BP神经网络在降雨侵蚀力预测预报中的应用研究

降雨侵蚀力反映由降雨引起土壤侵蚀的潜在能力，是建立通用土壤流失方程USLE的最基本因子之一。由于降雨侵蚀力计算过程中所需资料较难收集，给其计算增加了难度。运用BP神经网络方法对降雨侵蚀力与地理之间的关系进行研究，建立降雨侵蚀力BP神经网络模型。对福建省46个地域的降雨侵蚀力进行研究，结果表明:所建立的降雨侵蚀力BP神经网络模型对模拟预测福建不同地域的降雨侵蚀力，平均模拟精度为96.81%，平均预测精度为95.68%，达到了较为理想的效果。这不仅为降雨侵蚀力的预测预报提供了科学依据，而且也为BP神经网络在水土保持研究中的应用开辟了新的思路。

Application of Prediction for Rainfall Erosivity Based on BP Neural Network

Rainfall erosivity reflects the potential ability of the soil loss caused by rainfall and it is very important for predicting soil loss quantitatively.It is difficult to calculate rainfall erosivity because it is difficult to collect the data needed.This paper studied the relationships between rainfall erosivity and geographic factors by BP neural network,and brought forward rainfall erosivity BP neural network model.The model was applied to study the rainfall erosivity of 46 regions in Fujian Province,and the results showed that the accuracy of the mean simulative accuracy and the mean predictive accuracy of the model for different regions in Fujian Province were satisfactory,which were 96.81% and 95.68% respectively. Therefore,it not only provided a scientific basis in predicting rainfall erosivity,but also opened up train of thought in the application of BP neural network in soil and water conservation research.