山西高原降水量空间插值分析  

Spatial Interpolation Analysis of Annual Average Precipitation on Shanxi Plateau

作　　者：程朋[1] 张珺[1] 张茹[1] 闫美芳[2] 任鸿瑞[1] 

机构地区：[1]太原理工大学测绘科学与技术系,山西太原030024 [2]太原理工大学环境科学与工程学院,山西太原030024

出　　处：《人民黄河》2016年第2期24-27,31,共5页Yellow River

基　　金：山西省高等学校创新人才支持计划资助项目;山西省青年科技研究基金资助项目(2013021030-1);太原理工大学人才引进科研启动项目(TYUTRC-201240A);太原理工大学教育教学改革项目;太原理工大学2013年校青年团队项目(2013T007)

摘　　要：利用山西高原及其周边27个国家气象台站1981—2005年的降水资料,采用反距离权重法、全局多项式法、局部多项式法、径向基函数法、样条函数法、趋势面法、普通克里金法与经验贝叶斯克里金法共8种插值方法进行了空间插值,使用留一法交叉检验了不同方法的插值精度,并分析了山西高原年均降水量的空间分布特征。结果表明:1径向基函数法具有最佳的插值精度,其平均绝对误差、平均相对误差、均方根误差分别为25.1 mm、5.7%、31.4 mm;2各种插值方法按精度高低的排序依次为径向基函数法、样条函数法、趋势面法、普通克里金法、局部多项式法、全局多项式法、经验贝叶斯克里金法、反距离权重法;3山西高原降水量空间分布具有明显的地域性规律,一方面表现为由东南向西北递减,另一方面表现为东部太行山和西部吕梁山的降水量大于中部汾河谷地的。Based on the precipitation data from 27 meteorological stations in Shanxi Plateau and its surrounding areas, the paper used some spatial interpolation methods to interpolate the annual average precipitation of Shanxi Plateau from 1981 to 2005. These methods included Inverse Distance Weight, Global Polynomial Interpolation, Local Polynomial Interpolation, Radial Basis Function, Spline, Trend, Ordinary Kriging and Empirical Bayesian Kriging. The precision of these methods was evaluated by using leave?one?out cross validation. Finally, it analyzed the precipitation spatial distribution of Shanxi Plateau which was obtained by the optimal method. The results show that a) among these methods, Radial Basis Function yields the best performance, whose mean absolute error is 25.1 mm, mean relative error is 5.7% and root mean squared error is 31.4 mm;b) according to the accuracy, the interpolation methods is ranked as:Radial Basis Function, Spline, Trend, Ordinary Kriging, Local Polynomial Interpolation, Global Polynomial Interpolation, Empirical Bayesian Kriging and Inverse Distance Weight and;c) the precipitation spatial distribution in Shanxi Plateau presents the regional characters, on the one hand, which shows a decreased trend from southeast to northwest, on the other hand, the precipitation of Mount Lyuliang and Mount Taihang is much more than that of the Fenhe River.

关 键 词：分布规律 空间插值 降水量 山西高原 

分 类 号：TV211.1[水利工程—水文学及水资源] P208[天文地球—地图制图学与地理信息工程]