Non-local impacts of upwind vegetation on soil moisture across South America

Soil moisture variability and drought severity in South America are increasingly pressing challenges, driven by global climate change and extensive land use change. In particular, the biophysical effects of vegetation not only influence local water availability, but also have non-local impacts through atmospheric moisture transport. Understanding how upwind vegetation dynamics affect downwind soil moisture anomalies (SMA) is critical to addressing these challenges. In this study, we investigate the role of upwind vegetation in modulating SMA from 2001 to 2018 using a deep learning framework. We identified a pronounced sensitivity of downwind SMA to Amazonian vegetation, with water transport dominating during more than half of the drought events. Hotspots in the eastern Amazon were found where increased vegetation could significantly enhance atmospheric moisture supply to downwind regions, thereby buffering soil moisture variability in Brazilian agricultural zones. Overall, our results highlight the critical role of atmospheric moisture transport in shaping regional hydrology and emphasize the interconnectedness of land use change and hydrological processes. By integrating vegetation dynamics and non-local moisture transport into hydrological and land management strategies, this research provides actionable insights for improving drought resilience and managing the hydrological impacts of vegetation in a changing climate.