北京地区一次强降水过程的多种观测资料四维变分同化试验

采用MM5非静力预报模式及其伴随模式系统,对2005年汛期北京地区的一次局地强对流降水过程进行了包括地面每3小时一次的湿度观测资料、北京地区自动站逐小时降水资料、北京地区逐30分钟GPS可降水量观测在内的多种观测资料的四维变分同化试验。结果表明,采用四维变分方法同化多种非常规中尺度观测资料后,模式成功地模拟出此次降水过程,不仅很好地修正了变分同化窗内的降水预报,同时也使同化窗口以后的降水预报获得了较为明显的改善,并且发现同化了多种观测资料后获得的最优初始场较原模式初始场对流不稳定能量显著增加,具有更利于降水形成的动力及热力条件。另外还对各种观测资料在四维变分同化中的相对重要性进行了分析。

Four Dimensional Variational Data Assimilation of Multiple Types of Observations for a Severe Storm in Beijing Area

Several real-data assimilation experiments are performed for a severe convective storm occurring in the Beijing area,using the MM5 four-dimensional variational data assimilation(4DVAR) system with a full physics adjoin.Several kinds of data are assimilated,including surface moisture observations with 3-hr interval,local 1-hr accumulated AWS precipitation and ground-based GPS precipitable water observations with 30-minute interval.Assimilation results show that the MM5-4DVAR system is able to reproduce the observed rainfall in terms of precipitation pattern and amount.It is found that the model's forecast performance for precipitation are greatly improved not only during the assimilating window,but after that,especially when being compared with the 3-hr accumulated AWS precipitation observations.From the analysis innovates of the wind,moisture and temperature,the optimal initial conditions generated from the 4DVAR data assimilation of multiple observations may generate much more convective available potential energy,which is favorable to initiate convection.In addition,the significance of each type of data was tested.The surface moisture does not have significant impact on predicting the amount and pattern of precipitation.However,its assimilation was helpful to recover the vertical moisture.Ground-based GPS PW data had the maximum positive impact for model's performance.The assimilation of precipitation observations is important for improving precipitation forecasts,especially for that after the assimilating window.