Abstract: The three-dimensional variational data assimilation of radar reflectivity is studied using an indirect radar reflectivity assimilation scheme within the Weather Researeh and Forecasting model(WRF) three-dimensional data assimilation system(WRF 3D-Var).This scheme,instead of assimilating radar reflectivity directly, assimilates retrieved rainwater and water vapor derived from radar reflectivity.The impact of radar reflectivity from the radar network of the Beijing-Tianjin-Hebei area is assessed by conducting the numerical experiments using the operational Rapid Update and Cycling data assimilation and forecast system at the BMB.The four summertime connective cases, which produced severe rainfall in Beijing in June 2009,are studicd.The results show that both assimilating retrieved rainwater and derived water vapor can improve precipitation forecasts and assimilating derived water vapor play a more important role.On average, the assimilation of reflectivity data can significantly improve short-term precipitation forecast skill up to 6 hours,but the assimilation of radial velocity cannot make a stable improvement.The RUC mode is very important to short-term quantitative precipitation forecast (QPF) when radar data assimilated.