Novel Applications of Surface Water–Groundwater Modeling

1. Introduction

Water resources systems are inherently complex, governed by dynamic interactions between surface water and groundwater across multiple spatial and temporal scales [Contribution 1]. Traditionally, these two domains have often been studied separately due to methodological limitations and disciplinary boundaries. However, increasing pressures from climate change, population growth, and intensified water use have revealed the limitations of such compartmentalized approaches [Contribution 2]. As a result, integrated surface water–groundwater modeling has emerged as a critical tool for understanding hydrological processes and supporting sustainable water resources management.

Over the past two decades, significant advances have been made in the development of coupled modeling frameworks, including physically based models such as SWAT-MODFLOW [Contributions 1 and 2], HydroGeoSphere [Contribution 3], and integrated finite-element systems. These models have enabled researchers to simulate complex feedback mechanisms between rivers, aquifers, and land surface processes [Contributions 2 and 4]. At the same time, the rapid growth in data availability and computational power has facilitated the integration of data-driven approaches, such as machine learning and deep learning, into hydrological modeling [Contributions 1–4].

Despite these advancements, several challenges remain. Integrated models often involve high levels of uncertainty due to parameterization complexity, limited observational data, and scale mismatches [Contribution 4]. Moreover, the diversity of modeling approaches—ranging from fully coupled numerical models to hybrid data-driven frameworks—can make it difficult to compare results and derive generalizable insights [Contribution 3]. There is therefore a pressing need to explore novel applications of integrated models that not only advance methodological development but also address real-world water management issues.

This Special Issue, “Novel Applications of Surface Water–Groundwater Modeling,” aims to contribute to this effort by presenting a collection of studies that showcase innovative modeling approaches and applications across diverse hydrological environments [Contributions 1–8]. The contributions span a wide range of topics, including groundwater level forecasting, uncertainty analysis, water quality modeling, recharge estimation, analytical solutions for groundwater–surface water interactions, streamflow depletion assessment, artificial recharge planning, and the development of new integrated modeling tools.

2. Emerging Trends in Integrated Modeling

A key theme emerging from this Special Issue is the increasing convergence between physically based models and data-driven techniques. While traditional models rely on governing equations to represent hydrological processes, data-driven approaches leverage large datasets to identify patterns and make predictions without explicitly defining physical relationships. The integration of these approaches offers new opportunities to overcome the limitations associated with each method [Contribution 1].

For instance, deep learning models can capture nonlinear relationships and provide efficient forecasting tools, while physically based models ensure interpretability and adherence to conservation laws. Hybrid frameworks that combine these strengths are becoming increasingly important, particularly in contexts where data availability and computational constraints vary.

Another important trend is the growing emphasis on uncertainty quantification. Integrated models often involve numerous parameters and inputs, each associated with uncertainties that can propagate through simulations. Addressing these uncertainties is essential for improving model reliability and supporting decision-making under changing environmental conditions.

In addition, there is a noticeable expansion of modeling applications beyond traditional hydrological analysis. Integrated models are now being used to investigate water quality, ecosystem dynamics, and human–water interactions, reflecting the interdisciplinary nature of modern water resources research.

Scale Integration and Process Coupling Challenges

One of the most persistent challenges in integrated surface water–groundwater modeling lies in the issue of scale mismatch. Surface hydrological processes are often simulated at relatively coarse spatial resolutions (e.g., sub-basin scale), while groundwater processes require finer spatial discretization to capture aquifer heterogeneity. Temporal scales also differ significantly, with surface runoff responding rapidly to precipitation events, whereas groundwater systems exhibit delayed responses over longer timeframes.

Bridging these scale gaps requires advanced coupling strategies that can reconcile differences in spatial resolution, temporal discretization, and process representation. Recent developments in multi-scale modeling frameworks and adaptive coupling techniques provide promising solutions; however, further research is needed to ensure numerical stability, computational efficiency, and physical consistency.

In addition, process coupling itself introduces complexity. The bidirectional feedback between surface water and groundwater involves nonlinear interactions influenced by factors such as soil properties, hydraulic gradients, and anthropogenic interventions. Accurately representing these interactions remains a central challenge, particularly under transient and extreme conditions such as droughts and floods.

