Decadal prediction of East Asian winter monsoon with CMIP6 models: explainable predictability from the western North Pacific sea surface temperature

Abstract The East Asian winter monsoon (EAWM) profoundly impacts the societal and ecological systems, making its prediction of great significance. This study assesses the decadal prediction skill of the EAWM using the Coupled Model Intercomparison Project Phase 6 (CMIP6) Decadal Climate Prediction Project (DCPP) simulations. The evaluations demonstrate that the anomaly correlation coefficient (ACC) between the observed and predicted EAWM indices exceeds 0.60 in both the forecast lead year-2-5 (FY2-5) and year-6-9 (FY6-9), significant at the 95% confidence level, suggesting the corresponding skillful predictions. The analysis from the multi-member ensemble indicates that the decadal predictability of the EAWM could be possibly related to the skillful prediction of the western North Pacific sea surface temperature (NPSST). Notably, the higher the ACCs of the western NPSST anomaly, the more skillful prediction of the decadal EAWM. Further, the better and worse member ensembles are compared, which further indicates that the improvement of the EAWM decadal prediction is sensitive to the prediction skill of the western NPSST. The better member ensemble achieves higher ACCs (exceeding 0.70) for the EAWM index and reasonably reproduces the observed decadal weakening of the EAWM since the mid-1980s. In addition, the better member ensemble exhibits higher prediction skill for the East Asian winter temperature than the worse ones. Such results in this study indicate that western NPSST anomaly could serve as a crucial source for EAWM decadal prediction, providing a scientific basis to improve the East Asian winter climate prediction.