Habitat suitability of four threatened Himalayan species: Asiatic black bear, common leopard, musk deer, and snow leopard

Background Biodiversity conservation is becoming challenging day by day. For this, it is essential to understand the distribution, habitat, and impact of anthropogenic activities on animals at risk. We assessed the suitable habitats and anthropogenic impacts on Asiatic black bears, common leopards, musk deer, and snow leopards in and outside the protected areas of Gandaki Province, Nepal. Methods We collected the presence locations of Asiatic black bears, common leopards, musk deer, and snow leopards based on scats and other signs. We employed the Maximum Entropy (MaxEnt) tool to identify suitable habitats of our studied species and their anthropogenic impacts on them. Results The total suitable habitat of the common leopard was found to be 6,052 km2, followed by the Asiatic black bear (5,819 km2), snow leopard (4,447 km2), and musk deer (1,690 km2) in Gandaki Province. Most of the areas of suitable habitat for common leopards and Asiatic black bears were outside the protected areas, and for musk deer and snow leopards were inside the protected areas. Elevation was the most important variable determining habitat suitability of Asiatic black bear, common leopard, and musk deer, whereas the distance to water was the most important variable determining habitat suitability of snow leopard. Asiatic black bears, common leopards, and musk deer face significant anthropogenic impacts, but snow leopards face some anthropogenic impacts. Conclusion Managing these animals’ habitats inside and outside protected areas is essential. Hence, biodiversity conservation and livelihood opportunities should be balanced in the Himalayas on a win-win basis.

Common leopards live in various habitats ranging from deserts to forests to high mountains in Asia, Africa, and East Europe (Stein et al., 2020).The Wild Cat Status Survey (IUCN/SSC Cat Specialist Group) has categorized leopards as a vulnerable species (Stein et al., 2020).Due to the high risk of extinction by a trade of its parts, it is included in Appendix I of CITES (CITES, 2019).Wild ungulates cover the primary dietary composition of the common leopard, whereas livestock is also identified as a diet of this leopard in a small amount (Aryal & Kreigenhofer, 2009).The common leopard in Nepal has a conservation focus, with the population estimated at <1,000 individuals.This leopard is a significant conflict-causing wildlife in Nepal (Acharya et al., 2016;Adhikari et al., 2020).Similarly, the common leopard is at high mortality risk due to human-wildlife conflict in central Nepal (Adhikari et al., 2022).Habitat encroachment is a major reason for livestock depredation by this leopard (Maharjan et al., 2017).
Musk deer is a widely but discontinuously distributed flagship species throughout the Himalayas from 3,000 to 4,400 m (Green, 1985(Green, , 1986;;Shrestha, 1997).In Nepal, it is recorded from 2,300 to 4,300 m in forests dominated by birch and rhododendron (Shrestha, 1998).This species is found in mountainous regions of Nepal, covering 30,177.19km 2 , of which 5,815.08 km 2 are located inside protected areas (Aryal & Subedi, 2011).Despite the more potential habitat outside the protected areas, its specific management activities were lacking (Aryal & Subedi, 2011).This lack assisted in a surge in persistent population decline (Aryal et al., 2010).Because of anthropogenic activities such as over-exploitation of forest resources, habitat shrinkage, destruction, and degradation, this species is listed as endangered on the IUCN red list (Timmins & Duckworth, 2015).Also, it has been listed in CITES Appendix I (CITES, 2019) and is protected under the National Park and Wildlife Conservation Act of Nepal.Musk deer dwell in steep, forested, or shrub-covered slopes, mainly in the sub-alpine zones of mountain regions.Dense undergrowth of rhododendron (Rhododendron anthopogon, R. arboreum) bamboo (Arundinaria spp), and other shrubs constitute the typical habitat (Green, 1987).Habitat use depends on the availability of food, cover, and other factors.They are very shy and solitary animals and may not become active until dusk.
The endangered snow leopard inhabiting the rugged and fragile landscape of the Himalayas (Jackson & Ahlborn, 1989) is one of the predators in the energy-deficient environments of high altitudes.It is assessed as "vulnerable" since its global population is estimated to be between 2,500-10,000, with its projected decline of at least 10% in 23 years (McCarthy et al., 2017).Similar to other wildlife studies, the snow leopard is also included in Appendix I of CITES (CITES, 2019).Snow leopards are a significant conflict causing wildlife in the High Mountain region of Nepal.They attack the livestock near the livestock sheds in the rangelands of the High Mountain (Karki & Panthi, 2021).Livestock is also the prey of this leopard in the Mountainous region (Oli, Taylor & Rogers, 1993).Due to climate change, its habitat is expected to decline (Aryal et al., 2016).Similarly, about 30% of snow leopard habitat may be lost due to a shifting treeline and consequent shrinking of the alpine zone in the Himalayas (Forrest et al., 2012).Retaliatory killing, poaching, habitat degradation, and prey depletion are considered key factors leading to their population decline (McCarthy et al., 2017).
Given the importance and conservation threats of the Asiatic black bear, common leopard, musk deer, and snow leopard in Nepal, it is essential to have comprehensive information on the distribution, habitat, and anthropogenic impacts on these species of interest.Comprehensive information on the distribution and impact of human activities on threatened wildlife is still lacking in the central part of Nepal.Hence, this study attempts to solve these gaps for the Asiatic black bear, common leopard, musk deer, and snow leopard in the Gandaki Province of Nepal using a MaxEnt model.It also assesses suitable habitats of these species in and outside the protected areas.Habitats inside the protected area system face low anthropogenic activities compared to those outside this system.Furthermore, these areas are especially managed for conservation.Therefore we hypothesized that most of the habitat of these four species is covered by the protected area system.

