Thinking about predators: multi-scale factors influencing human attitudes towards snow leopards and the wolves

Transcription

1 Thinking about predators: multi-scale factors influencing human attitudes towards snow leopards and the wolves Kulbhushansingh R. Suryawanshi 1,2, Saloni Bhatia 1,2, Yash Veer Bhatnagar 1,2, Stephen Redpath 3, Charudutt Mishra 1,2 1 Nature Conservation Foundation, 3076/5, IV Cross Gokulam Park, Mysore , India 2 Snow Leopard Trust, 4649 Sunnyside Av. North, Suite 325, Seattle, WA 98103, USA 3 Aberdeen Centre for Environmental Sustainability, Macaulay Institute and University of Aberdeen, Craigiebuckler, Aberdeen AB15 8QH, UK

2 Thinking about predators: multi-scale factors influencing human attitudes towards snow leopards and the wolves Abstract 1. The threat posed by large carnivores to livestock and human life makes peaceful co-existence between them difficult. Effective implementation of conservation laws and policies is contingent upon the attitudes of local residents towards the species. While there are many known correlates of human attitudes towards carnivores in conflict, these have only been assessed at the scale of the individual. Because human societies are organized hierarchically, attitudes are presumably influenced by different factors at different scales of social organization. Yet, a multi-scale examination of factors influencing human attitudes towards large carnivores hasn t been attempted. 2. We used structured interview surveys to quantitatively assess the attitudes of a Buddhist pastoral people towards the snow leopard Panthera uncia and the wolf Canis lupus. We sampled 383 individuals from 24 villages in six study sites across the high altitude Spiti Valley of the Indian Trans-Himalaya. We gathered information on key explanatory variables that together captured variation in individual and village-level socio-economic factors. We examined how the importance of these factors in influencing human attitudes changes with the scale of analysis from the individual to the community. 3. Our results revealed notable differences in the individual and community-level correlates of human attitudes towards carnivores in conflict. Factors significant at the individual-level were gender, education, and the age of the respondent (for both wolves and snow leopards), number of income sources in the family (wolves), agricultural production and large-bodied livestock holdings (snow leopards). At the community-level, the significant factors included the number of smaller-bodied herded-stock killed by wolves and the mean agricultural 103

3 production (wolves), and village size and large livestock holdings (snow leopards). Presence of a conservation program was influential in improving attitudes at the scale of the study site. People had significantly more positive attitudes towards the snow leopard than the wolf. 4. Our work shows that scaling-up from the individual to higher levels of social organization can highlight important factors that influence attitudes of people towards wildlife in conflict, or formal conservation efforts in general. Such scalespecific information can help to efficiently target appropriate conservation measure at the appropriate scale. Our results reiterate the need for conflict management programs to be multi-pronged. Introduction As the human population continues to grow rapidly, finding effective ways to conserve large carnivores has become a challenge (Woodroffe et al. 2005). Large home ranges, and the threat they pose to livestock and human life make co-existence between large carnivores and humans difficult (Linnell et al. 2001). Despite relevant laws and conservation policies being in place, their implementation is fraught with challenges and is substantially influenced by the attitudes of local residents towards the species (Decker & Purdy 1988; West et al. 2006). Large carnivore conservation in a modern, democratic world depends on how readily these animals are accepted by local residents. Studies examining people s attitudes towards carnivores saw a surge in Europe and North America over the last two decades. This followed the recolonization by wolves Canis lupus and brown bears Ursus arctos of areas where they had been extirpated many decades ago (Kleiven et al. 2004). However, very few studies have assessed factors influencing attitudes of local people towards carnivores in Asia (Bagchi & Mishra 2006; Allendorf 2007; Badola 1998). Almost no systematic studies have been conducted in Central Asia, which has a unique faunal assemblage that inhabits land used traditionally by local communities for grazing livestock. The few existing studies have stopped at characterizing attitudes of local people towards carnivores or protected areas, and have not attempted to identify the drivers of those attitudes using robust empirical frameworks 104

4 (eg. Oli et al. 1994; Ikeda 2004; Bagchi and Mishra. 2006; Wang et al. 2006; Liu et al. 2011). Surveys in the developed world generally found that young, urban, educated people tended to have relatively positive attitudes towards carnivores as compared to older, rural and less educated people (Kellert 1985, 1991). Klieven et al. (2004) reported that people were more accepting of carnivores when they lived far away from them, and when the human communities were larger. Bjerke et al. (2001) and Røskaft et al. (2003) reported that Norwegian women were less accepting of carnivores due to fear of the species. Zinn et al. (2000), on the other hand, reported that in the USA, men were more agreeable to destroying mountain lion (Puma concolor) than women under various scenarios (sighting, loss of pet animal, injury or loss of human life). The identity of the carnivore species is also reported to be important in influencing peoples attitudes. In Norway, people were less accepting of bears and wolves than of lynx Lynx lynx and wolverines Gulo gulo (Klieven et al. 2004). While these authors proposed that body size of the carnivore was an important factor in acceptability, Kellert et al. (1996), suggested that cultural history was important in influencing attitudes to bears, wolves and cougars in the USA. Economic loss due to large carnivores generally contributes to the negative attitudes towards them (Williams et al. 2002; Bagchi & Mishra 2006). Clearly, attitudes to carnivores are influenced by diverse and complex factors from individual human attributes and socio-economic indices, to appearance and behavior of the carnivore species making it difficult to understand their drivers. Because human societies are organized hierarchically an individual is a part of a family group that lives within a community and so forth the process of decision-making takes place at multiple hierarchical levels. While managing carnivore-caused livestock damage or retaliating against predators, people either act individually or in cooperation with other members of the community, and in many cases, with the support and approval of the State (Treves & Naughton-Treves 2005). We therefore propose that factors influencing peoples attitudes should vary as one scales up from the level of the individual to the community. While certain personal traits such as age and gender can only influence attitudes at the level of the individual and factors such as culture and traditions can influence attitudes at the level of the community, the influence of socio-economic factors such as income, job, 105

