Ochotona nubrica
Univariate models - Comparison of univariate models at different scales prior to model building revealed that more relative percentage of bush cover, more number of shrubs, a greater height of shrubs, and large mature stands of tall grass and reeds were significant for predicting the presence of ON across all scales (Table C2). At the larger Ladakh scale, variables related to rocks (negatively) and NDVI across seasons (positively) were important in explaining the presence ofON (Table C2). In the North-West, presence was associated with NDVI (positively) and reflected greener habitats that the species occupies. Details of univariate models run can be found in Supplementary material (Table S6, Table S7, Table S8).
Multivariate models - At the Ladakh scale, the purposeful model (Model 2) produced the best results; the sub-modeling approach (Model 3) was best for the South-East, while at the West-Ladakh scale, all three modeling approaches produced equally good results (Table 2). At the Ladakh spatial scale, variables that predicted the absence of the species included cover and number of large-sized rocks, and variables that predicted the presence of the species, such as the cover of shrubs and NDVI, were important (Table 2). In the South-East, only variables related to vegetation and mainly shrubs were important in explaining the presence of the species (Table 2). In the North-West, different NDVI variables (for different modeling approaches) and mature stands of tall grass and reeds were necessary to explain the species’ presence, although the species was restricted to large mature stands of Sea buckthorn (Table 2).