FIGURE 7  Distribution pattern of potential species richness of 25 rare, endangered and/or national protected key higher plant species in Northwest Yunnan

3.5 Environmental explanations for potential distribution of species

According to the category of environmental factors, we employed the package ‘spgwr’ of the software R and constructed six different types of environmental factor models to explain the potential distribution of species through the GWR model (Table 4). The research results indicated that the human activity model had a lower interpretation rate than other models, which appeared a goodness of fit R 2 and AIC value of 0.83 and 8,850.15, respectively. However, the goodness of fit R 2 and AIC value of the water model were 0.88 and 7,703.82, respectively, which gave a higher explanation rate than other models, indicating that the water factor largely influenced the potential distribution of species. In addition, the goodness of fitR 2 and AIC value of the comprehensive model were 0.92 and 6,785.26, respectively, and the difference in the AIC value (∆AIC) with other models was greater than three, demonstrating that the comprehensive model had a better simulation effect on the potential distribution of species. The results also revealed that multiple environmental variables were not mutually exclusive, and the potential distribution of species was the result of the combined effects of various environmental factors.
TABLE 4  Fitting effects of different models on the potential distribution of species