基于地理因子的因旱饮水困难人口快速评估模型——以云南省2012年大旱为例

Geographical Factor Based Rapid Assessment Model of Population in Drinking Water Access Difficulties Because of Drought——A Case Study of 2012 Yunnan Extreme Drought

农村群众饮水困难一直是国家持续关注的问题,保障受旱灾区人民群众基本生活是重中之重。准确地预测因旱饮水困难人口的发生发展,并采取适当的减灾措施,将很大程度地降低经济损失与人员伤害。以2012年云南大旱为例,在全面收集研究区气象、基础地理信息、社会经济、灾情等资料的基础上,基于BP人工神经网络构建因旱饮水困难人口快速评估模型。以月均降水量、地形高程、水系密度、总人口、路网密度和GDP为网络输入,以因旱饮水困难人口数量为网络输出,选择30个受灾县样本进行网络训练,经试算不断优化模型参数,其MSE为0.003 6;通过训练好的模型来预测剩余55个受灾县的因旱饮水困难人口,其模拟值与实际值的线性拟合结果 R2为0.67,表明BP人工神经网络方法能有效预测因旱饮水困难人口,该方法能够为因旱饮水困难人口的快速评估和灾情的核查提供有效的借鉴。

The problem of rural drinking water difficulty has been continuingly concerned in China.Keeping the basic living conditions is most important for people in the drought-hit disaster areas.Predicting accurately the population in drinking water access difficulties because of drought and taking appropriate mitigation measures can minimize economic loss and personal injury.Taking 2012 Yunnan Extreme Drought as an example,on the basis of collecting the meteorological,basic geographic information,socio-economic data,and disaster effect data of the study area,a rapid assessment model based on BP neural network is constructed.The six factors are the input of network,which are the average monthly precipitation,DEM,river density,the total population,road density and GDP.The population in drinking water access difficulties because of drought is the output of network.Taking 30 drought-affected counties samples for network training,under optimizing the model parameters,the MSE is 0.003 6;by the trained model to predict the population in drinking water access difficulties of remaining 55 drought-affected counties.The fitting result of R2 between the analog value and the true value was 0.67.It shows that the BP artificial neural network method can effectively predict the population in drinking water access difficulties because of drought.The method may provide an effective reference for rapid assessment and disaster verification of the population in drinking water access difficulties because of drought.