Habitat selection of Himalayan Musk Deer Moschus leucogaster (Mammalia: Artiodactyla: Moschidae) with respect to biophysical attributes in Annapurna Conservation Area of Nepal

Himalayan or White-bellied Musk Deer Moschus leucogaster, an IUCN indexed endangered species, is distributed in isolated pockets in the Himalaya. The deer population is decreasing owing to several pressures that include habitat loss and fragmentation, and poaching. It is essential to identify preferred habitat characteristics to support appropriate management strategies for conserving this endangered species. This study was carried out in the Nysheang basin of Annapurna Conservation Area of Nepal to identify habitats preferred by the musk deer. Habitat field parameters were collected using transect surveys. To analyze vegetation use and availability, nested quadrate plots size 20 m2 were established. Ivlev’s electivity index (IV) (-1 to +1) was employed to determine habitat preference, and one-way ANOVA (F) and chi-square tests (χ2) were used to examine different habitat parameters. Similarly, the importance value index (IVI) of the vegetation was calculated. Our results showed that the Himalayan Musk Deer strongly preferred habitats at 3601–3800 m altitude (IV= 0.3, F= 4.58, P <0.05), with 21–30o slope (IV= 0.2, F= 4.14, P <0.05), 26–50 % crown cover (IV= 0.25, F= 4.45, P <0.05), 26–50 % ground cover (IV= 0.15, F= 4.13, P <0.05), and mixed forest (IV= 0.29, χ2= 28.82, df= 3, p <0.001). Among the trees, Abies spectabilis (IVI= 74.87, IV= 0.035) and Rhododendron arboretum (IVI= 55.41, IV= 0.02) were the most preferred, while Rhododendron lepidotum, Cassiope fastigiata (IV= 0.35) and Berberis aristata (IV= 0.25) were the most preferred shrubs, and Primula denticulata (IV= 0.87) and Primula rotundifolia (IV= 0.31) were the most preferred herbs. These preferred habitat conditions should be maintained and conserved to sustain a viable population of deer in the study area. Further studies will be required to assess the effects of climate change on habitat suitability.


INTRODUCTION
Musk Deer under genus Moschus are of taxonomic, biological, and commercial interest; the latter primarily arising from the value of the musk produced by adult male deer (Khadka & James 2016). Refined and improved knowledge has enabled the recognition of seven Moschus species (Li et al. 2016), with three occurring in Nepal (Satyakumar et al. 2015 The National Parks and Wildlife Conservation Act, 2029Act, (1973, Nepal (GoN 1973) includes the Musk Deer Moschus chrysogaster (Image 1) in Schedule-1 as a "Protected Wildlife" species. Earlier, M. chrysogaster was believed to be the only Musk Deer species of Nepal. M. fuscus was believed to be extinct, or not recorded in Nepal (Bhuju et al. 2007, page 30, 106), and M. leucogaster was earlier treated as subspecies of M. chrysogaster (Satyakumar et al. 2015). In the present study, we have treated the Musk Deer of Annapurna Conservation Area as Moschus leucogaster (hereby Musk Deer) in central Nepal. The species is categorized as 'Endangered' in the IUCN Red List (Harris 2016).
Habitat preference is an intrinsic behavior that determines the selection and fitness of species to particular habitat (Jaenike & Holt 1991). It is an element of natural factors which may prompt to the improvement of asset choice behavior (Boyce & McDonald 1999;Manly et al. 2007). An asset choice may be forever or briefly exhausted by the action of the creature (Green 1986). Moreover, habitat preference is the disproportionality among utilization and accessibility (Manly et al. 2007). Creatures are liable to contending requests and inspirations for example, must secure nourishment, discover mates, raise offspring, protect restricted assets, and maintain a strategic distance from predators. So as to achieve these goals, their decision of natural surrounding selection is influenced and balanced over their area in space (Hebblewhite & Merrill 2009). The majority of the wildlife conservationists have concentrated on natural surrounding selection for managing the populaces and anticipating impacts of natural surrounding disturbances (Boroski et al. 1996). Other than this, however, it can be utilized as an apparatus to see how environment, behavior and wellness are connected (McLoughlin et al. 2008;Gaillard et al. 2010). The growing anthropogenic weight and their following impacts on natural life has been well seen all around (Millenium Ecosystem Assessment 2005).
The population of Musk Deer is declining due to several anthropogenic pressures, including illegal hunting and habitat loss or degradation (Jnawali et al. 2011) due to human encroachment, firewood collection, etc. (Thapa et al. 2018). Suitable living space for deer is principally limited to protected areas in fragmented habitats (Singh et al. 2018a). As per Shrestha (2012), Musk Deer is one of the least studied mammals and its population is found in highly isolated areas. Hence taking all these considerations, our study was focused to identify and explore the state of the habitats in respect

