The reservoir is an important hydraulic engineering measure for human utilization and management of water resources. Additionally, a reasonable and effective reservoir operating plan is essential for realizing reservoir function. To explore the application of a deep learning algorithm on the field of reservoir operations, a recurrent neural network (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU) are employed to predict outflows for the Xiluodu (XLD) reservoir. Meanwhile, this paper summarized the law of the effect of parameter setting on model performance compared to the simulation performance of three models, and analyzed the main factors that affect reservoir operation to provide the reference for future model of application research. Results show (1) the number of iterations and hidden nodes mainly influence the model precision, and the former has more effect than the latter, and the batch size mainly affects the calculated speed; (2) all three models can predict the reservoir outflow accurately and efficiently; (3) the operating decision generated by three models can implement the flood control and power generation goal of the reservoir and meet the operating regulation; and (4) under different hydrological periods, the influence factors of reservoir operation and their importance are different.

Four major species of Chinese carp, namely black carp (Mylopharyngodon piceus), grass carp (Ctenopharyngodon idellus), silver carp (Hypophthalmichthys molitrix) and bighead carp (Hypophthalmichthys nobilis), are important economic freshwater fish varieties in China. They primarily inhabit and breed in the Yangtze River. Unfortunately, the construction and operation of the Gezhouba Dam and the Three Gorges Dam have dramatically changed the hydrodynamic conditions in the middle reaches of the Yangtze River, leading to a sharp decline in the reproduction rates of these carp. The egg abundance of the four species of carp downstream from the Three Gorges Dam reached 8.35 billion in 1965, but abundance during 2005–2012 was only 0.25 billion. One of the main reasons was that the hydrodynamic conditions of the spawning ground could not meet the four species’ breeding requirements. However, due to the limitations of traditional detection tools, the spawning characteristics of these four species of carp were still unclear. In this study, the ultrasonic telemetry and a three–dimensional hydrodynamic model were utilized to build the habitat suitability index (HSI) curves for the four species of carp. The habitat suitability model was then built based on HSI curves to assess spawning habitat quantity under different flow conditions. Finally, the habitat suitability model in the Yidu spawning ground was validated using 32 groups of sampling data in 2015 and 2017. The statistical results showed that the most suitable velocity ranged from 0.78 m/s to 0.93 m/s. The most suitable water depth ranged from 14.56 m to 16.35 m, and the most suitable Froude number ranged from 0.049 to 0.129. The habitat suitability model simulation results indicated that when the discharge was between 15,000 m³/s and 21,300 m³/s, the weighted usable area (WUA) values in both the Yidu and Zhicheng spawning grounds would remain at a high level. The validation results showed that most spawning activities occurred at a high level of WUA, and that the daily spawning egg numbers increased with the WUA value. Therefore, discharges of between 15,000 m³/s and 21,300 m³/s could be recommended as ecological operation target flows. We propose a feasible ecological operation scheme by setting the initial flow at 15,000 m³/s and maintaining the daily discharge increase at 1500 m³/s for 4 days.

Hydrodynamic characteristics of spawning grounds are important factors affecting the spawning of four major Chinese carps (Mylopharyngodon piceus, Ctenopharyngodon idella, Hypophthalmichthys molitrix, and Aristichthys nobilis). To investigate the relationship between the preferred hydrodynamic characteristics of spawning sites and the response of fish spawning behavior, we monitored the flow field of spawning sites during ecological operation of the Three Gorges Dam (i.e., man-made flood regulation) in 2014 and 2015. We used the data to explore the correlation between vorticity changes in spawning grounds and the spawning amount. Pearson correlation coefficients of the average vorticity in all cross-sections of the Yidu spawning ground and spawning amount in 2014 and 2015 were 0.730 and 0.822, respectively, indicating a significant positive correlation between vorticity and spawning activity. In some specific regions, this correlation was even stronger (Pearson correlations of the regional vorticity and egg production were >0.95). To further corroborate and analyze the relationship between these specific regions and the distribution of broodstock during the breeding season, an ultrasonic telemetry test of broodstock was conducted in the Yidu spawning ground in 2016. The results showed that the broodstocks were concentrated in the reach near the Quantong Pier (~76 km from the Three Gorges Dam). These regions were consistent with areas of river characterized by highly correlated vorticity and egg production levels, suggesting that these regions are areas preferred by four major Chinese carps for spawning.

