Quantifying the impact of vegetation changes on global terrestrial runoff using the Budyko framework

The global hydrological cycle is undergoing unprecedented changes with the intensifications of climate change and human activities. In particular, the impact of vegetation changes on the catchment hydrological processes has recently attracted widespread scientific attention. Although such impact in different regions of the globe has been extensively investigated, a comprehensive understanding of the impact across the entire globe remains elusive. In this study, we quantify the impact of vegetation change on surface runoff (Q) based on the Budyko framework over the global terrestrial environments (excluding Antarctica and northern high latitudes) during 1982-2014. Using long-term hydrometeorological observations and satellite-derived leaf area index (L) from 663 strictly selected catchments globally, we quantify the sensitivity of the Budyko parameter n to changes in L (S-n_L) at individual catchments. We find that S-n_L is closely correlated with catchment aridity index (phi) and soil water holding capacity (theta(p)), with higher (lower) S-n_L found in more arid (humid) catchments and catchments having smaller (larger) theta(p) values. Based on this, an empirical model relating S-n_L to phi and theta(p) is developed, which is further combined with the Budyko model to quantify the L and climate elasticities of Q and to attribute global Q changes over 1982-2014. Results show that the L elasticity of Q is higher and can be even greater than the climate elasticities of Q in relatively dry regions. Averaged over the globe, an increase in Q (+15.5%) induced by precipitation increases are largely offset by a decreased Q (-14.9%) caused by L increases over the last three decades. The vegetation greening-induced Q declines are manifested in global drylands. Over the study period, increases in L have led to Q reductions of 48.0%, 33.2% and 20.2% in global arid, semi-arid and sub-humid regions, respectively, which are larger than climate change-induced Q changes in these three regions. Our findings highlight the importance of vegetation change in regulating the long-term Q changes in relatively dry regions and call for cautions when conducting and/or planning future revegetation activities in these areas and needs for developing appropriate vegetation management policies to better inform water resources management in the context of global change.