基于克里金插值的耕地表层土壤有机质空间预测

基于环境变量与耕地表层土壤有机质的相关分析,分别采用普通克里金(OK)、逐步回归克里金(SWRK)和地理加权回归克里金(GWRK),对吉林省扶余市耕地表层土壤有机质进行了空间预测。结果表明:扶余市耕地表层土壤有机质分布存在中等空间自相关性,受资源开发、耕作方式等随机因素影响较小;受曲率和植被长势影响大,受一阶微观地形、复合宏观地形及汇流和沉积物运移影响小;在海拔高、植被少且土壤水分不足的阳面通风山丘区含量低;空间分布以西南部为低值中心向四周升高,高值区出现在东部边界。GWRK中环境变量参与了局部回归,其模拟精度最高且揭示了更多空间变化细节。本研究的结果可为较大空间尺度的耕地表层土壤有机质空间变异与预测研究提供可靠的方法借鉴。

Prediction of Spatial Distribution of Topsoil Organic Matter Content in Cultivated Land Using Kriging Methods

In this study,the Ordinary Kriging (OK),Step-wise Regression Kriging (SWRK) and Geographically Weighted Regression Kriging (GWRK) were used to predict the spatial distribution of topsoil organic matter content in cultivated land in Fuyu City,Jilin Province based on the correlation analysis of environmental variables and topsoil organic matter content.Results showed that there was a moderate spatial autocorrelation of topsoil organic matter content in cultivated land in the study area,the topsoil organic matter content was affected slightly by random factors,such as resources exploitation and cultivation measures,but strongly by curvature and vegetation growth.The topsoil organic matter content was also affected slightly by first-derivative micro topography,complex macro topography,confluence and sediment transport.The topsoil organic matter content in cultivated land was low in arid sunny open slope of high hill,it was gradually increased from the southwest part to the surrounding areas of the study area,and the high value occurred at the eastern boundary.Environmental variables in GWRK were taken into the local regression,the simulation precision was the highest,and more spatial variation was revealed.The results could be referred in studying the spatial variability and prediction of topsoil organic matter content at large scale.