha value of 0.05. I performed the data analyses with STATSTICA version 10.

RESULTS To quantify the time budgets of free-ranging Mountain Nyala in Munessa, I assessed 119 and 116 focal individuals for wet and dry seasons respectively. In the wet season, Mountain Nyala allocated significantly more time to some activities than to others

(F (3,472) = 81.82; p<0.001). Mountain Nyala spent most of their time vigilant (mean ≈ 42.34%; stdev ≈ 19.99%), followed by time moving (mean ≈ 30.12%; stdev ≈ 16.72%), time feeding (mean ≈ 23.32%; stdev ≈ 25.14%), and then time resting (mean ≈ 4.22%; stdev ≈ 12.97%) (Fig. 4.1). A post-hoc Tukey's HSD test for multiple comparisons across activity types showed a significant difference in time budgets between vigilance versus resting, feeding versus resting, and moving versus resting (Fig. 4.1). Similarly, during the dry season, Mountain Nyala devoted more time to some activities than to others (F (3,460) = 388.72; p<0.001). Mountain Nyala spent most of their time moving (mean ≈ 47.39%; stdev ≈ 12.40%), followed by time vigilance (mean ≈ 37.10%; stdev ≈ 16.34%), time feeding (mean ≈ 13.13%; stdev ≈ 8.55%), and time resting (mean ≈ 2.39%; stdev ≈ 4.85%) (Fig.

64

4.1). A post-hoc Tukey's HSD test for multiple comparisons across activity types showed a significant difference in time budgets between vigilance versus feeding, vigilance versus resting, moving versus feeding, moving versus resting, and feeding versus resting, (Fig. 4.1).

70.00

60.00

50.00 Vigilance 40.00 Feeding 30.00 Resting Moving 20.00

10.00

Time spent on activities (%) activities on spent Time 0.00 Wet Dry Season

Figure 4.1. General activity time budget patterns of Mountain Nyala during the wet and the dry seasons in Munessa pooled over four categories, adult males, adult females, sub- adults, and juveniles. The number of focal individuals (N) included in this analysis was 119 and 116 for the wet and the dry seasons respectively. The error bars represent +1 stdev. In the wet season, abundance of trees affected most behaviors of Mountain Nyala, including proportion of time vigilant (Table 4.1; Fig. 4.5) and feeding (Table 4.1; Fig. 4.6). However, abundance of trees did not affect any behaviors of Mountain Nyala in the dry season (Table 4.1). In the wet season, crown diameter of trees did not affect any behaviors of Mountain Nyala (Table 4.1). However, during the dry season, crown diameter of trees significantly affected some behaviors of Mountain Nyala, including proportion of time feeding (Table 4.1; Fig. 4.10). In the wet season, abundance of shrubs did not affect any behaviors of Mountain Nyala (Table 4.1). However, during the dry season, abundance of shrubs significantly affected most behaviors of Mountain Nyala, including proportion of time vigilant (Table 4.1; Fig. 4.9) and feeding (Table 4.1; Fig. 4.10).

65

Table 4.1. The seasonal effects of habitat variables, group size, habitat type, sex-age categories, and the interaction of habitat type and sex-age categories on the activity time budgets of Mountain Nyala in Munessa. The different statistical values are included. Variables with significant effect at 0.05 alpha values are marked with an asterisk (*).

Season Factors df MANOVA Vigilance Feeding Moving Abundance of trees 1 Pillai’s Trace value = 0.115; F = 4.00* F = 5.59* F = 0.39 F (3, 83) = 2.74; p = 0.051 Crown diameter of 1 Pillai’s Trace value = 0.012; F = 0.12 F = 0.78 F = 0.02 tree F (3, 83) = 0.26; p = 0.857 Abundance of shrubs 1 Pillai’s Trace value = 0.067; F = 1.58 F = 0.15 F = 1.77 F = 1.50; p = 0.223 (3, 83) Grass and herbs 1 Pillai’s Trace value = 1.039; F = 3.01* F = 4.11* F = 1.61 F = 3.871; p = 0.041 Wet (3, 83) Altitude 1 Pillai’s Trace value = 0.232; F = 0.07 F = 0.08 F = 0.94 F (3, 83) = 0.34; p = 0.744 Slope condition 1 Pillai’s Trace value = 0.096; F = 4.05* F = 4.95* F = 0.34 F (3, 83) = 2.22; p = 0.044 Group size 1 Pillai’s Trace value = 0.17; F = 0.12 F = 1.70 F = 1.78 F (3, 83) = 0.21; p = 0.069 F = 15.02*** Habitat type 2 Pillai’s Trace value = 0.269; F = 0.23 F = 7.99** F (3, 83) = 7.76; p = 0.002 Sex-age categories 3 Pillai’s Trace value = 0.027; F = 0.73 F = 0.22 F = 0.08 F (6, 126) = 0.29; p = 0.939 Habitat type X Sex- 6 Pillai’s Trace value = 0.073; F = 0.19 F = 0.77 F = 0.12 age categories F (15, 174) = 0.32; p = 0.993 Abundance of trees 1 Pillai’s Trace value = 0.009; F = 0.01 F = 0.39 F = 0.19 F (3, 71) = 0.16; p = 0.926 Crown diameter of 1 Pillai’s Trace value = 2.021; F = 0.01 F = 2.71* F = 0.23 trees F (3, 71) = 3.36; p = 0.042 Abundance of shrubs 1 Pillai’s Trace value = 2.034; F = 2.81* F = 2.01* F = 0.01 F (3, 71) = 3.59; p = 0.019 Grass and herbs 1 Pillai’s Trace value = 2.006; F = 2.09* F = 0.24 F = 0.01 Dry F (3, 71) = 3.11; p = 0.054 Altitude 1 Pillai’s Trace value = 0.008; F = 0.01 F = 0.28 F = 0.19 F (3, 71) = 0.142; p = 0.934 Slope condition 1 Pillai’s Trace value = 0.134; F = 0.41 F = 0.33 F = 1.82 F (3, 71) = 1.63; p = 0.060 Group size 1 Pillai’s Trace value = 2.044; F = 1.34 F = 6.94** F = 2.01* F (3, 71) = 3.79; p = 0.002 F = 31.33*** Habitat type 2 Pillai’s Trace value = 0.398; F = 0.63 F = 5.93* F (3, 71) = 11.22; p = 0.001 Sex-age categories 3 Pillai’s Trace value = 0.153; F = 4.68* F = 3.19* F = 1.03 F (6, 104) = 1.44; p = 0.028 Habitat type X Sex- 6 Pillai’s Trace value = 0.231; F = 1.93 F = 0.68 F = 0.82 age categories F (15, 159) = 0.89; p = 0.582 *** = p < 0.001 ** = p < 0.01 * = p < 0.05 In the wet season, percent cover of grass and herbs significantly affected most behaviors of Mountain Nyala, including proportion of time vigilant (Table 4.1; Fig. 4.5) and feeding (Table 4.1; Fig. 4.6). However, during the dry season, percent cover of grass

66

and herbs affected only proportion of time vigilant (Table 4.1; Fig. 4.9). Regardless of season, altitude of landscape had no effect on the behaviors of Mountain Nyala (Table 4.1). In the wet season, slope condition significantly affected most behaviors of Mountain Nyala, including proportion of time vigilant (Table 4.1; Fig. 4.5) and feeding (Table 4.1; Fig. 4.6). However, in the dry season, slope condition did not affect any behaviors of Mountain Nyala (Table 4.1). In the wet season, group size did not affect any behaviors of Mountain Nyala (Table 4.1). In contrast, in the dry season, group size affected most behaviors of Mountain Nyala, including proportion of time feeding (Table 4.1; Fig. 4.10) and moving (Table 4.1; Fig. 4.11). In the wet season, habitat type significantly affected most behaviors of Mountain Nyala, including proportion of time feeding and moving (Table 4.1). Accordingly, proportion of time feeding was the highest in the clear cut (mean = 0.35; stdev = 0.27), followed in the natural forest (mean = 0.27; stdev = 0.24), and in the plantation (mean = 0.04; stdev = 0.09). However, proportion of time moving was highest in the plantation (mean = 0.41; stdev = 0.17), followed in the clear cut (mean = 0.29; stdev = 0.13), and in the natural forest habitat (mean = 0.23; stdev = 0.16). Similarly, during the dry season, habitat type significantly affected most behaviors of Mountain Nyala, including proportion of time vigilant (Table 4.1; Fig. 4.2) and moving (Table 4.1). Accordingly, proportion of time vigilant was highest in the clear cut (mean = 0.71; stdev = 0.12), followed in the plantation (mean = 0.42; stdev = 0.10), and in the natural forest habitat (mean = 0.29; stdev = 0.07). Where as proportion of time moving was highest in the natural forest (mean = 0.53; stdev = 0.07), followed in the plantation (mean = 0.43; stdev = 0.08), and in the clear cut habitat (mean = 0.26; stdev = 0.14).

0.9 0.8 0.7 0.6 Clear cut 0.5 Plantation 0.4 Natural forest 0.3 0.2

0.1 Proportion of time vigilant time of Proportion 0 Wet Dry Season

Figure 4.2. The effect of habitat type on the vigilance level of Mountain Nyala during the wet and dry seasons in Munessa pooled over the four categories, adult males, adult females, sub-adults, and juveniles. The number of focal individuals included in this analysis was 119 and 116 for the wet and the dry seasons respectively. The error bars represent +1 stdev.

67

In the wet season, sex-age categories of the group containing the focal animal did not affect any behaviors of Mountain Nyala (Table 4.1). In contrast, during the dry season, sex-age categories of the group containing the focal animal significantly affected most behaviors of Mountain Nyala, including proportion of time vigilant (Table 4.1; Fig. 4.3) and feeding (Table 4.1; Fig. 4.10). Regardless of season, habitat type did not significantly interact with sex-age categories of the Mountain Nyala (see Table 4.1 above).

0.8 0.7 0.6 Adult male 0.5 Adult female 0.4 0.3 Sub-adult 0.2 Juvenile 0.1

Proportion of time vigilant time of Proportion 0 Wet Dry Season

Figure 4.3. The effect of sex-age categories on the proportion of time vigilant during the wet and the dry seasons. The number of focal individual Mountain Nyala considered for the wet season analysis was 17 adult males, 41 adult females, 27 sub-adults, and 31 juveniles. However, 17 adult males, 38 adult females, 28 sub-adults, and 36 juveniles were considered for the dry season analysis. The error bars represent +1 stdev.

In the wet season, Mountain Nyala did not show any significant difference in their group size across habitat types in Munessa (F (2, 46) = 1.43; p = 0.249) (Fig. 4.4). In contrast, during the dry season, Mountain Nyala showed significant difference in their group size across habitat types (F (2, 38) = 19.57; p < 0.001) (Fig. 4.4). Accordingly, the greatest mean group size was obtained in the natural forest (mean = 8.48 individuals; stdev = 2.50), but the mean group size was equal both in the plantation and the clear cut habitats (mean = 4.00 individuals in each habitat type; stdev = 2.20 in the plantation habitat; stdev = 1.63 in the clear cut habitat) (Fig. 4.4). In the dry season, the greatest maximum group size was recorded in the natural forest habitat (max = 13 individuals), followed by the plantation (max = 9 individuals), and the clear cut habitat (max = 6 individuals). A post- hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in group size between natural forest versus clear cut habitats, natural forest

68

versus plantation habitats (Fig. 4.4). Overall, group size of Mountain Nyala was significantly higher (F (1, 84) = 8.99; p = 0.004) in the wet season (mean = 7.35 individuals; stdev = 3.19) than in the dry season (mean = 6.29 individuals; stdev = 0.93).

14

12

10 Clear cut 8 Plantation 6 Natural forest

Groupsize 4

2

0 Wet Dry Season

Figure 4.4. Seasonal variation of group sizes in Mountain Nyala across habitat type in Munessa. The error bars represent +1 stdev.

Wet season Mountain Nyala behavioral models I developed the behavioral models by correlating each habitat variable, sex-age categories, and group size measured with the proportion of time budget quantified for each activity type of focal Mountain Nyala. To examine the effect of the coefficient of each independent variable on the vector of each dependent activity type, I defined the alpha value of 0.05 for all analyses. A summary of the detailed wet season behavioral models is shown in Table 4.2. Effects of each independent variable on the respective dependent variable are described briefly below. Proportion of time vigilant significantly increased with the abundance of trees and the percent cover of grass and herbs (Table 4.2; Fig. 4.5), but decreased with the percent cover of bare soils and slope condition (Table 4.2; Fig. 4.5). Proportion of time feeding significantly decreased with the abundance of trees and percent cover of bare soil (Table 4.2; Fig. 4.6), but significantly increased with the percent cover of grass and herbs and percent slope condition (Table 4.2; Fig. 4.6). None of the habitat variables significantly correlated with the moving behaviors of Mountain Nyala (Table 4.2; Fig. 4.7). Proportion of time resting by Mountain Nyala significantly increased with shrub abundance only (Table 4.2; Fig. 4.8).

69

Table 4.2. A summary of the behavioral models for Mountain Nyala in the wet and the dry seasons in Munessa. The intercept and the coefficient of each independent variable are included in the model. The independent variables included in each model are abbreviated as follows. Var.1 = Abundance of trees; Var.2 = Crown diameter of trees (m); Var.3 = Abundance of shrubs; Var.4 = Percent cover of grass and herbs; Var.5 = Percent cover of bare soils; Var.6 = Altitude (m); Var.7 = Slope condition (%); Var.8 = Group size; Var.9 = Sex-age categories of the focal animals. Variables whose coefficients significantly differ at 0.05 alpha values are marked with an asterisk (*). Season Behavioral activity Model (0.367 + 0.021 x Var.1)* + (0.421 - 0.001 x Var.2) + Proportion of time vigilant (0.429 - 0.07 x Var.3) + (0.352 + 0.002 x Var.4)* + (0.549 - 0.002 x Var.5)* + (-0.182 - 0.0003 x Var.6) + (0.4703 - 0.0049 x Var.7)* + (0.446 - 0.003 x Var.8) + (0.485 - 0.001 x Var. 9) (0.298 - 0.025 x Var.1)* + (0.273 - 0.006 x Var.2) + Proportion of time feeding (0.269 - 0.014 x Var.3) + (0.063 + 0.003 x Var.4)* + (0.327 - 0.003 x Var.5)* + (1.159 - 0.0004 x Var.6) + (0.1653 + 0.0086 x Var.7)* + (0.186 + 0.007x Var.8) + (2.537 - 0.023 x Var.9) Wet (0.269 + 0.009x Var.1) + (0.273 + 0.003 x Var.2) + Proportion of time moving (0.286 + 0.003 x Var.3) + (0.297 - 0.0002 x Var.4) + (0.279 - 0.0002 x Var.5) + (-0.048 + 0.0002 x Var.6) + (0.2994 - 0.0029 x Var.7) + (0.309 - 0.001 x Var.8) + (-1.341 + 0.016x Var.9) (0.066 - 0.007 x Var.1) + (0.034 + 0.003 x Var.2) + Proportion of time resting (0.016 + 0.018 x Var.3)* + (0.107 - 0.001 x Var.4) + (0.024 + 0.001 x Var.5) + (0.063 - 0.00001 x Var.6) + (0.065 - 0.0008 x Var.7) + (0.059 - 0.003 x Var.8) + (-0.682 + 0.007 x Var.9) (0.385 + 0.001 x Var.1) + (0.415 - 0.005 x Var.2) + Proportion of time vigilant (0.417 - 0.016 x Var.3)* + (0.344 + 0.002 x Var.4)* + (0.523 - 0.002 x Var.5)* + (0.656 - 0.0001 x Var.6) + (0.4453 - 0.0028 x Var.7) + (0.512 - 0.019 x Var.8) + (-1.352 + 0.017 x Var.9) * (0.138 - 0.002 x Var.1) + (0.116 + 0.003 x Var.2)* + Proportion of time feeding (0.116 + 0.009 x Var.3)* + (0.151 - 0.001 x Var.4) + Dry (0.076 + 0.001 x Var.5) + (0.326 - 0.0001 x Var.6) + (0.105 + 0.0002 x Var.7) + (0.005 + 0.093 x Var.8)* + (0.969 - 0.008 x Var.9) * (0.468 - 0.002 x Var.1) + (0.449 + 0.003 x Var.2) + Proportion of time moving (0.448 + 0.008 x Var.3) + (0.489 - 0.001 x Var.4) + (0.379 + 0.001 x Var.5) + (0.347 + 0.0002 x Var.6) + (0.4235 + 0.0029 x Var.7) + (0.357 + 0.016 x Var.8)* + (1.399 - 0.009 x Var. 9) (0.009 + 0.004 x Var.1)* + (0.020 - 0.001 x Var.2) + Proportion of time resting (0.019 - 0.001 x Var.3) + (0.016 + 0.00004 x Var.4) + (0.021 - 0.00004 x Var.5) + (-0.329 + 0.0002 x Var.6)* + (0.0263 - 0.0003 x Var.7) + (0.038 - 0.002 x Var.8) + (-0.017 + 0.0004 x Var.9)

70

Var. 1 Var. 2

1 1 0.8 0.8 0.6 0.6

0.4 0.4 vigilant vigilant

0.2 0.2 Time proportion Time proportion Time 0 0 0 2 4 6 8 10 0 5 10 15 20 25 30 35 Abundance of trees Crown diameter of trees (m)

Var. 3 Var. 4

1 1 0.8 0.8 0.6 0.6 0.4

vigilant 0.4 vigilant 0.2

Time proportion Time 0.2 proportion Time 0 0 0 2 4 6 8 10 12 0 10 20 30 40 50 60 70 80 90 100 Abundance of shrubs Grass and herbs (%)

Var. 5 Var. 6

1 1 0.8 0.8 0.6 0.6 0.4 vigilant 0.4 vigilant 0.2 Time proportion proportion Time 0.2 proportion Time 0 0 0 10 20 30 40 50 60 70 80 90 100 2100 2150 2200 2250 2300 2350 Soil substrates (%) Altitude (m)

Var. 8 Var. 7 1 1 0.8 0.8 0.6 0.4 0.6 vigilant 0.2 0.4 vigilant

Timeproportion 0 0 10 20 30 40 50 0.2 proportion Time Slope condition (%) 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Group size Var. 9 1.0

0.8

0.6

0.4

0.2

Time proportion vigilant proportion Time 0.0 Adult female Sub-adult Juvenile Adult male

Figure 4.5. A model for the vigilance behavior of Mountain Nyala in Munessa during the wet season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time vigilant are also shown in Table 4.2.

71

Var. 1 Var. 2

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0

Time proportion feeding proportion Time 0 Time proportion feeding proportion Time 0 2 4 6 8 10 0 5 10 15 20 25 30 35 Abundance of trees Crown diameter of trees (m) Var. 3 Var. 4 1 0.8 1 0.6 0.8 0.4 0.6 0.2 0.4 0 Time proportion feeding proportion Time 0.2 0 2 4 6 8 10 12 0 Abundance of shrubs feeding proportion Time 0 10 20 30 40 50 60 70 80 90 100 Grass and herbs (%) Var. 5 Var. 6 1 0.8 1 0.6 0.8 0.4 0.6 0.2 0.4 0 feeding proportion Time 0.2 0 10 20 30 40 50 60 70 80 90 100 0 Soil substrates (%) feeding proportion Time 2100 2150 2200 2250 2300 2350 Var. 7 Altitude (m) 1 0.8 Var. 8 0.6 0.4 1

feeding 0.2

Timeproportion 0 0.8 0 10 20 30 40 50 0.6 Slope condition (%) 0.4 0.2 Var. 9 0 1.0 feeding proportion Time 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 0.8 Group size 0.6 0.4 0.2

feeding Time proportion 0.0 Adult female Sub-adult Juvenile Adult male

Figure 4.6. A model for the feeding behavior of Mountain Nyala in Munessa during the wet season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time feeding are also shown in Table 4.2.

72

Var. 1 Var. 2

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0

Time proportion moving proportion Time 0 Time proportion moving proportion Time 0 2 4 6 8 10 0 5 10 15 20 25 30 35 Abundance of trees Crown diameter of trees (m) Var. 3 Var. 4

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0 Time proportion moving proportion Time 0 0 2 4 6 8 10 12 moving proportion Time 0 10 20 30 40 50 60 70 80 90 100 Abundance of shrubs Grass and herbs (%) Var. 5 Var. 6 1 1 0.8 0.6 0.8 0.4 0.6 0.2 0.4

0 0.2 Time proportion moving proportion Time

0 10 20 30 40 50 60 70 80 90 100 0 Time proportion moving proportion Time Soil substrates (%) 2100 2150 2200 2250 2300 2350

Var. 7 Altitude (m) Var. 8 1 0.8 1 0.6 0.4 0.8 moving 0.2 0.6 proportion Time 0 0 10 20 30 40 50 0.4 Slope condition (%) 0.2

0 Time proportion moving proportion Time 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Var. 9 Group size 1.0 0.8 0.6 0.4 0.2

Time Time proportion moving 0.0 Adult female Sub-adult Juvenile Adult male

Figure 4.7. A model for the moving behavior of Mountain Nyala in Munessa during the wet season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time moving are also shown in Table 4.2.

73 Var. 1 Var. 2

1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

Time proportion resting proportion Time 0

resting proportion Time 0 0 2 4 6 8 10 0 5 10 15 20 25 30 35 Abundance of trees Crown diameter of trees (m) Var. 3 Var. 4 1 1 0.8 0.6 0.8 0.4 0.6 0.2 0.4 0.2 resting proportion Time 0 0 2 4 6 8 10 12 resting proportion Time 0 Abundance of shrubs 0 10 20 30 40 50 60 70 80 90 100 Grass and herbs (%) Var. 5 Var. 6 1 0.8 1 0.6 0.8 0.4 0.6 0.2 0.4

Time proportion resting proportion Time 0 0.2 0 10 20 30 40 50 60 70 80 90 100

Time proportion resting proportion Time 0 Soil substrates (%) 2100 2150 2200 2250 2300 2350 Altitude (m) Var.7 Var. 8 1 0.8 1 0.6 0.8 0.4 0.2 0.6 0

resting proportion Time 0.4 0 10 20 30 40 50 Slope condition (%) 0.2

resting proportion Time 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Var. 9 Group size 1.0 0.8 0.6 0.4 0.2

0.0 resting Timeproportion Adult female Sub-adult Juvenile Adult male

Figure 4.8. A model for the resting behavior of Mountain Nyala in Munessa during the wet season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time resting are also shown in Table 4.2.

74

Dry season Mountain Nyala behavioral models The dry season behavioral models showed significant independent variables that were a bit different from the wet season behavioral models. A summary of the detailed dry season behavioral models also appears in Table 4.2. Effects of each independent variable on the respective dependent variable are described briefly below. Proportion of time vigilant significantly decreased with the abundance of shrubs and percent cover of bare soils, but increased with the percent cover of grass and herbs (Table 4.2; Fig. 4.9). Moreover, proportion of time vigilant was significantly affected by sex-age categories of the focal Mountain Nyala and it was the highest for the adult males (Table 4.2; Fig. 4.9). Proportion of time feeding significantly increased with crown diameter of trees, abundance of shrubs, and group size of Mountain Nyala (Table 4.2; Fig. 4.10). In addition, proportion of time feeding was significantly affected by sex and age categories of the focal Mountain Nyala and it was the highest for sub-adults (Table 4.2; Fig. 4.10). Proportion of time moving significantly increased with group size only (Table 4.2; Fig. 4.11). Proportion of time resting significantly increased with the abundance of trees and altitude of the landscape (Table 4.2; Fig. 4.12).

75

Var. 1 Var. 2 1 1 0.8 0.8 0.6 0.6

0.4 0.4 vigilant vigilant

0.2 0.2

Time proportion Time Time proportion Time 0 0 0 2 4 6 8 10 0 5 10 15 20 25 30 35 AbundanceVar. 3 of trees CrownVar. diameter 4 of trees (m) 1 1 0.8 0.8 0.6 0.6 0.4 vigilant 0.4

vigilant

0.2 Time proportion proportion Time prportion Time 0.2 0 0 0 2 4 6 8 10 12 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Abundance of shrubs Grass and herbs (%) Var. 5 Var. 6

1 1 0.8 0.8 0.6 0.6 0.4

vigilant 0.4 0.2 vigilant

Time proportion Time 0.2

proportion Time 0 0 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 2100 2150 2200 2250 2300 2350

Soil substrates (%) Altitude (m) Var. 8 Var. 7 1 1 0.8 0.8 0.6 0.6

0.4 0.4 vigilant 0.2 vigilant

0.2 Time proportion Time

Timeproportion 0 0 10 20 30 40 50 0 Slope condition (%) 2 3 4 5 6 7 8 9 10 11 12 13 Group size

Var. 9 1.0

0.8

0.6

0.4 0.2

Time vigilant proportion 0.0 Adult female Sub-adult Juvenile Adult male

Figure 4.9. A model for the vigilance behavior of Mountain Nyala in Munessa during the dry season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time vigilant are also shown in Table 4.2.

76

Var. 1 Var. 2 1 1 0.8 0.8 0.6 0.6

0.4 0.4

feeding feeding 0.2

0.2 proportion Time Time proportion Time 0 0 0 2 4 6 8 10 0 5 10 15 20 25 30 35 Abundance of trees Crown diameter of trees (m)

Var. 3 Var. 4

1 1 0.8 0.8 0.6 0.6 0.4

0.4 feeding feeding

0.2 prportion Time 0.2 proportion Time 0 0 0 2 4 6 8 10 12 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 Abundance of shrubs Grass and herbs (%) Var. 5 Var. 6 1 1 0.8 0.8 0.6 0.6 0.4 feeding 0.4 0.2 feeding

proportion Time 0.2 0 proportion Time 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 0 2100 2150 2200 2250 2300 2350 Soil substrates (%) Altitude (m)

Var. 8 Var.7 1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0 0.2 feeding proportion Time

0 5 10 15 20 25 30 35 40 45 50 0 Time proportion feeding proportion Time Slope condition (%) 2 3 4 5 6 7 8 9 10 11 12 13 Group size

Var. 9 1.0

0.8

0.6

0.4

0.2 Time proportion feeding 0.0 Adult female Sub-adult Juvenile Adult male

Figure 4.10. A model for the feeding behavior of Mountain Nyala in Munessa during the dry season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time feeding are also shown in Table 4.2.

77

Var. 1 Var. 2 1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

0 Time proportion moving proportion Time

Time prportion moving prportion Time 0 0 5 10 15 20 25 30 35 0 2 4 6 8 10 Crown diameter of trees Abundance of trees Var.3 Var. 4 1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0 moving proportion Time 0 Time proportion moving proportion Time 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 0 2 4 6 8 10 12 Grass and herbs (%) Abundance of shrubs Var. 6 Var. 5 1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0 0.2 moving proportion Time 2100 2150 2200 2250 2300 2350 0 moving proportion Time Altitude (m) 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Soil substrates (%)

Var. 8 Var.7 1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

0 0

Time proportion moving proportion Time Time proportion moving proportion Time 0 5 10 15 20 25 30 35 40 45 50 2 3 4 5 6 7 8 9 10 11 12 13 Slope condition (%) Group size Var. 9 1.0

0.8 0.6

0.4

0.2

Timemoving proportion 0.0 Adult female Sub-adult Juvenile Adult male

Figure 4.11. A model for the moving behavior of Mountain Nyala in Munessa during the dry season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time moving are also shown in Table 4.2.

78

Var. 1 Var. 2 0.25 0.25 0.2 0.2 0.15 0.15 0.1 0.1 0.05 0.05

Time proportion resting proportion Time 0

Time proportion resting proportion Time 0 0 2 4 6 8 10 0 5 10 15 20 25 30 35 Crown diameter of trees (m) Abundance of trees Var. 4 Var. 3 0.25 0.25 0.2 0.2 0.15 0.15 0.1 0.1 0.05 0.05

resting proportion Time 0

Time proportion resting proportion Time 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 0 2 4 6 8 10 12 Grass and herbs (%) Abundance of shrubs Var. 6 Var. 5 0.25 0.25 0.2 0.2 0.15 0.15 0.1 0.1 0.05 0.05

resting proportion Time 0

resting proportion Time 0 2100 2150 2200 2250 2300 2350 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Altitude (m) Soil substrates (%) Var. 8 Var. 7 0.25 0.2 0.25 0.2 0.15 0.15 0.1 0.1 resting 0.05 0.05

Timeproportion 0 resting proportion Time 0 0 10 20 30 40 50 2 3 4 5 6 7 8 9 10 11 12 13 Slope condition (%) Group size

Var. 9 1.0 0.8 0.6 0.4 0.2 0.0

Time proportion resting Adult female Sub-adult Juvenile Adult male

Figure 4.12. A model for the resting behavior of Mountain Nyala in Munessa during the dry season. The intercept and the slope of each variable, and those variables with significant effects on the proportion of time resting are also shown in Table 4.2.

79

DISCUSSION Behavioral modeling based on time budgets of focal individuals helps understand how animals perceive their environment (Manor and Saltz, 2003; Hochman and Kotler, 2006b; Rieucau, and Martin, 2008) and how they trade-off food acquisition and safety (Stephens and Krebs, 1986; Lima and Dill, 1990; Brown et al., 2001; Brown and Kotler, 2004). As an example, frightened prey allocate more time to vigilance and less time to feeding (Pulliam, 1973; Brown, 1999). As a means for managing risk, prey can also use time allocation and apprehension (Kotler et al., 2002; Hochman and Kotler, 2006b) or avoid exploiting risky patches (Pulliam, 1973; Brown, 1999; Altendorf et al., 2001). So, fear from risk of predation ultimately affects habitat selection, patch use, and activity time budgets of individuals of a species. I have shown here how measuring time budgets through focal animal observations yield behavioral indicators (Morris et al., 2009) to study the habitat selection and use behaviors of Mountain Nyala. The present study revealed that time budgets of free-ranging Mountain Nyala significantly varied among activity types in Munessa. In the wet season, Mountain Nyala allocated most of their time to vigilance behavior. A possible explanation for this outcome is that legal sport hunting in the Munessa hunting block is mostly practiced in the wet season. In addition, human and livestock disturbances are common in all habitat types in the wet season (see also Chapter 6). The hunting pressure coupled with human and livestock induced disturbances may compel the Mountain Nyala to allocate most of their time budgets to safety related activities. Manor and Saltz (2003) reported that increase their time devoted to vigilance during times of high human disturbances. The result further suggested that Mountain Nyala allocate relatively a small portion of their time to feeding and resting behaviors. In the dry season, the result showed that Mountain Nyala allocated the greatest portion of their time to moving behavior. One possible motivation for Mountain Nyala movement may be the scarcity of palatable forage in the dry season which in turn may force the Mountain Nyala to allocate their largest proportion of time to moving behavior in search of sufficient food. The Mountain Nyala also invested their highest proportion of time to moving behavior during the dry season in the Bale Mountains National Park (see Chapter 5). Like in the wet season, Mountain Nyala allocated small portions of their time

80

to feeding and resting in the dry season. Irrespective of season and habitat type, this suggests that impacts of risk of predation and human and livestock disturbances (see also Chapter 6 for the details) are common threats to Mountain Nyala in Munessa. The type and intensity of human and livestock disturbances on the habitats of Mountain Nyala are addressed and discussed in greater depth in Chapter 6. Night time biased activity of Mountain Nyala (see Chapter 3) also supports the notion that human and livestock induced problems have large impacts on the behaviors and habitat uses of Mountain Nyala in Munessa. Different parts of the environment represent habitats of varying quality, in terms of costs and benefits, for different organisms (Melton, 1987). During the wet season, the result showed that habitat type significantly affected time allocated to feeding and moving behavior by free-ranging Mountain Nyala (Table 4.1). In contrast, Mountain Nyala did not show significant difference in time allocated to vigilance behavior across habitat type. This may be because human and livestock related disturbances and threats are common throughout the available habitats in Munessa (see also Chapter 6). And this may be why Mountain Nyala allocated the greatest proportion of their time to vigilance as discussed above. In addition, in the wet season, all the local farm lands are in cultivation and not available for livestock grazing. This is when the local people often drive their livestock into the Munessa Forest for free-range grazing. Wood cutting by humans was prominent both in the natural forest and the plantation habitats in the wet season when the local people frequently collect wood for commercial purpose (see also Chapter 6). Thus, using the natural forest or the plantation habitat for concealment may not help much to reduce the vigilance level of Mountain Nyala compared with the clear cut habitat. The high level of human and livestock disturbances on the available habitats of Mountain Nyala in Munessa during the wet season is quantified and discussed in Chapter 6. For example, illegal wood cutting by humans was prominent both in the natural forest and plantation habitat in the wet season (see also Chapter 6). In the dry season, the proportion of time vigilant significantly differed among habitat types, with Mountain Nyala being most vigilant in the clear cut habitat. This may be because Mountain Nyala need good cover for concealment from risk of predation and human and livestock disturbances in the day time (Evangelista et al., 2007; Mamo, 2007).

