基于耦合预测模型的大通河月径流预测OA

The simulation and prediction of river runoff are of great significance for controlling basin water volume and ensuring optimal allo-cation of basin water resources.However,due to the influence of abnormal climate and human activities,the instability of medium-and long-term runoff sequences has increased the difficulty of runoff prediction.To improve prediction accuracy,a coupled deep learning model frame-work based on variational mode decomposition(VMD),mutual information(MI),and bidirectional long short-term memory(Bi-LSTM)networks,called the VMD-Bi-LSTM model,was established.First,VMD was used to decompose the original runoff data into intrinsic mode components;Then,Bi-LSTM was applied to each component to build prediction models,with the input lag time determined by the mutual in-formation method;Finally,the prediction results of each subsequence were superimposed to obtain the final prediction result.The paper ex-plored the performance of the proposed model in predicting the monthly runoff at Tiantang hydrological station in the Datong River Basin and compared it with other models.The results show that:Compared to other models,this model exhibits significant advantages in both point and interval predictions.The Nash-Sutcliffe efficiency coefficient(NSE)of the prediction results reaches 0.95,and the coverage rates of interval predictions are 0.92 and 0.85 at the 95%and 90%confidence intervals,respectively.