江西及其临近区域同化系统中不同季节背景误差协方差特征及其对降水预报影响  

Covariance Characteristics of Different Seasonal Background Errorsin the Assimilation System for Jiangxi and the Neighboring Regionsand Its Influence on Precipitation Forecast

作　　者：夏雪 金勇根 段和平 李嘉 Xia Xue;Jin Yonggen;Duan Heping;Li Jia(Jiangxi Provincial Meteorological Disaster Prevention Technology Center,Nanchang 330046,China)

机构地区：[1]江西省气象灾害防御技术中心,江西南昌330046

出　　处：《气象与减灾研究》2020年第4期241-248,共8页Meteorology and Disaster Reduction Research

基　　金：2017年度江西省气象局青年人才培养项目“背景误差协方差平衡约束在江西省暴雨预报中的适用性研究”.

摘　　要：在不同季节背景下,对比分析多元变量相关的背景误差协方差特征,了解其在天气过程中的作用,可以改进同化系统性能,提高降水数值预报水平。对比分析汛期和非汛期江西及其临近区域多元变量相关的背景误差协方差特征,分别选取2019年汛期和非汛期的2次降水过程,设计2组循环同化试验,探讨多元变量相关的背景误差协方差对江西降水预报效果的影响。结果表明,相较于非汛期,汛期分析变量对风场、温度场和水汽场的贡献值整体上要更大,且汛期各个控制变量的背景场误差更为显著。多元变量试验的降水预报评估效果整体上更好,其汛期预报效果改善更显著。汛期与非汛期的多元变量试验降水预报结果更接近实况,且相对非汛期降水而言,汛期多元变量试验模拟效果更好,模拟的降水分布及强度与实况更为接近。Under different seasonal backgrounds,the comparative analysis of background error covariance characteristics related to multiple variables to understand its role in the weather process can improve the performance of the assimilation system and the level of precipitation numerical forecast.In this paper,background error covariance characteristics related to multiple variables in flood season and non-flood season in Jiangxi and its adjacent regions were compared and analyzed.Two precipitation processes during flood season and non-flood season in 2019 were selected respectively,and two groups of cyclic assimilation experiments were designed to explore the influence of background error covariance related to multiple variables on the precipitation forecast effect in Jiangxi.The results show that compared with the non-flood season,the analysis variables of the flood season model contribute more to the wind field,temperature field and water vapor field as a whole,and the background field error of each control variable of the flood season model is more significant.The evaluation effect of precipitation forecast in multivariable variable test was better on the whole and the improvement of flood season forecast was more obvious.The distribution characteristics of cumulative precipitation showed that the forecast results of multivariate variable test in flood season and non-flood season were closer to the actual precipitation,and the simulation results of multivariate variable test in flood season were better than that of non-flood season precipitation,and the distribution and intensity of simulated precipitation were closer to the actual precipitation.

关 键 词：多元变量相关 背景误差协方差 平衡贡献 降水预报 

分 类 号：P456.7[天文地球—大气科学及气象学]