利用3维变分方法同化多普勒天气雷达资料的试验研究

Study on Doppler weather radar data assimilation via 3D-Var

以美国新近研发的天气研究预报模式（WRF）配置的3维变分同化系统WRF3D—Var为平台，针对2003年7月4-5日的一次暴雨过程进行了一组同化多普勒天气雷达径向风和回波强度资料的试验研究。其中同化径向风资料的方法是先用两步变分法由多普勒资料反演出水平风，再同化反演风场；同化雷达回波强度资料的方法是通过雷达回波强度与雨水混合比的经验关系以暖雨过程为约束的同化方法。试验结果表明同化雷达资料后对降水预报有显著的正作用。其中单独同化雷达回波强度资料的正作用在前3～6h效果非常显著；单独同化反演的水平风场对降水预报的正作用能持续12h；二者同时同化对提前9h的降水预报改进更加明显。

The horizontal winds retrieved from Doppler radar data by using a two-step variational method and Doppler reflectivity are assimilated with WRF 3 D-Var. A set of experiments assimilating Hefei radar data for the heavy rainfall case on 5 July 2003 are carried out. Results from the ease studies show positive impact of the Doppler radar data assimilation upon the precipitation forecast. The positive impact of Doppler reflectivity is mainly in the first 3-6 hours forecast. However, assimilation of retrieved wind can extend the positive impact on rainfall forecast up to 12h. Assimilation of both retrieved winds and reflectivity obtains the highest threat scores of 3-h rainfall simulations at the first 3 verification times for either threshold of 1 mm or 10 mm.