论文部分内容阅读
针对中长期水文预报因果规律不清楚,预报准确率低的问题,该文引入奇异谱分析方法(Singular SpectrumAnalysis,简称SSA),结合ARIMA模型,建立了基于SSA的分解预测校正模型。该模型通过SSA方法从年径流时间序列中提取对应着某些大气低频振荡的显著主分量序列,然后运用ARIMA模型对各显著分量序列分别进行预测,并对各序列预测结果的和进行校正。最后以大连市碧流河水库的年径流预报为例,对建立的SSA分解预测校正模型进行了应用检验。
Aiming at the problem that the cause and effect of hydrological forecast in medium and long term are not clear and the accuracy of forecasting is low, this paper introduces Singular Spectrum Analysis (SSA) and combined with ARIMA model to establish the model of decomposition forecast and correction based on SSA. In this model, significant principal component sequences corresponding to some atmospheric low-frequency oscillations are extracted from annual runoff time series by SSA method. Then ARIMA models are used to predict the significant component sequences respectively, and the sum of prediction results of each sequence is corrected. Finally, taking the annual runoff forecast of Biliuhe Reservoir in Dalian as an example, the established SSA decomposition prediction and correction model is tested.