Daily Runoff Forecasting Using Ensemble Empirical Mode Decomposition and Long Short-Term Memory

被引:35
作者
Yuan, Ruifang [1 ]
Cai, Siyu [2 ]
Liao, Weihong [2 ]
Lei, Xiaohui [2 ]
Zhang, Yunhui [2 ]
Yin, Zhaokai [3 ]
Ding, Gongbo [4 ]
Wang, Jia [5 ]
Xu, Yi [6 ]
机构
[1] China Univ Geosci, Sch Water Resources & Environm, Beijing, Peoples R China
[2] China Inst Water Resources & Hydropower Res, State Key Lab Simulat & Regulat Water Cycle River, Beijing, Peoples R China
[3] China Three Gorges Corp, Beijing, Peoples R China
[4] Sichuan Univ, Coll Water Resource & Hydropower, Chengdu, Peoples R China
[5] Tianjin Univ, Sch Environm Sci & Technol, Tianjin, Peoples R China
[6] Zhejiang Univ, Coll Civil Engn & Architecture, Hangzhou, Peoples R China
基金
国家重点研发计划;
关键词
ensemble empirical mode decomposition; long short-term memory; three gorges reservoir; runoff forecasting; streamflow series; hydrological model; ARTIFICIAL NEURAL-NETWORKS; STREAMFLOW; SWAT; PRECIPITATION; PERFORMANCE; BASIN;
D O I
10.3389/feart.2021.621780
中图分类号
P [天文学、地球科学];
学科分类号
07 ;
摘要
Hydrological series data are non-stationary and nonlinear. However, certain data-driven forecasting methods assume that streamflow series are stable, which contradicts reality and causes the simulated value to deviate from the observed one. Ensemble empirical mode decomposition (EEMD) was employed in this study to decompose runoff series into several stationary components and a trend. The long short-term memory (LSTM) model was used to build the prediction model for each sub-series. The model input set contained the historical flow series of the simulation station, its upstream hydrological station, and the historical meteorological element series. The final input of the LSTM model was selected by the MI method. To verify the effect of EEMD, this study used the Radial Basis Function (RBF) model to predict the sub-series, which was decomposed by EEMD. In addition, to study the simulation characteristics of the EEMD-LSTM model for different months of runoff, the GM(group by month)-EEMD-LSTM was set up for comparison. The key difference between the GM-EEMD-LSTM model and the EEMD-LSTM model is that the GM model must divide the runoff sequence on a monthly basis, followed by decomposition with EEMD and prediction with the LSTM model. The prediction results of the sub-series obtained by the LSTM and RBF exhibited better statistical performance than those of the original series, especially for the EEMD-LSTM. The overall GM-EEMD-LSTM model performance in low-water months was superior to that of the EEMD-LSTM model, but the simulation effect in the flood season was slightly lower than that of the EEMD-LSTM model. The simulation results of both models are significantly improved compared to those of the LSTM model.
引用
收藏
页数:12
相关论文
共 35 条
[1]   A continental-scale hydrology and water quality model for Europe: Calibration and uncertainty of a high-resolution large-scale SWAT model [J].
Abbaspour, K. C. ;
Rouholahnejad, E. ;
Vaghefi, S. ;
Srinivasan, R. ;
Yang, H. ;
Klove, B. .
JOURNAL OF HYDROLOGY, 2015, 524 :733-752
[2]   Comparison of the performance of SWAT, IHACRES and artificial neural networks models in rainfall-runoff simulation (case study: Kan watershed, Iran) [J].
Ahmadi, Mehdi ;
Moeini, Abolfazl ;
Ahmadi, Hassan ;
Motamedvaziri, Baharak ;
Zehtabiyan, Gholam Reza .
PHYSICS AND CHEMISTRY OF THE EARTH, 2019, 111 (65-77) :65-77
[3]   Simulation of karst spring discharge using a combination of time-frequency analysis methods and long short-term memory neural networks [J].
An, Lixing ;
Hao, Yonghong ;
Yeh, Tian-Chyi Jim ;
Liu, Yan ;
Liu, Wenqiang ;
Zhang, Baoju .
JOURNAL OF HYDROLOGY, 2020, 589
[4]   The Three Gorges Project [J].
不详 .
ENGINEERING, 2015, 1 (01) :11-13
[5]   Daily Reservoir Runoff Forecasting Method Using Artificial Neural Network Based on Quantum-behaved Particle Swarm Optimization [J].
Cheng, Chun-tian ;
Niu, Wen-jing ;
Feng, Zhong-kai ;
Shen, Jian-jian ;
Chau, Kwok-wing .
WATER, 2015, 7 (08) :4232-4246
[6]  
Ding G.B., 2019, China Rural Water and Hydropower, P66
[7]   The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis [J].
Huang, NE ;
Shen, Z ;
Long, SR ;
Wu, MLC ;
Shih, HH ;
Zheng, QN ;
Yen, NC ;
Tung, CC ;
Liu, HH .
PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES, 1998, 454 (1971) :903-995
[8]   A review on Hilbert-Huang transform: method and its applications to geophysical studies [J].
Huang, Norden E. ;
Wu, Zhaohua .
REVIEWS OF GEOPHYSICS, 2008, 46 (02)
[9]   Changes in monthly precipitation and flood hazard in the Yangtze River Basin, China [J].
Jiang, Tong ;
Kundzewicz, Zbigniew W. ;
Su, Buda .
INTERNATIONAL JOURNAL OF CLIMATOLOGY, 2008, 28 (11) :1471-1481
[10]   A new hybrid data-driven model for event-based rainfall-runoff simulation [J].
Kan, Guangyuan ;
Li, Jiren ;
Zhang, Xingnan ;
Ding, Liuqian ;
He, Xiaoyan ;
Liang, Ke ;
Jiang, Xiaoming ;
Ren, Minglei ;
Li, Hui ;
Wang, Fan ;
Zhang, Zhongbo ;
Hu, Youbing .
NEURAL COMPUTING & APPLICATIONS, 2017, 28 (09) :2519-2534