Abstract
Climate change may alter mountain bamboo distributions and giant panda food resources. We modeled potential habitat suitability for Bashania faberi, Yushania brevipaniculata, and Chimonobambusa szechuanensis in a defined Qionglai Mountains study area in Sichuan, China, using MaxEnt, ten bootstrap replicates, and three global climate models (GCMs) for the 2050s and 2070s under SSP1-2.6 and SSP5-8.5. We tested sensitivity to predictors, thresholds, spatial grain, connectivity, and future climate products. High omission in some spatial folds and 15 records make B. faberi projections exploratory and its patch metrics indicative. At species-specific 10th-percentile (10P) thresholds, the nine-variable ensemble projected gains for B. faberi and Y. brevipaniculata only under SSP5-8.5 in the 2070s; the five-variable configuration and fixed thresholds reversed both. Each expansion was supported by one GCM, while two projected contraction. Chimonobambusa szechuanensis contracted under the nine-variable and elevation-containing configurations but increased in the temperature-inclusive diagnostic at 10P. The combined suitable-area proportion fell from 34.20% currently to 21.00–31.05% in baseline future maps; mean patch area and aggregation also declined. Averaging outcomes could produce a different change direction from averaging predictions. These projections describe conditional suitability, not food supply or functional connectivity.