Modeling PFAS transport, persistence and source attribution in an esker aquifer : A case study on the Hammarby groundwater reservoir

Sammanfattning

Eskers are of vital importance to Swedish society, constituting a significant portion of Swedish drinking water supply. Their high permeability also makes them vulnerable to contaminants such as PFAS. Norrvatten is a producer of drinking water with several groundwater abstraction facilities along the Northern Stockholm esker that are used as reserve sources of water. In the Hammarby reservoir, PFAS are found consistently throughout most of the esker, with limited data suggesting a fully developed contamination plume. The source is unknown and there is a lack of knowledge regarding PFAS transport behavior. To investigate potential sources and their respective pathways, mass transport modeling was performed in Feflow, a groundwater modelling software. Two models of the reservoir were used, with varying hydraulic conductivities and meshes, and mass transport modules were implemented for this thesis. Simulated PFAS concentrations were compared with observations to identify which sources were the most probable. Sorption to the esker was also estimated, using an assumption that the source could be up to 90 years old and had reached steady state conditions recently. With these maximum partitioning coefficients (Kdmax) derived, the expected persistence of PFAS after hypothetical removal of sources was evaluated at the Hammarby water source. The PFAS simulated were perfluorobutanoic acid (PFBA) and perfluorooctane sulfonic acid (PFOS). To further increase knowledge of PFAS presence in the reservoir, sampling of previously untested groundwater wells was also done.The results suggest that at least two sources are present in the reservoir. One is leaching perfluorocarboxylic acids (PFCA), the other is leaching perfluorosulfonic acids (PFSA) and they are both likely positioned by the upstream edge of the reservoir. It is also possible but less likely that a third source is contributing both groups of PFAS further downstream. The sources must be old, between 50 – 90 years depending on location, to have reached the steady state conditions present. These large travel times led to low estimated sorption, with maximum partitioning coefficients ranging between 0.03 – 0.16 L kg-1 for PFOS and 0.00 – 0.20 L kg-1 for PFBA. These values are only upper bounds for sorption, and the true sorption could be any value within these intervals. With these derived sorption values and using PFOS and PFBA as representatives, concentrations of PFSA would have decreased significantly (down to 10% of today’s concentrations) within 55 – 95 years, concentrations of PFCA by longer. Even if the sources were to be removed soon, PFAS will persist for several decades in the Hammarby reservoir.