基于ECMWF细网格、GRAPES的短时强降水潜势预报和检验  

Short Time Heavy Rain Potential Prediction and Inspection Based on ECMWF Fine Grid and GRAPES_GRAPES

作　　者：吴迎旭[1] 孟莹莹[1] 周一[1] 赵柠[1] 李兴权[1] 刘松涛[1] 吴岩[1] WU Yingxu;MENG Yingying;ZHOU Yi;ZHAO Ning;LI Xingquan;LIU Songtao;WU Yan(Heilongjiang Meteorological Observatory,Harbin 150030,China)

机构地区：[1]黑龙江省气象台,黑龙江哈尔滨150030

出　　处：《沙漠与绿洲气象》2021年第5期30-40,共11页Desert and Oasis Meteorology

基　　金：中国气象局预报员专项“基于EC细网格和GRAPES_GRAPES短时强降水客观预报”(CMAYBY2018-024);黑龙江省自然科学基金项目“黑龙江省致灾强降水雷达回波物理机制研究”(LH2020D014)。

摘　　要：利用2016—2018年6—8月ECMWF细网格、GRAPES_GRAPES(简称GRAPES)、黑龙江省822个自动站资料研究黑龙江省6—8月短时强降水(一般短时强降水和极端短时强降水)的预报方法和各影响因子指标与它们之间的融合。采用双线插值法或临近格点法、分位数法、配料法、排除法、多重分析法等形成以水汽、不稳定、抬升为框架的客观预报方法。研究发现,强降水的环境背景不仅受限于各物理因子阈值,还与它们之间融合密切相关。各因子间存在一定旬差异和日较差,与水汽相关的各阈值夜间大于白天,热力不稳定性白天高于夜间。6月中上旬与水汽含量相关的各因子阈值小于其它时段。从检验结果看,由于强降水的突发性、局地性和研究方法以及模式本身的特征,预报的空报率非常大,漏报率较低,TS评分最低且随着强降水分布密度的降低而降低。一般性强降水检验中,两种模式点对点检验的TS评分为0.015左右,14和40 km为半径的点对面检验中TS评分约0.03和0.08。极端强降水检验中,两种模式点对点检验TS评分约0.004,14和40 km点对面检验TS约0.007和0.02。7月由于强降水分布密度相对较大,检验效果也相对较好。一般性强降水检验TS评分EC细网格高于GRAPES_GRAPES,而极端强降水检验TS评分刚好相反。Using the ECMWF fine gird,GRAPES_GRAPES and 822 automatic stations data from June to August during 2016-2018 to study the forecast methods of short-time heavy rainfall(general short time heavy precipitation and extreme short-term heavy precipitation)in Heilongjiang Province from June to August and the fusion of various impact factor indicators and them.Use double-line interpolation or adjacent grid method,quantile method,batching method,elimination method,multiple analysis method,etc.to form an objective forecast method based on water vapor,instability,and uplift.The results show that the environmental background of heavy rainfall is not only limited by the thresholds of various physical factors,but also closely related to their fusion.There are a certain ten-day difference and daily difference among various factors.The thresholds related to water vapor are greater at night than that during the day,and thermal instability is higher during the day than that at night.The thresholds of the factors related to water vapor content in mid-to-early June are smaller than that in other time periods.Due to the suddenness of heavy precipitation,local topography and research methods,and the characteristics of the model itself,the forecasting rate of false alarms is very high,the rate of underreporting is low,and the TS score is the lowest,and reduces with the decrease of the distribution density of heavy precipitation.In the general heavy precipitation test,the TS score of the point-to-point test of the two modes is about 0.015,and the TS scores of the point-to-surface test with radius of 14 km and 40 km are about 0.03 and 0.08.In the extreme heavy precipitation test,the TS score for the two models is about 0.004 for point-to-point tests,and the TS for points-to-point tests for 14 km and 40 km are about 0.007 and 0.02.Due to the relatively large distribution density of heavy rainfall in July,the test results are relatively good.The general heavy rainfall test TS score of EC fine grid is higher than that of GRAPES_GRAPES,while

关 键 词：短时强降水 ECMWF细网格 GRAPES_GRAPES 潜势预报 检验 

分 类 号：P457.6[天文地球—大气科学及气象学]