Baohong Lu

Groundwater pumping near large rivers is commonly used to prevent agricultural water shortages, yet the impacts of high extraction rates on groundwater systems require clarification. In Dalat County along the Yellow River, intensive extraction has reshaped groundwater flow patterns, forming a cone of depression that reversed groundwater discharge to recharge from the river. The values of groundwater δ18O, EC and Cl within the cone are lower than in the Yellow River irrigation zones. Paradoxically, groundwater in the cone exhibits an older groundwater age, confirming the shift of groundwater flow patterns. This stems from shallow, salinized aquifers in river-irrigated areas, where increased leaching after river recharge elevated EC/Cl despite younger groundwater ages. The cone's depleted δ18O signals suggest mountain-sourced recharge from southern highlands. These findings elucidate recharge dynamics critical for sustainable water management in changing environments.

Real-time control (RTC) can be applied to optimize the operation of the anaerobic–anoxic–oxic (A2O) process in wastewater treatment for energy saving. In recent years, many studies have utilized deep reinforcement learning (DRL) to construct a novel AI-based RTC system for optimizing the A2O process. However, existing DRL methods require the use of A2O process mechanistic models for training. Therefore they require specified data for the construction of mechanistic models, which is often difficult to achieve in many wastewater treatment plants (WWTPs) where data collection facilities are inadequate. Also, the DRL training is time-consuming because it needs multiple simulations of mechanistic model. To address these issues, this study designs a novel data-driven RTC method. The method first creates a simulation model for the A2O process using LSTM and an attention module (LSTM-ATT). This model can be established based on flexible data from the A2O process. The LSTM-ATT model is a simplified version of a large language model (LLM), which has much more powerful ability in analyzing time-sequence data than usual deep learning models, but with a small model architecture that avoids overfitting the A2O dynamic data. Based on this, a new DRL training framework is constructed, leveraging the rapid computational capabilities of LSTM-ATT to accelerate DRL training. The proposed method is applied to a WWTP in Western China. An LSTM-ATT simulation model is built and used to train a DRL RTC model for a reduction in aeration and qualified effluent. For the LSTM-ATT simulation, its mean squared error remains between 0.0039 and 0.0243, while its R-squared values are larger than 0.996. The control strategy provided by DQN effectively reduces the average DO setpoint values from 3.956 mg/L to 3.884 mg/L, with acceptable effluent. This study provides a pure data-driven RTC method for the A2O process in WWTPs based on DRL, which is effective in energy saving and consumption reduction. It also demonstrates that purely data-driven DRL can construct effective RTC methods for the A2O process, providing a decision-support method for management.

The Huaihe River Basin is an important ecological function conservation area in China, and it is also an important production area for national food, energy, minerals, and manufacturing. The groundwater storage and groundwater drought in this region are of great significance for ecological maintenance and water resources management. In this study, based on GRACE data and GLDAS data, a dynamic calculation method for groundwater storage in the Huaihe River Basin was developed, and a groundwater drought index (GRACE-GDI) was derived. By coupling GRACE-GDI with run theory, the quantitative identification of groundwater drought events, as well as their duration, intensity, and other characteristics within the basin, was achieved. The spatiotemporal changes in groundwater storage and groundwater drought in the Huaihe River Basin were analyzed using the developed method. The results showed that GRACE data are highly applicable in the Huaihe River Basin and is capable of capturing the spatiotemporal variations in groundwater storage in this region. Over the study period, mainly affected by rainfall, the terrestrial water storage and surface water storage in the Huaihe River Basin showed a decreasing trend, while groundwater storage showed a slight increasing trend. The duration of groundwater drought events in the basin ranged from 78 to 152 months, with an intensity of 82.77 to 104.4. The duration of drought gradually increased from north to south, while the intensity increased from south to north.

