Fig. 2: Soil bacterial α-diversity is shaped by abiotic environmental factors.

a Spearman’s rank correlation between environmental variables and the α-diversity of all taxa, each ecotype, and within each ecosystem. PIMP: percentage imperviousness. Positive and negative correlation is indicated by purple and orange, in the upper heatmap, and green and yellow, respectively, in the bottom heatmap. ****, ***, **, *, and ns denote adjusted two-sided P < 0.0001, 0.001, 0.01, 0.05, and not significant, respectively. b The top ten most predictive environmental variables for bacterial α-diversity (SHAP-based; X axis), sorted by descending importance. c Venn diagram of the variation partitioning analysis (VPA) showing the variation of the α-diversity of each ecotype and within each ecosystem explained by environmental factors. Residuals indicate unexplained variation. d Distance decay relationships for all taxa, ecotypes, and ecosystems. Bacterial composition dissimilarity is indicated by weighted UniFrac distance. ρ and P are Spearman’s rank correlation coefficient and P-value, respectively, and the colored lines and shaded areas depict the best-fit trend lines and the 95% confidence interval (mean ± 1.96 s.e.m.) of the linear regression, respectively.