Study area
This study was carried out in Gandaki Province, the central part of Nepal (Fig. 1).

Data collection
Occurrence points of species The field survey was conducted from February to June 2020 (approximately 85 days) throughout the Gandaki Province to collect the primary data for the study.The presence of wildlife was recorded based on the direct sight of the particular species or its indirect signs (scats, hairs, and footprints) in the field.Finally, we maintained at least 500 m in each presence point to lessen spatial auto-correlation and retain 62 points for Asiatic black bear, 109 points for common leopard, 62 points for musk deer and 61 points for snow leopard for modelling purposes.

Environmental variables
We downloaded environmental variables from online sources.The Digital Elevation Model (DEM) was obtained from the United States Geological Survey (https:// earthexplorer.usgs.gov/).This DEM was further used to prepare the slope and aspect with the help of ArcGIS software (release 10.5; Esri, Redlands, CA, USA).Water sources were acquired from the Geofabrik website (https://www.geofabrik.de/data/shapefiles.html) (GEOFABRIK, 2020) and converted to a distance raster file.These data were transformed to ASCII format and spatial resolution of 30 m using ArcGIS (release 10.5; Esri, Redlands, CA, USA) (Table 1).
Herbivores are expected to be directly affected by the vegetation characteristics in the habitat as they form a food source.In contrast, carnivores maybe affected directly (through its effect on shade and hiding areas for predators) or indirectly (through its effect on herbivore prey presence).This study used vegetation-related variables such as forest cover and enhanced vegetation index (EVI).Forest cover data were taken from the global forest change website (Hansen et al., 2013).Also, we used Landsat 8-based EVI time series data for 2018-2019.
This study used anthropogenic datasets, including roads; settlements; and land use land cover (LULC).Vector datasets containing paths and roads were obtained from Geofabrik (https://www.geofabrik.de/data/shapefiles.html) (GEOFABRIK, 2020), and the settlements were obtained from the Department of Survey, Nepal.ArcGIS transformed these files into distance raster files with a distance tool (release 10.5; Esri, Redlands, CA, USA).Distance to paths, roads, and settlements is a distance of a particular pixel of a raster file from paths, roads, and settlements.In addition, LULC data were acquired from the International Centre for Integrated Mountain Development website (ICIMOD; http://www.icimod.org)(Uddin et al., 2015).We selected these environmental variables for modeling based on availability and their importance.The help of existing literatures and expert knowledge identified the importance of variables.

