Abstract
Surface soil moisture-precipitation (SSM-P) coupling involves complex processes, with sensible heat (SH) and evapotranspiration as important mediators. However, these coupling pathways and their underlying mechanisms across the globe remain unclear, limiting hydrometeorological predictions and projections. Here, we employ an information flow technique to satellite observations and reanalysis, revealing strong local SSM impacts on precipitation across ~16% of analyzed global land. Among the eight identified coupling hotspots, the SH-mediated pathway emerges as a crucial mechanism, except for two African hotspots dominated by the evapotranspiration-mediated pathway. These pathway differences are linked to remote moisture availability and boundary layer height variability. Strong coupling preferentially occurs over regions with large SSM variability, particularly for SSM-SH-P. Most CMIP6 models fail to reproduce these coupling patterns, with only four successfully capturing the ERA5-derived variability-causality relationship. Our study offers additional insights into land-atmosphere coupling and proposes process-based metrics for model evaluation.