Dynamic Storage, Release, and Enrichment of some Per- and Polyfluoroalkyl Substances in the Groundwater Table Fluctuation Zone: Transport Processes Requiring Further Consideration

Per- and polyfluoroalkyl substances (PFAS) are a diverse group of contaminants that exhibit complex transport and fate behaviors that are becoming better understood (see McGarr et al. 2023 for a recent literature review). For example, it has been shown that some PFAS sorb to sediments (e.g., Schaefer et al. 2021) and microplastics (e.g., Pramanik et al. 2020; Cheng et al. 2021; Scott et al. 2021), exhibit self-assembly behavior (e.g., Dong et al. 2021), and partition into non-aqueous phase liquids (e.g., McKenzie et al. 2016; Van Glubt and Brusseau 2021; Liao et al. 2022). Further, while many PFAS are resistant to biological degradation processes, others can act as precursors and transform into other terminal PFAS (Harding-Marjanovic et al. 2015; Mejia-Avendaño et al. 2016; LaFond et al. 2023; Holly et al. 2024). Recent research has highlighted that the behavior of PFAS in the unsaturated zone is particularly complex. One of the unique characteristics of many PFAS is their surfactant properties and affinity to air-water interfaces (e.g., Brusseau 2020; Brusseau and Guo 2022; Stults et al. 2022). As a consequence, PFAS adsorption at air-water interfaces strongly influences PFAS leaching from shallow source areas (such as fire-training areas [FTAs] where aqueous film-forming foams [AFFF] were applied) through the vadose zone, and associated mass discharge to the saturated groundwater zone (Guo et al. 2020; Schaefer et al. 2021; Zeng et al. 2021; Anderson et al. 2022; Brusseau and Guo 2022; Gnesda et al. 2022; Schaefer et al. 2022; Hort et al. 2024; Stults et al. 2024).

Most of the understanding of PFAS transport behavior has been developed through controlled laboratory experiments and modeling evaluations and as more field data become available, this knowledge must be validated by the observed behavior of these contaminants in real world systems. In some cases, recent field data have illuminated important remaining knowledge gaps, highlighting the need for further study. In particular, we are unaware of any field investigations specifically designed to better understand how these compounds behave within the groundwater table (GWT) fluctuation zone (roughly defined as the zone between the seasonal high and low water table elevations). Because some PFAS preferentially accumulate at the air-water interface, it is expected that changes in moisture content and air-water interfacial area as a consequence of groundwater elevation changes will result in dynamic PFAS storage and release in the GWT fluctuation zone. This, in turn, may drive complex temporal changes in groundwater PFAS concentrations measured in monitoring wells and influence long-term PFAS mass flux/discharge within the plume. An analogous phenomenon has been well documented for light nonaqueous-phase liquid- (LNAPL-) impacted sites. The low density (relative to water) of the LNAPL and variations in water level result in temporary contaminant sequestration within the GWT fluctuation zone (i.e., the “smear zone”). In these cases, significant contaminant mass is present within this difficult-to-monitor zone, which can act as a long-term secondary source as contaminants are released over time back into groundwater. While PFAS are not generally present as a separate low-density phase, PFAS adsorbing to and releasing from the dynamically varying air-water interfaces as the GWT fluctuates presents unique complexities relative to those related to LNAPL. Importantly, this phenomenon may not be limited to high-concentration areas near the original release(s) but may also be occurring over large areas far downgradient of the original release areas.

In addition to this process, it is known that many PFAS sorb to metal oxide minerals (e.g., Alvez et al. 2020; Schaefer et al. 2021). The presence, species, and structure of these minerals are expected to change over time in the GWT fluctuation zone as a result of variable saturation and dynamic oxidation–reduction potential (ORP), or “redox” conditions (e.g., Machado-Silva et al. 2024). Consequently, sorption characteristics may vary temporally according to changes in mineral occurrence and structure, which may further complicate dynamic PFAS storage and release behavior from the GWT fluctuation zone.

