Chinese pangolin extinction-risk model
We successfully modeled the extinction risk of the Chinese pangolin across China with MaxEnt from 1970-2000 (Fig. 3). The average test AUC of 100 replicate runs was 0.935, and the standard deviation was 0.007. The prediction results had a satisfactory reference value. The population count gained a 68% contribution rate of the model: bio19 contributed 11%, bio4 contributed 6.9% and elevation contributed 5.3% (others included bio9 [3.1%], bio15 [2.1%], cropland [1.9%] and grazing [1.6%]). The extinction risk ranged from 0 to 0.93. The marginal response curve of a single variable indicated that as the population count increased, the extinction risk was elevated exponentially. When the population count reached approximately 7500 persons each grid (100 km2), the extinction risk remained almost unchanged (supersaturated status). The extinction risk first increased and then decreased with increasing precipitation in the warmest quarter and seasonal changes in temperature and elevation. The extinction risk model indicated that anthropogenic variables could be the principal causes of local extinctions of Chinese pangolins, followed by climatic variables.