Abstract
The interaction between surficial shallow aquifers of poorer quality and semi-confined water-supply aquifers poses a potential risk for degradation of the water supply. Groundwater engineers and hydrogeologists use groundwater models to synthesize field data, conceptualize hydrological processes, and improve understanding of the groundwater system to support informed decision-making. Models for decision-making, called management models, aid in the efficient planning and sustainable management of groundwater systems. Management models search for the best or least-cost management strategy satisfying hydrologic and environmental regulations. In management models, a simulation model is linked or coupled with an optimization formulation. Widely used optimization formulations are linear, non-linear, quadratic, dynamic, and global search models. Management models are applied but are not limited to maximizing withdrawals, minimizing drawdown, pumping costs, and saltwater intrusion, and determining the best locations for production wells. This paper theoretically presents the development of groundwater wellfield management strategies and the corresponding modeling framework for each strategy's evaluation. Depending on the strategy, the modeling effort applies deterministic (simulation) and stochastic (simulation-optimization) techniques. The goals of the optimization strategies are to protect wells from potential contaminant sources, identify optimal future well installation sites, mitigate risks, and extend the life of wells that may face water contamination issues.•Several management strategies are formulated addressing well depth, seasonal pumping operation, and mapping no-drilling or red zones for new well installation.•Modeling methodologies are laid down that apply thousands of numerical simulations for each strategy to simulate and evaluate recurring patterns of contaminant movement.•The simulation model integrates MODFLOW and MODPATH to simulate 3D groundwater flow and advective contaminant movement, respectively and is transferred via FloPy to couple with the optimization/decision model using a custom Python script.