可降水量资料反演及在暴雨数值模拟中的初步应用

The Retrieval of Precipitable Water and Its Application to the Numerical Simulation of Torrential Rain

通过同化可降水资料改善模式初始湿度场从而提高暴雨模拟与预报效果。首先应用适时回归方法反演大气可降水量资料,然后以1998年6月18—19日暴雨模拟为例,采用变分同化方法研究了可降水量资料在暴雨数值模拟中的作用。数值模拟结果表明:1)MM5模式中同化可降水量资料后,对于雨带位置的模拟效果起到很大的改善;亦可改进暴雨降水强度的模拟效果,改进程度与降水参数化方案有关。2)同化可降水量资料后,明显改善初始湿度场,在一定程度上缓解旋转加强(spinup)问题,从而有效地提高降水模拟效果。3)降水模拟对参数化方案的选取同样敏感,改进初始湿度场的同时应该选择相应的降水参数化方案。

Study shows that rain forecast is very sensitive to the initial humidity field.This paper aims at assimilating precipitable water(PW) data to improve the quality of model initial humidity fields,so as to improve the result of rain forecasting.At first,a regression method is applied to retrieve the PW,and then its effects on numerical simulation of torrential rain happened in June 18—19,1998 are assessed with comparasion experiments with and without the assimilation of retrieved PW.The results indicate that 1)the rain-belt position is closer to the observed after the assimilation of the PW data into the MM5,in addition the improvement in the intensity of torrential rainfall depends on cumulus parameterization schemes used;2)the effectiveness of precipitation simulation is enhanced because the improved initial humidity fields resolve the spin up problem in some sense;3)rain forecast is sensitive to cumulus parameterization schemes,and its essential to select corresponding cumulus parameterization scheme for different types of precitation processes when the initial humidity field is improved.