81

In addition, at this time crops have been harvested and crop residues are available for livestock feeding on farm lands in the dry season. So, livestock disturbances are less common in the natural forest during the dry season (see also Chapter 6). Moreover, the local people do not frequently cut wood in the natural forest for commercial purpose during the dry season. Instead local people mostly sell grain to earn money. Thus, the Mountain Nyala can use the natural forest for concealment and hence be less vigilant there than in the clear cut or the plantation habitat. This in turn may result in significant variation in the vigilance level of Mountain Nyala based on habitat type of occupation. The present study also revealed that Mountain Nyala significantly devoted their time to moving and vigilance behavior in the clear cut habitat during the dry season. In the dry season, human and livestock disturbances are high in the open habitat so that Mountain Nyala frequently run to the adjacent natural and plantation habitats to seek concealment. At this time, daylight activity density for Mountain Nyala was also biased to the natural forest (see also Chapter 3), further supporting this interpretation. In Munessa, the local people illegally hunt the Mountain Nyala mostly during the dry season. Moreover, the illegal hunting pressure in Munessa is more chronic than the legal hunting which takes place only during the wet season in the area (information obtained from local people). Thus, the openness of the habitat may help the local people to more successfully hunt the Mountain Nyala in the clear cut habitat than in the dense natural forest because they use guns to hunt the Mountain Nyala (information obtained from the Munessa hunting block scouts). In addition, massive hunting parties of the local people ride on horseback (horses run fast in the open habitat) to chase and kill the Mountain Nyala with spears during the illegal hunting in the dry season (e.g., Evangelista et al., 2007). In the wet season, time spent in vigilance did not differ among the different age and sex classes of Mountain Nyala. This suggests that all age and sex categories of Mountain Nyala are equally vulnerable to risk of predation, human nuisance, and livestock disturbances. However, this is contrary to the classical predation risk hypothesis which notes that adult females should spend much more time to vigilance and safety related activities than do adult males (e.g., Sukumar and Gadgil, 1988; Young and Isbell, 1991; Muller et al., 1995; Main et al., 1996; Prins, 1996; Ruckstuhl, 1998; Laundre´ et al., 2001; Ruckstuhl and Kokko, 2002; Ruckstuhl, 2007). Although adult females with kids have

82

additional costs arising from the risk of predation on the young in the wet season, adult males in the Munessa hunting block have additional costs due to legal trophy hunting that solely targets males in the wet season. Thus, in the wet season, both adult males and females allocated almost equal proportion of time to vigilance behavior in a departure from the predictions of the classical predation risk hypothesis. Similar departure from the predation risk hypothesis occurred in the dry season. At that time, adult males were even more vigilant than the other sex and age categories. While in the wet season, it was legal trophy hunting that targeted adult males, in the dry season it is also illegal hunting that does the same. Large males are targeted by local people for their spectacular horns. These horns are used for ritual ceremonies (from discussions held with local people), local medicines, and nipples for traditional milk bottles (Sillero-Zubiri, 2008). All this may have encouraged local people to kill more adult males than females in Munessa. Brown (1969b) noted that intensive hunting of Nyala by local people was openly admitted and discussed in the Arsi region. Evangelista et al. (2007) also noted that massive hunting parties of local people are organized using horseback, spears, and dogs on a regular basis in the dry season to kill the male Mountain Nyala in the Arsi region including Munessa. Illegal hunting pressure is more chronic than the legal hunting in Munessa (information obtained from local people). Consequently, adult males may have a reason for being especially vigilant. The result further revealed that, compared with the other sex and age categories, adult males allocated significantly the least time to feeding (see also Fig. 4.10), suggesting that adult males behavior is a reflection of a trade-off food acquisition and safety. In addition, in the dry season, adult males totally avoided using the clear cut habitat in the daylight hours. However, in areas such as the Bale Mountains National Park where there is no hunting pressure posed on Mountain Nyala, adult males devoted the least time to vigilance behavior compared with the other sex-age categories (see also Chapter 5, Fig. 5.8). This shows how the prevalence of legal and illegal hunting pressures affected the vigilance behavior of adult male Mountain Nyala in Munessa. Using environmental factors, it is possible to address the habitat selection and patch use behaviors of Mountain Nyala (see also Table 4.1, 4.2; Fig. 4.5 through Fig. 4.12). These should be places where opportunities are maximized and risks are minimized (Tadesse and Kotler, 2011) based on environmental and behavioral factors acting together

83

to influence the habitat use of Mountain Nyala. For example, the wet season feeding behavioral model suggests that Mountain Nyala spend a considerable proportion of their time to feeding where there is high percent cover of palatable grass and herbs, good sources of food for Mountain Nyala (e.g., Evangelista et al., 2007; Mamo, 2007). In addition, Mountain Nyala devoted more time to feeding and less time to vigilance in sloppy terrain (Table 4.2), because slope provides Mountain Nyala with safety so that they allocated their considerable time to feeding (e.g., Evangelista et al., 2007; Mamo, 2007). In contrast, Mountain Nyala devoted less time to feeding with the increase in the percent cover of soil substrate where there was less food to eat. The wet season behavioral model further revealed that the resting behavior of Mountain Nyala significantly increased with shrub abundance. This suggests that shrubs can serve as good cover to resting for Mountain Nyala. Evangelista et al. (2007) and Mamo (2007) also noted that Mountain Nyala need good cover for concealment and resting. In the dry season, the behavioral model revealed that the feeding behavior of Mountain Nyala significantly increased with the abundance of shrubs, suggesting that shrubs are good sources of food for Mountain Nyala during the dry season in Munessa. Other authors also noted that Mountain Nyala rely on shrubs as their typical sources of food in the Bale Mountains National Park during the dry season when grass and herbs are less abundant and poor in their nutritional quality and digestibility (e.g., Hillman and Hillman, 1987; Refera and Bekele, 2004; Mamo, 2007). In contrast, Mountain Nyala devoted less time to vigilance when the abundance of shrubs increases (see Table 4.2; Fig. 4.9), suggesting that shrubs may provide Mountain Nyala with good concealment from risk of predation (Evangelista et al., 2007). The dry season behavioral model suggests that the moving behavior of Mountain Nyala significantly increased with an increase in group size. This may result from the high competition for limited food resources in the dry season when group size increases. Previous field reports also noted that Mountain Nyala become highly mobile to search for sufficient food both in quantity and quality in the dry season (Ethiopia Wildlife Conservation Authority, unpublished field reports). Consequently, it was shown that Mountain Nyala devoted much of their time to feeding as group size increases in the dry season (see Table 4.2; Fig. 4.10). The dry season behavioral model further revealed that the resting behavior of Mountain Nyala significantly increased with an

84

increase in the abundance of trees and altitude of the landscape. This suggests that Mountain Nyala may require trees and higher grounds for resting in the dry season. Evangelista et al. (2007) also noted that Mountain Nyala use dense vegetation for camouflage, and higher grounds for detecting risk of predation and resting. So, both in the wet and the dry seasons, it appears that the behaviors of Mountain Nyala are related to foraging opportunities and predation risk characteristics of the habitat as revealed by the behavioral models in this study. Regardless of season, group size did not affect the vigilance level of Mountain Nyala in Munessa. The absence of the group size effect could be due to the fact that for females, dependent kids cannot act as dilution agents (e.g., Laundre´ et al., 2001; Rieucau and Martin, 2008). In regards to minimizing risk of predation or human nuisance, Dall and Valone (2005) noted that individual information is more important than group information in some social animals. When personal information is valued more than social information (Lima et. al., 1985), constant levels of vigilance regardless of group size is a more likely outcome. Lack of a group size effect on vigilance has been recorded in other species of social ungulates such as elk ( elaphus) (Laundre´ et al., 2001), (Bison bison) (Laundre´ et al., 2001), and Nubian Ibex (Capra nubiana) (Hochman and Kotler, 2006b; Tadesse and Kotler, 2011). Other studies have found either no effect or the opposite effect of group size on the level of vigilance in social animals (e.g., Lima, 1995; Treves, 2000; Beauchamp, 2003). The mean group size of Mountain Nyala did not differ among habitat types in the wet season. However, the greatest maximum group size (22 individuals) was recorded during the wet season in the clear cut habitat in Munessa. Refera and Bekele (2004) also reported the greatest maximum group size (62 individuals) of Mountain Nyala in the grassland habitat of the Bale Mountains National Park during the wet season. In contrast, during the dry season, the mean group size of Mountain Nyala significantly differed among habitat types, with both the greatest mean group size and maximum group size (13 individuals) occurring in the natural forest habitat. One possible explanation for this is that the availability of palatable forage coupled with good cover condition from risk of predation and hot weather may favor large aggregations in the natural forest habitat during the dry season. Dry season daylight activity densities from Chapter 3 were also consistent

85

with this result. Other authors noted that Mountain Nyala aggregate in a habitat which provisions major habitat requirements, such as food, water, and cover (e.g., Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). However, the availability of food in the wet season may favor the Mountain Nyala to have an overall larger mean group size than in the dry season (e.g., Refera and Bekele, 2004; Evangelista et al., 2007) which is consistent with the result of the present study. Dry season daylight activity densities from Chapter 3 were also consistent with this result. To summarize, behavioral models are important tools to predict the variations in the behavioral responses of Mountain Nyala related to foraging, human nuisance and livestock disturbances, season, and predation risk characteristics of the habitat. Social factors (e.g. group size) and individual characteristics (e.g. sex and age classes) that may affect the behavioral responses of Mountain Nyala were successfully considered and built into the behavioral models. This may help generate more realistic outputs that reflect a range of factors influencing the habitat selection and patch use behaviors of Mountain Nyala in Munessa. I believe that the present behavioral study is helpful in giving clues to resolve conservation problems for the endangered Mountain Nyala and their habitats in Munessa. For example, to conserve Mountain Nyala, wildlife managers in the Munessa hunting block need to promote novel behaviors that enhance survivorship in response to novel hazards. From a management perspective, searching for the fitness consequences of behavioral responses may provide insights into the impacts of human and livestock on Mountain Nyala.

86

CHAPTER 5 HABITAT USE BY MOUNTAIN NYALA DETERMINED USING BITE DIAMETERS, BITE RATES, AND TIME BUDGETS IN THE BALE MOUNTAINS NATIONAL PARK ABSTRACT Measuring and quantifying the natural food items consumed by free-ranging wild herbivores is crucial for habitat and species management. I hypothesized that Mountain Nyala perceive differences in their habitat quality that can be evaluated and determined by applying foraging theory. Where bite diameters will be greatest, I also expect bite rates (number of bites per minute) to be highest, vigilance rates the lowest, and the most time devoted to feeding. Use of open grasslands versus dense woodlands by free-ranging Mountain Nyala was assessed and determined through quantifying bite diameters at point of browse, bite rates, and time budgets. Bite diameters provide a natural measure of giving- up density and can be used to assess foraging costs and foraging efficiencies, with greater bite diameters corresponding with lower costs and greater efficiencies. Along each transect through each habitat patch, the browse species cropped by free-ranging Mountain Nyala were identified and the diameters of all browsed twigs were measured. Focal observations were carried out and bite rates were recorded for each sex-age categories of target animal. Time budgets were also quantified for focal animals according to sex-age categories and habitat type. The numbers of individuals in the group were also recorded. The results revealed that bite diameters, bite rates, vigilance rates and proportion of time spent foraging all differed between habitats. In particular, Mountain Nyala had greater bite diameters, higher bite rates, and spent a greater proportion of their time foraging and less in vigilance in the grassland habitat. In addition, adult females had the highest bite rates, and the browse species Solanum marginatum had the greatest bite diameter. The results show that grasslands are a higher quality habitat than woodlands, offering lower foraging costs, greater safety, and more time for foraging. The results further show how behavioral indicators and natural giving-up densities can reveal habitat quality for endangered wildlife through the use of non-invasive techniques. Key words: bite diameters, bite rates, foraging theory, habitat quality, natural giving-up densities, non-invasive techniques, time budgets, vigilance

87

INTRODUCTION The landscape where an animal lives is composed of habitats that vary in quality (Druce, 2005), and the use of those habitats can affect individual fitness (Melton, 1987). So, habitat selection by individuals of a species is affected by the presence of habitat essentials (Stephens and Krebs, 1986) and suitability of escape terrain (e.g., Stephens and Krebs, 1986; Kotler et al., 1994; Shrader et al., 2008a). As foraging behavior is related to environmental factors that influence energetic or predation costs (Hochman and Kotler, 2006a), the way in which an animal perceives and uses its environment is a central point of its ecology (Johnson, 1980). Previous studies demonstrated that habitat quality and characteristics influence temporal activity patterns, foraging behavior, anti-predator behavior, and social organization of a species (e.g., Brown and Alkon, 1990; Rosenzweig, 1981; Cresswell, 1994; Kotler et al., 1994; Arenz and Leger, 1997; Druce, 2005). Optimal foraging theory provides a unique avenue to investigating the adaptive behaviors of animals related with resource densities over different habitat types (Brown, 1988, 1992; Brown and Kotler, 2004; Druce et al., 2006; Morris and Mukherjee, 2007; Morris et al., 2009). Measuring the patch use behaviors of a forager can reveal its costs and benefits from foraging (e.g., Brown, 1988, 1992) as well as the environmental factors influencing them (e.g., Kotler et al., 1994; Brown and Kotler, 2004; Hochman and Kotler, 2006a), and the significance of energy, nutrients, and predation risk to its fitness (e.g., Tilman, 1982; Kotler et al., 1994; Hochman and Kotler, 2006a). For example, foragers extract more food from safe patches than they do from risky patches (Brown, 1988, 1992; Kotler et al., 1994). Food availability, food preferences, and foraging impacts of animals can be quantified through measurement of patch use and giving-up densities. The giving-up- density (GUD) is the amount or density of food resources left in a food patch when the most efficient forger leaves the resource patch (Brown, 1988; Kotler et al., 1994). When harvest rates in a patch are a diminishing function of food density, GUDs provide an estimate of foraging costs and provide a measure of foraging efficiency and habitat quality. Foragers that experience diminishing returns while exploiting such patches should exploit them until their harvest rates fall to equal their energetic, predation, and missed opportunity costs of foraging (Brown 1988, 1992; Kotler et al., 1994).

88

GUDs can be readily measured for free-ranging animals in their natural environment. This is often done by placing out feeding trays in which food has been mixed with an inedible substrate to produce diminishing returns. GUDs have been measured in this way for animals ranging from desert rodents (e.g., Kotler et al., 2001, 2002) to various ungulates (e.g., Altendorf et al., 2001; Druce, 2005; Hochman and Kotler, 2006a; 2006b), to granivorous birds (e.g., Soobramoney and Perrin, 2008 ), and even to fruit eating bats (e.g., Sanchez, 2006; Sanchez et al., 2008). But sometimes it isn’t practical to use feeding trays because of the feeding manner of the forager, neophobia, or even conservation regulations. In these circumstances, it may be possible to measure natural GUDs. Bite diameters at point of forage can be used to quantify foraging for free-ranging herbivore species such as the Mountain Nyala in this study (Wilson and Kerley, 2003). Such a measure provides a natural GUD (Brown 1988, 1992) by providing a physical record of the amount of fibrous, indigestible materials that the forager is willing to process. Here, the more fibrous the bite, the longer will be the processing time, the less will be the digestible content, and less will be the ingestion rate of energy that should ultimately lead foragers to have smaller bite diameters (i.e., higher GUDs and higher quitting harvest rates). In contrast, the less fibrous the bite, the shorter will be the processing time, the higher will be the digestible content, the higher will be the ingestion rate of energy which encourages foragers to have larger bite diameters (i.e., more food removed and consumed) should correspond with lower GUDs and lower quitting harvest rates. Measurements of bite size and other possible measures of natural GUDs have been carried out for several species. The approach of measuring bite diameters was successfully applied to domestic goats and other wild ungulates in the Eastern Cape of South Africa (e.g., Haschik and Kerley, 1997a, 1997b; Wilson and Kerley, 2003). Shipley et al. (1999) suggest that bite size selection may be influenced by the browsers’ level of hunger or satiation and foraging time available for free-ranging browsers under natural conditions. Similarly, Houle et al. (2006) estimated natural giving-up densities by counting the fruit on the branches of the trees which primates, i.e. Red-tailed monkeys (Cercopithecus ascanius) versus Blue monkeys (Cercopithecus mitis) left behind in Kibale National Park, Uganda. Mountain Nyala are social animals. Mountain Nyala require two kinds of habitat, i.e. foraging grounds and cover in which to hide themselves from predators and shelter

89

from extreme weather (Hillman and Hillman, 1987; Malcolm and Evangelista, 2005). Mountain Nyala are found only in a handful of locations in the highlands of Ethiopia and are critically endangered. Management policy prevents any sort of invasive manipulations or feeding. Other non-invasive approaches are necessary if one is to obtain the information necessary for conservation and management. I hypothesized that Mountain Nyala perceive differences in their habitat quality that can be evaluated and determined by applying foraging theory (e.g., Brown, 1988, 1992; Kotler et al., 1994; Hochman and Kotler, 2006a). This can be practically achieved by measuring bite diameters (e.g., Haschik and Kerley, 1997a, 1997b; Wilson and Kerley, 2003), bite rates (e.g., Owen-Smith, 1994; Neuhaus and Ruckstuhl, 2002), and time budgets (e.g., Hochman and Kotler, 2006b; Tadesse and Kotler, 2011) of free-ranging Mountain Nyala. Bite diameters can be used to reveal the safest habitat for Mountain Nyala. There, where bite diameters will be greatest, I also expect bite rates (number of bites per minute) to be highest, vigilance rates the lowest, and the most time devoted to feeding. In this manner, behavioral indicators can be used to reveal habitat quality and patch use behavior. The objectives of this study were to: (1) measure and quantify habitat-specific bite diameters of free-ranging Mountain Nyala foraging on common natural plant species in two main habitat types (grassland versus woodland), (2) quantify the bite rates (number of bites per minute) of free-ranging Mountain Nyala as functions of habitat type and sex-age categories, and (3) measure the activity time budgets of Mountain Nyala as functions of habitat type, group size, and sex-age categories in the Bale Mountains National Park, Ethiopia.

METHODS Data Collection Techniques Habitat inventory Mountain Nyala provided the study species and is fully described in Chapter 2. All fieldwork in this study was carried out in January 2011. The study site is situated in the northern edge of the Bale Mountains National Park (i.e., Dinsho Sanctuary and Gaysay area), which harbors the largest herd of Mountain Nyala in the world (Refera and Bekele,

90

2004; Mamo, 2007). The study site in the Bale Mountains National Park is thoroughly described in Chapter 2. I conducted a reconnaissance survey to identify major habitat types based on topographic features and vegetation types (Mamo, 2007). Accordingly, I identified two major types of habitat over the study site. These included woodland and grassland habitats. I briefly describe each of these habitat types as follows. The woodland habitat comprises three isolated habitat patches, the Dinsho sanctuary, the Adele Ridge, and the Boditi Ridge. The Dinsho Sanctuary is characterized by dominant plant species such as Juniperus procera, Hagenia abyssinica, Hypericum revolutum, Solanum marginatum, Rosa abyssinica, Carduus nyassanus, and Achyranthes aspera. However, the invasive and unpalatable dumalis is becoming a dominant plant species there. The Adele and the Boditi Ridges are characterized by dominant tree species such as Juniperus procera, Hagenia abyssinica, Hypericum revolutum, and Rapanea simensis. The understory is dominated by Solanum marginatum, Artemisia afra, Senecio ragazzi, Cineraria abyssinica, Carduus nyassanus, Achyranthes aspera, spp., and Rosa abyssinica. The Mountain Nyala mostly use the woodland habitat as source of forage, cover, and shelter from risk of predation and extreme weather (Hillman and Hillman, 1987; Malcolm and Evangelista, 2005; Mamo, 2007). The grassland habitat in the Gaysay Valley is characterized by dominant plant species such as Nepeta azurea, Artemisia afra, Helichrysum spp., spp., Rosa abyssinica, Solanum marginatum, Cineraria abyssinica, and other grass species. Shrubby Hypericum revolutum is also found at the edge of the grasslands. The Mountain Nyala mostly use the grassland habitat as a source of forage (Refera and Bekele, 2004; Malcolm and Evangelista, 2005; Mamo, 2007). However, the grassland habitat is highly overgrazed by livestock around its border (Refera and Bekele, 2004; Malcolm and Evangelista, 2005; Mamo, 2007; personal observation during the fieldwork). Along the altitudinal gradient, I randomly aligned a total of three transects through the three isolated woodland habitat patches (i.e. one transect through each habitat patch). Three more transects were randomly aligned through the grassland habitat in north-south direction. Transects varied in length from 0.7 km to 1 km, i.e. the length of each transect varied with the size of each habitat patch.

91

Bite diameter measurements The intensity of resource patch use provides a measure of habitat quality (Hochman and Kotler; 2006a; Olsson and Molokwu, 2007). However, it is often not feasible to train individuals of target species to use artificial food patches especially in areas where there is high human and livestock disturbances, or conservation regulations do not allow the use of experimental feeding trays. In such cases, using natural measures of patch use would be preferable. Along each transect, free-ranging Mountain Nyala were observed in the field with the aid of binoculars. Focal animal observations were carried out to identify freshly browsed plants. By measuring the twig diameter at points of browsing, I obtained habitat- based twig diameters of bites for common but important natural browse species cropped by free-ranging Mountain Nyala. Following Wilson and Kerley (2003), for each individual animal that foraged during the focal observation, I measured with aid of a Vernier caliper the diameter of the twigs in the bites cropped from each browsed plant species using the diameter at point of browsing (dpb). All bites on a single branch were measured; observations were repeated until sufficient numbers of bites were measured per plant species per habitat type. Moreover, tracks and feeding signs of Mountain Nyala were thoroughly examined along each transect, and the bite diameters of any browsed natural plant species by Mountain Nyala were measured and recorded. Track surveys and plant bites taken by the species may also help to study the habitat selection of a species (e.g., Martinka, 1968; Hansen and Reid, 1975). I treated these bite diameters as habitat-specific natural giving-up densities of free- ranging Mountain Nyala, analogous to artificial giving-up densities that are measured and quantified from feeding trays. The bite diameter measurement was conducted early in the morning from 6:00am to 9:00am and late in the afternoon from 4:00pm to 6:30pm local time when Mountain Nyala were actively foraging in their natural habitat patches.

92

Plate 5.1. A picture showing how Verner caliper is used to measure the bite diameters of Carduus nyassanus plant species browsed by Mountain Nyala in the woodland habitat, Bale Mountains National Park, Ethiopia. (Photo by the author, 2011).

Bite rate quantification Bite rate method was adopted from Owen-Smith (1994) in the study of . Focal individual observations were carried out along each of the six transects to quantify the bite rates (number of bites per minute) of foraging focal Mountain Nyala. To do so, each focal individual (i.e., adult male, adult female, sub-adult, or juvenile) was observed with binoculars at an average sighting distance of approximately 50 meters while the animals were foraging in their natural food patches. I took care to sample Mountain Nyala on different days, in different locations throughout their range, and in different habitats. This reduced the chances of sampling the same individual twice. In addition, I took care to collect bite rates from only a single

93

individual of each age and sex class from a group each day and to collect data in different locations and at different times of the day on different days. The bite rate measurement was carried out early in the morning from 6:00am to 9:00am and late in the afternoon from 4:00pm to 6:30pm local time when Mountain Nyala were actively foraging. A total of 259 focal individual Mountain Nyala were observed, and the number of bites taken by each were separately counted and recorded.

Focal animal observations Following the group scan method proposed by Altmann (1974) and Martin and Bateson (1993), I took field observations on focal individual Mountain Nyala encountered while walking along each of the six transects. In order to minimize any effect caused by the presence of the observer, I waited for the groups of Mountain Nyala calm down for a minimum of 10 minutes before starting the focal observations. Focal observations were conducted with the aid of binoculars. Following Brown (1969b) and Refera and Bekele (2004), I classified focal observations according to sex and age class of the focal individual. The total number of individuals in a group was recorded. Individuals were considered to be in the same group if the separation distance was approximately less than 50 meters (Hillman and Hillman, 1987; Refera and Bekele, 2004). If the group contained a mix of ages and genders, stratified random sampling according to age and gender was employed to select focal individuals. Then I took focal observations of animals representing each category of sex and age. I took care to sample Mountain Nyala on different days, in different locations throughout their range, and in different habitats. This necessarily reduced the chances of sampling the same individual twice. In addition, I took care to collect focal observations from only a single individual of each sex-age class from a group each day and to collect data in different locations each day and to collect data at different times of the day on different days. These spread observations out across sex-age classes, habitats, time, and location as much as was possible for this population. The interspersion of data collection over days, locations, sex-age classes, and habitats make me confident that observations were independent.

94

Following Altman (1974) and Martin and Bateson (1993), I classified the behavioral activities of each focal Mountain Nyala into four types: feeding, vigilance, moving, and resting (for the details, please see Chapter 4). For each focal observation, I noted the type of activity in which the focal Mountain Nyala was engaged at the start of the observation period and recorded the length of time spent in different activities as mentioned above. Focal animal observations were carried out early in the morning from 6:00am to 9:00am and late in the afternoon from 4:00pm to 6:30pm local time (ideal times to sample group size and activity time budgets) when Mountain Nyala were most active. I carried out the focal observations for 10 minutes for each focal individual Mountain Nyala. I recorded the habitat type and time for each focal observation. I observed a total of 106 individual Mountain Nyala.

Data Analysis Measuring bite diameters of free-ranging browsers such as the Mountain Nyala in this study came with logistical and experimental problems that limited the precision of assigning data to particular age and sex classes. This happened when a group of Mountain Nyala containing different sex and age classes all foraged from the branches of a single plant. Consequently, bite diameter measurements were lumped together across age and sex classes for each habitat type and browse species. I used two-way ANOVA to analyze the effects of habitat type (i.e. woodland versus grassland), natural forage plant species, and the interactions of habitat type and natural browse plant species on the twig bite diameters cropped by Mountain Nyala. I used a post- hoc Tukey's HSD test for the multiple comparisons across natural browse plant species. I used two-way ANOVA to analyze the effect of habitat type, sex-age categories, and the interaction of habitat type and sex-age categories on bite rates. I used a post-hoc Tukey's HSD test for the multiple comparisons across sex-age categories. I calculated the proportion of time for each activity type obtained from the focal animal observations. To analyze the general activity time budget patterns of Mountain Nyala, I pooled together data for time allocation by all the four animal categories (i.e. adult males, adult females, sub-adults, and juveniles). I square root transformed the data to ensure normality and to better meet the assumption of ANOVA and used one-way

95

ANOVA to analyze the data. Activity type (i.e., feeding, vigilance, moving, and resting) was a predictor, a proportion of time for each activity type was a dependent variable. I used a post-hoc Tukey's HSD test for the multiple comparisons across activity types. I used Generalized Linear Modeling (GLM) with MANOVA (Multivariate Analysis of Variance) along with univariate statistics to analyze the effect of habitat type, sex-age categories of the focal animal, and the interaction of habitat type and sex-age categories of the focal animal on the behavioral time budgets of Mountain Nyala. The MANOVA approach considered the entire set of dependent and independent variables at the same time in the same analysis. Then, univariate statistics allowed for better interpretation of the multivariate results. In addition, I used linear regression to determine the effect of group size on the vigilance level of Mountain Nyala. For all analyses, I defined the alpha value of 0.05. The analyses were performed by STATSTICA version 10.

RESULTS Bite Diameters To quantify bite diameters (natural giving-up densities), I measured a total of 9,366 twigs from 15 natural browse plant species cropped by free-ranging Mountain Nyala in the Bale Mountains National Park (Table 5.1). This included twig bite diameters from all 15 browse species from the woodland habitat, and all 6 species from grassland habitat (Table 5.1). The 6 grassland species also occurred in the woodland habitat and allowed cross- habitat comparisons.

96

Table 5.1. Mean bite diameters and standard deviations of twigs cropped and consumed by free-ranging Mountain Nyala in the Bale Mountains National Park. N = number of twigs measured. Scientific names Habitat N Mean twig diameter (mm) Standard deviation type Achyranthes aspera Woodland 459 2.54 0.87 Grassland - - - Artemisia afra Woodland 283 2.48 1.02 Grassland 774 3.23 1.07 Asparagus abyssinicus Woodland 113 2.00 1.22 Grassland - - - Carduus nyassanus Woodland 533 3.94 1.48 Grassland - - - Cineraria abyssinica Woodland 266 2.22 0.92 Grassland - - - Cupressus lusitanica Woodland 124 3.20 1.06 Grassland - - - Woodland 308 2.25 0.87 Grassland 396 1.67 0.72 Hypericum revolutum Woodland 1227 2.64 1.11 Grassland 972 3.02 1.28 procera Woodland 456 2.95 0.98 Grassland - - - Maytenus spp. Woodland 278 2.93 0.98 Grassland - - - Nepeta azurea Woodland 215 2.18 1.03 Grassland 362 2.20 0.88 Rapanea simensis Woodland 522 6.95 1.69 Grassland - - - Rosa abyssinica Woodland 476 3.65 1.73 Grassland 437 3.34 1.42 Senecio ragazzi Woodland 134 3.00 1.24 Grassland - - - Solanum marginatum Woodland 745 5.97 1.41 Grassland 286 6.52 4.70

Habitat type significantly affected bite diameter (F (1, 6469) = 9.79; p = 0.0018) for the 6 common browse species cropped by Mountain Nyala (Fig. 5.1). Independent of browse species, Mountain Nyala cropped browse plants to larger dpb in the grassland than in the woodland (Fig. 5.1).

97

3.45

3.40

3.35

3.30

3.25

3.20

3.15

Diameter point of browsing (mm) Diameter of browsing point

3.10 Grassland Woodland Habitat type

Figure 5.1. The effect of habitat type on the diameters of twigs browsed by free-ranging Mountain Nyala in the Bale Mountains National Park (BMNP). Twig diameters of Artemisia afra, Helichrysum splendidum, Hypericum revolutum, Nepeta azurea, Rosa abyssinica, and Solanum marginatum cropped by Mountain Nyala in each habitat type were separately pooled and included in the analysis. The error bars represent ± 1 stdev.

Mountain Nyala browsed the 6 species common to both grasslands and woodlands to different dpb (F (5, 6469) = 761.42; p < 0.001), leaving Solanum marginatum (mean = 6.12mm) at the largest stem diameter (Fig. 5.2). A post-hoc Tukey's HSD test for multiple comparisons across browsed species showed a significant difference in dpb between Solanum marginatum versus Rosa abyssinica, Rosa abyssinica versus Artemisia afra, and Hypericum revolutum versus Nepeta azurea (Fig. 5.2).

98

Figure 5. 2. The effect of natural forage plant species on the diameters of twigs browsed by free-ranging Mountain Nyala in the BMNP. The twig diameters of common natural browse species measured in the two habitat types (i.e. grassland and woodland) were lumped together for each species type in the analysis. The error bars represent ± 1 stdev. Bars labeled with different letters significantly differed.

Furthermore, habitat type and natural browse species interacted (F (5, 6469) = 23.73; p < 0.001), with Mountain Nyala showing higher dpb in the grassland for only half of the browse species (Fig. 5.3).

99

12

11

10

9

8

7 Woodland habitat 6 Grassland habitat 5

4

3 Diameter point of browsing (mm) browsing of point Diameter 2

1

0

Artemisia afra Nepeta azurea Rosa abyssinica

Hypericum revolutum Solanum marginatum Helichrysum splendidum Browsed plant species

Figure 5.3. The effect of habitat type and natural browse plant species on the diameters of twigs browsed by free-ranging Mountain Nyala in BMNP. Twig diameters of common natural forage plant species cropped by Mountain Nyala in each habitat type were separately included in the analysis. The error bars represent +1 stdev.

Bite Rates

Habitat type significantly affected (F (1,251) = 98.26; p < 0.001) the bite rates for Mountain Nyala, with higher bite rates in the grassland habitat (mean = 17.72 bites per minute) than in the woodland habitat (mean = 14.86 bites per minute) (Fig. 5.4).

100

25

20

15

10

5

Number of bites per minute per bites of Number 0 Grassland Woodland Habitat type

Figure 5.4. The effect of habitat type on the bite rates of free-ranging Mountain Nyala in BMNP pooled over bite rates by four categories, adult males, adult females, sub-adults, and juveniles. The number of focal individuals included in this analysis was 127 for the grassland habitat and 132 for the woodland habitat. The error bars represent +1 stdev.

Bite rates differed significantly according to sex and age class (F (3,251) = 83.31; p < 0.001) with adult females taking the most bites per minute (mean = 19.03 bites per minute), followed by juveniles (mean = 17.14 bites per minute), then sub-adults (mean = 15.16 bites per minute), and finally adult males (mean = 13.11 bites per minute). A post- hoc Tukey's HSD test for multiple comparisons across sex-age classes showed a significant difference in bite rates between adult females versus sub-adults and juveniles versus adult males. However, the bite rates comparison between adult females versus juveniles and sub- adults versus adult males did not significantly differ.

Habitat type and sex-age class interacted significantly (F (3, 251) = 3.11; p = 0.027), with greater differences among classes occurring in the grasslands and the differences between grassland and woodland being greater for adults than for sub-adults and juveniles (Fig. 5.5).

101

25

20 Adult male 15 Adult female 10 Sub-adult Juvenile 5

Number of bites per minute per bites of Number 0 Grassland Woodland Habitat type

Figure 5.5. The effect of habitat type and sex-age categories on the bite rates of free- ranging Mountain Nyala in BMNP. The number of focal individual Mountain Nyala considered for this analysis was 21 adult males, 38 adult females, 41 sub-adults, and 28 juveniles for the grassland habitat; and 31 adult males, 34 adult females, 35 sub-adults, and 31 juveniles for the woodland habitat. The error bars represent +1 stdev.

Focal Animal Observations To quantify the general time budget patterns, I assessed a total of 106 focal individuals. Mountain Nyala differed in the proportion of time they devoted to the different activities, (F (3,420) = 27.69; p < 0.001). They spent most of their time moving (mean ≈ 35.55%), followed by time feeding (mean ≈ 34.15%), vigilance (mean ≈ 22.28%), and resting (mean ≈ 8.02%) (Fig. 5.6). A post-hoc Tukey's HSD test for multiple comparisons across activity types showed a significant difference in time budgets between moving versus resting, and feeding versus resting (Fig. 5.6).

102

a 80 a 70 ab 60 50 b 40 30

20 spent time Percent 10 0 Vigilance Feeding Resting Moving Activity type

Figure 5.6. General activity time budget pattern of Mountain Nyala in BMNP pooled over four categories, adult males, adult females, sub-adults, and juveniles. The number of focal individuals (N) included in this analysis was 106. The error bars represent +1 stdev. Bars labeled with different letters significantly differed.

