一次江淮梅雨锋暴雨的数值模拟可预报性研究

利用中尺度非静力MM5模式研究不同初始扰动（误差）对2003年7月4—5日发生在江淮流域的一次梅雨锋暴雨数值预报不确定性的影响,并着重分析了提前36h定量降水的可预报性。结果表明,利用常规观测资料和NCEP/NCAR分析资料形成初始场的控制试验能够提前36h做出较好的模拟。扰动温度场的敏感性试验表明,扰动温度的均方差愈大,降水预报不确定性也愈大。误差演变特征和增长机制分析表明,误差增长具有升尺度特征,误差首先在对流层低层和高层增长,然后大值区向对流层中层扩展;湿降水过程是对流层中低层误差增长的主要机制;对流层高层的误差增长是大气干动力与湿过程共同作用的结果,前期以干过程为主,后期以湿过程为主。

Mesoscale Predictability Study on a Mei-Yu Frontal Heavy Rainfall over Changjiang-Huaihe Basin

The predictability of a Mei-Yu frontal heavy rainfall over Changjiang-Huaihe basin on July 4-5,2003 is studied with a regional nonhydrostatic mesoscale model MMS. The paper focus on the quantitative precipitation forecast uncertainty out to a lead time of 36 h. The numerical results of the control run, with the initial fields form and the lateral boundary condition updated in a time interval of 6-h with the NCEP/NCAR 1° x 1 °reanalyses and the conventional observations, show that the heavy rainfall is well simulated. The results of sensitivity experiments with the temperature field perturbed in the initial condition show that the greater the standard deviation of temperature perturbation, the greater the precipitation forecast uncertainty. Analyses of error evolution and growth indicate that the error growth possessed an upscale character; the error increased at first in the upper and lower troposphere and the large value areas of errors extended towards the middle troposphere; the error growth mainly resulted from wet processes in the lower-middle troposphere, but from dry as well as wet processes in the upper troposphere. As far as the error growth in the entire process of the heavy rainfall is concerned, it mainly resulted from dry processes in the early stage, but from wet processes in the late stage.