The Role of Vegetation Growth Dynamics in Accelerating the Early Development of Flash Drought

Abstract

Flash droughts, characterized by rapid soil moisture depletion and severe ecosystem stress, are primarily driven by atmospheric anomalies. Vegetation can also modulate drought evolution, but its role in such rapid intensification remains largely unknown. Based on satellite‐derived Normalized Difference Vegetation Index (NDVI) data and multi‐source soil moisture data sets, this study quantifies the contribution of vegetation growth dynamics to the early‐stage onset speed of flash droughts at a global scale from 1982 to 2022 using a machine learning model interpreted with the Shapley Additive Explanations (SHAP) method. Our results demonstrate an overall positive contribution of vegetation growth to flash drought onset speed across most vegetated regions worldwide, with mean contribution rates exceeding 10% in several hotspot regions, including Eastern Asia, the Mediterranean, Northeastern Brazil, and Southern Asia. Furthermore, the influence of vegetation growth varies with background climate and vegetation conditions, with stronger impacts in relatively humid croplands and woody ecosystems. In these environments, the positive contribution can be largely attributed to enhanced evapotranspiration partitioning toward transpiration. These findings advance our understanding of the critical role of vegetation dynamics in accelerating flash drought development and indicate the necessity of incorporating dynamic vegetation representations into drought monitoring and early warning systems.