基于不同采样方式棉田养分空间变异特征分析

Analysis of the Characteristics of Nutrient Spatial Variability in Cotton Fields Based on Different Sampling Methods

【目的】基于GIS和地统计学,对比网格采样法和随机采样法条件下棉田养分的空间变异特征。【方法】分别用两种采样方式采集研究区域耕层土壤样品,以土壤全氮、速效磷和速效钾含量为基础数据进行描述性统计分析,并运用地统计学技术方法对其进行半方差函数模型的拟合;进而结合Arc GIS中的Kriging插值法进行插值,获得棉田养分分布图。【结果】研究区域的棉田土壤属性服从正态分布,且网格采样法中的N、P、K的变异系数均大于随机采样法;并且网格采样法中的速效磷具有强烈的空间自相关性;全氮和速效钾具有中等的空间自相关性,而随机采样法中的N、P、K三种养分的空间自相关性都很弱;网格采样法中土壤养分的相关变程均大于土壤采样间距。网格采样法插值结果表现较随机采样法的更细微,其中速效磷反应的最为明显。【结论】与随机采样法相比,网格采样法中的土壤养分的空间变异结果能更好的反应研究区域的空间变异特征,可为棉田后期土壤养分分区管理、变量施肥等提供可靠的科学依据。

[ Objective ] The purpose of this project aims to compare the characteristics of the nutrient spa- tial variation under grid sampling method and random sampling method in cotton fields according to GIS and statistics. [ Method] By means of two kinds of sampling method to collect soil samples in the arable layer of the study area and based on soil total nitrogen, available phosphorus and available potassium content, the descriptive statistical analysis of data was carried out. And what is more, the geostatistical techniques was used to fit their method semivariogram model; Then with the Kriging interpolation method was applied for interpolation in ArcGIS to obtain nutrients cotton field distribution. [Result] Cotton soil property of the study area was normally distributed and grid sampling method of N, P, K coefficient of variation was greater than the random sam- piing method; and available P in grid sampling method had a strong spatial autocorrelation; nitrogen and potassium displayed moderate spatial autocorrelation, while the random sampling method in the space of three kinds of N, P, K nutrient spatial autocorrelation were weak. In grid sampling method, the relative variation of soil nutrients was greater than that of the sampling interval. Grid sampling interpolation results showed that a random sampling method was more subtle, and available phosphorus reaction was the most obvious. [ Conclusion ] Compared with the random sampling method, the spatial variation analysis results of the spatial variability of soil nutrient was regional grid sampling method, which could be better for providing a reliable scientific basis for the regional management of soil nutrients, fertilizer and other variables in cotton fields.