sign in sign up
<<
Home , Indian grey mongoose, Ruddy mongoose

Assessing sloth as surrogates for carnivore conservation in Sri Lanka

Shyamala Ratnayeke1,3 and Frank T. van Manen2,4

1University of Tennessee, Department of Forestry, Wildlife and Fisheries, 274 Ellington Plant Sciences Building, Knoxville, TN 37996, USA 2United States Geological Survey, Leetown Science Center, University of Tennessee, Department of Forestry, Wildlife and Fisheries, 274 Ellington Plant Sciences Building, Knoxville, TN 37996, USA

Abstract: Bears are large, charismatic whose presence often garners conservation attention. Because healthy populations typically require large, contiguous areas of habitat, land conservation actions often are assumed to benefit co-occurring , including other mammalian carnivores. However, we are not aware of an empirical test of this assumption. We used remote camera data from 2 national parks in Sri Lanka to test the hypothesis that the frequency of detection of sloth bears (Melursus ursinus) is associated with greater richness of carnivore species. We focused on mammalian carnivores because they play a pivotal role in the stability of ecological communities and are among Sri Lanka’s most endangered species. Seven of Sri Lanka’s carnivores are listed as endangered, vulnerable, or near threatened, and little empirical information exists on their status and distribution. During 2002–03, we placed camera traps at 152 sites to document carnivore species presence. We used Poisson regression to develop predictive models for 3 categories of dependent variables: species richness of (1) all carnivores, (2) carnivores considered at risk, and (3) carnivores of least conservation concern. For each category, we analyzed 8 a priori models based on combinations of detections, sam- ple year, and study area and used Akaike’s information criterion (AICc) to test our research hypothesis. We detected sloth bears at 55 camera sites and detected 13 of Sri Lanka’s 14 species. Species richness of all carnivores showed positive associations with the number of sloth bear detections, regardless of study area. Sloth bear detections were also positively associated with species richness of carnivores at risk across both study years and study areas, but not with species richness of common carnivores. Sloth bears may serve as a valuable surrogate species whose habitat protection would contribute to conservation of other carnivores in Sri Lanka.

Key words: Carnivora, conservation surrogate, Melursus ursinus, Poisson regression, remote camera, sloth bear, species richness, Sri Lanka

Ursus 23(2):206–217 (2012)

Challenges associated with managing growing that they can serve as flagships for conservation numbers of threatened and endangered species have (Simberloff 1998). If species have area and resource prompted various approaches for prioritizing what requirements that would encompass those of other species and ecosystems should be protected (Noss species, they could serve as conservation umbrellas 1990, Lambeck 1997, Simberloff 1998). Certain (Frankel and Soule´ 1981, Noss 1990, Ozaki et al. species may serve as indicators of biodiversity, 2006). The umbrella species concept has been used to threatened habitats, or ecological processes (Caro delineate protected areas (Caro 2003) and to identify and O’Doherty 1999), or are sufficiently charismatic or predict regions of high biodiversity (Cardoso et al. 2004, Thorne et al. 2006). 3Current address: School of Biology, The University of Sri Lanka is one such country where conservation Dodoma, PO Box 338, Dodoma, Tanzania; sratnaye- challenges may be effectively addressed using the [email protected] 4Current address: US Geological Survey, Northern Rocky sloth bear (Melursus ursinus) as a surrogate for Mountain Science Center, Interagency Grizzly Bear Study conservation. Fourteen species of Carnivora occur Team, Bozeman, MT 59715, USA in Sri Lanka (Phillips 1984; Table 1). Mammalian

206 SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen 207

Table 1. Carnivore species of Sri Lanka (Phillips 1984), International Union for Conservation of Nature/Species Survival Commission (IUCN/SSC) Red List 2011 status, and population trend based on entire geo- graphic distribution. Family Common name Scientific name IUCN/SSC Red List Population trend jungle cat chaus least concern declining fishing cat viverrinus endangered declining rusty spotted cat Prionailurus rubiginosus vulnerable declining leopard pardus near threatened declining aureus least concern increasing Ursidae sloth bear Melursus ursinus vulnerable declining European lutra near threatened declining Viverricula indica least concern stable common palm civet hermaphroditus least concern stable Paradoxurus zeylonensis vulnerable declining Herpestidae Indian grey edwardsi least concern unknown Indian brown mongoose H. fuscus vulnerable declining Indian ruddy mongoose H. smithii least concern declining stripe-necked mongoose H. vitticollis least concern stable carnivores tend to occur at low densities and in small Subsistence farming is the most widespread means of populations, which makes them particularly vulner- livelihood for rural families, who also exploit forests able to habitat loss, poaching, and local extinction to supplement incomes (Forest Resources Assess- (Woodruffe and Ginsberg 1998, Woodruffe 2001, ment 2000), increasing the potential for forest Cardillo et al. 2004). Indeed, 5 Carnivora species in degradation and loss of wildlife habitat. This trend Sri Lanka are listed as Endangered or Vulnerable is expected to accelerate in the aftermath of a 30- and 2 as Near Threatened (International Union for year civil war that ended in 2008, with subsequent Conservation of Nature [IUCN] 2011; Table 1), but resettlement of displaced families in the north and little empirical data exist on past or recent distribu- east. Sloth bears are still distributed widely in tions of these species. All forms of sport hunting remaining forests of Sri Lanka’s dry lowlands, where were banned in Sri Lanka in 1956, but conservation human densities and disturbance are low (Ratnayeke measures for carnivores have been mostly incidental et al. 2007a,b), and thus could be a potentially useful through in-situ conservation of habitat reserves to species to identify, prioritize, and manage areas for protect and facilitate the movement of Asian biodiversity protection (Fleishman et al. 2004). elephants (Elephas maximus). The presence of intact Areas occupied by sloth bears in the north and east carnivore assemblages is important because carni- of Sri Lanka contain the largest extents of unpro- vores play an important role in structuring ecological tected, contiguous forest on the island (Ratnayeke communities (Eisenberg 1989, Prugh et al. 2009). et al. 2006). Establishing protected areas for bears in However, biological knowledge of the habitat this region might also serve to protect habitat for requirements, current distribution, and population numerous carnivore species. Detection of sloth bears status of Sri Lanka’s carnivores is poor, the sloth with remote cameras, for example, may provide an bear being a notable exception (Santiapillai and effective tool to identify priority areas for carnivore Santiapillai 1990; Ratnayeke et al. 2006, 2007a,b). conservation. Therefore, we tested the research Our objective was to determine whether the sloth hypothesis that carnivore richness is associated with bear could serve as a conservation surrogate for the frequency of detection of sloth bears. We tested other mammalian carnivores in Sri Lanka. The sloth this hypothesis for all carnivores, carnivores at risk bear, found on the and Sri (IUCN categories Endangered, Vulnerable, and Near Lanka, occurs in a global region with high human Threatened), and common carnivores in Sri Lanka. densities, widespread poverty, and few resources for conservation planning (Garshelis et al. 1999). For example, human densities in Sri Lanka exceed 300 Study area people/km2 with almost 80% of the population living The national parks of Yala (126,781 ha) and in rural areas (United Nations Secretariat 2011). Wasgomuwa (39,385 ha) are in the dry zone

