水土保持研究2011，Vol.18Issue(2)：1-5,5.
基于BP神经网络模型的川中丘陵区坡耕地土壤侵蚀预测研究
Research of Soil Erosion Prediction Based on Back Propagation Neural Network Model in the Hilly Region of Central Sichuan
摘要
Abstract
Based on back propagation (BP) neural network, the soil erosion was predicted according to the observational data from 1991 to 2000 of the six runoff plots at Suining Soil and Water Conservation Experiment Station in the central Sichuan Hilly Region. Five factors (rainfall, rainfall duration, rainfall intensity, vegetation cover, slope),nine factors (five factors plus early rainfall, early rainfall duration, pre-rainfall intensity and time interval before and after the rain), and ten factors (nine factors added to water conservation measures) are input respectively, and erosion is output in the model. The results showed as follows: the determination coefficient of 5 factors on soil erosion was 0.52, and the coefficient of efficiency of those was 0.48, which were not satisfactory, while those of 9 factors on soil erosion were 0.53 and 0.48, which were not good either, but those of 10 factors on soil erosion were 0.57 and 0.70, which achieved a satisfactory prediction results and business forecasting.关键词
川中丘陵区/坡耕地/土壤侵蚀/BP神经网络模型Key words
Central Sichuan hilly region/ slop farmland/ soil erosion/ back propagation neural network model分类
农业科技引用本文复制引用
朱雪梅,冯海峰,林立金,杨远祥,邵继荣,朱波,冯济敏..基于BP神经网络模型的川中丘陵区坡耕地土壤侵蚀预测研究[J].水土保持研究,2011,18(2):1-5,5.基金项目
国家科技支撑计划项目(2008BAD98B05) （2008BAD98B05）