As the largest hydropower project in the world, the Three Gorges Reservoir (TGR) exerts significant backwater effects, leading to various water quality issues. The Xiangxi River, the largest tributary near the dam, has experienced algal bloom phenomena for several years. However, the characteristics and influencing factors of water quality during different operational periods of the reservoir remain unclear. This study analyzes 12 physical and chemical parameters of the Xiangxi Bay (XXB) from 2019 to 2023, employing a data-driven approach to explore the temporal and spatial characteristics of water quality under varying operational conditions of the TGR. It discusses the influencing factors and proposes countermeasures to address water quality challenges. The results indicate that: (1) Water quality parameters in XXB exhibit significant spatial and temporal variability, peaking in summer, with higher values observed in the middle and upper reaches. The Water Quality Index (WQI) indicates overall good water quality, while the Trophic Level Index (TLI) suggests medium eutrophication, particularly evident in spring and summer. (2) The backwater effects of the TGR have altered hydrodynamic conditions and mixing processes in XXB, significantly impacting water quality and promoting eutrophication. (3) Water quality is influenced by nutrient, organic matter, algae, and temperature, especially during low water levels in summer, with the middle and upper reaches being most affected. Key factors driving algal growth during this period include Total Nitrogen (TN), Transparency (SD), and Water Temperature (WT). (4) Effective control of eutrophication and algal blooms in XXB should focus on long-term upstream source management, close monitoring and prediction during spring and summer, and ecological operation of reservoir. This study enhances our understanding of the temporal and spatial characteristics of water quality in XXB under the operational dynamics of the TGR, providing guidance for eutrophication management and bloom prevention.

Water quality plays a pivotal role in human health and environmental sustainability. However, traditional water quality prediction models are limited by high model complexity and long computation time, whereas AI models often struggle with high-dimensional time series and lack physical interpretability. This paper proposes a two-dimensional water quality surrogate model that couples physical numerical models and AI. The model employs physical simulation results as input, applies spectral proper orthogonal decomposition to reduce the dimensionality of the simulation results, utilizes a long short-term memory neural network for matrix forecasting, and reconstructs the two-dimensional concentration field. The simulation and predictive performance of the surrogate model were systematically evaluated through four design scenarios and three sampling dataset lengths, with a particular focus on the convection–diffusion zone and high-concentration zone. The results indicated that the model achieves high prediction accuracy for up to 7 h into the future, with sampling dataset lengths ranging from 20 to 80 h. Specifically, the model achieved an average R² of 0.92, a MAE of 0.38, and a MAPE of 1.77%, demonstrating its suitability for short-term water quality predictions. The methodology and findings of this study demonstrate the significant potential of integrating spectral proper orthogonal decomposition and deep learning for water quality prediction. By overcoming the limitations of traditional models, the proposed surrogate model provides high-accuracy predictions with enhanced physical interpretability, even in complex, dynamic environments. This work offers a practical tool for rapid responses to water pollution incidents and supports improved watershed water quality management by effectively capturing pollutant diffusion dynamics. Furthermore, the model’s scalability and adaptability make it a valuable resource for addressing intelligent management in environmental science.

To meet the demand for water temperature prediction in the ecological operation of reservoirs, this study presents a long-term water temperature forecasting model based on machine learning algorithms using the Yidu spawning grounds, located downstream of the Three Gorges Reservoir (TGR), as a case study. The results demonstrated that the transfer learning model outperformed conventional models in terms of prediction accuracy, forecast duration, and computational efficiency. Specifically, the transfer learning model achieved an average error of 0.16–0.35 °C, outperforming the conventional model (with an error range of 0.15–0.6 °C), and exhibited superior capability in capturing complex water temperature variations. Regarding computational efficiency, the transfer learning model required significantly less training time and adapted rapidly to new data, enhancing the practical applicability of the model. During the critical ecological operation period (May–June), the transfer learning model’s average absolute error was 0.2–0.3 °C, effectively supporting optimal selection of the reservoir’s ecological operation timing. The findings provide a scientific basis for decision making in the integrated operation of reservoirs.