J TT
of topographic and vegetation highlights that portray their habitat preferences.

Study Area
The Annapurna Conservation Area (ACA) is located in the hills and mountain of west-central Nepal (28.231-29.336 (Figure 1), within the north-east portion of ACA is one of the major pocket areas for Musk Deer (Singh et al. 2018a). It occupies an area 689.6 km 2 and elevation ranging 2,900-7,939 m.

Data Collection
The study was conducted during March of 2018. At that time, the snowfall had decreased and the melting of snow had accelerated, which aided our investigation. To identify habitat parameters, a random sampling technique was utilized. Throughout the study area 'habitat use plots' (U) and availability plots (A) were adopted. On each location where indirect signs of Musk Deer such as latrine, hair, pugmark, and bed site were observed; 'habitat use plot' was established within 50 m

J TT
distance. Habitat parameters, in particular the gradient, altitude, crown cover, ground cover and land features were noted from each plot. 'Habitat availability plots' were chosen at 100 m distance from the use plots in a random direction (Panthi et al. 2012) and the similar habitat parameters were noted as recorded in the use plots. 'Availability plots' were renamed as 'use plots' if signs of the deer were present in availability plots. Vegetation analysis was performed within both the use and availability plots. Quadrats of size 20 × 20 m were placed on each transect at the intervals of 100 m (Singh et al. 2018a). Within the quadrats, nested structured small quadrats of size 5 × 5 m and 1 × 1 m were laid ( Figure 2). Trees (dbh >10cm) were measured in each 20 × 20 m quadrat, shrubs and sapling (tree species >1 m height and <10 cm diameter) were measured in 5 × 5 m quadrats and seedlings (tree <1 m in height) were measured in 1 × 1 m quadrats and those measurements were recorded. Besides, information such as the tree diameter at breast height (DBH), height, crown cover, number of trees, ground cover, frequency of tree, shrub and herb as well as signs of animals were collected within the quadrats.

Data Analysis
Using Ivlev's electivity index (IV), habitat preference of deer was analyzed. The IV value ranges from -1.0 to + 1.0. Habitat preference is indicated by the positive value, whereas negative value indicates avoidance and finally, 0 values indicate random use (Ivlev 1964). For this purpose, following relation was used.
Regarding vegetation analysis, the field data was utilized to calculate the species richness, frequency and relative frequency, density, and relative density of tree using following formulae (Smith 1980). Importance value index (IVI) was calculated as IVI = Relative density + relative frequency + relative dominance.
Besides, one-way ANOVA and Chi-square test were used to identify the significances of different habitat variables; crown cover, ground cover, forest types with respect to Musk Deer presence at 5% level of significance.