Ursus 23(2):206–217 (2012) 208 SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen lowlands of Sri Lanka and represent 2 different sampling species that are secretive and difficult to landscapes typical of sloth bear range (Panwar and capture (Karanth 1995, Wemmer et al. 1996, Karanth Wickramasinghe 1997, Pabla et al. 1998, Ratnayeke and Nichols 1998, Kelly et al. 2008). We overlaid the et al. 2007b). Both parks are composed of a mosaic national parks with a 1-km2 grid, and demarcated of vegetation types ranging from dense forest to regions that covered a range of habitat types and open grassland. Yala National Park is situated along could be accessed by road or by foot. We numbered the southeastern coastline (81u209E, 6u339N) and those grid cells and randomly selected 98 cells at Wasgomuwa National Park is in the central low- Wasgomuwa and 90 at Yala for field sampling. The lands (80u559E, 7u459N; Fig. 1). The climate is size of the grid cells was appropriate for our analysis classified as Tropical Dry Zone (Domro¨s 1974), because sloth bear home ranges we documented in with an extended dry period from June through mid Wasgomuwa National Park are some of the smallest October, although the southernmost region of Yala reported for any bear species (Ratnayeke et al. 2007a). receives much less rainfall (,50 cm annually) than We sampled during the dry season (Jun–Sep). We Wasgomuwa (.180 cm annually). Most precipita- sampled 49 sites at Wasgomuwa and 50 sites at Yala tion occurs from November through January. in 2002. In 2003, we sampled an additional 49 sites at At Wasgomuwa, elevations range from 60 to 200 m Wasgomuwa and 40 sites at Yala. We deployed 3–5 of undulating terrain separated by a long ridge (300– cameras/session; not all sites were sampled concur- 1,000 m) extending north–south. In contrast, Yala’s rently. In a few instances, we placed cameras in topography is mostly flat undulating plain (30–100 m) adjacent grid cells if recent illegal activity (e.g., interspersed with rock outcrops (Panwar and Wick- poaching, logging) at the original site was evident. ramasinghe 1997). Temperatures are uniformly high This resulted in 6 grid cells being sampled twice, throughout the year, with an annual mean of 32uC. although the site of camera placement and time that Both national parks supported high faunal diver- the cell was sampled differed. We recorded the sity including large herbivores such as Asian locations of camera sites with a global positioning elephants, buffalo (Bubalus bubalis), sambar (Rusa system (GPS) receiver. We placed 1 camera within a unicolor), and spotted deer (Axis axis). Vegetation at grid cell along an trail for 4 consecutive nights Wasgomua and Yala was broadly classified as dry and programmed all cameras using a 1-minute delay monsoonal forest (Jayasingham et al. 1992, Jaya- between pictures. We placed sensors to maximize singham and Vivekanantharajah 1994, Panwar and detection of sloth bears and small and medium-sized Wickramasinghe 1997) dominated by Drypetes carnivores by positioning the infrared beam at a sepiaria and Manilkara hexandra trees. Owing to its height of 12–15 cm above ground level across animal coastal proximity, Yala had greater habitat hetero- trails. Camera systems recorded an event when the geneity, including mangrove and sand dune commu- infrared beam was intercepted for .0.15 seconds. nities, and greater expanses of short grasslands We calculated carnivore species richness at each interspersed with thorn scrub and Salvadora persica camera site as the number of detected mammalian trees. Wasgomuwa National Park’s southern region carnivore species, excluding sloth bears. For each consisted of a patchwork of grassland, scrub, and camera site, we used activity data from the Trail- secondary forest, reflecting various stages of ecolog- master receiver to determine detection (presence of 1 ical succession resulting from human occupation or more bears) or nondetection of sloth bears for 4 before establishment of the national park in the early consecutive sampling occasions of 24 hr. 1980s. Legal entry into both national parks required Thus, the maximum number of detections of sloth permits and a Department of Wildlife Conservation bears at a camera site was 4. We used the number of (DWLC) guide. sloth bear detections at camera sites rather than sloth bear presence or absence because detections contained more information, not only measuring Methods presence of bear habitat but also the relative Field sampling importance of habitat. We used remote cameras (TrailmasterH, Lenexa, Kansas, USA; Kucera and Barrett 1993) triggered Analysis by active infrared sensors to detect sloth bears and We used Poisson regression (Jones et al. 2002) to other carnivores. Remote cameras are effective for determine if carnivore species richness (the dependent

Ursus 23(2):206–217 (2012) SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen 209

Fig. 1. Location of (a) Wasgomuwa National Park and (b) Yala National Park, Sri Lanka, displaying sampling sites to assess carnivore species distributions in 2002–03. Original habitat classes were derived from Pabla et al. (1998) and Panwar and Wickramasinghe (1997).