Habitat type significantly affected most behaviors of Mountain Nyala (MANOVA:

Pillai’s Trace value = 0.103; F (3,96) = 3.68; p = 0.015), including the proportion of time vigilant (ANOVA: F (1, 98) = 8.12; p = 0.005) (Fig. 5.7) and feeding (ANOVA: F (1, 98) =

8.21; p = 0.005), but not proportion of time moving (ANOVA: F (1, 98) = 0.01; p = 0.931). Accordingly, Mountain Nyala showed higher proportion of time vigilant in the woodland habitat (mean = 0.281) than in the grassland habitat (mean = 0.140) (Fig. 5.7). In contrast, Mountain Nyala had higher proportion of time feeding in the grassland habitat (mean = 0.435) than in the woodland habitat (mean = 0.267).

103

0.6 0.5 0.4 0.3

vigilant 0.2 0.1

0 Proportion of time time of Proportion Grassland Woodland Habitat type

Figure 5.7. The effect of habitat type on the vigilance level of Mountain Nyala in the Bale Mountains National Park pooled over four categories, adult males, adult females, sub- adults, and juveniles. The number of focal individuals included in this analysis was 47 for the grassland habitat and 59 for the woodland habitat. The error bars represent +1 stdev.

Age and sex categories of the focal animal did not affect any behaviors of

Mountain Nyala (MANOVA: Pillai’s Trace value = 0.108; F (9,294) = 1.22; p = 0.279), including the proportion of time vigilant (ANOVA: F (3, 98) = 0.74; p = 0.529) (Fig. 5.8), feeding (ANOVA: F (3, 98) = 0.71; p = 0. 547), and moving (ANOVA: F (3, 98) = 0.56 ; p = 0.644). Habitat type and sex-age categories did not interact to affect any behaviors of

Mountain Nyala (MANOVA: Pillai’s Trace value = 0.056; F(9,294) = 0.63; p = 0.775), including proportion of time vigilant (ANOVA: F(3, 98) = 0.33; p = 0.801) (Fig. 5.8), feeding (ANOVA: F(3, 98) = 0.76; p = 0.518), or moving (ANOVA: F(3, 98) = 0.01; p = 0.998).

104

0.7 0.6 0.5 Adult male 0.4 Adult female 0.3 Sub-adult 0.2 Juvenile 0.1

Proportion of time vigilant time of Proportion 0 Grassland Woodland Habitat type

Figure 5.8. The effect of habitat type and sex-age categories on the proportion of time vigilant by free-ranging Mountain Nyala in BMNP. The number of focal individual Mountain Nyala considered for this analysis was 10 adult males, 20 adult females, 8 sub- adults, and 9 juveniles for the grassland habitat, and 15 adult males, 25 adult females, 9 sub-adults, and 10 juveniles for the woodland habitat. The error bars represent +1 stdev.

Group size did not have a significant effect on the vigilance level of Mountain Nyala in the Bale Mountain National Park (Fig. 5.9).

1.2 y = 0.0101x + 0.1628 1 p = 0.1108; r2 = 0.0243

0.8

0.6

0.4

0.2 Proportion of timeProportion vigilant

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Group size

Figure 5.9. The effect of group size on the vigilance level of Mountain Nyala in the Bale Mountains National Park.

105

DISCUSSION Food is one of the most important habitat essentials determining the distribution and abundance of individuals of any species in a given habitat (Spellerberg, 1992; Houle et al., 2006; Tadesse and Kotler, 2010). Free-ranging herbivores constantly make foraging that affects their fitness. Variation in qualitative and quantitative food demands and the characteristics of the available food are the main reasons for the necessity of these foraging decisions (Stephens and Krebs, 1986). In addition, food resources in the natural environments are distributed in patches of variable size and shape with fluctuating quality, quantity, and safety (e.g., Brown, 1999; Houle et al., 2007). Based on foraging theory, I studied the habitat quality and habitat selection behaviors of the critically endangered Mountain Nyala (IUCN, 2008) using non-invasive natural behavioral indicators. The study was carried out in two major habitats (grassland versus woodland) of the Bale Mountains National Park which harbors the largest herd of the unique Mountain Nyala in the world (Evangelista et al., 2007; Mamo, 2007). I collected the field data through measurement of twig bite diameters combined with quantification of bite rates and time budgets of focal Mountain Nyala stratified according to their sex-age categories and habitat type. The study revealed that Mountain Nyala feed on a variety of vegetation species, especially in the woodland habitat. However, lower natural GUDs revealed by higher dpb in the grassland habitat make it more selected by Mountain Nyala. In addition, higher bite rates, higher proportion of time feeding, and lower proportion of time vigilant in the grassland habitat make it more attractive to Mountain Nyala, probably because it is safer. In this way, behavioral indicators can be used to assess habitat selection and habitat quality of animals in their natural environment even when invasive or experimental techniques are not available to researchers. The details of the findings are discussed as follows. The present study revealed that Mountain Nyala forage on a wide range of natural browse species to meet their nutrient and energy requirements in the Bale Mountains National Park. However, the diversity of natural browse species cropped by Mountain Nyala varied with habitat types. For example, 15 common natural browse species were foraged by Mountain Nyala in the woodland habitat patches (i.e., Dinsho Sanctuary, Adele, and Boditi ridges). In addition, 6 of these browse species were also cropped by Mountain

106

Nyala in the Gaysay grassland habitat. This suggests that food preference and foraging behavior of Mountain Nyala may vary with food availability, nutritive quality, and predation risk characteristics of the habitat (Evangelista et al., 2007; Mamo, 2007). Houle et al. (2007) also noted that fruit quantity and quality on the branches of trees affect the food preference and foraging behavior of frugivorous primates in the Kibale National Park, Uganda. The 15 browsed plant species, their growth forms, occurrence of habitat type, and parts of the plant cropped by Mountain Nyala are shown in Appendix I - Table 5.1. Through focal animal observations, the diameters of twigs of common natural browse species cropped by free-ranging Mountain Nyala were assessed, identified, and measured along the six transects used in this study (three transects over each habitat type). The result revealed that bite diameters of free-ranging Mountain Nyala were significantly affected by habitat type. Generally, bite diameters of common natural browse species cropped by Mountain Nyala were higher in the grassland habitat than in the woodland habitat. This showed that foraging costs differ between the two habitat types and suggests that the differences are most likely due to difference in the risk of predation between the two habitats. Evangelista et al. (2007) noted that Mountain Nyala use open sightlines to detect and locate approaching predators. So, the openness of the Gaysay grassland may favor the foraging Mountain Nyala by allowing them to more easily locate predators, leading to lower foraging costs arising from the risk of predation and higher quitting harvest rates, resulting in larger overall bite diameters. Other studies also support my findings. For example, Risenhoover and Baily (1985) reported that Bighorn sheep ( canadensis) prefer to feed and harvest more food in open areas where they can easily detect approaching predators. In another instance, Kotler et al. (1994) and Hochman and Kotler (2006b) noted that Nubian Ibex (Capra nubiana) mostly prefer to feed in open areas which are situated on the cliff or near the cliff as this strategy was found to help them easily locate and escape from dangerous predators. Furthermore, blocking sightlines cause ibex to increase their GUDs and alter their vigilance behavior (Iribarren and Kotler, unpublished data). The present study showed that browse species also affects the bite diameters of free-ranging Mountain Nyala. I compared the twig diameters of six browse species which were commonly cropped by Mountain Nyala in both woodland and grassland habitats,

107

namely Solanum marginatum, Artemisia afra, Helichrysum splendidum, Hypericum revolutum, Nepeta azurea, and Rosa abyssinica. The greatest mean bite diameter was measured and quantified for Solanum marginatum. Mountain Nyala commonly forage on the fruits and young shoots of Solanum marginatum. Assefa (2003) also noted that Mountain Nyala is the main seed dispersal agent for Solanum marginatum in the Bale Mountains National Park, suggesting that there is a strong association between Mountain Nyala and Solanum marginatum where the former is the main browser and the latter is a good browse species. Another plant frequently browsed by Mountain Nyala is Hypericum revolutum. This shrubby species is especially found in the ecotone between the woodland and the grassland (Mamo, 2007) where it is heavily exploited by Mountain Nyala (Hillman and Hillman, 1987; Mamo, 2007). This makes the ecotone especially attractive to Mountain Nyala. I observed in many instances that Mountain Nyala foraged on flowers, twigs, and leaves of Hypericum revolutum. This suggests that vegetation type also influences the habitat selection behavior of Mountain Nyala. In places where Hypericum revolutum is not available, I mostly observed Mountain Nyala to forage on Solanum marginatum, Artemisia afra, Helichrysum splendidum, Nepeta azurea, Rosa abyssinica, and other palatable browse species (see also Appendix I - Table 5.1). I frequently observed Mountain Nyala feeding on the fallen leaves of the tree species Hagenia abyssinica. However, the twigs on the branches of this species occur high out of reach of Mountain Nyala, so, I was not able to include Hagenia abyssinica in my assessment of natural GUDs. I also observed Mountain Nyala feeding on other plant species. These included herbs such as Achyranthes aspera and Carduus nyassanus that were fed upon after other palatable browse species had been depleted in the woodland habitat patches. It also includes woody species such as Juniperus procera, Rapanea simensis, and Cupressus lusitanica that are exploited in the extended dry season when other preferable natural browse species are scarce in the woodland habitat patches (discussion held with some wildlife professionals in the park). Moreover, I observed in many instances that the Mountain Nyala browse on Achyranthes aspera and Carduus nyassanus in the woodland Dinsho Sanctuary where the population density of Mountain Nyala is especially high. However, the expansion of the invasive but inedible Euphorbia dumalis shrub in the

108

Dinsho Sanctuary has reduced the availability of natural browse species there (Mamo, 2007; personal observation and discussion held with some wildlife professionals in the park). The present study revealed that bite rates (number of bites per minute) by free- ranging Mountain Nyala were significantly affected by habitat type. Regardless of sex-age categories, bite rates were higher in the grassland habitat than in the woodland habitat. A possible explanation for this finding is that Mountain Nyala use open sightlines to detect approaching predators, including humans (e.g., Evangelista et al., 2007). As a result, the openness of the grassland habitat allowed Mountain Nyala to devote more of their time for feeding and less to vigilance, resulting in higher bite rates. The data from the focal animal observation of the present study supported this result in which Mountain Nyala were more vigilant in the woodland than in the grassland habitat. The finding of this study further revealed that sex-age categories significantly affected bite rates. Accordingly, adult females had the highest number of bites per minute compared with adult males and other age groups (i.e., sub-adults and juveniles). This suggests that high nutritional demand of reproductive and lactation could be reflected in the bite rates of adult female Mountain Nyala. Reid (2005) also reported that adult female (Antidorcas marsupialis) in Augrabies National Park devoted a greater proportion of their time feeding than adult male and other age groups. It is well- documented that reproductive state has a significant effect on the nutritional requirements, with reproductive individuals having higher nutritional needs than non-reproductive males (e.g., Beier, 1987; Young and Isbell, 1991; Ginnett and Demment, 1997). Neuhaus and Ruckstuhl (2002) in their field study on the foraging behavior of free-ranging plains zebra (Equus burchelli) also reported that lactating females take more bites per minute when foraging than either pregnant or non-reproducing females or males. Savannah yellow baboons (Papio cynocephalus) that feed on higher quality and quantity of food during pregnancy and lactation gain greater reproductive success (Altmann, 1998). The body size of Mountain Nyala may also have played a role in determining bite rates. Adult male Mountain Nyala have larger mouths and should have correspondingly greater bite mass than do the adult females. Consequently, adult females should need a greater number of bites than the adult males to achieve the same food intake rate. Ginnett

109

and Demment (1997) in their field study on the foraging behavior of free-ranging giraffes in Tanzania reported that males took larger bites than females, but females cropped bites more quickly and chewed faster. Males had longer per-bite handling times than females so that males have lower bite rates than females (Ginnett and Demment, 1997). Regarding the effect of age on bite rates of Mountain Nyala, next to adult females, the result revealed that juveniles have the highest bite rates. This may be a consequence of their high nutrient and energy demand to maintain elevated metabolic costs due to rapid body growth (e.g., Harrison, 1983; Padmalal and Takatsuki, 1994). In addition, the small mouth size of juvenile Mountain Nyala may also lead to smaller bite mass. In order to compensate the low intake rate, the juveniles would then need higher bite rates. Regardless of habitat type, Mountain Nyala in the Bale Mountains National Park allocated the largest portion of their time budget to moving. The large time allocation to movement by Mountain Nyala could be due to the shortage of quality forage during the dry season in the park. This should lead to high roaming behavior in Mountain Nyala in quest for food and nutrients. Previous field reports also suggest that the home range of free- ranging Mountain Nyala becomes larger during the dry season (e.g., Ethiopian Wildlife Conservation Authority, unpublished data), which likely results in greater movement. Habitat type significantly affected the vigilance level of Mountain Nyala. Accordingly, Mountain Nyala were more vigilant in the woodland habitat than in the grassland habitat. The possible explanation for this outcome is that Mountain Nyala use open sightlines to more easily detect and escape from approaching predators, including humans (Evangelista et al., 2007). The openness of the grassland may make vigilance especially effective in reducing mortality risk, and therefore animals may require little for safety (Brown, 1999). Consequently, Mountain Nyala have more time to allocate to foraging and so devote a greater proportion of time to foraging in the grasslands which is also supported by the present focal animal observation data. Group size often helps determine vigilance levels in ungulates (e.g., Lagory, 1986; Lima and Dill, 1990; Quenette, 1990; Roberts, 1996; Hunter and Skinner, 1998; Manor and Saltz, 2003), but does not appear to do so in Mountain Nyala. This suggests that group size does not lead to safety for Mountain Nyala, either because larger groups are more easily encountered by predators or possibly because individual information is more

110

important than public information in regards to risk of predation or human nuisance disturbances. In addition, the absence of the group size effect could be due to the fact that for females dependent kids cannot act as dilution agents (Rieucau and Martin, 2008). Consequently, these females may rely mostly on predator detection rather than on the numerical dilution of risk to improve safety, thus masking the classical group size effect (e.g., Tadesse and Kotler, 2011). In regards to minimizing risk of predation or human nuisance disturbances, Dall and Valone (2005) noted that individual information is more valuable and reliable than group information in some social free-ranging animals. When personal information is valued more than social information (Lima et. al., 1985), constant levels of vigilance regardless of group size is a more likely outcome. In summary, Mountain Nyala browse on a wide diversity of woody vegetation, especially in the woodland habitat. Nonetheless, lower GUDs revealed by larger dpb, higher bite rates, higher proportion of time feeding, and lower vigilance level in the grassland habitat make it more selected than the woodland, probably because it is safer. The results of the present study predicted the habitat selection, foraging efficiency, and foraging impacts of free-ranging Mountain Nyala in the Bale Mountains National Park using non-invasive techniques. Houle et al. (2006) similarly used natural giving-up densities in fruit trees to measure foraging efficiency and to evaluate possible mechanisms of coexistence between two different frugivorous primate species, i.e. Red-tailed monkeys versus Blue monkeys in Kibale National Park, Uganda. Likewise, the results from bite diameters represent habitat specific natural giving-up densities of free-ranging Mountain Nyala, analogous to artificial giving-up densities are measured and quantified from feeding trays. In effect, the combination of the results obtained from bite diameters, bite rates, and time budgets could indicate the habitat quality of free-ranging Mountain Nyala through reflecting the habitat characteristics.

111

CHAPTER 6 IMPACTS OF HUMANS AND LIVESTOCK ENCROACHMENTS ON THE HABITATS OF MOUNTAIN NYALA IN MUNESSA ABSTRACT Measuring the extent of humans and livestock encroachments on the habitats of the endangered Mountain Nyala is crucial to ensuring effective conservation, but empirical evidence is lacking. In order to examine the risk-disturbances hypothesis, I assessed and quantified the impacts of humans and livestock encroachments on the habitats utilized by Mountain Nyala in Munessa, Ethiopia. I estimated the activity density of livestock along transects aligned through three major habitat types utilized by Mountain Nyala. In addition, stem damage, crown damage, evidence of wood use, number of stumps cut, sign of habitat use by livestock, and level of grazing by livestock were quantified on each transect. I collected the field data in the wet and the dry seasons. Activity density of livestock was significantly highest in the natural forest habitat during the wet season. Regarding the impacts of humans on the habitats of Mountain Nyala, both in wet and dry seasons: stem and crown damages were significantly highest in the plantation habitat, the evidence of wood use and the number of stumps cut by humans were significantly highest in the natural forest habitat. Sign of habitat use by livestock did not differ among habitat types -- rather it was dispersed throughout all habitats. Overall, sign of habitat use by livestock was significantly higher in the wet season. The intensity of livestock grazing / browsing was significantly heaviest in the natural forest habitat during the wet season; however, it did not differ among habitats during the dry season. The study revealed that humans and livestock encroachments forced and excluded the Mountain Nyala from using their optimum habitats in Munessa. These results have important management implications for the critically endangered Mountain Nyala in Ethiopia. Proper maintenance and management of the Munessa Forest enhances the availability and quality of habitats for Mountain Nyala. In addition, zonation of the core habitats of Mountain Nyala may reduce and / or avoid humans and livestock disturbances. Key words: behaviors, habitat use, humans encroachments, livestock activity density, livestock grazing / browsing, Mountain Nyala, Munessa, risk-disturbances hypothesis

112

INTRODUCTION Humans have induced many severe and irreversible changes to the natural environment and have most probably had the biggest influence on the distribution patterns of almost all wildlife species on the surface of the earth (Saunders and Hobbs, 1991; Brooks and Balmford, 1996; Oates, 1999; Mace and Balmford, 2000; Morris and Kingston, 2002). Human activities such as hunting and poaching, wildlife trade, and habitat alteration and fragmentation further increase species vulnerability to extinctions (Morris et al., 2009). The extinction of species as a result of the negative impacts of human activities is estimated to occur more than 100 times faster than the natural rate of extinction and more rapidly than new species can evolve (Primack, 2002). These effects are often manifested through effects on population size and on animal behavior. Small population size increases vulnerability to extinction especially when human disturbances increase (Primack, 2002). Habitat loss and fragmentation affect the survival of wildlife species in various ways including influencing animal behavior (e.g., Nour, et al., 1997; Moenting and Morris, 2006; Ukizintambara, 2008), reducing of the total amount of usable habitats, degrading habitat quality, and creating edge effects (e.g., Moenting and Morris, 2006; Evangelista et al., 2007; Atickem et al., 2011). The need for cultivation and grazing lands, settlement, charcoal production, commercial wood, and construction materials (e.g., Tedla, 1995; Hundessa, 1997; Evangelista et al., 2007; Mamo, 2007) has contributed much to the reduction of forest cover in Ethiopia. Deforestation and uncontrolled burning of vegetation is causing climate change, alteration of hydrological processes, soil erosion, and degradation of biodiversity and impoverishment of unique ecosystems (Teketay, 1992; EFAP, 1994; Berry, 2003; Evangelista et al., 2007; Kebede, 2009). In this way, the mis-utilization of wood products by the rural human communities contributes much to the degradation of forests and wildlife habitats in Ethiopia (Hundessa, 1992; Tedla, 1995; Evangelista et al., 2007; Mamo, 2007; Kebede, 2009). Habitat loss and fragmentation caused by human impacts is by far the largest threat to the vast majority of Ethiopia’s wildlife species, including the endangered Mountain Nyala (Hillman, 1986, 1988, 1993; Hundessa, 1992, 1997; Tedla, 1995; Evangelista et al., 2007; Evangelista et al., 2008). Such threats are widely documented

113

(Hundessa, 1992, 1997; Tedla, 1995; Brooks and Balmford, 1996; Forman, 1996). For example, since Mountain Nyala was discovered by western science in 1908, its range has probably shrunk 10-20% due to deforestation and habitat destruction (Malcolm and Evangelista 2005). In addition, human populations through settlement and expanded agricultural activities have considerably reduced the availability and quality of habitats for Mountain Nyala (Evangelista et al., 2007; Mamo, 2007). As a result, Mountain Nyala are confined to “sky islands” on a handful of peaks and ridges (Evangelista et al., 2007; Sillero-Zubiri, 2008). Illegal hunting of wild animals for consumptive and commercial purposes has helped to cause severe population decreases and even the extirpation of several wild animal species in Ethiopia (Hundessa, 1992, 1997; Hillman, 1993; Tedla, 1995; Evangelista et al., 2007). For example, following the fall of the Dergue regime in 1991, civil unrest from years of government oppression resulted in a dramatic rise of wildlife persecution and massive environmental destruction in the country (Hundessa 1992, 1997; Hillman, 1993; Tedla, 1995; Woldegebriel, 1996; Evangelista et al., 2007). At that time, there was little control over illegal human activities in most wildlife protected areas of Ethiopia (Hillman, 1993; Tedla, 1995; Hundessa, 1997; Evangelista et al., 2007). As a result, illegal hunting posed heavy burdens on wildlife species (Hundessa, 1992, 1997; Hillman, 1993; Tedla, 1995; Evangelista et al., 2007; Mamo, 2007). In addition, bush meat is often sought to avoid the slaughter of livestock. Intensive traditional hunting of Mountain Nyala involving massive hunting parties of local people has been regularly organized using horseback, spears, and dogs in the Arsi and the Bale regions including Munessa (Woldegebriel, 1996; Evangelista et al., 2007). Free-range livestock grazing has strong negative impacts on native wild herbivores, their habitats, and overall ecosystem function and structure (Mishra and Rawat, 1998; Morrison et al., 1998; Mamo, 2007; Kebede, 2009). Livestock usually intensively compete with wild animal species for different habitat resources including forage, water sources, and space (Hansen and Reid 1975; Augustine, 2004; Evangelista et al., 2007; Mamo, 2007; Kebede, 2009). The problem can be exasperated by the local people’s consideration of excess numbers of livestock as generating prestige in Ethiopia (Hillman, 1993; Tedla, 1995; Hundessa, 1997; Mamo, 2007; Kebede, 2009). The

114

resultant high livestock population densities put strains on the habitats of many wild animal species (Hundessa, 1992, 1997; Tedla, 1995; Mamo, 2007; Kebede, 2009). For example, when livestock population densities increase beyond the carrying capacity of the natural environment, they degrade the habitat and make it unsuitable for the prevailing wild herbivores (Stephens et al., 2001; Mamo, 2007; Kebede, 2009). This process is now affecting the higher elevations since agriculture is dominantly practiced in lower elevations in the highlands of Ethiopia, and livestock often move to higher elevations for free-range grazing (Stephens et al., 2001; Evangelista et al., 2007). Livestock grazing often alter the structure and species composition of natural vegetation (Morton, 1990; Kebede, 2009). For example, in sites where livestock grazing is more extensive, the vegetation types and composition change from more diverse and suitable to less diverse and unpalatable (Morton, 1990; Stephens et al., 2001; Mamo, 2007; Kebede, 2009). By selectively reducing palatable plant species, livestock grazing ultimately results in the proliferation of invasive and unpalatable plant species for wild herbivores (Stephens et al., 2001; Mamo, 2007; Kebede, 2009). For example, the unpalatable shrubby Euphorbia dumalis is now much more common in and around the Dinsho Sanctuary of Bale Mountains National Park as a result of livestock grazing (Mamo, 2007). Grazing pressure also puts an enormous strain on the fragile ecosystems as a whole and causes severe soil erosion (Wesche et al., 2000; Kebede, 2009). Close contact between livestock and wild animal species inevitably results in various related problems such as increased predation of wild animals by guard dogs (e.g., Woldegebriel, 1996) and loss of genetic viability through hybridization (Sillero-Zubiri et al., 1996; Stephens et al., 2001; Evangelista et al., 2007). In addition, livestock may act as reservoir hosts and vectors (Hudson et al., 2002) for diseases that can also affect wildlife (Jessup et al., 1995; Sillero-Zubiri et al., 1996; Stephens et al., 2001), and raise the risk of extinction of endangered wildlife species (e.g., Saunders and Hobbs, 1991; Sillero-Zubiri et al., 1996). As a result, management plans typically recommend buffer zones to separate endangered wildlife species from domestic livestock that carry potentially threatening infectious disease (e.g., Jessup et al., 1995; Sillero-Zubiri et al., 1996). Humans and livestock disturbances modify the natural behaviors of wild animals. For example, the continuous humans and livestock disturbances in Dinsho and Gaysay

115

areas of Bale Mountains National Park resulted in the modification of the natural behaviors of Mountain Nyala (Stephens et al., 2001; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). According to the risk-disturbance hypothesis, the effect of humans and livestock disturbances on wild animals is frequently measured in terms of shifts in habitat use and other behavioral responses (Gill et al., 2001; Frid and Dill, 2002; Gilroy and Sutherland, 2007) including reduction in the quality and availability of habitats (Mamo, 2007). Understanding the negative impacts of humans and livestock encroachments on the habitat quality and availability for Mountain Nyala can inform decision makers and also promote conservation efforts to mitigate the conflicts between human needs and the conservation of Mountain Nyala. I examined the risk-disturbance hypothesis by quantifying the activity densities of livestock, and also by assessing the number and intensity of disturbances caused by humans and livestock in habitats utilized by Mountain Nyala in Munessa, Ethiopia.

METHODS Livestock Activity Densities In this study, I used the same survey routes used to carry out the habitat inventory and population censusing of Mountain Nyala (see also Chapter 3). Livestock densities are often used as indicators of grazing pressure on habitats (e.g., Mamo, 2007; Kebede, 2009) used by Mountain Nyala, so I quantified the presence of livestock. Along each line transect over the three habitat types of natural forest, plantation, and clear cut (see also Chapter 3 for the details), I estimated activity densities of livestock (i.e., numbers of cattle, horses, donkeys, sheep, and goats) using similar techniques to those applied to estimate the activity densities of Mountain Nyala. Accordingly, each transect was assessed four times in the wet season and six times in the dry season.

Humans and Livestock Disturbances on the Habitats of Mountain Nyala in Munessa Circular plots (each with a radius of 5 meters) that were used to collect data on local habitat variables (see also the habitat variable data collection methods in Chapter 3) were inspected for the extent of humans’ disturbances and livestock grazing/browsing

116

effects. Following Silori and Mishra (2001) and Mamo (2007), variables including stem damage (i.e., presence or absence of peeled and/or damaged bark of trees and shrubs caused by humans), crown damage (i.e., branches removed or broken by humans), evidence of wood use (i.e., presence or absence of any leftover materials after wood collection by humans), number of cut stumps, and sign of livestock (i.e., presence or absence of livestock droppings and/or foot prints) were quantified and recorded in each plot laid along each transect over the three habitat types. A total of 109 plots were assessed (31 plots in the clear cut habitat, 41 plots in the plantation habitat, and 37 plots in the natural forest habitat). Following Smit et al. (2006) and Mamo (2007), the level of browsing / grazing by livestock was inspected and quantified in each nested circular plot and categorized either as: no evidence of browsing / grazing (0% browsed / grazed), lightly browsed (1-25% browsed / grazed), moderately browsed / grazed (26-50% browsed / grazed), or heavily browsed / grazed (>50% browsed / grazed). The field data were collected in two separate seasons: June through August 2010 and December 2010 through February 2011 for the wet and dry season respectively.

Data Analysis I incorporated the transect sample area information from Chapter 3 (Table 3.1) (please see Chapter 3 for the details) with the livestock census data obtained from transect walks during the wet and the dry seasons. As grazing / browsing pressures may vary with the livestock type (i.e., cattle, donkey, horse, sheep, and/or goat), livestock densities were calculated and converted into Adult Cattle Units (ACU), as an indicator of grazing pressure on the study area. ACU conversion followed Silori and Mishra (2001), where 1 adult cow or bull =1 ACU; 1 calf = 0.5ACU; 1 adult horse or donkey = 1.5 ACU and 1 sheep or goat = 0.5 ACU. In this way, I estimated the densities of livestock in each habitat type for the wet and the dry seasons as shown in Table 6.1. As the data were composed of both categorical independent and continuous dependent variables, I used ANOVA to test for whether livestock were exhibiting a choice in their habitat use (Snedecor and Cochran, 1989) during the wet and the dry seasons in Munessa, with habitat type as a predictor and activity density of livestock as the dependent variable. Humans and livestock disturbances

117

on the habitats of Mountain Nyala were also analyzed using ANOVA. For all analyses, I defined the alpha value of 0.05. I performed the analysis using STATSTICA version 10.

RESULTS Livestock Activity Densities In the wet season, the activity density of livestock in the natural forest habitat was greater than for all other habitats (F (2, 45) = 13.99; p < 0.001; Fig. 6.1). A post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in activity density between natural forest versus clear cut and natural forest versus plantation habitats (Fig. 6.1). Accordingly, livestock were seen in the natural forest in every transect survey (≈100.00%) and achieved their greatest maximum activity density there (11.77 livestock/ha; Table 6.1). In addition, livestock achieved their highest mean activity density in the natural forest (5.80 livestock/ha), followed by plantation (2.85 livestock/ha), and clear cut habitat (2.68 livestock/ha) (Table 6.1). In contrast, during the dry season, livestock did not show any significant difference in their habitat selection behavior (F (2, 69)

= 1.36; p = 0.264). Overall, livestock activity density was significantly higher (F (1, 114) = 18.27; p < 0.001) in the wet season than in the dry season.

Table 6.1. A summary of the activity density of livestock in different habitat types during the wet and the dry season field surveys: minimum, mean, and maximum activity densities. Data are given in Adult Cattle Units (ACU) to account for different species of livestock. N = number of transects sampled that habitat type.

Minimum Maximum Number of Sightings per Season Habitat type N Density Mean Density Density livestock transect walk (Livestock/ha) (Livestock/ha) (Livestock/ha) sightings (%) Clear cut 16 1.28 2.68 4.82 16 100.00 Plantation 16 1.75 2.85 6.31 16 100.00 Wet Natural forest 16 1.63 5.80 11.77 16 100.00 Total 48 4.66 11.33 22.29 48 100.00 Clear cut 24 0.01 2.03 6.76 24 100.00 Plantation 24 0.00 1.70 4.05 23 95.83 Dry Natural forest 24 0.46 2.37 5.96 24 100.00 Total 72 0.47 6.10 16.77 71 98.61

118

10

8

6 Clear cut Plantation 4 Natural forest

2 Density (Livestock/ha) Density 0 Wet Dry Season

Figure 6.1. Seasonal habitat use by livestock in Munessa. The error bars represent +1 stdev.

Humans and Livestock Disturbances on the Habitats of Mountain Nyala in Munessa In the wet season, the extent of stem damage by humans differed significantly among habitats (F (2, 106) = 15.82; p < 0.001). Accordingly, it was highest in the plantation (mean = 59%), followed by the natural forest (mean = 28%), but nonexistent in the clear cut habitat (mean = 0%) (Fig. 6.2). Similarly, in the dry season stem damage differed significantly among habitats (F (2, 106) = 10.38; p < 0.001), with highest stem damage in the plantation (mean = 49%), followed by the natural forest (mean = 41%), but nonexistent in the clear cut habitat (mean = 0%) (Fig. 6.2). Both in the wet and dry seasons, a post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in stem damage by humans between natural forest versus clear cut and plantation versus clear cut habitats (Fig. 6.2).

120.00 100.00 80.00 Clear cut 60.00 Plantation 40.00 Natural forest

20.00 % stem% damage 0.00 Wet Dry Season

Figure 6.2. Seasonal stem damage on trees across habitat types in Munessa. The error bars represent +1 stdev. Presence or absence of peeled and / or damaged barks was assessed to quantify humans caused stem damages in each plot. No tree was observed in plots laid in the clear cut habitat type so that no stem damage was recorded in this habitat. The number of plots assessed during each season in the clear cut, plantation, and natural forest was 31, 41, and 37 respectively.

119

During the wet season, the intensity of crown damage by humans differed among habitats (F (2, 106) = 27.31; p < 0.001). Accordingly, it was highest in the plantation (mean = 73%), followed by the natural forest (mean = 63%), nonexistent in the clear cut habitat (mean = 0%) (Fig. 6.3). Similarly, in the dry season crown damage was significantly different among habitats (F (2, 106) = 10.19; p < 0.001), being highest in the plantation (mean = 49%), followed by the natural forest (mean = 38%), nonexistent in the clear cut habitat (mean = 0%) (Fig. 6.3). Both in the wet and dry seasons, a post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in crown damage by humans between natural forest versus clear cut and plantation versus clear cut habitats (Fig. 6.3).

140.00 120.00 Clear cut 100.00 Plantation 80.00 Natural forest 60.00 40.00

% crown% damage 20.00 0.00 Wet Season Dry

Figure 6.3. Seasonal crown damage on trees across habitat types in Munessa. The error bars represent +1 stdev. Presence or absence of removed or broken branches was assessed to quantify humans caused crown damages in each plot. No tree was observed in plots laid in the clear cut habitat type so that no crown damage was recorded in this habitat. The number of plots assessed during each season in the clear cut, plantation, and natural forest was 31, 41, and 37 respectively.

In the wet season, the intensity of wood use by humans differed among habitats (F

(2, 106) = 22.08; p < 0.001). Accordingly, it was highest in the natural forest (mean = 94%), followed by the plantation (mean = 78%), and the clear cut habitat (mean = 28%) (Fig.

6.4). Similarly, in dry season wood use also differed among habitats (F (2, 106) = 5.54; p =

120

0.005) and was highest in the natural forest (mean = 50%), followed by the plantation (mean = 46%), and the clear cut habitat (mean = 12%) (Fig. 6.4). Both in the wet and dry seasons, a post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in the intensity of wood use by humans between natural forest versus clear cut and plantation versus clear cut habitats (Fig. 6.4).

140.00 120.00 100.00 Clear cut 80.00 Plantation 60.00 40.00 Natural forest 20.00

0.00 % evidence% woodof use Wet Dry Season

Figure 6.4. Seasonal wood use across habitat types in Munessa. The error bars represent +1 stdev. Presence or absence of any left-over materials after wood collection was assessed to quantify evidence of wood use by humans. The number of plots assessed during each season in the clear cut, plantation, and natural forest was 31, 41, and 37 respectively.

