单因素土壤墒情预测模型研究  

The Study on Soil Moisture Forecast Model with Single Factor

作　　者：王勇志[1] 赵燕东[1] 马扬飞[1] 

机构地区：[1]北京林业大学工学院,北京100083

出　　处：《灌溉排水学报》2013年第2期134-137,共4页Journal of Irrigation and Drainage

基　　金：北京市教育委员会共建项目建设计划科学研究与科研基地建设项目(2008BJKY01)

摘　　要：以土壤10、30、50cm深度处的土壤墒情各当前和历史数据不同组合为输入,以30cm处1h后的土壤墒情为预测输出,建立了基于BP神经网络的单因素土壤墒情预测模型。结果表明,模型预测误差约为10%,取得了较好的预测效果。A soil moisture forecast model with single factor, based on BP neural network, was eatablished by taking different combinations of current and past soil moisture contents at a depth of 10 cm, 30 cm and 50 cm as input, and the soil moisture content an hour later at a depth of 30 cm as output. Results showed that the prediction error was about 10%, and the forecast model obtained better forecasting results. It could provided a new way for the establishment of soil moisture prediction model, and laid the foundation for precision irrigation.

关 键 词：土壤墒情 BP网络 单因素 预测模型 

分 类 号：S152.7[农业科学—土壤学]