TL;DR
This paper presents a new analytical model to accurately predict how groundwater levels in confined aquifers respond to fluctuating river levels, using the Yangtze River in Wuhan as a case study. By combining field monitoring data with a refined theoretical approach, the authors demonstrate a significant improvement in prediction accuracy compared to existing methods, offering valuable insights for riverside engineering projects.
What Problem Does This Paper Address?
Understanding the dynamic interaction between surface water bodies (like rivers) and groundwater is crucial for effective water resource management and infrastructure development, especially in flood-prone areas. Rivers and aquifers are not isolated systems; they influence each other. In regions like Wuhan, situated on the Yangtze River's first terrace, the river experiences significant seasonal fluctuations, particularly heavy flooding during summers. These fluctuations can directly impact the groundwater levels in adjacent confined aquifers. While existing analytical and numerical models exist for stream-aquifer interactions, many struggle to accurately capture local-scale processes and the complex dynamics of confined aquifers influenced by variable river stages. This paper aims to bridge this gap by developing a more precise model for predicting confined aquifer response to stream-stage fluctuations in such a context.
Methods
The study employs a two-pronged approach:
- Field Hydrological Monitoring: Extensive monitoring of both the Yangtze River's water level and the confined groundwater head was conducted over three hydrological years (2017-2019) using 60 dedicated monitoring wells. This provided a rich dataset for understanding the real-world behavior of the system.
- Analytical Modeling:
- Stream-Stage Fluctuation Model: Based on the observed data, a sinusoidal equation of the form $g(t) = a + b \sin(\omega t + c)$ was developed to represent the temporal variation of the Yangtze River's water level. The parameters $a$, $b$, $\omega$, and $c$ represent the average water level, amplitude of fluctuation, period, and phase shift, respectively, and were determined through fitting the observed data.
- New Confined Aquifer Response Model: Building upon the classical 1-D groundwater flow equation, the authors derived a novel analytical solution. This new model incorporates the proposed sinusoidal stream-stage fluctuation equation and addresses limitations of previous models, particularly their tendency to underestimate groundwater heads at the onset of fluctuations and their reduced accuracy with increasing distance from the river. The derivation involves Fourier sine transforms and leads to an integral solution that accounts for the stable water level and the dynamic variation.
Datasets
The study utilizes two primary datasets:
- Yangtze River Water Level Data: This data was collected from the Changjiang Maritime Safety Administration and is described as publicly available.
- Confined Groundwater Head Data: This data was collected from the Houhu groundwater monitoring system, established by Wuhan Geotechnical Engineering and Surveying Co., Ltd. This dataset is noted as being protected by company procedures and not publicly available.
The data spans three hydrological years (2017-2019) and was used for both fitting the stream-stage model and validating the proposed confined aquifer response model.
Key Insights and Ideas
- Sinusoidal River Fluctuations: The Yangtze River's water level in the Wuhan section can be effectively modeled as a sinusoidal function, with parameters that have clear physical interpretations (average level, amplitude, period, phase). This provides a predictable input for groundwater models.
- Limitations of Classical Models: The standard 1-D analytical solution for stream-aquifer interaction (Equation 10 in the paper) struggles to accurately predict absolute groundwater heads, especially at the beginning of a fluctuation period. It also shows a significant decrease in accuracy with distance from the river, likely due to not fully accounting for the damping effect of pressure conduction over distance.
- Importance of Stable Water Level: The proposed model explicitly separates the confined groundwater head into a stable component ($h_0$) and a dynamic variation ($\Delta h$). This acknowledges that aquifers have a baseline water level influenced by factors beyond immediate river fluctuations, and that the dynamic response is superimposed on this baseline.
- Improved Accuracy: The new analytical model demonstrates a markedly better fit to observed data compared to the classical model, particularly at greater distances from the river. This suggests it better captures the complex interplay of factors influencing groundwater response.
Key Results
- The proposed sinusoidal equation for Yangtze River water level achieved a relative error of 8.4% and an R² value of 0.889 when fitted to observed data.
- The new analytical model for confined aquifer response showed a significant reduction in prediction error. For distances of 600 m, 800 m, and 1700 m from the river, the relative errors were 8.88%, 8.71%, and 6.14%, respectively.
- In contrast, the classical model (Equation 10) yielded much higher errors: 19.15%, 17.67%, and 41.33% for the same distances.
- Overall, the proposed model reduced the prediction error from approximately 26% (for the previous model) to about 8%.
Limitations and Future Work
- Data Availability: While the Yangtze River data is public, the confined groundwater data is proprietary, limiting broader validation without access.
- Model Complexity: The proposed model, while analytical, involves an integral that might require numerical integration for practical application in some scenarios.
- Generalizability: The model was developed and validated for a specific case study (Yangtze River in Wuhan). Further testing across different geological settings and river types would be beneficial.
- Future Work: The authors suggest that their improved analytical model can be integrated into larger groundwater flow models like MODFLOW and FEFLOW. This would allow for more sophisticated simulations of stream-aquifer interactions in practical engineering applications, such as dewatering designs for foundation pits in riverside areas. Further research could explore the application of this model to other river systems and investigate the impact of other factors like rainfall and pumping on the stream-aquifer dynamics.
Summary updated