摘要
基于灾害天气短时临近系统(SWAN)的定量降水预报(QPF)产品进行0~2 h QPF外推试验研究。利用贵州省地面自动站数据、SWAN输出的1 h-QPF产品、SWAN 1 h定量降水估测(QPE)产品进行融合订正外推。选取2018年3次典型暴雨个例进行试验,采用融合订正技术,利用相似离度算法对降水强度位相进行调整,Weibull分布算法对降水极值分布进行订正,改进后的交叉相关法(COTREC)进行降水外推预报。通过主客观检验对比,初步说明基于Weibull分布算法的降水极值订正对降水极值分布有很好的模拟效果。融合订正外推的贵州省短时临近定量降水预报产品(GZ_NPF)在>5 mm以上量级降水的TS检验中评分更高。GZ_NPF的空报率明显降低,但空报率降低的代价是漏报率相对提高。GZ_NPF总的相对误差减小,格点总降水量与实况总降水量更为接近。试验说明新的融合订正外推算法提高了0~2 h短时临近降水的预报能力。
Based on the QPF of the Severe Weather Automatic Nowcast System(SWAN),the quantitative precipitation forecast(QPF)for 0-2 h duration extrapolation test was researched.The data were used including:regional meteorological stations and national meteorological stations,QPF product of the SWAN,quantitative precipitation estimation(QPE)product of the SWAN.Three intense precipitation events("6·20"、"6·22"、"9·26")in 2018 were selected to make extrapolation test.New demands for better short-term forecasting and nowcasting used the fusion correction extrapolation forecast,which was based on the similarity divergence algorithm to rectify the precipitation phase and intensity,Weibull distribution to rectify the precipitation extremum distribution and to use the improved cross-correlation extrapolation(COTREC)method extrapolation.Through our research found that the QPF of extrapolation test were better than the original QPF of the SWAN on phase and intensity.The study in this paper preliminary certified that Weibull distribution could rectify the precipitation extremum distribution greatly.The QPF extrapolation test of three cases had the similar consequence,the error corrections of precipitation intensity to be the best(the error less 5 mm in most areas and cases).Nowcasting precipitation forecast of Guizhou(GZ_NPF)by using the fusion correction had higher threat score(TS).The GZ_NPF had lower NH,but it got a higher PO.The total absolute error numbers of GZ_NPF were smaller than the original QPF,and the Cumulative total precipitation of GZ_NPF were more close to the actual total precipitation.The experiments testified that the new fusion error corrections raised the precipitation precision rate of the short-term forecasting and nowcasting.
作者
苏华英
曾莉萍
罗乃兴
王国松
周皞
SU Huaying;ZENG Liping;LUO Naixing;WANG Guosong;ZHOU Hao(Power Dispatching Center of Guizhou Power Grid Liability Co.LTD,Guiyang 550002,China;Guizhou Meteorological Bureau,Guiyang 550002,China;Guiyang Meteorological Bureau,Guiyang 550001,China)
出处
《沙漠与绿洲气象》
2020年第5期27-35,共9页
Desert and Oasis Meteorology
基金
南方电网公司科技项目(066500KK52160006)
贵州省气象局重要业务科研项目(黔气标合ZY[2018]01号)
贵州省气象局科研业务项目(黔气科登[2018]05-04号)。
关键词
短时临近降水预报
QPF融合外推技术
相似离度算法
WEIBULL分布
short-term forecasting and nowcasting
blending extrapolation techniques of QPF
similarity divergence algorithm
Weibull distribution
