Abstract:Doppler radar reflectivity and radial velocity data are added directly to the mesoscale model ARPS (The Advanced Regional Prediction System) in numerical simulation by its data processing system ADAS(ARPS Data Analysis System). The control experiment, reflectivity assimilated experiment, velocity assimilated experiment, simultaneous assimilation of the two kinds of data, and continuous assimilion experiment at multiple times are included in this study. By simulating one rainstorm process around Jiangsu Province, we analyzed the impact of improving the initial field and of getting better forecast results due to the two kinds of radar data added in differernt combinations, and came to the following conclusions: reflectivity and radial velocity data mainly separately adjust the moisture and fields, both the two kinds of radar data assimilated performed better than one kind of radar data added within 2 h. Compared with other assimilative experiments, the predicted wind speed from the multiple times'assimilating experiment with two kinds of radar data was more close to the sounding wind speed, and the predicted rainfall area from this experiment had better relationship with radar product for 3h precipitation, but there was a considerable error of precipitation amount.