Fresh-saline water dynamics in coastal aquifers: Sand tank experiments with MAR-wells injecting at intermittent regimes: Sand tank experiments with MAR-wells injecting at intermittent regimes

Seawater intrusion (SWI) has been widely investigated in physical sandbox-field experiments and analytical–numerical mathematical models. However, to the best of our knowledge, there were no studies of a commingled dynamics of the triad of saline water, natural fresh groundwater and freshwater injected in managed aquifer recharge (MAR) using sand tank experiments similar to the one presented in this paper. The main objective of this research is to investigate the encroachment and retreat of saline water in a coastal unconfined aquifer under various MAR-injection schemes. The effectiveness of MAR in mitigating SWI is explored using sand tank experiments under a spectrum of hydrological conditions that include various locations-depths of injection wells, including injection and post-injection evolution of the fresh and saline lenses-bubbles, when a pulse of treated wastewater (TWW) was inoculated into an initially static dense sea water tongue. Regimes (volumes) of freshwater jabs were also varied in our experiments. Our essentially transient flows showed that injecting water close to the toe of a “background” saltwater-freshwater interface was more effective to reduce SWI by 32% to 61.4% (depending on the injected volume, which varied from 0.5 L to 1.4 L), compared with the scenarios of wells placed away from the toe. Injection of water within the initially static saline water tongue (behind the interface) retarded the re-intrusion of saline water because the created “freshwater bubble” (topologically isolated from fresh groundwater) built an evanescent hydrological barrier against intrusion. Injection of water near the saline water boundary (SWB) and above the interface resulted in large direct loss of the injected water to the ocean, by seeping above the aquifer-intruded saline tongue, with only minimal impact on it. Injection at a shallow depth was more effective in reducing the salinity in the intruded zone of the aquifer compared with the scenario when the injection was performed deeper (near the bottom of the aquifer/sandbox). Injecting a larger volume has caused further recession of the saltwater-freshwater interface seawards (increasing the injected volume by 32%, the tip of the interface retreated by 43% more). However, our experiments showed that it is important to monitor the height of the created water table mound because a steeper hydraulic gradient in the “adverse” inland direction may provoke local farmers to pirate the injected water by dug-drilled illegal wells and creating an uncontrolled recovery, rather than a designed MAR-combatting, system. We monitored a rapid discharge (loss) of the injected water through both the vertical boundaries of the sand tank. In field MAR conditions when aquifer storage and recovery are practiced, abstraction wells can be the discharge (or drain) facility and so injection and abstraction rates should be strictly administered-balanced to increase the effectiveness of injection in mitigation of SWI. Optimization of MAR (by using the location of injection wells, rates and volumes of injection, depths of screens, duration of injection as control variables, along with maintaining the fresh water balance for the aquifer as a constraint) is necessary to avoid ineffectiveness or failure of MAR aimed at combating SWI.