Introduction
Asia's mountains support a rich diversity of ungulates that sustain large carnivores such as the Endangered snow leopard Panthera uncia (Schaller, Reference Schaller1977, Reference Schaller1998). Relatively little information is available on the abundance and population trends of these mountain ungulates, however, because of the remoteness of their habitats and a lack of robust scientific studies (Singh & Milner-Gulland, Reference Singh and Milner-Gulland2011). Changing livestock husbandry practices (Berger et al., Reference Berger, Buuveibaatar and Mishra2013), mining and associated development (Farrington, Reference Farrington2005; Kaczensky et al., Reference Kaczensky, Kuehn, Lhagvasuren, Pietsch, Yang and Walzer2011), and poaching (Pratt et al., Reference Pratt, Macmillan and Gordon2004) threaten the survival of many of these ungulate species and the predators that depend on them.
The availability of wild-ungulate prey is a critical determinant of carnivore density (Karanth et al., Reference Karanth, Nichols, Kumar, Link and Hines2004). However, in the mountains of Asia the rugged habitat and lack of sufficient expertise have posed a challenge to the estimation and monitoring of ungulate populations (Singh & Milner-Gulland, Reference Singh and Milner-Gulland2011). Many of the established methods of ungulate population estimation, such as distance sampling, are difficult to implement in mountainous terrain because of the inapplicability of their assumptions (Suryawanshi et al., Reference Suryawanshi, Bhatnagar and Mishra2012). Alternatives such as aerial surveys are expensive and often unsafe in mountainous landscapes.
The primary occupation of the local pastoralists is cashmere production from goats. In response to demand from the global market for cashmere, the goat population in large parts of the mountains of Central Asia is increasing (Bhatnagar et al., Reference Bhatnagar, Wangchuk, Prins, Van Wieren and Mishra2006; Berger et al., Reference Berger, Buuveibaatar and Mishra2013). Livestock species such as goats are reported to compete with wild ungulate species for forage in other regions (Bagchi et al., Reference Bagchi, Mishra and Bhatnagar2004; Mishra et al., Reference Mishra, Van Wieren, Ketner, Heitkönig and Prins2004), and the presence of herders and herding dogs creates competition for space (Namgail et al., Reference Namgail, Fox and Bhatnagar2007). Berger et al. (Reference Berger, Buuveibaatar and Mishra2013) reported that the biomass of native wild ungulates across seven study areas in Mongolia, India, and China's Tibetan Plateau is now < 5% that of livestock. Conservation of native ungulates alongside a pastoral industry such as cashmere production will require rigorous monitoring and incentives for wildlife-friendly pastoralism (Berger et al., Reference Berger, Buuveibaatar and Mishra2013). Monitoring of wild ungulate abundance is therefore fundamental to any such effort (Ale et al., Reference Ale, Yonzon and Thapa2007; Xu et al., Reference Xu, Jiang, Li, Guo, Da and Cui2008), and will also facilitate assessment of the size of the snow leopard population that could be sustained by the available prey.
The Tost Local Protected Area is an important corridor between the Great Gobi Strictly Protected Area A in the west and the Gobi Gurvan Saikhan National Park in the north. An ongoing radio-telemetry study of the snow leopard has already recorded one case of a subadult snow leopard dispersing from the Tost Mountains in the Local Protected Area to the Nemegt Mountains in the Gobi Gurvan Saikhan National Park. Local Protected Area status affords a relatively low level of protection under Mongolian law, however, and it is difficult to protect the area from mining. Supported by the local government authorities, local pastoralists and conservationists, the elevation of the protection status of Tost Local Protected Area to that of Nature Reserve is being considered by the National Government. Thus, there is an urgent need for robust information on the conservation value and potential of this region.