5 property, and risk could span the depth of human social hierarchy. However, human attitudes to species in conflict have only been assessed at the scale of the individual, and a multi-scale examination of the factors influencing human attitudes towards large carnivores hasn t been attempted. Using a multi-scale approach, we examine the attitudes of local people towards the snow leopard Panthera uncia and the sympatric wolf Canis lupus, and attempt to understand the importance of various factors in influencing attitudes at the scale of the individual and the village. These carnivores occur across the mountain ranges of Central Asia. Negative interactions with pastoralists over livestock depredation threaten their survival across their distribution range (McCarthy & Chapron 2003). Studies report that local pastoral communities may suffer average annual losses ranging from 3% to 18% of their livestock holding to snow leopards and wolves (Mishra 1997; Mishra et al. 2003; Hussain 2000; Namgail et al. 2007). High losses sometimes create such level of hatred towards these carnivores that local people view their complete extermination as the only solution (Oli et al. 1994). Both these carnivores have large home-ranges. While a snow leopard's home range is spread over several hundred to a few thousand square kilometers (McCarthy et al. 2005, 2010), the wolf can range over several thousand square kilometers (Kaczensky et al. 2008). Even though human density within their distributional ranges is relatively low, pastoral communities are spread across the mountains (Mishra et al. 2010). The Protected Areas are few and far apart and are often not free of human and livestock presence (Mishra 2000). Thus, effective conservation of the snow leopard and the wolf is contingent upon co-existence with humans in multiple use landscapes (Mishra et al. 2010). This will require enhancing peoples tolerance towards these species. Understanding the factors influencing the attitudes of local people is important for framing and implementing policies that can help improve the social carrying capacity for these carnivores. We quantified attitudes of the local pastoral community inhabiting the high altitude Spiti Valley (lat 32 00' ' N; Long 77 37' 'E) of the Indian Trans-Himalaya and identified the key variables influencing human attitudes towards carnivores at the scale of the individual and the village. Using structured interview surveys, we quantitatively 106

6 assessed the attitudes of local pastoral people towards the snow leopard and the wolf. We gathered information on key explanatory variables that together captured variation in individual and village-level socio-economic factors. Additionally we examined the influence of a fifteen year long community-based conservation program in influencing peoples attitudes towards the snow leopard and the wolf. Our work highlights the value of understanding the drivers of human attitudes towards wildlife at various relevant scales of social organization to best inform management decisions. Methods STUDY AREA The Spiti Valley (lat 32 00' ' N; Long 77 37' 'E) in Himachal Pradesh, India covers an area of about 7000km² in the Trans-Himalayan region. Annual temperatures range from 40 degree Celsius in peak winter to c. 30 degree Celsius in summer, with the minimum temperature remaining sub-zero from September to April. Vegetation is classified as 'Alpine scrub' or 'dry alpine steppe (Champion & Seth 1968). Large mammalian fauna of Spiti Valley includes bharal Pseudois nayaur, ibex Capra sibirica, and their predators, the snow leopard and the wolf. Other mammalian species include the red fox Vulpes vulpes, pale weasel, Mustela altaica, stone marten, Martes foina, and pika Ochotona spp. Agro-pastoralist communities, currently of Buddhist denomination, have inhabited this region for two to three millennia. Presently, the population of region is approximately people in 75 permanent villages and a township. The livestock assemblage includes sheep, goat, donkey, cow, cow-yak hybrid, horse and yak. Livestock graze in pastures except during extreme winter when they are stall-fed. Based on herding practices, livestock can be classified as 1) Large-bodied free-ranging, henceforth referred to as large stock (yaks and horses), and, 2) Medium and small-bodied herded, henceforth referred to as herded stock (cow, donkey, cow-yak hybrid, goat and sheep). Large stock is free ranging throughout the year, except for a few weeks during extreme winters. Herded stock is shepherded to the pastures every morning and brought back to the stocking pens inside the villages in the evening. Families take turns at shepherding the entire village's 107

7 herded stock along with a designated village-shepherd. Families mostly own small agricultural land holdings (1-2 ha). Primary crops are barley Hordeum vulgare, green pea Pisum sativum, and a local variety of green pea called black pea. Over the last three decades, the still ongoing adoption of green pea crop in many villages, and apple orchards in the relatively lower altitude ones, have marked a watershed in the economy of the region. Due to the high elevation, this region produces green peas during a time of high market prices. The crop is sown in late April or early May around the melting of the snow and harvested in the months of August-September. Figure 5.1: Nested sampling design used to conduct interview surveys. We sampled 383 individual participants in 24 villages across six study sites. 108

