基于BP网络的土壤墒情预报模型研究

Study on Forecasting Soil Moisture Content Based on BP web

本文对土壤墒情(旱情)预测预报的工作方法做了叙述。通过北京大兴芦城试验基地的不同深度土壤水分传感器采样试验,以容易测量得到的4个深度的土壤体积含水率和日平均气温为输入量,提出了一个基于BP网络的、可以对年内各个季节墒情做出短期预报的模型,以1周、1d为时间间隔,仿真结果表明预报误差很小,表明模型用于土壤墒情预报(旱涝预报)是可行的,是一种简单、易于推广的墒情(旱情)预报方法。

This paper narrates how to forecast soil moisture content. By the experiments through moisture sensor from different depths of the soil in Lucheng Experiment Base in Daxing Town of Beijing, the model based on BP web and that could give forecast of soil moisture content of every season in a year is studied according to the moisture rate in soil body of four soil depths and daily average temperature. By the intermission of a week or a day, the study result shows that forecast error is little, that is the forecast soil moisture content model can be used, and it is an easy method that can be spread to forecast soil moisture content.