Managed Aquifer Recharge (MAR) systems have supplied drinking water to rural communities in southwestern Bangladesh since 2009. Although MAR enhances water availability, there are concerns about the potential mobilization of iron (Fe), manganese (Mn), and arsenic (As) during storage. Fourteen push-pull tests (PPTs) were performed under oxidative and reductive conditions at four MAR sites. These tests involved injecting filtered and O2-saturated pond water for oxidative conditions, and sucrose-amended anoxic stored MAR water for reductive conditions, via a well in the stored MAR water. During oxidative PPTs, repeated aeration, injection, and abstraction cycles resulted in rapid consumption of dissolved oxygen (DO) with first-order rate constants of ∼52 to 72 day^-1 across all sites. DO was mainly consumed by adsorbed and dissolved Fe, with no apparent signs of pyrite and organic matter (OM) oxidation. The consistently high rate constant across the cycles suggests that heterogeneous Fe oxidation dominates. DO oxidizes Fe(II) to form Fe-(oxyhydr)oxides, resulting in the temporary removal of dissolved Fe (∼98 %), Mn (∼70–80 %), and As (60–70 %) at sites GMF11 and JJS91 due to sorption onto newly formed Fe-(oxyhydr)oxides. At sites MGS and MF05, increased As concentrations were noted due to the desorption of As from the Fe-(oxyhydr)oxides surface during abstraction. During reductive PPTs, the sucrose degraded over time, resulting in increased bicarbonate (HCO₃) and acetate concentrations and decreased pH and (sucrose-derived) DOC in abstracted water. These conditions led to the reductive dissolution of Fe-(oxyhydr)oxides, mobilizing Fe, Mn, and As, resulting in concentration peaks up to 70 mg/L Fe, 3.5 mg/L Mn, and 120 µg/L As. At MGS and MF05, similar trends for Fe and Mn were observed, while As levels did not increase. Peak concentrations were observed after about one day at JJS91, and two days at the other sites. Regular infiltration of O2-saturated water may limit mobilization of Fe, Mn, and As, while the occurrence of reduced conditions should be prevented, as they could result in mobilization of these geogenic metals and endanger the provision of safe drinking water.

Managed Aquifer Recharge (MAR) has been applied as Aquifer Storage, Transfer, and Recovery (ASTR) to provide fresh drinking water for local communities at 99 locations in southwest Bangladesh since 2009. Aerobic freshwater from ponds is filtered and subsequently infiltrated into anaerobic shallow brackish aquifers. At approximately 45% of these sites, relatively higher levels of Fe and As were observed in recovered water, which requires a better understanding of the hydrogeochemical processes that govern the Fe, Mn, and As levels in these MAR systems. Therefore, two representative sites with As above (74 ± 11 μg/L at site GMF11) and below (19 ± 6 μg/L at site JJS91), the Bangladesh drinking water standard of 50 μg/L were weekly monitored on hydrochemical changes from Dec 2017 to Dec 2018. Hydrogeochemical processes occurring during storage were quantified with inverse and forward geochemical mass balance models developed with PHREEQC. The following processes explained the changes in water quality: 1) mixing of infiltration water with native groundwater (∼90%:∼10%); 2) consumption of O₂ by a) dissolved Fe²⁺ that subsequently precipitated as Fe(OH)₃ at GMF11 and by b) dissolved and sedimentary organic matter (OM) at site JJS91; 3) reduction of SO₄ coupled to the oxidation of OM at both sites; and 4) mixing corrosion and freshening induced cation-exchange (Ca sorption; Na desorption) triggering calcite and siderite dissolution at GMF11. Dissolution of these carbonate minerals occurred to a lesser extent at JJS91, while cation exchange (Na sorption; Ca desorption) suggested that the freshwater was displaced by brackish groundwater because of inadequate infiltration at JJS91. Distinct pH values in recovered water reflected the dominance of Fe²⁺ versus OM oxidation. Siderite dissolution led to 4.3 ± 3.1 and 1.0 ± 0.5 mg/L Fe in recovered water at GMF11 and JJS91, respectively. Elevated As and Mn levels in recovered water were caused for max. 20% by mixing with native groundwater and for min. 80% by mobilization processes, mainly by desorption of As from Fe-oxides and by the dissolution of Mn-bearing siderite. Recommendations are provided to improve recovered water quality.