基于EMD分解的AR模型在年径流预测中的应用

为提高年径流中长期预测的精度,提出了一种新的时间序列预测方法——基于EMD分解的AR模型,以汾河上游上静游、汾河水库、寨上和兰村四座水文站1956~2000年的年径流序列为例,首先利用经验模态分解(EMD)方法将四座水文站的年径流序列分解为若干个固有模态函数(IMF)分量和一个残余项分量,然后运用自回归(AR)模型分别对各阶IMF进行预测,最后将各阶预测值重构得到年径流量预测值与单独运用AR模型的预测结果进行比较。结果表明,运用基于EMD分解的AR模型对汾河上游年径流进行预测,其预测精度比单独运用AR模型的预测精度有明显提高,表明该方法可行、有效。

Application of Auto Regressive Model to Annual Runoff Forecasting Based on Empirical Mode Decomposition

In order to improve the accuracy of mid-long annual runoff forecasting, a new method of time series prediction, which auto regressive (AR) model based empirical mode decomposition (EMD) is used to annual runoff forecasting, is proposed in this paper. Taking the annual runoff series covering 1956-2000 in the upper reaches of the Fenhe River as an example, firstly the annual runoff series of Shangjingyou, Fenhe Reservoir, haishang and Lancun hydrologic stations are decomposed by the EMD and generated several intrinsic mode functions and one residue component. And then all the intrinsic mode functions are predicted by using AR model respectively. Finally, the predicted values are reconstruct as the predicted values of annual runoff. Compared with the results obtained by AR model, it shows that the prediction accuracy of the EMD-AR model as improved significantly than the single AR model, and it is feasible and effective.