论文部分内容阅读
精确预测紫色土区土壤有机质含量的空间分布,对于指导紫色土区农业生产和培肥土壤具有重要意义。以杜家沟小流域为研究区,以遥感影像作辅助变量,采用回归克里格法,预测土壤有机质含量的空间分布,并与参照方法的预测精度进行比较。结果表明:(1)Landsat ETM+影像的波段2和波段5是土壤有机质含量多元线性回归预测的最佳辅助变量,回归残差的最优半方差函数模型为球状模型,模型的拟合精度较高;(2)土壤有机质含量呈由沟谷逐渐向坡顶递减的趋势,空间变异的细节信息表达较好;(3)回归克里格法在验证点的预测值与实测值的拟合能力更好,预测结果更倾向于无偏的,MAE、RMSE和R2均优于参照方法。因此,回归克里格法是紫色土区土壤有机质含量高精度空间预测的有效方法。
Precise prediction of the spatial distribution of soil organic matter in purple soil is of great significance for guiding the agricultural production and fertilizing soil in purple soil area. Taking Dujiagou small watershed as the study area and the remote sensing image as the auxiliary variable, the regression Kriging method was used to predict the spatial distribution of soil organic matter content, and compared with the prediction accuracy of the reference method. The results show that: (1) Band 2 and Band 5 of Landsat ETM + image are the best auxiliary variables for multivariate linear regression prediction of soil organic matter content. The optimal semivariance function model of regression residual is spherical model, and the fitting accuracy of the model is high ; (2) The content of soil organic matter tends to decrease gradually from the valley to the top of the slope, and the details of the spatial variability are well expressed; (3) The regression kriging method has better fitting ability between the predicted value and the measured value , The prediction results tend to be unbiased, MAE, RMSE and R2 are better than the reference method. Therefore, returning to Kriging is an effective method to predict the high-precision soil organic matter content in purple soil.