Habitat modeling of species
We used Maximum Entropy (MaxEnt) software, version 3.4.1, to predict the species distributions and the anthropogenic impact on these species using species presence points and environmental variables (Table 1) (Phillips, 2017).MaxEnt is a standard and widely used species distribution software.This software is already used to model the suitable habitat of several faunas in Nepal (Aryal et al., 2016;Bista, Panthi & Weiskopf, 2018;Panthi, 2018;KC et al., 2019;Panthi et al., 2019;Sharma et al., 2020).Variance Inflation Factors (VIFs) less than 10 indicate that multicollinearity is not a severe problem (Forthofer, Lee & Hernandez, 2007).Therefore, we calculated the VIFs of variables with the help of R Studio software version 4.1.1 and used those variables having VIF < 10 in the model (Forthofer, Lee & Hernandez, 2007;Panthi, Pariyar & Low, 2021).Of the total presence points, 70% were used for training, and the rest 30% were used for testing and validation.We used 5,000 background points, 10 replications, 1,000 iterations, and subsample replicated run type to model the suitable habitat for all four species.Discriminatory ability showed a slight peak in performance around the default regularization multiplier (Radosavljevic & Anderson, 2014), so we used the default (1) regularization multiplier during the modelling.

Accuracy assessment
Accuracy assessment is a prerequisite to understandingand validating the models' performance.In this study, we used two methods such as threshold-independent and threshold dependent, for model evaluation.The threshold-independent method involves the area under the receiver-operator curve (AUC) elucidating higher than the AUC, better the model performance (Pearce & Ferrier, 2000).AUC values above 0.9 explain excellently, 0.7-0.9explains moderate, and lower than 0.7 explains poor model performances (Pearce & Ferrier, 2000).However, researchers criticized this method as a classical one, which is supposed to be affected by the non-uniform spatial distribution of samples (Lobo, Jiménezvalverde & Real, 2008).Hence, this study used a threshold-dependent method using the True Skill Statistic (TSS) as a model evaluator (Merow, Smith & Silander, 2013).TSS = Sensitivity + Specificity − 1, and ranges from −1 to 1, where values less than 0 indicate a performance no better than random and 1 indicates a perfect fit of the model (Allouche, Tsoar & Kadmon, 2006;Panthi, Pariyar & Low, 2021).Of the 10 models (0-9 replications) used in the study, TSS was calculated for all models using the PresenceAbsence package in R studio version 4.1.1 and averaged for the final TSS (Jiang et al., 2014;Panthi, Aryal & Coogan, 2019).The sum of sensitivity and specificity was used as a threshold to convert the habitat suitability map (raw output of MaxEnt) into a suitable/unsuitable binary map and to calculate the TSS (Liu, White & Newell, 2013).

RESULTS
The suitable habitats of the species The species-specific suitable habitats and their overlap habitats have been presented in Table 2.In the study area, 6,052 km 2 was identified as a suitable habitat for the common leopard.Moreover, 5,819 km 2 was identified as a suitable habitat for Asiatic black bears, 4,447 km 2 was identified as a suitable habitat for snow leopards, and 1,690 km 2 was identified as suitable habitat for musk deer.Out of them, more than four-fifth of the suitable habitat of musk deer and three-quarters of the suitable habitat of snow leopards were recorded in protected areas.Most suitable habitats of the common leopards and Asiatic black bears were recorded outside the protected areas, with 88% and 61%, respectively.Substantial suitable habitats for all four species were found in Annapurna Conservation Area.In Annapurna Conservation Area, snow leopards have predicted suitable habitat of 2,472 km 2 , followed by an Asiatic black bear (1,676 km 2 ), musk deer (905 km 2 ), and common leopard (701 km 2 ).While considering habitat overlap between two species, it was observed that overlapped habitat of musk deer and snow leopard inside the protected areas was 85%, followed by Asiatic black bear and musk deer (79%), common leopard and musk deer (78%), Asiatic black bear and snow leopard (63%), common leopard and snow leopard (51%), Asiatic black bear and common leopard (27%).The areas of overlapped habitats for all species were found to be the highest in Annapurna Conservation Area.In addition, the same areas were observed to be suitable for three species of interest simultaneously.These represent overlapped habitat of the Asiatic black bear, musk deer, and snow leopard in Annapurna Conservation Area 355 km 2 ; followed by an Asiatic black bear, common leopard, and musk deer (67 km 2 ); and common leopard, musk deer and snow leopard (32 km 2 ).All four species of Asiatic black bear, common leopard, musk deer, and snow leopard can share an area of 28 km 2 as their typical suitable habitat.

