2.2 Environmental Variables
We gathered environmental variables associated with bioclimatic, soil, and topographic factors as potential predictors of species distribution (Table 1). 19 climate variables spanning the last glacial, last glacial maximum, Mid-Holocene, current periods, and future scenarios were sourced from the World Climate Database (http://worldclim.org) (Gao et al., 2018). Future scenarios included low-concentration emissions (SSPs1-2.6) and high-concentration emissions (SSPs5-8.5) of greenhouse gases. 16 soil variables pertaining to the soil surface were acquired from the Chinese Soil Dataset within the Harmonized World Soil Database (HWSD, http://www.fao.org/faostat/en/#data), while elevation data were obtained from the same source. The spatial resolution was set at 2.5 m. Map data were represented in SHP format based on a 1:1 million scale Chinese map obtained from the National Center for Basic Geographic Information.
To mitigate multicollinearity and potential model overfitting (Graham, 2003), we conducted Spearman’s correlation analysis within ArcGIS (Yang et al., 2013) to examine relationships among environmental factors. Variables demonstrating a correlation coefficient of ≥ 0.8 were considered highly correlated, and the less influential factor was excluded from subsequent analysis. Consequently, a total of 16 environmental factors were retained for calculation and analysis within the MaxEnt model.
Climate and elevation variable data were clipped according to the vectograph of a 1:1 million scale Chinese administrative map and then converted to ASC format using ArcGIS software. Soil variable data were integrated by importing the China soil file and HWSD DATA file into ArcGIS software. Subsequently, the grid layer comprising the 16 soil variables within the MU_GLOBAL layer was extracted and converted into ASC format. Finally, all environmental variable layers underwent batch processing using ArcGIS software, resulting in environment layers with non-overlapping extents.