The Role of Ocean Initialization in Improving the Decadal Prediction Skill of East Asian Summer Monsoon

Abstract

The East Asian summer monsoon (EASM) exhibits significant decadal variations and influences the summer rainfall in China, which has a great impact on the economic and social activities there. Thus, the accurate decadal prediction of the EASM has important scientific significance and application value. One of the key factors affecting the decadal prediction skills is initialization, which incorporates the observed internal variability into the initial condition (IC). However, few studies have pointed out the importance of ocean initialization in improving the EASM prediction skill. This study focuses on the decadal prediction of the EASM by FGOALS-g2, a coupled climate model, initialized by incorporating the oceanic analysis data using the dimension-reduced projection four-dimensional variational (DRP-4DVar) data assimilation method. High decadal prediction skill is obtained, which outperforms those achieved by most decadal prediction systems participated in the fifth and sixth phases of Coupled Model Intercomparison Project (CMIP5&6). Additionally, the spatial distribution of summer rainfall anomalies in China exhibits notable improvement compared to the uninitialized simulation. The high prediction skill is attributed to an accurate representation of land-sea thermal contrast between East Asia and the western Pacific as that represented in the NCEP-NCAR reanalysis data, which benefits from the accurate prediction of the Pacific Decadal Oscillation (PDO) due to realistic ocean IC obtained through ocean initialization. This study emphasises the crucial role of ocean initialization based on a DRP-4DVar method in enhancing the decadal prediction skill of the EASM.