Important variables used in the models
Out of 14 variables used in the model, elevation, LULC, distances to water, and minimum EVI were found to be the most important variables determining the habitat suitability of the Asiatic black bear.Anthropogenic variables (distance to path, distance to building, distance to settlement, and distance to road) were moderately essential to model the suitable habitat of this species (Fig. 2).Similarly, elevation, minimum EVI, distance to a path, and LULC were found to be the most important variables to model the suitable habitat of musk deer (Fig. 3).Moreover, anthropogenic variables (distance to building, distance to settlement, and road distance) were moderately important to model the suitable habitat of this speceis.
Also, elevation, distances to water, slope, and maximum EVI were found to be the most important variables to model the suitable habitat of common leopard (Fig. 4).Furthermore, anthropogenic variables (distance to building, distance to settlement, distance to road, and distance to path) were moderately important to model the suitable habitat of this speceis.
Of the 14 variables used in the model, distance to water, minimum EVI, distances to settlement, and aspect were found to be the most important variables to model the suitable habitat of snow leopard (Fig. 5).Anthropogenic variables such as distance to road, distance to path, distance to building, and LULC were moderately important to model the suitable habitat of this speceis.Response of habitat suitability of species to anthropogenic variables The species-specific models of Asiatic black bears indicate that their existing habitats are near four anthropogenic variables such as building, path, road, and settlement.The habitat suitability of this species increases with decreasing distance from buildings, paths, roads, and settlements (Fig. 6).The map of suitable habitats of this species is shown (Fig. 7).
The species-specific models of musk deers indicate that their existing habitats are near four anthropogenic variables such as building, path, road, and settlement, and habitat suitability of this species increases with decreasing distance from building, path, road, and settlement (Fig. 8).The map of suitable habitats of this species is shown (Fig. 9).The species-specific models of common leopards indicate that their existing habitats are near four anthropogenic variables such as building, path, road, and settlement, and habitat suitability of this species increases with decreasing distance from building, path, road, and settlement (Fig. 10).The map of suitable habitats of this species is shown (Fig. 11).In the case of snow leopards, habitat suitability is the maximum at a certain distance from the building, road, and settlements (Fig. 12).The map of suitable habitats of this species is shown (Fig. 13).

Model accuracy
The accuracies of the models were found to be relatively suitable for all models.We obtained 0.828 ± 0.053 AUC and 0.607 ± 0.086 TSS for the model, which predicts the suitable habitat of the Asiatic black bear (Table 3).Similarly, we obtained 0.964 ± 0.016 AUC and 0.864 ± 0.072 TSS for the model predicting the suitable habitat of musk deer (Table 4); 0.858 ± 0.040 AUC and 0.608 ± 0.081 TSS for the model predicting suitable habitat of common leopard (Table 5); and 0.879 ± 0.014 AUC and 0.655 ± 0.038 TSS for the model predicting suitable habitat of snow leopard (Table 6).

