关键词: 土壤养分, 空间插值, 普通克里格法, 协同克里格方法, 交叉检验

Abstract:

Spatial interpolation is a main method for obtaining the spatial distribution of soil nutrient which is an important part in agricultural research field. There are many different methods for spatial interpolation with different results. Taking three kinds of soil nutrient elements including total nitrogen, available phosphorus and available potassium in Shijiazhuang rurbania as main variables, soil organic matter as auxiliary variable. We obtained the spatial distribution of soil nutrient content by using two types of representative interpolation method, including Ordinary Kriging and Cokriging. There were 85 samples used for spatial interpolation and we used cross validation to analyze and appraise the results. The results showed that compared with the ordinary Kriging under the same sampling numbers the Cokriging is more suitable for spatial interpolation of soil nutrients. The interpolation accuracy of original data reduced to 70 using Cokriging was still higher than original 85 data using ordinary Kriging and their spatial distributions were quite similar. The results showed that the Cokriging was more suitable for spatial interpolation of soil nutrients. We proposed scientific management measures of soil nutrient content based on the spatial distribution of soil nutrient. It provided scientific theoretical guidance to use and protect land resources rationally and improve economic efficiency of crops．

Key words: soil nutrient, spatial interpolation, ordinary Kriging, Cokriging, cross validation