基于暴雨空间自相关性的降水量插值方法比较  

Comparison of precipitation interpolation methods considering spatial autocorrelation of rainstorm precipitation

作　　者：谈灿灿 张润润[1] 邓志民[2] 宁玮 布庆月 章阳 TAN Cancan;ZHANG Runrun;DENG Zhimin;NING Wei;BU Qingyue;ZHANG Yang(College of Hydrology and Water Resources,Hohai University,Nanjing 210098,China;Changjiang Water Resources Protection Institute,Wuhan 430051,China;Fuhe River Hydrology and Water Resources Monitoring Center,Fuzhou 344000,China;Poyang Lake Key Laboratory of Hydroecological Monitoring,Nanchang 330002,China)

机构地区：[1]河海大学水文水资源学院,江苏南京210098 [2]长江水资源保护科学研究所,湖北武汉430051 [3]抚河水文水资源监测中心,江西抚州344000 [4]鄱阳湖水文生态监测研究重点实验室,江西南昌330002

出　　处：《河海大学学报（自然科学版）》2024年第6期8-14,29,共8页Journal of Hohai University(Natural Sciences)

基　　金：中央高校基本科研业务费专项资金资助项目(B220202035);国家自然科学基金项目(41701016)。

摘　　要：为研究空间自相关性对暴雨事件降水量空间插值精度的影响,以乌江流域2009-2018年102个雨量站点的日降水量数据为基础,筛选出88场暴雨事件,分别采用反距离权重法(IDW)、径向基函数法(RBF)、普通克里金法(OK)、协同克里金法(CoK)对流域日降水量进行空间插值,利用莫兰指数(Moran’s I)定量分析了102个雨量站各场暴雨事件降水量的空间自相关性,对比分析了不同插值方法对不同空间自相关性暴雨事件降水量的插值精度。结果表明:随着暴雨事件降水量空间自相关性的增强,莫兰指数增大,4种插值方法精度均有不同程度提升,OK法和CoK法的精度提升效果更为明显;对于流域平均而言,莫兰指数小于0.28时推荐使用RBF法,莫兰指数大于或等于0.28时推荐使用OK法;对于暴雨中心而言,莫兰指数小于0.4时推荐使用RBF法,莫兰指数大于或等于0.4时推荐使用OK法或CoK法。To investigate the impact of spatial autocorrelation on the accuracy of spatial interpolation for the precipitation of rainstorm events,based on the daily precipitation data from 102 rainfall stations in the Wujiang River Basin from 2009 to 2018,88 rainstorm events were screened out.These events were interpolated using the inverse distance weighting(IDW),radial basis function(RBF),ordinary Kriging(OK),and co-Kriging(CoK)methods.Moran’s I was employed to quantitatively analyze the spatial autocorrelation of each rainstorm event precipitation at the 102 rainfall stations,and this study compared the interpolation accuracy of different methods for rainstorm event precipitation with varying spatial autocorrelation.The results show that with the increase of spatial autocorrelation of rainstorm event precipitation,indicated by the Moran’s I index,the accuracy of all four interpolation methods improves to varying degrees,and the accuracy improvement of OK method and CoK method is more obvious.For the average basin,when the Moran’s I is less than 0.28,the RBF method is recommended,whereas when the Moran’s I is greater than or equal to 0.28,OK is recommended.For the rainstorm center,when the Moran’s I is less than 0.4,RBF is recommended,when it is greater than or equal to 0.4,OK or CoK is recommended.

关 键 词：暴雨事件 降水量 空间自相关性 莫兰指数 空间插值 插值精度 乌江流域 

分 类 号：P426.6[天文地球—大气科学及气象学]