Ursus 23(2):206–217 (2012) 210 SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen variable) can be predicted by the frequency of sloth as a covariate, we implicitly tested the hypothesis that bear detections, study year, and study area. Poisson relationships between carnivore richness and sloth regression is particularly suited for data from bear detections varied according to the coarse-scale faunal or floral surveys, which often involve count habitat differences between the 2 study areas. We used data that tend to have skewed frequency distribu- Akaike’s information criterion adjusted for small tions. We conducted separate analyses for 3 sample size (AICc: Hurvich and Tsai 1989, Burnham different groupings of carnivores, excluding bears: and Anderson 2002) to examine the evidence for (1) all carnivore species reported from Sri Lanka competing models. We based inference on the entire (Table 1), (2) carnivores considered at risk (species set of models and used 95% confidence intervals of listed as Endangered, Vulnerable, or Near Threat- model-averaged parameter estimates to identify ened by IUCN; Table 1), and (3) carnivores that informative variables (Anderson 2008). We used were common. We used IUCN rankings to residual plots to examine if Poisson regression designate species to these categories, but there were assumptions were met. We used R (version 2.14.1; R a few exceptions for common carnivores based on Development Core Team 2007) statistical software to data from Sri Lanka. For common carnivores, we perform the Poisson regressions. considered 4 species whose conservation status was listed as Least Concern by the IUCN (2011): small Indian civet, common palm civet, golden jackal, Results and grey mongoose. These 4 species were common Data loss at several sites was caused by stolen outside protected areas and in areas occupied by cameras, camera malfunction, or cameras being humans (Phillips 1984, Duckworth et al. 2008a,b), dislodged by , thus yielding ,4 nights of and population trends were thought to be stable or data. Because it was logistically difficult to resample increasing, with the exception of the grey mongoose those sites, we excluded them from analysis. At whose population trends were ‘‘unknown’’ (IUCN Wasgomuwa, we obtained adequate camera data 2011). Three other species were listed by the IUCN from 36 sites in 2002 and 39 different sites in 2003, as Least Concern (jungle cat, stripe-necked mon- resulting in 300 camera nights (Fig. 1a). At Yala we goose, ruddy mongoose; Table 1) but we disagreed obtained usable data from 46 camera sites in 2002 with this status for Sri Lanka. The ruddy mongoose and 31 in 2003, totaling 308 camera nights (Fig. 1b). is not common outside protected areas in Sri Lanka We detected all of Sri Lanka’s 14 carnivora species (S. Ratnayeke, personal observation), nor near (Fig. 2), except for the endangered fishing cat human settlements (Shekhar [2003], cited in (Prionailurus viverrinus). We recorded the small Choudhury et al. 2008). The jungle cat and stripe- Indian civet and ruddy mongoose most frequently necked mongoose are uncommon within and at Yala (Fig. 2). At Wasgomuwa, we photographed outside protected areas in Sri Lanka (Phillips the small Indian civet and the sloth bear most 1984; S. Ratnayeke, personal observation). Finally, frequently (Fig. 2). We detected 12 carnivore species, populations of the jungle cat and ruddy mongoose excluding bears, at 128 of 152 remote camera sites, are perceived as declining (Choudhury et al. 2008, carnivores at risk (n 5 5) at 70 sites, and common IUCN 2011). Therefore, we did not include these 3 carnivores (n 5 4) at 89 sites (Fig. 3). More camera species in the category of common carnivores. sites were visited by carnivore species in Yala (87%) For each of the 3 categories of carnivore species than at Wasgomuwa (76%), but the proportion of richness, we fitted 8 a priori models using combina- sites visited at Yala (46%) and Wasgomuwa (47%) tions of the 3 independent variables (number of sloth was similar for carnivores at risk. We detected a bear detections, study area, study year). Our main mean overall carnivore species richness of 1.73 (SD question was whether carnivore species richness can 5 1.45, range 5 0–6) per site at Wasgomuwa and be predicted by the number of sloth bear detections. 1.79 (SD 5 1.22, range 5 0–5) at Yala. Species Because the relationship between carnivore species richness of carnivores at risk (x¯ 5 0.65, SD 5 0.83, richness and sloth bear detections may vary with range 5 0–3) and common carnivores (x¯ 5 0.67, SD study year and study area, we also examined the 5 0.64, range 5 0–2) at Wasgomuwa were similar. influence of year and study area on carnivore species At Yala, species richness of common carnivores (x¯ 5 richness and their 2-way interactions with the number 0.75, SD 5 0.71, range 5 0–2) was greater than that of sloth bear detections (Table 2). In using study area of carnivores at risk (x¯ 5 0.51, SD 5 0.60, range 5

Ursus 23(2):206–217 (2012) SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen 211

