Purpose of Review
Climate change and hydrogeological perturbations significantly impact the availability and utility of global water resources, necessitating the need for advanced modelling approaches for effective management. Numerical models (NM) and machine learning (ML) approaches have already been explored in surface water-groundwater linkages, but limited attention has been given to their integrated application. The present study aims to bridge that gap by reviewing existing models for understanding linkages and evaluating the strengths and limitations of both approaches. This study also highlighted the potential of their combined use to enhance model accuracy, efficiency, and adaptability in complex hydrological settings.
Recent Findings
Groundwater models, including the widely used numerical models like Finite Difference Model (FDM), Finite Element Model (FEM), & Finite Volume Model (FVM), along with surface water models, are considered the most accepted models in observing the linkages. FVM, in particular, is noted for its flexibility in grid selection and conservation properties, providing ~ 90% accuracy; however, this flexibility comes at the cost of long simulation times. This review indicated that these numerical models performed simulations within seconds with an accuracy of ~ 95% when integrated with artificial intelligence and machine learning. These advancements suggest that integrating AI/ML with numerical models is a promising approach for efficient, robust GW-SW interaction modelling.
Summary
This review paper discusses the existing numerical models in groundwater-surface water (GW-SW) interactions studies and the challenges they face. This paper also highlights the need to integrate ML with these models to overcome these challenges and enhance predictive performance by improving accuracy and reducing simulation time.