Understanding Multi-Scale and Multi-Species Habitat Selection by Mammals in the Eastern Himalayan Biodiversity Hotspot.

Human-induced habitat loss and fragmentation threaten biodiversity in the Eastern Himalayas, a crucial part of the Indo-Myanmar biodiversity hotspot. This study examines the distribution of 10 mammal species in Arunachal Pradesh using a multi-scale ensemble modeling approach, integrating Generalized Linear Models (GLM), Generalized Additive Models (GAM), and MaxEnt to assess habitat suitability. By analyzing 57 environmental predictor variables across multiple spatial scales, we found that elevation is a key determinant for carnivores such as the dhole and the Asiatic golden cat, while herbivores like the northern red muntjac and the mainland serow prefer broadleaf forests. Species distributions showed distinct patterns, with most carnivores concentrated in the south, except for the widely distributed yellow-throated marten. Dhole and leopard cat preferred elevated broadleaf forests, while the Asiatic golden cat favored mixed forests. Herbivores like the northern red muntjac and mainland serow were found at higher elevations, whereas the Indian wild pig preferred grasslands and degraded habitats near human settlements. While protected areas (PAs) exhibited higher species richness, significant suitable habitats also exist outside these regions, underscoring the need for landscape-level conservation strategies. Precipitation seasonality and human population density emerged as significant predictors, highlighting the influence of climatic and anthropogenic factors on habitat suitability. Our findings emphasize the necessity of conserving large, connected landscapes to mitigate human-induced pressures and climate change impacts on these species. By combining spatial modeling with ecological insights, this study provides a framework for prioritizing conservation efforts. Future research should expand data collection across broader temporal and geographic scales and incorporate climate change projections to anticipate species distribution shifts. These findings are critical for guiding effective conservation planning and habitat management in this ecologically rich yet vulnerable region.