基于CMA-REPS小时降水的邻域集合预报应用试验  

Application experiment of neighborhood ensemble forecasting based on CMA-REPS hourly precipitation

作　　者：刘莹[1,2] 陈朝平 陈莹[1,2] 龙柯吉 周秋雪[1,2] Liu Ying;Chen Chaoping;Chen Ying;Long Keji;Zhou Qiuxue(Sichuan Meteorological Observatory,Chengdu 610072,China;Sichuan Provincial Key Laboratory of Heavy Rain,Drought and Flood Disasters in Plateaus and Basins,Chengdu 610072,China)

机构地区：[1]四川省气象台,成都610072 [2]高原与盆地暴雨旱涝灾害四川省重点实验室,成都610072

出　　处：《气象研究与应用》2022年第2期98-104,共7页Journal of Meteorological Research and Application

基　　金：四川省气象局重点实验室重大项目(2018-重点-06);四川省气象局重点实验室研究型业务面上专项(SCQXKJYJXMS202201);国家重点研发计划重点专项项目(2017YFC1502000)、(2017YFC1502004)。

摘　　要：基于四川2020年汛期CMA区域集合模式的小时降水产品开展了集合预报邻域法订正试验,试验方案分为四种:邻域平均的集合平均预报(ENM)、邻域平均的集合成员最大预报(MNM)、邻域概率的集合平均概率预报(ENP)、邻域概率的集合最大概率预报(MNP)。结果表明,小时降水超过0.1mm·h^(-1)的ENM在部分预报时效内TS评分略优于原集合平均预报;小时雨强分级检验的结果均显示出邻域概率预报优于原概率预报,随小时降水阈值的增大,最优TS评分越趋于更高邻域半径和更低的概率,且10mm·h^(-1)以上MNP预报更具参考性;MNP能显示集合预报的大范围降水中心,可以从另一个角度给预报员预报提供一定的参考。Based on the hou rly precipitation products of the CMA regional ensemble model in the flood season of Sichuan in 2020,the correction experiment of the ensemble forecast neighborhood method was carried out.There are four kinds of test schemes:neighborhood average ensemble average prediction(ENM),neighborhood average ensemble member maximum prediction(MNM),neighborhood probability ensemble average probability prediction(ENP),and neighborhood probability ensemble maximum probability prediction(MNP).The results show that the TS score of the ENM with hourly precipitation exceeding 0.1mm·h;is slightly better than the original ensemble average forecast within the partial forecast period.The results of the hourly rainfall intensity classification test all show that the neighborhood probability prediction is better than the original probability prediction.With the increase of hourly precipitation threshold,the optimal TS score tends to higher neighborhood radius and lower probability,and MNP prediction above 10mm·h;is more referential.MNP can display the large-scale precipitation center of ensemble forecast,and can provide a certain reference for forecasters from another angle.

关 键 词：小时降水 集合预报 邻域法 

分 类 号：P456[天文地球—大气科学及气象学]