---

作 者 ：高峻;黄元仿;李保国;张世熔;张凤荣

期 刊 ：植物营养与肥料学报 2003年 9卷 2期 页码：151-157

关键词：普通Kriging；土壤颗粒组成；空间变异；趋势分析；

Keywords:ordinary kriging, soil particle composition, spatial variability, trend analysis,

摘 要 ：对60m×55m的农田尺度上100个取样点土壤颗粒组成及剖面层次的空间变异性的传统统计分析表明,各属性的变异系数属中等强度。半方差函数模型均为球状模型,各属性具有一定的空间相关性,并表现有一定的各向异性。在对土壤颗粒组成及剖面层次空间分布趋势分析的基础上,根据半方差函数模型,对各属性选择不同趋势和异向性的普通Kriging内插值比较的结果表明,考虑变量在空间分布趋势和异向性的插值结果比不考虑要好。在以上分析基础上,利用Kriging内插值绘制各层深度和颗粒组成的等值线图,并分析了其空间分布规律。

Abstract:The spatial variability of soil particle compositions and soil profile horizons were analyzed in the 60 m×55 m field. Conventional statistical analysis of 100 sampled points showed that coefficients of variation of those properties are medium. Semivariance model analysis for those soil parameters showed that their semivariance structures were spherical. These properties were correlated in different spatial ranges with different anisotropy. Base on Semivariogram and the trend analysis for those soil parameters, this paper compared interpolation method of Ordinary Kriging with different trend and anisotropy. The results showed that the Kriging methods considering trend and anisotropy were better than unconsidering trend and anisotropy. Finally, Ordinary Kriging with trend and anisotropy was applied to workout the contour map of soil particle compositions and soil profile horizons of unknown points, and their spatial distribution was analyzed.

---