The snow leopard population of the Tost Mountains is being monitored closely using camera trap-based multi-season models (Sharma et al., Reference Sharma, Bayrakcismith, Tumursukh, Johansson, Sevger, McCarthy and Mishra2014). However, little information has so far been available on the status of their primary wild-ungulate prey, the ibex Capra sibirica and argali Ovis ammon. The ibex is one of the preferred prey species of the snow leopard across its range (Jumabay-Uulu et al., Reference Jumabay-Uulu, Wegge, Mishra and Sharma2014; Lyngdoh et al., Reference Lyngdoh, Shrotriya, Goyal, Clements, Hayward and Habib2014).
Up to 20% of the snow leopard's diet in the Tost Local Protected Area consists of livestock (Shehzad et al., Reference Shehzad, McCarthy, Pompanon, Purevjav, Coissac, Riaz and Taberlet2012), and retaliatory persecution of snow leopards by local herders is an important conservation concern. The availability of wild prey has significant implications for the livestock predation behaviour of wild carnivores (Suryawanshi et al., Reference Suryawanshi, Bhatnagar and Mishra2012; Odden et al., Reference Odden, Nilsen and Linnell2013; Gervasi et al., Reference Gervasi, Nilsen, Odden, Bouyer and Linnell2014). Thus, understanding the availability and trends of wild ungulate prey over time is important for conservation management in the region.
The primary objective of this study was to estimate the population of ibex and argali in the Tost Mountains, which comprise the snow leopard habitats of the Tost Local Protected Area. We also examined the demography of the two species, to understand the survival of various age–sex categories of the ibex.
Study area
Our study area comprised 1,400 km2 of the mountainous region of Tost, in South Gobi province, Mongolia (Fig. 1). The Tost Local Protected Area was declared by the administration of Gurvantes soum (district) in October 2010, after mining leases were issued for the region. The Protected Area has high conservation value for species such as the Endangered snow leopard, and argali and ibex (McCarthy et al., Reference McCarthy, Murray, Sharma and Johansson2010), and is valued by the local pastoralists for grazing livestock. The temperature in the region ranges between −35°C in winter and +35°C in summer, and the altitude is 1,600–2,400 m. The Nemegt Mountains of the Gobi Gurvan Saikhan National Park are north of the surveyed area. Towards the west the study area is connected to the Great Gobi Strictly Protected Area A via a fragmented network of hillocks. The international border with China lies c. 60 km south of the study area. The town of Gurvantes and the Noyon Mountains lie east of the study area. Livestock herding is the primary occupation of the local people. Goats (22,070) and sheep (1,263) are the most common livestock, followed by camels (414), horses (329) and cattle (47), based on a census conducted in 2013 (LT, unpubl. data). Ibex, argali and tolai hare Lepus tolai are the common wild herbivores in the region, and snow leopard, wolf Canis lupus, red fox Vulpes vulpes, corsac fox Vulpes corsac and lynx Lynx lynx are the large predators.
Fig. 1 Map of the survey area (dark shade). Existing Protected Areas are shaded (light). The Tost Local Protected Area is denoted by dotted line in the map and solid black in the inset of Mongolia.
Methods
Ibex and argali populations were estimated in 2012 and 2013 using the double-observer survey method (Forsyth & Hickling, Reference Forsyth and Hickling1997; Suryawanshi et al., Reference Suryawanshi, Bhatnagar and Mishra2012). In 2012 the study area was divided into 15 blocks of c. 100 km2, depending upon the topography. Survey blocks were separated by flat areas between mountain chains. The 100 km2 blocks were further divided into cells of c. 25 km2. The shape and size of each cell depended on the terrain and the logistics of accessing and surveying the area. The surveys were conducted during late autumn (10–24 November 2012 and 9–19 November 2013), which is the mating season for the ibex and argali in the region. It is usually easier to detect ibex during the mating season as they form larger groups, remain active for longer periods and are less easily disturbed. Eight trained observers were divided into four teams, and each team surveyed one cell in 1 day, walking along predetermined trails and scanning using binoculars. Based on initial trials, the second observer started the survey c. 10–15 minutes after the first. This duration was chosen to balance potential overestimation as a result of evasive movement of ungulates after the first survey and potential underestimation if the time between the two surveys was too short, resulting in a lack of independence between the two surveys. If the first observer came within the sight of the second, the second observer waited to increase the distance between the two. The protocol precluded any unintentional visual cues; for example, prolonged interest in a particular direction by one observer could influence the other observer's survey.
