Yu et al. (2025) Impacts of the mega cascade reservoirs on riverine hydrothermal regimes based on deep learning

Identification

• Journal: Journal of Hydrology
• Year: 2025
• Date: 2025-11-23
• Authors: Meixiu Yu, Jiexin Guan, Jianyun Zhang, Junliang Jin, Qingmei Meng, Hanlin Song, Zixuan Xu, Xiaolong Liu, Jiayi He, Ting Fu
• DOI: 10.1016/j.jhydrol.2025.134633

Research Groups

• College of Hydrology and Water Resources, Hohai University, Nanjing, China
• The National Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing, China
• Yangtze Institute for Conservation and Development, Nanjing, China
• Research Center for Climate Change of Ministry of Water Resources, Nanjing Hydraulic Research Institute, China
• Production Management Department of the Three Gorges Cascade Regulation, China Changjiang Power Co., Ltd, Yichang, China
• School of Architecture and Art Design, Nanjing Polytechnic Institute, Nanjing, China

Short Summary

This study investigates the impacts of four mega cascade reservoirs on the Lower Jinsha River's downstream hydrological and water temperature regimes using an LSTM-based hydro-thermal model, revealing significant alterations in flow, temperature, and their coupling, with implications for ecological risks.

Objective

• To scrutinize the ramifications of the operational dynamics of four colossal cascade reservoirs situated on the Lower Jinsha River with respect to their downstream hydrological and water temperature regimes.
• To comprehensively assess the systematic disturbance patterns and intrinsic attributions of giant cascade reservoir operations on the river’s natural hydrological and thermal regimes.

Study Configuration

• Spatial Scale: Lower Jinsha River, focusing on the downstream impacts of four colossal cascade reservoirs.
• Temporal Scale: Multi-year period corresponding to reservoir operation.

Methodology and Data

• Models used: Integrated hydro-thermal model based on Long Short-Term Memory (LSTM) deep learning method.
• Data sources: Observation data (streamflow, water temperature) used for model training and validation, and for comparison with natural baselines.

Main Results

• Hydrological regimes exhibited intensified peak-shaving and flow-supplementing; observed water temperature showed an overall warming trend but a persistent cooling tendency relative to natural conditions, accompanied by an increasing seasonal thermal lag.
• Hydrological alteration escalated from moderate to high, increasing ecological risks, while thermal alteration remained persistently high.
• Reservoir operations restructured the hydro-thermal coupling, reversing the flow-temperature hysteresis from clockwise to counter-clockwise, indicating a shift to an anthropogenically dominated regime.
• Giant cascade reservoirs dominate downstream hydrothermal variability, especially under extreme drought conditions, underscoring their significant regulatory and mitigation power.

Contributions

• Development and application of an integrated hydro-thermal model based on deep learning (LSTM) to accurately simulate natural streamflow and water temperature processes under reservoir operation.
• Comprehensive assessment of systematic disturbance patterns and intrinsic attributions of giant cascade reservoir operations on riverine hydrothermal regimes.
• Quantification of hydrological and thermal alterations and their associated ecological risks.
• Identification of a reversal in the flow-temperature hysteresis loop, indicating a fundamental shift in hydro-thermal coupling due to anthropogenic influence.
• Highlighting the dominant role of mega cascade reservoirs in controlling downstream hydrothermal variability, particularly during extreme events.

Funding

• Not explicitly mentioned in the provided text.

Citation

@article{Yu2025Impacts,
  author = {Yu, Meixiu and Guan, Jiexin and Zhang, Jianyun and Jin, Junliang and Meng, Qingmei and Song, Hanlin and Xu, Zixuan and Liu, Xiaolong and He, Jiayi and Fu, Ting},
  title = {Impacts of the mega cascade reservoirs on riverine hydrothermal regimes based on deep learning},
  journal = {Journal of Hydrology},
  year = {2025},
  doi = {10.1016/j.jhydrol.2025.134633},
  url = {https://doi.org/10.1016/j.jhydrol.2025.134633}
}