3. Overview of Contributions

The eight papers included in this Special Issue collectively illustrate the breadth and depth of current research in integrated surface water–groundwater modeling.

3.1. Data-Driven Groundwater Forecasting

Kim et al. [Contribution 5] present a deep learning-based approach for groundwater level forecasting in the Gyorae area of Jeju Island, a region characterized by highly permeable volcanic geology and a strong dependence on groundwater resources. By employing long short-term memory (LSTM) models and incorporating hydro-meteorological variables, the study demonstrates the feasibility of using data-driven techniques for operational groundwater management.

A particularly notable contribution is the introduction of a derivative-based learning approach, which focuses on predicting changes in groundwater levels rather than absolute values. This method improves model stability and forecasting performance. In addition, the ensemble forecasting strategy enhances robustness by reducing the impact of short-term variability. This study highlights the potential of artificial intelligence in complementing traditional modeling approaches, especially in data-rich environments.

3.2. Uncertainty Reduction in Integrated Models

Lee et al. [Contribution 3] address the critical issue of uncertainty in integrated watershed modeling by investigating the impact of observation errors on model calibration. Using the HydroGeoSphere model applied to the Sapgyo watershed, the authors compare different weighting schemes that account for observation error variances.

Their results show that incorporating observation error-based weighting significantly improves parameter estimation and model performance. Furthermore, the study demonstrates that groundwater plays an important role in buffering hydrological responses to climate variability, particularly under future climate scenarios. This work underscores the importance of rigorous uncertainty analysis in enhancing the credibility of integrated models.

3.3. Integration of Machine Learning and Water Quality Modeling

Rodríguez-López et al. [Contribution 6] extend the application of integrated modeling to water quality assessment by employing machine learning algorithms to estimate chlorophyll-a concentrations in Lake Llanquihue, Chile. Using a long-term dataset spanning more than three decades, the study evaluates multiple algorithms, including XGBoost, LightGBM, and AdaBoost.

The results indicate high predictive accuracy, demonstrating the potential of machine learning techniques for water quality monitoring and management. This study illustrates how integrated modeling frameworks can be expanded to include ecological and biogeochemical processes, thereby providing a more comprehensive understanding of aquatic systems.

3.4. Groundwater Recharge Estimation

Ware et al. [Contribution 1] focus on one of the most challenging components of the hydrological cycle: groundwater recharge. By coupling the SWAT-MODFLOW model with the transient water table fluctuation method, the authors provide a robust framework for estimating spatiotemporal recharge distribution in the Anyang watershed.

The study reveals significant seasonal variability in recharge, with higher values during wet periods. It also demonstrates that integrated modeling can improve the accuracy of recharge estimation compared with standalone methods. This work highlights the importance of combining multiple approaches to address uncertainties in water balance analysis.

3.5. Analytical Advances in Groundwater–Surface Water Interaction

Wu et al. [Contribution 2] contribute to the theoretical understanding of groundwater–surface water interactions by developing an analytical solution for unsteady seepage under conditions of linearly varying river water levels. By simplifying the Laplace transform process and introducing a linearization approach, the authors derive solutions that closely match numerical results.

This analytical framework provides a valuable tool for understanding fundamental processes governing water exchange between rivers and aquifers. Such theoretical advancements are essential for improving the conceptual foundations of integrated models.

3.6. Streamflow Depletion Under Groundwater Withdrawal

Lee et al. [Contribution 7] investigate the impact of groundwater withdrawal on streamflow using a modified SWAT model that explicitly accounts for pumping. Applied to the Bokhacheon watershed, the model quantifies reductions in low and drought flows, highlighting the significant influence of groundwater extraction on hydrological systems.

The study also reveals changes in other water balance components, such as increased soil moisture and evapotranspiration due to irrigation return flows. These findings emphasize the importance of considering human activities in integrated modeling and water management.

3.7. Artificial Groundwater Recharge Planning

Gururani et al. [Contribution 8] utilize remote sensing and GIS techniques to identify potential zones for artificial groundwater recharge in a watershed in India. By integrating multiple thematic layers, including slope, land use, and stream order, the study provides a spatial prioritization framework for recharge planning.

The proposed approach offers practical insights for water resource managers and highlights the role of integrated modeling in supporting sustainable groundwater management strategies.

