Short-term river flow forecasting provides a critical time window and decision-making support for disaster prevention, mitigation, and sustainable utilization of water resources. To address the limitations of the single long short-term memory (LSTM) model, such as difficulties in capturing local features, sensitivity to hyperparameters, and poor prediction performance, a novel river discharge prediction model integrating convolutional neural network (CNN), LSTM, and a self-attention mechanism was proposed. The Bayesian optimization (BO) algorithm was employed for hyperparameter tuning, resulting in the construction of a coupled BO-CNN-LSTM-Att model. Taking the Yangtze River Basin in Jingzhou, Hubei Province as a case study, the prediction accuracy of various models was compared and evaluated. The results demonstrate that the proposed integrated model reduces the root mean square error (RMSE) and mean absolute error (MAE) by 51.55% and 53.56%, respectively, compared with the traditional LSTM model, and achieves the highest coefficient of determination (R2) of 0.972 8. The prediction performance is significantly superior to that of incomplete integrated models and other algorithmic models.
LIU Heng, LI Zhi, FAN Yang.
Short-term Runoff Prediction in the Yangtze River Basin Based on Bayesian Optimized CNN-LSTM-Att Model[J].
Science Technology and Engineering, 2026, 26(16): 6978-6984 DOI:10.12404/j.issn.1671-1815.2503853
MaK, HeD, LiuS, et al. Novel time-lag informed deeplearning framework for enhanced streamflow prediction and flood early warning in large-scale catchments[J]. Journal of Hydrology, 2024, 631: 130841.
[2]
KhandK, SenayG B. Evaluation of streamflow predictions from LSTM models in water-and energy-limitedregions in the United States[J]. Machine Learning with Applications, 2024, 16:100551.
[3]
WeiY, HashimH B, LaiS H, et al. Comparative analysis of artificial intelligence methods for streamflow forecasting[J]. IEEE Access, 2024, 12: 10865-10885.
HanYunxi, ZhangXin, WangZifa, et al. Watershed scale flood forecasting method based on coupled meteorological hydrological hydrodynamic model: a case study of Beijiang River Basin[J]. Science Technology and Engineering, 2024, 24(11): 4397-4408.
[6]
AkshayK, BasudevB. A model combination approach for improving streamflow prediction[J]. Water Resources Management, 2022, 36(15): 5945-5959.
[7]
ParisoujP, MohebzadehH, LeeT. Employing machine learning algorithms for streamflow prediction: a case study of four river basins with different climatic zones in the United States[J]. Water Resources Management, 2020, 34(13): 1-19.
[8]
DaneshM, GharehbaghiA, MehdizadehS, et al. A comparative assessment of machine learning and deep learning models for the daily river streamflow forecasting[J]. Water Resources Management, 2024, 39(4): 1911-1930.
HuLeyi, JiangXiaolei, ZhouJiahui, et al. Application of LSTM considering time steps in runoff prediction of Ganjiang River Basin[J]. Journal of Lake Sciences, 2024, 36(4): 1241-1252.
[11]
YuQ, TolsonB A, ShenH, et al. Enhancing long short-term memory (LSTM)-based streamflow prediction with a spatially distributed approach[J]. Hydrology and Earth System Sciences, 2024, 28(9): 2107-2122.
[12]
LiuD, JiangW, MuL, et al. Streamflow prediction using deep learning neural network: case study of Yangtze River[J]. IEEE Access, 2020, 8: 90069-90086.
HeYan, LuYing, DengLixian, et al. Daily runoff prediction in the Nanguo River Basin based on the LSTM model[J]. Journal of China Hydrology, 2025, 45(3): 78-85.
[17]
LiangJ, DuY, XuY, et al. Using adaptive chaotic grey wolf optimization for the daily streamflow prediction[J]. Expert Systems with Applications, 2024, 237: 121113.
[18]
LivierisI E, PintelasE, PintelasP. A CNN-LSTM model for gold price time-series forecasting[J]. Neural Computing and Applications, 2020, 32: 17351-17360.
[19]
ZhaW, LiuY, WanY, et al. Forecasting monthly gas field production based on the CNN-LSTM model[J]. Energy, 2022, 260: 124889.
[20]
HuangR, WeiC, WangB, et al. Well performance prediction based on long short-term memory(LSTM) neural network[J]. Journal of Petroleum Science and Engineering, 2022, 208:109686.
XuDongmei, WangYiyang, WangWenchuan. Prediction of annual runoff using long short term memory neural network model based on Bayesian optimization algorithm[J]. Water Resources and Power, 2022, 40(12): 42-46.
LiuHeng, ChenYinan, GeYihua. et al. Design and experiment of a real-time flow velocity measurement system for river channels based on acoustic tomography[J]. Journal of Transduction Technology, 2024, 37(11): 2003-2009.