DISCUSSIONS
This comprehensive study identifies the suitable habitat of the four threatened Himalayan wildlife in Gandaki Province, Nepal, with the help of MaxEnt modelling.The total suitable habitat of the Asiatic black bear is found to be 5,819 km 2 .Similarly, suitable habitats for common leopards, musk deer, and snow leopards are 6,052, 1,690, and 4,447 km 2 , respectively.Unlike our hypothesis, many patches of suitable habitat for the threatened fauna are identified outside the protected areas.
Our study shows that most of the suitable habitat of Asiatic black bears is found outside the protected area (61%).Similarly, Mohammadi et al. (2021) identified most brown bears' habitat outside the protected area in Iran.It was reported in the village that the black bears mainly inhabit the forests and visit agricultural fields close to the forests (Ali et al., 2017).The Asiatic black bear shares a habitat with the rest of the studied species in the study area.Previous studies also recorded the habitat sharing of Asiatic black bears with red pandas in Makalu Barun National Park, Nepal (Bista, Panthi & Weiskopf, 2018).
Most of the signs of Asiatic black bears have been recorded at an elevation ranging from 1,900 to 3,100 m (Bista & Aryal, 2013).Elevation and bioclimatic variables were major contributors to model the habitat of the Asiatic black bear in the Gandaki River Basin.A similar study area of this study used bioclimatic variables to predict the climate change effect (Rai et al., 2022).However, our study did not use these bioclimatic variables considering a coarse resolution to model the wildlife habitat in a small study area.So we might find the elevation as the most important variable, which could work as a proxy for climatic variables.Water resource availability was the most important variable in modeling the brown bear distribution in Iran (Ansari & Ghoddousi, 2018).
In recent years, anthropogenic activities and infrastructure development, such as road construction, have jeopardized the habitats and life forms of Asian elephants (Sharma et al., 2020).Similarly, the habitat of Asiatic black bears is declining due to human encroachment (Escobar, Awan & Qiao, 2015).Cereal crops such as maize and millet were identified in the diet of Asiatic black bears in two protected areas of Nepal (Panthi, Aryal & Coogan, 2019).This evidence proves that the Asiatic black bear is facing anthropogenic impact.
We found elevation, distances to water, slope, and maximum EVI were found to be the most important variables in modeling the suitable habitat for common leopard.Leopards showed a high tolerance to infrastructure and fragmented patches of forests used by humans for agriculture and other land uses (Sarkar et al., 2018).We also find anthropogenic variables (distance to building, distance to settlement, distance to road, and distance to path) moderately important to model the suitable habitat of common leopards.
The common leopards are widely distributed wild cats and occupy various habitats which includes rainforest, deserts, fringes of urban areas and remote mountains (Dickman & Marker, 2005;Baral et al., 2021).They are known to exhibit a high tolerance to human activities in their habitat (Athreya et al., 2013).In Nepal, many common leopards were recorded in community-managed forests outside protected areas, threatening human lives living closer to the forests (Baral et al., 2021).Habitats of the common leopard and tiger (P.tigris) overlapped in India (Rather, Kumar & Khan, 2020).The leopard prefers agricultural land until it is covered with bush and forest, as these areas have a greater affinity towards the prey base (Maharjan et al., 2017).Among the six variables chosen deliberately for the model, the most important variables were settlement area, sparse forest, bush, and roads (Maharjan et al., 2017).Habitat suitability modeling results of the common leopard in a representative Himalayan landscape of Kailash Sacred Landscape of India assured that the common leopard habitat is not only influenced by topographic and environmental features but also by their combined spatial arrangements at different spatial scales (Sarkar et al., 2018).Avoidance of humans was the primary influence on the distribution of the Persian leopard (Panthera pardus saxicolor) in the Montane areas of West Asia (Farhadinia et al., 2015).The contribution of the variable "distance from settlement area" was the highest (52.4%) to impact the model while calculating the habitat suitability of the common leopard in Shivapuri Nagarjun National Park (Maharjan, 2013).The probability of the presence of P. pardus also increases up to a distance of 750 m and whereas decreases up to 1,200 m (Maharjan et al., 2017).Also, anthropogenic factors such as the extent of 31 spots of deforestation (decrease in forest cover) negatively affected leopard occupancy in Nepal's Terai Arc Landscape region (Thapa et al., 2020).
We identified most of the habitat of musk deer inside the protected areas.In Nepal, very few studies were conducted to identify the habitat of musk deer.Based on the government documents, Annapurna Conservation Area, Manaslu Conservation Area, and Dhorpatan Hunting Reserve are the habitat of musk deer (Department of National Parks and Wildlife Conservation of Nepal (DNPWC), 2017).Musk deer prefers high-altitude regions above 2,500 m and is found mostly in the central and eastern parts of the country (Lamsal et al., 2018).In our study, most of the habitat of the musk deer falls inside the Annapurna Conservation Area, Manaslu Conservation Area, and Dhorpatan Hunting Reserve.Most parts of these protected areas are situated above the 2,500 m elevation.Temperature is the most influential variable in the distribution of musk deer in Nepal Himalaya since annual mean temperature and isothermality are two key contributors to the model (Lamsal et al., 2018).The vegetation type/land use was the most important variable for the habitat suitability assessment of the Himalayan musk deer in Kedarnath Wildlife Sanctuary (Nandy, Lakshmi & Kushwaha, 2020).Rocky areas, ridges, mixed coniferous forest (Red pine and Mongolian oak), southern and southeastern slopes, elevation above 800 m, and distance from water course less than 300 m were dominant habitat types of Siberian musk deer.This study also found DEM and mean EVI to be the most important variables determining habitat suitability of musk deer.In this study, temperature was not taken as the separate variable but accompanied by elevation.Musk deer population has been declined in its native regions owing to various anthropogenic threats such as habitat fragmentation and illegal hunting (Khanal, 2020).For the selection of habitat of the musk deer habitat types, fuel wood and timber cutting, rock cover, litter cover and distance to settlements were the main factors (Shrestha & Xiuxiang, 2014).Habitat of the musk deer was associated with contrasting species composition of trees and forbs, and certain species of shrubs.Main causes for reducing the population of musk deer was poaching and human induced habitat alterations (Shrestha & Xiuxiang, 2014).
Our study claims that most of the suitable habitat of the snow leopard was recorded inside the protected area.Based on the habitat model, Jackson & Ahlborn (1989) concluded that 65% of Nepal's snow leopard population is inhabited outside the protected areas.After that study, two conservation areas (Annapurna and Manaslu) were established in our study area.Now, most of the snow leopard habitat is covered by these conservation areas.Annapurna and Manaslu Conservation areas are snow leopard habitats (Aryal et al., 2014; Department of National Parks and Wildlife Conservation of Nepal (DNPWC), 2017).DEM and LULC were found to be the most important variables determining the habitat suitability of snow leopards.The snow leopard's presence ranged from 2,965 to 5,831 m and a maximum distance to water of 1,575 m (Watts, McCarthy & Namgail, 2019).However, in our study, elevation is not identified as the most important environmental variable (Fig. 5).Land cover and aspect contributed less to the model, and land cover and prey were the least important ones (Watts, McCarthy & Namgail, 2019).In our study, distance to water, minimum EVI, distances to settlement, and aspect are found to be the most important variables in modeling the suitable habitat of snow leopards.
Similarly, we observed the habitat overlap of snow leopards with the other three studied species in the study area.Wild herbivores, marmots, and livestock (blue sheep and other wild ungulates) were major snow leopard prey (Oli, Taylor & Rogers, 1993).Our study identified 1,211 km 2 as overlapped snow leopard and musk deer habitats.In overlapped habitats, musk deer may be prey for the snow leopard.Snow leopards are significant conflict-causing wildlife in the High Mountain region of Nepal; they attack the livestock near the livestock sheds in rangelands of the High Mountain region (Oli, Taylor & Rogers, 1993;Karki & Panthi, 2021).Retaliatory killing, poaching, smuggling, habitat degradation, and prey depletion are considered key factors leading to their population decline (McCarthy et al., 2017).
We found the anthropogenic impact on the habitat of our studied animals in the study area.The habitat suitability of Asiatic black bears, common leopards, and musk deer increases with decreasing distance from buildings, paths, roads, and settlements (Figs. 6,8 and 10).This means these animals' habitats are near the path, road buildings, and settlements.Therefore we can assume that these three wildlife face an enormous anthropogenic impact on their habitat.In the case of snow leopards, habitat suitability is maximum at a certain distance from the building, road, and settlements (Fig. 12).This information can indicate that the snow leopard is facing a moderate anthropogenic impact on its habitat.Habitats of snow leopards are grassland and bare land of the High Himalayan region, which are very far from human activities.Therefore anthropogenic variables might be less influential in predicting this species' habitat.