8 Table 5.1: Questions and scoring system used to quantitatively assess attitude of local pastoralists towards the snow leopard and the wolf. The attitude scores were summed across the seven questions. Attitude scores ranged from -8 being the most negative and 8 being the most positive. Columns 4 & 5 summarize the responses in percentage. % response to snow leopard % response to wolf Questions Replies Scores Would you like to see this animal in the rangelands of Spiti? Yes No Unsure Should it be protected inside or outside Spiti? Both Inside Don't know Outside Nowhere If SL/Wolf were to be conserved in Spiti, would you support it? Yes No Unsure Should kids be taught about SL/Wolf at school? Yes No Don't know Do you think the conservation of these animals is beneficial for the environment of Spiti? Yes Don't know Both No Where should SL/Wolf be protected? Rangelands Everywhere Only National Parks Don't know Zoo Nowhere What should be done when SL/Wolf kills your livestock? They also need food Nothing, I can bear it Can't do anything Chase it away Kill it

9 SAMPLING DESIGN We used a nested sampling design (figure 5.1). We selected six study sites in Spiti Valley, each comprising of a cluster of villages; namely, Kibber, Langza, Lingti, Pin, Tabo and Lossar. Study sites ranged from km 2 and were separated from each other by km along a straight line, except for Langza and Kibber which were adjacent to each other but separated by a deep gorge. The only road access between these two sites is 50 km long, which is operational only for about six months each year. The study site of Kibber represented a cluster of five villages that has a community-based conservation program run by the Nature Conservation Foundation and Snow Leopard Trust for over fifteen years (Mishra et al. 2003). The conservation program largely involves conflict management through a multi-pronged approach. It aims at (i) facilitating wild preyrecovery by creating livestock-free areas, (ii) sharing and offsetting economic losses due to livestock depredation by large carnivores through a community-run livestock insurance program and (iii) improving the social carrying capacity for carnivores through sustained conservation education. One village in Lossar and two villages in Langza also had similar conservation programs by the same organizations, running for the previous two years. We sampled in all the villages (1, 5, 4, 5, 4, 5 in Lingti, Lossar, Pin, Tabo, Langza and Kibber, respectively) within each of the six sites except in Tabo, where we sampled in five of the six villages. One village could not be sampled due to time limitation. We randomly sampled a third of the households in each village except for Lingti where we interviewed respondents from all the households. We interviewed one adult member in every household. Village sizes ranged from ATTITUDE SCORES We used structured interviews to quantify attitudes. The interview survey contained seven questions that were scored and added to quantify attitudes (table 5.1). We did not offer any options to choose from. The questions were designed such that there could only be a limited set of replies. All the questions were scored (details of the scoring system in table 5.1). The sum of scores could potentially range from -8 to 8 with -8 representing the most negative attitude on a relative scale, and 8 representing the most positive. We chose 110

10 not to follow the popular 5-point Likert scale of replies ranging from Strongly agree to Strongly disagree as we found that people experienced difficulties in deciding between Agree and Strongly agree or Disagree and Strongly disagree. The questionnaire was initially tested in five villages with thirty participants. The data were not included in the final analysis. No changes to the questionnaire were deemed necessary after the initial testing. The interviewers (KS and SB) were accompanied by a local translator. We asked the questions in Hindi, since most respondents spoke the language. The translator repeated the question in Spitian (a dialect of Tibetan) in the rare case when an interviewee did not know Hindi, or found a particular question difficult to understand. The formal interview began by testing if the interviewee could identify the predator species (snow leopard or wolf) from photographs. If the interviewee correctly identified the predator, the interviewer proceeded by informing the interviewee that the following questions pertained to that predator. Such a method had to be adopted as there are multiple names for snow leopards and wolves and some of these names are interchangeable. The interviewer then followed the same procedure for the second predator. We alternated the sequence of carnivore photographs between the interviewees. We interviewed 381 individuals across the six study sites. Of these, 164 (43%) were women. Age of the respondents ranged from 19 to 85 years with a median of 40 years. We interviewed 54, 65, 61, 66, 70 and 65 individuals across Kibber, Langza, Lingti, Lossar, Pin and Tabo study sites, respectively. Based on the photographs shown for first question, a total of 368 and 332 out of 381 people correctly identified the snow leopard and wolf, respectively. Of these, we could not gather a complete set of covariate information for 3 participants which had to be dropped from the analysis. The final analysis included 365 and 329 interviewees for the snow leopard and wolf, respectively. The responses of only these interviewees were used for further analysis. The replies to each individual question are summarized in table 5.1. We used students paired t-test to examine the difference between mean attitude score towards snow leopard and wolf. We used linear mixed effect models to identify the factors influencing the attitudes of people towards the snow leopard and the wolf (Gelman & Hill 2007). We used the lmer command from the package 111

11 lme4 in R (Bates et al. 2012; R Development core team 2011). Study site and village were used as fixed effects throughout the analysis. EXPLANATORY VARIABLES We quantified and analysed a total of 17 control variables that could potentially influence peoples attitudes at the individual and village levels. Of these, 11 were individual level and 7 were village level variables (Table 5.2). We used a combination of interviews, key informant surveys and censuses. After the interview surveys for attitude scores were completed, a local representative conducted a door-to-door survey for information on participant's education level, livestock holding and green pea production. Such a method was used as it was observed that participant were generally shy of talking about their land and livestock holding or education to outsiders (KS & SB). We requested the professional herders (key informants) of each village to maintain a record of the total livestock in the village and all livestock mortality on a monthly basis since the year 2009 (Suryawanshi et al. 2013). The herders maintained records of the species and age of livestock, month, cause of mortality and the total number of livestock in each village throughout January 2009 to December In cases of livestock depredation by large carnivores, the identity of the predator was confirmed through direct sighting or signs around the kill. By cross-checking their knowledge, we ensured that the herders keeping records were experienced at identifying signs of snow leopards and wolves. A local representative maintained contact with each herder every 4 6 months and recorded all the livestock mortalities. We expect the herder s records to be accurate, as they also had to report mortality events to the livestock owner. 112

