土壤水分特征曲线模型参数的非线性传输函数研究

A Study on Nonlinear Transfer Function of Soil Moisture Characteristic Curve Model

以黄土高原区土壤为研究对象,通过土壤水分特征曲线与土壤基本理化参数系列试验,获得了Van-Genuchten模型参数的实测数据样本,建立了土壤基本理化参数与Van-Genuchten模型参数之间一一对应的关系,创建了以土壤黏粒含量、粉粒含量、容重、有机质含量、全盐量为输入变量,Van-Genuchten模型参数为输出变量的非线性传输函数预报模型。研究表明:以土壤黏粒含量、粉粒含量、容重、有机质含量、全盐量为输入变量,对Van-Genuchten模型参数进行非线性预报是可行的;所建立的非线性预报模型精度较高,预测样本下Van-Genuchten模型参数α的实测值与预测值的相对误差的平均值为9.66%,参数n的实测值与预测值的相对误差的平均值为6.83%,检验样本参数α的实测值与预测值的相对误差的平均值为7.34%,参数n的实测值与预测值的相对误差的平均值为5.45%。研究成果为黄土地区提供一种便捷获取土壤水分特征曲线的途径。

Taking soil in Loess Plateau region as the research object, the soil water characteristic curve and soil basic physical and chemical parameters are tested and the measured data samples of Van-Genuchten model parameters are obtained. The relationship between the basic physical and chemical parameters of soil and the parameters of Van-Genuchten model is established. A nonlinear transfer function prediction model with soil clay content, silt content, bulk density, organic matter content, total salt content as input variable, and Van-Genuchten model parameter as output variable is established. The research result shows that taking soil clay content, silt content, bulk density, organic matter content and total salt as input variables is feasible to carry out nonlinear prediction of Van-Genuchten model parameters; the established nonlinear prediction model has higher accuracy, the average relative error between the measured and predicted values of the Van -Genuchten model parameter α in the prediction sample is 9.66%, the average relative error between the measured and predicted values of parameter n is 6.83%, the average value of the relative error between the measured and predicted values of the sample parameter a is 7.34%, and the average relative error between the measured and predicted values of the parameter n is 5.45%. The research result provides a convenient way to obtain soil water characteristic curve for the Loess area.