Abstract:The forecast skill of summer ranifall over Huaihe River Basin (HRB) was evaluated based on the 30-year ensemble hindcasting experiment from 1981 to 2010 of IAP Atmospheric General Circulation Model Version 4.1 (IAP AGCM4.1). The results show that the IAP AGCM4.1 can generally reproduce the observed spatial distribution of summer-averaged rainfall over the HRB, the spatial correlation coefficient between precipitation and observation simulated by the model reaches 0.93 for June. However, systematic bias can be found in rainfall intensity and observation, and the rainfall interannual variability simulated by the model is also significantly weaker than the observation. The deterministic predictive skill of summer rainfall over HRB by IAP AGCM4.1, in terms of anomaly correlation coefficient (ACC), was found to be relatively higher in Southwestern HRB, where the ACC is larger than 0.2. The deterministic rainfall forecast skill for the whole HRB is the highest for June, and July is the second. For the mean areal rainfall of main sub-catchments in Huaihe River, IAP AGCM4.1 does show certain forecast skill, with ACC reaching 0.11 for Bengbu drainage basin, 0.13 for Lutaizi drainage basin, and 0.16 for Wangjiaba drainage basin, respectively. Based on the probabilistic forecast skill using rainfall, it's found that the IAP AGCM4.1 shows a high skill in forecasting the below-normal rainfall events occurred in Southern part of HRB in July, and the above-normal rainfall events in the central part of HRB in June.