The observers recorded ungulate group size, age × sex categorization, distance to group, the name of the pasture where the group was encountered, and any other matters that could help identify the observed groups. At the end of the survey the observers discussed and identified the groups seen by both observers and those seen by only one, based on group size and composition, location and any other notes. Density was estimated by dividing the abundance estimate by the area surveyed (Oli, Reference Oli1994; Forsyth & Hickling, Reference Forsyth and Hickling1997; Suryawanshi et al., Reference Suryawanshi, Bhatnagar and Mishra2012).
The demography of the ibex population was surveyed eight times during May 2012–November 2013. The surveys were carried out over 7–10 days in the same 15 blocks used in the double-observer surveys. We surveyed 10–20 km in each block, for even coverage of the study area. Areas that were inaccessible by all-terrain vehicle or motorcycle were surveyed on foot. Groups of ibex were observed from a distance of 100–500 m using 8 × 32 binoculars and a 20–60 × 60 spotting scope, and categorized according to the following age–sex categories: kid (< 1 year old), yearling (1–2 years old), adult female and adult male (Schaller, Reference Schaller1977). We conducted 10,000 bootstraps to assess the 95% confidence intervals of the young-to-female and male-to-female ratios within each survey, using herd as the sampling unit. The median values were used as the estimates of young-to-female and male-to-female ratios, and the 0.025 and 0.975 quantiles were used as 95% confidence intervals.
Results
Double-observer survey estimates of ibex and argali abundance were 899 (95% CI 727–1,071) and 108 (95% CI 47–169) individuals in 2012 and 1,051 (95% CI 857–1,245) and 230 (95% CI 120–340) in 2013, respectively. The mean group sizes of ibex encountered during the surveys were 5.24 and 5.04 in 2012 and 2013, respectively. Mean group sizes of argali were 7.72 and 7.5 in 2012 and 2013, respectively. This yields density estimates of 0.64 and 0.75 ibex km−2 and 0.08 and 0.16 argali km−2 in 2012 and 2013, respectively. The difference in ibex population estimate for the 2 years was not statistically significant (z = 1.16, one-tailed P = 0.12). The increase in the argali population estimate was significant (z = 1.9, one-tailed P = 0.03). All parameters are summarized in Table 1.
Table 1 Results of spaced double-observer surveys of ibex Capra ibex and argali Ovis ammon in the Tost Mountains, South Gobi, Mongolia (Fig. 1), in 2012 and 2013.
The ratio of kids (< 1 year) to adult females was 0.51 (95% CI 0.39–0.65) in May 2012 and 0.78 (95% CI 0.6–1.0) in November 2013 (Fig. 2a). The ratio of yearlings (1–2 years) to adult females was 0.09 (95% CI 0.03–0.17) in September 2012 and 0.37 (95% CI 0.23–0.54) in June 2013. The ratio of young to female decreased in summer (August) and again after winter, with a significant drop after weaning at 12 months of age. The adult male-to-female ratio showed a general decreasing trend from 0.94 (95% CI 0.6–1.44) in May 2012 to 0.67 (95% CI 0.56–0.81) in November 2012 and increased to 1.08 (95% CI 0.81–1.49) in November 2013 (Fig. 2b). This suggests male-biased mortality in 2012 and female-biased mortality in 2013. The highest mean group sizes were recorded in May 2012 (7.3) and June 2013 (7.5). The lowest mean group sizes were recorded in August 2012 (4.0) and August 2013 (4.7; Table 2).
