论文部分内容阅读
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。
Taking the surface soil organic matter content in Fengqiu County of Henan Province as an example, the authors discussed the descriptive statistical characteristics of regional soil variables, the fitting effect of the theoretical model of semi-variance function, the accuracy of the prediction results of ordinary Kriging interpolation and the spatial variability of the target variables Ability and many other effects. The results show that the number of samples dropped significantly from 5000 to 20, and the mean value of organic matter in surface samples of the study area did not change significantly. When the number of soil samples is more than 625, the semivariogram function model of surface soil organic matter content has a good fitting effect. Kriging interpolation can be used to obtain the prediction results with high accuracy and strong ability to explain the spatial variability of the target variables. When the soil sample size is less than or equal to 78, the semivariogram function model can not be obtained theoretically by fitting. The Kriging interpolation method can not obtain the ideal prediction result of surface soil organic matter content in the study area.