Table 2. Model selection results of Poisson regressions to predict 3 dependent variables of carnivore species richness (all carnivores, carnivores at risk, common carnivores) from the number of sloth bear detections (bear) at camera sites, year of study (year), study area (area), and their 2-way interactions (bear x year, bear x area), Yala and Wasgomuwa National Parks, Sri Lanka, 2002–03.

a b c Model AICc DAICc AICc weight K All carnivores bear, year 482.03 0 0.45 3 bear, year, bear x year 483.42 1.39 0.23 4 bear, area 484.74 2.71 0.12 3 bear 485.33 3.30 0.09 2 bear, year, area, bear x year, bear x area 485.87 3.84 0.07 6 bear, area, bear x area 486.47 4.44 0.05 4 year 492.30 10.27 0.00 2 area 502.09 20.06 0.00 2 Carnivores at risk bear 292.76 0 0.39 2 bear, year 293.56 0.80 0.26 3 bear, area 294.82 2.05 0.14 3 bear, year, bear x year 295.34 2.58 0.11 4 bear, year, area, bear x bear, bear x area 296.36 3.59 0.07 6 bear, area, bear x area 298.86 6.09 0.02 4 year 300.28 7.51 0.01 2 area 302.24 9.48 0.00 2 Common carnivores bear, year 313.69 0 0.31 3 area 315.00 1.31 0.16 2 bear 315.05 1.35 0.16 2 bear, area 315.13 1.44 0.15 3 bear, year, bear x year 315.41 1.71 0.13 4 bear, area, bear x area 317.19 3.50 0.05 4 bear, year, area, bear x year, bear x area 318.51 4.81 0.03 6 area 319.84 6.14 0.01 2 aAkaike’s information criterion adjusted for small n. b Difference in AICc compared with lowest AICc model. cNumber of model parameters.

0–2; t 5 2.33, 148 df, P 5 0.020). For Wasgomuwa, average, 1.46 carnivore species occurred at camera frequencies of grid cells with 0, 1, 2, 3, and 4 sloth sites if sloth bears were not present; with sloth bear bear detections were 39, 22, 9, 3, and 2, respectively. detections mean carnivore species richness increased For Yala, these frequencies were 58, 16, 3, 0, and 0, from 1.91 for 1 detection to 4.29 for 4 detections. respectively. Carnivore species richness was greater in 2002 (x¯ 5 At different levels of covariates, variances were 2.07) than 2003 (x¯ 5 1.40). Models for carnivores at similar to or slightly smaller than the means for all 3 risk also suggested that species richness was posi- categories of carnivore species richness. Thus, we tively associated with bear detections and was found no evidence of overdispersion. Residual plots greater during 2002 (x¯ 5 0.68) than 2003 (x¯ 5 of fitted values indicated constant variances. Con- 0.46; Table 2); unlike models based on all carnivores, sidering all carnivores, models that received the most the frequency of bear detections was the most support suggested carnivore species richness was important correlate of carnivore species richness positively associated with the number of bear (Table 3). An average of 0.46 at-risk carnivores were detections and varied by year (Table 2). Confidence observed at camera sites without sloth bear presence, intervals for model-averaged parameter estimates which increased to 0.64 for 1 sloth bear detection excluded zero for bear detections and study year but and to 1.79 for 4 detections. A variety of models not for the bear x year interaction, suggesting the for common carnivores indicated broad support relationship between carnivore richness and bear (Table 2). Greater species richness of common detections did not vary by year (Table 3). On carnivores was associated with the number of bear

Ursus 23(2):206–217 (2012) 212 SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen

Fig. 3. Number of camera sites at Yala National Park Fig. 2. Relative frequency expressed as a percent- (total n = 77) and Wasgomuwa National Park (total age of all camera captures (n = 713) of Carnivora at n = 75), Sri Lanka, 2002–03, where we recorded Yala and Wasgomuwa national parks, Sri Lanka, species of Carnivora, excluding bears. 2002–03, for stripe-necked mongoose (SNM), brown mongoose (BRM), ruddy mongoose (RM), grey mongoose (GM), sloth bear (SB), small Indian civet the relationship between carnivore species richness (SIC), common palm civet (CPC), golden palm civet and sloth bear detections was similar for both (GPC), golden jackal (JAC), jungle cat (JC), rusty landscapes. Our models indicated species richness spotted cat (RSC), leopard (LEO), and European of carnivores was greater at camera sites in 2002 otter (OTT). than in 2003, which corresponded with our expecta- tions. Precipitation during the second study year was detections, study year, study area, and bear x year greater, which may have resulted in more dispersed interaction. However, confidence intervals for model- movements and consequently less reliance on heavily averaged parameter estimates overlapped zero for all used animal trails that often form the shortest route parameters (Table 3), suggesting no association between sources of water for many species, includ- between species richness of common carnivores and ing carnivores. Although bear detections were fewer sloth bear detections. during the second year of the study, the lack of a distinct interaction effect with study year indicated the relationship with species richness of all carni- Discussion vores or carnivores at risk was similar both years. Our study provides the first empirical data on These results suggest that the sloth bear may serve as carnivore species occurrence and distribution in 2 a useful indicator of areas that are also occupied by protected areas in the low country dry zone of Sri other carnivores in Sri Lanka, particularly carni- Lanka. Carnivore guilds that include 10 or more vores of conservation concern. Furthermore, large species are rare (Schaller 1996). Our documentation carnivores such as sloth bears can garner substantial of 13 of the island’s 14 carnivore species indicates the conservation attention, and because they leave global significance of Yala and Wasgomuwa Na- conspicuous sign or are easily observed, their tional Parks. Our measures of carnivore richness and populations are more effectively monitored than frequency of sloth bear detections varied substan- small species. tially among camera sites, thus providing a suitable We considered 3 caveats interpreting the results of data set to test our research hypothesis. Sloth bears our analyses: (1) factors that may bias species were detected at a lower proportion of sites at Yala richness estimates, (2) application of the observed than at Wasgomuwa, but species richness of all relationship between sloth bear detections and carnivores, and at-risk carnivores in particular, was carnivore richness outside protected areas of Sri positively associated with the frequency of bear Lanka, and (3) measuring species richness may detections regardless of study area. Because we used diminish the important role of rare species to the study area covariate as a coarse-scale measure of conservation. Behavioral differences among species sloth bear habitat conditions, this finding suggests or between sexes of the same species likely influenced

