Abstract: The one dimensional variational data assimilation system (1D var), with ability of assimilating the rain rate retrieved from radar reflectivity, is established by utilizing two linearized physical parameterizations of large-scale condensation and convection as the observational operator. A realistic water vapor profile, obtained by 1D var, is assimilated by the three dimensional data assimilation system (3D var) to produce a better quality of water vapor field for high resolution models. To preliminarily estimate the impact of the 1D+3D var assimilation of the rain rate upon short-term precipitation forecast, a case study of the severe rainfall occurred on 9 June 2010 and one month experiment (May 2011) are performed, by applying the “1D+3D var” to get the realistic water vapor profile and a better initial condition for fine resolution models.