12 Table 5.2: Explanatory variables considered as potential predictors of attitudes of local people towards the snow leopard and the wolf. Variable Label Level Source of data Data type Range Perception of snow leopard/wolf population size? Pop.perception Individual Interview Categorical Increasing, decreasing, stable Education in number of years of formal schooling Education Individual Census Continuous 0-18 Green pea production in number of sacks pea.harvest Individual Census Continuous Gender Gender Individual Interview Categorical Male, Female Age in years Age Individual Interview Continuous Large-bodied free-ranging stock holding (yak and horse) Largestock Individual Census Continuous 1 to 12 Proportion of large-bodied free-ranging stock (yak and horses) killed by the snow leopard and wolf from 2009 to 2011 recorded separately for both the species Largestock-killed Individual Key informants Proportion 0-1 Small-bodied herded-stock holding (sheep, goat, donkey, cattle) Herded-stock Individual Census Continuous 0 to 61 Proportion small-bodied herded-stock (sheep, goat, donkey, cattle) killed by snow leopard and wolf from 2009 to 2011 recorded separately for both the species Herded-stock killed Individual Key informants Proportion 0-1 Encounter with carnivore; separate for snow leopard and wolf Carnivore.enc Individual Interview Categorical Yes, no Number of sources of income for family Jobs Individual Census Continuous 1 to 4 Village size as number of families Village.size Village Census Continuous 6-76 Large-bodied free-ranging stock holding Largestock Village Key informants Continuous 4 to 167 Small-bodied herded-stock holding Herded-stock Village Key informants Continuous Proportion of large-bodied free-ranging stock (yak and horses) killed by the snow leopard and wolf from 2009 to 2011 recorded separately for both the species Prop.large. killed Village Key informants Proportion 0-1 Proportion small-bodied herded-stock (sheep, goat, donkey, cattle) killed by snow leopard and wolf from 2009 to 2011 recorded separately for both the species Prop.herded. killed Village Key informants Proportion 0-1 Mean green pea production in the village Mean.pea.production Village Census Continuous

13 We built two candidate sets of models for individual and village level analyses separately. Each candidate set contained 15 models (table 5.3 & 5.4). We used Akaike Information Criterion (AIC) for model selection. We chose best model for each level (viz. individual, village). We then integrated the variables from the best model at individual and village levels, respectively, into a single model. We estimated parameter for this model alone. We interpreted the effect of a control variable on the attitudes based on the parameter estimate from this model. Such a strategy was used as there were over 40 meaningful potential models using all the 17 control variables from across the two scales. Johnson and Omland (2004) recommend no more than 20 candidate models based on philosophical considerations. Using the above strategy we identified the control variables best explaining human attitudes towards carnivores at individual and village levels. We integrated these variables into a single hierarchical model. We interpreted the variables included in the final model using their 95% CI as calculated using their standard errors. The same candidate set of models was used to examine the determinants of attitudes towards snow leopards and wolves. All continuous variables were normalized to the mean for linear mixed-effect modeling. Results Overall, the mean attitude scores towards the snow leopard and the wolf were positive for four of the six study sites (figure 5.2). Attitude scores were negative and neutral for one site each for both the species. The mean attitude score towards snow leopard (1.62) was significantly higher than mean attitude score towards the wolf (0.81) (t = 5.48, df = 323; P < 0.005). The correlation coefficient between attitudes towards snow leopards and wolf was The mean attitude score towards both predators varied across the six landscapes (Figure 5.2). The attitude score towards snow leopard was highest in the conservation site of Kibber at 3.07 (±1 se 0.5) and lowest in neighboring site of Langza at (0.47) (Table 5.5). Mean attitude score towards wolves was highest for the conservation site of Kibber 2 (0.59) and lowest for the neighboring site of Langza (0.43) (Table 5.5). 114

14 Between , snow leopards killed 17, 19, 7, 12, 38, and 228 herded-stock at Kibber, Langza, Lingti, Lossar, Pin and Tabo, respectively. During the same duration, snow leopards also killed 42, 36, 3, 6, 39, 4 large-bodied stock at Kibber, Langza, Lingti, Lossar, Pin and Tabo respectively. Wolves killed 35, 109, 3, 54, 11, 7 herded-stock and 42, 36, 3, 6, 39, 4 large-bodied stock at Kibber, Langza, Lingti, Lossar, Pin and Tabo respectively (Table 5.5). Figure 5.2. Mean attitudes score of local people towards large carnivores in six study sites in the Spiti Valley, India. Higher values indicate positive attitudes. Error bars represent 95% CI. Circles and squares represent attitudes towards snow leopards and wolves, respectively. 115

