PFAS in stormwater: How clean is clean?

Air emissions, deposition, and subsequent stormwater runoff are emerging as important pathways for per- and polyfluoroalkyl substances (PFAS) to the environment. However, guidance on addressing PFAS-impacted stormwater, including treatment, and criteria for allowable concentrations of PFAS in stormwater, have largely not yet been established.

In the absence of guidance or criteria, facilities where stormwater is identified as a pathway of PFAS to the environment may need to work closely with regulatory agencies to identify appropriate source control or mitigation activities, and to develop appropriate criteria to apply for stormwater discharges.

How clean is clean? was a central question at an industrial facility that has applied polytetrafluoro-ethylene (PTFE) and other fluoropolymer coatings onto medical devices for over 40 years. The conceptual site model (CSM) identified air emissions as a primary source of PFAS to the environment that occurred primarily during historical former operations.

PFAS were deposited and accumulated onto nearby soils, the facility roof, and process equipment. These accumulations continue to be the primary and secondary sources, respectively, of PFAS to groundwater via leaching through the vadose zone and infiltration of stormwater runoff from the facility roof.

The facility is in a groundwater resource area where groundwater is extracted for nearby public and private drinking water supplies, and groundwater concentrations exceeded state regulatory thresholds near the facility.

The facility needed to identify and eliminate or mitigate the sources of PFAS to the groundwater and then derive and achieve a runoff level that would still be protective of stormwater. To assist in mitigating the ongoing impacts of PFAS in stormwater, Ramboll derived a site-specific threshold for PFAS in the facility stormwater that would be protective of groundwater as a drinking water resource.

Extensive multimedia investigations completed at the facility identified stormwater runoff discharged from the facility roof onto the permeable ground surface via 14 individual scuppers as a potential ongoing source of PFAS to groundwater. The stormwater impacts were initially mitigated by the removal and replacement of original roofing materials impacted with emitted and ‘chunked’ PFAS-containing coating residues, cleaning of the spray booth stacks, and upgrades made to the emission control equipment.

However, PFAS continued to be detected in stormwater. The average PFAS concentration, measured as a sum of six PFAS or ‘PFAS6’¹, in the stormwater was 530 nanogram/liter (ng/L), and ranged from 17 to 3,400 ng/L as measured from the individual scuppers. These PFAS6 concentrations exceeded the groundwater standard of 20 ng/L of PFAS6 for drinking water use areas.

In response to the ongoing elevated PFAS6 concentrations, Ramboll developed a plan for and implemented a comprehensive source investigation at the facility to identify the ongoing sources of PFAS in stormwater.

Working with the facility personnel, Ramboll evaluated the processes employed at the facility and their potential for generating air emissions, and then identified equipment associated with those processes for further investigation.

The investigation included collecting residue samples from spray booth stacks, oven stacks, accumulated residues or dust from the roof, rain runoff from stacks and other equipment on the roof, among others. Ramboll also measured PFAS in ambient rainfall.

The source investigation identified certain spray booth stacks as the potential predominant ongoing source of PFAS to stormwater. The spray booth stack residues contained more than 410,000 micrograms per kilogram (ug/kg) of PFAS6 six and rain runoff from the spray booth stacks contained more than 8,000 ng/L of PFAS6.

The residues and runoff samples were dominated by PFOA, consistent with the historical use of PFOA-rich fluoropolymer formulations at the facility.

A key criterion Ramboll sought to develop to assist in analyzing the data collected as part of the comprehensive PFAS source investigation was an acceptable PFAS concentration in stormwater runoff from the roof that would be protective of the groundwater resource.

While a criterion of 20 ng/L of PFAS6 had been established by the regulatory agency overseeing the investigations, the agency had not established a stormwater criterion. And, in the absence of such a criterion, planning for mitigation activities was challenging.

Therefore, Ramboll proposed a site-specific concentration threshold for PFAS6 in stormwater that would not result in greater than 20 ng/L of PFAS6 in groundwater at the downgradient edge of the facility building (i.e., within the footprint of the building).

The stormwater threshold was derived using numerical groundwater flow and fate and transport simulations that have already been developed for the facility and immediate area to evaluate remedial options for the entire site. The model was repurposed for the stormwater concentration threshold estimation.

To estimate the stormwater threshold concentration, the facility was modeled as a one-foot wide and 360-foot long rectangle, corresponding to the length of the facility building parallel to the groundwater flow direction.

MODFLOW and MT3D were used to simulate the groundwater flow and fate and transport of PFAS, respectively. MODFLOW was calibrated using site-specific data, including hydraulic conductivity. Stormwater discharge was simulated as recharge with varying PFAS concentrations, and therefore, SESOIL was used to derive a groundwater recharge rate for MODFLOW.

To simplify the scope of the modeling effort, PFOA, which was the predominant PFAS present at the site, was used as a surrogate for PFAS6. Consistent with the historical use of PFOA as a fluoropolymer processing aid, PFOA represents more than 90% of the PFAS6 concentration in groundwater within the groundwater plume and nearly 80% of the PFAS6 concentration in the facility roof runoff.

To determine the threshold concentration in stormwater resulting solely from the input stormwater, the model assumed that there were no other contributing PFAS sources to groundwater except for PFAS present in the upgradient groundwater. Other major assumptions of the model included:

• Steady state (that is, equilibrium) conditions exist such that PFAS from stormwater instantaneously impacts groundwater.
• A net groundwater recharge of 80% of average annual rainfall per year exists within the model domain.
• PFAS are not degraded or generated from precursors.
• The upgradient concentration of PFAS in groundwater is approximately 5 ng/L.
• No PFAS are present in rainfall, consistent with site-specific measurements from an upwind location.

The groundwater models were run iteratively by increasing the PFOA concentrations in the stormwater from the facility roof until the PFOA concentration in groundwater at the downgradient edge of the facility met the groundwater standard of 20 ng/L.

Using PFOA as a surrogate, the highest allowable PFAS6 concentration in the stormwater that would not result in an exceedance of 20 ng/L of PFAS6 in groundwater at the downgradient edge of the facility was developed and proposed to the regulatory agency.

Sensitivity analyses evaluated the effect of various input parameters on the model output and included the effect of 1) upgradient concentrations of PFOA in groundwater, 2) lower hydraulic conductivities and upgradient groundwater flow rates, and 3) the stormwater infiltration rate on the PFOA concentration in groundwater at the downgradient edge of the facility structure (i.e., location of interest).

None of the evaluated parameters were found to be sensitive for the model output:

• PFOA concentrations in groundwater upgradient of the facility showed minimal variability.
• Lower hydraulic conductivities, and therefore lower groundwater flow rates, which would have resulted in a lower proposed allowable stormwater concentration, were not consistent with field observations.
• The stormwater infiltration rate was estimated to be approximately 33% of the groundwater flow through the model domain and was not expected to vary significantly based on historical precipitation data.

The proposed stormwater threshold concentration was reasonable because the model conservatively assumes no dispersion, no retention, and no degradation of PFAS6.

The stormwater threshold was used to identify equipment for removal from the facility. Using data on PFAS concentrations that were measured by Ramboll in the equipment-specific runoff, those stacks and other equipment on the roof that had runoff concentrations exceeding the stormwater threshold were identified and replaced.

¹PFAS6 is the sum of perfluoroheptanoic acid (PFHpA), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), perfluorohexane sulfonic acid (PFHxS) and perfluorooctane sulfonic acid (PFOS).