In the wet season, the number of stumps cut by humans differed significantly among habitats (F (2, 106) = 23.17; p < 0.001). It was highest in the natural forest (mean = 12.16 cut stumps per plot), followed by the clear cut (mean = 5.48 cut stumps per plot), and the plantation habitat (mean = 1.73 cut stumps per plot) (Fig. 6.5). Similarly, in the dry season the number of stumps cut differed among habitats (F (2, 106) = 24.31; p < 0.001) and was highest in the natural forest (mean = 9.25 cut stumps per plot), followed by the clear cut (mean = 5.48 cut stumps per plot), and the plantation habitat (mean = 2.00 cut stumps per plot) (Fig. 6.5). Both in the wet and dry seasons, a post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in the number of stumps cut by humans between natural forest versus clear cut, natural forest versus plantation, and clear cut versus plantation habitats (Fig. 6.5).

121

25

20

15 Clear cut Plantation 10 Natural forest 5

Number of cut stumps / plot stumps cut / of Number 0 Wet Dry Season

Figure 6.5. Cut stumps in different seasons and habitat types in Munessa. The error bars represent +1 stdev. The number of cut stumps was quantified across habitat type. The number of plots assessed during each season in the clear cut, plantation, and natural forest was 31, 41, and 37 respectively.

Signs of livestock presence did not differ among habitats in either the wet (F (2, 106)

= 1.76; p = 0.177; Fig. 6.6) or the dry season (F (2, 106) = 0.17; p = 0.846); instead they were dispersed throughout all habitats (Fig. 6.6). Overall, signs of habitat use by livestock were significantly higher in the wet season than in the dry season (F (1, 212) = 48.85; p < 0.001) (Fig. 6.6; see also Appendix I - Table 6.2).

140.00 120.00 100.00 Clear cut 80.00 Plantation 60.00

livestock Natural forest 40.00 20.00 % sign of habitat use by use habitat signof % 0.00 Wet Dry Season

Figure 6.6. Seasonal habitat use by livestock in Munessa. The error bars represent +1 stdev. Each plot was assessed for the presence or absence of signs of habitat use by livestock (e.g., presence or absence of foot prints, droppings, etc.). The number of plots assessed during each season in the clear cut, plantation, and natural forest was 31, 41, and 37 respectively.

122

During the wet season, the level of grazing / browsing by livestock differed among habitats (F (2, 106) = 10.985; p < 0.001), with heaviest foraging occurring in the natural forest (mean = 57.75%), followed by the clear cut (mean = 52%), and the plantation habitat (mean = 33%) (Fig. 6.7). A post-hoc Tukey's HSD test for multiple comparisons across habitat types showed a significant difference in the level of grazing / browsing by livestock between natural forest versus plantation habitats (Fig. 6.7). In contrast, the dry season the level of grazing / browsing by livestock did not differ among habitats (F (2, 106) = 0.33; p = 0.724). Overall, the intensity of livestock grazing / browsing was significantly higher in the wet season than in the dry season (F (1, 212) = 63.78; p < 0.001) (Fig. 6.7; see also Appendix I - Table 6.2).

90.00 80.00 70.00 60.00 Clear cut 50.00 Plantation 40.00 30.00 Natural forest

livestock per plot per livestock 20.00 10.00

0.00 % level of grazing / browsing by browsing / grazing of level % Wet Dry Season

Figure 6.7. Seasonal level of grazing / browsing by livestock in Munessa. The error bars represent +1 stdev. Each nested circular plot with a radius of 0.5m was assessed for the level of grazing / browsing by livestock. This was carried out to quantify the extent of seasonal livestock disturbances on the quality and availability of habitats for Mountain Nyala in Munessa. The number of nested plots assessed during each season in the clear cut, plantation, and natural forest was 31, 41, and 37 respectively.

123

DISCUSSION Disturbance by human activity and livestock foraging differed at Munessa across habitats and seasons. In the wet season, livestock activity density was significantly highest in the natural forest habitat where there is more diverse palatable forage compared with the clear cut and the plantation habitats. However, during the dry season, livestock did not show significant habitat selection, but their mean activity density was still nominally highest in the natural forest habitat. Cattle constituted the highest densities both in the wet and dry seasons and horses the least during both seasons in all habitat type (see also Appendix I - Table 6.1). Overall, livestock activity densities in all habitat types were higher in the wet season than in the dry season (see also Appendix I - Table 6.1). One possible explanation is as follows. In the wet season, much of the lands of the local people are occupied by crop cultivation so that the local people have few options but to drive their livestock into the Munessa Forest for free-range grazing. However, in the dry season, the crops would have been harvested and crop residues available for livestock feeding, and thus there would be little need to take them to the forest for free-range grazing. The present study showed that commercial fuel wood collection, tree cutting for construction materials, logging, free-range livestock grazing, illegal hunting, and human nuisance are the most common threats to Mountain Nyala and their habitats in Munessa. The study further revealed that the problems are more prevalent and chronic in the wet season of the year. On top of this, it was frequently observed that many of the local people go to the Munessa Forest to cut trees and collect wood for commercial sale during the crop growing season, thus intensifying disturbance and decreasing habitat quality. In this manner, the ever-increasing demand for forest products coupled with the rapid increase in human population is putting intensified pressure on the remaining forest fragments and wildlife habitats in Munessa (Teketay, 1992; Evangelista et al., 2007). At the beginning of the dry season, I observed that there were few humans and livestock disturbances in any of the habitats of Mountain Nyala. Instead, this is a time when the local people drive their livestock to nearby agricultural fields to feed on crop residues. In addition, crop harvesting and marketing during this period engages most of the time of local people so that they cannot go to the forest so often to collect wood. Thus, the early dry season seems to be relatively a safe time for Mountain Nyala. However, this

124

situation is short-lived and by the late dry season, humans and livestock disturbances had returned to high levels in all habitats. Livestock induced disturbances might be among the major factors constraining regeneration and recruitment of woody species (Wassie et al., 2009) in the natural forest of Munessa. The present study revealed that the extent of habitat use and level of grazing /browsing by livestock was most intensive in the natural forest habitat especially in the wet season. Intensive livestock grazing and browsing may contribute to the reduction of primary production and the decline of woody species richness (Kebede, 2009) which otherwise could be good sources of forage and cover for wild herbivores including the Mountain Nyala in Munessa. For example, frequent disturbance due to livestock grazing and browsing strongly affect forest structure, species composition, and the ability of understorey species to regenerate (Neptali et al., 2001; Mamo, 2007; Wassie et al., 2009). Moreover, previous studies suggested that trampling and grazing of plant seedlings by livestock may lead to low seedling survival (e.g., Carolina and Javier, 2001; Wassie et al., 2009). Natural regeneration could be presumably low due to livestock grazing, uprooting, and the high fruit/seed predation that it causes (e.g., Neptali et al., 2001; Wassie et al., 2009). In addition, unrestricted livestock movements in the forest may destroy seeds or press them deep into the soil where they are not able to emerge easily (e.g., Hulme and Borelli, 1999; Smit et al., 2006; Wassie et al., 2009). Livestock trampling can also compact soil and aggravate erosion, and thereby reduce forest regeneration (Kebede, 2009; Wassie et al., 2009). In Munessa, lack of regeneration of dominant tree species such as Prunus africana, Syzygium guineense, and Hagenia abyssinica in the natural forest may be due to excessive grazing / browsing and trampling pressure on young seedlings and saplings by domestic herbivores. Prunus africana, Syzygium guineense, and Hagenia abyssinica are good sources of forage for Mountain Nyala in Munessa (personal communication with local people). Unless the interference of livestock in the core habitats of Mountain Nyala is controlled, canopy trees and shrubs that are potential cover and food sources for Mountain Nyala could become rare or locally extinct. Most importantly, during free-range grazing, all livestock in Munessa were observed being followed by their sheepherders. This reduces the risk of predation to livestock (e.g., Shrader et al., 2008a) so that livestock often leave

125

the foraging patch with very low giving-up densities which ultimately results in overgrazing of the best habitats for Mountain Nyala. In addition, perennial palatable grasses cannot tolerate the repeated impacts of livestock grazing and trampling, which eventually lead them to be replaced by unpalatable grass, forb, and shrub species (e.g., Mamo, 2007; Kebede, 2009). The present study revealed that, both in the wet and the dry seasons, tree stem damage and crown removal by humans were the most prevalent problems in the plantation habitat (see also Appendix I - Table 6.2). For example, I observed that the local people frequently de-bark the standing trees in the plantation habitat to get materials for traditional beehive building. In addition, I frequently noted that the local people remove the crowns of standing trees in the plantation habitat to get wood to sell. I also observed that the local people often cut the branches of palatable shrubs and trees in the natural forest to feed their livestock on site. The presence of leftover materials after wood collection suggests that human disturbance is a common problem in all habitats of Mountain Nyala in Munessa, but is most prevalent in the natural forest habitat both in the wet and the dry seasons. The number of cut stumps quantified in this study also suggests that illegal tree cutting activity is a common problem in all habitats, but the problem is still most severe in the natural forest both in the wet and dry seasons (see also Appendix I - Table 6.2). Ethiopia, which stands first in cattle population and second in human population in Africa, has been facing chronic problems for the management of most of its wildlife protected areas (Hillman, 1986, 1993; Tedla, 1995; Hundessa, 1992, 1997; Berry, 2003; Evangelista et al., 2007; Mamo, 2007; Evangelista et al., 2008; Kebede, 2009). The distribution and habitat use of wild animals are often negatively affected by the presence and proximity of people and livestock (e.g., Austin and Urness, 1986; Wallace and Krausman, 1987; Augustine, 2004; Young et al., 2005; Kebede, 2009). Habitats which are adjacent to human settlement are less used by Mountain Nyala in Munessa, largely due high human nuisance and livestock disturbances. In addition, the clearing of natural vegetation for the expansion of agricultural land and settlements (Teketay, 1992) has directly reduced the ranges, availabilities, and qualities of habitats for Mountain Nyala. In a recent study, Mamo (2007) noted that the mineral spring locally known as ‘Hora’, which was once used as salt licks by Mountain Nyala, was largely encroached by expanding

126

human settlements in the Bale Mountains National Park (BMNP). As a result, Mountain Nyala avoided human settlement areas in the BMNP. Atickem et al. (2011) also confirmed that the presence, distribution, and habitat use of Mountain Nyala in the Bale Mountains was negatively affected by human developments such as agriculture and human settlements. The temporal partitioning in habitat use between livestock and Mountain Nyala in Munessa should be a direct indicator of the exclusion of the latter by the former. In the present study, livestock used all three habitat types in both seasons, but Mountain Nyala always avoided at least one habitat, usually that with heavy livestock use. In addition, I have never seen any Mountain Nyala in any of the three habitat types while livestock were there in either season. Even in the less disturbed BMNP, I have never seen the Mountain Nyala browsing among livestock during my field work. Hansen and Reid (1975) and Mishra and Rawat (1998) note that when there is high overlap in habitat and forage requirements by livestock and wild herbivores, overstocking can cause competitive exclusion of the latter. In his recent study, Mamo (2007) reported a negative correlation between density versus droppings of livestock and Mountain Nyala in the Bale Mountains National Park. This also supports the hypothesis that Mountain Nyala avoid areas dominated or over-utilized by humans and livestock. Mamo (2007) also noted that one of the potential negative impacts of livestock on Mountain Nyala in the Bale Mountains National Park may be as a source of ecto-parasites such as ticks. Evangelista et al. (2007) have stated that examination of Mountain Nyala carcasses taken by legal hunting at lower elevations indicates that Mountain Nyala are hosts to flies, ticks, and a variety of ecto- parasites that can easily serve as vectors for disease. So, the risks of disease and parasites to Mountain Nyala are likely to be greater in populations in close proximity to livestock (e.g., Evangelista et al., 2007). During the focal group discussions held with local people (see also Chapter 7), the local participants noted that the population of Mountain Nyala in Munessa is currently facing multi-dimensional threats. These include competition from free-range livestock grazing, habitat destruction due to illegal tree cutting for fuel wood and construction materials, predation of calf and juvenile Mountain Nyala by Spotted Hyena and Anubis Baboon, and massive illegal hunting by the highlanders of the Arsi people. Many of the

127

local people noted that Mountain Nyala usually move to the inaccessible and remote natural forest habitat during the day time to avoid human noise and livestock disturbances. Malcolm and Evangelista (2005) also noted that Mountain Nyala are confined in the steep woodland areas which are unsuitable for human settlement and crop cultivation in the Arsi and the Bale Mountains. The local discussants further emphasized that humans and livestock induced disturbances made viewing Mountain Nyala a difficult task during the day time. Thus, the impact of livestock and human activity is manifold, affecting range quality for Mountain Nyala and reducing the potential for ecotourism by reducing the quality of wildlife viewing. Stephens et al. (2001) noted that the high number of livestock has negative impact on the Bale Mountains National Park in terms of tourist attraction. In summary, even though Mountain Nyala are known to persist in areas heavily affected by human activities (e.g., Evangelista et al., 2008), the current study revealed that the presence of people and livestock often excludes Mountain Nyala from using their optimum habitats in Munessa. Other authors also noted that Mountain Nyala in the Bale Mountains avoid areas frequently used by humans and livestock (e.g., Malcolm and Evangelista, 2005; Mamo, 2007; Atickem et al., 2011). Mountain Nyala never graze an area which is overgrazed by livestock (e.g., Mamo, 2007). As selective feeders, Mountain Nyala need to feed on palatable and fresh forage species. Mountain Nyala usually browse on leaves, twigs, flowers, and fruits of palatable and fresh plant species (e.g., Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). Thus, the negative impacts of human and livestock disturbances on Mountain Nyala could be measured and understood in terms of shifts in their behaviors and habitat use. Unless urgent and consistent conservation measure and management action is taken against the negative impacts of humans and livestock encroachments on Mountain Nyala and their habitats, the anthropogenically pressured small population of the unique Mountain Nyala in Munessa will locally go extinct in the foreseeable future.

128

CHAPTER 7 ATTITUDES OF LOCAL PEOPLE TOWARD MOUNTAIN NYALA AND THEIR HABITATS IN MUNESSA ABSTRACT Wildlife conservation in developing countries with high numbers of people living in poverty is often viewed as being in conflict with development goals. The attitudes of local people toward Mountain Nyala and their habitats in the Munessa Forest of Ethiopia were studied through interview questionnaires and focal group discussions. For the interview questionnaires, pretested questionnaires were developed that considered the different demographic and socio-economic characteristics of the local people, such as marital status, sex, age, family size, length of local residence, livelihood strategy, land ownership, livestock ownership, and knowledge about Mountain Nyala and their habitats. Based on their proximity to the study site, three peasant associations and one village were selected and involved in the interview questionnaires. For the focal group discussions, key community members who are rich in indigenous knowledge were selected and participated. The results revealed that the attitudes of local people toward Mountain Nyala, its habitats, and population increase were significantly affected by several socio- economic variables such as livelihood strategy, land ownership, livestock ownership and knowledge. The present study revealed that Mountain Nyala is an observable component of the local landscape. As a result, respondents did not have difficulty in formulating positive attitudes toward Mountain Nyala and its population increase. In addition, through focal group discussions, key community members shared their indigenous knowledge about the different behaviors of Mountain Nyala and their habitats in Munessa. By comparing attitudes quantified in the baseline study presented here and results from future replication of this study, researchers can provide relevant information for decision makers and conservation managers to deal with potential conflict of interests between the conservation of the endangered Mountain Nyala and the needs of the local people. Key words: attitudes, demographic and socio-economic characteristics, focal group discussion, indigenous knowledge, interview questionnaires, key community members, wildlife conservation

129

INTRODUCTION Successful species and habitat conservation and management often relies on the support of local people (Morzillo et al., 2007). The disconnection between ecological knowledge and conservation success has led to a growing sense among ecologists and field practitioners that social factors are often the primary determinants of success or failure (Mascial et al., 2003). For example, Folke et al. (1996) and Foley et al. (2005) recognize increased prospect of failure in biodiversity conservation unless the processes include the surrounding landscape dynamics and associated human conflicts. In the past, failures of management implementation by wildlife managers often occurred when they did not assess social feasibility in the management process (e.g., Jacobson et al., 2006). So, wildlife managers must provide both the ecological needs of a species and the needs and wants of humans who may be affected by management decisions (Kleiman, 1989; Riley et al., 2002). In this way, engagement with local people is clearly a key component of any conservation activity to resolve human-wildlife conflict (Hillman, 1993; Agrawal, 1997; Woodroffe et al., 2005). Local communities are often rich in indigenous knowledge and appreciation of their natural and cultural heritages. Yet pressure for rapid development can alienate people from these heritages and degrade the local environment. For example, lack of public awareness and absence of economic benefit-sharing schemes from wildlife resources to the local people contributed much to the loss of wildlife in Ethiopia (Hillman, 1993; Tedla, 1995; Hundessa, 1997; Evangelista et al., 2007; Kebede, 2009). The extent to which local people participate in wildlife conservation may vary; it can be as little as outreach which acknowledges local communities’ concerns, or may include full delegation of conservation and management authority to local people (Hillman, 1993; Woodroffe et al., 2005; Tessema et al., 2007). Regardless of the land use type within an area, it is less likely that far-ranging species, such as Mountain Nyala, will remain completely isolated from humans (Evangelista et al., 2007). Hillman (1993) suggests that the successful conservation and management of our wildlife is dependent upon the goodwill and co-operation of the local people who are inherently connected with the wildlife and its habitats. So, the conservation and utilization measures of wildlife and its habitats must be directed

130

towards this, particularly through their active participation in wildlife management as a form of land use (Hillman, 1993). Management interventions which could be technical measures, economic incentives, or policy initiatives can help resolve conflicts and allow sustainable co-existence of people and wildlife (Agrawal, 1997; Woodroffe et al., 2005). For example, revenue generating measures such as ecotourism and legal sport hunting might be more successful than simple compensation (Agrawal, 1997; LWAG, 2002; Woodroffe et al., 2005; Lindsey, 2008) to resolve conflicts created between local people and problematic wildlife species. But the success of such actions depends on being able to quantify and track the attitudes of local people. In wildlife management, attitudes of people are positive or negative responses to a particular species (Bath, 1989; Bright and Manfredo, 1996). So, attitudes can vary on the basis of key demographic and socioeconomic characteristics such as knowledge about a species, sex, age, length of local residence, and household income (Kellert et al., 1996; Bowman et al., 2001; Bowman et al., 2004). For example, in a study conducted to examine the attitudes of people toward black bears, Shropshire (1996) and Bowman et al. (2004) noted that income, level of education, sex, age, and knowledge about black bears were among the significant predictors of support for increasing the Mississippi black bear population in the USA. The charismatic Mountain Nyala is an endemic flagship species in Ethiopia, but lives with considerable human and livestock pressures throughout its ranges of distribution, including the Munessa Forest (Evangelista et al., 2007). Of course, Mountain Nyala are protected under the Ethiopian wildlife laws both at federal and regional levels; however, anthropogenic factors are increasingly causing problems for their conservation and management in Ethiopia (Hillman, 1985; Hillman and Hillman, 1987; Woldegebriel, 1996; Stephens et al., 2001; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007; Sillero-Zubiri, 2008). The Ethiopian Wildlife Conservation Authority (EWCA) is unable to enforce laws intended to protect Mountain Nyala and their habitats in Ethiopia including the Munessa Forest (Evangelista et al., 2007; Sillero-Zubiri, 2008). Since its discovery in 1908, Mountain Nyala has declined substantially in number and shrunk in range, mainly because of habitat destruction and uncontrolled hunting (Refera and Bekele, 2004; Evangelista et al., 2007). Because of high humans and livestock

131

encroachments, Mountain Nyala are known to have been extirpated from some parts of Ethiopia (e.g., Wondo Genet) (Malcolm and Evangelista, 2005). Due to the steady decline in numbers, Mountain Nyala is listed as an endangered C1 species by the IUCN (2008). Legal trophy sport hunting in the Munessa hunting block generates revenue. The use of licensed sport hunting as a means of conservation may allow the wild animals in the hunting block to make dramatic recoveries (Malcolm and Evangelista, 2005; Atickem et al., 2011) because part of the money can be allocated to promote infrastructure developments that enhance wildlife and habitat conservation activities (Evangelista et al., 2007; Atickem et al., 2011). However, the local people expressed their frustration over little economic benefit they received from the legal sport hunting in Munessa. Ensuring continued local support for Mountain Nyala conservation over the long term needs proactive programs of sustainable economic benefit-sharing and local awareness of conservation values (Tessema et al., 2007). The objective of this study was to assess and quantify the knowledge and attitudes of local people about the management of Mountain Nyala and its habitats in the Munessa Forest of Ethiopia. This in turn helps identify the opportunities for managers to address existing concerns about Mountain Nyala, assist residents with learning more about Mountain Nyala, and could minimize potential conflicts between conservation and management of Mountain Nyala and humans in Munessa. Based on previous literature and studies dealing with Mountain Nyala and its habitats in Ethiopia (e.g., Hillman and Hillman, 1987; Woldegebriel, 1996; Stephens et al., 2001; Evangelista et al., 2007, Mamo, 2007; Atickem et al., 2011), I expect that: (1) local residents hold generally positive attitudes toward Mountain Nyala, (2) efforts to increase knowledge about Mountain Nyala among local residents will lead to greater support for increasing the population size of Mountain Nyala in Munessa, (3) demographic and socioeconomic characteristics, such as sex, age, marital status, length of local residence, livelihood strategies, and knowledge should contribute to differences in attitudes toward the conservation and management of Mountain Nyala and its habitats in Munessa.

132

METHODS The Study Area The study was conducted in Munessa, Ethiopia (see Chapter 2). For this particular study, based on their proximity to the Munessa Forest, three adjacent peasant associations, i.e. Senbero-Rogicha Peasant Association, Degaga Peasant Association, and Argeda-Shaldo Peasant Association, and one village, i.e. Goljota village, were selected. I briefly describe each peasant association and/or village as follows.

Senbero-Rogicha Peasant Association Senbero-Rogicha Peasant Association is situated northwest of the Munessa Forest. The peasant association is characterized by “Dega”, “Weina Dega”, and “Kolla” climate types (Senbero-Rogicha Peasant Association office). The peasant association has a total area of 2,230 ha, and a human population of 5,575 living in 1060 households. They mainly engage in crop production, animal husbandry, mixed farming, and traditional bee keeping activities. The major crops grown include maize, wheat, sorghum, teff, and barley. Domestic livestock include 5,929 cattle, 2,752 sheep and goats, and 904 horses and donkeys. The major tree species grown in the area include Cupressus lusitanica, eucalyptus spp., acacia spp., Celtis africana, Juniperus procera, Afrocarpus falcatus, and Croton macrostachyus (Source: Senbero-Rogicha Peasant Association office). According to Negash and Ermias (1995) and Dejene (2003), the traditional climate classification can be described as follows. “Dega” refers to cool and humid highlands with a mean annual temperature of 11.5° C – 16° C, mean annual rainfall of 900 – 1200 mm, and elevation of 2300 m – 3200 m above sea level; “Weina Dega” refers to temperate and cool sub-humid highlands with a mean annual temperature of 16° C – 20° C, mean annual rainfall of 800 mm – 1200 mm, and elevation of 1500 m – 2300 m above sea level; and “Kolla” refers to warm semi-arid lowlands with a mean annual temperature of 20° C – 27.5° C, mean annual rainfall of 200 mm – 800 mm, and elevation of 500 m – 1500 m above sea level.

133

Degaga Peasant Association Degaga Peasant Association is situated west of the Munessa Forest. The peasant association is characterized by “Weina Dega” and “Kolla” climate types (Degaga Peasant Association office). The peasant association has a total area of 2,280 ha, and a human population of 8,330 (4,020 males and 4,310 females) residing in 1,190 households. They mainly engage in crop production, animal husbandry, and mixed farming activities. The major crops grown include maize, wheat, sorghum, teff, and barley. Livestock include 5,982 cattle, 1,039 sheep, 1,105 goats, 105 horses, and 905 heads of donkeys. Major tree species grown in the peasant association include Cupressus lusitanica, eucalyptus spp., Celtis africana, Prunus africana, Cordia africana, Afrocarpus falcatus, and Croton macrostachyus (Source: Degaga Peasant Association office).

Argeda-Shaldo Peasant Association Argeda-Shaldo Peasant Association is situated southwest of the Munessa Forest. The peasant association is characterized by (3,362 males and 3,087 females) living in 993 households. They mainly engage in crop prod “Weina Dega” and “Kolla” climate types (Argeda-Shaldo Peasant Association office). The peasant association has a total area of 3,205.5 ha, and a human population of 6,449 uction, animal husbandry, and mixed farming activities. The major crops grown include maize, wheat, sorghum, teff, and barley. Livestock include 5,555 cattle, 985 sheep, 970 goats, 145 horses, and 558 donkeys. The major tree species grown in the area include Cupressus lusitanica, eucalyptus spp., acacia spp., Celtis africana, Prunus africana, Syzygium guineense, Aningeria adolfi-friederici, Afrocarpus falcatus, Mellittia ferrugenea, and Croton macrostachyus (Source: Argeda-Shaldo Peasant Association office).

Goljota Village Goljota Village is situated west of the Munessa Forest. The village is characterized by “Weina Dega” climate type (Goljota Village office). The total human population size is 5,558 (2,752 males and 2,806 females) living in 1013 households. They mainly engage in crop production, animal husbandry, and retailing activities. Major crops include maize, wheat, and sorghum. Livestock include 971 cattle, 336 sheep, 310 goats, 7

134

horses, and 252 donkeys. The major tree species grown in the village include Cupressus lusitanica, eucalyptus spp., Celtis africana, Afrocarpus falcatus, Cordia africana, and Croton macrostachyus (Source: Goljota Village office).

Data Collection Techniques Interview questionnaires Properly designed interview questionnaires can be used as effective and affordable tools for gathering large amounts of data within a relatively short period of time (Hague, 1993). However, setting meaningful and interpretable questions is a complex process and requires careful word selection to ensure that the questions are administered in such a way that they are fully understood by the respondents (Hague, 1993; Gendall, 1998). This technique is useful for collecting both qualitative and quantitative data on the attitudes of local people toward Mountain Nyala and their habitats in Munessa. In addition, it is also helpful for querying local communities about general trends in the Mountain Nyala population and the major problems they face. The questionnaires considered key demographic and socioeconomic characteristics, such as sex, age, marital status, family size, length of local residence, livelihood strategies, livestock ownership, land ownership, and knowledge about Mountain Nyala and their habitats. A key aim of the survey is to obtain information on the needs of the local people. It also encourages the local residents to become involved in the conservation and management activities of Mountain Nyala and their habitats. This in turn better informs decision makers to formulate community-based conservation strategy and management plan which guide conservation efforts toward Mountain Nyala and their habitats in Munessa (Evangelista et al., 2007). Considering adequate level of reliability and costs of data collection (Green, 1991), a sample size of 5% of randomly selected households from each of the peasant associations and/or village included in the study area was chosen to receive pretested open and closed ended interview questionnaires. The questionnaires were administered in December 2010 to a total of 214 randomly selected households (i.e., 53 households from Senbero-Rogicha

135

Peasant Association, 60 households from Degaga Peasant Association, 50 households from Argeda-Shaldo Peasant Association, and 51 households from Goljota Village). Independent variables: Independent variables were derived from 22 survey questions focusing on respondent demographics, socioeconomics, and familiarity with Mountain Nyala in Munessa. These include:(a) peasant association and / or village type, (b) sex, (c) age, (d) marital status, (e) number of family members (number of adult males, number of adult females, and number of children (<18 years of age)) in a household, (f) occupation type, (g) do you have livestock? (yes, no), (h) if yes to question (g), how many livestock do you have? (cattle, horse, donkey, sheep, and goat), (i) do you have enough grazing land for your livestock? (yes or no), (j) do you want to keep more livestock than you have at present? (yes or no), (k) if yes to question (j), why do you want to keep more number of livestock? (l) do you have a shortage of fodder for your livestock? (yes or no), (m) if yes to question (l), how do you manage to satisfy the forage requirement for your livestock?, (n) place of settlement, (o) how long have you been here (in years)?, (p) how did you settle here?, (q) do you plan to stay here in the future? (yes, unsure, or no), (r) do you have your own land? (yes or no), (s) if yes to question (r), have you allocated any of your land for woodlot plantations? (yes or no), (t) if yes to question (s), how much have you allocated for woodlot plantation?, (u) do you have a shortage of fuel wood? (yes or no), (v) if yes to question (u), how do you manage your fuel wood shortage? Knowledge of respondents about Mountain Nyala and their habitats. Respondents were asked to share their indigenous knowledge about Mountain Nyala and its habitats in Munessa with special emphasis on the following 15 questions: (a) have you ever seen Mountain Nyala in Munessa? (yes or no), (b) if you said yes to question (a), which sex and age have you commonly observed?, (c) in which habitats do you commonly observe Mountain Nyala in Munessa?, (d) what do you know about the trend in the Mountain Nyala population in Munessa in the last decade?, (e) if you said decreasing to question (d), what is the reason for the decline of the Mountain Nyala population in Munessa?, (f) have you ever seen any dead / killed Mountain Nyala in the past years in Munessa? (yes or no), (g) mention if there are problems created by Mountain Nyala on your farm crops or livestock, (h) do you think that the Mountain Nyala forage on areas grazed / browsed by livestock? (yes, unsure, or no), (i) have you ever observed Mountain Nyala while

136

browsing / grazing among livestock in the field in Munessa? (yes or no), (j) do you know any common disease or parasite affecting both Mountain Nyala and livestock? (yes, unsure, or no), (k) if you said yes to question (j), mention the type of common diseases or parasites affecting both domestic livestock and Mountain Nyala in Munessa, (l) other than humans, what main predators kill the Mountain Nyala in Munessa?, (m) do you think that Mountain Nyala migrate or travel long distances out of Munessa to other areas? (yes, unsure, or no), (n) if you said yes to question (m), daily, seasonally, or both?, (o) if you said yes to question (m), mention the areas of movements and the reasons behind for such movements by Mountain Nyala. Activities. Respondents were asked to tell about their participation in 5 activities that could potentially put them in either direct or indirect contact with Mountain Nyala: (a) do you think that you get benefits from Mountain Nyala in Munessa? (yes, unsure, or no), (b) if you said yes to question (a), what are the perceived benefits to you?, (c) do you have any beliefs that encourage killing of Mountain Nyala when you encounter them in your land holdings or communal lands? (yes, unsure, or no), (d) is there any ritual ceremony or cultural practice in which you use the products of Mountain Nyala, such as horns and skins in Munessa? (yes, unsure, or no), (e) how would your livelihoods be affected if all the benefits (e.g. no wood, no grazing, etc.) from the Munessa forest are stopped? (negatively affect me, unsure, or do not affect me). Dependent variables: I examined two groups of dependent variables. First, “attitudes toward Mountain Nyala,” was constructed using principal component analysis with varimax rotation (Fabrigar et al., 1999) from five basic attitude and/or belief based statements related to Mountain Nyala and two related to the conservation and management of Mountain Nyala and its habitats. A varimax rotation is a change of coordinates used in principal component analysis and/or factor analysis that maximizes the sum of the variances of the squared loadings (Kaiser, 1958). Respondents indicated agreement / disagreement with each of the following statements: (i) the Mountain Nyala in Munessa is the heritage of all Ethiopians, (ii) the presence of Mountain Nyala in Munessa is the sign of a healthy environment, (iii) the Mountain Nyala have the right to exist wherever they may occur, (iv) predators such as leopards are forced to kill our livestock when their natural prey, such as Mountain Nyala, disappear from Munessa, (v)

137

the quality and availability of habitats for Mountain Nyala in Munessa would be affected by human and livestock disturbances, (vi) tourism in the Munessa Forest is a good opportunity for promoting the conservation of Mountain Nyala and its habitats, and (vii) the conservation and management of the Mountain Nyala population and its habitats in Munessa should be supported by the local people. Responses to each of the seven statements were measured using 5-point scales indicating level of agreement (5 = strongly agree, through 1 = strongly disagree). Items were coded so that larger values reflected greater support for Mountain Nyala. The second group dependent variable, “attitudes toward Mountain Nyala population increase,” was evaluated by asking respondents to answer the following three questions: (i) do you wish to see more numbers of Mountain Nyala in the future in Munessa?, (ii) do you think that the areas occupied by Mountain Nyala at present would be sufficient enough to see increased number of Mountain Nyala in the future?, and (iii) do you have any traditional practices or taboos that restrict local people from killing of Mountain Nyala in Munessa? Responses were measured on a 3-point scale (3 = yes, 2 = unsure, 1 = no). Items were coded so that larger values reflected greater support for Mountain Nyala population increase.

Focal group discussion Focal group discussion is a kind of qualitative research data acquisition technique which helps collect relevant information through group interactions on a specific topic (Bernard, 2002; Neumann, 2003). In this study, the data collected through focal group discussions allowed me to quantify the opinions and knowledge of local people toward the conservation and management of Mountain Nyala and their habitats in Munessa. This also assists to assess what basic information is currently available and what new information is needed for managing Mountain Nyala and their habitats in Munessa. I contacted the leaders of the aforementioned communities and thoroughly briefed them on the purpose of the focal group discussion. Accordingly, I requested the leader of each community to assign key community members who are rich in their indigenous knowledge and can share with me much information about Mountain Nyala and their habitats in Munessa. Consequently, all the selected community members agreed to

138

participate in the focal group discussions. A check list was developed to guide the focal group discussions. I developed the check list that consists of a range of questions based on available literature and my past field experiences (see also Appendix II). Two focal group discussion sessions, with each group comprised of 12 people, were held. In each group session, different people participated. This helps acquire a range of new information from different local people. I used a note book to compile the opinions of the selected key community members who participated in each focal group session. In each focal group discussion session, the same topics of discussion were addressed. This kind of technique helps cross check the validity and consistency of information collected from participants of different focal groups (Tessema et al., 2007). I conducted the focal group discussion sessions in December 2010.

