不同同化方案在台风“白鹿”强降水预报中的应用  

The application of different data assimilation systems in Typhoon“Bailu”rainstorm forecast

作　　者：梁钧怡 王东海 张宇 姚乐宝 LIANG Junyi;WANG Donghai;ZHANG Yu;YAO Lebao(School of Atmospheric Sciences/Guangdong Province Key Laboratory for Climate Change and Natural Disaster Studies,Sun Yat-sen University,Zhuhai 519082,China;Southern Laboratory of Ocean Science and Engineering(Zhuhai),Zhuhai 519082,China;Meteorological Bureau of Foshan,Foshan 528000,China;South China Sea Institute of Marine Meteorology,Guangdong Ocean University,Zhanjiang 524088,China;Inner Mongolia Autonomous Region Meteorological Centre,Hohhot 010051,China)

机构地区：[1]中山大学大气科学学院/广东省气候变化与自然灾害研究重点实验室,广东珠海519082 [2]南方海洋科学与工程广东省实验室(珠海),广东珠海519082 [3]佛山市气象局,广东佛山528000 [4]广东海洋大学南海海洋气象研究院,广东湛江524088 [5]内蒙古自治区气象台,内蒙古呼和浩特010051

出　　处：《中山大学学报（自然科学版）（中英文）》2022年第4期104-118,共15页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家重点研发计划项目(2019YFC1510400);国家自然科学基金(41861164027,41775097);广东省科技计划(2017B020244002)。

摘　　要：基于WRF模式和GSI同化系统,构建了一套同化预报系统,研究常见的两种同化方法3DVar(threedimensional variational)和3DEnVar(three-dimensional ensemble-variational)的应用,开展不同的嵌套区域采用不同的同化方法对模式预报效果的影响研究。基于本系统,分别设计三组试验,评估不同的同化方案对2019年8月第11号台风“白鹿”在华南地区造成强降水的模式预报和分析同化效果。个例试验结果表明:在双层嵌套模式中,两层均采用3DVar同化方法的试验对“白鹿”登陆前后所带来的小雨、中雨量级降水落区的模拟存在优势;而两层均采用3DEnVar同化方法的试验则对强降水落区和强度预报把握度更好,特别是对暴雨、大暴雨等强降水事件中心位置的模拟与实况更接近,且降水评分最优。此外,3DEnVar同化方法所构建的背景误差协方差具有明显的流依赖特性,并且对形成强降水的相关气象要素场包括垂直风场、温度场和湿度场的预报中均有较好的表现。The WRF model and the GSI assimilation system were used to build an assimilation prediction system regarding problems of data application and forecasting effect between the 3DVar data assimilation and the 3DEnVar(Hybrid)data assimilation.Based on this system,three tests of comparative experiments were designed to perform simulation analysis and assessment of these two data assimilation methods on a case of heavy rain process caused by Typhoon"Bailu"in August 2019.The results show that:1)In the scheme of using 3DVar assimilation with two nested domains,there is an advantage in simulating the approximate distribution of rainfall;and it can capture the location of the light rain and moderate rain.2)Using 3DEnVar(Hybrid)assimilation with two nested domains can effectively improve the accuracy of the center position of typhoon rainstorms and the forecasted positions are closer to the actual situation.3)The 3DEnVar(Hybrid)assimilation has obvious flow dependence on the background field;meanwhile,the simulations of related tactics such as wind vector,temperature,and humidity that form heavy rainfall are more accurate.

关 键 词：三维变分同化 三维集合变分混合同化 WRF-GSI预报系统 台风降水 台风“白鹿” 

分 类 号：P445[天文地球—大气科学及气象学]