---

作 者 ：孙义祥^1,2;吴传洲³;朱克保³;崔振岭¹;陈新平¹;张福锁¹

期 刊 ：应用生态学报 2009年 20卷 03期 页码：673-678

关键词：土壤；

摘 要 ：与欧美大规模农场经营不同，土地分散经营使我国县域土壤养分空间变异特征评价更加困难.本研究以安徽芜湖土壤有效磷为例，系统地评价插值方法与样点数对县域土壤养分空间变异特征评价准确性的影响.结果表明：局部多项式、普通克里格、简单克里格和析取克里格插值方法的评价效果优于反距离加权法、全局多项式、径向基插值和泛克里格等插值方法，考虑到实际操作简单，推荐用普通克里格方法进行县域土壤有效磷空间变异特征评价.随着参与空间插值样点数的增加，县域土壤有效磷空间变异特征预测的准确性增加，充分考虑评价的准确性和田间取样费用，建议县域土壤有效磷空间变异特征评价的适宜样点数应介于500～1000个.

Abstract:Different from the large scale farm management in Europe and America, the scattered farmland management in China made the spatial variability of soil nutrients at county scale in this country more challenging. Taking soil Olsen-P in Wuhu County as an example, the influence of interpolation method and sampling number on the spatial prediction accuracy of soil nutrients was evaluated systematically. The results showed that local polynomial method, ordinary kriging, simple kriging, and disjunctive kriging had higher spatial prediction accuracy than the other interpolation methods. Considering of its simplicity, ordinary kriging was recommended to evaluate the spatial variability of soil Olsen-P within a county. The spatial prediction accuracy would increase with increasing soil sampling number. Taking the spatial prediction accuracy and soil sampling cost into consideration, the optimal sampling number should be from 500 to 1000 to evaluate the spatial variability of soil Olsen-P at county scale.

---