Data Analysis I summarized the data using descriptive statistics organized into tabular format. Weights were applied to descriptive analyses to reflect the actual population distribution of the entire study area (Kalton, 1983; Morzillo et al., 2007). For the two groups of dependent variables, I used principal component analysis with varimax rotation for data reduction (Morzillo et al., 2007). Cronbach’s alpha (α) was used as the test for internal consistency of the components (Cortina, 1993; Gliem and Gliem, 2003). Cronbach's alpha (α) is a coefficient of reliability used to measure the internal consistency or reliability of a test score for a sample of examinees (Cortina, 1993; Gliem and Gliem, 2003). I used one-way ANOVA and linear regression to compare sample means and test the variable relationships, respectively. For all analyses, I defined the alpha value of 0.05. After accounting for multiple comparisons in bivariate analyses (28 tests per dependent variable) with a Bonferroni correction, p ≤ 0.002 was considered significant. The Bonferroni correction is a safeguard against multiple tests of statistical significance on the same data falsely giving the appearance of significance (Gliner et al., 2001). I computed effect size (eta) for all variables identified as significant by the bivariate statistical analysis to examine the strength of relationship between variables (Gliner et al., 2001). I used multiple linear regression to predict dependent variables and to develop the

139

model. Information collected from the focal group discussions was organized, summarized, and rated using text analysis (Bernard, 2002; Neumann, 2003) to discover the consistency of the responses of the discussants with respect to the research objectives (Tessema et al., 2007). Later, I used the most important ones to enrich the discussion part of this study.

RESULTS Most of the respondents (79.91%) were males, the average age of the respondents was 40.59 years, and the greatest percentage (92.06%) of the respondents was married. The average family size in a household was 7.19 people, i.e. 2.09 adult males, 1.79 adult females, and 3.31 children, and the greatest percentage (37.85%) of the respondents engaged in mixed farming (see also Table 7.1). Most of the respondents (92.52%) had livestock. Of these, on average, each respondent had 5.01 cattle, 0.80 horses, 1.17 donkeys, 3.05 sheep, and 1.98 goats. Most of the respondents (86.54%) did not have enough grazing land. In contrast, 83.05% of the respondents felt a need to keep more livestock than they had at present. Most of the respondents (55.87%) noted that having more livestock serve as insurance during crop failure. However, 78.18% confirmed that they had a shortage of fodder for their livestock. Concerning management of fodder shortage for their livestock, most of the respondents (63.80%) noted that they used free-range grazing in the Munessa Forest (see also Table 7.1). Most of the respondents (76.17%) confirmed that they lived in peasant associations adjacent to the Munessa Forest. On average, respondents had lived in the area for about 30.93 years. Regarding history of settlement in the area, most of the respondents (50.99%) noted that they had inherited land from their ancestors. Most of them (84.03%) planned to stay in the area in the future, and 80.49% confirmed that they owned their own private lands. However, most of the respondents (64.50%) confirmed that they had allocated none of their land holdings for woodlot plantations. So, most of the respondents (69.18%) noted that they had a shortage of fuel wood, and that they often collected fuel wood from the Munessa Forest to manage their fuel wood shortage (66.73%; see also Table 7.1).

140

Most of the respondents (97.20%) confirmed that they had seen Mountain Nyala in Munessa. Of these, 93.75% noted that they had seen both male and female Mountain Nyala. About 98.56% of the respondents had commonly seen adult Mountain Nyala. In addition, most of the respondents (94.86%) confirmed that they commonly used to see Mountain Nyala in the natural forest habitat. Regarding the trend of Mountain Nyala population in Munessa during the last decade, most of the respondents (53.86%) were of the opinion that the population size of Mountain Nyala had been decreasing. For example, most of the respondents (65.70%) confirmed that they had seen dead / killed Mountain Nyala in Munessa. In contrast, many of the respondents (48.06%) noted that Mountain Nyala do not create any problems for either their farm crops or livestock (see also Table 7.1). Most of the respondents (77.79%) noted that Mountain Nyala forage on areas grazed / browsed by livestock. For example, most of the respondents (75.51%) argued that they had observed Mountain Nyala while browsing /grazing along with livestock in the field in Munessa. However, most of the respondents (63.95%) knew common diseases and parasites affecting both Mountain Nyala and livestock in Munessa, particularly ticks. In addition, most of the respondents (94.73%) noted that, other than humans, leopards are the main predators preying on Mountain Nyala in Munessa. Regarding the movement behavior of Mountain Nyala, most of the respondents (70.65%) noted that Mountain Nyala migrate long distances out of Munessa. For example, most of the respondents (59.03%) noted that the Mountain Nyala regularly move to Lake Langano to drink salty water “Hora” during the dry Season (see also Table 7.1). Most of the respondents (94.33%) confirmed that they got benefits from Mountain Nyala in Munessa, where aesthetic and recreational values were among the most prominent ones. So, most of the respondents (85.98%) stated that they did not have any beliefs that encouraged killing of Mountain Nyala when they encountered them in their land holdings or communal lands. In contrast, most of the respondents (58.09%) confirmed that they had often used the products of Mountain Nyala, such as horns and skins during ritual ceremonies. Thus, most of the respondents (88.30%) confirmed that their livelihoods would have been negatively affected if all the benefits from the Munessa Forest had been prohibited through strict conservation (see also Table 7.1).

141

Table 7.1. Sample characteristics and descriptive results for independent variables

Variable (n) Descriptive Resultsa Percent Senbero-Rogicha (1060 households) 24.77 Peasant association and / or village Degaga (1,190 households) 28.04 type (214) Argeda-Shaldo (993 households) 23.36 Goljota (1013 households) 23.83 Male 79.91 Sex (214) Female 20.09 Age (214) Mean = 40.59 years; stdev = 15.15 Married 92.06 Marital status (214) Single 6.07 Divorced 1.87 Adult males  mean = 2.09; stdev = 1.78 Number of family members in a Adult females  mean = 1.79; stdev = 1.23 household (214) Children (<18 years of age)  mean = 3.31; stdev = 2.65 Crop cultivation 36.45 Livestock rearing 1.87 Occupation type (214) Mixed farming 37.85 Others (e.g., guarding, daily employment, government jobs, etc.) 23.83 Livestock ownership (214) Yes 92.52 No 7.48 Cattle  mean = 5.01 ; stdev = 4.33 Horse  mean = 0.39 ; stdev = 0.80 Number of livestock (198) Donkey  mean = 1.30 ; stdev = 1.17 Sheep  mean = 2.64 ; stdev = 3.05 Goat  mean = 0.93 ; stdev = 1.92 Yes 13.46 Enough grazing land (198) No 86.54 Want to keep more livestock than Yes 83.05 have at present (198) No 16.95 Prestige 2.90 Reasons to keep more number of Insurance during crop failure 55.87 livestock (165) Enough grazing land 12.14 To improve income 29.09 Shortage of fodder for livestock Yes 78.18 (198) No 21.82 Free-range grazing 63.80 Cut and carry system 18.42 Management of shortage of fodder Traditional pastoral transhumance system 28.85 for their livestock (155) Purchasing additional fodder 16.72 Crop residue 54.41 Residue of local liquor 5.02 Inside the Munessa Forest 0.00 Place of settlement (214) In the adjacent peasant association or village 76.17 Outside the Munessa Forest 23.83 Time in the area (214) Mean = 30.93 years; Standard deviation = 14.99 Inherited land from my ancestor 50.99 Personal history of settlement in Bought land 18.90 the area (214) Settled by my own interest in search of land 12.08 Settled by the state 18.03 Yes 84.03 Plan to stay in Munessa in the Unsure 11.65 future (214) No 4.32 Yes 80.49 Private land ownership (214) No 19.51 Allocated land for woodlot Yes 35.50 plantation (172) No 64.50

142

Three-fourth of land holdings 3. 90 How much have you allocated your Half of land holdings 15.58 land holdings for woodlot Quarter of land holdings 67.53 plantation (77) One-eighth of land holdings 8.97 One-tenth of land holdings 7.92 Yes 69.18 Shortage of fuel wood (214) No 30.82 Collection of fuel wood from the Munessa Forest 66.73 Methods to manage fuel wood Using cow dung 2.87 shortage (148) Using crop residue as a source of fuel wood 32.95 Purchasing additional fuel wood 10.73 Seen Mountain Nyala in Munessa Yes 97.20 (214) No 2.80 Sex type of Mountain Nyala Only male Mountain Nyala 4.81 commonly seen in Munessa by Only female Mountain Nyala 1.44 respondents (208) Both male and female Mountain Nyala 93.75 Old age 75.00 Age class of Mountain Nyala Adults 98.56 commonly seen by respondents Sub-adults 87.98 (208) Juveniles 81.73 Calves 79.32 Habitat types where you commonly Natural forest habitat 94.86 see Mountain Nyala in Munessa Clear cut or open habitat 24.04 (208) Plantation habitat 18.75 Increasing 38.37 The trend of Mountain Nyala Decreasing 53.86 population in Munessa in the last Stable 2.08 decade (208) Unsure 7.19 Illegal hunting 47.95 Licensed trophy sport hunting 37.14 Predators 48.50 Accidents 5.66 Confirm the reasons for the decline Poor breeding abilities of the species 4.72 of Mountain Nyala population in Diseases and parasites 28.60 Munessa (112) Habitat loss, destruction, and fragmentation 51.86 Toxic plants 6.76 Emigration of Mountain Nyala to other areas 40.87 Habitat resource competition by livestock due to free-range 43.67 grazing Lack of public awareness and conservation education 49.31 Seen dead / killed Mountain Nyala Yes 65.70 in Munessa (214) No 34.30 No problem 48.06 Mention if there are problems Graze on young food crops, such as maize, wheat, and barley 40.35 created by Mountain Nyala on your Adult male Mountain Nyala inbreed with domestic cow 11.10 farm crops or livestock (214) Mountain Nyala transmit diseases and parasites to our livestock 0.49 Yes 77.79 Mountain Nyala forage on areas Unsure 10.10 grazed / browsed by livestock (214) No 12.11 Observed Mountain Nyala while Yes 75. 51 browsing/ grazing along with livestock in the field (166) No 24.49 Know common disease or parasite Yes 63.95 affecting both Mountain Nyala and Unsure 6.98 livestock in Munessa (214) No 29.08 Type of common diseases and Anthrax and Blackleg 14.07 parasites affecting both domestic Diarrhea 19.57 livestock and Mountain Nyala in Bovine tuberculosis 1.43 Munessa (137) Ticks 41.32

143

Spotted Hyena 64.97 Other than humans, what main Leopard 94.73 predators kill Mountain Nyala in Lion 3.36 Munessa (214) Jackal 5.33 Anubis baboon 39.19 Lammergeyer 10.71 Mountain Nyala migrate or travel Yes 70.65 long distances out of Munessa to Unsure 3.42 other areas (214) No 25.93 Daily 42.35 Kinds of temporal movements by Seasonally 35.38 Mountain Nyala in Munessa (151) Both daily and seasonally 22.27 Throughout the Munessa Forest 40.83 Areas of movement by Mountain To Gambo Forest 4.42 Nyala (151) To Mount Kubsa, Chike, and Badira 1.91 To Lake Langano 59.03 In search of palatable forages 47.14 Reasons for movements by To drink salty water "Hora" during the dry season 59.03 Mountain Nyala in Munessa (151) To seek for safe covers to escape from risk of predation and 28.97 human disturbances Do you think that you get benefits Yes 94.33 from Mountain Nyala in Munessa? Unsure 1.47 (214) No 4.20 Employment opportunities 82.72 What are the perceived benefits to Infrastructure development 64.40 you from Mountain Nyala in Mountain Nyala products 67.63 Munessa? (177) Aesthetic and recreational values 90.30 Ecological values 5.72 Do you have any beliefs that Yes 10.74 encourage the killing of Mountain Nyala when you encounter them in Unsure 3.27 your land holdings or communal No 85.98 lands? (214) Is there any ritual ceremony or Yes 58.09 cultural practice in which you use 2.45 the products of Mountain Nyala, Unsure such as horns and skins in Munessa? (214) No 39.46 How would your livelihoods be Negatively affect me 88.30 affected if all the benefits from the Unsure 0.00 Munessa Forest get stopped? (214) Do not affect me 11.70 aDescriptive results were weighted to account for oversampling of the three peasant associations and one village. The principal components analysis revealed that 5 out of 7 belief-attitude statements grouped together (55.32% variance explained) for the dependent variable “attitudes toward Mountain Nyala, and its conservation and management” (α = 0.85) (Table 7.2). These statements included (factor loading scores in parentheses): (a) the presence of Mountain Nyala in Munessa is the sign of a healthy environment (0.74), (b) Mountain Nyala have the right to exist wherever they may occur (0.73), (c) the quality and availability of habitats for Mountain Nyala in Munessa would be affected by humans and livestock disturbances (0.74), (d) predators such as leopards are forced to kill our livestock

144

when their natural prey such as Mountain Nyala disappear from Munessa (0.72), and (e) the conservation and management of the Mountain Nyala population and its habitats in Munessa should be supported by the local people (0.71). The mean “attitudes toward Mountain Nyala” score across all respondents was 21.92 (SD = 4.08; possible score range = 5 - 25) (Table 7.2).

Table 7.2. Descriptive results (percent of responses) of multiple items combined to measure “attitudes of local people toward Mountain Nyala, and their conservation and management in Munessa”a

Sample Strongly Strongly Belief statements size Agree Agree Unsure Disagree Disagree M (SD)b (n) The presence of Mountain Nyala in Munessa is the sign of a healthy environment 214 5 4 3 2 1 4.47 (0.69) The Mountain Nyala have the right to exist wherever they may occur 214 5 4 3 2 1 4.26(1.03) The quality and availability of habitats for Mountain Nyala in Munessa would be affected by 214 5 4 3 2 1 4.25 human and livestock disturbances (0.97) Predators such as leopards forced to kill our livestock when their natural prey such as Mountain Nyala 214 5 4 3 2 1 4.19(0.89) disappear from Munessa The conservation and management of Mountain Nyala population in Munessa should be supported by the 214 5 4 3 2 1 4.75 local people (0.50)

aDescriptive results were weighted to account for oversampling of the three peasant associations and one village. bScale values (strongly agree = 5 through strongly disagree = 1) were used to calculate mean (M) and standard deviation (SD) values, where higher values indicate more favorable attitudes toward Mountain Nyala and their conservation.

The principal component analysis further grouped 2 out of 3 belief-attitude statements together to explain 32.07% of the variance for the dependent variable “attitudes toward Mountain Nyala population increase” (α = 0.81) (Table 7.3). These statements

145

included (factor loading scores in parentheses): (a) wish to see more numbers of Mountain Nyala in the future in Munessa (0.74), and (b) the areas occupied by Mountain Nyala at present would be sufficient enough to see increased number of Mountain Nyala in the future (0.71). The mean “attitudes toward Mountain Nyala population increase” score across all respondents was 5.62 (SD = 1.05; possible score range = 2 - 6) (Table 7.3).

Table 7.3. Descriptive results (percent of responses) of statements combined to measure “attitudes of respondents toward Mountain Nyala population increase in Munessa” a Sample Belief statements size Yes Unsure No M (SD)b (n) Wish to see more numbers of Mountain Nyala in the future in Munessa 214 3 2 1 2.93 (0.36) The areas occupied by Mountain Nyala at present would be sufficient enough to see increased number 214 3 2 1 2.69 (0.69) of Mountain Nyala in the future aDescriptive results were weighted to account for oversampling of the three peasant associations and one village. bScale values (Yes = 3, Unsure = 2, No = 1) were used to calculate mean (M) and standard deviation (SD) values, where higher values indicate more favorable attitudes toward Mountain Nyala population increase.

Bivariate results revealed that respondents who live in Senbero-Rogicha Peasant

Association (F (3, 210) =40.22, eta = 0.45), inherited land from their ancestors (F (3, 210)

=41.41, eta = 0.46), had seen Mountain Nyala in Munessa (F (1, 212) =45.41, eta = 0.52), benefited from Mountain Nyala (F (2, 211) = 49.13, eta = 0.55), knew that Mountain Nyala do not forage on areas grazed / browsed by livestock (F (2, 211) = 35.33, eta = 0.37), knew about the presence of common disease or parasite affecting both Mountain Nyala and livestock (F (2, 211) = 38.60, eta = 0.42), knew about the movements of Mountain Nyala out of the Munessa Forest (F (2, 211) = 37.06, eta = 0.41), and knew of beliefs that encourage the killing of Mountain Nyala when one encounters them on his/her land holdings or communal lands (F (2, 211) = 37.99, eta = 0.41) held significantly more positive attitudes toward Mountain Nyala and their habitats in Munessa. Bivariate results further showed that respondents who lived in Senbero-Rogicha (F (3, 210) = 35.96, eta = 0.37), engaged in crop cultivation (F (3, 210) = 33.87, eta = 0.34), allocated land for woodlot plantation (F (3, 212) = 32.38, eta = 0.33), had seen Mountain Nyala in Munessa (F (1, 212) =

146

35.33, eta = 0.36), benefited from Mountain Nyala (F (2, 211) = 45.16, eta = 0.52), and opposed the use of the products of Mountain Nyala in ritual ceremony or cultural practice in Munessa (F (2, 211) = 35.58, eta = 0.36) were more supportive of increasing the Mountain Nyala population size in Munessa. Multivariate analyses showed that attitudes toward Mountain Nyala and their habitats were significantly affected by peasant association / village type, personal history of settlement in Munessa, previous sightings of Mountain Nyala in Munessa, knowledge of Mountain Nyala foraging in areas grazed / browsed by livestock, knowledge of common disease or parasite affecting both Mountain Nyala and livestock, knowledge of the movements of Mountain Nyala out of the Munessa Forest, previous benefits from Mountain Nyala, and knowledge of beliefs that encourage the killing of Mountain Nyala when one encounters them in his land holdings or communal lands (Table 7.4). Moreover, attitudes of the respondents toward an increase in the population size of Mountain Nyala was significantly affected by peasant association / village type, occupation type, personal allocation of land for woodlot plantation, lack of sufficient fuel wood, previous sighting of Mountain Nyala, previous benefits from Mountain Nyala, and knowledge of the use of Mountain Nyala products in ritual ceremony or cultural practice (Table 7.4). Overall, the multiple linear regression model revealed that the independent demographic and socio-economic variables considered in this study had significant impacts on the two groups of the dependent variables i.e. “attitudes toward Mountain Nyala” (19.5% variance explained by the multivariate model), and “attitudes toward Mountain Nyala population increase” (15.0% variance explained by the multivariate model) (Table 7.4).

147

Table 7. 4. Multiple linear regression modela for attitudes toward “Mountain Nyala and their habitats”b, and “Mountain Nyala population increase” c. Significant variables with + ß coefficients indicated positive attitude toward Mountain Nyala or its population increase whereas significant variables with - ß coefficients revealed negative attitude toward Mountain Nyala or its population increase. Attitudes toward Mountain Attitudes toward Mountain Nyala and their habitats in Nyala population increase in Variable Munessa Munessa ß t (176) p value ß t (176) p value Peasant association and / or village type 0.137 2.819* 0.002 -0.466 -2.365* 0.001 Sex (Male = 1) 0.118 1.528 0.128 -0.056 -0.719 0.473 Age -0.002 -0.777 0.438 -0.002 -0.690 0.491 Marital status -0.005 -0.069 0.945 0.106 1.482 0.140 Family size per household 0.019 1.753 0.081 -0.0002 -0.013 0.989 Occupation type -0.005 -0.159 0.873 0.454 3.280* 0.001 Livestock ownership (Yes =1) -0.127 -0.562 0.575 -0.284 -1.256 0.211 Enough grazing land (Yes =1) 0.159 0.774 0.440 0.176 0.853 0.395 Want to keep more livestock than have at present -0.039 -0.404 0.686 -0.067 -0.679 0.498 (Yes = 1) Shortage of fodder for livestock (Yes = 1) 0.221 1.0912 0.276 0.179 0.881 0.379 Place of settlement 0.095 0.689 0.321 0.038 0.325 0.402 Time in the area 0.004 1.259 0.209 0.002 0.627 0.531 History of settlement in the area 0.532 3.162* 0.002 0.021 0.714 0.476 Plan to stay in the area in the future -0.074 -1.141 0.255 -0.034 -0.516 0.609 Land ownership (Yes = 1) 0.053 0.583 0.561 -0.015 -0.168 0.867 Allocated land for woodlot plantation (Yes = 1) 0.118 1.058 0.292 0.429 3.037* 0.002 Shortage of fuel wood (Yes = 1) 0.171 1.528 0.128 -0.415 -2.899* 0.002 Seen Mountain Nyala (Yes = 1) 0.362 3.817* 0.001 0.449 3.747* 0.001 The trend of Mountain Nyala population in 0.069 0.907 0.366 -0.019 -0.891 0.374 Munessa in the last decade Seen dead / killed Mountain Nyala in Munessa 0.052 0.768 0.601 -0.020 -0.299 0.765 (Yes = 1) Mountain Nyala forage on areas grazed / browsed -0.145 -2.033* 0.002 -0.079 -1.104 0.271 by livestock Observed while Mountain Nyala browsing/ grazing among with livestock in the field 0.077 0.139 0.434 -0.044 -0.336 0.737 (Yes = 1) Know common disease or parasite affecting both 0.547 3.218* 0.001 0.019 0.485 0.628 Mountain Nyala and livestock in Munessa Mountain Nyala migrate or travel long distances 0.541 3.096* 0.002 0.072 1.915 0.067 out of Munessa to other areas Benefited from Mountain Nyala in Munessa 0.351 2.330* 0.001 0.592 3.802* 0.0001 (Yes = 1) Beliefs that encourage the killing of Mountain Nyala when one encounters them in his land - 0.503 - 2.921* 0.001 0.037 0.685 0.494 holdings or communal lands Any ritual ceremony or cultural practice in which 0.067 0.014 0.346 -0.409 -3.296* 0.001 you use the products of Mountain Nyala My livelihoods will be affected if all the benefits 0.069 1.264 0.208 -0.0002 -0.003 0.998 from Munessa forest are stopped

aStandardized coefficients reported. An asterisk (*) represents significance at the 95% confidence level. b 2 2 R = 0.305 (Adj. R = 0.195), F (28,176) = 2.761, overall p < 0.0001. c 2 2 R = 0.267 (Adj. R = 0.150), F (28,176) = 2.287, overall p < 0.0006

148

DISCUSSION Laws that prohibit the illegal killing of threatened wildlife species have made a huge contribution to the conservation and protection of endangered species (Woodroffe et al., 2005). However, strict legal protection alone is not sufficient to achieve wildlife conservation objectives. For example, a large numbers of countries offer African wild dogs (Lycaon pictus) full legal protection, yet they have frequently become locally extinct (Woodroffe et al., 2005). This is because legal protection often creates a feeling of disenfranchisement among local people which can make them hostile to conservation efforts (Agrawal, 1997). However, less strict legal protection may help alleviate such feelings and stimulate greater public assistance for the conservation of conflict species (Hillman, 1993; Agrawal, 1997; Woodroffe et al., 2005). After all, it is the local people who are directly experiencing the costs of living with wildlife, and hence are those most likely to kill wildlife, legally or illegally. Legislation that is perceived to exclude the local stakeholders can often create aggression and may lead to reduced tolerance for wildlife and increased lethal control (Agrawal, 1997; Woodroffe et al., 2005). This is the first study to quantify the attitudes of local people toward Mountain Nyala population and its habitats in Munessa. The results of the present study are strongly skewed in a positive direction toward Mountain Nyala and its population increase, which is consistent with my expectations. The bivariate results revealed that several socio-economic variables significantly affected the dependent variables. Female respondents were as knowledgeable as male respondents and had similar attitudes toward Mountain Nyala. Interestingly, age, length of residence in the area, family size, and marital status did not affect the knowledge and attitudes of respondents toward Mountain Nyala. Based on effect sizes, the impact of the independent variables was medium, though the statistics were significant. This may be due to the similarity of knowledge among the different respondents about Mountain Nyala and strong agreement among the local respondents. In addition, the explained variances by the two groups of the dependent variables were also quite large. Similarly, the multivariate analyses and the subsequent multiple linear regression model revealed that several socio- economic characteristic variables showed significant effect on “attitudes toward Mountain

149

Nyala” and / or “attitudes toward Mountain Nyala population increase” as to my expectation, though the direction is not consistent (for the details, see Table 7.4). Wildlife management often demands integrated knowledge on the biology of the species, its habitat, and the needs of the local people living in the surrounding landscape (Anderson et al., 2002). So, wildlife managers should often incorporate both ecological and societal issues into their research designs and management plans (Hillman, 1993). While focusing on the attitudes of local people toward Mountain Nyala and its habitats, the present study contributes to improved knowledge of human dimensions in Munessa. This may motivate wildlife managers to include the local people while formulating wildlife conservation strategy and management plans for the Munessa hunting block. The present study revealed that Mountain Nyala is an observable component of the local landscape. As a result, respondents did not have difficulty in formulating positive attitudes toward Mountain Nyala. In addition, local people value Mountain Nyala for aesthetic and recreational reasons, which are thought to be resulted from historic links between wildlife and traditional tribal culture (Tessema et al., 2007). The respondents shared their abundant indigenous knowledge on Mountain Nyala. For example, the local people emphasized that Mountain Nyala usually go to Lake Langano in search of salty water “Hora” during the dry season. This is because the salt licks become very hard during the dry season in Munessa so that the Mountain Nyala cannot use them satisfy their sodium requirement. Moreover, during the dry season, the focal group participants noted that Mountain Nyala satisfy their sodium requirement by feeding on salty plant species, such as the bark of Cupressus lusitanica. Mamo (2007) also noted that salt is one of the important food components for Mountain Nyala in the Bale Mountains National Park. In addition, the local people discussed that depending on the scarcity of habitat resources in Munessa, Mountain Nyala seasonally migrate to Gambo Forest which is situated several kilometers from Munessa. The focal group participants noted that the Mountain Nyala were formerly distributed in the Arsi Mountains including Mountain Kubsa, Badira, Korme, Chike, Galama, and Kaka. However, the reforestation program in Munessa, such as the establishment of the plantation forest together with the proper protection of the natural forest was believed to attract the Mountain Nyala back to the Munessa Forest from the Arsi

150

Mountains. Evangelista et al. (2007) also noted that the Mountain Nyala were believed to have arrived to the Munessa Forest from the Arsi Mountains. The respondents discussed important issues about the common diseases and parasites attacking both Mountain Nyala and their livestock in Munessa. They noted that ticks are one of the common parasites of both Mountain Nyala and livestock in Munessa. Other authors noted that ticks are potential problems for the management of threatened and endangered wildlife species (e.g. Hudson et al., 2002). Ticks feed on a wide variety of vertebrate hosts and transmit a range of parasites that cause suffering and mortality to both livestock and wildlife (Hudson et al., 2002). Mamo (2007) noted that one of the potential negative impacts of livestock on Mountain Nyala in the Bale Mountains National Park may be as a source of ecto-parasites such as ticks and flies that can easily serve as vectors for disease. Evangelista et al. (2007) also noted that risks of diseases and parasites to Mountain Nyala are likely to be greater in populations in close proximity to livestock. This suggests that ticks cause tick-borne infectious diseases in wildlife populations. Ticks are extensive in their blood sucking habits, so they can act as the principal disease reservoir and often carry a community of pathogens (Hudson et al., 2002). For example, management plan for bighorn sheep in California recommends a fifteen kilometer buffer zone to separate the endangered bighorn sheep populations from domestic sheep carrying potentially threatening infectious disease (Jessup et al., 1995). The local respondents noted that Spotted Hyena, Anubis baboon, Leopard, Caracal (Caracal caracal), Lammergeyer (Gypaetus barbatus), and Serval cat (Felis serval) are potential predators of Mountain Nyala in Munessa. The local people discussed that the Mountain Nyala become very vigilant when they live in a habitat with high risk of predation. This is because the predators prey on their calves. The local people also noted that the Mountain Nyala become very vigilant when they live in disturbed habitats. The present study revealed that the perception of the local community is that population size of Mountain Nyala seems to be decreasing in Munessa during the last decade. The focal group discussions held with key respondents also support this. For example, the local participants during the focal group discussion emphasized that the population of Mountain Nyala in Munessa is currently facing multi-faceted threats, such as competition for habitat resources due to free-range livestock grazing, habitat destruction

151

due to illegal tree cutting for fuel wood and construction materials, predation of calf and juvenile Mountain Nyala by predators, such as Spotted Hyena and Anubis baboon, and massive traditional illegal hunting of Mountain Nyala by the highlanders of the Arsi people. The focal group participants noted that the local people are eager to protect the Mountain Nyala in Munessa because they all believe that the Mountain Nyala have cultural values to local people and they are also the heritage of all Ethiopians. So, the participants emphasized that the federal and the regional governments should work hand in hand with the local people to better protect and manage the endangered Mountain Nyala population and its habitats in Munessa. Community support for increasing Mountain Nyala population size would be dramatically greater if local respondents were to be given economic benefit-sharing schemes as part of the management of Mountain Nyala and their habitats in the Munessa hunting block (Hillman, 1993; Tedla, 1995; Hundessa, 1997; Anderson et al. 2002; Mamo, 2007). For example, in Namibia and Zimbabwe, community-based wildlife management practices have brought significant employment and income-generating opportunities, empowerment, and governance impacts to some remote communities through wildlife- based tourism (LWAG, 2002). In fact, the focal group discussions revealed that some local people in Munessa have earned some economic benefits from the Mountain Nyala in the form of employment opportunities being employed in the Munessa hunting block. However, the benefit is not fairly distributed among the local people. Of course, all of the key community members admitted that they used to select and kill old male Mountain Nyala and distribute the bush meat among the local communities during extreme drought periods and economic crises. The focal discussants noted that Mountain Nyala generates high revenue to the regional and the national economy through tourism and legal trophy sport hunting. Providing that all the local people will get equitable benefits from Mountain Nyala in the future, the key community members strongly agreed that they should support the conservation and management of Mountain Nyala in Munessa. Promoting the direct participation of the local residents in decision-making and implementation of wildlife management can mitigate potential conflicts and assure long-term public support (e.g., Raick et al., 2003; Riley et al., 2003; Fulton et al., 2004).

152

Compared to other Munessa wild animal species such as Spotted Hyena and Anubis Baboon, the present sociological study revealed that Mountain Nyala is a less problematic species to the local people in terms of crop raiding. Behaviors of people can be influenced by increasing their knowledge (e.g., Kollmuss and Agyeman, 2002) suggesting that informing the local communities about the different values of Mountain Nyala through conservation education and advocating the need for conservation and sustainable utilization may increase the support of local people for the management activities of Mountain Nyala and their habitats in Munessa. More importantly, public awareness programs and conservation education can assist in improving the attitudes of young people toward Mountain Nyala (Tessema et al., 2007). Integrating indigenous knowledge of the local people should be one of the key components for better management and sustainable utilization of Mountain Nyala in the Munessa hunting block (Hillman, 1993; Agrawal, 1997; Anderson et al. 2002; Woodroffe et al., 2005; Lindsey, 2008). By comparing attitudes quantified in the baseline study presented here and results from future replication of this study, researchers may provide relevant information for wildlife conservation managers to deal with potential conflicts of interests between the conservation of the endangered Mountain Nyala in the Munessa hunting block and the surrounding local people. Based on the results of the present social study, I recommend that part of the revenue collected from the professional outfitter in the Munessa hunting block should be allocated to improve social infrastructures such as clinics, schools, clean water supply, etc. to the local people. This kind of conservation strategy and management plan will emphasize the economic value of wildlife resources to the local people through promoting benefit-sharing schemes. This will in turn encourage the local people to fully involve themselves in the conservation and management of Mountain Nyala in order to realize sustainable wildlife utilization in the Munessa hunting block. In addition, full responsibility should be given to the local people to conserve and manage the Mountain Nyala as their own property.

