NDMA and NDMA precursor attenuation in environmental buffers prior to groundwater recharge for potable reuse

Highlights: • NDMA in percolation basins was completely removed by natural photolysis. • NDMA removal in other environmental buffers is related to capacity for photolysis. • NDMA photolysis did not generate significant amounts of NDMA precursors. • Filtration and wetland treatment did not effectively remove NDMA or NDMA precursors. Natural attenuation of N-nitrosodimethylamine (NDMA) and NDMA precursors was evaluated in infiltration basins, a riverbed filtration system, and constructed wetlands operated as part of a managed aquifer recharge system. Initial NDMA concentrations up to 9.0 ng/L in infiltration basins (advanced purified, recycled water) before sunrise declined to non-detect (<1.5 ng/L) by 10:00 A.M due to natural photolysis (half-life of 33 to 86 min dependent on solar irradiance). NDMA fortified controls adjacent to the infiltration basin showed similar results, while concentrations in dark controls did not change over the basin's hydraulic retention time. NDMA precursor concentrations did not change significantly in the basin containing advanced-treated water from a potable reuse treatment plant, indicating that photolysis did not remove NDMA precursors nor did photolysis produce a significant amount of precursors. For the other environmental buffers evaluated, NDMA removal was variable through laboratory scale soil columns (22 cm height), in full-scale riverbed filtration system that pre-filters water prior to infiltration basin recharge, and in the constructed wetland. Variability in NDMA removal through the wetlands is attributed to high turbidity. In the case of the riverbed filtration system, variability is likely due to short exposure times to sunlight. For the soil columns, limited NDMA removal is attributed to inefficacy of soil aquifer treatment in removing NDMA over short travel times/distances. NDMA precursors were also ineffectively removed in these systems, with effluent concentrations occasionally exceeding influent concentrations. Overall, the removal of NDMA in environmental buffers utilized for planned or de facto indirect potable reuse is dependent on the system's capacity for photolysis, while NDMA precursors are more recalcitrant and unlikely to be removed in such systems without enhancement or sufficient hydraulic residence times.