复杂地形影响下四川省逐月气温空间插值方法研究  

Study on Interpolation Model of Monthly Temperature in Sichuan Province Under the Influence of Complex Topography

作　　者：何鹏[1,2] 蹇东南 李晓[1,2] 林正雨 HE Peng;JIAN Dong-nan;LI Xiao;LIN Zheng-yu(Agricultural Information and Rural Economy Institute,Sichuan Academy of Agricultural Sciences,Chengdu 610066,China;Big Data Center,Sichuan Academy of Agricultural Sciences,Chengdu 610066,China;College of Resources,Sichuan Agricultural University,Chengdu 611130,China)

机构地区：[1]四川省农业科学院农业信息与农村经济研究所,四川成都610066 [2]四川省农业科学院大数据中心,四川成都610066 [3]四川农业大学资源学院,四川成都611130

出　　处：《生态与农村环境学报》2019年第6期801-808,共8页Journal of Ecology and Rural Environment

基　　金：四川省科技支撑计划软科学项目(2017ZR0045),四川省科技支撑计划(2016NZ0114);四川省财政创新能力提升工程(2016GYSH-004)

摘　　要：为了探索适合四川省的逐月气温数据空间插值方法,在已有研究的基础上运用样条函数法(SP)、反距离权重法(IDW)、普通克立格法(OK)和加入海拔影响因子的样条函数法(SPE)、反距离权重法(IDWE)、普通克立格法(OKE)及多元回归分析法(MRM)对全省144个气象站点的多年逐月平均气温进行空间插值,并利用交叉检验方法对不同方法的插值精度进行评估。结果表明:四川省月平均气温与海拔高度之间呈显著相关关系,且相关性呈明显季节性变化,夏季相关系数大于冬季。各月平均气温随海拔高度的增加均呈递减趋势,夏季气温垂直变化率明显大于冬季,各月气温垂直变化率在0. 308~0. 443 ℃·(100 m)^-1 之间;考虑了海拔影响因子的插值结果明显好于不考虑海拔影响因子的插值结果,在考虑海拔影响因子的4种空间插值方法中,MRM效果最好,其次是OKE和IDWE,最差的为SPE。Sichuan Province has a great variety in topography,which has significant influences on regional climate distribution. Based on the data of monthly air temperature of 144 meteorological stations in Sichuan Province,seven interpolation methods including spline(SP),inverse distance weighting(IDW),ordinary kriging(OK),spline function method considering elevation effect(SPE),inverse distance weighing considering elevation effect(IDWE),ordinary kriging considering elevation effect(OKE)and multivariable linear regression method(MRM)were applied in current study to spatialize the monthly temperature. Meanwhile,the cross validation method was used to evaluate the accuracy of seven interpolation methods. The results show that the monthly temperature in Sichuan Province was significantly correlated with altitude,and correlation coefficient changed seasonally,with greater change in summer than in winter. The average temperature of each month decreased with the increase of altitude,and lapse rate of summer was greater than that of winter. The lapse rate for different months varied from 0. 308 to 0. 443 ℃·hm^-1. The interpolation methods considering elevation effect achieved better accuracy than that without considering elevation effect. Among four interpolation methods considering elevation effect,MRM had the best accuracy,followed by IDWE and OKE,and SPE had the worst accuracy.

关 键 词：月平均气温 空间插值方法 数字高程模型 四川省 

分 类 号：X83[环境科学与工程—环境工程]