Hydrological modeling is a crucial tool in hydrology and water resource management for analyzing runoff evolution patterns. In this study, the process-driven soil and water assessment tool (SWAT) model and data-driven machine learning techniques (XGBoost, random forest, LSTM, BILSTM, and GRU) were employed to simulate runoff at monthly and daily intervals in the Fenhe River basin, situated in the middle reaches of the Yellow River, respectively. The SWAT model demonstrated effective performance in simulating runoff at various scales, with the coefficient of determination (R²) exceeding 0.80 and the Nash–Sutcliffe efficiency (NSE) surpassing 0.79. Sensitivity analysis reveals varying degrees of sensitivity among the model parameters. Furthermore, the deep learning techniques (LSTM, BILSTM, and GRU) exhibited superior simulation generalization capabilities compared to the SWAT model across various scales. Additionally, the generalization abilities of traditional machine learning techniques (XGBoost and random forest) were comparable to the SWAT model. This indicates that deep learning techniques demonstrate remarkable stability and generalization capabilities across various scales. This analysis was motivated by the use of external continuous time series data as input and the application of deep learning techniques to internal mechanisms. Moreover, an integrated modeling approach was used to enhance simulation accuracy by combining the SWAT model with machine learning techniques. The results indicate that the integrated modeling approach improves simulation performance across various scales compared to the single-model approach. This research is significant for improving the efficiency of water resource utilization and management in the Fenhe River basin.

The construction of dams destroys the integrity of a watershed system and the continuity of natural water flow, creating a watershed with segmented and fragmented rivers. This, in turn, affects and even destroys the health and stability of the watershed ecosystem. This study selected the downstream area of Bengbu Sluice in the Huai River Basin of China as the study area. To address the increasingly prominent ecosystem degradation in the Huai River Basin, ecological flow thresholds were determined using habitat simulation and hydrological approaches for mutual validation. A multi-objective synergistic decision model incorporating ecological and socioeconomic objectives was developed to coordinate the economic and ecological water use conflicts in the study area. The optimal coordinated solution for the ecological flow of important biological habitats in the basin was determined with the multi-objective synergistic method. The results demonstrated that a coordinated solution could guarantee the ecological and economic water demands of the basin. The findings of this study can be used as a reference for scientific guidelines on future ecological operations in dam-controlled rivers.

To study the vertical water temperature structure and thermodynamic characteristics of tropical lake-like reservoirs, a water temperature model was developed by a vertical one-dimensional numerical model for Songtao Reservoir in Hainan Province, China. The model was verified by the measured water temperature data, and sensitivity analysis of key model parameters was carried out. The results show that water temperature simulated by the model in Songtao Reservoir agreed with the observations quite well, and the model is feasible for water temperature simulations in large reservoirs in tropical zones. The sensitivity of vertical water temperature structure to different model parameters varied. For example, the extinction coefficient greatly affected surface water temperature, which is important for the formation and development of the surface water temperature hybrid layer. The vertical mixing coefficient significantly influenced the inflection point position and thickness of the thermocline. The vertical water temperature structure in Songtao Reservoir was stratified. Reservoir surface water temperature varied from 19.4 °C to 33.8 °C throughout a year. The hypolimnion mainly appeared in elevation below 150 m, where the water temperature is basically maintained at 19 °C throughout the year. This study also found that the surface water temperature of Songtao Reservoir in the tropical zone was higher than the air temperature throughout a year, with an annual average of 3.5 °C higher than that of air temperature. The preliminary analysis found out that the higher surface water temperature may be caused by the strong air temperature and solar radiation in tropical zones, in addition to the enhanced capacity of heat absorption and heat storage due to the slow water flow in the reservoir.

In view of the low efficiency of the particle swarm algorithm under multiple constraints of reservoir optimal operation, this paper introduces a particle swarm algorithm based on strongly constrained space. In the process of particle optimization, the algorithm eliminates the infeasible region that violates the water balance in order to reduce the influence of the unfeasible region on the particle evolution. In order to verify the effectiveness of the algorithm, it is applied to the calculation of reservoir optimal operation. Finally, this method is compared with the calculation results of the dynamic programming (DP) and particle swarm optimization (PSO) algorithm. The results show that: (1) the average computational time of strongly constrained particle swarm optimization (SCPSO) can be thought of as the same as the PSO algorithm and lesser than the DP algorithm under similar optimal value; and (2) the SCPSO algorithm has good performance in terms of finding near-optimal solutions, computational efficiency, and stability of optimization results. SCPSO not only improves the efficiency of particle evolution, but also avoids excessive improvement and affects the computational efficiency of the algorithm, which provides a convenient way for particle swarm optimization in reservoir optimal operation.