Two numerical experiments of ensemble hindcasting for 17 years are performed here, initial conditions are NCEP-DOE Reanalysis 2 data (adopted in the real-time prediction )and that in the second are supplied by the system of three-dimensional variational assimilation (3D-VAR),and there are five-member ensembles in each experiment. The model is used here which is a grid-point Atmospheric General Circulation Model with 9 vertical levels and a resolution of 2 0 in latitudinal and 2.5 0 in longitudinal direction, developed by the Institute of Atmospheric Physics, the Chinese Academy of Sciences (IAP9L2x2.
5-AGCM). The technique of Correlation Analysis is applied. Comparison results show that (1) the difference of temporal anomaly coefficient is very small in the tropics under two different initial conditions, whereas it is large in middle and high latitudes, especially in East Asia (China) region, (2) temporal anomaly coefficient in the experiment 2 is bigger than that in the experiment 1 in most of China region, (3) spatial anomaly correlation of many variables has also different degree improvement by use of the initial conditions after data assimilation(in the experiment 2).That is to say, the results demonstrate quantitatively that the role of initial conditions on short-term climate prediction in East Asia is more important than in the global, what's more, the initial conditions after data assimilation are better than that in the real-time prediction to some extend. So, the effect of initial conditions on atmospheric general circulation model is not neglected.