Abstract Seasonal forecasting of precipitation variability over High Mountain Asia (HMA) remains a big challenge, as current dynamical models struggle to accurately capture its seasonal evolution. In this study, we identify key spatiotemporal patterns of rainy‐season precipitation anomalies across HMA. These patterns capture joint spatial and temporal signals of precipitation anomalies, rather than only spatial or temporal patterns individually. Consideration of the joint patterns provides a physical constraint that helps identify those more robust predictors, which as a result, are closely linked to the full seasonal evolution of precipitation anomalies. Based on these spatiotemporal patterns and their predictors reflecting slowly varying atmospheric boundary conditions, we develop a physical‐statistical prediction model that significantly improves predictions of both the detailed seasonal evolution and the rainy‐season mean of HMA precipitation anomalies. This advancement provides valuable insights for enhancing early warning capabilities and improving water resource management in vulnerable high mountain regions.

Read original article

Read original article