CONCLUSION
Some species can utilize a wider range of habitats than others, resulting in a more extensive suitable habitat.Unlike our hypothesis, most of the areas of suitable habitat for common leopards and Asiatic black bears were outside the protected areas.Suitable habitats of musk deer and snow leopards were inside the protected areas.This spatial distribution of the suitable habitat of the threatened fauna suggests managing the wildlife habitat outside the protected areas.Annapurna Conservation Area, Dhorpatan Hunting Reserve, and Manaslu Conservation Area are situated in the mountainous region of Gandaki Province.These protected areas cover most of the suitable musk deer and snow leopard habitats.Habitats located inside and outside the protected areas are equally important for this wildlife.Therefore, we recommend managing these animals' habitats inside and outside protected areas.This study emphasizes maintaining land use land cover based on the need of particular animal species, which is crucial to halt the rate of population decrease and an important step to biodiversity conservation.Furthermore, protecting and managing water sources is also recommended to conserve the wildlife in the study area.
Hari Adhikari analyzed the data, authored or reviewed drafts of the article, and approved the final draft.Shiva Pariyar conceived and designed the experiments, performed the experiments, analyzed the data, prepared figures and/or tables, and approved the final draft.Rishi Baral conceived and designed the experiments, analyzed the data, authored or reviewed drafts of the article, and approved the final draft.Rukmagat Subedi conceived and designed the experiments, performed the experiments, analyzed the data, authored or reviewed drafts of the article, and approved the final draft.Bishnu Prasad Adhikari performed the experiments, authored or reviewed drafts of the article, and approved the final draft.Mahesh Poudel conceived and designed the experiments, performed the experiments, prepared figures and/or tables, and approved the final draft.Nischal Sedhai conceived and designed the experiments, performed the experiments, prepared figures and/or tables, and approved the final draft.Megharaj Poudel performed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the article, and approved the final draft.

