Despite its recognized importance, understanding of land–atmosphere (L–A) coupling remains fragmented, with most studies limited to regional analyses or single variables. A systematic global assessment of dominant drivers and their interactions is still lacking. This study utilized reanalysis data from 1958 to 2022, and employed SOM and K -means algorithms to detect the dominant spatial pattern of global L‒A coupling based on 12 relevant factors, thereby dividing the globe into five regions: Cold Snowy Region (CSR), Cold Wet-Soil Region (CWR), Temperate Continental Region (TCR), Hot Arid Region (HAR), and Hot Evaporative Region (HER). Building upon the initial pattern identification, a multi-scale temporal analysis further revealed that interannual signals dominate the variability of global L‒A coupling in most regions, whereas decadal signals remain comparatively weak—with the notable exception of pronounced decadal variability observed in HER during summer and autumn. Finally, the dominant factors of L‒A coupling in each region were identified using the Deep Forest model and SHAP algorithm, while key coupling processes were preliminarily understood through Process Network. The analysis indicates: radiative and surface energy factors dominate in CSR, eco-hydrothermal factors play a central role in CWR, multiple factors jointly regulate in TCR, water availability is the limiting factor in HAR, and temperature and humidity factors jointly influence coupling in HER. The combined importance of these dominant factors substantially exceeds that of the remaining variables, jointly accounting for no less than 40% of the overall variability. This study provides new perspectives and important references for understanding the complexity and regional variations of global L‒A coupling.
WU et al. (Sun,) studied this question.