A Risk Assessment Model of PFAS Transport and Fate from the Beneficial Reuse of Biosolids at Renewable Water Resources

This assessment and report was performed by Cameron Colby, P.E., a Master's in Environmental Engineering candidate from Clemson University and employee of Renewable Water Resources, with expected program completion and graduation on December 17, 2020. Dr. Timothy DeVol, PhD served as the primary advisor for this project. Background and Introduction Poly-and perfluoroalkyl compounds (PFAS) are an emerging category of synthetic organic contaminant first included in the EPA's Third Unmonitored Contaminant Monitoring Rule 3 (UCMR 3) in 2012 (Environmental Protection Agency, 2016). Six PFAS compounds, perfluorooctanesulfonic acid (PFOS), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorohexanesulfonic acid (PFHxS), perfluoroheptanoic acid (PFHpA), and perfluorobutanesulfonic acid (PFBS) were named to UCMR 3 (Environmental Protection Agency, 2016). PFAS compounds have historically been used in firefighting foams and, more recently, widely used in numerous consumer items such as textiles, non-stick cookware, paper products, carpets, and cleaning materials (Hamid, 2016). Industrial uses are also prevalent, incorporating PFAS compounds into processes such as metal plating, electronics production, and photography (Hamid, 2016). At the time of this report, no studies were identified that assessed the fate and transport of PFAS from the land application of wastewater biosolids to a human receptor. It is imperative for water resource recovery facilities to be able to quantify and understand, on a scientific basis, the potential impact of PFAS contamination from the biosolid product stream as it pertains to groundwater, vegetation, livestock, and a human receptor. Where lifecycle quantification of a contaminant are not measured or modeled, managerial and operational decisions involving biosolids could be made arbitrarily or unnecessarily, leading to costly program burdens. This risk assessment sought to quantify the possible cumulative oral risk posed to a human receptor from the land application of biosolids from water resource recovery facilities owned and operated by Renewable Water Resources (ReWa). Six biosolids samples from each of four water resource recovery facilities were analyzed for the presence of PFOS, PFOA, PFNA, PFHxS, PFBS, and pentafluorobutanoic acid (PFBA), with the average concentration of each compound from each facility used in a risk assessment. Transport and fate of PFAS from the land application of municipal Class B liquid biosolids in the Piedmont region of South Carolina were modeled in one dimension using derivations of the contaminant transport equation through the vadose zone, groundwater, and vegetative pathways. Chemical partioning, diffusion, bioaccumulation, and transfer parameters were compiled from multiple peer reviewed studies. The level of impact to a human receptor was quantified as a dose in milligrams per kilogram per day of contaminant and compared to the oral chronic or sub-chronic established or proposed reference dose, where available. Summary findings Vadose zone transport velocity was found to be retarded by the compounds' calculated soil partitioning coefficients (KD), which are dependent on the fraction of organic carbon present in the soil. In general, compounds with longer carbon chains (C>7) experienced longer retardation than those with shorter chains and are expected to be present at higher concentrations in shallower soil than in deeper soil. PFOA was an exception, and the assessment predicted it reaching the groundwater table in approximately 40 years as opposed to PFOS, also containing an eight-carbon chain, which was predicted to reach the groundwater table in 128 years. Additionally, vertical transport appeared to occur more quickly than horizontal transport, which is a factor of percolation rate and can vary based on region specific rainfall amounts and soil attributes. Saturated transport below the groundwater table was less retarded than vadose zone transport, suggesting the air-water interface in the vadose zone may play an important role in a compounds' mobility. In general, and similarly to vadose transport, soil partitioning played a large role in the retardation of groundwater transport, with longer carbon chain compounds taking longer to reach a well located ten meters from the point of application when compared to shorter chain compounds. PFOA was again an exception to this pattern. Transport through the vadose zone and groundwater presented the most temporally varied behavior amongst PFAS compounds, largely due to differing organic carbon sorption behavior. Biosolid concentration levels resulted in a maximum modeled groundwater contamination level an order of magnitude below the EPA's current health advisory level of 70 parts per trillion for PFOS or PFOA, signifying a nominal risk. Groundwater consumption did not contribute to the exceedance of any oral chronic or sub-chronic reference doses. PFOS and PFHxS exhibited modeled concentration in cow's milk from the consumption of fescue and Bermuda grass on land application fields in excess of the oral chronic RfD and sub-chronic RfD respectively. PFHxS also exceeded the oral sub-chronic RfD in cow muscle tissue. This suggests that livestock and livestock products may be a source of potential exposure for utilities land applying biosolids and having those fields' products serve as feedstock or grazing space. Having a grasp of the end-use of vegetation from land application fields could become a critical component in biosolid program decisions. Additional studies are needed to firmly establish compound bioaccumulation and transfer factors. Municipalities should consider, where feasible, understanding the lifecycle of biosolid impact to local, regional, or nationally distributed livestock products, i.e. milk and meat, produced from animals fed hay and wet feed from biosolid amended fields. Model verification Soil samples in 3 locations from the field used in the risk assessment were collected to verify the model. Two six-inch core samples were taken at the outer boundaries of the field, with each separated into 3-inch sections for analysis. A third core sample was taken at the middle of the field and measured 24-inches, with analysis being conducted on 1-3-inches, 4-6 inches, 6-12 inches, and 12-24 inches. At the time of this abstract, results have not yet been returned. The expected result should verify that compounds with longer carbon chains (C>7) will have higher concentrations in the shallower samples than when compared to deeper samples, and shorter chain carbons (C<7) should be present at deeper soil depths. Final results would be incorporated into the presentation and final manuscript.