基于GIS的滴灌棉田土壤养分空间变异及预测方法比较研究  

Comparative Study of Soil Nutrient Spatial Variability and Prediction Methods in Drip-irrigated Cotton Field Based on GIS

作　　者：吕宁[1] 吕新[1] 马莉[1] 王海江[1] 陈剑[1] 张泽[1] 李新伟[1] 董楠 张涛[1] 

机构地区：[1]新疆生产建设兵团绿洲生态农业重点实验室,新疆石河子832003

出　　处：《土壤通报》2013年第2期403-408,共6页Chinese Journal of Soil Science

基　　金：新疆产学研重大专项--精准农业与信息技术应用与示范(2007ZX03);国家自然基金(30760104)资助

摘　　要：在GIS和地统计学方法支持下,对新疆农五师81团滴灌棉田土壤养分空间变异特征进行定量评价。结果表明:研究区有效铜、有效铁、有效锌、有效硼、有效锰、总盐等空间变异主要由结构性因素引起,有机质、碱解氮、速效磷、速效钾空间变异主要由结构性和随机性因素共同引起;在普通克里格、反距离加权、全局多项式、局部多项式、径向基函数预测方法中,用局部多项式插值法来预测有机质、碱解氮、速效磷、有效铁、有效锌、有效锰、有效硼的空间分布精度较高;用普通克里格插值法来预测有效铜空间分布精度较高;用径向基函数插值法来预测速效钾的空间分布精度较高。Based on the GIS and geo statistics methods, the spatial variability of soil nutrients was quantitatively evaluated in one drip-irrigated cotton field of 81st Regiment of 81st Agricultural Construction Division in Xinjiang. The results showed that the spatial variability of the available Cu, Fe, Zn, B, and Mn, as well as total salt were mostly influenced by soil structural factors, while the spatial variability of organic matter, available N and P and K were mostly impacted both by the structural and random factors. Among the methods of ordinary kriging （OK） and inverse distance weighting （IDW） and global polynomial （GPI） and radial basis function （RBF） and local polynomial （LPI） prediction method, the local polynomial interpolation （LPI） had the highest accuracy in predicting the spatial variability of the organic matter, as well as available N, P, Fe, Zn, Mn and B; the ordinary kriging （OK） had the highest accuracy for available Cu and the radial basis function （RBF） had the highest accuracy for available K.

关 键 词：土壤养分 滴灌棉田 GIS 空间变异 插值方法 

分 类 号：S158.9[农业科学—土壤学]