Fig. 3: Regional mechanisms underlying the latitudinal divergence in runoff responses due to forest-atmosphere feedbacks of forestation.
From: Latitudinal divergence in runoff responses to global forestation due to forest-atmosphere feedbacks
a, b Geographical patterns of runoff changes associated with precipitation changes (\(\triangle {R}_{P}\); a) and PET changes (\(\triangle {R}_{{PET}}\); b), respectively. Shown are the mean results of the two Penman-based PET algorithms. The insets display the bar plots of global mean of \(\triangle {R}_{P}\) and \(\triangle {R}_{{PET}}\) based on the mean result of two PET methods, respectively. Numbers show the percent change in runoff relative to CTL-simulated runoff. c Regional annual mean of runoff relative changes due to forest-atmosphere feedbacks (\(\triangle {R}_{{atm}}\)), precipitation changes (\(\triangle {R}_{P}\)) and PET changes (\(\triangle {R}_{{PET}}\)) in both warm (> 12.5 °C) and cold (< 12.5 °C) regions. The bar plots are the mean results of the two Penman-based PET algorithms. d Regional components of precipitation relative changes associated with four physical processes in both warm and cold regions. The four main processes are thermodynamic effect (ΔWV), altered horizontal advection (ΔAdv), altered vertical transport (ΔVert) and land ET changes (ΔET). The residual (Res) represents differences between precipitation changes (ΔP) and the combined contribution from the four processes, reflecting the impacts of transient eddies. e Regional annual mean of forestation-induced relative changes in PET (ΔPET), and relative changes in its two main drivers, including surface available energy (Δ(Rn−G)), atmospheric vapour pressure deficit (ΔVPD), respectively. The bar plots of ΔPET are the mean results of two Penman-based PET algorithms. Estimates by Penman-OW and PM-RC in (a–c) and (e) are shown as circles and triangles, respectively. Error bars in (d) and (e) indicate the 1-standard deviation of the last 60-year series. ** p < 0.05; * p < 0.1.
