Subscribe Now Subscribe Today
Research Article

MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir

Muzaffar A. Kichloo and Neeraj Sharma
Facebook Twitter Digg Reddit Linkedin StumbleUpon E-mail

Background and Objective: Asiatic black bear, one of the flagship species in Kishtwar high altitude National Park is least surveyed for its current distribution, thus limiting the conservation planning and management actions at landscape level. To have first-hand information about the possible distribution and area occupancy, an attempt was made to map its spatial distribution to earmark the areas demanding specific management attention. Materials and Methods: Maximum Entropy (MaxEnt) algorithm was used to delineate the areas of maximum probability of black bear distribution based on selective environmental correlates. Fifty-one presence points and six environmental variables were used to generate probability distribution maps in MaxEnt. The analysis gave a deep insight into the degree of influence of environmental variables on its distribution. Results: The average value of 0.92 obtained for the area under receiver operating characteristic curve (AUC) substantiated that the model performed very well. The study indicated that black bears mostly preferred forested landscapes (2200-3300 m) in Kibber, Nanth, Kiyar and Renai watersheds. The land use/land cover and elevation emerged as the most critical variables to the model building with individual contribution of 48.1 and 45.2%, respectively. Conclusion: The results suggested that dense temperate forests, the ideal habitats of black bear mostly along the perennial streams and rivers should be given due attention in long-term conservation and planning. The area outside the park should also be included to have a better understanding of habitat requirements of the animal and its likely dispersal routes.

Related Articles in ASCI
Search in Google Scholar
View Citation
Report Citation

  How to cite this article:

Muzaffar A. Kichloo and Neeraj Sharma, 2021. MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir. Asian Journal of Animal Sciences, 15: 19-26.

DOI: 10.3923/ajas.2021.19.26

Received: March 22, 2020; Accepted: June 22, 2020; Published: November 14, 2020

Copyright: © 2021. This is an open access article distributed under the terms of the creative commons attribution License, which permits unrestricted use, distribution and reproduction in any medium, provided the original author and source are credited.


For the effective management of large mammals, knowledge about their distribution is very important so that managers and conservationists can delineate and prioritize areas for the implementation of conservation policies and practices1. In the face of increasing anthropogenic pressures, determining distributions is crucial for long term survival of the species2. Long term data on space-use and abundance are necessary, but conventional approaches like transect surveys, camera traps and non-invasive genetic sampling are cost and labor-intensive as well as time consuming3-7. Consequently, the conservation biologists are increasingly relying on predictive models for estimating patterns of species distribution and thereby defining conservation strategies8-10.

Whereas a large number of statistical methods are in practice to construct species distribution models, MaxEnt was used in the current study. MaxEnt is a machine learning program that uses a statistical technique called maximum entropy which makes predictions from incomplete information11-12. MaxEnt models the geographic distribution of species using occurrence records and different environmental variables on the principle of a uniform probability distribution. MaxEnt models have been considered ideal for predictive performance of Species Distribution Models.

The Asiatic black bear Ursus thibetanus, (here onwards as black bear) sometimes also referred to as Himalayan black bear is a vulnerable species as per the International Union of Nature and Natural Resources (IUCN) red list13 with bodyweight average of 135 kg (60-200 kg) for males and 40-125 kg for females. It has been reported to be continuously distributed through southern and eastern Asia from westward through Pakistan and Afghanistan to Baluchistan province of Iran; east to Indo-China, Korea and Japan and an isolated population in Taiwan14. Schaller15 reported a widespread distribution for black bears from Korea and Russia to Indo-China and from the forests of the Himalayas below an altitude of 3,750 m west as far as Afghanistan and Iran. The Himalayan region and hills of northeast India probably holds one of the largest populations of black bear in Asia14,16, with best-known population being reported from the Jammu and Kashmir state16,17.

Kishtwar high altitude National Park, (here onwards as Kishtwar National Park) was notified in the year 1986 in Kishtwar district of the erstwhile state of Jammu and Kashmir. Located between 33°20 and 34°04 N and 75°40 and 76°10 E, the park initially covered an area of 1700 Km2 (1250 Km2 under permanent snow/glacier and 450 Km2 under forest cover) in an elevational range of 1800 m to 6000 m. Subsequently, in an extensive demarcation exercise, all habitations, hamlets and villages from the park were excluded and park boundaries were redefined with the final area coverage of 2193 Km2. The National Park provides representative and distinct habitats for large carnivores like snow leopard, common leopard, black bear, brown bear and Himalayan fox. The current status and extent of distribution of large mammals from this region is largely unknown. The first animal survey in the protected area was however undertaken by Kichloo18 followed by Parsa19 and Baba20 while, Hilaluddin and Naqash21,22 conducted preliminary census and population/density estimates of large mammals. No more studies / exhaustive census surveys have been reported since then. The spatial extent, past distribution and current occupancy of mammals, including the black bear has not so far been recorded for the park. This study thus aimed at understanding the current distribution of black bears, identify its potential habitat and delineate such areas for further conservation and management.


