基于北斗/GNSS PWV数据构建的短临强降雨模型及其应用  

Constructing a short term heavy rainfall model based onBeiDou/GNSS PWV data and its application

作　　者：邓南山 刘洋[2] 李小伟[1] 斯小华[1] 雷磊 DENG Nanshan;LIU Yang;LI Xiaowei;SI Xiaohua;LEI Lei(Central South Institute of Metallurgical Geology,Yichang 443000,China;School of Geodesy and Geomatics,Wuhan University,Wuhan 430079,China)

机构地区：[1]中南冶金地质研究所,湖北宜昌443000 [2]武汉大学测绘学院,湖北武汉430079

出　　处：《测绘通报》2024年第9期96-100,共5页Bulletin of Surveying and Mapping

基　　金：宜昌市自然科学研究项目(A23-2-039);湖北省重点研发计划(2022BAA047);湖北省自然科学基金(2022CFB557)。

摘　　要：传统降雨预警模型仅局限于预警降雨事件是否发生,忽略了对短临强降雨事件的预警。基于降雨发生前大气可降水量(PWV)表现出明显增长的趋势,本文提出了基于北斗/GNSS PWV的短临强降雨预警模型。模型包含PWV值、PWV变化量和PWV变化率3种预测因子,并引入百分位法确定预测因子关键参数的最优阈值。选取湖北省宜昌市5个GNSS站点2022年逐小时PWV和降雨数据进行验证,统计结果显示,本文提出的短临强降雨预警模型可预测未来2~6 h内94%的大雨事件,误报率仅为32.84%。The traditional rainfall warning model is only limited to warning whether rainfall events have occurred,ignoring the warning of short-term and imminent heavy rainfall events.Based on the significant increase in precipitable water vapor(PWV)before rainfall occurs,this paper proposes a short-term and imminent heavy rainfall warning model based on BeiDou/GNSS PWV.The model includes three predictive factors:PWV value,PWV change amount and PWV change rate,and introduces the percentile method to determine the optimal threshold for key parameters of the predictive factors.Selecting five GNSS stations in Yichang city,Hubei province for hourly PWV and rainfall data validation in 2022,the statistical results show that the proposed short-term and imminent heavy rainfall warning model can predict 94%of heavy rain events in the next 2~6 hours,with a false alarm rate of only 32.84%.

关 键 词：强降雨预警 北斗/GNSS PWV 预警模型 

分 类 号：P228[天文地球—大地测量学与测量工程]