Extended Data Fig. 2: Climate control on the vegetation-energy interplay.

(a, c, e) Sensitivity of latent heat \(\left(\frac{{\partial LE}}{{\partial LAI}}\right)\), sensible heat \(\left(\frac{{\partial H}}{{\partial LAI}}\right)\) and Bowen ratio \(\left(\frac{{\partial B}}{{\partial LAI}}\right)\) to LAI changes (on the y-axis) separately computed for the 1982–1999 and 2000–2016 periods and binned as a function of the aridity index (on the x-axis) (Eq. (3), Methods). Results of the Kolmogorov-Smirnov test is shown in label and reflects the significance level (p_ks) to reject the null hypothesis of dissimilar curves. (b, d, f) Temporal variations of sensitivities extrapolated from changes in precipitation and climate for moisture supply- and atmospheric demand-limited regions and the whole globe displayed with respect to the first year sensitivity (year 1982). Numbers at the bottom of the panel report the sensitivity values of the first year, while numbers on top report the coefficient of regression (R²) and the significance (Mann-Kendall test; p_mk) of the fitting linear regression models. The spatial domains of supply- and demand-limited regions are shown in Supplementary Fig. 5.