以延河流域为例,运用主成分分析法筛选出影响流域社会经济需水的主要指标,并根据1980-2000年各项指标的统计数据,在Matlab平台下,采用具有自学习、自适应特点的BP神经网络模型,预测了流域在2010、2020以及2030年的社会经济需水量,以期为完善流域水资源信息系统,平衡水资源供给与需求,实现未来规划年水资源合理调配做出科学指导。结果表明:随着流域人口的持续增长和经济的快速发展,未来年其社会经济各项需水量均呈现逐年增加的趋势。2010-2030年,流域居民的生活需水将由1785.40万m^3增加为3231.98万m^3,而且城镇居民的需水百分比不断增大;农业生产需水将由1018.05万m^3增加为1193.27万m^3,主要原因在于农业内部结构中林果产业和畜牧业等相对高用水产业比重有所增加;工业生产需水将由2741.16万m^3增加为2853.41万m^3,集中表现为高用水工业生产需水增加迅速。
In this research,Yanhe watershed which is as the representative arid and semi-arid region is chosen to study.We firstly select the main indices that influence the water need for socioeconomic development by using principal component analysis.Additionally,based on these indices from 1980-2000,we predict and analyze the water need for socioeconomic development in the year of 2010,2020 and 2030 by employing the Back-Propagation Network at the Matlab worktable.These jobs are important to perfect the water resource information system,balance the supply and requirement of water resource and allocate it scientifically.The results show:(1) In the future,the water demand for socioeconomic progress will increase year by year with the development of population and economy.(2) During 2010-2030,the water demand for people life will increase from 1785.40×10^4m^3 to 3231.98×10^4m^3,and the proportion of water demand for city life will ascend continuously;the water demand for agriculture production will increase from 1018.05×10^4m^3 to 1193.27×10^4m^3,which is because of the expansion of forestry and stockbreeding;the water demand for industry production will befrom 2741.16×10^4m^3 to 2853.41×10^4m^3,which is due to the augment of some industrial departments needing more water.