153

CHAPTER 8 CONCLUDING DISCUSSION AND RECOMMENDATIONS Concluding Discussion Habitat quality is a measure of the contribution of an area to individual fitness and population persistence (Van Horne, 1983; Block et al., 1998; Morris, 1998). However, for better practical understanding, habitat quality must be defined by the species and populations of interest, and in a manner that reflects underlying processes operating at appropriate spatial and temporal scales (Morris, 2003). The decline in habitat quality may often represent transformation from optimal toward unsuitable environmental conditions for survival and reproduction of a species (Evangelista et al., 2007; Morris et al., 2009). Habitat quality is affected by costs, such as risk of predation (e.g., Brown, 1988; Kotler et al., 1994) and benefits, such as availability of the right kinds of food (e.g. Gross et al., 1995), and cover associated with residing in that habitat (e.g., Brown and Alkon, 1990). Mountain Nyala is one of the endemic flagship species found in the south-eastern highlands of Ethiopia (Hillman and Hillman, 1987). Mountain Nyala are sexually dimorphic social animals. They are commonly reported to be found in a mosaic of high- altitude woodland, bush land, heathland, moorland, and valley bottom grassland, ranging from 2700 m up to 4300 m above sea level, but mostly prefer heathland and alpine habitats (Brown, 1969b; Yalden and Largen, 1992). However, their range is much broader than previously reported, and they are distributed across a variety of montane forest habitats. For example, Evangelista et al. (2007) observed Mountain Nyala at elevations as low as 1,600 m occupying dense forests. Mountain Nyala require two kinds of habitat, i.e. foraging grounds and cover in which to hide themselves from predators and shelter from extreme weather (Malcolm and Evangelista, 2005). Since its discovery in 1908, Mountain Nyala has become an important trophy species to sport hunting in Ethiopia (Lydekker, 1912; Sanford and Legendre, 1930; Mellon, 1975; Evangelista et al., 2007; Atickem et al., 2011). However, compared with its closest relatives such as the and the Nyala of southeastern Africa, the Mountain Nyala is less studied in Ethiopia (Brown, 1969a, 1969b; Hillman, 1985; Hillman and Hillman, 1987; Shuker, 1993; Woldegebriel, 1996; Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007; Evangelista et al., 2008; Atickem et al., 2011). The

154

overall aim of these studies was to examine the major environmental and anthropogenic factors affecting the habitat quality, habitat use, and foraging ecology of Mountain Nyala. The studies were carried out in the Munessa Forest and the Bale Mountains National Park, Ethiopia. Both study areas are found in the south-eastern highlands of Ethiopia, and they were primarily designated as wildlife protected areas to manage and conserve the endemic Mountain Nyala and other wildlife species of interest. Different approaches were used to accomplish the present studies. I would like to summarize and briefly discuss each approach as follows. Habitat suitability index (HSI) models and isodar analyses were applied to determine the seasonal habitat selection and habitat uses of Mountain Nyala in Munessa (Chapter 3). Separate habitat suitability models were developed for the wet and the dry season field data. The results revealed that Mountain Nyala did not show significant habitat selection behavior during the wet season. In addition, none the microhabitat variables considered for the HSI model significantly affected the habitat suitability of Mountain Nyala in the wet season. However, in the dry season, the natural forest was the most suitable habitat for Mountain Nyala, with crown diameters of trees significantly affecting the habitat suitability. The slopes of the isodar analyses also revealed that the natural forest habitat was qualitatively, but not quantitatively, better than either the plantation or the clear cut habitat during the dry season. Perhaps this is because the natural forest provides the Mountain Nyala with more diversified food, shelter from extreme weather, and cover from predation risk including human nuisance and livestock disturbances (e.g., Evangelista et al., 2007). However, the result with spotlight censusing suggested that Mountain Nyala selected the clear cut habitat during the night time when people and livestock are absent in the area. This showed that Mountain Nyala avoided human and livestock disturbances by becoming active mostly during the night time. Other authors noted that introduction of livestock changes the spatial and temporal distribution and habitat use patterns of indigenous wild herbivores (e.g., Austin and Urness, 1986; Wallace and Krausman, 1987; Augustine, 2004; Young et al., 2005; Kebede, 2009). The impacts of humans and livestock encroachments on the habitats of Mountain Nyala quantified in Chapter 6 supports the above scenario.

155

The habitat use of Mountain Nyala was further evaluated with behavioral indicators using time budgets of focal individuals stratified according to habitat type, group size, sex- age class, and seasons (Chapter 4). The time budget results revealed that Mountain Nyala allocated most of their time to vigilance behavior during the wet season in Munessa, but there was no significant difference in vigilance level across habitat types. This may be because there is high humans and livestock pressure imposed on Mountain Nyala and their habitats throughout Munessa (Chapter 6). Moreover, in the wet season, the behavioral model revealed that there is no significant difference in time allocation to vigilance among the different sex-age classes of Mountain Nyala. This suggests that all age and sex categories of Mountain Nyala are equally vulnerable to risk of predation, human nuisance, and livestock disturbances in Munessa. However, this is contrary to the classical predation risk hypothesis which predicts that adult females should allocate more time to vigilance and safety related activities than do adult males (e.g., Sukumar and Gadgil, 1988; Young and Isbell, 1991; Muller et al., 1995; Main et al., 1996; Prins, 1996; Ruckstuhl and Kokko, 2002; Ruckstuhl, 2007). The suggested reason is that, like adult females that have additional costs to care for their calves, adult males have additional costs from targeted legal hunting during the wet season in Munessa. The dry season behavioral results revealed that Mountain Nyala were most vigilant in the clear cut habitat. In contrast to my prediction, adult males were the most vigilant compared with the other sex and age categories. So, adult males allocated greater proportion of time to vigilance behavior than did adult females. The possible explanation for such departure from the predation risk hypothesis is that the targeted legal and illegal hunting pressure imposed on adult males in the Munessa hunting block forces them to allocate the greatest proportion of their time to vigilance. Group size of Mountain Nyala was significantly highest in the natural forest habitat during the dry season, corresponding to the highest activity density there; see also (Chapter 3). Regardless of season, the results revealed that group size did not affect the vigilance level of Mountain Nyala. Measuring and quantifying the natural food items eaten by free-ranging wild herbivores is crucial for habitat and species management (Hansen and Reid, 1975; Fitzgerald and Waddington, 1979). The cost of habitat use by free-ranging Mountain Nyala was assessed and determined through quantifying bite diameters of browsed twigs, bite

156

rates, and time budgets (Chapter 5). The study was carried out in the northern tip of the Bale Mountains National Park where the largest population size of Mountain Nyala is found (Refera and Bekele, 2004; Evangelista et al., 2007; Mamo, 2007). Two major habitat types were identified that the Mountain Nyala regularly use in the study area. These were grassland and woodland habitats (Refera and Bekele, 2004; Mamo, 2007). The results revealed that the bite diameters of free-ranging Mountain Nyala were significantly affected by habitat type, browse species, and the interactions of habitat type and browse species. For across habitat type comparison, larger mean bite diameter was measured when free-ranging Mountain Nyala were foraging in the grassland habitat, i.e., lower GUD. For across browse species comparison, the largest mean bite diameter of Mountain Nyala was measured for Solanum marginatum, possibly indicating greatest choice and use. The results further revealed that bite rates of Mountain Nyala were significantly affected by habitat type, sex-age categories, and the interactions of habitat type and sex-age categories. For across habitat type comparison, larger mean bite rates were quantified when free-ranging Mountain Nyala were foraging in the grassland habitat. For across sex-age classes’ comparison, the highest mean bite rates were measured and quantified for adult females. The results further showed that the vigilance level of Mountain Nyala was significantly affected by habitat type. Accordingly, higher mean vigilance level was measured and quantified when free-ranging Mountain Nyala were in the woodland habitat. In contrast, Mountain Nyala allocated their greater proportion of time to feeding when they were in the grassland habitat. Generally, the results in Chapter 5 revealed that bite diameters, bite rates, time feeding, and vigilance level of free-ranging Mountain Nyala are all significantly affected by habitat type. And, the means of bite diameters, bite rates, and proportion of time feeding were all higher for the grassland habitat. Likewise, the vigilance level was higher in the woodland habitat. This shows that the perceived risk of predation by free-ranging Mountain Nyala significantly differed between these two habitat types, suggesting that the grassland is a safer habitat. The reason may be that Mountain Nyala use open sightlines to detect potential predators early enough to allow for easy escape (Evangelista et al., 2007). Humans and livestock disturbances modify the habitats and natural behaviors of wild animals. For example, since Mountain Nyala was discovered by western science in

157

1908, its range has probably shrunk 10-20% due to deforestation and habitat destruction (Malcolm and Evangelista, 2005). In addition, human populations through settlement and agricultural expansion activities have considerably reduced the availability and quality of habitats for Mountain Nyala in Ethiopia (Evangelista et al., 2007; Mamo, 2007). Free- range livestock grazing has strong negative impacts on Mountain Nyala and their habitats in the Bale Mountains National Park (Stephens et al., 2001; Malcolm and Evangelista, 2005; Mamo, 2007). So, measuring the extent of humans and livestock disturbances on the quality and availability of habitats for Mountain Nyala is a contentious issue for ensuring their conservation (Stephens et al., 2001; Mamo, 2007). In this study, the impacts of humans and livestock encroachments on the habitats of Mountain Nyala were assessed and quantified in Munessa (Chapter 6). The results revealed that the activity density of livestock was significantly highest in the natural forest habitat in the wet season. In contrast, during the dry season, livestock did not show any significant difference in their habitat use. Overall, livestock activity densities in all habitat types were higher during the wet season than in the dry season. Regarding the impacts of humans on the habitats of Mountain Nyala, both in the wet and the dry seasons, the results revealed that the extent of stem and crown damages were significantly highest in the plantation habitat. Both in the wet and the dry seasons, the evidence of wood use and the numbers of stumps cut by humans was significantly highest in the natural forest. However, both in the wet and dry seasons, sign of habitat use by livestock did not differ among habitat types; rather it was dispersed throughout the available habitats of Mountain Nyala. In the wet season, the intensity of grazing / browsing by livestock was significantly heaviest in the natural forest habitat. However, intensity of livestock grazing / browsing did not differ among habitats in the dry season. Overall, the intensity of livestock grazing / browsing was heavier during the wet season, reflecting higher seasonal livestock densities. A possible reason for greater use of forest habitats in the wet season is that all the lands of the local people were occupied by crop cultivation so that the local people were forced to drive their livestock into the Munessa Forest for free- range grazing. However, in the dry season, the crops had been harvested and livestock had access to feed on crop residues. Generally, the results revealed that the impacts of humans and livestock disturbances were persistent throughout the habitats of Mountain Nyala in Munessa. As a

158

result, Mountain Nyala avoided both humans and livestock disturbances by becoming active during times (especially in the night time) when people and livestock were absent. This also suggests that humans and livestock disturbances negatively affect the availability and quality of habitats for Mountain Nyala as well as their behaviors. In their recent studies, Mamo (2007) and Atickem et al. (2011) also reported that Mountain Nyala avoided humans and livestock disturbances in the Bale Mountains. The behavioral time budget study conducted in Munessa (Chapter 4) suggests that the Mountain Nyala in Munessa are more sensitive to humans and livestock disturbances than they are in Bale Mountains National Park (Chapter 5). The greater time allocation to vigilance by the Mountain Nyala in Munessa should be the direct reflection of the severity of humans and livestock disturbances in the area. Moreover, the prevalence of legal and illegal hunting in Munessa makes the problem more chronic and complex (e.g., Ethiopian Wildlife Conservation Authority, unpublished data). As a result, the Mountain Nyala in Munessa are usually shy and when seen are usually fleeing from such disturbing agents. However, as there is relatively better control over humans and livestock disturbances in the Bale Mountains National Park (e.g., Atickem et al., 2011), the Mountain Nyala in the park are less shy. Moreover, the absence of both legal and illegal hunting incidence in the park is reflected in the habituation of these Mountain Nyala to humans and livestock presence in the area (e.g., Evangelista et al., 2007). Contrary to adult male Mountain Nyala in Munessa, adult males in the Bale Mountains National Park devoted the least time to vigilance behavior compared with the other sex-age categories (see Chapter 5, Fig. 5.8). This shows how the prevalence of human and livestock induced disturbances as well as legal and illegal hunting pressures affected the behaviors of Nyala in Munessa. The distance at which individuals respond to disturbance is often thought to indicate their sensitivity to disturbance (e.g., Mori et al., 2001): individuals that flush at greater distances are more sensitive than individuals that do not flush until the disturbance is near (e.g., Stillman et al., 2007). In all instances for my fieldwork with these two distinct populations of Mountain Nyala, the flush distances in Munessa were farther (approximately a minimum flush distance of 50 meters) than in the Bale Mountains National Park (approximately a minimum flush distance of 25 meters). This suggests that the Mountain Nyala in Munessa are more sensitive to humans and livestock disturbances

159

than the Mountain Nyala in the Bale Mountains National Park. This corresponds to further field studies indicating that the Mountain Nyala in Munessa are behaviorally more apprehensive than the Mountain Nyala in the Bale Mountains National Park (e.g., Ethiopian Wildlife Conservation Authority, unpublished field reports). As a consequence of hunting, the Mountain Nyala in Munessa are sensitive to disturbance and shift their activity to times (especially in the night time) when people and livestock are absent. The results from the spotlight in Chapter 3 and the humans and livestock disturbances in Chapter 6 support this contextual interpretation. Regardless of the presence of people and livestock, the Mountain Nyala in the Bale Mountains National Park are active throughout the daylight hours. Atickem et al. (2011) found that the probability of habitat use by Mountain Nyala was higher in areas that were well patrolled by wardens, e.g. Gaysay area, than the open wildlife areas. The attitudes of local people toward Mountain Nyala and their habitats were studied through interview questionnaires and focal group discussions in Munessa (Chapter 7). Wildlife conservation in developing countries with high numbers of people living in poverty is often viewed as being in conflict with development goals. The questionnaires were developed considering the different demographic and socio-economic characteristics of the local people, such as marital status, family size, sex, age, length of local residence, livelihood strategy, land ownership, livestock ownership, and knowledge about Mountain Nyala and their habitats. The results revealed that the attitudes of local people toward Mountain Nyala and their population increase were significantly affected by several socio- economic variables including livelihood strategy, land ownership, livestock ownership and knowledge, as expected. However, contrary to my expectation, demographic characteristics, such as sex, age, marital status, family size, and length of residence in the area did not affect the attitudes of the local people toward Mountain Nyala and their population increase. Through focal group discussions, key members of the local communities shared their abundant indigenous knowledge about the different behaviors of Mountain Nyala and their habitats in Munessa. Overall, the results suggest that the social methodologies can be used to improve the limited knowledge on human dimensions that then later can be included in the conservation strategy and management plans.

160

Legal trophy sport hunting in the Munessa hunting block generates revenue. When sustainably implemented, trophy sport hunting has the potential to provide economic incentives for conservation (Lindsey et al., 2007; Lindsey, 2008), including economic benefits to local communities (Atickem et al., 2011). This kind of conservation strategy helps emphasize the economic value of wildlife resources to the local people through promoting benefit-sharing schemes (Kubsa, 1999; Evangelista et al., 2007). In principle, revenues generated by safari hunting are shared between the federal and the regional governments in Ethiopia. For example, the federal government receives 15% of the license fees, while each regional government receives 85% of revenues generated by sport hunting licenses and 100% of the concession fees (Kubsa, 1999). Of the hunting revenues received by the regions, 20% are allocated to local communities residing in the hunting concessions (Kubsa, 1999). The use of licensed sport hunting as a means of conservation may also allow the wild animals in the hunting areas to make dramatic recoveries (Malcolm and Evangelista, 2005; Atickem et al., 2011) because part of the money can be allocated to promote infrastructure developments that enhance wildlife and habitat conservation activities (Evangelista et al., 2007; Atickem et al., 2011; EWCO, unpublished field reports). However, in the case of the Munessa hunting block, nothing goes to the local people in the form of direct benefit-sharing. Many local people in wildlife areas do not legally receive economic benefits; however, they bear the costs of living with wildlife (e.g., Kiss, 1990; Tessema et al., 2007). In the present study, most of the respondents wished to see both Mountain Nyala and their habitats protected; however, they expressed their frustration over the little economic benefit they received from Mountain Nyala in Munessa. This suggests that the absence of economic benefit-sharing to the local people has created an unfriendly relationship between the local people and the conservation of Mountain Nyala and its habitats in the Munessa hunting block. A rapid decline of wildlife has been documented in areas where economic benefits are not accrued to the local community (e.g., Hillman, 1993; Tedla, 1995; Tessema et al., 2007). So, some forms of economic incentives to encourage tolerance may include compensation, performance payments, and both consumptive and non-consumptive use of wildlife (e.g., Agrawal, 1997; Woodroffe et al., 2005).

161

Ensuring continued local support for wildlife conservation over the long term needs proactive programs of sustainable economic benefit-sharing and local awareness of conservation values (Tessema et al., 2007). The result from the social part of the current study suggested that part of the revenue collected from the professional outfitter in the Munessa hunting block should be allocated to improve social infrastructures such as clinics, schools, clean water supply, etc. to the local people. This kind of management strategy will encourage the local people to fully involve in the conservation and management of Mountain Nyala so that sustainable utilization will be realized in the area (Hillman, 1993; Agrawal, 1997; Woodroffe et al., 2005). During my fieldwork, I observed that conservation facilities in the Munessa hunting block are almost non-existent. So, part of the revenue collected from the professional outfitter should be allocated to improve management facilities for the hunting block. To summarize, I believe that the different approaches used in the present studies give a general picture about the conservation and management of Mountain Nyala. The results obtained from Chapters 3, 4, and 5 are relevant to knowledge gained about the habitat quality, behaviors, and foraging ecology of Mountain Nyala. The present study provides measures of habitat suitability and habitat selection that are based on behavioral variables (using activity densities and time budgets as behavioral assays). Because behavior is adaptive, the resulting measures are leading indicators of habitat change and can form the basis for a more proactive management approach. These can then be used for monitoring and evaluating the impacts of both predicted and actual anthropogenic habitat alterations. However, the traditional habitat suitability models are constructed based on population densities which are liable to human perceptions, observations, and scales of references and may even make use of expert knowledge (e.g., Stoms et al., 1992; Store and Kangas, 2001; Yamada et al., 2003) or best estimates (e.g., Burgman et al., 2001). The intensity of resource patch use provides a measure of habitat quality (Olsson and Molokwu, 2007). However, it is often not feasible to train individuals of target species to use artificial food patches. In such cases, using natural measures of patch use would be preferable. In the current study, I developed techniques for measuring natural giving-up- densities (GUDs) using measurement of bite size based on twig diameter of branches browsed by free-ranging Mountain Nyala on natural forage species. The diameters of

162

browsed twigs are in effect natural giving-up-densities, and can be used just like GUDs measured from artificial food patches to compare food choice, compare habitat quality, look for impact of risk of predation and human disturbance, and track changes in range quality. Such measures are especially important in that they provide a measure of range quality perceived by the foraging Mountain Nyala. The field information may also direct habitat conservation priorities to the habitats that require greater levels of protection from humans and livestock use (Evangelista et al., 2007; Mamo, 2007). Most importantly, natural giving-up densities quantified through the measurement of bite diameters coupled with bite rates and time budgets of free-ranging Mountain Nyala provide better understanding of species habitat selection from the animals’ perspective rather than ours, as they allow organisms to demonstrate their own choices and costs. They also have the potential to provide leading indicators of change (e.g., Kotler et al., 2001, Morris et al., 2009). Behavioral indicators based on foraging theory are often fast, inexpensive, accurate, and simple to implement. Therefore, the use of natural giving-up densities as behavioral assays would allow wildlife managers to objectively ascertain when species are under shortage of food and allow for preventive management actions, rather than taking actions only once the problem emerges. This is because natural giving-up densities are not observer dependent and hence provide consistent results across habitats and management regime. Although the use of natural giving-up densities requires some skills, the method generates large data sets in a relatively short period of time, with relatively moderate efforts. Natural giving-up densities are easier and cheaper to use than the traditional methods such as measuring demographic parameters or assessing movement using GPS collaring, and also provide immediate indication of conditions faced by the study animals. The results presented in Chapter 6 help to assess, quantify, and understand the impacts of humans and livestock encroachments on the habitats and behaviors of Mountain Nyala. The results from Chapter 7 help appraise, measure, and acknowledge the indigenous knowledge and attitudes of local people toward Mountain Nyala and their habitats. Combining the entire knowledge obtained through these five related Chapters can help policy makers and wildlife managers to formulate an informed conservation strategy and management plan that gives the needed consideration to the ecology and habitats of

163

Mountain Nyala while still taking into account the needs and wants of the local people in the management processes. This will ultimately help establish ecologically sustainable, economically feasible, and socially acceptable conservation and management systems for the endangered Mountain Nyala in Ethiopia.

Recommendations Compared with the Mountain Nyala in Munessa, the population of Mountain Nyala in the Bale Mountains National Park (BMNP) has been relatively well studied by other scholars (e.g., Brown, 1969a, 1969b, Hillman, 1985, Hillman and Hillman, 1987; Hillman, 1988; Woldegebriel, 1996; Stephens et al., 2001; Refera and Bekele, 2004; Mamo, 2007; Atickem et al., 2011). In addition, there are also several unpublished field reports on Mountain Nyala in the BMNP by the staff of the Ethiopian Wildlife Conservation Authority. So, my recommendations solely focus on the little studied population of Mountain Nyala in Munessa. I believe that the present studies have implications for future conservation and management of Mountain Nyala and their habitats. Based on the results of these studies, I would like to recommend the following. The present studies revealed that Mountain Nyala are facing several humans and livestock induced challenges which likely threaten their fitness. Introducing regular patrolling activities in the Munessa hunting block may improve the conservation and protection of Mountain Nyala and their habitats through controlling poaching of Mountain Nyala and illegal habitat destruction activities. Particularly, the forest management in the Munessa Forest District should give due emphasis to control illegal tree cutting activities for fuel wood and construction materials, free-range livestock grazing, illegal seasonal settlement inside the Munessa Forest by the highlanders, and pit-sawers who illegally harvest trees both in the natural and the plantation forests. Maintenance and management of the forests enhances the availabilities and qualities of habitats for Mountain Nyala. Zonation of the core habitats may reduce and/or avoid humans and livestock disturbances on Mountain Nyala and their habitats. Introducing and promoting community-based conservation efforts that allow communities to derive economic benefits from ecotourism may promote conservation while at the same time providing a solution to resource use conflicts. Ecotourism activity

164

can also improve and diversify the incomes of the local people through creating job opportunities, such as tourist guiding services, souvenir selling, horse renting, etc., all of which can help make ecotourism economically viable in Munessa. For example, Fetene et al. (2012) noted that ecotourism is an important industry to create self-employment opportunities for the local community and also to enhance greater partnership for sustainable management of wildlife protected areas. Improving tourist facilities in Munessa helps promote wildlife-based tourism. Munessa is rich in biodiversity resources including mosaic of landscapes, natural and plantation forests, mosaic of cultures and traditions, the presence of Rift Valley lakes in the vicinity of Munessa, the rich variety of bird species, and the presence of different wild mammal species including the large endemic Mountain Nyala and Menelik's Bush Buck. All of these are top tourist attractions in Munessa. Most importantly, Munessa is situated very close to the famous tourist destination, Lake Langano. A number of comfortable private safari lodges have been built around Lake Langano (e.g., Bishan Gari lodge, Wenney lodge, Sabbana beach resort, etc.). Munessa is also found at the junction of several routes that tourists today typically follow when travelling to other attractions in the south and south-eastern parts of Ethiopia. This makes it an attractive location for the establishment of more ecotourism opportunities. So, developing Munessa for wildlife- based tourism seems to be a promising business in the future. Developing hiking trails and interpretive materials including field guides to birds, mammals, and plants would be valuable assets for ecotourism. The supplementary questions addressed to the local people during the social survey revealed that poverty is one of the rampant problems around the Munessa hunting block. Thus, future conservation and management activities should be geared toward the mutual benefits of local people and the conservation of Mountain Nyala population in Munessa. Introducing and advocating economic benefit-sharing systems with full participation of the community in the conservation and management processes could be helpful in this regard. For example, part of the revenue collected from legal sport hunting could be allocated to develop social infrastructures such as clinics, roads, clean water supplies, schools, etc. to the local people living around the Munessa hunting block. However, there should be an equitable benefit-sharing system to all local people residing in the area. Full responsibility

165

should be given to the local people to conserve and manage the Mountain Nyala as their own property. Developing a comprehensive management plan whose purpose is to guide and promote conservation and sustainable utilization of wildlife in the Munessa hunting block is mandatory. For example, the regional and the federal governments should work hand in hand with the local people to promote conservation and participatory management operations toward Mountain Nyala and their habitats. Promoting scientific research strategies and initiatives toward Mountain Nyala and their habitats may help devise a good conservation and management approach toward the sustainable utilization of Mountain Nyala. The habitat conditions for Mountain Nyala can be improved by implementing appropriate management prescriptions. For example, providing salty soils “Hora” to Mountain Nyala especially in the dry season is a possible habitat management activity. This is because the natural salt licks in Munessa become very hard for the Mountain Nyala to lick during the dry season. In order to meet their sodium requirements, Mountain Nyala often go to Lake Langano (which is situated several kilometers from Munessa) to drink salty water in the dry season during the night time (information from focal group discussion with key community members). However, on their way to Lake Langano, the Mountain Nyala may be susceptible to illegal hunting and predation. In addition, the beach of Lake Langano is already occupied by local settlers so that the Mountain Nyala are denied their access to Lake Langano (information from focal group participants). Habitat management may also need to include creating forest openings to make available foraging grounds for Mountain Nyala especially in the wet season. The social part of the present studies revealed that the rapid increase in the numbers of Anubis Baboon and Spotted Hyena is becoming a serious threat to the survival of both the endangered Mountain Nyala and the livestock of the local people in Munessa. Anubis Baboon and Spotted Hyena are thought to kill calf and/or juvenile Mountain Nyala. I first recommend that the numbers of potential predators in the core habitats of the endangered Mountain Nyala should be quantified. If their numbers are beyond the carrying capacity of the habitats, their number should be regulated. For example, every predator that preys on readily surveyed wild prey species can have its carrying capacity predicted and determined

166

based on the abundance and biomass of its preferred wild prey species (Hayward et al., 2007 and the references there in). Knowledge on the biomass of the preferred prey is important because a single predator species may have more than one preferred prey species whose biomass largely varies with the body size of the prey species. Where the prey densities are known, Hayward et al. (2007) developed the methods and equations to estimate the carrying capacity of Africa’s large predators in desert, savannah, and thicket biomes of South Africa. Permanently assigning qualified and experienced wildlife professionals who are capable to carry out wildlife research, monitoring, habitat inventory, etc. activities in the Munessa hunting block is necessary. Provisioning regular training opportunities could build the capacity and also improve the efficiency of the wildlife professionals. Moreover, recruiting wildlife scouts among the local people can strengthen the control over illegal activities against Mountain Nyala and their habitats in Munessa. However, the wildlife scouts should get training opportunities at regular intervals because the training improves their performance. Allocating sufficient running budgets and also providing vehicle support would improve regular wildlife patrolling, monitoring, habitat inventory, and wildlife research activities. Behavioral indicators should be integrated into this part. Training the wildlife scouts in these methods and using them regularly should provide the data necessary to managers. An integration of indigenous knowledge with modern conservation approaches in the planning and implementation process is crucial to improve and promote local participation in conservation and management. Local knowledge not only provides relevant information on the use of the species, but also contributes valuable information on how to maintain and conserve the Mountain Nyala and their habitats in the Munessa hunting block. Effective conservation and sustainable use of Mountain Nyala and their habitats, therefore, needs the full involvement of many stakeholders including local communities. The critical issue concerning the conservation of the Munessa Forest and wildlife populations is the development of alternative livelihoods for the local communities who largely depend on forest and forest products. Poverty is the major problem in the area and hence development strategies that address both poverty alleviation and sustainable

167

utilization of the forest are required. For example, the local communities can be supported through enhancing non-timber forest products so that their income will be improved. Establishing management systems involving the government and local communities, including certification of the forest and wildlife, makes it possible for local authorities to better monitor natural resource utilization and conservation. Sustainable use of non-timber forest products, such as wild honey collection and medicinal plant extraction should be supported and encouraged by the federal and regional governments as well as by non- governmental organizations. The most important component of forest and wildlife conservation approach is the rehabilitation of denuded areas in and around the Munessa Forest. On these degraded areas, multi-purpose tree species can be planted and managed so that it creates alternative livelihoods for the local communities. At the same time, these areas buffer the Munessa Forest and wildlife conservation zone and also promote sustainable use of natural resources. At a national level, Ethiopia should develop and implement land-use policies that promote land uses according to land suitability and characteristics. Any rural development strategy should focus toward multi-faceted approaches which consider rural development based on the carrying capacity of the natural environment. Issuing good wildlife policy could also guide and enhance conservation and management of Mountain Nyala and their habitats in the Munessa hunting block.

Proposed Future Researchable Areas As the present studies are limited in their scopes to cover many aspects of Mountain Nyala, I therefore recommend that future research directions should give due emphasis to some of the following researchable areas. Small populations are generally at risk of extinction due to genetic (e.g., inbreeding depression, drift), environmental (e.g., catastrophes), demographic (e.g., single sex biased ratio), and/or anthropogenic factors (e.g., hunting) (Primack, 2002). Small populations may also have their fitness severely affected by all of those factors. The population size of Mountain Nyala in Munessa is small, probably not exceeding 200 individuals (Evangelista et al., 2007). Thus, the population of Mountain Nyala in Munessa is inevitably at risk of

168

extinction in the future due to one or the combined effects of the aforementioned small population problems. So, the genetics of the Mountain Nyala population in Munessa should be studied and documented so as to enhance future management prescriptions and operations. Population dynamics and demography of Mountain Nyala should be well studied because this knowledge is important to plan and practice sustainable trophy sport hunting in the Munessa hunting block. For example, East (1988) noted that demographic information has become one of the crucial tools in the conservation of endangered antelopes. Carrying out population viability analysis (PVA) may also be complementary if due consideration is given to the different scenarios such as demographic and/or environmental stochasticity, management regime, and disease or the different combinations of the aforementioned. Empirical knowledge on energetic and reproductive costs of humans and livestock disturbances on the two distinct populations of Mountain Nyala in the Munessa Forest and the Bale Mountains National Park is indispensable, but lacking. So, future research should be directed to quantify energetic and reproductive costs of humans and livestock disturbances on these two distinct populations of Mountain Nyala in Ethiopia. The social study revealed that Mountain Nyala seasonally move long distances out of the Munessa Forest. For example, most of the respondents (59.03%) noted that Mountain Nyala usually go to Lake Langano in search of salty water “Hora” during the dry season. So, movement types (i.e., daily and seasonal movements), patterns, and ranges by Mountain Nyala should be studied in the future possibly with GPS collaring or marking- recapturing technique. The outputs of such a study may help in designing corridors.

169

REFERENCES Agrawal, A. 1997. Community in conservation: beyond enchantment and disenchantment. Conservation and Development Forum, Gainesville. Altendorf, K.B., Laundré, J.W., Gonzalez, C.A.L., and Brown, J.S. 2001. Assessing effects of predation risk on foraging behavior of mule . J. Mam. 82:430 – 439. Altmann, J. 1974. Observation study of behavior: sampling methods. Behavior. 49: 227 – 265. Altmann, S.A. 1998. Foraging for survival: yearling baboons in Africa. Chicago: Chicago University Press. Anderson, T.J., Forcey, M.G., Osbourne, D.J., and Spurgeon, B.A. 2002. The importance and use of wildlife management plans: an example from the camp Dawson collective training area. West Virginia Academy of Sciences.74:8 – 17. Arenz, C.L. and Leger, D.W. 1997. Artificial visual obstruction, anti-predator vigilance, and predator detection in the Thirteen-Lined Ground Squirrel (Spermophilus tridecemlineatus). Behavior. 134:1101 – 1114. Assefa, A. 2003. Dispersal agent of Solanum marginatum jacq. in the Bale Mountains National Park. Addis Ababa, Ethiopia. Atickem, A., Loe, L.E., Langangen, Ø., Rueness, E.K., Bekele, A., and Stenseth, N.C. 2011. Estimating population size and habitat suitability for Mountain Nyala in areas with different protection status. Animal Conservation. The Zoological Society of London. 14: 409 – 418. Augustine, D.J. 2004. Influence of cattle management on habitat selection by on Central Kenyan Rangeland. Journal of Wildlife Management. 68(4): 916-923. Austin, D.D. and Urness, P.J. 1986. Effects of cattle grazing on mule deer diet and area selection. J. Range Manage. 39: 18 – 21. Bath, A.J. 1989. The public and wolf reintroduction in Yellowstone National Park. Society and Natural Resources. 2: 297 – 306. Beale, C.M. 2007. The behavioral ecology of disturbance responses. International Journal of Comparative Psychology. 20:111 – 120. Beauchamp, G. 2003. Group size effects on vigilance: a search for mechanisms. Behav. Process. 63: 111 – 121.

170

Beier, P. 1987. Sex differences in quality of white-tailed deer diets. Journal of Mammalogy. 68: 323 – 329. Belanger, L. and Bedard, J. 1990. Energetic cost of man-induced disturbance to staging snow geese. Journal of Wildlife Management. 54: 36 – 41. Bell, D.V. and Owen, M. 1990. Shooting disturbance-A review. Pages 159 – 171 in Managing Waterfowl Populations: Proceedings of the IWRB Symposium, Astrakhan 1989 (G.V.T. Matthews. Ed.). IWRB Special Publication 12. International Waterfowl and Wetlands Research Bureau, Slimbridge, U. K. Berger-Tal, O., Polak, T., Oron, A., Lubin, Y., Kotler, B.P., and Saltz, D. 2011. Integrating animal behavior and conservation biology: a conceptual framework. Behavioral Ecology. 22(2):236 – 239. Bernard, H.R. 2002. Research methods in anthropology: qualitative and quantitative approaches.3rd ed. Lanham, Md. Altamira Press. Berry, L. 2003. Land degradation in Ethiopia: its extent and impact. A case study prepared for the Food and Agriculture Organization of the United Nations. Retrieved from http://www.fao.org. Black, H., Scherzinger, R.J., and Thomas, J.W. 1976. Relationships of rocky mountain elk and rocky mountain mule deer habitat to timber management in the Blue Mountains of Oregon and Washington. Pages 11 – 13 in S.R. Heib, editor. Proceedings of the Elk-Logging-Roads Symposium. Forest, Wildlife, and Range Experiment Station, University of Idaho, Moscow, Idaho, USA. Block, W.M., Morrison, M.L., and Scott, P.E. 1998. Development and evaluation of habitat models for herpetofauna and small mammals. Forest Science. 44(3):1 – 8. Boshoff, A.F. and Kerley, G.I.H. 2001. Potential distributions of the medium- to large- sized mammals in the Cape Floristic Region based on historical accounts and habitat requirements. African Zoology. 36: 245 – 273. Boshoff, A.F., Kerley, G.I.H., and Cowling, R.M. 2002. Estimated spatial requirements of the medium- to large-sized mammals, according to broad habitat units, in Cape Floristic Region, South Africa. African Journal of Range and Forage Science. 19: 29 – 44.