15 Table 5.3: Structure of models used in a Linear Mixed-effect Modelling framework with AIC as model selection criterion used to identify the variables influencing attitudes of local pastoralists towards the snow leopard and wolf at the scale of an individual. All model included study site (SS) and village as fixed effects. Variables and their abbreviations are defined in table 5.2. AIC Snow leopard Random effect Fixed effect Age + gender + education + job + largestock + pea.production + large.killed + largestock:large.killed (1 Village) + (1 SS) Age + gender + education + job + largestock + herded-stock + pop.perception + pea.production (1 Village) + (1 SS) Age + gender + education + job + largestock + herded-stock + carnivore.enc + pop.perception + pea.production + large.killed + herded.killed + largestock:large.killed + herded-stock:small.killed (1 Village) + (1 SS) Age + gender + education + job + largestock + herded-stock + pop.perception + pea.production (1 + Pea.production Village) + (1 SS) Pop.perception + Education + pea.production + gender + Age + largestock + herded-stock + large.killed + herded.killed + carnivore.enc + Job + largestock:large.killed + herded-stock:herded.killed + Largestock:carnivore.enc + smallstock:carnivore.enc + gender:carnivore.enc (1 Village) + (1 SS) age + gender + education (1 Village) + (1 SS) Age + gender + education + job + largestock + herded-stock + carnivore.enc + pea.production (1 Village) + (1 SS) age + gender + education (1 + Education Village) + (1 + SS) Age + gender + education + job + largestock + herded-stock + pop.perception + pea.production (1 + Education Village) + (1 + SS) Pop.perception + Education + pea.production + gender + age + largestock + herded-stock + large.killed + herded.killed + carnivore.enc + Job + largestock:large.killed + herded-stock:herded.killed + Largestock:carnivore.enc + smallstock:carnivore.enc + gender:carnivore.enc AIC Wolf (1 + Education + Pea.production Village) + (1 SS) Largestock + herded-stock + large.killed + herded.killed + largestock:large.killed + herded-stock:herded.killed + gender + carnivore.enc (1 Village) + (1 SS) Largestock + large.killed + largestock:large.killed + gender + carnivore.enc + pop.perception + gender:pop.perception (1 Village) + (1 SS) Largestock + herded-stock + large.killed + herded.killed largestock:large.killed + herded-stock:herded.killed (1 Village) + (1 SS) Null (1 Village) + (1 SS) Largestock + herded-stock + large.killed + herded.killed (1 Village) + (1 SS)

16 Table 5.4: Structure of models used in a Linear Mixed-effect Modelling framework with AIC as model selection criterion used to identify the variables influencing attitudes of local pastoralists towards the snow leopard and wolf at the scale of a village. All model included study site (SS) and village as fixed effects. Variables and their abbreviations are defined in table 5.2. Random effect Fixed effect AIC SL AIC W Village.size + largestock + herded.stock + mean.pea.production (1 Village) + (1 SS) Village.size + largestock + herded.stock + prop.large.killed + prop.herded.killed + mean.pea.production (1 Village) + (1 SS) largestock + prop.large.killed + Village.size (1 Village) + (1 SS) Village.size + largestock + herded.stock + prop.large.killed + prop.herded.killed + mean.pea.production + largestock:prop.large.killed + herded.stock:prop.herded.killed (1 Village) + (1 SS) largestock + prop.large.killed + largestock:prop.large.killed + Village.size (1 Village) + (1 SS) Village.size + mean.pea.production (1 Village) + (1 SS) largestock + prop.large.killed + largestock:prop.large.killed + mean.pea.production (1 Village) + (1 SS) herded.stock + prop.herded.killed + herded.stock:prop.herded.killed + mean.pea.production (1 Village) + (1 SS) herded.stock + prop.herded.killed + Village.size (1 Village) + (1 SS) Null (1 Village) + (1 SS) herded.stock + prop.herded.killed + herded.stock:prop.herded.killed + Village.size (1 Village) + (1 SS) herded.stock + prop.herded.killed (1 Village) + (1 SS) largestock + prop.large.killed + largestock:prop.large.killed (1 Village) + (1 SS) herded.stock + prop.herded.killed + herded.stock:prop.herded.killed (1 Village) + (1 SS) largestock + prop.large.killed (1 Village) + (1 SS)

17 Figure 5.3: Parameter estimates for all the variables included in the multi-level model explaining human attitudes towards the snow leopard. These variables were chosen from the best individual and village level candidate models. Error bars indicate 95% CI calculated using the standard error. * indicates village level variables. MIXED EFFECTS MODELS For the snow leopard, the model including variables from the best models (least AIC) at individual level (table 5.3) and village levels (table 5.4) had a lower AIC (906.6) than either of them used separately. The best models at village and individual levels had an AIC of and 909.3, respectively. Similarly for the wolf, the models with individual and village level variables from the respective best models had a lower AIC (861.9) than the best individual level model (865.37) and village level model (924.6). Individual level At the level of an individual, the best model (least AIC) explaining attitudes towards snow leopards included the total large-bodied stock holding, proportion of large-bodied stock killed by carnivores, gender, age, education, jobs and total green pea production 118

