基于快速滚动更新的无缝隙定量降水预报模型  

Seamless quantitative precipitation forecasting model based on rapid rolling update technique

作　　者：曹勇 包红军 张恒德[1,2] 郭云谦 刘凑华[1,2] 陈双 CAO Yong;BAO Hongjun;ZHANG Hengde;GUO Yunqian;LIU Couhua;CHEN Shuang(National Meteorological Center,China Meteorological Administration,Beijing 100081,China;CMA-HHU Joint Laboratory for Hydrometeorological Studies,Beijing 100081,China)

机构地区：[1]国家气象中心,北京100081 [2]中国气象局-河海大学水文气象研究联合实验室,北京100081

出　　处：《河海大学学报（自然科学版）》2021年第4期303-308,共6页Journal of Hohai University(Natural Sciences)

基　　金：国家重点研发计划(2018YFC1508102);国家自然科学基金(41775111,41875131);中国气象局气象预报业务关键技术发展专项(YBGJXM(2019)03)。

摘　　要：针对中小尺度天气系统极易导致突发性短时强降水而难以定量化预报的问题,提出了一种基于快速滚动更新的无缝隙定量降水预报模型。模型基于改进光流法技术实现临近(0~2 h)定量降水预报,基于GRAPES-3 km实时频率匹配订正技术构建短时(2~12 h)定量降水预报,基于最优背景场生成技术形成短中期(12~240 h)定量降水预报;依据3种预报时效定量降水预报技术的技巧差异,结合实时预报检验评估,提出了基于经验权重函数的不同预报时效之间的预报融合技术,并根据最新降水实况逐1 h调整预报,构建了0~240 h快速滚动更新的无缝隙定量降水预报模型。选取2018年5—10月24 h定量降水预报进行检验,并与传统定时预报对比,结果表明该模型24 h降水预报准确率高于传统定时预报,特别是对中小尺度天气系统突发性短时强降水预报时空精度明显提高,有利于中小河流洪水预报与灾害防控。The sudden short-term severe precipitation,frequently caused by mesoscale and microscale weather systems,is very complicated to forecast quantitatively.This study develops a seamless quantitative precipitation forecasting(QPF)model based on the rapid rolling update technique.In the developed model,the quantitative precipitation nowcasting(0-2 h)is based on the improved optical flow method,the short-term quantitative precipitation forecasting(2-12 h)is based on the GRAPES-3 km and real-time frequency matching correction method,and the short and medium range quantitative precipitation forecasting(12-240 h)is based on the generation technique of optimal background precipitation field.According to forecasting skills of three methods in different forecasting time,the forecast fusion technology based on the empirical weight function is developed for the seamless quantitative precipitation forecasting(0-240 h)with the real-time precipitation verification and QPF.The 24-hour quantitative precipitation forecast from May 2018 to October 2018 is taken as the experimental case to test the developed model,and the result is compared with forecasts of the traditional model.Results show that the 24-hour precipitation forecast accuracy of the developed model is higher than that of the traditional forecast model,especially for the short-term quantitative sudden severe precipitation forecasting caused by mesoscale and microscale weather systems.It is conducive to the flood forecasting,disaster prevention and control of small to medium-sized rivers.

关 键 词：无缝隙定量降水预报 快速滚动更新 改进光流法 实时频率匹配订正 最优背景场 

分 类 号：P338[天文地球—水文科学]