【摘要】 本文提出一种结合元胞自动机的区域水土流失预测模型,采用线性神经网络方法预测区域土地利用总体变化趋势;同时考虑道路、商服中心对城市土地利用覆盖的影响,制定元胞自动机转换规则,预测区域未来土地利用覆盖情况,预测精度达到72%。最后,将CA预测数据作为变量输入通用土壤流失方程,预测区域水土流失情况。并以武汉市东湖流域为例对其水土流失问题进行了现状分析及预测,为其水土流失整治提供参考。更多还原

【Abstract】 This paper puts forth a new tool,cellular automata(CA),and a universal soil loss equation(USLE) for forecasting of strong spatial-temporal dynamic evolution of soil loss.Linear neural network was adopted to develop an USLE model for prediction of the tendency of land use change,considering the impact of road and commercial center construction on the development of land use.Based on the previous work,transformation rules in CA model were identified to improve prediction accuracy.Calculations show that the predicted land use coverage has an accuracy of 72%.Finally,the predicted land use data was inputted into the USLE model to predict the distribution of future soil erosion.As a case study,the existing soil erosion in the Donghu watershed in Wuhan and its future development were calculated and analyzed.更多还原