It is important to note that, while the general concept of the GWT is commonly understood and typically defined as the elevation where the pore water pressure is equal to atmospheric pressure (which can be measured by the free surface in a well), the characteristics and processes operating in and near the GWT are quite complex (see Baird and Low 2022, for an excellent summary and insightful discussion). For example, trapped gas bubbles can be present within the “saturated” zone below the water table, and at some sites perched and inverted water tables (the wetting front beneath a perched water-saturated zone) occur. Furthermore, there is lateral seepage within the capillary fringe (although at a lower velocity than at the GWT due to reduced effective hydraulic conductivity) and the thickness of the capillary fringe can vary from inches to feet depending on soil type. Additionally, air-water interfaces within the GWT fluctuation zone may or may not be immobile. A more nuanced appreciation of the complexities and subtleties of the GWT fluctuation zone is likely warranted by practitioners when considering PFAS transport.

Many PFAS-impacted sites are in the early phases of characterization activity and therefore field data sets are still relatively limited. Furthermore, because the primary contaminants that have been the focus of remediation for the past four decades (e.g., petroleum hydrocarbons, chlorinated solvents, metals) are not surfactants, the impact of dynamic groundwater conditions and partitioning to air-water interfaces on contaminant transport has not been a specific focus of most contaminant assessments. However, results from a few sites provide evidence that the processes discussed above are relevant and should be considered. For example, positive correlations between PFAS concentrations and groundwater level fluctuations were observed for a field study conducted at a treated wastewater recharge facility (Cáñez et al. 2021). Below we present additional empirical data sets collected from two field sites that exhibit features that indicate these processes are important and require further study.

As a limited effort to evaluate whether the conceptual processes (Figure 1) and mechanisms discussed could result in behaviors consistent with characteristics of the observed field data, numerical simulations of PFAS transport in the subsurface beneath a synthetic AFFF-impacted source zone were conducted with and without the fluctuation of a GWT. The numerical simulations were generated by the mathematical model reported in Guo et al. (2020) and Zeng and Guo (2021). More systematic model investigations and analyses of the impact of GWT fluctuation on PFAS subsurface transport are reported in a separate study by Zeng et al. (2024). Figure 9 presents simulated PFOS and PFOA concentrations in a groundwater well near the source zone. The average GWT elevation at the center of the source zone is approximately 9 ft below ground surface. The GWT fluctuation data were obtained from a field site in Burlington County, NJ, USA. The long-term lateral hydraulic gradient is set to 1%. Long-term rainfall data measured from a site in NJ, USA were used as forcing. Additional details of the model setup and parameters can be found in Zeng et al. (2024). Compared to the simulation with no GWT fluctuation, GWT fluctuation introduces strong temporal variations to the PFOS and PFOA concentrations, with RSDs of about 190% for PFOS and 17% for PFOA (in some cases predicted monthly variations exceed one order of magnitude). As expected, PFOS has a stronger variability due to its greater K_i value and resultant interfacial activity at air-water interfaces. Furthermore, at times of low GWT elevations, the ratio between the PFOS mass in the fluctuation zone and the groundwater underneath exceeds 3.2 (the two subzones are assumed to have similar thicknesses). This is an important result, and while this analysis is not intended to be broadly representative or rigorously conclusive, it implies that at times and under certain conditions, there can be more PFAS mass (for high K_i species) stored within the GWT fluctuation zone than in the saturated plume.

The model simulations, while relatively simple, confirm storage and release behavior expected to be significant for some site conditions. However, the specific impact is expected to be highly dependent on the properties of the subsurface porous media properties (e.g., air-water interfacial area as a function water the wetting state), the nature of the GWT fluctuation (e.g., driving forces, amplitude, and frequency), the interfacial activity of PFAS, and other factors such as subsurface heterogeneity, as analyzed and discussed in detail in Zeng et al. (2024).

In our view, the development of a more sophisticated model of PFAS behavior in the GWT fluctuation zone is a time-urgent challenge. The research and practitioner communities must collaborate to understand the importance of these processes on site characterization, CSM development, risk assessment, remedy design and performance expectations, and long-term performance monitoring.