中文摘要：

鉴于云南观测信息相对不足、局地强降水突出的现状,利用WRF（Weather Research and Forecasting）模式及其变分同化系统进行雷达反射率因子和反演风场的三维变分同化试验.通过对2012年9月12日00：0013日00：00发生在云南的一次强降水过程进行数值模拟和对比分析,结果表明,同时同化雷达反演风场和基本反射率因子,对区域模式同化系统中风矢量、相对湿度、位势高度几个基本分析量都有明显影响.雷达资料的同化,有利于区域模式初始场中强降水区域的上游中低层空气湿度增加、水汽输送增强和强降水发生区域的风场辐合加强,从而改善区域模式对强降水落区、强度的预报质量.对于切变线等天气尺度系统影响下的强降水过程,雷达资料的同化持续时间选取3h、同化间隔为1h较适宜.另外,雷达反演风场和基本反射率因子的同化均对降水预报改善有明显贡献,且多种资料的同化效果好于单一资料同化.

英文摘要：

In view of the lack of observations and high frequency of local torrential rain in Yunan province,basic reflectivity and Doppler Radar retrieved wind data are assimilated to WRF （Weather Research and Forecasting） model with WRF-3DVar （3-Dimensional Variational Data Assimilation System）.With different assimilation schemes,a torrential rain that occurred in the province from 0000UTC on 12 to 0000UTC on 13 September,2012 is numerically simulated and compared and analyzed.The results indicated that the assimilation can markedly improve the initial basic variables such as the vector,relative humidity,and geopotential height.The assimilation of the basic radar reflectivity data and the retrieved wind data help to increase upstream relative humidity on low levels,strengthen water vapor transportation and air convergence in the heavy rain region.Then the assimilation improves the forecast of torrential rain area and quantitative precipitation.The quantitative test of forecast rainfall shows that the scheme with 3-hour assimilation time and 1-hour assimilation frequency is preferable in the torrential rain case influenced by shear line.On the other hand,both the basic radar reflectivity data and the retrieved wind data contribute to a better rainfall forecast,and the accuracy with multiple data assimilation schemes at once is better than any single data assimilation scheme.