饲草料地土壤水分动态变化规律及其预测的人工神经网络模型的研究

Dynamic Patterns of Soil Moisture in Forage Land and Modelling of Soil Water Prediction with Neural Network

根据2005～2006年实测土壤水分资料和气象资料,研究分析了饲草料地土壤水分的动态变化规律和建立了考虑多个因素对土壤水分影响的BP人工神经网络模型,结果表明:表层土壤的含水率变幅较大,主要是受大气降水的影响,20～40cm和40～60cm土层土壤含水率的波动,除大气降水的影响外,还与植物的生长发育状况有关;土壤水分预测模型具有较好的预测效果,用神经网络建立土壤水分预测模型的方法是可行的,对于不同条件的地区具有广泛的适应性和推广应用前景。

Based on the data of soil water contents in forage land and weather obseration obseration nearby during 2005 to 2006, a back-propagation（BP） neural network model for soil water content forecast was established.The results showed that the water content of surface soil moisture fluctuated violently, due to the precipitation and developing of plants. The predictied model could forecast soil water changes with neural network.