Observed declining strength of vegetation-atmosphere coupling

Land-atmosphere coupling (LAC) directly influences the occurrence of extreme climate events. Traditionally, the studies of LAC strength have primarily used soil moisture as a proxy for land conditions. However, recent research has highlighted the significant role of vegetation–atmosphere coupling (VC) in the evolution of extreme climate events through its regulation of the water and energy cycles. Despite this progress, the global patterns and driving mechanisms of VC remain unclear. In this study, the index with a clear physical meaning, ω, defined as the relationship between the canopy conductance (gc) and aerodynamic conductance (ga), was introduced to represent VC values. Long-term (1981–2018) global annual VC values were derived using two high-quality reanalysis datasets (ERA5 and MERRA2) based on two different gc models. Both gc models exhibited similar spatial distributions that the highest VC values in Arid regions, the lowest in Humid regions, and intermediate values in Transition zones. Results showed 38.84–61.98 % of global land with decreasing VC trend. An attribution analysis using a nonlinear machine learning approach revealed that leaf area index (LAI) and wind speed dominated the VC changes across different climate zones. An increase in LAI reduced VC strength, whereas enhanced wind speed increased VC values. LAI was the dominant factor influencing VC through transpiration regulation (i.e., gc) over Transition and Arid regions, while wind speed controlled VC variations via ga over Humid regions. Our study analyzed the spatiotemporal changes in VC values and their driving mechanisms across global land areas. These findings contribute to a deeper understanding of vegetation-climate feedback and its role in amplifying extreme climate events.