Species Distribution Modeling of High-Altitude Mammals

Abstract

High-altitude mammal communities -- comprising charismatic megafauna such as snow leopards, Himalayan wolves, yaks, Tibetan antelopes, pikas, and mountain ungulates -- face disproportionate climate vulnerability because their thermal niches are compressed between the upper limit of physiological tolerance and the physical summit ceiling, leaving upward range shift as the only available response to warming and foreclosing when terrain runs out. Despite this recognised vulnerability, the spatial distribution models required to quantify projected range loss and identify conservation priority areas for high-altitude mammals remain incomplete for most Himalayan, Tibetan, Andes, and Central Asian alpine species. This study presents the largest ensemble species distribution modelling analysis for high-altitude mammals yet conducted, fitting MaxEnt + BRT + GLM ensemble models for 84 species from 24 families across four major mountain systems (Himalaya-Tibet, Central Asian ranges, Ethiopian Highlands, and Andes) using 184,700 quality-filtered occurrence records. Current suitable habitat averaged 84,700 +- 38,400 km2 per species. Under RCP 4.5 (2070), mean projected habitat loss was -34.7% +- 14.7%; under RCP 8.5, -57.4% +- 18.4%. Summit-trap species -- those whose upper range limit is within 200 m of local summit elevations -- show 2.84-fold greater projected habitat loss than species with upward dispersal capacity. Ensemble models identified 247 climate refugia -- areas retaining > 50% of current suitability under both RCP 4.5 and RCP 8.5 by 2070 -- concentrated in the eastern Himalayan massif, Tibetan Plateau interior, and tropical Andes high-altitude zones. Only 38.4% of these refugia overlap with existing protected areas, leaving 61.6% as unprotected climate refugia requiring priority conservation designation.