Enhancing Seasonal Predictability of Haihe River Basin Summer Precipitation Through ENSO and North Pacific–Atlantic Precursors

Abstract

Escalating flood risks highlight the need for a better understanding of the predictability of Haihe River Basin summer precipitation (HRBSP). Here, we investigate the distinct roles of ENSO-related and ENSO-independent sea surface temperature (SST) precursors in modulating interannual HRBSP variability. Using Extended Empirical Orthogonal Function analysis, we extract two modes of ENSO evolution, with the developing mode emerging as the dominant ENSO-related driver. After removing ENSO-related signals, we reveal a pronounced interdecadal shift in the ENSO-independent SST precursors around the early 2000s. Before 2003, HRBSP variability was associated with opposite-signed SST anomalies between the North Pacific and North Atlantic during the preceding autumn, whereas after 2003 it was primarily linked to a winter meridional SST gradient over the North Pacific. These two precursors exhibit distinct cross-seasonal propagation but are associated with similar summer Rossby wave trains affecting East Asia. The decadal shift of precursors is associated with weakened winter persistence of North Atlantic anomalies and altered North Pacific evolution tied to a phase transition of the Interdecadal Pacific Oscillation. Building on these insights, we develop an empirical prediction model for interannual HRBSP variability using only autumn–winter SST observations. In the context of a retrospectively identified regime framework, this empirical model achieves higher prediction skill than dynamical model forecasts issued over the same period and successfully predicted the extremely wet HRBSP anomaly in 2025. These findings provide new mechanistic understanding of HRBSP variability and highlight the potential of physically based SST precursors for improving seasonal precipitation prediction.