论文部分内容阅读
为了对不同土壤入渗模型参数多元非线性预测模型的精度进行对比分析,基于区域性黄土田间耕作土壤水分入渗实测资料,分别对Kostiakov、Kostiakov-Lewis以及Philip入渗模型的参数建立了以土壤体积含水率、干密度,黏粒、粉粒和有机质含量等土壤基本理化参数为输入变量的多元非线性传输函数,并进行了精度分析和比较。结果表明:Kostiakov入渗的二参数模型参数的平均误差在9%以下,预测精度明显高于Kostiakov-Lewis三参数入渗模型参数10.78%和Philip入渗模型参数11.72%。因此,推荐形式简单且预测精度高的二参数模型参数多元非线性预测模型为土壤入渗模型参数预测模型。
In order to compare the accuracy of multivariate nonlinear prediction models of different soil infiltration model parameters, based on the measured data of tillage soil moisture infiltration in regional loess fields, the parameters of Kostiakov, Kostiakov-Lewis and Philip infiltration models were established respectively. The soil physical and chemical parameters, such as volumetric water content, dry density, clay, grain and organic matter content, are the multivariate nonlinear transfer functions of input variables, and the accuracy analysis and comparison are made. The results show that the average error of Kostiakov infiltration parameters is less than 9%, and the prediction accuracy is significantly higher than that of Kostiakov-Lewis parameter 10.78% and Philip infiltration model 11.72%. Therefore, the multivariate nonlinear prediction model of two-parameter model with simple recommendation and high prediction accuracy is the parameter prediction model of soil infiltration model.