不同采样密度的土壤水分特征参数预测

Predicting Parameters of Soil Moisture at Different Sampling Intensities

利用不同取样精度的土壤，将土壤质地（砂土、淤泥、粘土含量）和容重作为输入值，探讨了使用基于土壤转换函数的BP神经网络模型来预测0～20cm表层土壤水分特征曲线参数，用甘肃省称钩河流域小流域的土样进行预测并进行了误差分析。结果表明，使用线性回归能够减小预测误差与实测值差距；使用BP神经网络来预测饱和体积含水量，其准确性比使用BP神经网络预测剩余体积含水量和田间持水量要高。为了进一步提高预测精度，还应尽可能地包括土壤结构、有机质含量等信息。

Soil was sampled at different precision, and soil texture （sand, silt, clay content） and bulk den- sity were used as inputs to predict soil water retention curve parameters of depths （0-20 cm） in top-soil by BP artificial neural networks, which is based on pedotransfer functions. Furthermore, soil samples collected from watershed of Chenggouhe in Gansu province were used to predict soil water retention curve parameters and to analysis on their bias. The result indicated that linear regression can be used to reduce the bias of prediction parameters and tested parameters; the accuracy of predicting saturated volumetric capacity was better than the accuracy of predicting soil available water capacity and field moisture capacity by BP artificial neural networks. At last, we think more information was needed to improve the accuracy, such as soil structure, organic matter contents and so on.