基于BP神经网络模型的三江平原湿地面积预测研究

以相关资料和文献中1950-2009年的数据为依据,利用相关分析、灰色预测方法和BP神经网络模型,定量分析了三江平原湿地面积变化的影响因素,并对未来20年三江平原的湿地面积进行预测。结果表明,耕地面积、年均气温、人口数量、径流深度、相对湿度、年降水量、政策因子与湿地面积变化关系密切,可以作为BP神经网络模型的预测变量。BP神经网络模型预测2010s三江平原的湿地面积为58.58万hm2,2020s为45.86万hm2,湿地面积在未来一段时间还有减少的趋势,但减少趋势明显减缓。

Forecasting research of Sanjiang Plain wetland area based on the BP neural network model

The global wetlands area has decreased during the recent decades.Wetlands changes and trend prediction become the research hotspot now.Based on the correlation analysis,Grey prediction method and BP neural networks,we did a quantitative analysis at the factors that affect the changing of Sanjiang Plain wetland and forecasted of area of wet land in the following 20 years.The result indicated that the changing of wetland had close relationship with cultivated area,temperature,population quantity,runoff depth,relative humidity,rainfall and policy,which could be used as the forecasting evidence of the BP neural network model.The BP neural network forecasted that the area of Sanjiang Plain wetland was 585,800 hm2 in 2010s and will be 458,600 hm2 in 2020s.The area showed the reducing tendency in the future and the tendency will slow down.