Seasonal prediction system using CFES and comparison with SINTEX-F2

In this study, we introduce a new seasonal prediction system using an atmosphere–ocean-coupled general circulation model called CFES (hereafter referred to as CFES ESPreSSO). We compare its prediction skill of the interannual variability of the surface air temperature (SAT) and precipitation anomalies with that of the SINTEX-F2 seasonal prediction system. We find that CFES ESPreSSO has a higher skill in predicting the SAT variability in January-February-March over East Asia and northeastern North America than SINTEX-F2, while the following season (April-May-June), SINTEX-F2 provides better predictions of the SAT variability over the Maritime Continent and subtropical North Pacific. Meanwhile, CFES better predicts the SAT variability in July-August-September over Eurasia and Arctic, and it continues to be so over the following season (October-November-December) over Eurasia. However, the prediction skill of SINTEX-F2 is generally better in the tropics (e.g., SAT in the subtropical North Pacific, SAT and precipitation in the Maritime Continent). Regarding climate indices, CFES shows a better prediction skill for the Atlantic Niño and Ningaloo Niño indices, whereas SINTEX-F2 is generally better for El Niño and the Indian Ocean dipole mode. These results suggest that for improved seasonal forecasting, it is beneficial to consider a multi-model approach, leveraging the respective strengths of each model.