Abstract: Predicting the spatial variation of suitable habitat of endangered plants under climate change in the future can provide theoretical guidance for protection and sustainable utilization of populations. Illicium difengpi, an endemic species in karst regions, is a China national secondary protected plant with considerable medicinal, ecological and ornamental value. With 30 geographical distribution points of I. difengpi and 10 environmental variables, we analyzed the potential distribution areas and dominant variables of I. difengpi under the current climate scenario using the Biomod2 ensemble model. The potential distribution area in the future (2050s, 2070s, and 2090s) under three climate scenarios (SSP126, SSP245, and SSP585) was simulated. The results showed that: (1) The ensemble model constructed with three optimal single models (Generalized Additive Models, Maximum Entropy Model, and Random Forest) showed better performance than single models in predicting the suitable areas of I. difengpi, with the true skill statistics (TSS) and the area under receiver operating characteristic curve (AUC) values of 0.890±0.024 and 0.982±0.007, respectively, indicating high prediction accuracy. (2) The mean diurnal range, temperature annual range, precipitation of warmest quarter, and elevation were the dominant variables affecting the distribution of I. difengpi. (3) Under the current climate scenario, the potential suitable area of I. difengpi was approximately 3.28×10⁴ km², concentrated in central, northwest and southwest of Guangxi. (4) Under future climate scenarios, the suitable distribution area of I. difengpi would initially expand and then contract, with the distribution centroid shifting towards the southeast. The results provide a reference for in situ conservation, ex situ conservation, introduction, and domestication of I. difengpi under the background of climate change.

Key words: Illicium difengpi, ensemble model, climate change, species distribution range