Ursus 23(2):206–217 (2012) SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen 213

Table 3. Model-averaged parameter estimates and their standard errors and confidence intervals for Poisson regression parameters to determine relationships of 3 dependent variables of carnivore species richness (all carnivores, carnivores at risk, common carnivores) with number of sloth bear detections (bear) at camera sites, year of study (year), study area (area), and their interactions (bear x year, bear x area), Yala and Wasgomuwa National Parks, Sri Lanka, 2002–03. Model-averaged Parameter parameter estimate Standard error 95% lower CI 95% upper CI All carnivores intercept 0.29 a 0.14 0.01 0.58 bear 0.27 a 0.11 0.06 0.48 year 0.24 a 0.12 0.00 0.49 area 20.05 0.05 20.16 0.05 bear x year 20.04 0.06 20.15 0.07 bear x area 0.01 0.02 20.03 0.05 Carnivores at risk intercept 20.84 a 0.18 21.20 20.48 bear 0.34 a 0.14 0.06 0.61 year 0.12 0.12 20.12 0.35 area 0.00 0.06 20.11 0.11 bear x year 20.02 0.04 20.09 0.06 bear x area 0.02 0.03 20.04 0.08 Common carnivores intercept 20.57 a 0.21 20.98 20.17 bear 0.20 0.12 20.04 0.45 year 0.26 0.17 20.06 0.59 area 20.07 0.08 20.22 0.08 bear x year 20.03 0.05 20.12 0.07 bear x area 0.00 0.02 20.04 0.05 aConfidence interval excludes zero capture probabilities. We placed cameras along bear detections and carnivore species richness may animal trails and across dry or partially dry stream not be the same outside protected areas where beds. Species such as the common palm civet and anthropogenic influences are greater. golden palm civet (Paradoxurus zeylonensis) are Thirdly, although species that are rare (and thus of semi-arboreal (Phillips 1984) and are possibly greatest conservation concern) may not be well underrepresented in our camera data. Several represented by measures of species richness, our species, including the grey mongoose, jungle cat, data indicate this may not be a major concern. We and golden jackal, frequent open habitats and may used the IUCN Red List to determine the conserva- have little need for trails. Season and camera tion status of carnivores. We recorded almost all placement may have reduced the probability of species that were considered at risk at Yala and obtaining photos of fishing cats and . Each of Wasgomuwa and many with greater frequency (e.g., these factors would result in potential underestima- the sloth bear, leopard, golden palm civet, rusty tion of carnivore richness, but this bias is not likely spotted cat, brown mongoose; Fig. 2) than species of to vary with the number of sloth bear detections. Least Concern (e.g., grey mongoose, stripe-necked Therefore, the relationships we observed between mongoose, golden jackal; Fig. 2). We recognize, sloth bear detections and carnivore species richness however, this may not be the case for the otter (1 would not be affected. detection) and the endangered fishing cat (no The second caution is that carnivore species detections). Although poor detection likely was a distributions at Yala and Wasgomuwa may not function of their rarity, narrow niche requirements represent the status of carnivore species in other may have played a role as well. Thus, it is important parts of the low country dry zone, particularly to document which species are represented when unprotected areas. Richness and population sizes of using surrogate species for conservation. at-risk carnivores are likely larger in protected areas. We also note that carnivores rarely encountered in Thus, the relationship we observed between sloth protected areas might not necessarily be rare outside

