摘要
以榆阳地区土壤有机质、水解性氮、有效磷、速效钾、全氮等5种养分元素为例,选取常用且具有代表性的反距离加权法、样条函数法、普通克里格法3种空间插值方法,对土壤元素进行空间插值,并对其结果进行验证分析和评价。在对各方法参数优化的基础上,采用全交叉验证法对1 459个土壤样本的插值结果进行验证分析。插值结果表明,插值效果受采样点密度影响较大,在样点密集区可得到较好效果,在采样点稀疏的地区插值结果较差。对比研究表明,普通克里格法对刻画区域土壤养分元素的空间分布趋势效果最佳,但其半变异函数模型及参数的优化较为复杂,仍有待进一步研究;反距离加权法和样条函数法对土壤元素分布的空间插值精度一般,但其简单易用、插值最优参数易于选择。
Taking five kinds of soil nutrient elements including soil organic matter,hydrolyzable nitrogen,available phosphorus,available potassium,total nitrogen in Yuyang County of Shaanxi Province as an example,we made spatial interpolation for soil nutrient elements by using three types of representative interpolation methods,including Inverse Distance Weighted（IDW）,Spline and Ordinary Kriging and validated and appraised the results.And a comparison is made between methods with optimized parameters.There were 1459 samples used for interpolation,and cross validation method was used to verify and analyze interpolation.It is showed that the density of samples has greater effects on interpolation results： better results in intensive samples area while worse results in sparse samples area.The results of comparison indicate that ordinary Kriging exhibits best effect in characterizing spatial distribution trend of soil nutrient elements, yet the optimized models and parameters of semi-variogram are still pending for further study;Inverse Distance Weighted and Spline interpolate less accurately for the spatial distribution of soil nutrient elements,but they are easy to use and to select optimized parameters.
出处
《干旱地区农业研究》
CSCD
北大核心
2010年第2期177-182,共6页
Agricultural Research in the Arid Areas
基金
国家重点基础研究发展计划"973"项目"区域水土流失过程与趋势分析"(2007CB407203)
"十一五"国家科技支撑计划项目(2006BAD09B0603)
国家自然科学基金(30872073)
