Abstract:  To improve forecast accuracy of precipitation phase state, the variation of sleet day was analyzed based on the daily meteorological data from 20:00 to 20:00 during January of 1951 to April of 2011 from the southern suburb observatory station of Beijing as a representative station. The cases of rain, sleet, snow were selected at 08:00 and 20:00 from October to April during 2000 to 2011, and the effects of conventionally meteorological elements in each layer on precipitation phase state were analyzed according to their corresponding ground and upper meteorological data, and the main influencing factors and the discrimination index of precipitation phase state were found out. A forecast equation of precipitation phase state was established by a multiple regression method. The equation and the corresponding discrimination index were tested. The results indicate that the averaged temperature, dew-point values from rain and snow at lower layer (from the ground to 850 hPa) differ significantly. Both differences decrease with the increase of height. The mean dew-point difference of temperature in each layer is not significant and hardly to use for distinguishing the precipitation phase state. By the test of precipitation phase state from November of 2011 to March of 2012, it shows that the discrimination index and the forecast equation could provide references for forecast of precipitation phase state.

Key words: Precipitation phase state, Discrimination indicator, Statistical equation, Verification