Habitat Preferences
Altitude Preference: The Musk Deer mainly preferred altitudinal ranges of 3,601-3,800 m with (IV= 0.3) (Figure 3). Altitudinal preference increased from 3201 m to 3800 m in a gradual manner. The altitudinal range of 3,801-4,000 m (IV= 0.2) was least preferred. Similarly, the region beneath the elevation 3,200 m (IV= -0.25) and above 4,000 m (IV= -0.8) was avoided. The utilization of different altitude intervals in extent to their availabilities was statistically significant (F= 4.58, P <0.05).
Slope Preference: Primarily, the Musk Deer preferred the slope 21º to 30º (IV= 0.2) (Figure 4). Preference slope expanded in continuous way from 11º to 30º and somewhat diminished up to 40º. It avoided the slope <10º (IV= -0.25) and >40º (IV= -0.71). The use of different slopes in extent to their availability was statistically significant (F= 4.14, P <0.05).
Crown Cover Preference: Mainly, the Musk Deer favored the crown cover of 26 to 50 % (IV= 0.25) followed by crown cover of 51 to 75 % (IV= 0.05), while 76 to 100 % (IV= -0.65) crown cover was evaded ( Figure 5). The utilization of different crown cover in extent to their availability was statistically significant (F= 4.45, P <0.05).
Ground Cover Preference: Initially ground cover was partitioned in 4 classes for the analysis. Ground cover having 26-50 % (IV= 0.15) and 0-25% (IV= 0.09) was mostly preferred by Musk Deer while it completely avoided 76-100 % cover (IV = -0.75) (Figure 6). This suggests that it preferred scarce and modest ground cover. The use of different ground cover in extent to their availability was statistically significant (F= 4.13, P <0.05).
Since most of pellet was documented in forest, it was figured out that the Musk Deer preferred forest (IV= 0.15) ( Figure 7). The cliff (IV= 0) and rock (IV= 0) were utilized randomly and the stream-bed (IV= -0.43) was totally dodged. The use of different ground features in extent to their availability was statistically significant (F= 3.29, P <0.05).

DISCUSSION
Habitat usage relies upon factors like the creature's behavior, length of the day and the time of year in relation to accessibility of food, shelter, and cover (Green & Kattel 1997). Anthropogenic and natural factors may also influence accessibility to habitats and modify habitat preference (Pulliam & Daielson 1991). It is also possible that preferences vary among species of the same genus. In this context, without attempting to specify species level differences, we observed that our base-line findings (Table 1) on habitat preference by Musk Deer from ACA are comparable to certain extents with other studies in Nepal and neighborhood.
Khadka & James (2016) found that Musk Deer preferred small patch of pine and fir forest in the central Himalayas. While in ACA the preferences were the maximum in mixed forest to the minimum in Betula forest, and the preference for Rhododendron forest was low, close to that of Betula forest. The preference for

