秦巴山地县域土壤碱解氮空间变异与合理采样数的确定

目的]了解秦巴山地县域土壤碱解氮的空间分布规律,确定合理的采样密度,为研究区农田养分管理提供科学依据。方法]运用地统计学与GIS结合的方法,随机抽取不同采样密度的样点数据,进行插值分析,采用交叉验证法对插值精度进行评价。结果](1)土壤碱解氮的变异系数为42.95%,属于中等变异;(2)块金值与基台值的比值约为1/2,具有中等强度的空间相关性,空间最大相关距离为9 171m;(3)样点数目从1 060个到742个时,变程以及块金值与基台值之比出现明显偏差,其相对误差分别为152.32%和36.1%,均方根误差(RMSE)、相关系数(R)同样出现明显偏差。结论]汉滨区土壤碱解氮空间连续性较好,适当地减少采样密度,仍可以满足插值分析的需要,考虑到土壤碱解氮的空间变异评价的精度分析结果,县域土壤碱解氮的合理采样数应该控制在1 060个以上,即最大以345.5hm2为一个采样单元。

County-scale Spatial Variability of Soil Available Nitrogen Distribution and Determination of Reasonable Sampling Density in Qinba Mountains

Objective] Research on the county-scale spatial variability of soil available nitrogen(AN) and determination of reasonable sampling density in order to provide scientific knowledge for nutrient management in the Qinba mountain areas.Methods] Sampling schemes with different sample intensities were simulated by extracting data from an AN database of intensive samples, and then made the spatial interpolation analysis of AN based on geostatic combined on GIS technology, and the interpolation accuracy was evaluated by the cross validation method.Results] (1) The content of AN showed moderate variation with the coefficient of 42.95%; (2) The moderate spatial variability of the AN was found with the ratio of Nugget and Sill of about 1/2 and the maximum spatial self-related distance of 9 171 m; (3) Significant changes were detected when sample number reduced from 1 060 to 742, meanwhile, great changes were found of variation distance and the ratio of Nugget and Sill with the relative errors of 152.32% and 36.1%, and the root mean square error (RMSE) and relative coefficient changed significantly.Conclusion] The range of the AN in Hanbin District was longer than other regions, so spatial continuity was better. It is possible to reduce the number of sample points to satisfy the need of the research regarding spatial variability of AN. Taking the results of spatial prediction accuracy into consideration, more than 1 060 samples should be the suitable number to evaluate the spatial variability of AN on county scale, and the maximum sampling unit is 345.5 hm².