摘要
针对江西省气象站点分布不均,密度过小,无法做到对江西整个区域进行观测,利用2020年江西省及邻近区域115站点气温月平均数据,采用反距离权重法、普通克里金法和ANUSPLIN法对江西省月平均气温进行空间插值获取插值栅格数据,并采用交叉验证法对插值精度进行对比检验,选出了对江西地区气温插值效果最好的插值方法。结果表明:在精度方面,ANUSPLIN软件插值法的插值精度最高(MAE:0.626℃,RMSE:0.751℃)。在普通克里金法次之(MAE:1.116℃,RMSE:1.348℃),反距离权重法精度最低(MAE:1.008℃,RMSE:1.217℃)。插值表观效果方面,ANUSPLIN软件法插值表面文件纹理清晰,气温走势与海拔走势息息相关;普通克里金法在站点数目少时插值表面文件易出现分层现象;反距离权重法易受极值点影响,插值表面文件易出现“牛眼”现象。综合分析得出3种插值方法对江西省气温插值的精度从高到低依次为:ANUSPLIN软件法,普通克里金法,反距离权重法。
In view of the uneven distribution of meteorological stations in Jiangxi Province and the small density to observe the entire area of Jiangxi,this paper uses the monthly average temperature data of 115 stations in Jiangxi Province and adjacent areas in 2020.The inverse distance weight method,ordinary kriging method and ANUSPLIN method are used to obtain the interpolation raster data of the monthly average temperature of Jiangxi Province.The cross-validation method is used to compare and test the interpolation accuracy,and the interpolation method is used to select the highest interpolation accuracy for the temperature in Jiangxi region.The results show that the interpolation accuracy of ANUSPLIN software interpolation method is the highest(MAE:0.626℃,RMSE:0.751℃),followed by ordinary kriging(MAE:1.116℃,RMSE:1.348℃),and the inverse distance weighting method has the lowest accuracy(MAE:1.008℃,RMSE:1.217℃).From the analysis of interpolated surface files,the texture of ANUSPLIN software interpolated surface files is clear,and the temperature trend is closely related to the altitude trend.Ordinary kriging interpolates surface files when the number of sites is small.The inverse distance weighting method is easily affected by extreme value points,and the interpolated surface file is prone to the phenomenon of"bull's eye".Comprehensive analysis shows that the accuracy of three interpolation methods for temperature interpolation in Jiangxi Province is:ANUSPLIN software method,ordinary kriging method,and inverse distance weight method.
作者
付艳平
幸松
王妍
曹良中
Fu Yanping;Xing Song;Wang Yan;Cao Liangzhong(Jiujiang University,Jiujiang,Jiangxi 332005,China)
出处
《绿色科技》
2023年第4期51-56,共6页
Journal of Green Science and Technology
关键词
月平均气温
空间插值
普通克里金
反距离权重
ANUSPLIN软件
average monthly temperature
spatial interpolation
ordinary kriging
inverse distance weighting
ANUSPLIN software