Ursus 23(2):206–217 (2012) 214 SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen national parks. Camera records and direct observations ry habitats are associated with shifting agriculture of grey mongoose, stripe-necked mongoose, and (Fernando et al. 2006), which is unlikely to be jungle cat were infrequent during this study. It is optimal for threatened carnivore populations. Home possible that these species do better in areas ranges of sloth bears are small and despite its small occupied by humans, or experience ecological size, Wasgomuwa National Park supports an abun- release where top predators or competitors they dance of sloth bears and other carnivores of would encounter in protected areas are absent. conservation concern (Ratnayeke et al. 2007a). Even These 3 species are listed as Least Concern (IUCN establishing small protected areas (e.g., similar in 2011) and have a wide geographic distribution, but size to Wasgomuwa National Park) within sloth bear their apparent rarity in the 2 national parks we range in northern and eastern Sri Lanka would likely surveyed suggests their distribution and status in Sri protect habitat for most carnivores of conservation Lanka need to be assessed in more detail. concern. Our objective was to assess the potential for sloth Bears are generally viewed as a charismatic, bears to serve as a focal or surrogate species that flagship species whose presence contributes positive- would provide the incentive to protect other ly to conservation efforts (Simberloff 1998). That carnivores in Sri Lanka. Simberloff (1999) specifi- characteristic is also important if the sloth bear is cally addressed the role of bears as potential to serve as a conservation surrogate or umbrella umbrella species whose protection would also species. However, sloth bears frequently attack safeguard species that coexist with bears. The humans, often resulting in serious injuries or even umbrella species concept represents a hypothesis death (Santiapillai and Santiapillai 1990, Rajpurohit that has not been thoroughly tested (Simberloff and Krausman 2000, Bargali et al. 2005, Ratnayeke 1998, 1999) and thus has received substantial et al. 2006). Consequently, people who live in sloth criticism. For example, there is debate whether the bear habitat greatly fear bears (Ratnayeke et al. requirements of a single species can effectively meet 2007b), posing considerable challenges to sloth bear those of many co-occurring species (Berger 1997, conservation. Effective umbrellas, however, need Linnell et al. 2000, Caro 2003), particularly species not exclude species that are not good conservation from different orders and classes (Roberge and flagships. Species associated with endangered bio- Angelstam 2004, Seddon and Leech 2008). Fleish- logical communities can be effective umbrellas if their man et al. (2004) suggested that a species could serve protection automatically extends protection to other as an effective umbrella for conservation if its species (Ozaki et al. 2006). Thus, if the sloth bear were distribution overlapped with a relatively large to be used as a conservation surrogate for the dry percentage of its close taxonomic relatives or species lowlands of Sri Lanka, where most of the island’s of conservation interest (see Seddon and Leech threatened carnivores remain, the significance of 2008). Our analysis supports the notion that sites carnivores in ecological communities and the sloth where sloth bears occur tend to have greater richness bear’s role as an indicator of healthy carnivore of carnivores, particularly those of greatest concern communities would need to be emphasized. Such an for conservation. Sloth bear detections were associ- approach may be effective within the cultural ated with 4 threatened carnivore species (leopard, framework of rural Sri Lanka where peas- rusty spotted cat, brown mongoose, and golden ant traditions and livelihoods are closely linked with palm civet). Large portions of sloth bear range in the healthy, diverse ecosystems. The 1974 Chipko move- north and east of the island remain unprotected and ment and 1995 Jungle Jeevan Bachao Yatra (Save the currently experience heavy exploitation with the Forests, Save our Lives) march are examples in which recent end of a 30-year civil war and the establish- rural communities have successfully championed the ment and resettlement of human communities. conservation of natural forests (Rangarajan 1996). About 800,000 ha in Sri Lanka exist in the form of Finally, beyond the relationships we observed sanctuaries, nature reserves, and national parks (Sri between sloth bears and other carnivores, we Lanka Department of Wildlife Conservation, Co- hypothesize that sloth bears may play a significant lombo, unpublished data), the majority of which ecological role in tropical dry forests. Most regions function as protected areas for Sri Lanka’s conser- inhabited by sloth bears experience long, dry periods vation flagship species, the Asian elephant. Although where waterholes and streambeds dry out. Sloth elephants have large area requirements, their prima- bears dig deep holes in dry streambeds, some of which