Study area: The Kishtwar National Park lies at a distance of around 70 km north of Kishtwar town on the north-eastern side of the Union Territory of Jammu and Kashmir bordering with Kargil district of Union Territory of Ladakh. It lies between 33°43 and 33°99 N and 75°65 and 76°28 E with area coverage of 2193 km2 and elevational range of 2224 m to 6293 m (Fig. 1). The study area is characterized by vivid landscapes comprising vast and narrow valleys, rugged mountains, broken cliffs, snow-clad peaks and a vast drainage network. The wide range of habitats viz., temperate, sub-alpine, alpine and rocky outcrops support a good population of carnivores and their prey base. The Park is well drained by four major streams/rivulets i.e., Kibber, Nanth, Kiyar and Renai. Renai, the largest watershed bounds Kashmir in north-west and Ladakh in north, while Kiyar share the northern boundary with Ladakh, and, Kibber and Nanth with Paddar division of Kishtwar district in the east. Renai stream drains into Marusudar River at Marwah while Kiyar, Kibber and Nanth streams join Marusudar at Dachan. Marusudar joins river Chandrabhaga at Bhandarkot, Kishtwar and flow as river Chenab beyond this point draining Akhnoor further south towards Pakistan.

Species occurrence data: The trails within forests, streams, ridgelines, narrow valleys in different landscapes were walked for signs of black bear in the elevational range of 2200 m to 4800 m during the period from March to October, 2018.

Image for - MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir
Fig. 1:False color composite of Kishtwar National Park
  Source: Department of Wildlife Protection, J and K

Table 1: Pre-selected environmental variables for the Modeling process
Image for - MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir

The direct sightings or presence signs (footprints, scats, scrapes, predation on wild and domesticated animals, etc.) were observed using binoculars and digital cameras and the geo-coordinates were recorded using Global Positioning System (GPS). Out of total 67 presence points recorded during the surveys, only 51 were used in modeling as others lie outside the park boundaries.

Environmental variables: Different environmental variables were considered as potential predictors affecting the distribution of black bears in Kishtwar National Park (Table 1). These were based on the knowledge and understanding of species in the park as well as from the existing literature. The selected variables included terrain (elevation, slope, aspect and distance to water bodies/rivers), vegetation (Normalized Difference Vegetation Index, NDVI) and land use/land cover. Distance to the roads and settlements were not used in this case as our investigations were restricted to the park boundaries and not beyond.

Modeling: MaxEnt (Maximum Entropy Modeling) version 3.4.111 was used to generate the potential distribution maps for black bear in Kishtwar National Park. The environmental layers, along with occurrence layer were loaded into MaxEnt as the model inputs. The presence data was divided into two sets, 75% for training and 25% for testing the model.

The model was run in ten replications using sub-sampling method with 5000 iterations and the percent contribution of each environmental variable was estimated for each iteration in MaxEnt algorithm. The Jackknife test that evaluates the importance of each environmental variable by subtracting the gain value of each variable in isolation to the total gain value with all variables was also performed. Environmental variables with highest gain value convey most meaningful information to the model.

Evaluation of accuracy of any distribution model is essential to validate models and to understand the model performance. Model was evaluated by calculating the area under the Receiver Operator Characteristic (ROC) Curve (AUC)11,23. As ROC (AUC) analysis is independent of both threshold setting and prevalence, it is one of the highly effective methods to assess the performance of ordinal score (presence-only) species distribution models24-26. AUC is derived by plotting sensitivity (the likelihood that a model correctly classifies a presence) versus 1-specificity (the probability that a model correctly classifies an absence). AUC has a specified range value of 0.5 to 1, where AUC >0.9 reveals the excellent, 0.7-0.9 the moderately useful and AUC <0.7 poor model performance, respectively.


The AUC value obtained for ten replicate runs was 0.922 with a standard deviation of 0.025 as reflected in Receiver Operating Characteristic (ROC) curve (Fig. 2).

Image for - MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir
Fig. 2:ROC curves for models of black bear

The higher AUC values substantiate high accuracy of model generated for predicting the distribution of black bear in Kishtwar National Park. The warmer colours (P = 1) in continuous coloured distribution map generated for the species indicate areas with high probability of the species while blue (P = 0) suggested the least likelihood (Fig. 3).

