摘要
采用时间序列自回归滑动平均（ARIMA）预测模型，对伊洛瓦底江支流大盈江拉贺练水文站1980～2005年平均含沙量资料进行建模预测。综合AIC值、相对误差，确定模型的阶数，运用Marquardt非线性最小二乘法估计模型参数，建立ARIMA预测模型。经检验，AIC=-114，相对误差全部低于20％，残差序列为白噪声序列，表明ARIMA（1，3，2）模型较为合理。应用模型对2006～2009年拉贺练水文站的年平均含沙量进行了预测，实现河流输沙状况的短期预报：
ARIMA(Auto Regressive Integrated Moving Model)model was used to forecast average sediment content from 1980 to 2005 at Lahelian hydrologic station on the Dayingjiang river,a branch of Irrawddy river.The order of the model was determined by AIC value,relative error and the parameter of the model were estimated by Marquardt least square method for the model establishment.After check,AIC=-114,all relative errors is less than 20%,the residual errors is white noise series,it is demonstrated that ARIMA(1,2,3) is reasonable.The established model is used to forecast the annual sediment content from 2006 to 2009 at Lahelian hydrologic station and short-term forecast for river sediment transportation is realized.
出处
《人民长江》
北大核心
2008年第1期43-45,共3页
Yangtze River
基金
国家重点基础研究发展973规划项目(2003CB415100)
