用小波神经网络耦合模型预测流域年均产沙量  

Prediction of average annual sediment yield of river basin based on wavelet neural network coupling model

作　　者：丁文荣[1] 周跃[2] 周晓路 

机构地区：[1]云南师范大学旅游与地理科学学院,云南昆明650092 [2]云南财经大学气候流域管理中心,云南昆明650211 [3]国立新加坡大学地理系,新加坡119260

出　　处：《人民长江》2010年第1期74-78,共5页Yangtze River

基　　金：国家重点基础研究发展计划"973"项目(2003CB415105);云南省自然科学基金项目(2009CD053)

摘　　要：引入小波神经网络耦合模型对流域年均产沙量进行了定量研究。由于流域的地质、地貌、土壤在一定时间尺度内具有相对稳定的特性,选出年降雨量、年均气温、年径流量、大雨降雨量、暴雨降雨量、蒸发量、日照时数和汛期降雨量等8个要素作为模型的气候水文输入因子,而以耕地面积、林地面积、水库库容、修建公路、水土保持面积、裸地面积、年采矿量及年末总人口等8个要素作为模型的人类活动输入因子,对流域年均产沙量进行了定量建模预测。结果表明:小波神经网络耦合模型不仅拟合精度高,而且预测效果好,为流域产沙的定量研究提供了新的途径。The application of wavelet neural network coupling model in quantitative study on yearly average sediment yield of a river basin is described. We select annual rainfall, air temperature and runoff, as well as heavy rain, rainstorm, amount of evaporation, sunshine hours and rainfall in flood season as main climate - hydrological factors of the model, and choose cultivated area, woodland area, reservoir storage capacity, road construction, water and soil conservation area, bare land area, annual mining amount and total population as main human activity factors. The result shows that the model provides a new approach for quantitative study on sediment yield of a river basin,which has a high fitting precision and an ideal prediction effect.

关 键 词：年均产沙量 小波神经网络耦合模型 盘龙河流域 预测 

分 类 号：TV121[水利工程—水文学及水资源]