中文摘要：

以河南省封丘县表层土壤有机质含量为例，探讨土壤样点密度对区域化土壤变量描述性统计特征、半方差函数理论模型拟合效果、普通Kriging插值预测结果的精度与表现目标变量空间变异的能力等多方面的影响。研究结果表明，样点数量从5000个大幅减少至20个，研究区表层样品有机质含量均值未发生显著变化。当土壤样点≥625个时，表层土壤有机质含量半方差函数模型具有较好的拟合效果，可以通过Kriging插值手段获得精度较高且对目标变量空间变异特征解释能力较强的预测结果；当土壤样点≤78个时，半方差函数模型理论上无法通过拟合获得，通过普通Kriging插值手段不能获得研究区表层土壤有机质含量理想的预测结果。

英文摘要：

Taking the variation of organic matter contents in topsoil of Fengqiu County of Henan Province as an example, the influences of soil sample density on the descriptive statistics and semivariogram model-fitting of the regionalized soil variables were discussed. And, the relationship between sample density and the precision and interpretative capacity of Kriging interpolation was demonstrated. The results indicated that, no evident change was found in the comparison of mean content of organic matter in topsoil samples of the studied area, when sample numbers decreasing sharply from 5000 to 20. Under a condition of soil-sampled sites more than 625, the semivariogram of the organic matter content in topsoil of the study area could be properly fitted and the soil attribute could be predicted with a desirable precision and a great interpretative capacity to spatial variation. When the numbers of soil-sampled sites were less than 78, however, the semivariogram could not be fitted through the scatter diagrams of soil sample variograms and their separation distances, and Kriging interpolation could not bring an acceptable predictive output of organic matter content in topsoil of the studies area.