利用NCAR MM5V3对1999年6月长江流域的极端异常降水事件进行了模拟,主要研究不同水平和垂直分辨率对极端区域气候事件模拟的影响。数值模拟试验表明:模式能够模拟出极端强降水的主要分布特征;水平分辨率的提高降低了模式模拟的强降水偏差,对逐日降水变化的模拟更加合理,而垂直分辨率的提高基本上也都减小了模拟的强降水过程的偏差,改善对强降水的模拟能力;模式水平、垂直分辨率的提高在一定程度上增强了对强降水过程的模拟能力。水平分辨率的提高能够改善模式对海平面气压的模拟,而垂直分辨率的提高可以改善模式模拟的地面气温和低层环流。分辨率对中层大气环流的影响不是很敏感。不同积云对流参数化方案模拟的对流降水比率随水平分辨率的变化是不同的,Grell方案对流降水比例随分辨率的提高而增加,而Kain-Fritsch方案的结果相反。
The fifth-generation Pennsylvania State University-NCAR Mesoscale Model Version 3 (MM5V3) is used to simulate extreme heavy rain events over the Changjiang River Basin in June 1999, and the effects of model's horizontal and vertical resolution on the extreme climate events are investigated in detail. In principle, the model is able to characterize the spatial distribution of monthly heavy precipitation. The results indicate that the increase in horizontal resolution could reduce the bias of the modeled heavy rain and more reasonably simulate the change of daily precipitation during the study period. A finer vertical resolution leads to pronouncedly improved rainfall simulations with smaller biases, hence, better resolving heavy rain events. The increase in both horizontal and vertical resolution could produce better predictions of heavy rain events. In regard to the simulation of other meteorological variable fields, the increase in both horizontal and vertical resolutions raises to certain extent the ability of the model for simulating heavy rain processes. The increase in horizontal resolution is able to improve the simulation of sea level pressure, and the increase in vertical resolution can improve the simulation of surface temperature and lower level circulation, however, the simulation of the middle and upper tropospheric circulations is insensitive to the model's resolution. In addition, the effect of model's resolution on the ratio of convective precipitation to total rainfall differs when different cumulus parameterization schemes are used: the ratio of convective precipitation increases with increasing horizontal resolution in the Grell scheme experiments, by contraries deceases in the Kain-Fritsch scheme experiments.