黑龙江省2020-2022年汛期多模式降水预报检验评估  

Verification and Estimate of Multi-model Precipitation Forecast in Heilongjiang Flood Season 2020-2022

作　　者：刘松涛[1] 高梦竹 王承伟[1] 张惠君[1] LIU Song-tao;GAO Meng-zhu;WANG Cheng-wei;ZHANG Hui-jun(Meteorological Observatory of Heilongjiang Province,Heilongjiang Harbin 150001)

机构地区：[1]黑龙江省气象台,黑龙江哈尔滨150001

出　　处：《黑龙江气象》2022年第4期1-5,共5页Heilongjiang Meteorology

基　　金：国家重点研发计划“重大自然灾害监测预警与防范”重点专项“我国北方局地突发性强降水机理及预报方法研究”专题五(2018YFC1507305);黑龙江省气象局竞争性科技攻关项目(HQGG202006);黑龙江省气象局“智能网格预报及数值模式释用”创新团队共同资助。

摘　　要：本文利用黑龙江省考核站降水观测数据,对2020-2022年汛期6-8月全球尺度模式预报、中尺度模式预报、客观订正预报等在黑龙江省预报性能、误差进行了对比分析。对比检验分析发现,定量降水预报误差与观测降水量密切相关,汛期降水量较大,预报定量误差较大。黑龙江省气象台客观预报和融合预报误差相对较小,中尺度模式的定量误差相对较大。五种降水预报存在预报范围偏大的问题,EC预报范围偏大现象最为显著。针对一般性降水,两种订正预报随机误差表现最好,其次为EC预报,两种中尺度模式降水预报的随机误差表现相对较差。订正预报pm_fm在大雨及以下量级随机误差表现最好,暴雨量级CMA-SH9最好。订正预报不但可以有效消除模式系统性误差,也可以改善原始模式随机误差。降水分级检验结果表明,CMA-MESO在大暴雨量级有一定的预报能力,fuse在大雨量级上表现较好,pm_fm在小雨、暴雨量级上预报能力明显优于其它模式。In this paper,the performance and error of global scale model forecast,mesoscale model forecast and objective correction forecast in Heilongjiang flood season from June to August 2020-2022 are compared and analyzed by using the precipitation observation data.The comparison results show that the quantitative precipitation prediction error is closely related to the observed precipitation,the precipitation in summer is large,and the quantitative prediction error is also large.The quantitative error of objective correction forecast and fusion forecast is relatively small,and the quantitative error of mesoscale model is relatively large.There is a problem that the range of the five precipitation forecasts is too large,especially the EC forecast.For general precipitation,the random errors of the two correction forecasts are the best,followed by the EC forecast,and the two mesoscale model forecasts are relatively poor.For random errors,the corrected forecast pm_fm performs best in heavy rain and below,and CMA-SH9 performs best in rainstorm.The correction forecast can not only effectively eliminate the systematic error of the model,but also improve the random error of the original model.The precipitation classification test results show that CMA-MESO has certain forecasting ability in downpour,the correction forecast fuse performs well in heavy rain,the corrected forecast pm_fm is superior to other models in light rain and rainstorm.

关 键 词：多模式预报 检验评估 随机误差 

分 类 号：P456.7[天文地球—大气科学及气象学]