基于分位数映射法的强降水预报——以四川短时强降水为例  

Heavy Rain Forecast Based on Quantile Mapping Method:a Case Study of Short-Duration Heavy Rain in Sichuan

作　　者：曹萍萍[1,2] 肖递祥 王佳津[1,2] 冯良敏 CAO Pingping;XIAO Dixiang;WANG Jiajin;FENG Liangmin(Sichuan Meteorological Observatory,Chengdu 610072,China;Heavy Rain and Drought-Flood Disasters in Plateau and Basin Key Laboratory of Sichuan Province,Chengdu 610072,China)

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

出　　处：《沙漠与绿洲气象》2025年第1期112-119,共8页Desert and Oasis Meteorology

基　　金：中国气象局2024年复盘总结专项(FPZJ2024-111);川西南(雅安)暴雨实验室科技发展基金项目(CXNBYSYSYWZD202405);四川省气象局智能网格预报创新团队。

摘　　要：利用逐小时四川地面观测降水资料及同时段ECMWF模式各要素预报场资料,分析四川地区短时强降水分布特征,从动力、热力、水汽等方面选取物理意义明确的预报因子,基于分位数映射法计算得到各因子分位值,结合配料法研发四川省短时强降水概率预报产品并投入业务使用。结果表明:(1)盆地短时强降水时空变化特征明显,12:00─17:00(北京时,下同)短时强降水频率较低,18:00后逐步上升,峰值出现在02:00,随后大幅下降,12:00发生短时强降水的比例最低;短时强降水主要发生在以陡坡地形为主的盆地西部沿山。(2)预报评估显示,概率35%以上区域对短时强降水落区有明确指示意义。概率预报产品能在ECMWF模式3h累计降水明显偏弱的情况下,有效指示短时强降水易发区域,且空报较小。Based on hourly surface precipitation observation data from Sichuan Province and ECMWF model data,the distribution characteristics of short-duration heavy rain were analyzed.Forecast factors with clear physical significance were selected from the dynamical,thermal,and moisture-related predictors,and the corresponding quantile values for each predictor were obtained using the quantile mapping method.Combining this with the ingredients-based methodology,probability forecast products for short-duration heavy rain were developed and applied for operational purposes.The results showed the following:(1)Temporal and spatial variations of short-duration heavy rain in the basin are significant.The frequency of short-duration heavy rain is low from 12:00 to 17:00,gradually increasing after 18:00,with a peak at 02:00,followed by a significant decrease thereafter.The lowest proportion of short-duration heavy precipitation occurs at 12:00.The majority of short-duration heavy rain events are concentrated in the western part of the basin,which is characterized by steep slopes.(2)Forecast evaluations demonstrated that regions with a probability greater than 35%show a significant correlation with short-duration heavy rain events.The probability forecast products effectively identify areas prone to short-duration heavy rain,with fewer false alarms,particularly when the ECMWF 3-hour cumulative precipitation forecast is weak.

关 键 词：短时强降水 分位数映射法 “配料”法 概率预报 

分 类 号：P457.6[天文地球—大气科学及气象学]