土壤连续属性空间插值方法及其精度的研究进展

作为土壤变化的时空定量监测方法,土壤属性空间插值方法及其精度是计量土壤学和"数字土壤"领域的重要研究内容。文章首先介绍了土壤属性空间插值的常用方法,包括克立格插值法(Kriging)、反距离权重法(IDW)、样条插值法(Spline)、贝叶斯最大熵(BME)、地理加权回归(GWR)以及高精度曲面建模方法(HASM);其次阐述了土壤属性空间插值精度验证的方法和指标;再次总结了能够提高土壤属性插值精度的6种途径,包括合理选择插值方法、准确设定插值方法参数、合理设计采样数目和密度、注意空间自相关程度和范围的影响、科学安排实验分析顺序以及结合辅助信息进行插值;最后从插值方法的选择、验证指标的选取以及辅助信息的结合三个方面指出了土壤属性空间插值方法及其精度的未来研究方向。

Research Progress in Soil Property Interpolators and Their Accuracy

Soil property interpolators and their accuracy are regarded as an important topic for the fields of pedometrics and "digital soil".Soil property interpolation is the quantitative monitoring method for spatio-temporal change of soil,and their accuracy is one of the important future research considerations in pedometrics.High accuracy of soil property interpolation can support technique for micro-decision of soil available use and environmental protection.This paper introduces six methods for soil property interpolation,including Kriging,IDW,Spline,Bayesian maximum entropy(BME),geographically weighted regression(GWR) and high accuracy surface modeling(HASM),and presents the validation methods and indices for soil property interpolation.Then we identify six methods for improving the performance of soil property interpolation,including selection of interpolator,setting of interpolator parameter,sampling numbers and density,degree and extent,sequence of analysis in library and combination of secondary information.Finally,we present the future research orientation of soil property interpolation from three aspects,including selection of interpolator,indices and secondary information.