18 (table 5.3). When used together with the village level correlates, education (0.55; ±0.44 CI) and green pea production (0.49; ± 0.40) showed a significant positive association with attitudes towards snow leopards while large-bodied stock holding (-0.50; ±0.39) and age (- 0.04; ±0.02) showed a significantly negative association. Males (1.75; ± 0.7) had significantly more positive attitudes than females (Figure 5.5). The interaction between large-bodied stock holding and large-bodied stock killed by carnivores was also significant (-0.54; ± 0.48). This suggests that people with more large-bodied stock and greater stock depredation had disproportionately lower attitudes towards the snow leopard. The remaining variables Proportion of large-bodied stock killed by carnivores (- 0.18; ± 0.39) and number of sources of income (0.36; ± 0.54) did not have a significant effect (figure 5.3). Individual model best explaining attitudes towards wolves included total herdedstock holding, large-bodied stock holding, gender, perception of wolf population, education, number of sources of income, age and green pea production (table 5.3). When used together with the significant variables at the village-level, education (0.59; ± 0.48) and number of sources of income (0.7; ± 0.61) showed a significant positive association with attitudes towards wolves. Age (-0.05; ±0.02) showed a significant negative relation. Males (1.49; ± 0.8) had significantly higher attitude scores than females (Figure 5.5). Herded-stock holding (-0.37; ±0.43), large-bodied stock holding (-0.24; ± 43), perception of wolf population (Don't know 0.3; ±1.1; increasing -0.76; ±1.1), and green pea production (0.35; ±0.43) did not have significant effect on attitudes towards wolves (figure 5.4). Village level The best model (least AIC) explaining the attitudes towards snow leopards at the scale of the village included the size of the village, total large-bodied stock holding of the village, total herded-stock holding of the village, and mean green pea production of the village. When used together with variables from the best individual level model, size of the village (0.9; ± 0.5) showed a significant positive relationship with attitudes towards snow leopards and total large-bodied stock holding of the villages (-0.7; ±0.54) showed a significant negative relationship. Total herded-stock holding and mean green pea 119

19 production of the village did not have a significant effect on attitudes towards snow leopards at the scale of the village (figure 5.3). Figure 5.4: Parameter estimates for all the variables included in the multi-level model explaining attitudes towards the wolf. These variables were chosen from the best individual and village level candidate models. Error bars indicate 95% CI calculated using the standard error. * indicates village level variables. The best model (least AIC) explaining human attitudes towards wolves at the scale of the village included the total herded-stock holding of the village, the proportion of herdedstock killed by the wolf and the mean green pea production. It also included the interaction between total herded-stock holding and proportion of herded-stock killed by wolves. When used together with variables from the best model at the individual level, only mean green pea production (0.6; ± 0.5) of the village showed a significant positive relationship with attitudes towards wolves at the village level. Proportion of herded-stock killed (-0.7; ± 0.45) by wolves showed a significant negative relation with attitudes. Total herded-stock holding of the village (0.08; ± 0.43) and its interaction with the proportion of 120

20 herded-stock killed by carnivores (-0.35; ± 0.4) did not have a significant effect on attitudes to wolves at the village level. Figure 5.5: Mean attitude scores of male and female respondents (circle and square, respectively) towards wild carnivores in six sites in the Spiti Valley, India. Attitudes towards snow leopard a) and wolf b). Error bars represent 95% CI. 121

21 Table 5.5. Attitudes of local pastoral people towards snow leopards and wolves, and livestock depredation over by these carnivores across six study sites in Spiti Valley, India Study site Mean attitude to snow leopard Mean attitude to wolf Largestock damage by snow leopard Largestock damage by wolf Herdedstock damage by snow leopard Herdedstock damage by wolf Kibber * 3.07 (0.5) 2 (0.59) Langza (0.47) (0.43) Lingti 0.43 (0.44) 0.37 (0.46) Lossar 2.25 (0.49) 1.23 (0.51) Pin 1.85 (0.40) 1.59 (0.53) Tabo 2.45 (0.44) 1.86 (0.52) * indicates study sites with long-term conservation program Discussion HUMAN ATTITUDES TOWARDS SNOW LEOPARDS AND WOLVES In four of our five study sites that did not have long-term conservation programs, people had positive to neutral attitudes towards carnivores. This suggests that local peoples attitudes need not always be at odds with wildlife conservation objectives despite high levels of livestock losses to wild carnivores. In our study sites, the generally positive attitude could be due to the prevalence of Buddhism, which has been known to positively influence attitudes and on-ground conservation status in other areas such as Tibet (Salick et al. 2007). We recorded significant differences in the human attitudes towards snow leopards and wolves, with people being generally more positive towards snow leopards. This pattern was consistent across all the six study sites, including the one with a long-term conservation program. Similar differences in attitudes towards carnivore species have been recorded elsewhere. In the North American Rockies, local people were more positive towards puma Puma concolor than towards the wolf (Kellert et al. 1996). In Spiti Valley itself, Suryawanshi et al. (2013) reported that local people perceived wolves to be a larger problem for livestock than snow leopards even though data on actual livestock damage by snow leopard and wolves suggested otherwise. A strong cultural bias against the wolf is reflected in strong negative attitudes and relatively high persecution compared to the 122

22 snow leopard (Mishra 1997). Human attitudes towards a species are reportedly influenced by the physical and behavioural characteristics of the carnivore, and their cultural and historical associations (Kellert et al. 1996; Kleiven et al. 2004). Notwithstanding the differences in human attitudes towards the two species, the high correlation in attitudes towards snow leopards and wolves recorded in our study suggests that conservation programs aimed at a charismatic species like the snow leopard could benefit other sympatric carnivores. Indeed, at the scale of the study site, we found that the presence of a conservation program seemed to have a significant positive impact on human attitudes towards both snow leopards and wolves, though attitudes towards wolves remained less positive compared to snow leopards. Specifically, the neighboring sites of Kibber and Langza suffered amongst the highest losses to large carnivores among all our study sites. Yet, people in the conservation site of Kibber had the most positive attitude to snow leopards and wolves amongst all our six study sites, while people in Langza had the most negative attitude. With the continuation and expansion of the relatively recent conservation program in Langza, we expect peoples attitudes to turn significantly positive over time. IMPORTANCE OF MULTI-SCALE ANALYSES Humans make decisions at multiple levels of social organization for instance, from the individual, to the community, to the State. We therefore expected the importance of factors that influence human attitudes towards carnivores in conflict to vary with scale from the level of the individual to the level of the community (village). Indeed, we found considerable differences in the factors as the scale of analysis changed. For instance, the most important factors affecting attitudes towards the carnivores at the individual level were gender, education, and the age of the respondent (for both wolves and snow leopards), number of income sources in the family (wolves), agricultural production and large stock holdings (snow leopards). At the level of the community, the most important factors influencing human attitudes included the number of herded-stock killed by wolves and the mean agricultural production (wolves), and village size and large stock holdings (snow leopards). 123