3.8. Development of New Integrated Modeling Tools

Yang et al. [Contribution 4] present a newly developed open-source integrated surface–subsurface flow model designed for plain farmland environments. By coupling a two-dimensional surface flow model with a one-dimensional subsurface module, the model captures the effects of micro-topography on runoff and infiltration processes.

The study demonstrates high predictive accuracy and provides a flexible tool for agricultural water management. The development of open-source models is particularly important for promoting accessibility and collaboration within the research community.

3.9. Bridging Science and Policy Through Integrated Modeling

Beyond methodological advancements, an important contribution of the studies in this Special Issue is their relevance to real-world water management and policymaking. Integrated models provide a quantitative basis for evaluating management scenarios, such as groundwater abstraction limits, artificial recharge strategies, and climate adaptation measures.

In particular, the ability to quantify streamflow depletion due to groundwater pumping and to identify recharge zones has direct implications for regulatory frameworks and water allocation policies. These tools enable decision-makers to move beyond qualitative assessments and adopt evidence-based approaches grounded in scientific modeling.

Furthermore, the integration of data-driven approaches enhances the usability of models in operational contexts. For example, real-time groundwater forecasting systems can support early warning and adaptive management. As such, integrated modeling is increasingly positioned not only as a research tool but also as a practical component of water governance systems.

4. Implications for Future Research

The contributions in this Special Issue highlight several important directions for future research in integrated surface water–groundwater modeling.

First, the integration of physical and data-driven approaches should be further explored. Hybrid models that combine process-based understanding with machine learning capabilities have the potential to improve predictive accuracy while maintaining interpretability.

Second, uncertainty quantification must remain a central focus. As models become more complex, robust methods for handling uncertainty are essential for ensuring reliability and supporting decision-making.

Third, the representation of human–water interactions should be enhanced. Water use, land management, and policy interventions play critical roles in shaping hydrological systems, and their inclusion in models is necessary for realistic simulations.

Finally, the development of open-source modeling tools and standardized frameworks will facilitate knowledge sharing and collaboration. Such efforts are crucial for advancing the field and addressing global water challenges.

4.1. Toward Digital Twin and AI-Driven Water Systems

A promising future direction for integrated modeling is the development of digital twin systems for water resources. Digital twins combine real-time monitoring data, predictive modeling, and advanced analytics to create dynamic representations of physical systems. When applied to surface water–groundwater systems, digital twins can enable continuous simulation, forecasting, and optimization.

The integration of artificial intelligence further enhances these capabilities. AI-driven models can assimilate large volumes of data, identify hidden patterns, and improve predictive performance. When coupled with physically based models, these systems can provide both accuracy and interpretability, addressing a key limitation of purely data-driven approaches.

Such advancements are particularly relevant in the context of climate change, where uncertainty and variability are increasing. Digital twin frameworks can support adaptive management by continuously updating model states and predictions based on new data.

4.2. Incorporating Human–Water Systems

Another critical research direction is the explicit incorporation of human activities into integrated models. Traditional hydrological models often treat human influences as external inputs, such as boundary conditions or forcing variables. However, water use, land use change, and policy decisions are dynamic processes that interact with natural systems.

Future models should aim to represent these interactions more explicitly, incorporating socio-economic factors, institutional constraints, and behavioral dynamics. This will enable a more comprehensive understanding of water systems and support the development of sustainable management strategies.

4.3. Toward Global Applicability and Transferability

While many integrated modeling studies are conducted at local or regional scales, there is a growing need to develop models that are transferable across different climatic and geological settings. Achieving this requires standardized modeling frameworks, open-source tools, and shared datasets.

The contributions in this Special Issue demonstrate progress in this direction, particularly through the development of open-source models and the application of machine learning techniques. Continued efforts in model standardization and data sharing will be essential for advancing global water research.

5. Conclusions

This Special Issue demonstrates that integrated surface water–groundwater modeling is rapidly evolving into a multidisciplinary field that combines hydrology, data science, and environmental management. The studies presented here illustrate the diverse applications of integrated models, from theoretical developments to practical decision-support tools.

As water systems become increasingly stressed by environmental and societal pressures, the need for comprehensive and reliable modeling approaches will continue to grow. By advancing both methodological innovation and practical application, the research in this Special Issue contributes to the ongoing effort to achieve sustainable water resources management.