辣椒不同深度土壤墒情预报的BP神经网络模型

BACK-PROPAGATION NEURAL NETWORK MODEL FOR RED PEPPER SOIL MOISTURE FORECAST AT DIFFERENT DEPTHS

依据内蒙古河套灌区治丰试验基地的不同深度土壤水分传感器采样土壤含水率以及气象、地下水、灌水(降雨)数据,建立了基于BP人工神经网络的辣椒同时输出不同深度土层体积含水率预报模型,并用实测土壤含水率数据对模型进行了检验。结果表明,模型具有较好的预报效果,应用于不同深度土壤墒情分析和预报是可行的。

A Back -propagation neural network model for Red Pepper soil moisture forecast at different depths was developed using the soil moisture data measured with soil moisture sensors at different depths and meteorological, groundwater and irrigation （rainfall） data at Zhifeng Experimental Base in Inner Mongolia Hetao Irrigation District. The model was verified with observed soil moisture data. The results showed that the model has a satisfactory forecasting capability and can be used for the analysis and forecast of soil moisture at different depths.