Fig. 2 (a) Young : adult female ratio and (b) male : female ratio of ibex Capra ibex in Tost Local Protected Area (Fig. 1), from May to November 2012 and from June to November 2013. Survival of young ibex dropped significantly during winter and post-weaning. Points are the median values of 10,000 bootstraps to estimate the 95% confidence interval (error bars), with herd as the sampling unit.
Table 2 Numbers of ibex assigned to various age–sex categories during surveys in the Tost Mountains, South Gobi, Mongolia.
Discussion
We used the statistically robust double-observer survey method to estimate the abundance of ibex and argali in the mountainous regions of the Tost Local Protected Area. The detection probabilities recorded (0.56 and 0.36 for ibex and argali, respectively) were lower than in other places (0.79, 0.74; Suryawanshi et al., Reference Suryawanshi, Bhatnagar and Mishra2012) where this method has been used but our confidence intervals were reasonable because we used a larger sample size.
Suryawanshi et al. (Reference Suryawanshi, Bhatnagar and Mishra2012) recommended that for double-observer surveys of mountain ungulates the two observers be separated in space or time to minimize potential under-estimation as a result of autocorrelation between the observers. However, this recommendation was based on their study in the Himalayas (Suryawanshi et al., Reference Suryawanshi, Bhatnagar and Mishra2012). In Tost the animals were comparatively shy and our initial observations indicated that such a separation was likely to result in overestimation because of animals leaving the survey area between the two surveys. We therefore used a minimal 10–15 minute separation between the observers to minimize overestimation. Despite this change, a lower detection probability was recorded consistently for the second observer, which suggests there was evasive movement by animals after the first survey. For future surveys we suggest the use of Bayesian approaches proposed by Barker et al. (Reference Barker, Forsyth and Wood2014) to address the potential biases in population estimates.
The estimated ibex population size remained similar across the 2 years of monitoring. The argali population estimate, however, doubled from 2012 to 2013. We suspect this increase may be attributed to the movement of argali from smaller hills in the north and west of the survey area, which are used by argali but do not contain ibex habitat. By November 2013 a drought in the southern Gobi region had restricted the movement of ungulates to around the remaining waterholes. Occasionally we observed aggregations of argali in the vicinity of a waterhole on the edge of our study site in late October (LT, pers. obs.).
Our demographic surveys cannot be used to estimate the ibex kid mortality for the first few weeks after birth; beyond this time up to 12 months, however, mortality rates among ibex kids are low. A lower young-to-female ratio, with wide confidence intervals, recorded in August may be attributable to sampling error, as ibex become difficult to sight as summer progresses. Importantly, the young-to-female ratio declined after 1 year, which we suspect was largely a result of predation when the yearlings started feeding by themselves in the herd. The adult female-to-male ratios were lowest at the start (May 2012) and end of the monitoring (November 2013), peaking in November 2012. Such a pattern could be a result of either sighting bias in the demographic surveys (separate from the double-observer surveys) or selective (sex-biased) predation by carnivores. We consider our surveys to be unbiased as they covered the entire study area and all ibex habitats. The surrounding hills in the north and west are not suitable habitat for ibex. Such a pattern could also arise as a result of long-distance movement by male ibex. We do not think this was the cause, however, as we surveyed a relatively large area. We recommend further investigation of the hypothesis of gender-biased predation, using marked animals.