J TT
forests of mixed stands and Rhododendron in our study appears similar to the findings by Shrestha & Meng (2014) in Gaurishankar Conservation Area, Nepal.
Concerning preferences for altitude range, Timmins & Duckworth (2015) suggested that 2,500-4,800 m is the most preferred for M. leucogaster, while Thapa et al. (2019) mentioned that 3,700-3,800 m was the foremost favored altitudinal extent for Moschus in Khaptad National Park, Nepal. Ilyas (2015) observed that a majority of the latrines of M. chrysogaster in Uttarakhand Himalaya, India occurred from 4,200 m down to 2,500 m. A study carried out by Srivastava & Kumar (2018) revealed that Musk Deer preferred the habitat within the altitude range 3,600-3,900 m in Sikkim Himalaya. Likewise, the Musk Deer highly preferred that altitude range 3,600-3,900 m in Api-Nampa Conservation Area, Nepal (ANCA 2018). In our study, the species favored the altitudes of 3,600-3,800 m, which is similar to the altitudinal preference in Api-Nampa Conservation Area, Nepal and Himalaya of Sikkim. However, elevation alone does not directly affect the Musk Deer's distribution. Instead, elevation is correlated with other climatic predictors like precipitation, temperature and solar radiations (Elith & Leathwick 2009) that lead to the change in habitat features and its quality to support the occurrence of the species.
In Api-Nampa Conservation Area, the slopes of 21-30º are highly preferred followed by slopes >40º by Musk Deer and avoid the slope of 0-10º (ANCA 2018). The study carried by Singh et al. (2018b) recorded the majority of latrines of Musk Deer in the slope of 20-40º in ACA. Our study in ACA coincides with these two studies as the principally preferred slope lie at 20-30º and completely avoid the slopes of 0-10º and >41º. Plain slope in our study was avoided due to presence of cattle grazing. Shrestha (2012) also suggested that Musk Deer avoid areas with high human disturbances like fuel wood collection and cattle grazing. And the slope >41º might have been avoided because of difficult terrain that resist them escaping from their predator.
Study carried out by Singh et al. (2018b) reported that Musk Deer prefer greater crown cover with high shrub diversity. In contrast to this, Musk Deer preferred moderate crown cover, i.e., 26-50 % in Api-Nampa Conservation Area (ANCA 2018), which is similar to our study. This is because the dense cover suppresses the growth of the ground level vegetation due to low light penetration, which might create the food shortage for the Musk Deer. This insight is supported by the study of Awasti et al. (2003) who recognized Musk Deer as the mixed feeder, i.e., grazers and browsers.
The thickness of ground cover governs the habitat preference of Musk Deer. The study carried out by Ilyas (2015) stated that Musk Deer prefer sparse ground cover. This study is supported by the study carried out in Api-Nampa Conservation Area where Musk Deer principally prefer the ground cover of 26-50 % (ANCA 2018), which is similar to our study in ACA. The dense ground cover is avoided; the reason could be that it is less friendly since it resists the rapid movement of Musk Deer that hinders to escape from predator. Singh et al. (2018b) reported that 69 % of the Musk Deer latrines were observed under tree, 26.4 % under canopy, and 4.6 % under rock. Similar to this study, forest and cave were found to be preferred and stream bed was found to be avoided in our study, which may be because the forest and caves are used for thermal requirements and escape whereas the streams are difficult to move across.
According to Khadka & James (2016), the Himalayan Musk Deer seems to utilize the region featured by presence of Pinus species and Abies species forest with moderately thick canopy cover (26-50 %) on higher elevation zone (≥ 3600 m) of the northern aspect. These choices are apparently social and structural adjustments (Futuyma & Moreno 1988). Musk Deer are shy and elusive creatures (Kattel 1993) with longer rear appendages compared to forelimbs, an adaptation for

J TT
living in rough terrain at high elevations. The domination of Abies species, which have dense crown cover, protects the area from snow, while the rivers flowing through the area serve as major water sources for Musk Deer throughout the year. Data on habitat parameters and their levels of preference recorded from different protected areas provide valuable baseline data, and offer the scope for determining micro-habitat for different species of Moschus in Nepal. Correlations in future when camera traps or molecular studies enable to have clear knowledge on the profile of species in each protected area.

CONCLUSION
The Musk Deer appear to have habitually utilized mixed and Rhododendron stands for defecation and foraging. Deer occurrence is sparse at lower elevations and higher elevations close to the tree line, and they are mostly distributed between 3,600 and 4,000 m. Thus altitudinal ranges of 3,800-4,000 m with mixed and Rhododendron woods adjacent to water sources are appropriate regions to execute conservation programs to protect Musk Deer and their environment. The likelihood of pellet presence diminished with the rise in ground elevation. A total of 15, 10 and 18 species of tree, shrub and herb were recorded, respectively, in the study area. The occurrence of Musk Deer was more around the forested area with crown cover of 26-50 %, and the tree species Abies spectablis, Betula utilis, Acer spp., Rododendron spp., Spruce spp., Taxus bacata, Honey suckle, Berberis spp. etc. The terrain with Pinus wallichiana, Cupresus spp. and Sorbus spp. appear to have been avoided. Likewise, the deer appear to have preferred areas where we have listed four species of shrub and nine species of herb, and further studies are required to assess the habitat suitability of the Musk Deer in response to climate change.