Seawater intrusion occurs when saltwater infiltrates freshwater sources like coastal aquifers and rivers driven by the contrast in density between seawater and freshwater, which is due to differences in dissolved salt content. The process is typically accentuated by changes in natural recharge conditions and by groundwater pumping.

Coastal regions worldwide have been under increasing pressure due to factors such as growing populations, rising standards of living, and the effects of climate change and sea level rise. These pressures have led to aquifer over-drafting, resulting in seawater intrusion and the deterioration of groundwater quality. To address these issues, water administrators and stakeholders must explore combinations of more efficient pumping schemes, demand reduction strategies, and technological interventions like desalination.

It is then vital to investigate the sustainable management of water resources in coastal regions facing challenges such as seawater intrusion, groundwater quality deterioration, and the impacts of climate change and sea level rise. We address the development of effective solutions for managing coastal groundwater systems through the development and application of simulation-optimization (SO) frameworks, which rely on the integrating calibrated groundwater simulators with optimization algorithms, subject to various constraints. While SO models offer potential in enhancing water supply resilience to SWI, their application has been limited due to extensive computational requirements of variable density groundwater flow simulators.

To address this, "surrogate" models, which mimic full-scale model responses at a fraction of the computational cost, have been proposed. Surrogates may be categorized into data-driven and model-driven types. They aim to minimize the number of full-model runs needed for accurate predictions. While data-driven surrogates, such as artificial neural networks, have been explored in coastal aquifers, challenges persist in accurately modelling nonlinear objective functions and large numbers of decision variables typical in SWI problems. Model-driven surrogates, though less explored, show promise. Realistic SWI problems involve various hydrogeological complexities and nonlinearities, which influence the computational costs of building accurate model surrogates. While surrogate models hold potential for informed management decisions regarding coastal groundwater resources, further research is needed to address these challenges.

Another key aspect is the need to engage with water agencies, policymakers, administrators, and stakeholders to bridge the gap between scientific research and practical applications in coastal water management. By fostering collaboration, sharing knowledge, and involving key stakeholders in the research process, the goal is to create meaningful impacts, promote sustainable practices, and contribute to the resilience of coastal water resources.