Oli (Reference Oli1994) recorded a 1 : 114–159 ratio of snow leopard to blue sheep Pseudois nayaur, by weight. Blue sheep is another preferred prey of the snow leopard (Lyngdoh et al., Reference Lyngdoh, Shrotriya, Goyal, Clements, Hayward and Habib2014) in other parts of the snow leopard's range. Considering the mean weight of ibex to be 60 kg (Hess, Reference Hess and Parker1990), argali 90 kg (Valdez, Reference Valdez and Parker1990) and snow leopard 40 kg (Oli, Reference Oli1994), the estimated biomass of wild herbivores in the Tost Mountains was 63,720 kg in 2012 and 83,760 kg in 2013. Using Oli's (Reference Oli1994) predator : prey ratios we estimate that the ibex and argali populations of the Tost Mountains could have supported c. 10–14 and 13–18 adult snow leopards in 2012 and 2013, respectively. Carbone & Gittleman (Reference Carbone and Gittleman2002) suggested that 10,000 kg of prey could support c. 90 kg of carnivore biomass. Using their conversion, our estimated prey biomass would have been adequate to support 14 adult snow leopards in 2012 and 18 in 2013. Both these estimates of snow leopard populations are comparable to the estimates of 12 (95% CI 11–15) and 14 (95% CI 14–21) determined by means of camera-trap-based mark–recapture in the same (but larger, 1,684 km2) study area (Sharma et al., Reference Sharma, Bayrakcismith, Tumursukh, Johansson, Sevger, McCarthy and Mishra2014).
Although the estimated density of wild ungulates (0.71 and 0.91 km−2 in 2012 and 2013, respectively) may seem sufficient to support 14–18 adult snow leopards, these densities are an order of magnitude lower than the stocking density of livestock (17.1 km−2) in this region. Consistent with Berger et al. (Reference Berger, Buuveibaatar and Mishra2013), the biomass of native ungulates in Tost Local Protected Area is c. 6% of the livestock biomass. It is known that there is a significant overlap in the diet of domestic goats and sheep and that of ibex (Bagchi et al., Reference Bagchi, Mishra and Bhatnagar2004). Other studies have found that competition for forage from livestock and disturbance from herding dogs could reduce the density of wild ungulate herbivores (Mishra et al., Reference Mishra, Van Wieren, Ketner, Heitkönig and Prins2004; Namgail et al., Reference Namgail, Fox and Bhatnagar2007). This could be a cause of relatively low ibex biomass compared to livestock in the region, and requires further investigation.
Our results suggest that the Tost Mountains are an important site for the conservation of ibex, argali and the Endangered snow leopard. This area also serves as a corridor connecting the snow leopard and ibex habitats of the Great Gobi Strictly Protected Area and Gobi Gurvan Saikhan National Park. We are completing a management plan for the Local Protected Area, and continue to work with relevant stakeholders, including the local community and the local government, to have the Local Protected Area elevated in status to a Nature Reserve by the Mongolian parliament. Despite its significance for conservation and pastoralism, the Local Protected Area remains vulnerable to the threat of mining; designation as a Nature Reserve would protect it against this threat.
Acknowledgements
We thank People's Trust for Endangered Species for funding this work, and the Fondation Segré–Whitley Fund for Nature, David Shepherd Wildlife Foundation, and Disney Worldwide Conservation Fund for supporting our programmes in Mongolia and across the snow leopard range. We thank Panthera, a partner organization in the long-term ecological study, who provided core funding, equipment, staff and expertise. We thank Odbayar Tumendemberel, Koustubh Sharma, Purevjav Lkhagvajav and Bayarjargal Agvaantseren for help during the various stages of the work, and the rangers from Gobi Gurvan Saikhan National Park who participated in the surveys.
Biographical sketches
Lkhagvasumberel Tumursukh studies the ecology of the snow leopard and its prey, with applications for the conservation of these species in Mongolia. Kulbhushansingh Suryawanshi is interested in the ecology and conservation of large carnivores and their ecosystems. Charudutt Mishra studies the ecology of large carnivores and herbivores, rangelands, human ecology and conservation conflicts in the Himalayas and Central Asia. Tom McCarthy has spent more than 2 decades studying and seeking to conserve snow leopards across their range in Central Asia. Bazartseren Boldgiv is an experimental ecologist focusing on the effects of climate change and land use on biodiversity and ecological system dynamics in Mongolia.
