摘要
利用中国气象局全国综合气象信息共享平台(CIMISS)提供的1981~2018年重庆地区35个国家气象站逐日降水资料,采用最小曲率插值法、径向基函数插值法、克里金插值法、局部多项式插值法、反距离加权插值法以及改进的Shepard插值法,对平均累计降水、降水极值以及夏季极端降水空间插值模拟结果进行交叉检验。结果表明:空间插值方法在渝西和渝中大部分地区以及长江沿线地区适用性较好,在渝东北、渝东南地区北部以及渝西地区南部的山区适用性较差。对于平均累计降水,各插值方法的均方根误差随着降水量的增加而增大,其中最小曲率插值法整体表现最为稳定,最小曲率插值法在渝东南、渝中以及渝西地区最为适用,径向基函数插值法在渝东北地区最为适用。对于降水极值,量级模拟最好的插值方法为径向基函数插值法,位置模拟最好的方法为最小曲率插值法。对于夏季极端降水,最小曲率插值法插值结果与实况最为接近。
Based on the daily precipitation data of 35 meteorological stations in Chongqing from 1981 to 2018 by China Integrated Meteorological Information Service System,the minimum curvature(MC),radial basis function(RBF),kriging(KR),local polynomial(LP),inverse distance weighting(IDW)and Shepard interpolation method(MS)are implemented for the simulation and cross examination of average accumulative precipitation,extreme precipitation and summer extreme precipitation.The test results show that the spatial interpolation method has good applicability in most areas of western and central Chongqing and along the Yangtze River,and has poor applicability in the mountainous areas of northeast,southeast and southwest of Chongqing.For the average cumulative precipitation,the root mean square error of each interpolation method increases with the increase of precipitation,among which the MC is the most stable as a whole,and it is the most suitable in the southeast,middle and west of Chongqing.the RBF is the most suitable in the northeast of Chongqing.In terms of extremum precipitation,the RBF is the best method for magnitude simulation,and the MC is the best method for position simulation.The interpolation result of the MC is the closest to actual precipitation of summer extreme precipitation.
作者
刘祥
庞玥
何军
陈欢
LIU Xiang;PANG Yue;HE Jun;CHEN Huan(Chongqing Meteorological Service Centre,Chongqing 401147,China;Chongqing Meteorological Observatory,Chongqing 401147,China)
出处
《高原山地气象研究》
2022年第2期75-81,共7页
Plateau and Mountain Meteorology Research
基金
重庆市气象局青年基金项目(QNJJ-202005)。
关键词
空间插值
交叉检验
降水
适用性分析
Spatial interpolation
Cross-validation
Precipitation
Applicability analysis