Ursus 23(2):206–217 (2012) SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen 215 exceed 2–3 m in depth, to reach water (Phillips 1984; theoretic approach. Springer-Verlag, New York, New S. Ratnayeke, personal observation). The powerful York, USA. forelimbs and long claws, adapted for breaking into CARDILLO, M., A. PURVIS,W.SECHREST, J.L. GITTLEMAN,J. termite mounds, undoubtedly help sloth bears with BIELBY, AND G.M. MACe. 2004. Human population this ability. These watering sites are also used by a density and extinction risk in the world’s carnivores. PloS Biology 2:909–914. multitude of vertebrates and invertebrates. Six camera CARDOSO, P., I. SILVA, N.G. dE OLIVEIRA, AND A.R.M. sites were near such waterholes, resulting in 21 photos SERRANO. 2004. Higher taxa surrogates of spider of sloth bears, 35 photos of threatened carnivores, (Aranea) diversity and their efficiency in conservation. and 86 photos of other mammals, birds, and reptiles. Biological Conservation 126:453–459. This potential keystone function should be explored CARO, T.M., AND G. O’DOHERTY. 1999. On the use of to elucidate the role of this unique species of bear in surrogate species in conservation biology. Conservation tropical dry forests. Biology 13:805–814. ———. 2003. Umbrella species: Critique and lessons from East Africa. Animal Conservation 6:171–181. Acknowledgments CHOUDHURY, A., C. WOZENCRAFT,D.MUDDAPA, AND P. Fieldwork was supported by Grant 0107293 from YONZON. 2008. Herpestes smithii. In IUCN Red List of Threatened Species. International Union for Conser- the US National Science Foundation. D. Branden- vation of Nature, Cambridge, UK and Gland, Switzer- burg designed and constructed metal frames to land, Version 2011.2. http://www.iucnredlist.org/apps/ protect camera units from bear and elephant redlist/details/41617/0, accessed 14 June 2012. damage. We are indebted to Park Wardens B.V.R. DOMRO¨ S, M. 1974. The agroclimate of Ceylon. Franz Jayaratne, W.S. Weragama, and L. Pieris for Steiner Verlag GMBH, Wiesbaden, Germany. assisting with logistics, and to wildlife officers G.G. DUCKWORTH, J.W., R.J. TIMMINS, AND D. MUDDAPA. Karunaratne and K. Chamindha for accompanying 2008a. Viverricula indica. In IUCN Red List of us on remote camera surveys. Permits to conduct Threatened Species. International Union for Conser- camera sampling at Wasgomuwa National Park vation of Nature, Cambridge, UK and Gland, Switzer- were made to K. Padmalal. A. Bandara and R. Pieris land, Version 2012.1. http://www.iucnredlist.org/apps/ assisted with fieldwork. C. Thatcher and L. Thomp- redlist/details/41710/0, accessed 21 June 2012. son assisted with GIS analyses. J.D. Clark helped ———, P. WIDMANN,C.CUSTODIO,J.C.GONZALEZ,A. JENNINGS, AND G. VERON. 2008b. Paradoxurus her- with review of model results. R. Pethiyagoda and R. maphroditus. In IUCN Red List of Threatened Species. Nadaraja generously provided advice, assistance, Version 2012.1. International Union for Conservation and resources during the project. We thank R. of Nature, Cambridge, UK and Gland, Switzerland, Harris, J. McDonald, D. Garshelis, C. Parker, and http://www.iucnredlist.org/apps/redlist/details/41693/0, an anonymous reviewer for valuable comments on accessed 21 June 2012. previous drafts of this manuscript. Any use of trade, EISENBERG, J.F. 1989. An introduction to the Carnivora. product, or firm names is for descriptive purposes Pages 1–9 in J.L. Gittleman, editor. Carnivore behav- only and does not imply endorsement by the US ior, ecology and evolution. Volume 1. Cornell Univer- government or the University of Tennessee. sity Press, Ithaca, New York, USA. FERNANDO, P., E.D. WIKRAMANAYAKE,D.WEERAKOON, H.K. JANAKA,M.GUNAWARDENA,L.K.A.JAYASINGHE, H.G. NISHANTHA, AND J. PASTORINI. 2006. The future of Asian Literature cited elephant conservation: Setting sights beyond protected ANDERSON, D.R. 2008. Model based inference in the life area boundaries. Pages 252–260 in J.A. McNeely, T.M. sciences. Springer, New York, New York, USA. McCarthy, A. Smith, L. Olsvig-Whittaker, and E.D. BARGALI, H.S., N. AKHTAR, AND N.P.S. CHAUHAN. 2005. Wikramanayake, editors. Conservation Biology in Characteristics of sloth bear attacks and human Asia. Paper 17. Society for Conservation Biology casualties in North Bilaspur Forest Division, Chhattis- Asia Section, Resources Himalaya Foundation, Kath- garh, India. Ursus 16:263–267. mandu, Nepal. BERGER, J. 1997. Population constraints associated with FLEISHMAN, E., J.B. DUNHAM, D.D. MURPHY, AND P.F. the use of black rhinos as an umbrella species for desert BRUSSARD. 2004. Explanation, prediction, and mainte- herbivores. Conservation Biology 11:69–78. nance of native species richness and composition. BURNHAM,K.P.,AND D.R. ANDERSON.2002.Modelselection Pages 232–260 in J.C. Chambers and J.R. Miller, and multimodel inference: A practical information editors. Great Basin riparian ecosystems: Ecology,

