复杂地形条件下WRF模式的适用性：以滇西地区降水模拟为例
Applicability of the WRF Model in Complex Topographic Conditions: A Case Study on Precipitation Simulation in the Western Yunnan Region
中文摘要：
使用WRF模式和GSI同化系统对2014年7月13日云南省西部地区的一次天气过程进行控制模拟和同化模拟试验，探讨WRF模式在复杂地形条件下的适用性，得到以下结论：控制试验和同化试验都能模拟出此次降水过程中云南雨带的位置，但强度偏大。同化试验中雨带的降水强度有所减弱，但位置略有南移。控制试验和同化试验的强降水与TRMM相比，表现为开始时间晚，持续时间长。控制试验的强降水落区与TRMM相似，同化试验的强降水落区比TRMM的强降水落区偏西。同化探空数据能够显著减小模式降水模拟中各站点的24 h累积降水量及其RMSD，改进此次降水模拟中站点累积降水量的误差。控制试验对于此次过程的小时降水量的模拟值普遍偏大，同化试验模拟的各域的站点小时降水量相应减小，并且对于单峰型的日降水类型具有一定的模拟能力，峰值与实况较为接近。但是，对于多峰值降水区，以及无降水或者弱降水地区，控制试验和同化试验均有较大误差。WRF模式对于滇西复杂地形区的降水具有一定模拟能力，耦合GSI同化系统后，能够对模式输入场各变量分布进行调整，从而影响模式模拟性能，但总体模拟效果还有待进一步提高。
中文关键词：
WRF模式，GSI，探空资料，TRMM资料，复杂地形，降水
KeyWords：
WRF model, GSI, sounding data, TRMM, complex terrain, precipitation
Abstract：
This study investigated the applicability of the WRF model in complex topographic conditions by modeling a weather incident that occurred in the west Yunnan region on July 13, 2014, using the WRF model and GSI assimilation system. The results showed that the location of the precipitation belt could be predicted by both the control model and assimilation model, although the intensity predicted by the control model was a little larger. The intensity of precipitation predicted by the assimilation model was somewhat weaker and the location of precipitation moved southward. In comparison to the results from TRMM, the precipitation predicted by the control model and assimilation model started late and had a longer duration. The precipitation region predicted by the control model was similar to the results from TRMM, while that predicted by the assimilation model was a little westward. In combining with the assimilated sounding data, the model signifcantly decreased the 24h-accumulated rainfall and RMSD of each site. The 1.0-h rainfall at each site simulated by the control model was larger, while the 1.0-h rainfall of each site simulated by the assimilation model was relatively small. This suggests that the assimilation model had the ability to predict single-peak precipitation. The simulation results for both the control and assimilation models for the area with multi-peak rain, and no or weak rain, had large errors. The WRF model has the ability to simulate precipitation incidents in complex terrain in the West Yunnan region. Coupling the model with GSI, it could adjust the distribution of each variable of the mode input feld to in?uence the model simulation performance, however, the overall simulation results need to be further improved.