黄河上游天然年径流长期变化趋势预测  

Prediction of Long-Term Variation Trend of Natural Annual Runoff of the Upper Yellow River

作　　者：张少文[1] 张学成[2] 德格吉玛 王玲[2] 丁晶[1] 

机构地区：[1]北京市水利科学研究所 [2]黄河水文水资源科学研究院,河南郑州450004 [3]北京水利水电学校,北京100024

出　　处：《人民黄河》2007年第1期27-29,共3页Yellow River

基　　金：国家自然科学重点基金资助项目(50239050);国家自然科学基金资助项目(50249024;40271024)

摘　　要：基于神经网络理论与传统分析方法，采用改进的BP网络算法，利用黄河青铜峡水文站278年（1723～2000年）的天然年径流时间序列，建立了黄河上游年径流长期变化的BP网络预测模型，并以此模型对2001～2050年黄河上游天然年径流变化趋势进行预测分析，结论认为：①黄河上游天然年径流在未来50a变化的大趋势是丰水时段占优势，且黄河干流天然年径流变化要经历3个不同的时段，即2001—2013年以多年平均水平为基准上下波动的时段、2014—2037年相对丰水时段和2038，2050年相对枯水时段；②用神经网络理论建立了黄河天然年径流中长期BP网络预测模型，模型预测分析的未来几十年黄河天然年径流变化的大趋势是可信的。It establishes BP network predictive model of long-term variation of annual runoff of the upper Yellow River based on the theory of neural network and conventional analytical method and by using the improved BP network arithmetic and 278-year time series(1723~2000) of natural annual runoff of Qingtongxia Hydrometric Station on the Yellow River.It predicts and analyzes the variation trend of natural annual runoff of the upper Yellow River in 2001~2050 period and concludes that a) the major trend of variation in the coming 50 years is mainly wet-year time interval and the variation of natural annual runoff of the main river will go through three different time intervals,i.e.2001~2013 fluctuation time interval of taking many years average as a basis,2014~2037 time interval of relatively wet and 2038~2050 time interval of relatively dry and;b) the BP network predictive model of medium and long-term natural annual runoff of the Yellow River is established by using neural network theory.The major trend of the natural annual runoff in the coming decades of the river predicted and analyzed by the model is credible.

关 键 词：BP网络 天然年径流 长期预测 黄河上游 

分 类 号：P333[天文地球—水文科学] TV882.1[水利工程—水文学及水资源]