Assessment of soil buffer capacity on nutrients and pharmaceuticals in nature-based solution applications

The ability of a soil to sustain infiltration rates and to attenuate pollutants is critical for the design and operation of Managed Aquifer Recharge/Soil Aquifer Treatment and phyto-treatment schemes, also referred to as "Blue Infrastructures". We investigated the buffering capacity of a sediment sample and a peat soil sample for nutrients and selected pharmaceutical compounds and its evolution under continuous infiltration of secondary treated wastewater (TWW) in column experiments. Samples were obtained from two blue infrastructures, the Sant'Alessio Induced River Bank Filtration plant and the San Niccolò large-scale phyto-treatment plant in Italy, and were mainly different in their organic carbon contents (0.9 and 48%, respectively). In the column experiments, a constant infiltration rate of about 0.5 L/d was maintained for 6 months. After 4 months of operation, diclofenac and carbamazepine were spiked into the TWW to evaluate their fate. Water quality was monitored by periodic water sampling from the column inflow, at sampling ports along the column length, and at the outflow. Hydraulic conductivity (K) was also monitored. The hydraulic conductivity of the Sant'Alessio sediment decreased by a factor of 10 during the first 10 days of infiltration and then stabilized, while for the San Niccolò K (initially lower) remained constant for 50 days until it decreased following a change of the redox condition in the column. The different redox conditions, due to the two different soils tested, influenced also the concentration and mobility of PO₄^3-, Fe, Mn, and NPOC, and the speciation of the redox sensitive elements (nitrogen and sulfur). NOPC and phosphate were enriched during the filtration through San Niccolò peat soil (from 2 to 4 times, respectively), while they were buffered by the Sant'Alessio sediment (from 0.2 to 0.4 times, respectively). Diclofenac removal (69% and below 20% for San Niccolò and Sant'Alessio, respectively) was related to sorption and degradation processes and it was lower than the removal of carbamazepine in both soils (76 and 35%). The buffer capacity differences between the two soils were higher for diclofenac (62%) than carbamazepine (35%). Nevertheless, since no apparent degradation of carbamazepine was detected in both soils, its persistence in the soil may have a larger impact in case of desorption, posing contamination risk to groundwater. The results highlight the importance of the soils or sediments to be used as medium in such nature-based solutions for their operations. They also offer an approach to, e.g., tailor man-made soil layers in infiltration basins. We strongly suggest that soil characteristics and test duration are carefully considered in designing these infrastructures, when nature-based processes are the choice for dealing with reuse of treated wastewater management issues.