Ursus 23(2):206–217 (2012) 216 SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen

management, and restoration. Island Press, Washing- forests: Large carnivores as flagships, umbrellas, ton, DC, USA. indicators, or keystones? Biodiversity and Conserva- FOREST RESOURCES ASSESSMENT. 2000. Forest resources of tion 9:857–868. Sri Lanka: Country report. Forest Resources Assess- NOSS, R.F. 1990. Indicators for monitoring biodiversity: ment Program, Working Paper 17, Rome, Italy. A hierarchical approach. Conservation Biology 4: FRANKEL, O.H., AND M.E. SOULE´ . 1981. Conservation and 355–364. evolution. Cambridge University Press, Cambridge, OZAKI, K., M. ISONO,T.KAWAHARA,S.IIDA,T.KUDO, AND UK. K. FUKUYAMA. 2006. A mechanistic approach to GARSHELIS, D.L., A.R. JOSHI, J.L.D. SMITH, AND C.D. evaluation of umbrella species as conservation surro- RICE. 1999. Sloth Bear Conservation Action Plan gates. Conservation Biology 20:1507–1515. (Melursus ursinus). Pages 225–240 in C. Servheen, S. PABLA, H.S., V.B. MATHUR, AND W.R.M.S. WICKRAMA- Herrero, and B. Peyton, editors. Bears: Status survey SINGHE. 1998. Management plan. Wasgomuwa National and conservation action plan. IUCN/SSC Bear and Park and Riverine Nature Reserve. Project UNO/SRL/ Polar Bear Specialist Groups, International Union for 001/GEF-SRL/92/G3. Department of Wildlife Conser- Conservation of Nature, Cambridge, UK and Gland, vation, Sri Lanka. Switzerland. PANWAR, H.S., AND W.R.M.S. WICKRAMASINGHE. 1997. HURVICH, C.M., AND C.L. TSAI. 1989. Regression and time Management plan. Yala Protected Area Complex. series model selection in small samples. Biometrika Project UNO/SRL/001/GEF-SRL/92/G3. Department 76:297–307. of Wildlife Conservation, Sri Lanka. INTERNATIONAL UNION FOR CONSERVATION OF NATURE. PHILLIPS, W.W.A. 1984. The sloth bear. Pages 290–296 in 2011. IUCN red list categories. International Union Wildlife and Nature Protection Society of Sri Lanka, for Conservation of Nature and Natural Resources, editor. Manual of mammals of Sri Lanka, Colombo, IUCN, Gland, Switzerland. Sri Lanka. JAYASINGHAM, T., S. BALASUBRAMANIAM, AND S. VIVEKA- PRUGH, L.R., C.J. STONER, C.W. EPPS, W.T. BEAN, W.J. NANTHARAJAH. 1992. Vegetation survey of Wasgomuwa RIPPLE, A.S. LALIBERTE, AND J.S. BRASHARES. 2009. The National Park: Reconnaissance. Vegetatio 101:171–181. rise of the mesopredator. BioScience 59:779–791. ———, AND S. VIVEKANANTHARAJAH. 1994. Vegetation survey RDEVELOPMENT CORE TEAM. 2007. R: A language and of Wasgomuwa National Park, Sri Lanka: Analysis of environment for statistical computing. R Foundation the Wasgomuwa Oya forest. Vegetatio 113:1–8. for Statistical Computing, Vienna, Austria, http://www. JONES, M.T., G.J. NIEMI, J.M. HANOWSKI, AND R.R. R-project.org, accessed 22 December 2011. REGAL. 2002. Poisson regression: A better approach RAJPUROHIT, K.S., AND P.R. KRAUSMAN. 2000. Human– to modeling abundance data? Pages 411–418 in J.M. sloth bear conflicts in Madhya Pradesh, India. Wildlife Scott, P.J. Heglund, M.L. Morrison, J.B. Haufler, Society Bulletin 28:393–399. M.G. Raphael, W.A. Wall, and F.B. Samson, editors. RANGARAJAN, M. 1996. The politics of ecology: The debate Predicting species occurrences: Issues of accuracy and on wildlife and people in India, 1970–95. Economic and scale. Island Press, Washington DC, USA. Political Weekly 31:2391–2409. KARANTH, K.U. 1995. Estimating tiger Panthera tigris RATNAYEKE, S., S. WIJEYAMOHAN, AND C. SANTIAPILLAI. populations from camera-trap data using capture– 2006. The status of sloth bears in Sri Lanka. recapture models. Biological Conservation 71:333–338. Pages 35–40 in T. Oi, T. Mano, K. Yamazaki, T. Aoi, ———, AND J.D. NICHOLS. 1998. Estimation of tiger M. Carr, M. Durnin, C.B. Imaki, A. Takayanagi, and densities in India using photographic captures and T. Tsubota, editors. Understanding Asian bears to recaptures. Ecology 79:2852–2862. secure their future. Japan Bear Network, Gifu, Japan. KELLY, M.J., A.J. NOSS, M.S. DI BITETTI,L.MAFFEI, R.L. ———, F.T. VAN MANEN, AND U.K.G.K. PADMALAL. ARISPE,A.PAVIOLO, C.D. DE ANGELO, AND Y.E. DI 2007a. Home ranges and habitat use of the sloth bear BLANCO. 2008. Estimating densities from camera (Melursus ursinus inornatus) at Wasgomuwa National trapping across three study sites: Bolivia, Argentina, Park, Sri Lanka. Wildlife Biology 13:272–284. and Belize. Journal of Mammalogy 89:408–418. ———, ———, R. PIERIS, AND V.S.J. PRAGASH. 2007b. KUCERA, T.E., AND R.H. BARRETT. 1993. The TrailmasterH Landscape characteristics of sloth bear range in Sri camera system for detecting wildlife. Wildlife Society Lanka. Ursus 18:189–202. Bulletin 21:505–508. ROBERGE, J., AND P. ANGELSTAM. 2004. Usefulness of the LAMBECK, R.J. 1997. Focal species: A multi-species umbrella species concept as a conservation tool. Con- umbrella for nature conservation. Conservation Biolo- servation Biology 18:76–85. gy 11:849–856. SANTIAPILLAI, A., AND C. SANTIAPILLAI. 1990. Status, LINNELL, J.D.C., J.E. SWENSON, AND R. ANDERSEN. 2000. distribution and conservation of the sloth bear (Me- Conservation of biodiversity in Scandinavian boreal lursus ursinus) in Sri Lanka. Tiger Paper 1:13–15.

Ursus 23(2):206–217 (2012) SLOTH BEARS AS CONSERVATION SURROGATES N Ratnayeke and van Manen 217

SCHALLER, G. 1996. Introduction: Carnivores and conser- Secretariat, New York, New York, USA, ,http://esa. vation biology. Pages 1–10 in J.L. Gittleman, editor. un.org/unpd/wup/index.htm, accessed 17 Sep 2011. Carnivore behavior, ecology and evolution. Volume5 WEMMER, C., T.H. KUNZ,G.LUNDIE-JENKINS, AND W.J. 2. Cornell University Press, Ithaca, New York, USA. MCSHEA. 1996. Mammalian sign. Pages 157–176 in SEDDON, P.J., AND T. LEECH. 2008. Conservation short cut, D.E. Wilson, F.R. Cole, J.D. Nichols, R. Rudran, and or long and winding road? A critique of the umbrella M.S. Foster, editors. Measuring and monitoring species criteria. Oryx 42:240–245. biological diversity. Standard methods for mammals. SIMBERLOFF, D. 1998. Flagships, umbrellas, and keystones: Smithsonian Institution Press, Washington DC, USA. Is single species management passe´ in the landscape WOODRUFFE, R., AND J.R. GINSBERG. 1998. Edge effects era? Biological Conservation 83:247–257. and the extinction of populations inside protected ———. 1999. Biodiversity and bears–a paradigm shift. areas. Science 280:2126–2128. Ursus 11:21–28. ———. 2001. Strategies for carnivore conservation: THORNE, J.H., D. CAMERON, AND J.F. QUINN. 2006. A Lessons from contemporary extinctions. Pages 61–92 conservation design for the central coast of California in J.L. Gittleman, S. Funk, D. Macdonald, and R.K. and the evaluation of mountain lion as an umbrella Wayne, editors. Carnivore Conservation. Cambridge species. Natural Areas Journal 26:137–148. University Press, London, UK. UNITED NATIONS SECRETARIAT. 2011. World population prospects: The 2010 revision and world urbanization Received: 28 November 2011 prospects. Population Division of the Department of Accepted: 6 September 2012 Economic and Social Affairs of the United Nations Associate Editor: J. McDonald

Ursus 23(2):206–217 (2012)