Analysis of variable contributions: While comparing the relative contributions of several environmental variables in the model, land use/land cover with individual contribution of 48% emerged as a significant factor influencing the spatial distribution of black bear in Kishtwar National Park.

It was followed by elevation (45.2%) while others viz., slope, aspect, NDVI and distance to the water body carried minuscule weightage (Fig. 4).

Likewise in Jackknife test, elevation (elev) appeared to be the prime environmental variable influencing the distribution of black bear in the park with highest gain value (Fig. 5) followed by lulc i.e., land use/landcover. The test suggested that elevation had highest gain when used in isolation, which therefore appears to have the most useful information by itself.

Habitat suitability: The probability distribution map generated using MaxEnt was imported into Q GIS27 and re-classified based on the 10-percentile logistic threshold (of 0.064) to a binary suitable/unsuitable habitat map for black bear in Kishtwar National Park (Fig. 6).

Image for - MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir
Fig. 3:Species distribution map of Black bear

Image for - MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir
Fig. 4:Contribution of environmental variables in the model building

Image for - MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir
Fig. 5:
Jackknife test of variable importance for Black bear in Kishtwar National Park

Image for - MaxEnt Modeling of Distribution and Habitat Preferences of Asiatic Black Bear in Kishtwar High Altitude National Park, Jammu and Kashmir
Fig. 6:
Binary habitat suitability map for Black bear in Kishtwar National Park

The suitable category included areas with environmental conditions conducive for the species to survive and vice versa. The binary habitat suitability map can serve as a valuable ingredient for setting up the conservation and management practices for black bears.


The current study was first of its kind in Kishtwar National Park where the data on mammals in general and black bear in particular is scanty22. The outputs produced by MaxEnt generally reflected the established facets of bear ecology and biology associated with habitat suitability. The results indicated that distribution of black bears is restricted to lower elevations between 2200-3300 m. Garshelis and Steinmetz13 reported black bear habitat in both coniferous and broad-leaved forested landscapes at an altitude of up to 4500 m, whereas, Bista and Aryal28 found the distribution of black bears confined to elevation range 1400-3500 m in Annapurna Conservation Area in Nepal. The elevation is one of the most influential environmental variables which act as a limiting factor for the distribution of terrestrial carnivores by indirectly affecting the distribution through temperature29,30. The Jackknife test also suggested that elevation had the highest gain value, which, therefore, appears to have the most useful information by itself which is the case for black bear as well as other large mammals29,31.

The land use/landcover emerged as a significant factor affecting the distribution of black bears in Kishtwar National Park, with maximum probability in temperate broad-leaved and coniferous forests along the major river courses of Kibber, Nanth, Kiyar and Renai watersheds. It signifies a direct relationship between the vegetation, water bodies and species distribution. Forest cover has been known to play an important role in habitat selection by black bears wherein they occupy a variety of forested areas from temperate evergreen conifers to subtropical broadleaved deciduous forests14,28,32,33. Liu et al.34 found that the forest cover was the only significant environmental variable that positively affected black bear distribution in south-western China. Black bears are also reported to use multiple land covers as opportunistic omnivores33,35. Another important variable which affected the black bear habitat selection in KNP was the distance to water bodies (rivers). Water indirectly affected the black bears in two ways; first their survival and secondly, the areas close to rivers were mostly covered by forests which primarily affected their distributions. MaxEnt modeling has proven to be very effective at determining species distributions for a variety of species and localities, especially when the sample size is very small36-38 and a suggested Species Distribution Modeling tool for black bear.


The present study investigated the spatial distribution of black bear in Kishtwar National Park. The probability distribution maps were prepared by using 51 presence points with six environmental variables in MaxEnt. The model indicated that black bear prefers the temperate broad-leaved and coniferous forests along with an elevation range of 2200-3300 m in all four watersheds. Study points out the importance of forest conservation in order to conserve black bears in KNP.


Land use/land cover dynamics and other environmental variables help to delineate the potential and preferred habitats of a particular species on broader spatial and temporal scales. Current study describes the drivers of spatial distribution and habitat preferences of black bear, one of the key mammal species of Kishtwar National Park that many researchers were not able to explore. The maps so generated will help the park managers to prioritize their habitat conservation efforts in the park.


Authors are grateful to the Department of Wildlife Protection, Govt. of Jammu and Kashmir for necessary permits and to the Wildlife Warden, Chenab Circle for providing necessary support in field surveys. The authors thankfully acknowledge the Rufford Foundation UK (Ref:18107-1) for providing the initial financial support in the surveys. The acknowledgments are also due to Rector, Bhaderwah Campus and the University of Jammu for providing the necessary support for the smooth conduct of the research work.


