三峡库区紫色土陡坡耕地土壤侵蚀量预测模型研究

Study on the prediction model of soil loss in purple soil sloping land in Three-gorges Reservoir area

以径流小区观测资料为基础,运用BP神经网络(Back Propagation Neural Network)方法对三峡库区紫色土陡坡耕地土壤流失量预测模型进行了模拟研究;建立了以次降雨量、径流量、前期土壤含水量、土壤容重和土壤孔隙度为输入层,以土壤侵蚀量为输出量的3层BP神经网络预测模型;通过模型学习和模拟计算进行检验,显示出了较好的预测效果。

The application of back-propagation neural network to prediction of soil loss is discussed in in purple soil sloping land in Three Gorges Reservoir area,based on the study of field experience,we established a model to predict the soil loss amount which includes five inputs and a output.The results is examined and showed the neural network method had a high prediction capability.