地区线性矩法在江苏省短历时极值降水频率分析中的应用  

Application of regional linear moment method in the analysis of the frequency of short-duration extreme precipitation in Jiangsu Province

作　　者：李敏 王啸华[2] 邵月红[3] 王坤 LI Min;WANG Xiaohua;SHAO Yuehong;WANG Kun(Haimen Meteorological Bureau,Jiangsu Nantong 226100,China;Jiangsu Meteorological Observatory,Nanjing 210041,China;Nanjing University of Information Science&Technology,Nanjing 210044,China;Nantong Meteorological Bureau,Jiangsu Nantong 226000,China)

机构地区：[1]海门区气象局,江苏南通226100 [2]江苏省气象台,南京210041 [3]南京信息工程大学,南京210044 [4]南通市气象局,江苏南通226000

出　　处：《气象科学》2024年第6期1038-1046,共9页Journal of the Meteorological Sciences

基　　金：江苏省水利科技资助项目(2022042);北极阁开放研究资助基金项目(BJG202211)。

摘　　要：本文以江苏省为研究区域,采用地区线性矩法进行一致区划分、最优频率分布曲线选择和极值降水频率估计值计算,并分析其时间和空间分布特征。结果表明:根据不和谐检测和异质性检验的分析结果将研究区划分为7个合理的子区域,并在此基础上,通过蒙特卡洛模拟和均方根误差综合判断得出各时段的最优频率分布曲线。其中,12 h时段下的各子区域的最优频率分布曲线分别是GLO、GNO、PE3、GEV、GLO、PE3、GNO。利用最优分布函数计算出各站点一整套短历时(1、3、6、12 h)不同重现期下的极值降水频率估计值,同一降水时段下,频率估计值随着重现期的增加而增加。同一重现期下,频率估计值的空间分布差异明显,大值区与实际观测结果一致。In this study,Jiangsu Province was used as the research area.The regional linear matrix method was used for the division of consistent areas,the selection of optimal distribution,and the calculation of extreme rainfall frequency estimates.Results show that seven consistent areas are divided in the study area.On this basis,optimal distribution was selected through monte-carlo simulation and Root Mean Square Error(RMSE)comprehensivejudgment.The12-hour timeinterval under optimal distribution is GLO,GNO,PE3,GEV,GLO,PE3,and GNO.Finally,the calculation of extreme rainfall frequency estimates for each station under a single time interval(1,3,6,12 h)and different reoccurrences was carried out.Under the same precipitation period,the frequency estimates increase with the increase of the return period.The spatial distribution of frequency estimates under the same return period varies significantly,with large value areas consistent with actual observations.

关 键 词：地区线性矩 短历时极值降水 频率分析 频率估计值 

分 类 号：P426.6[天文地球—大气科学及气象学]