Coupling Climate Downscaling With Species Distribution Models to Identify Potential Climate Refugia for Giant Panda Forage Bamboos

ABSTRACT

Climate change profoundly reshapes montane biodiversity patterns. A fine‐grained climatic resolution is indispensable to anticipate species‐specific habitat redistributions for effective conservation planning. Focusing on 20 giant panda forage bamboo (GPFB) species across 5 genera that constitute the bulk of the giant panda's ( Ailuropoda melanoleuca ) diet, we developed a “species‐distribution modelling based on climate‐delta downscaling” framework to project potential geographic shifts at a 30 m resolution across the 21st century under SSP1‐2.6, SSP2‐4.5, and SSP5‐8.5 scenarios. We found three primary results: (1) while the total climatically suitable area for GPFB expands under future scenarios, localized species richness within these habitats declines. (2) A mid‐elevation bamboo buffer zone (3000–4500 m) emerges as a critical thermal sanctuary where 7 out of 20 species expand under an unlimited dispersal assumption. However, incorporating a realistic biological dispersal constraint (10 m/year) severely truncates this elevational advantage, precipitating an overall habitat contraction for 11 species and reducing accessible refugia to merely 44%–65% of their theoretical potential. At localized scales, range shifts exhibit non‐monotonic trajectories and multidirectional movements rather than uniform upslope or northward migrations. (3) Integrating baseline and 12 future projections, we delineated stable macroclimate refugia governed by steep topographies, high hydrological availability, long‐term annual precipitation stability, and distinct seasonal environmental variations. 78.6% to 89.2% of the bamboo climate refuges overlap with the current (in 2025) forest distribution. Alarmingly, only 38.46%–44.11% of these crucial refugia fall within the current boundaries of the Giant Panda National Park, leaving extensive, contiguous refugia in the Liangshan mountain region largely unprotected. We recommend the establishment of a climate‐buffered conservation network that integrates extra‐park bamboo resources into unified spatial planning and deploys targeted connectivity corridors to mitigate dispersal bottlenecks. This 30 m workflow provides a high‐fidelity, transferable technical template for climate‐adaptation forecasting of threatened montane specialized diet species globally.