A Hybrid Coupled Ocean-Atmosphere Model and ENSO Prediction Study

A hybrid coupled ocean－tmosphere model is designed, which consists of a global AGCM and a simple anomaly ocean model in the tropical Pacific。 Retroactive experimental predictions in-itiated in each season from 1979 to 1994 are performed. Analyses indicate that: (1) The overall predictive capability of this model for SSTA over the central-eastern tropical Pacific can reach one year, and the error is not larger than 0.8?C. (2) The prediction skill depends greatly on the season when forecasts start. However, the phenomenon of SPB (spring prediction barrier) is not found in the model. (3) The ensemble forecast method can effectively improve prediction results. A new initialization scheme is discussed.