Figure 2
Figure 2 Variables importance for model training for Asiatic black bear.The regularized training gain shows how much better the model distribution fits the presence data relative to a uniform distribution."Withoutvariable" denotes the effect of removing that single variable from the model "with only variable" denotes the results of the model when an only that variable is run; "with all variables" indicates the results of the model when all variables are run(Phillips, 2017).Full-size  DOI: 10.7717/peerj.16085/fig-2

Figure 3
Figure 3 Variables importance for model training for musk deer.The regularized training gain shows how much better the model distribution fits the presence data relative to a uniform distribution."Withoutvariable" denotes the effect of removing that single variable from the model "with only variable" denotes the results of the model when an only that variable is run; "with all variables" indicates the results of the model when all variables are run(Phillips, 2017).Full-size  DOI: 10.7717/peerj.16085/fig-3

Figure 4
Figure 4 Variables importance for model training for common leopard.The regularized training gain shows how much better the model distribution fits the presence data relative to a uniform distribution."Withoutvariable" denotes the effect of removing that single variable from the model "with only variable" denotes the results of the model when an only that variable is run; "with all variables" indicates the results of the model when all variables are run(Phillips, 2017).Full-size  DOI: 10.7717/peerj.16085/fig-4

Figure 5 26 Figure 6 Figure 7 26 Figure 9 Figure 8 26 Figure 13
Figure 5 Variables importance for model training for snow leopard.The regularized training gain shows how much better the model distribution fits the presence data relative to a uniform distribution."Withoutvariable" denotes the effect of removing that single variable from the model "with only variable" denotes the results of the model when an only that variable is run; "with all variables" indicates the results of the model when all variables are run(Phillips, 2017).Full-size  DOI: 10.7717/peerj.16085/fig-5 (WWF, 2013)e's total area is 21,976.34km 2 , which is 14.9% of the total area of Nepal.This province has five distinct geographical regions: Himalaya, high mountains, middle mountains, Siwalik, and Terai.The Himalayan region covers the upper part of the Gandaki Province.Dhawalagiri (8,167 m), Manaslu (8,136 m) and Annapurna (8,091 m) are major mountains of the Gandaki Province.Around 45% of this province is covered by protected areas (Ministry of Industry, Tourism, Forest and Environment of Nepal (MoITFE), 2018).Annapurna Conservation Area, Manaslu Conservation Area, some parts of Dhorpatan Hunting Reserve, and Chitwan National Park are located in this province.ACA Dhorpatan Hunting Reserve and Manaslu Conservation Area are habitats of Himalayan threatened Moreover, the diversity of orchids is high in the Panchase forest conservation area(WWF, 2013).Being situated at the divide of the Eastern and Western floristic regions, the Kali Gandaki Gorge is a recognized corridor for birds to migrate (WWF, 2013).
fauna like Asiatic black bear, common leopard, the grey wolf (Canis lupus), musk deer, red panda, snow leopard and wild dog (Cuon alpinus).These protected areas are famous for trekking, unique landscape, and mountain biodiversity (Department of National Parks and Wildlife Conservation of Nepal (DNPWC), 2017; Ministry of Industry, Tourism, Forest and Environment of Nepal (MoITFE), 2018).

Table 2
Suitable habitat of the species.

Table 3
Thresholds and accuracies of different replications during the modelling of habitat suitability of Asiatic black bear in 10 (0-9) replications.