Abstract: GPS/PWV (Precipitable Water Vapor) data,sounding and surface data are assimilated into the WRF model together with a 3DVar assimilation system to investigate their effects on a regional rainstorm on May 10 and a local one on May 24,2015,respectively,over Anhui province.The results show that GPS/PWV data assimilated to WRF can improve the relative humidities in an initial field,but have little influence on wind.However,sounding and surface data assimilation can improve both relative humilities and wind,with a slightly worse result than that of GPS/PWV for the humidities.The forecast capacities of rainstorm locations and intensities may be enhanced with the assimilated GPS/PWV data only,superior to those with sounding and surface data.If all these data were assimilated to the WRF,the rainstorm locations and intensities can be significantly improved and closer to the observations,leading to higher TS scores and lower false-alarm rates.