171

Bowman, J.L., Leopold, B.D., Vilella, F.J., and Gill, D.A. 2004. A spatially explicit model derived from demographic variables to predict attitudes toward black bear restoration. Journal of Wildlife Management. 68: 223 – 232. Bowman, J.L., Leopold, B.D., Vilella, F.J., Gill, D.A., and Jacobson, H.A. 2001. Attitudes of landowners toward American black bear compared between areas of high and low bear populations. Ursus. 12:153 – 160. Boyce, M. and McDonald, L. 1999. Relating populations to habitats using resource selection function. Trends in Ecology and Evolution. 14:268 – 272. Bright, A. D. and Manfredo, M.J. 1996. A conceptual model of attitudes toward natural resource issues: A case study of wolf reintroduction. Human Dimensions of Wildlife. 1: 1 – 21. Brooks, T. and Balmford, A. 1996. Atlantic forest extinctions. Nature. 380: 115 – 116. Brown, J.S. 1988. Patch use as an indicator of habitat preference, predation risk, and competition. Behavioral Ecology and Sociobiology. 22:37 – 47. Brown, J.S. 1992. Patch use under predation risk: models and predictions. Ann Zool Fenn. 29:301 – 309. Brown, J.S. 1999. Vigilance, patch use, and habitat selection: foraging under predation risk. Evolutionary Ecology. 1: 49 – 71. Brown, J.S. and Alkon, P.U. 1990. Testing values of crested porcupine habitats by experimental food patches. Oecologia. 83: 512 – 518. Brown, J.S. and Kotler, B.P. 2004. Hazardous duty pay and the foraging cost of predation. Ecol. Lett. 7: 999 – 1014. Brown, J.S., Arel, Y., Abramsky, Z., and Kotler, B.P. 1992. Patch use by gerbils (Gerbillus allenbyi) in sandy and rocky habitats. J.Mam. 73:821 – 829. Brown, J.S., Kotler, B.P., and Bouskila, A. 2001. Ecology of fear: foraging games between predators and prey with pulsed resources. Ann. Zool. Fennici 38: 71 – 87. Brown, L. 1969a. Ethiopia’s elusive Mountain Nyala. Animals. 12: 340 – 344. Brown, L. 1969b. Observation of the status, habitat, and behavior of Mountain Nyala, Tragelaphus buxtoni in Ethiopia. Mammalia. 33:545 – 597. Burger, J. 1981. The effect of human activity on birds at a coastal bay. Biological Conservation. 21: 231 – 241.

172

Burgman, M.A., Breininger, D.R., Duncan, B.W., and Ferson, S. 2001. Setting reliability bounds on habitat suitability indices. Ecological Applications. 11: 70 – 78. Burnham, K.P. and Anderson, D.R. 2002. Model selection and multi-model inference: a practical information-theoretic approach. 2nd edition. Springer. pp488. Bussman, R. W. 1997. The forest vegetation of the Harrena escarpment (Bale Province, Ethiopia) syntaxonomy and phytogeographical affinities. Phytocoenologia. 27(1):1 – 23. Carolina, A. H and Javier, A. S. 2001. The effect of introduced herbivores upon an endangered tree (Beilschmiedia miersii, Lauraceae). Bio. Conservation. 98: 69 –76. Conradt, L. and Roper, T.J. 2000. Activity synchrony and social cohesion: a fission-fusion model. Proceedings of the Royal Society of London Series B-Biological Sciences. 267:2213 – 2218. Conway, C.J. and Martin, T.E. 1993. Habitat suitability for Williamson’s sapsuckers in mixed conifer forests. Journal of Wildlife Management. 57: 322 – 328. Cortina, J.M. 1993. What is coefficient alpha? An examination of theory and applications. Journal of Applied Psychology. 78:98 – 104. Cresswell, W. 1994. Flocking is an effective anti-predation strategy in Red Shanks (Tringa tetanus). Animal Behavior. 47:433 – 442. Dall, S. and Valone, T. 2005. Defining the concept of public information. Science, Vol. 308. No. 5720:353 – 356. Dejene, A. 2003. Integrated natural resources management to enhance food security: The case for community-based approaches in Ethiopia. Food and Agriculture Organization, Rome. Pp 58. Done White, J., Kendall, K.C., and Picton, H.D. 1999. Potential energetic effects of mountain climbers on foraging Grizzly Bears. Wildlife Soc. Bulletin. 27:146 – 151. Druce, D.J. 2005. Species requirements, coexistence and habitat partitioning at the community level: Rock Hyrax and . PhD dissertation. Pp 122. University of KwaZulu-Natal, Durban- South Africa. Druce, D.J., Brown, J.S., Castley, J.G., Kerley, G.I.H., Kotler, B.P., Slotow, R., and Knight, M.H. 2006. Scale-dependent foraging costs: habitat use by rock hyraxes (Procavia capensis) determined using giving-up-densities. Oikos.115:513 – 525.

173

Druce, D.J., Brown, J.S., Kerley, G.I.H., Kotler, B.P., Mackey, R.B., and Slotow, R. 2009. Spatial and temporal scaling in habitat utilization by determined using giving-up-densities. Austral Ecology. 34(5):577 – 587. East, R. 1988. Antelope global survey and regional action plans. Part I: East and North East Africa. Gland: IUCN/ SSC Antelope Specialist Group. Elgar, M.A. 1989. Predator vigilance and group size in mammals and birds. Biol. Rev. 64: 13 – 33. Ethiopian Forestry Action Program (EFAP).1994. Final report. Vol. 3-Issues and Actions. EFAP Secretariat, Addis Ababa.

Evangelista, P., Norman, J., Lakew, B., Kumar, S., and Alley, N. 2008. Predicting habitat suitability for the endemic Mountain Nyala in Ethiopia. Wild. Res. 409 – 416. Evangelista, P., Swartzinski, P., and Waltermire, R. 2007. A profile of the Mountain Nyala (Tragelaphus buxtoni). African Indaba. 5(2).Pp 47. Fabricius, C. and Mentis, M.T. 1991. The use of habitat suitability models in game ranch management. In: Renecker, L.A. and Hudson, R.J. (Eds.) Wildlife production: conservation and sustainable development. AFES misc. pub. 91 – 6. Univ. of Fairbanks, Alaska, USA. Fabrigar, L.R., Wegener, D.T., MacCallum, R.C., and Strahan, E.J. 1999. Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods. 4:272 – 299. Farmer, A.H., Armbruster, M.J., Terrell, J.W., and Schroeder, R.L. 1982. Habitat models for land use planning: assumptions and strategies for development. Transactions of the North American Wildlife and Natural Res. Conference. 47: 47 – 56. Fernandez-Juricic, E. and Telleria, J.L. 2000. Effects of human disturbance on the spatial and temporal feeding patterns of Blackbird (Turdus merula) in urban parks in Madrid, Spain. Bird Study. 47:13 – 21. Fetene, A., Bekele, T., and Tiwari, P.K. 2012. The contribution of ecotourism for sustainable livelihood development in the Nech Sar National Park, Ethiopia. International Journal of Environmental Sciences. 1(1): 19 – 25.

174

Fish and Wildlife Services. 1980. Standards for the development of suitability index models. Ecological service manual 103. US Department of the Interior, Fish and Wildlife Service, Division of Ecological Services. Washington, DC.68. Fitzgerald, A.E. and Waddington, D.C. 1979. Comparison of two methods of fecal analysis of herbivores. Journal Wildlife Management. 43: 468 – 470. Foley, J.A., DeFries, R., Asner, P.G., Barford, C., Bonan, G., Carpenter, R.S., Chapin, S. T., Coe, T.M., Daily, C.G., Gibbs, K.H., Helkowski, H.J., Holloway, T., Howard, A.E., Kucharik, J.C., Monfreda, C., Patz, A.J., Prentice, C.I., Ramankutty, N., and Snyder, K.P. 2005. Global consequences of land use. Science. 309: 570 – 574. Folke, C., Holling, C.S., and Perrings, C.A. 1996. Biological diversity, ecosystems, and the human scale. Ecological Applications.6: 1018 – 1024. Forman, T.T. 1996. Land mosaics. The ecology of landscapes and regions. Cambridge University Press. New York. Fretwell, S. D. and Lucas, H. L. 1969. Territorial behavior and other factors influencing habitat distribution in birds. I. Theor.Deve. Acta Biotheor. 19:16 – 36. Frid, A. and Dill, L.M. 2002. Human-caused disturbance stimuli as a form of predation risk. Conservation Ecology, 6, http://www.consecol.org / Journal / vol6 / iss1 /art11 / print.pdf. Fulton, D.C., Skerl, K., Shank, E.M., and Lime, D.W. 2004. Beliefs and attitudes toward lethal management of deer in Cuyahoga Valley National Park. Wildlife Society Bulletin. 32:1166 – 1176. Gashaw, M. and Fetene, M. 1996. Plant communities of afro alpine vegetation of Sanetti Plateau, Bale Mountains, Ethiopia. SINET: Eth. J. Sci. 19: 65 – 86. Gendall, P. 1998. A framework for questionnaire design: Labaw revisited. Marketing Bulletin. 3(9): 28 – 39. Gill, J.A. 1996. Habitat choice in wintering pink-footed geese: quantifying the constraints determining winter site use. Journal of Applied Ecology. 33: 884 – 892. Gill, J.A. and Sutherland, W.J. 2000. Predicting the consequence of human disturbance from behavioral decisions. Behavior and Conservation (eds L.M. Gosling & W.J. Sutherland), pp. 51–64. Cambridge Univ. Press, Cambridge, UK.

175

Gill, J.A., Norris, K., and Sutherland, W.J. 2001. Why behavioral responses may not reflect the population consequences of human disturbance. Bio. Cons. 97:265 – 268. Gill, J.A., Sutherland, W.J., and Watkinson, A.R. 1996. A method to quantify the effects of human disturbance on animal populations. J. App. Eco. 33: 786 – 792. Gilroy, J.J. and Sutherland, W.J. 2007. Beyond ecological traps: perceptual errors and undervalued resources. Trends in Ecology and Evolution. 22: 351 – 356. Ginnett, T.F. and Demment, M.W.B. 1997. Sex differences in foraging behavior at two spatial scales. Oecologia.110:291 – 300. Gliem, J.A. and Gliem, R.R. 2003. Calculating, interpreting, and reporting Cronbach’s alpha reliability coefficients for Likert-type scales. Paper presented at the Midwest Research-to-Practice Conference in Adult, Continuing, and Community Education, Ohio State University, Columbus, OH, Oct. 8 – 10, 2003. Gliner, J.A., Vaske, J.J., and Morgan, G.A. 2001. Null hypothesis significance testing: Effect size matters. Human Dimensions of Wildlife. 6: 291 – 301. Green, S.B. 1991. How many subjects does it take to do a regression analysis? Multivariate behavioral Research. 26 (3): 499-510. Gross, J. E., Alkon, P. U., and Demment, M. W. 1995. Grouping patterns and spatial segregation by Nubian Ibex. Journal of Arid Environments. 30:423 – 439. Gutzwiller, K.L. and Anderson, S.H. 1986. Improving vertebrate-habitat regression models. In: Verner J., Morrison M.L. and Ralph, and C. J. (Eds) Wildlife 2000: modeling habitat relationships of terrestrial vertebrates. University of Wisconsin Press, Wisconsin. 161 – 164. Hague, P. 1993. Questionnaire design. London: Kogan Page. Hansen, R.M. and Reid, L.D. 1975. Diet overlap of deer, elk, and cattle in southern Colorado. Journal of Range Management. 28(1):43 – 47. Harrison, M.J.S. 1983. Age and sex differences in the diet and feeding strategies of the green monkey, Cercopithecus sabaeus. Animal Behavior. 31: 969 – 977. Haschick, S.L. and Kerley, G.I.H. 1997a. Browse intake rates by bushbuck (Tragelaphus scriptus) and Boer goats (Capra hircus). Afr. J. Ecol. 35:146 – 155.

176

Haschick, S.L. and Kerley, G.I.H. 1997b. Factors influencing forage preference of bushbuck and Boer goats for subtropical thicket plants. Afr. J. Range For. Sci. 14(2):49 – 55. Hayward, M.W., O’Brien, J., and Kerley, G.I.H. 2007. Carrying capacity of large African predators: Predictions and tests. Biological Conservation. 139:219 – 229. Henley, S.R. 2001. Habitat suitability modeling for ungulates in the thicket biome, Eastern Cape. Unpublished doctoral dissertation, University of Port Elizabeth, Port Elizabeth, South Africa. Hillman, J.C. 1985. The Mountain Nyala. Acacia. 6: 39 – 46. Hillman, J.C. 1986. Bale Mountains National Park management plan. Ethiopian Wildlife Conservation Organization (EWCO), Addis Ababa, Ethiopia. Pp 250. Hillman, J.C. 1988. The Bale Mountains National Park area, southwest Ethiopia, and its management. Mountain Research and Development. 8:253 – 258. Hillman, J.C. 1993. Compendium of wildlife conservation information. Vol. 1 NYZS and EWCO, Addis Ababa, Ethiopia. Pp 580. Hillman, J.C. and Hillman, S.M. 1987. The Mountain Nyala (Tragelaphus buxtoni) and Simien fox (Canis simensis) in the Bale Mountains National Park. Walia. 10:3 – 6. Hochman, V. and Kotler, B.P. 2006a. Effects of food quality, diet preference, and water on patch use by Nubian Ibex. Oikos. 112:547 – 554. Hochman, V. and Kotler, B.P. 2006b. Patch use, apprehension, and vigilance behavior of Nubian Ibex under perceived risk of predation. Behavioral Ecology. Pp 368 – 374. doi:10.1093/beheco/arl087. Houle, A., Chapman, C.A., and Vickery, W.L. 2007. Intra-tree variation in fruit production and implications for primate foraging. Int. J. Primatol. 28: 1197 – 1217. Houle, A., Vickery, W.L., and Chapman, C.A. 2006. Testing mechanisms of coexistence among two species of frugivorous primates. J. of Animal Ecology. 75: 1034 – 1044. Hudson, P.J., Rizzoli, A., Grenfell, B.T., Heesterbeek, H., and Dobson, A.P. 2002. The ecology of wildlife diseases. Oxford University Press. Pp 197. Hulme, P.E and Borelli, T. 1999. Variability in post-dispersal seed predation in deciduous woodland: relative importance of location, seed species, burial, and density. Plant Ecology. 145: 149 – 156.

177

Hundessa, T. 1992. Management problems of protected areas in Ethiopia. UNESCO: World Heritage Fund, Paris. Hundessa, T. 1997. Major causes for the loss of wildlife resource in Ethiopia. Addis Ababa, Ethiopia. Walia, 18:3 – 6. Hunter, T.B. and Skinner, J.D. 1998. Vigilance behavior in African ungulates: the role of predation pressure. Behavior. 135: 195 – 211. Inglis, I. R. and Lazarus, J. 1981. Vigilance and group size in Brent geese: the edge effect. Zietschrif van Tuerpsychologie. 52: 193 – 200. IUCN. 2008. Red list of threatened species in 2008. Available online at www.iucnredlist.org. Jacobson, S.K., McDuff, M.D., and Monroe, M.C. 2006. Conservation education and outreach techniques. Oxford University Press. New York. Jeffers, J.N.R. 1982. Modeling. Chapman and Hall, New York, U.S.A. Jessup, D.A., Boyce, W.M., and Torres, S.G. 1995. Bighorn sheep health management plan in California: a fifteen year retrospective proceedings of the joint conference of the American association of zoo veterinarians, the wildlife disease association, and the American association of wildlife veterinarians. Johnson, C.J., Boyce, M.S., Schwartz, C.C., and Haroldson, M.A. 2004. Modeling survival: application of the Anderson-Gill model to Yellowstone grizzly bears. Journal of Wildlife Management. 68: 966 – 978. Johnson, D.H. 1980. The comparison of usage and availability of measurements for evaluating resource preference. Ecology. 61:65 – 71. Jokimäki, J. and Huhta, E. 1996. Effects of landscape matrix and habitat structure on a bird community in northern Finland: a multi-scale approach. Ornis Fennica. 73:97 – 113. Kaiser, H.F. 1958. The varimax criterion for analytic rotation in factor analysis. Psychometrika 23 (3):87 – 200. Kalton, G. 1983. Introduction to survey sampling. Newbury Park, CA: Sage Publications, Inc. Kebede, A.T. 2009. Sustaining the Allideghi grassland of Ethiopia: influence of pastoralism and vegetation change. PhD dissertation. Pp 309. Utah State University, USA.

178

Kellert, S.R., Black, M., Rush, C.R., and Bath, A.J.1996. Human culture and large carnivore conservation in North America. Conser. Biology. 10:977 – 990. Kingdon, J. 1997. The Kingdon field guide to African mammals. Princeton University Press, New Jersey. Kiss, A. 1990. Living with wildlife. Draft report of World Bank Environment Division, The World Bank. Washington, DC. Kitchen, A.M., Gese, E.M., and Schauster, E.R. 1999. Changes in coyote activity patterns due to reduced exposure to human persecution. Can. J. Zool. 78:853 – 857. Kleiman, D.G. 1989. Reintroduction of captive mammals for conservation. Bioscience. 39:152 –161. Klein, M.L., Humphrey, S.R., and Percival, H.F. 1995. Effects of ecotourism on distribution of water birds in a wildlife refuge. Conservation Biology. 9:1454 – 1465. Kohlmann, S.G., Muller, D.M., and Alkon, P.U. 1996. Anti-predator constraints on lactating Nubian Ibex. Journal of Mammal.774: 1122 – 1131. Kollmuss, A. and Agyeman, J. 2002. Mind the gap: why do people act environmentally and what are the barriers to pro-environmental behavior? Environmental Education Research. 8:239 – 260. Kotler, B.P., Brown, J.S., Dall, S.R.X., Gresser, S., Ganey, D., and Bouskila, A. 2002. Foraging games between gerbils and their predators: temporal dynamics of resource depletion and apprehension in gerbils. Ev. Eco. Research. 4: 495 – 518. Kotler, B.P., Brown, J.S., Oldfield, A., Thorson, J., and Cohen, D. 2001. Foraging substrate and escape substrate: patch use by three species of gerbils. Ecology. 82: 1781 – 1790. Kotler, B.P., Gross, J.E., and Mitchell, W.A. 1994. Applying patch use to assess aspects of foraging behavior in Nubian Ibex. J. of Wildlife Manage. 58(2):299 – 307. Kubsa, A. 1999. Need assessment and recommendations for active community involvement in the conservation of Galama Controlled Hunting Area. Report on behalf of advisory assistance to the Forest Administration Project (GTZ). Lagory, K.E. 1986. Habitat, group size, and the behavior of white-tailed deer. Behavior. 98: 168 – 179.

179

Laundre´, W., Hernandez, L., and Altendorf, K.B. 2001. Wolves, Elk, and Bison: reestablishing the ‘‘landscape of fear’’ in Yellowstone National Park, USA. Can. J Zool. 79(8):1401 – 1409. Lima, S.L. 1995. Back to the basics of anti-predatory vigilance: the group size effect. Animal Behavior. 49: 11 – 20. Lima, S.L. and Dill, L.M. 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Can. J. Zool. 68: 619 – 640. Lima, S.L., Valone, T.J., and Caraco, T. 1985. Foraging efficiency predation-risk trade-off in the grey squirrel. Anim. Behav.33:155 – 165. Lindsay, K., Craig, J., and Low, M. 2007. Tourism and conservation: the effects of track proximity on avian reproductive success and nest selection in an open sanctuary. Tourism Management. 29:730 – 739. Lindsey, P.A. 2008. Trophy hunting in Sub-Saharan Africa: economic scale and conservation significance. Best Practices in Sustainable Hunting. Pp, 41 – 47. Lindsey, P.A., Roulet, P.A., and Romanach, S.S. 2007. Economic and conservation significance of the trophy hunting industry in Sub-Saharan Africa. Biological Conservation. 134:455 – 469. LWAG. 2002. Wildlife and poverty study. Livestock and Wildlife Advisory Group (LWAG). Department for International Development's (DFID), Rural Livelihoods Department. Lydekker, R. 1912. On the Mountain Nyala, Tragelaphus buxtoni. Proceedings from the Zoological Society of London. 16:348 – 353. Mace, G.M. and Balmford, A. 2000. Patterns and processes in contemporary mammalian extinction. pp. 27 – 52. In: A. Entwistle and N. Dunstone (eds.) Priorities for the conservation of mammalian diversity: Has the Panda dad its day? Cambridge University Press, Cambridge. Main, M.B., Weckerly, F.W., and Bleich, V.C. 1996. Sexual segregation in ungulates: new directions for research. J. Mammal. 77:449 – 461. Malcolm, J. and Evangelista, P. 2005. The range and status of the Mountain Nyala. A report to the Ethiopian Wildlife Conservation Department, Addis Ababa, Ethiopia. Pp. 43.

180

Mamo, Y. 2007. Ecology and conservation of Mountain Nyala (Tragelaphus buxtoni) in Bale Mountains National Park, Ethiopia. PhD Dissertation. Univ. of Aberdeen, UK. Manor, R. and Saltz, D. 2003. Impact of human nuisance disturbance on vigilance and group size of a social ungulate. Ecological Applications. 13:1830 – 1834. Martin, P. and Bateson, P. 1993. Measuring behavior: An introductory guide. 2nd edition. Cambridge University Press, Cambridge. Martin, R.B. 1991. Integrating the needs of local people into the design of conservation and wildlife management: the CAMPFIRE program in Zimbabwe. Yale School of Forestry and Environmental Studies. pp. 11. Martinka, C.J. 1968. Habitat relationships of White-tailed and Mule deer in northern Montana. Journal of Wildlife Management.32 (3): 558 – 656. Mascial, M.A., Brosius, J.P., Dobson, T.A., Forbes, B.C., Horowitz, L., McKean, M.A., and Turner, N.J. 2003. Conservation and the social science. Cons. Biol. 17(3): 649 - 650. McLoughlin, P.D., Dunford, J. S., and Boutin, S. 2005. Relating predation mortality to broad-scale habitat selection. J. Animal Ecology. 74:701 – 707. Mech, L.D. 1977. Wolf-pack buffer zones as prey reservoirs. Scie. 198:320 – 321. Mellon, J. 1975. African Hunter. Safari Press, Long Beach, USA. Melton, D.A. 1987. Habitat selection and resource scarcity. South African Journal of Science.83:269 – 288. Mengesha, G. and Bekele, A. 2008. Diversity, distribution, and habitat association of large mammals of Alatish, North Gonder, Ethiopia. Acta Zoologica Sinica. 54(1): 20 – 29.

Miehe, S. and Miehe, G. 1994. Ericaceous forests and heath lands in the Bale Mountains of South Ethiopia. Siftung Walderhaltung in Afika. Hamburg. Mishra, C. and Rawat, G.S. 1998. Livestock grazing and biodiversity conservation: comments on Saberwal. Conservation Biology. 12: 712 – 717. Moenting, A.E. and Morris, D.W. 2006. Disturbance and habitat use: is edge more

important than area? Oikos. 115: 23 – 32.

181

Mori, Y., Sodhi, N.S., Kawanishi, S., and Yamagishi, S. 2001. The effect of human disturbance and flock composition on the flight distances of waterfowl species. Journal of Ethology. 19:115 – 119. Morris, D.W. 1998. State-dependent optimization of litter size. Oikos. 83. 518 – 528. Morris, D.W. 2003. How can we apply theories of habitat selection to wildlife conservation and management? Wildlife Research. 30:303 – 319. Morris, D.W. 1987. Ecological scale and habitat use. Ecology. 68: 362 – 369. Morris, D.W. 1988. Habitat dependent population regulation and community structure. Evolutionary Ecology. 2: 253 – 269. Morris, D.W. 1995. Habitat selection in mosaic landscapes, in Mosaic landscapes and ecological processes (eds L. Hansson, L. Fahrig and G. Marriam), Chapman and Hall, London. 5:110 – 135. Morris, D.W. 2005. Habitat-dependent foraging in a classic predator-/prey system: a fable from snowshoe hares. Oikos. 109: 239 -/254. Morris, D.W. and Kingston, S.R. 2002. Predicting future threats to biodiversity from habitat selection by humans. Evol. Ecol. Res. 4: 787 – 810. Morris, D.W. and Mukherjee, S. 2007. Is density-dependent resource harvest a reliable habitat indicator for conservation and management? Israel Journal of Ecology and Evolution. 53: 371 – 387. Morris, D.W., Kotler, B.P., Brown, J.S., Sundararaj, V., and Ale, S.B. 2009. Behavioral indicators for conserving mammal diversity. The Year in Ecology and Conservation Biology, 2009: Ann. N.Y. Acad. Sci. 1162:334 – 356. Morrison, M.L., Marcot, B.G., and Mannan, R.W. 1998. Wildlife habitat relationships: concepts and applications. Univ. Wisconsin Press, Madison, USA. Morton, S.R. 1990. The impact of European settlement on the vertebrate animals in arid Australia: a conceptual model. Proc. Ecol. Soc. Aust. 16: 201-213. Morzillo A.T., Mertig, A.G., Garner, N., and Liu, J. 2007. Resident attitudes toward Black Bears and population recovery in East Texas. Human Dimensions of Wildlife.12:417 – 428. Muller, D.M., Kohlmann, S.G., and Alkon, P.U. 1995. Nubian Ibex nursery: creche or natural trap. Israel Journal of Zoology. 41:163 – 174.

182

Negash, M. and Ermias, B. 1995. Towards a comprehensive agro-ecological zonation of Ethiopia. LUPRD, Agro-ecology Section. Ministry of Agriculture, Ethiopia. Pp1– 20. Neptali, R., Mario, G., and Guadalupe, W. 2001. Anthropogenic disturbance and tree diversity in montane rain forest in Chiapas, Mexico. Forest Ecology and Management. 154: 311 – 326. Neuhaus, P. and Ruckstuhl, K.E. 2002. The link between , activity budgets, and group cohesion: the case of the plains zebra (Equus burchelli).NRC Research Press. Pp 1437 – 1441. Available at http://cjz.nrc.ca. Neumann, W.L. 2003. Social research methods: qualitative and quantitative approaches (5th ed.). Pearson Education, Inc. Nigatu, L. and Tadesse, M. 1989. An ecological survey of the vegetation of the Harenna Forest, Bale, Ethiopia. SINET. 12(1):63 – 93. Nour, N., Matthysen, E., and Dohondt, A.A. 1997. Effects of habitat fragmentation on the foraging behaviors of tits and related species: does niche space vary in relation to the size and degree of isolation of forest fragments? Ecography. 20:263 – 278. Oates, J.F. 1999. Myth and reality in the rain forest: how conservation strategies are

failing in West Africa. Berkeley: University of California Press.

Olsson, O. and Molokwu, M.N. 2007. On the missed opportunity cost, GUD, and estimating environmental quality. Israel J. Ecology and Evolution. 53: 263 – 278. Owen-Smith, S. 1994. Foraging responses of kudus to seasonal changes in food resources: elasticity in constraints. Ecology. 75(4): 1050 – 1062. Padmalal, U.K.G.K. and Takatsuki, S. 1994. Age-sex differences in the diets of Sika deer on Kinkazan Island, northern Japan. Ecological Research. 9:251 – 256. Patton, D.R. 1992. Wildlife habitat relationships in forested ecosystems. Timber Press, Portland, Oregon, USA. Primack, R.B. 2002. Essentials of conservation biology. Inc. Publishers, Massachusetts, USA. 3rd edition. Pp 698. Prins, H.H.T. 1996. Ecology and behavior of the . Social inequality and decision making. 1st ed. Chapman and Hall, London.

183

Provenza, F.D. 1997. Feeding Behavior of Herbivores in Response to Plant Toxicants.

Handbook of Plant and Fungal Toxicants. 16: 231 – 242.

Pulliam, H.R. 1973. On the advantages of flocking. J. Theor. Biol. 38: 419 – 422. Quenette, P.Y. 1990. Functions of vigilance behavior in mammals: a review. Acta Oecol.

11:801 – 818.

Raik, D.B., Decker, D.J., and Siemer, W.F. 2003. Dimensions of capacity in community- based suburban deer management: the managers’ perspective. Wildlife Society Bulletin. 31: 854 – 864. Refera, B. and Bekele, A. 2004. Population status and structure of Mountain Nyala in Bale Mountains National Park, Ethiopia. African Journal of Ecology. 42:1 – 7. Reid, C. 2005. Habitat suitability and behavior of Springbok (Antidorcas marsupialis) at Augrabies National Park, South Africa. Masters Thesis. University of Port Elizabeth, Port Elizabeth, South Africa. Pp 106. Rieucau, G. and Martin, J.G.A. 2008. Many eyes or many ewes: vigilance tactics in female bighorn sheep (Ovis canadensis) vary according to reproductive status. Oikos. 117: 501 – 506. Riley, S.J., Decker, D.J., Carpenter, L.H., Organ, J.F., Siemer, W.F., Mattfeld, G.F., and Parsons, G. 2002. The essence of wildlife management. Wildlife Society Bulletin. 30: 585 – 593. Riley, S.J., Siemer, W.F., Decker, D.J., Carpenter, L.H., Organ, J.F., and Berchielli, L.T. 2003. Adaptive impact management: an integrative approach to wildlife management. Human Dimensions of Wildlife. 8: 81 – 95. Ripple, W.J. and Beschta, R.L. 2003. Wolf reintroduction, predation risk, and cottonwood recovery in Yellowstone National Park. For. Ecol. Mange. 184: 299 – 313. Ripple, W.J. and Beschta, R.L. 2004. Wolves and the ecology of fear: can predation risk structure ecosystems? Vol. 54 No. 8. Bioscience. 755. Risenhoover, K.L. and Bailey, J.A. 1985. Foraging ecology of mountain sheep: implications for habitat management. J. Wildlife Management. 49: 797 – 804. Roberts, G. 1996. Why individual vigilance declines as group size increases. Animal Behavior. 51: 1077 – 1086.

184

Rosenzweig, M.L. 1974. On the optimal aboveground activity of banner-tail kangaroo rats. Journal of Mammalogy. 55: 193 – 199. Rosenzweig, M.L. 1981. A theory of habitat selection. Ecology. 62: 327 – 335. Rosenzweig, M.L. 1987. Community organization from the point of view of habitat selectors. In 'Organization of Communities Past and Present'. (Eds J. H. R. Gee and P. S. Giller.) pp. 469-490. Blackwell Scientific Publications: Oxford. Ruckstuhl, K. E. 2007. Sexual segregation in vertebrates: proximate and ultimate causes. Integrative and Comparative Biology. 47(2): 245 – 257. Ruckstuhl, K.E. 1998. Foraging behavior and sexual segregation in bighorn sheep. Animal Behavior. 56: 99 – 106. Ruckstuhl, K.E. and Kokko, H. 2002. Modeling sexual segregation in ungulates: effects of group size, activity budgets, and synchrony. Anim. Beha. 64:909 – 914. Sanchez, F. 2006. Harvest rates and patch use strategy of Egyptian fruit bats in artificial food patches. Journal of Mammalogy. 87(6):1140 –1144. Sanchez, F., Kotler, B.P., Korine, C., and Pinshow, B. 2008. Sugars are complementary resources to ethanol in foods consumed by Egyptian fruit Bats. Journal of Experimental Biology. 211:1475 – 1481. Sanford, G. and Legendre, S. 1930. In quest of the Queen of Sheba’s antelope. Natural History. 30:17-32 and 161 – 176. Saunders, D.A. and Hobbs, R.J. 1991. Nature conservation 2: The role of corridors. Chipping Norton, Australia, Surrey Beatty and sons. Schamberger, M.L. and O’Neil, L.J. 1986. Concepts and constraints of habitat model testing. Pp 5-10 in Verner, J., Morrison, M.L. and Ralph, C.J. (Eds.) Wildlife 2000. Modeling habitat relationships of terrestrial vertebrates. University of Wisconsin Press, Madison. Senft, R.L., Rittenhouse, L.R., and Woodmansee, R.G. 1985. Factors influencing patterns of cattle grazing behavior on short grass steppe. J. Range Management. 38: 82– 87. Shettleworth, S.J., Hampton, R.R., and Westwood, R.P. 1995. Effects of season and

photoperiod on food storing by black-capped chickadees, Parus atricapillus. Anim.

Behav.49: 989 – 998.

185

Shipley, L.A., Illius, A.W., Danell, K., Hobbs, N.T., and Spalinger, D.E. 1999. Predicting bite size selection of mammalian herbivores: a test of a general model of diet optimization. Oikos. 84: 55 – 68. Shrader, A.M., Brown, J.S., Kerley, G.I.H., and Kotler, B.P. 2008a. Do free-ranging domestic goats show ‘landscapes of fear’? Patch use in response to habitat features and predator cues. Journal of Arid Environments. 72: 1811 – 1819. Shrader, A.M., Kotler, B.P., Brown, J.S., and Kerley, G.I.H. 2008b. Providing water for goats in arid landscapes: effects on feeding effort with regard to time period, herd size, and secondary compounds. Oikos. 117: 466 – 472. Shropshire, C.C. 1996. History, status, and habitat components of black bears in Mississippi. Doctoral dissertation, Mississippi State University, Mississippi State. Shuker, K.P.N. 1993. The lost ark: new and rediscovered animals of the twentieth century. London, Harper Collins Publishers. Pp 287. Siegfried, W.R. 1980. Vigilance and group size in Springbok. Madoqua.12:151 - 154. Sih, A. 1980. Optimal behavior: can foragers balance two conflicting demands? Science. 210: 1041 – 1043. Sillero-Zubiri, C. 2008. Tragelaphus buxtoni. In: IUCN 2011. IUCN red list of threatened species. Version 2011.1. Available at www.iucnredlist.org. Sillero-Zubiri, C., King A.A., and Macdonald, C.W. 1996. Rabies and mortality in Ethiopian wolves (Canis simensis). Journal of Wildlife Diseases.32:80 – 86. Silori, C.S. and Mishra, B.K. 2001. Assessment of livestock grazing pressures in and around the elephant corridors in Mudumalai Wildlife Sanctuary, South India. Biodiversity and Conservation. 10: 2181 – 2195. Smit, C., Ouden, J.D., and Mueller-Schaerer, H. 2006. Unpalatable plants facilitate tree sapling survival in wooded pastures. J. Applied Ecology. 43: 305 – 312. Smith, M. 1993. Patch use by the gerbil Gerbillurus paeba exilis in the Alexandria Dune field. MSc thesis, University of Port Elizabeth, Port Elizabeth. Snedecor, G.W. and Cochran, W.G. 1989. Statistical methods. Eighth edition, Iowa State University Press. Soobramoney, S. and Perrin, M.R. 2008. A comparison of giving-up densities of five species of granivorous birds. Journal of African Ornithology. 79(1): 101–104.

