Fig. 4: Summary of climate effects on masting, derived from the GJAM model that included coefficient of variation (CV) and temporal autocorrelation (AR1) as responses (n = 517 species).

a Boxplots show the marginal posterior distributions of the GJAM-derived coefficients. Specifically, boxes show mean effect size as vertical lines and are bounded by 80% credible intervals (CI), with 95% CI as whiskers. Colors highlight signs of the correlation (green for positive and purple for negative), with opacity increasing from 80% to 95% of the distribution outside of zero. Gray indicates coefficients that overlap zero. b Effects of mean annual temperature (MAP, in ^∘C) and mean annual precipitation (MAT, in cm) on CV and AR1. The surface shows the conditional relationship between CV/AR1 and MAT across levels of MAP. Convex hull is defined by species observations (red dots). MAT and MAP are defined for each species' distribution derived from the Global Biodiversity Information Facility (GBIF, www.gbif.org). Biplots of relationships between CV/AR1 and MAT and MAP are in Fig. S6. Climate effects on functional traits are in Fig. S7.