1:  Lu C.Y., G.W., A.H. Dai and H.Y. Wei, 2012. Assessing habitat suitability based on geographic information system (GIS) and fuzzy: A case study of Schisandra sphenanthera Rehd. et Wils. in Qinling Mountains, China. Ecol. Model, 242: 105-115.
CrossRef  |  Direct Link  |  

2:  Ebrahimi, A., A. Farashi and A. Rashki, 2017. Habitat suitability of Persian leopard (Panthera pardus Saxicolor) in Iran in future. Environ. Earth Sci., 76: 697-707.
CrossRef  |  Direct Link  |  

3:  McCarthy, T.M., T.K. Fuller and B. Munkhtsog, 2005. Movements and activities of snow leopards in Southwestern Mongolia. Bio. Cons., 124: 527-537.
CrossRef  |  Direct Link  |  

4:  Jackson, R.M., J.D. Roe, R. Wangchuk and D.O. Hunter, 2006. Estimating snow leopard population abundance using photography and capture–recapture techniques. Wildl. Soc. Bull., 34: 772-781.
CrossRef  |  Direct Link  |  

5:  Janečka, J.E., B. Munkhtsog, R.M. Jackson, G. Naranbaatar, D.P. Mallon and W.J. Murphy, 2011. Comparison of noninvasive genetic and camera-trapping techniques for surveying snow leopards. J. Mammol., 92: 771-783.
CrossRef  |  Direct Link  |  

6:  Zeller, K.A., S. Nijhawan, R. Salom-perez, S.H. Potosme and J.E. Hines, 2011. Integrating occupancy modeling and interview data for corridor identification: A case study for Jaguars in Nicaragua. Bio. Cons., 144: 892-901.
CrossRef  |  Direct Link  |  

7:  Sharma, K., R. Bayrakcismith, L. Tumursukh, O. Johansson, P. Sevger, T. McCarthy and C. Mishra, 2014. Vigorous dynamics underlie a stable population of the endangered snow leopard Panthera uncia in tost mountains, south gobi, mongolia. PLoS ONE,
CrossRef  |  Direct Link  |  

8:  Peterson, A.T. and C.R. Robins, 2003. Using ecological niche Modeling to predict Barred Owl invasions with implications for Spotted Owl conservation. Cons. Bio., 17: 1161-1165.
CrossRef  |  Direct Link  |  

9:  Araújo, M.B., M. Cabeza, W. Thuiller, L. Hannah and P.H. Williams, 2004. Would climate change drive species out of reserves? An assessment of existing reserve-selection methods. Global Change Biol., 10: 1618-1626.
CrossRef  |  Direct Link  |  

10:  Sánchez-Cordero, V., V. Cirelli, M. Munguía and S. Sarkar, 2005. Place prioritization for biodiversity content using species ecological niche modeling. Bio. Info., 2: 11-23.
CrossRef  |  Direct Link  |  

11:  Phillips, S.J., R.P. Anderson and R.E. Schapire, 2006. Maximum entropy modeling of species geographic distributions. Ecol. Modell., 190: 231-259.
CrossRef  |  Direct Link  |  

12:  Elith, J., S.J. Phillips, T. Hastie, M. Dudik, Y.E. Chee and C.J. Yates, 2011. A statistical explanation of MaxEnt for ecologists. Divers. Distributions, 17: 43-57.
CrossRef  |  Direct Link  |  

13:  Garshelis, D.L. and R. Steinmetz, 2008. Ursus thibetanus. The IUCN Red List of Threatened Species 2008: e.T22824A9391633.

14:  Sathyakumar, S., 2001. Status and management of asiatic black bear and himalayan brown bear in India. Ursus, 12: 21-30.
Direct Link  |  

15:  Schaller, G.B., 1977. Mountain Monarchs: Wild Sheep and Goats of the Himalaya. University of Chicago Press, Chicago, IL., USA., ISBN-13: 9780226736419, Pages: 425
CrossRef  |  

16:  Sathyakumar, S. and A. Choudhury, 2007. Distribution and Status of Asiatic black bear Ursus thibetanus India. J. Bom. Nat. His. Soc., 104: 316-323.
Direct Link  |  

17:  Charoo, S.A., L.K. Sharma and S. Sathyakumar, 2011. Asiatic black bear-human interactions around Dachigam National Park, Kashmir, India. Ursus, 22: 106-113.
Direct Link  |  

18:  Kichloo. N.A., 1992. Census report of Kishtwar high altitude National Park. Department of Wildlife Protection, Kishtwar, Govt. of Jammu and Kashmir.