23 This scale-dependence in the importance of factors influencing human attitudes towards carnivores has important implications. For instance, the role of the extent of livestock depredation in influencing attitudes towards the wolf at the level of the community implies that people weigh the risk of carnivore predation on their livestock based on cumulative livestock damage suffered by the larger community. Thus, while education and having multiple sources of income seem to help in improving the attitudes of individuals towards wolves, these effects may be overridden if the individual lives in a community that experiences relatively high levels of livestock losses. Here, we would like to underscore that a simple individual based examination of the correlates of human attitudes towards carnivores, as has been done in previous studies (e.g. Lohr et al. 1996; Stronen et al. 2007), would have missed this insight and the importance of processes that operate at higher scales of social organization. OTHER DRIVERS OF HUMAN ATTITUDES TOWARDS SNOW LEOPARDS AND WOLVES The importance of village-level agricultural production and of family-level sources of income suggests that a lowered dependence on livestock for income helps in improving attitudes towards carnivores. Interestingly, in the case of the snow leopard, livestock depredation did not emerge as an important factor influencing peoples attitudes at either the individual or the village level. The most important factors influencing the attitudes of local people towards snow leopards at the level of the individual were gender, education, age, agricultural production and holdings of yak and horses (large stock). At the village level, the most important factors were village size and holdings of yak and horses. This is consistent with Suryawanshi et al. (2013) where we found villages with higher holdings of yak and horses perceived the snow leopard to be a greater threat to livestock compared to villages with smaller holdings of large stock. In general, at the level of the individual, literate people are expected to be relatively more aware of wildlife protection laws and conservation value of these carnivores. With growing tourism in our study sites, young literate people may also be sensing economic 124

24 opportunities leading to improved attitudes. Thus access to formal and conservation education programs in areas where people are expected to co-exist with large carnivores is likely to be effective in increasing the social carrying capacity for the carnivores. In human-wildlife conflict situations, women are known to bear many hidden costs, such as reduced nutrition, increased labour, reduced security, and depression (Ogra 2008). This presumably explains the relatively negative attitudes of women compared to men towards wild carnivores. Empowerment of women is generally not the conservation agenda. However, our results suggest that while implementing conflict management projects, or others such as notification of a new protected area or implementation of conservation programs, working with women and helping improve their attitudes to wildlife may have far-reaching benefits for wildlife conservation. People with higher income from cash crop agriculture are expected to be less dependent on the income from livestock, and often find it hard to provide enough care for their livestock. Such people are likely to be less affected by the damage caused by carnivores. Conversely, people with lower crop income and larger livestock holding may be more affected by the damage caused by carnivores. While this may suggest promoting agriculture over livestock husbandry in Trans-Himalayan region, such a shift in production system could lead to a shift in human-wildlife conflicts from the current problem of livestock depredation by large carnivores to crop depredation by large herbivores. Instead, we recommend diversification of income sources. Altogether, our work suggests that scaling-up from the individual to higher levels of social organization can highlight important factors which would have been otherwise ignored that influence attitudes of people towards wildlife in conflict, or towards conservation efforts in general. Such a distinction of factors based on scale should be helpful in targeting appropriate conservation measures at the appropriate scale. This insight also reiterates the need for conflict management programs to be multi-pronged, directed towards reducing the extent of livestock damage by carnivores, buffering the economic losses through diversification of income opportunities and/or off-setting losses, and increasing the social carrying capacity through education. 125

25 Acknowledgements We thank the Forest Department of Himachal Pradesh for permissions. Whitley Fund for Nature, Conservation Leadership Program for funding. Neha, Nisarg, Dincy, Rinchen Tobgay, Tandup cherring for help in the field and all the 383 participants. References Allendorf TD, (2007). Residents' attitudes towards three protected areas in South Western Nepal. Biodiversity and Conservation. 16: Badola R, (1998). Attitudes of local people towards conservation and alternatives to forest resources: A case study from the lower Himalayas. Biodiversity and Conservation. 7: Bagchi, S. & Mishra, C. (2006). Living with large carnivores: predation on livestock by the snow leopard (Uncia uncia). Journal of Zoology. 268: Bates D, Maechler M and Bolker B (2012). lme4: Linear mixed-effects models using S4 classes. R package version Bjerke T, Kaltenborn BP, & Thrane C. (2001). Sociodemographic correlates of fear-related attitudes toward the wolf (Canis lupus lupus). A survey in southeastern Norway. Fauna norvegica. 21: Champion FW, Seth SK. (1968). A revised survey of the forest types of India. Government of India Press, Nasik Decker DJ, & Purdy KG. (1988). Toward a concept of wildlife acceptance capacity in wildlife management. Wildlife Society Bulletin. 16: Gelman A, & Hill J. (2007). Data analysis using regression and multilevel/hierarchical models. Cambridge University Press. Hussain S. (2000). Protecting the Snow Leopard and Enhancing Farmers Livelihoods. Mountain Research and Development. 20: Ikeda N. (2004). Economic impacts of livestock depredation by snow leopard Uncia uncia in the Kanchenjunga Conservation Area, Nepal Himalaya. Environmental Conservation, 31: Johnson JB, & Omland KS. (2004). Model selection in ecology and evolution. Trends in Ecology and Evolution, 19:

26 Kaczensky P, Enkhsaikhan N, Ganbaatar O, & Walzer C. (2008). The Great Gobi B Strictly Protected Area in Mongolia-refuge or sink for wolves Canis lupus in the Gobi. Wildlife Biology. 14: Kellert SR. (1985). The public and the timber wolf in Minnesota. Yale University, School of Forestry and Environmental Studies. Kellert SR. (1991). Public views of wolf restoration in Michigan. In Transactions of the North American Wildlife and Natural Resources Conference. 56: Kellert SR, Black M, Rush CR. & Bath AJ. (1996). Human Culture and Large Carnivore Conservation in North America. Conservation Biology. 10: Kleiven J, Bjerke T. & Kaltenborn BP. (2004). Factors influencing the social acceptability of large carnivore behaviours. Biodiversity and Conservation. 13: Linnell JDC, Swenson JE, & Andersen R. (2001). Predators and people: conservation of large carnivores is possible at high human densities if management policy is favourable. Animal Conservation 4: Liu F, Mcshea WJ, Garshelis DL, Zhu X, Wang D, & Shao L. (2011). Human-wildlife conflicts influence attitudes but not necessarily behaviors: Factors driving the poaching of bears in China. Biological Conservation. 144, Lohr C, Ballard WB, & Bath A. (1996). Attitudes toward gray wolf reintroduction to New Brunswick. Wildlife Society Bulletin McCarthy TM, & Chapron G. (2003) Snow Leopard Survival Strategy. ISLT and SLN, Seattle, USA. McCarthy TM, Fuller, TK, & Munkhtsog B. (2005). Movements and activities of snow leopards in Southwestern Mongolia. Biological Conservation. 124: McCarthy TM, Murray K, Sharma K, Johansson O. (2010). Preliminary results of a longterm study of snow leopards in South Gobi, Mongolia. Cat News. 53: Mishra, C. (1997). Livestock depredation by large carnivores in the Indian Trans- Himalaya: conflict perceptions and conservation prospects. Environmental Conservation. 24: Mishra C. (2000). Socioeconomic transition and wildlife conservation in the Indian Trans- Himalaya. Journal-Bombay Natural History Society. 97: Mishra C, Allen P, McCarthy T, Madhusudan MD, Bayarjargal A. & Prins HH. (2003). The Role of Incentive Programs in Conserving the Snow Leopard. Conservation Biology. 127

27 17: Mishra C, Bagchi S, Namgail T, & Bhatnagar YV. (2010). Multiple Use of Trans-Himalayan Rangelands: ReconcilingHuman Livelihoods withwildlife Conservation. Wild rangelands: conserving wildlife while maintaining livestock in semi-arid ecosystems. 6: Namgail T, Bhatnagar YV. & Fox JL. (2007). Carnivore-caused Livestock Mortality in Trans-Himalaya. Environmental Management Ogra M, (2008). Human-wildlife conflicts and gender in protected area borderlands: A study of costs, perceptions and vulnerabilities from Uttarakhand, India. Geoforum. 39: Oli, M.K., Taylor, I.R. & Rogers, E.M. (1994). Snow leopard Panthera uncia predation of livestock: An assessment of local perception in the Annapurna Conservation Area, Nepal. Biological Conservation. 68: R Development Core Team (2011). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN , URL Salick J, Amend A, Anderson D, Hoffmiester K, Gunn B, Zhendong F. (2007). Tibetan sacred sites conserve old growth trees and cover in the Eastern Himalayas. Biodiversity and Conservation. 16: Stronen AV, Brook RK, Paquet PC, & McLachlan S. (2007). Farmer attitudes toward wolves: Implications for the role of predators in managing disease. Biological conservation. 135: Suryawanshi, K. R., Bhatnagar, Y. V., Redpath, S., & Mishra, C. (2013). People, predators and perceptions: patterns of livestock depredation by snow leopards and wolves. Journal of Applied Ecology. DOI: / Treves A, & Naughton-Treves L. (2005). Evaluating lethal control in the management of human-wildlife conflict. Conservation Biology Series-Cambridge- 9: Wang SW, Lassoie JP, Curtis P, (2006) Farmer attitudes towards conservation in Jigme Singye Wangchuck National Park, Bhutan. Environmental Conservation. 33: West P, Igoe J, Brockington D (2006). Parks and people: Social impact of protected areas. Annual review of Anthropology. 35: Williams CK, Ericsson G, & Heberlein TA. (2002). A quantitative summary of attitudes 128

28 toward wolves and their reintroduction ( ). Wildlife Society Bulletin. 30: Woodroffe R, Thirgood SJ, & Rabinowitz A. (2005). People and Wildlife: Conflict or coexistence. Cambridge University Press, Cambridge, UK. Zinn HC, Manfredo MJ. & Vaske JJ. (2000). Social psychological bases for stakeholder acceptance capacity. Human Dimensions of Wildlife. 5:

29 Picture 8. The overall attitudes of local pastoralists in Spiti Valley towards carnivores such as the snow leopard (Panthera uncia) and wolf (Canis lupus) were generally positive. This is probably due the prevalence of Buddhism in this region which is known to positively influence attitudes towards wildlife.