Research on Slope Soil Erosion Based on Manpower Neural Network

Based on the complex nonlinear characteristics of slope soil erosion, method of artificial neural network was used, and three-layer feed-forward back-propagation network model for slope soil erosion in different tillage measures (contour tillage, manpower digging, manpower hoeing, linear slope) was established. The structure of the model has five input variables including rainfall intensity, gradient, length of slope, percentage of prophase soil moisture content and soil volume weight and one output variable for the sediment yield of secondary rainfall of slope soil erosion. The network model was trained and predicted by using the observed data of the field simulated rainfall experiment. The results showed that back-propagation network model was reasonable and can be referred as an effective method for studying slope soil erosion laws.