基于WRF模式的清江流域降雨-径流模拟研究

RAINFALL-RUNOFF SIMULATION OF QINGJIANG RIVER BASIN BASED ON WRF MODEL

基于WRF模式,采用4层嵌套方案,选取3种积云参数化方案和7种微物理方案组成21种组合,对清江流域2016—2018年6—10月6次典型降雨事件进行数值预报,结合CMORPH卫星-地面自动站-雷达三源融合降水产品,采用TS评分和FSS评分,分析不同分辨率和云微物理方案的降雨预报效果;基于较优组合方案的WRF模式与WRF-Hydro水文模式耦合进行径流模拟,分析WRF模式在水文模拟中的应用效果。结果表明:3 km和1 km分辨率对降雨中心位置及强度预报的差别不大,对降雨落区都有较好的预报能力;在积云参数化方案中,KF方案和BMJ方案的降雨预报效果优于GF方案;在微物理方案中,WSM3、WSM5、WSM6、Thompson方案的预报结果与融合数据有较好的一致性;基于较优组合方案BMJ_WSM3,将WRF模式与WRF-Hydro模式耦合,耦合模式能较好地模拟洪水过程,径流模拟相关系数都在0.67以上,且NSE最高可达0.79。

In this study,based on the WRF model,a four-layer nested scheme with 3 cumulus parameterization schemes and 7 microphysical schemes selected to form 21 combinations was used to forecast 6 typical rainfall events in the Qingjiang River basin from June to October in 2016—2018,combined with high-resolution radar-satellite-gauge merged precipitation products,TS scoring and FSS scoring were used to analyze the rainfall forecasting effects of different resolutions and parameterization schemes;the WRF model with a better combination scheme was coupled with the WRF-Hydro model to conduct runoff simulation to verify the application effects of the WRF model in hydrological forecasting.The results show that the difference between the 3km and 1km resolutions is not significant for predicting the location and intensity of rainfall centers,and both have good forecasting ability for rainfall areas;Among the cumulus parameterization schemes,the KF scheme and BMJ scheme have better rainfall forecasting results than the GF scheme;Among the microphysical schemes,the forecasts of WSM3,WSM5,WSM6,and Thompson schemes are in good agreement with the merged precipitation products.Based on the better combination scheme BMJ_WSM3,the WRF model is coupled with the WRF-Hydro model,the coupled model can simulate the flood process well,and the correlation coefficients of runoff simulation are all above 0.67,and the NSE can reach up to 0.79.