Potential for the desalination of a brackish groundwater aquifer under a background of rising sea level via salt intrusion prevention river gates in the coastal area of the Red River Delta, Vietnam

Additional freshwater sources are required in many parts of the world, including the coastal areas of the Red River Delta (RRD), where the groundwater (GW) is generally brackish. Determining a feasible method for desalinating brackish aquifers would help provide additional freshwater sources. However, substantial desalination of brackish aquifers cannot be achieved under the natural conditions of GW flow and precipitation recharge. Although rainfall recharge to the shallow Holocene aquifer has occurred for hundreds of years, the aquifer still remains brackish since the natural hydraulic conditions do not allow a complete mixing between the fresh recharged water and aquifer salinized water or the discharging of the aquifer salinized water. The planned salt intrusion prevention gates in the Red River, Tra Ly River and Hoa River in the RRD coastal area, combined with increased GW abstraction and associated aquifer recharge with fresh river water, could result in the gradual desalination of the shallow Holocene aquifer. These effects would help improve the area’s resilience to freshwater shortages and sea level rises and would allow for the creation of a long-term sustainable water resource development plan to manage the salinization of water resources caused by sea level rises. Finite element (FE) modeling of GW flow, solute transport via GW flow and dynamic programming (DP) have been used to study the potential desalination of brackish aquifers, the magnitude of GW abstraction quantities and the spatial and temporal aspects of desalination. FE modeling of GW flow coupled with DP was utilized to identify the magnitude of sustainable abstraction quantities and the GW flow field, which is required in salt transport models. Multiple sizes of elements and time steps were used to adapt to the unsteady state of GW flow and hydraulic head variables between the elements in the FE meshes in order to ensure reasonable accuracy of numerical modeling. The GW flow and salt transport modeling and DP allowed determining quasi-steady-state GW abstraction rates and aquifer salinity levels for conditions that did and did not include the shallow Holocene unconfined aquifer recharge from rainfall. The aquifer modeled domain which is supposed to serve the pumping well field is 1.5 km2. The results showed that the Holocene aquifer may provide a stable abstraction rate of 100 m3/day starting in the 6th year (for the worst-case scenario with zero aquifer recharge from rainfall) to 130 m3/day starting in the 3rd year (for the scenario with aquifer recharge equal to 3% of the rainfall levels). During the first years of GW abstraction, the desalination of the brackish upper Holocene aquifer will mainly occur in the area close to the river, and at the 18th year of abstraction, almost the entire area between the river and line of pumping wells would be desalinized. From the 10th year of abstraction, the abstracted water has a total dissolved solids content lower than 0.5 g/l for the worst-case scenario with zero aquifer recharge from rainfall and lower than 0.42 g/l for the scenario with aquifer recharge equal to 3% of the rainfall. The modeling results indicate the simulated process by which abstraction of groundwater adjacent to the Tra Ly River could desalinize the brackish aquifer via freshwater recharge from the river.