Fig. 4.

Determinants of pine species diversity. (A–I) Relationships between predictor variables and species richness based on the global multipredictor model across grid cells of 100 km × 100 km. (J–R) The relative importance of four types of variables (climate, topography, soil, and landcover) in explaining species richness based on global, two subgenera, and six regional analyses. The bar plots were implemented in the “ggplot2” package (97) using R (v.3.6.2) (84). The positive and negative correlations are indicated as + and −, respectively. MCAC represents Mexico, Central America, and Caribbean. Abbreviations and explanations of predictor variables refer to SI Appendix, Table S10.