空间变异在土壤性质长期定位观测及取样中的应用

通过对中国科学院沈阳生态实验站30m×42m样地进行网格法分层取样,采用经典统计学与地统计学相结合的方法,以土壤pH和有机碳(TOC)为例,探讨了空间变异在土壤性质定位试验及取样中的应用。结果表明,49对样本的pH和TOC变异系数均小于10%,变量在0~10cm和10~20cm两个层次的基本统计数值基本接近,最佳理论模型的参数显示各变量空间变异主要受结构性因素的影响,且以135°方向的变异为主;相对误差为10%时采集6个和10个土样可分别达到95%和99%的置信水平,结合空间分布格局可以确定更为精确的取样方案;综合分析表明待选样地可以作为土壤性质长期定位观测试验用地。

Application of Spatial Variability in Long-term Site-specific Observatory Study of Soil Properties and Sampling Strategy

Taking soil pH and total organic carbon (TOC) as test parameters and using traditional statistics combined with geostatistics, the application of spatial variability in site-specific experiment of soil properties was examined in the paper. A sampling grids of 30 m42 m was established in Shenyang Experimental Station of Ecology, Chinese Academy of Sciences. The results showed that the variance coefficients of 49 pairs of soil pH and TOC were less than 10%, and the fundamental statistical values of the variables were similar. Parameters derived from the best-fitted models showed that the spatial variability of soil properties were influenced mainly by structural factors, and the anisotropic variability was mainly taken place in 135direction. Taken 10% as relative error, collecting 6 and 10 samples could reach the significance level of 95% and 99%, respectively. A more precise sampling scheme could be made based on spatial distribution pattern of test soil properties. As a result, the test field was regarded as a suitable one for long-term experiment of soil properties.