186

Sparrowe, R.D. 1990. Developing harvest regulations strategies for Wood Ducks. In Proceedings of the 1988 North American Wood Duck Symposium. Fredrickson, L.H., Burger, G.V., Havera, S.P., Graber, D.A., Kirby, R.E., and Taylor, T.S. (Eds), St. Louis: University of Missouri. Spellerberg, I.F. 1992. Evaluation and assessment for conservation: ecological guidelines for determining priorities for nature conservation. Chapman and Hall, London. Stephens, D.W. and Krebs, J.R. 1986. Foraging Theory. Princeton University Press. New Jersey. Stephens, P.A., Sillero-Zubiri, C., and Leader-Williams, N. 2001. Impact of livestock and settlement on the large mammalian wildlife of the Bale Mountains National Park, Southern Ethiopia. Biological Conservation. 100: 307 – 322. Stillman, R.A., West, A.D., Caldow, R.W.G. and Durell, S. 2007. Predicting the effect of disturbance on coastal birds. Ibis. 149:73 – 81. Stoms, D.M., Davis, F.W., and Cogan, C.B. 1992. Sensitivity of wildlife habitat models to uncertainties in GIS data. Photogrammetric Engineering and Remote Sensing, 58: 843-850. Store, R. and Jokimäki, J. 2003. A GIS-based multi-scale approach to habitat suitability modeling. Ecological Modeling. 169:1 – 15. Store, R. and Kangas, J. 2001. Integrating spatial multi-criteria evaluation and expert knowledge for GIS-based habitat suitability modelling. Landscape and Urban Planning. 55: 79 – 93. Stuart, C. and Stuart, T. 2000. Field guide to the larger mammals of Africa. 2nd Edition. Cape Town (South Africa): Struik Publishers. Sukumar, R. and Gadgil, M. 1988. Male-female differences in foraging on crops by Asian Elephants. Animal Behavior. 36:1233 – 1235. Sutherland, W.J. 1996. From individual behavior to population ecology. Oxford University Press, Oxford. Sutherland, W.J. and Crock-ford, N.J. 1993. Factors affecting the feeding distribution of red-breasted geese Branta ruficollis wintering in Romania. Biological Conservation. 63:61 – 65.

187

Tadesse, S.A. and Kotler, B.P. 2010. Habitat choices of Nubian Ibex (Capra nubiana) evaluated with a habitat suitability modeling and isodar analysis. Israel Journal of Ecology and Evolution. 56: 55 – 74. Tadesse, S.A. and Kotler, B.P. 2011. The seasonal responses of habitat use by Nubian Ibex (Capra nubiana) evaluated with behavioral indictors. Israel J. of Ecology and Evolution.57:223 – 246. Tedla, S. 1995. Protected areas management crises in Ethiopia. Addis Ababa, Ethiopia. Walia. 16:17 – 30. Teketay, D. 1992. Human impact on a natural montane forest in south-eastern Ethiopia. Mountain Resources Development. 12:393 – 400. Teketay, D. and Granström, A. 1995. Soil seed banks in dry afro-montane forests of Ethiopia. Journal of Vegetation Science. 6:777 – 786. Ter-Braak, C.J.F. and Looman, C.W.N. 1987. Regression. In: Jongman, R.H., Ter Braak, C. J. F. and Van Tongeren, O. F. R. (eds). Data analysis in community and landscape ecology. Pudoc, Wageningen. pp 29 – 77. Tessema, M.E., Ashenafi, Z.T, Lilieholm, R.J., and Williams, N.L. 2007. Community attitudes towards wildlife conservation in Ethiopia. Proceedings of the George Wright Society Conference. Durrell Institute of Conservation and Ecology, University of Kent, Canterbury, Kent, United Kingdom. Pp 287 – 292. Third World Conference on Mountain Ungulates. 2002. Nubian Ibex response to human presence in ‘Eretz Ha’Machteshim region in the Israeli Negev Desert. Saragossa: Aragon, Spain. Tilman, D. 1982. Resource competition and community structure. Princeton Univ. Press. Treves, A. 2000. Theory and method in studies of vigilance and aggregation. Animal Behavior. 60: 711 – 722. Uhmann, T.V., Kenkel, N.C., and Baydack, R.K. 2001. Development of a habitat suitability index model for burrowing owls in the eastern Canadian prairies. Journal of Raptor Research. 5(4):378 – 384. Ukizintambara, T. 2008. Edge effects on ranging and foraging behavior of L’hoest’s monkey (Cercopithecus lhoesti) in Bwindi Impenetrable National Park, Uganada. Final Report to the Rufford Small Grant Foundation. Pp 11.

188

Van Horne, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management. 47: 893 – 901. Van Horne, B. and Wiens, J.A. 1991. Forest bird habitat suitability index models and the development of general habitat models. Fish and Wildlife Research 8. U.S. Fish and Wildlife Service, Washington, D.C., USA. Verhulst, S., Oosterbeek, K., and Ens, B.J. 2001. Experimental evidence for effects of human disturbance on foraging and parental care in oystercatchers. Biological Conservation.101:375 – 380. Verner, J., Morrison, M.L. and Ralph, C.J. editors. 1986. Wildlife 2000: modeling habitat relationships of terrestrial vertebrates. University of Wisconsin Press. Madison, USA. Wallace, M.C. and Krausman, P.R. 1987. Elk, mule deer, and cattle habitats in central Arizona. J. Range Manage. 40: 80 – 83. Waltermire, R. 1975. A national park in the Bale Mountains. Walia, 6, 20 – 23. Wassie, A., Sterck, F.J., Teketay, D, and Bongers, F. 2009. Effects of livestock exclusion on tree regeneration in church forests of Ethiopia. Forest Ecology and Management. 257: 765 – 772. Wesche, K., Miehe, G., and Kaeppeli, M. 2000. The significance of fire for Afro-alpine ericaceous vegetation. Mountain Research and Development. 20:340 – 347. Whittaker, D. and Knight, R.L. 1998. Understanding wildlife responses to humans. Wildlife Society Bulletin. 26:312 – 317. Williams, S. 2002. Bale Mountains: a guidebook. Ethiopian Wolf Conservation Program. United printers, Addis Ababa, Ethiopia. pp 52. Wilson, S.L and Kerley, G.I.H. 2003. Bite diameter selection by thicket browsers: the effect of body size and plant morphology on forage intake and quality. Forest Ecology and Management. 181: 51 – 65. Woldegebriel, G.K. 1996. The status of the Mountain Nyala (Tragelaphus buxtoni) in Bale Mountains National Park 1986-1994. Walia. 17: 27 – 37. Woodroffe, R., Thirgood, S., and Rabinowitz, A. 2005. People and wildlife: conflict or co- existence? Cambridge University Press. Pp 497.

189

Yalden, D.W. and Largen, M.J. 1992. Endemic mammals of Ethiopia. Mam. Review. 22(3/4):115 – 150. Yamada, K., Elith, J., McCarthy, M., and Zerger, A. 2003. Eliciting and integrating expert knowledge for wildlife habitat modeling. Ecological Modelling. 165: 251–264. Young, T.P. and Isbell, L.A. 1991. Sex differences in giraffe feeding ecology: energetic and social constraints. Ethology. 87: 79 – 89. Young, T.P., Palmer, T.M., and Gadd, M.E. 2005. Competition and compensation among cattle, zebras, and elephants in a semi-arid savanna in Laikipia, Kenya. Biol. Conserv. 122: 351 – 359.

190

APPENDICES Appendix I Table 3.1. A summary of habitat variables measured in Munessa. Means and standard deviations are included. N = total number of plots sampled that habitat type.

Season Habitat type N Parameters Mean Stdev Elevation / altitude (m) 2204.68 90.85 Slope condition (%) 10.32 11.11 % cover of grass and herbs per plot 89.80 15.71 Clear cut 31 % cover of bare soils per plot 10.20 15.71 Number of trees per plot 0.00 0.00 Number of shrubs per plot 0.80 0.76 Sample crown diameter per plot 0.00 0.00 Elevation / altitude (m) 2240.93 56.13 Slope condition (%) 6.2 5.08 % cover of grass and herbs per plot 62.39 31.20 Wet Plantation 41 % cover of bare soils per plot 37.61 31.20 Number of trees per plot 3.98 2.63 Number of shrubs per plot 1.44 1.47 Sample crown diameter per plot 6.44 2.01 Elevation / altitude (m) 2179.69 50.99 Slope condition (%) 9.5 8.39 % cover of grass and herbs per plot 57.38 20.83 Natural forest 37 % cover of bare soils per plot 42.62 20.83 Number of trees per plot 1.97 1.23 Number of shrubs per plot 3.50 2.71 Sample crown diameter per plot 10.61 6.20 Elevation / altitude (m) 2204.68 90.85 Slope condition (%) 10.32 11.11 % cover of grass and herbs per plot 41.40 17.71 Clear cut 31 % cover of bare soils per plot 58.60 17.71 Number of trees per plot 0.00 0.00 Number of shrubs per plot 0.80 0.76 Sample crown diameter per plot 0.00 0.00 Elevation / altitude (m) 2240.93 56.13 Slope condition (%) 6.2 5.08 Dry % cover of grass and herbs per plot 17.24 14.24 Plantation 41 % cover of bare soils per plot 82.76 14.24 Number of trees per plot 3.98 2.63 Number of shrubs per plot 1.44 1.47 Sample crown diameter per plot 6.44 2.01 Elevation / altitude (m) 2179.69 50.99 Slope condition (%) 9.5 8.39 % cover of grass and herbs per plot 16.88 16.30 Natural forest 37 % cover of bare soils per plot 83.12 16.30 Number of trees per plot 1.97 1.23 Number of shrubs per plot 3.50 2.71 Sample crown diameter per plot 10.61 6.20

191

Table 3.2. List of tree and shrub species recorded in plots which were considered to collect habitat variables data for Mountain Nyala in Munessa Scientific names Growth form Habitat type of occurrence Acokanthera schimperi Shrub Natural forest Adhatoda schimperiana Shrub Natural forest Afrocarpus falcatus Tree; Shrub Natural forest; Plantation; Clear cut Albizia gummifera Tree; Shrub Natural forest; Plantation Aningeria adolfi-friederici Tree Natural forest Arundinaria alpina Shrub Natural forest Bersama abyssinica Tree; Shrub Natural forest; Plantation Buddleja polystachya Shrub Natural forest Calpurnia auria Shrub Natural forest; Plantation Carissa edulis Shrub Natural forest Celtis africana Tree; Shrub Natural forest Combretum spp. Tree Natural forest Croton macrostachyus Tree; Shrub Natural forest; Plantation; Clear cut Cupressus lusitanica Tree; Shrub Plantation; Clear cut Ekebergia capensis Tree; Shrub Natural forest Eucalyptus globulus Tree Plantation Eucalyptus grandis Tree Plantation Grevillea robusta Tree Plantation Hagenia abyssinica Tree Natural forest Millettia ferruginea Tree; Shrub Natural forest Myrtus communis Tree Natural forest Maytenus senegalensis Tree; Shrub Natural forest; Clear cut; Plantation Olea hochstetteri Tree; Shrub Natural forest Phytolacca dodecandra shrub Plantation Pinus patula Tree; Shrub Plantation; Clear cut Prunus africana Tree Natural forest Rhus glutinosa Shrub Natural forest Rosa abyssinica Climber Natural forest Rytigynia neglecta Shrub Plantation; Clear cut; Natural forest Syzygium guineense Tree Natural forest Teclea nobilis Shrub Natural forest

192

Table 5.1. List of plant species which were found to be browsed by free-ranging Mountain Nyala in the Bale Mountains National Park where bite diameters were measured and recorded

Scientific names Growth form Habitat type of Foraged parts by occurrence Mountain Nyala Achyranthes aspera Herb Woodland Twigs Artemisia afra Shrub Grassland and open Twigs woodland Asparagus abyssinicus Herb Open woodland Twigs and branches Carduus nyassanus Herb Woodland Flower heads Cineraria abyssinica Herb Open woodland Twigs and branches Twigs and needle leaves Cupressus lusitanica Tree and shrub Woodland of shrubs because fully grown tree branches are out of the attainable height of Mountain Nyala Helichrysum splendidum Herb Grassland and open Twigs, flower heads woodland Hypericum revolutum Tree and shrub Woodland and edge of Flowers, twigs, leaves, grassland and sometimes barks Junipers procera Tree Twigs and needle leaves Woodland from lodged branches at the attainable height of Mountain Nyala Maytenus spp. Spinescent scrambling Woodland Twigs and leaves Shrub Nepeta azurea Herb Grassland and edge of Twigs, branches, and open woodland flower heads Rapanea simensis Tree and shrub Woodland Twigs Rosa abyssinica Spinescent scrambling Both grassland and Twigs, leaves, and flower Shrub woodland heads Senecio ragazzi Herb Open woodland Twigs and branches Solanum marginatum Spinescent scrambling Both grassland and Commonly fruits and Shrub woodland twigs but sometimes young leaves

193

Table 6.1. A summary of the density of livestock type (i.e. cattle, horse, donkey, sheep, and goat) in different habitat types during the wet and the dry season field surveys: minimum, mean, and maximum densities were shown as the number of times each livestock type were seen in each habitat type and as a percent of total number of times that habitat was sampled. The standard deviation for each livestock type in all habitat types in the wet and the dry seasons were also included. N = total number of transects sampled that habitat type.

Season Habitat Livestock Minimum Mean Maximum Number of Sightings per type N type density/ha density/ha density/ha stdev livestock type transect walk sightings (%) Cattle 0.63 1.40 2.55 0.60 16 100.00 Horse 0.00 0.09 0.30 0.10 9 56.25 Clear cut 16 Donkey 0.00 0.53 1.30 0.38 15 93.75 Sheep 0.00 0.46 1.02 0.30 14 87.50 Goat 0.00 0.19 0.65 0.22 9 56.25 Cattle 0.56 1.39 4.43 0.96 16 100.00 Horse 0.00 0.05 0.32 0.08 7 43.75 Wet Plantation 16 Donkey 0.00 0.47 1.01 0.31 15 93.75 Sheep 0.34 0.71 1.43 0.28 16 100.00 Goat 0.00 0.23 0.69 0.18 14 87.50 Cattle 0.46 2.93 7.23 1.90 16 100.00 Horse 0.00 0.03 0.24 0.06 6 37.50 Natural 16 Donkey 0.00 0.56 1.64 0.48 13 81.25 forest Sheep 0.26 1.46 2.82 0.71 16 100.00 Goat 0.35 0.81 1.51 0.36 16 100.00 Cattle 0.00 0.86 3.52 0.77 22 91.67 Horse 0.00 0.17 0.79 0.21 15 62.50 Clear cut 24 Donkey 0.00 0.46 1.20 0.33 23 95.83 Sheep 0.00 0.38 1.58 0.45 15 62.50 Goat 0.00 0.16 0.93 0.24 13 54.17 Cattle 0.00 1.06 2.68 0.76 23 95.83 Horse 0.00 0.07 0.51 0.12 12 50.00 Dry Plantation 24 Donkey 0.00 0.21 1.01 0.20 22 91.67 Sheep 0.00 0.26 0.66 0.23 17 70.83 Goat 0.00 0.11 0.40 0.11 15 62.50 Cattle 0.31 1.51 5.25 1.01 24 100.00 Horse 0.00 0.02 0.39 0.08 2 8.33 Natural 24 Donkey 0.00 0.23 1.18 0.25 21 87.50 forest Sheep 0.00 0.37 1.33 0.46 12 50.00 Goat 0.00 0.24 0.87 0.30 12 50.00

194

Table 6.2. Seasonal humans and livestock uses and disturbances across habitat types in Munessa. Mean and standard deviation values are included. N = number of plots assessed to quantify that parameter.

Season Habitat type N Parameters Mean Stdev 31 Stem damage / plot 0.00 0.00 31 Crown damage / plot 0.00 0.00 31 Evidence of wood use / plot 0.28 0.46 Clear cut 31 Number of cut stumps / plot 5.48 1.91 31 Sign of habitat use by livestock / plot 0.92 0.28 31 Level of grazing / browsing by livestock / plot 2.08 1.00 41 Stem damage / plot 0.59 0.50 41 Crown damage / plot 0.73 0.45 41 Evidence of wood use / plot 0.78 0.42 Plantation 41 Number of cut stumps / plot 1.73 1.84 Wet 41 Sign of habitat use by livestock / plot 0.80 0.40 41 Level of grazing / browsing by livestock / plot 1.32 0.96 37 Stem damage / plot 0.28 0.46 37 Crown damage / plot 0.63 0.45 37 Evidence of wood use / plot 0.94 0.25 Natural forest 37 Number of cut stumps / plot 12.16 11.05 37 Sign of habitat use by livestock / plot 0.94 0.25 37 Level of grazing / browsing by livestock / plot 2.31 0.90 31 Stem damage / plot 0.00 0.00 31 Crown damage / plot 0.00 0.00 31 Evidence of wood use / plot 0.12 0.33 Clear cut 31 Number of cut stumps / plot 5.48 1.91 31 Sign of habitat use by livestock / plot 0.48 0.51 31 Level of grazing / browsing by livestock / plot 0.84 0.99 41 Stem damage / plot 0.49 0.51 41 Crown damage / plot 0.49 0.51 41 Evidence of wood use / plot 0.46 0.50 Dry Plantation 41 Number of cut stumps / plot 2.00 1.94 41 Sign of habitat use by livestock / plot 0.41 0.50 41 Level of grazing / browsing by livestock / plot 0.68 0.96 37 Stem damage / plot 0.41 0.50 37 Crown damage / plot 0.38 0.49 Natural forest 37 Evidence of wood use / plot 0.50 0.51 37 Number of cut stumps / plot 9.25 7.22 37 Sign of habitat use by livestock / plot 0.47 0.51 37 Level of grazing / browsing by livestock / plot 0.84 0.95 Notability:  Stem damage (0 = absent; 1= present).  Crown damage (0 = absent; 1= present).  Evidence of wood use (0 = absent; 1= present).  Sign of habitat use by livestock (0 = absent; 1= present).  Level of browsing / grazing by livestock: 0 = No evidence (0% browsed/grazed); 1 = lightly browsed/grazed (1 - 25% browsed/grazed); 2 = moderately browsed/grazed (26 - 50% browsed/grazed); and 3 = heavily browsed/grazed (>50% browsed/grazed).

195

Appendix II Guided questions used for focal group discussions to assess the attitudes and indigenous knowledge of local people about Mountain Nyala and their habitats in Munessa

Introductory remarks about:  personal introduction to the local participants  the objective of the focal group discussions  the composition of the group and why they were selected for the focal group discussion

Check list of questions 1. What do you know about Mountain Nyala and their habitats in Munessa? 2. Where is the former distribution of Mountain Nyala in the Arsi Mountains? 3. How do you evaluate the trend of Mountain Nyala population over the last decade in Munessa? If you say that the population size of Mountain Nyala is decreasing, what do you think the major reasons behind? 4. Do the Mountain Nyala move out of the Munessa Forest? If so, where do they move out of the Munessa Forest? When and how often do they move out of Munessa? Do they move daily or seasonally? 5. What are the common predators preying on Mountain Nyala and your livestock in Munessa? 6. Do you know any common diseases and parasites affecting both your livestock and Mountain Nyala in Munessa? If so, please mention the name of the diseases and parasites. 7. Do you get any benefits from Mountain Nyala in Munessa? 8. Do you hunt Mountain Nyala in Munessa? If so, why, when and how often do you hunt the Mountain Nyala? Which sex and age class of Mountain Nyala do you hunt? 9. Are there any problems faced with the Mountain Nyala in Munessa? If so, what are those problems? How do you rate them based on severity?

196

10. Do the Mountain Nyala need salt in their diet? If yes, where do they get? Is the availability of salt lick similar in the wet and the dry seasons? Do Mountain Nyala feed on salty plants to satisfy their salt requirements? If so, mention the name of plant species that the Mountain Nyala forage to satisfy their salt requirements. 11. When do your livestock leave their foraging patches with lower giving-up densities during free-range grazing i.e. when they feed alone or followed by their sheepherders? What are the reasons for such outcomes? 12. Do you think that the Mountain Nyala in Munessa create any problems on your livestock or food crops? If yes, what are those problems and how do you control them? 13. What do you suggest to be done to conserve and manage the Mountain Nyala and their habitats in Munessa?

197

תקציר ההבנה כיצד האיכות והמאפיינים של בית הגידול משפיעים על השימוש בבית הגידול ועל התנהגויות השיחור של בעל חיים היא בעלת חשיבות עליונה לאקולוגיה וממשק של מינים. ניתן ליישם תיאוריות של בחירת בית הגידול וניצול הכתם בדרכים יצירתיות רבות כדי לבנות סדרה של מבחני התנהגות פשוטים יחסית שמספקים אינדיקטורים המצביעים על איכות וניצול בתי גידול. אני עבדתי באתיופיה, בפארק הלאומי של יער מונסה )Munessa( והרי באלה )Bale(, העשירים במגוון הביולוגי, על אוכלוסיות ניאלת הרים (Tragelaphus buxtoni), מין בסכנת הכחדה, החיות תחת לחץ רב מבני האדם והמקנה שלהם. המטרה הכוללת של מחקר זה הייתה לבחון את הגורמים הסביבתיים והאנתרופוגניים העיקריים המשפיעים על איכות בית הגידול ,שימוש בבית הגידול ,ואקולוגית השיחור של הניאלה הכריזמטית. ערכתי עבודת שדה בעונות היבשות ועונות הגשמים במונסה. השתמשתי בשיטות שונות כדי להשיג את הנתונים בשטח. באופן קבוע ערכתי ספירת מלאי וצנזוס )מפקד אוכלוסין( של אוכלוסיות לאורך קווי חיתוך )transects( קבועים שנחתכו על פני סוגי בתי הגידול העיקריים על הנוף .אלה הכילו צנזוס אוכלוסיות ביום ובלילה, כאשר הצנזוסים הליליים התבצעו רק במהלך העונה היבשה בעזרת זרקור. הצנזוסים הניבו הערכות של צפיפות פעילות על פני בתי הגידול וקטגוריות גיל ומין. מדדתי משתני מיקרו-בית-גידול חשובים בחלקות עגולות שהונחו לאורך כל קו חיתוך קבוע וערכתי קורלציה בין המשתנים הללו לצפיפות הפעילות המקומית של ניאלת ההרים כדי לייצר מודלים של התאמת בית גידול. ביצעתי גם תצפיות ממוקדות בבעלי חיים על מנת להעריך כיצד תגובות התנהגותיות של ניאלת ההרים משתנות בהתאם לגודל הקבוצה, קטגוריות מין וגיל, וסוגי הגידול בכל העונות במונסה. הערכתי וכימתתי את ההשפעות של פגעי האדם והמקנה שלהם על בתי גידול של ניאלת ההרים במונסה. בהתאם לכך, לאורך קווי החיתוך הקבועים, הערכתי את צפיפות פעילות של המקנה. בדקתי וכימתתי גם את ההיקף של פגעי האדם והמקנה על בתי גידול של ניאלת ההרים בחלקות העגולות שהונחו לאורך כל קו חיתוך קבוע. בנוסף ,פיתחתי שאלונים פתוחים וסגורים וחילקתי אותם לאנשים המקומיים המתגוררים בשלושה ארגוני איכרים סמוכים לשטח עבודת השדה וכפר אחד במונסה. ערכתי גם דיונים קבוצתיים עם דמויות מפתח מבין חברי הקהילה המקומית .לבסוף, עקבתי וצפיתי בניאלת ההרים באופן קבוע כדי לכמת את מידת הערנות שלהם ,קצב הנגיסה, וקוטר נגיסה בזמן שיחור על מיני צמחים טבעיים נפוצים בפארק הלאומי הרי באלה (BMNP). אלה הניבו מדדים של מאמץ שיחור וצפיפות וויתור טיבעית ) :GUDs giving-up-densities- כמות או צפיפות של מקורות מזון שנותרו בכתם המזון כאשר המשחר היעיל ביותר עוזב את כתם המשאב(. מדידות של צפיפות פעילות ומשתנים סביבתיים אפשרו לי לבנות מודלים של התאמת בית גידול ולהעריך איזודארים ) isodars ( המתארים בחירה תלוית צפיפות בבית הגידול. מודל התאמת בית הגידול הראה כי ניאלת ההרים לא הראתה העדפה משמעותית לבית גידול במהלך עונת הגשמים במונסה. עם זאת , בעונה היבשה ,יער טבעי היה בית הגידול המועדף ביותר. לגבי משתני בית גידול ,רק הקוטר של צמרות

198

העצים השפיע באופן משמעותי על התאמת בית גידול עבור ניאלת ההרים בעונה היבשה .השיפועים של האיזודארים הראו כי בית הגידול יער טבעי הוא בצורה איכותית ,אך לא באופן כמותי ,טוב יותר מאשר המטעים או בתי הגידול המבוראים, במהלך העונה היבשה במונסה. עם זאת ,תוצאות הצנזוסים הליליים עם הזרקור הראו כי ניאלת ההרים בברור העדיפה בתי גידול מבוראים בזמן לילה כאשר אנשים ובעלי חיים נעדרים ממונסה. מחקר ההתנהגות עולה כי ניאלת ההרים הקדישה חלק ניכר מזמנה להתנהגות הערנות במהלך העונת הגשמים במונסה, אך סוג הגידול לא להשפיע באופן משמעותי על ערנות. בנוסף, בעונת הגשמים, לא היה הבדל משמעותי במשך הזמן המוקדש לערנות בקרב קטגוריות המין-גיל השונות. בעונה היבשה, ניאלת ההרים הייתה באופן משמעותי ערנית ביותר בבית הגידול המבורא. אמנם חיות פגיעות יותר, במיוחד נקבות עם צאצאים, צפויות להיות ערניות יותר, זכרים בוגרים היו הערניים ביותר בהשוואה עם בני המין השני בקטגורית הגיל בעונה זו. תוצאה זו עשויה לשקף את השפעת לחץ הציד על ידי בני אדם, שמכוון בלעדית כלפי זכרים בוגרים. גודל הקבוצה הממוצע של ניאלת ההרים היה משמעותית הגבוה ביותר בבית הגידול יער טבעי במהלך העונה היבשה. ללא קשר לעונות, גודל הקבוצה לא השפיעה באופן משמעותי על רמת ערנות של ניאלת ההרים במונסה. מידת פלישת המקנה והאדם אל בתי גידול של ניאלת ההרים השתנתה עונתית. צפיפות הפעילות של המקנה הייתה הגבוהה ביותר באופן משמעותי בבית הגידול יער טבעי במהלך עונת הגשמים במונסה. לעומת זאת, במהלך העונה היבשה, בעלי חיים המבויתים לא הראו הבדל משמעותי בהעדפת בית הגידול שלהם. באופן כללי, צפיפות הפעילות של המקנה בכל סוגי בתי הגידול הייתה גבוהה יותר בעונה הגשומה מאשר בעונה היבשה. בשתי העונות, גשומות ויבשות, היקף הנזק לגבעולים ולצמרות הצמחים על ידי אדם היה באופן משמעותי גבוה ביותר בבית הגידול מטע. בשתי העונות, עדות לשימוש בעץ ומספר הגדמים שנחתכו על ידי בני אדם היו הגבוהים ביותר באופן משמעותי ביער טבעי. בשתי העונות, סימנים של ניצול המקנה את בית הגידול לא היו שונים בין סוגי בתי הגידול; אלא הם היו מפוזרים ברחבי כל בתי הגידול. עם זאת, בעונת הגשמים, עוצמת הרעייה על ידי מקנה הייתה הכבדה ביותר באופן משמעותי בבית הגידול יער טבעי. באופן כללי, התוצאות הראו כי ההשפעות של פגעי האדם והמקנה היו גבוהים ומתמשכים לאורך זמן רב במונסה. כתוצאה מכך, נמנעת ניאלת ההרים מהשפעות האדם והמקנה על ידי מעבר לפעילות בתקופות )למשל במשך הלילה(, כאשר בני אדם והמקנה שלהם נעדרים. המחקר החברתי הראה כי יחסם של המקומיים כלפי ניאלת ההרים וגידול אוכלוסייתה הושפעו באופן משמעותי על ידי מספר משתנים סוציו-אקונומיים כגון אסטרטגית פרנסה, הבעלות על הקרקע, בעלות על מקנה, וידע. בנוסף, באמצעות דיונים קבוצתיים מרכזיים, חברי מפתח בקהילה שיתפו בידע העממי השופע שלהם על התנהגויות שונות של ניאלת ההרים ובתי הגידול שלהם במונסה.

199

מתוצאות מדידת קטרי נגיסה עולה כי השיחור של ניאלות ההרים בפארק הלאומי של הרי באלה הושפע באופן משמעותי מסוג בית הגידול, מיני הצמחים הנאכלים, ויחסי הגומלין בין סוג בית הגידול ומיני הצמחים הנאכלים. בהשוואות בין סוגי בתי גידול, קוטר נגיסה גדול יותר )התואם לעלויות שיחור נמוכות יותר ויעילות שיחור גבוהות יותר( נמצא בבית הגידול מרעה. בהשוואות בין מיני הצמחים הנאכלים, קוטר הנגיסה הגדול ביותר הממוצע נמצא עבור Solanum marginatum. קצב הנגיסות של ניאלת ההרים הושפע באופן משמעותי מסוג בית הגידול, קטגורית מין וגיל, ויחסי הגומלין בין סוג בית הגידול וקטגורית המין והגיל. בהשוואות בן סוגי בתי הגידול, קצב הנגיסות הממוצע הגבוה ביותר נצפה כאשר ניאלות ההרים שיחרו בבית הגידול עשבוני. בהשוואות בין קטגוריות מין וגיל, קצב הנגיסות הממוצע הגבוה ביותר נצפו בנקבות בוגרות. רמת הערנות של ניאלת ההרים הייתה גבוהה יותר בבית הגידול יער. לעומת זאת, פרופורציית הזמן המוקדש לשיחור הייתה גבוהה יותר כאשר ניאלת ההרים שיחרה בבית גידול עשבוני. אינדיקאטורים התנהגותיים מראים כי ניאלת ההרים מעדיפה בית גידול מרעה פתוח שבו היא משקיעה יותר זמן באכילה ופחות זמן בערנות, ומנצלות הזדמנויות שיחור באופן יסודי יותר. המחקרים הנוכחיים הראו כי ניאלת ההרים נתקלות בכמה אתגרים מכיוון בני האדם והמקנה שלהם אשר עשויים לאיים על הכושר )fitness( שלהם במונסה. מערכות יחסים תומכות והדדיות בין התושבים המקומיים ומערכת מכסות הציד החוקי במונסה הן קריטיות להצלחה ארוכת טווח של המאמצים לשימור אוכלוסיית ניאלת ההרים. לכן ,מחקר זה מציע מספר המלצות ניהול שיש ליישם כדי לשמר ולנצל באורך בר קיימא את האוכלוסייה הנמצאת בסכנת הכחדה של ניאלת ההרים במונסה. והכי חשוב ,המחקר מציג ומקדם מאמצי שימור מבוססי קהילה, שמאפשרים לקהילות להפיק תועלת כלכלית מאקו-תיירות, ועשויים לקדם את שימור ובאותו זמן לתת פתרון לקונפליקטים של ניצול המשאבים בין התושבים המקומיים ושימור אוכלוסיית ניאלת ההרים במונסה. הצגת וקידום מערכת כלכלית מבוססת התחלקות ברווחים בהשתתפות מלאה של הקהילה המקומית בתהליכי השימור והניהול היא חשובה באותה המידה כמו תכנון ויישום בר קיימא של ציד ספורטיבי של ניאלת ההרים במונסה. לסיכום, ממכלול התוצאות שהתקבלו מהגישות השונות במחקרים הנוכחיים יכולות לסייע לקובעי מדיניות מקומיים ומנהלי שמורות טבע להבין אקולוגיה של בתי גידול של ניאלת ההרים. התוצאות צריכות גם לעודד את מקבלי ההחלטות המקומיים ומנהלי שמורות הטבע לתת דגש נאות על הצרכים והדרישות של האנשים המקומיים בתהליכי ניהול. בסופו של דבר גישה זו תעזור להקים באתיופיה מערכת ניהול, עבור שימור ניאלת ההרים הנמצאת בסכנת הכחדה, שתהיה ברת קיימא מבחינה אקולוגית, מבחינה כלכלית, ומקובלת חברתית.

200

העבודה נעשתה בהדרכת

פרופסור ברט קוטלר

במחלקה לאקולוגיה המדברית

במכונים ללימודי מדבר על שם יעקב בלאושטיין

201

איכות בית גידול ואקולוגית שיחור מזון בקרב ניאלת הרים )Tragelaphus buxtoni) ביער מונסה ובפארק הלאומי הרי באלה, דרום מזרח אתיופיה

מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה"

מאת

סולומון איילה טדסה

הוגש לסינאט אוניברסיטת בן גוריון בנגב

אישור המנחה ______

אישור דיקן בית הספר ללימודי מחקר מתקדמים ע"ש קרייטמן ______

אייר התשע"ב מאי 2102

באר שבע

202

איכות בית גידול ואקולוגית שיחור מזון בקרב ניאלת הרים )Tragelaphus buxtoni) ביער מונסה ובפארק הלאומי הרי באלה, דרום מזרח אתיופיה

מחקר לשם מילוי חלקי של הדרישות לקבלת תואר "דוקטור לפילוסופיה"

מאת

סולומון איילה טדסה

הוגש לסינאט אוניברסיטת בן גוריון בנגב

אייר התשע"ב מאי 2102

באר שבע