19:  Parsa, M.A., 1999. Survey of animals in Kishtwar high altitude National Park. Department of Wildlife Protection, Kishtwar, Govt. of Jammu and Kashmir.

20:  Baba, M.M., 2003. Annual animal census report of Kishtwar high altitude National Park 2002-03. Chenab Division, Kishtwar, Department of Wildlife Protection, Govt. of Jammu and Kashmir.

21:  Hilaluddin and R.Y. Naqash, 2006. Survey and census report of Kishtwar high altitude National Park. Chenab Wildlife Division, Department of Wildlife Protection, Govt. of Jammu and Kashmir.

22:  Hilaluddin and R.Y. Naqash, 2013. Densities and population sizes of large mammals in Kishtwar high altitude National Park, Jammu and Kashmir, India. Indian For., 139: 872-878.
CrossRef  |  Direct Link  |  

23:  Hanley, J.A. and B.J. McNeil, 1982. The meaning and use of the area under a Receiver Operating Characteristic (ROC) curve Radiology, 143: 29-36.
CrossRef  |  Direct Link  |  

24:  Fielding, A.H. and J.F. Bell, 1997. A review of methods for the assessment of prediction errors in conservation presence/absence models. Enviorn. Cons., 24: 38-49.
CrossRef  |  Direct Link  |  

25:  Allouche, O., A. Tsoar and R. Kadmon, 2006. Assessing the accuracy of species distribution models: Prevalence, kappa and the True Skill Statistic (TSS). J. Appl. Ecol., 43: 1223-1232.
CrossRef  |  Direct Link  |  

26:  Peterson, A.T., M. Papes and J. Sober, 2008. Rethinking receiver operating characteristic analysis applications in ecological niche modeling. Ecol. Modell., 213: 63-72.
CrossRef  |  Direct Link  |  

27:  QGIS Development Team, 2020. QGIS geographic information system. Open Source Geospatial Foundation Project.

28:  Bista, R. and A. Aryal, 2013. Status of the Asiatic black bear Ursus thibetanus in the south-eastern region of the Annapurna conservation area, Nepal. Zool Ecol., 23: 83-87.
CrossRef  |  Direct Link  |  

29:  Bista, M., S. Panthi and S.R. Weiskopf, 2018. Habitat overlap between Asiatic black bear Ursus thibetanus and red panda Ailurus fulgens in Himalaya. PLoS ONE,
CrossRef  |  Direct Link  |  

30:  Watts, S.M., T.M. McCarthy and T. Namgail, 2019. Modelling potential habitat for snow leopards (Panthera uncia) in Ladakh, India. PLoS ONE,
CrossRef  |  Direct Link  |  

31:  Dorji, S., K. Vernes and R. Rajaratnam, 2011. Habitat correlates of the red panda in the temperate forests of Bhutan. PLoS ONE,
CrossRef  |  Direct Link  |  

32:  Hiller, T. L., J. L. Belant, J. Beringer, and A. J. Tyre, 2015. Resource selection by recolonizing American black bears in a fragmented forest landscape. Ursus, 26: 116-128.
CrossRef  |  Direct Link  |  

33:  Takahata, C., N. Sadanori, K. Kirara and I. Shigeyuki, 2019. An evaluation of habitat selection of Asiatic black bears in a season of prevalent conflicts. Ursus, 24: 16-26.
Direct Link  |  

34:  Liu, F., M. William, G. David, Z. Xiaojian, W. Dajun, G. Ji’en and C. Youping, 2009. Spatial distribution as a measure of conservation needs: an example with Asiatic black bears in south-western China. Div. Dist., 15: 649-659.
CrossRef  |  Direct Link  |  

35:  Mcfadden-Hiller, J.E. and J. L. Belant, 2018. Spatiotemporal shifts in distribution of a recolonizing black bear population. Ecosphere,
CrossRef  |  Direct Link  |  

36:  Hernandez, P.A., C.H. Graham, L.L. Master and D.L. Albert, 2006. The effect of sample size and species characteristics on performance of different species distribution modeling methods. Ecography, 29: 773-785.
CrossRef  |  Direct Link  |  

37:  Pearson, R.G., C.J. Raxworthy, M. Nakamura and A.T. Peterson, 2007. Predicting species distributions from small numbers of occurrence records: A test case using cryptic geckos in Madagascar. J. Biogeogr., 34: 102-117.
CrossRef  |  

38:  Baldwin, A.R., 2009. Use of maximum entropy modeling in wildlife research. Entropy, 11: 854-866.
CrossRef  |  Direct Link  |  

©  2021 Science Alert. All Rights Reserved