PFAS Exposure, Soil, Groundwater and Drinking Water Risk Assessment

PFAS Chemicals and Screening Levels

Risk-based screening levels for the PFAS most often encountered in soil, groundwater, and residential drinking water investigations. The numerical values below are transcribed from the November 2024 EPA Regional Screening Level (RSL) workbook, from its Residential Soil, Residential Tapwater, and Soil-to-Groundwater sheets. EPA states that RSLs provide comparison values for residential and commercial/industrial soil, air, and tapwater. A dash (—) indicates that no value is published for that chemical and pathway.

Sources and notes

• Uploaded workbook (405298.xlsx) — Nov-2024 RSL workbook supplied by user; numerical screening values in the table are transcribed from its Residential Soil, Residential Tapwater, and Soil-to-Groundwater sheets.
• EPA RSL page — EPA says RSLs provide comparison values for residential and commercial/industrial soil, air, and tapwater; current page last updated March 13, 2026.
• EPA RSL equations — EPA describes resident soil, resident tapwater, and soil-to-groundwater pathways and equations.
• EPA PFAS drinking water — Current EPA PFAS drinking-water page; as of Aug 2026 it describes the 2024 final MCLs and 2026 proposed changes.

PFAS Exposure Pathways

The screening levels above are organised by the three residential exposure pathways EPA defines in its RSL equations:

• Residential soil — ingestion. Incidental ingestion of PFAS in surface and near-surface soil at a residence, reported here as both a cancer screening level and a child noncancer screening level, in mg/kg.
• Residential tapwater — ingestion. Ingestion of PFAS in water delivered to a residence, reported as a cancer screening level and a child noncancer screening level, in µg/L, alongside the workbook’s maximum contaminant level (MCL) entry for the same compound.
• Soil to groundwater. Migration of PFAS from soil into the underlying aquifer, reported as a risk-based soil screening level (SSL) and, where an MCL exists, an MCL-based SSL, in mg/kg.

Pathway analysis, exposure factors, and exposure reconstruction as applied in RoseyFields matters are described in Section 7, Exposure Assessment on this page and on the Exposure and Risk Assessment page.

PFAS Toxicology and Risk Assessment

Each pathway in the table is reported two ways: a cancer screening level, based on a target excess lifetime cancer risk, and a child noncancer screening level, based on a reference dose. Both are derived using the EPA RSL equations linked above. The chemical-specific toxicity values behind the PFOA, PFOS, and GenX (HFPO-DA) entries are documented in the RoseyFields reference library:

Risk characterisation, cumulative assessment across co-occurring PFAS, and dose reconstruction as applied in RoseyFields matters are described in Section 8, Risk Assessment on this page and on the Epidemiology and Toxicology page.

PFAS Case Studies and Related Chemicals

Representative PFAS matters are summarised in Section 13 on this page. Full case studies, and the chemical-specific regulatory history, epidemiology, toxicology, and risk assessment pages for PFAS and related compounds, are in the RoseyFields reference libraries:

RoseyFields LLC — Paul E. Rosenfeld, Ph.D., Principal Forensic Environmental Chemist

1. What Dr. Rosenfeld Does on PFAS Matters

Per- and polyfluoroalkyl substances present a problem that most environmental methodology was not built for. The compounds do not degrade, they move through soil and aquifers far more readily than the chlorinated solvents that shaped a generation of remedial practice, they travel in air and deposit to land in quantities sufficient to contaminate groundwater miles from a stack, they bind to serum proteins rather than partitioning to fat, they have half-lives in the human body measured in years, and they are present at some concentration in nearly every person in the country — which means there is no clean background against which to measure an increment.

Dr. Rosenfeld’s PFAS practice addresses four connected questions.

Where did it come from and how did it get there? Source identification and characterization, air dispersion and deposition modeling, groundwater and surface water fate and transport, and forensic differentiation among co-located sources.

Who was exposed, to how much, and when? Exposure reconstruction from reconstructed water and air concentrations, from product and occupational contact, and — working backward — from measured serum concentrations using compound-specific toxicokinetics.

What does that dose mean? Toxicity value derivation from the primary epidemiological and toxicological literature, hazard and cancer risk characterization, cumulative assessment across co-occurring PFAS, and derivation of Proposition 65 safe harbor levels where none has been adopted.

What was known, by whom, and when — and what should have been done? The state of scientific and industrial knowledge over time, product stewardship practice, and whether the investigation, monitoring, notification, and response actions taken by a manufacturer, user, facility, or water system met the standard of care prevailing at the time.

The work is performed for plaintiffs and defendants, for municipalities and water utilities, and for state agencies. Methodologies are evaluated under Daubert and analogous standards.

2. PFAS-Specific Credentials and Experience

Dr. Rosenfeld holds a Ph.D. in Soil Chemistry from the University of Washington (1999), an M.S. in Environmental Science from U.C. Berkeley (1995), and a B.A. in Environmental Studies from U.C. Santa Barbara (1991). He has more than 25 years of experience in exposure assessment, dose reconstruction, environmental forensics, and air dispersion modeling, is founder and principal of RoseyFields LLC, and was a founding partner and principal of SWAPE from 2003 to 2025.

His PFAS work dates to the early 2000s, not to the recent wave of litigation. He presented on the fate, transport, and persistence of PFOA and related chemicals at the Mealey’s C8/PFOA Science, Risk and Litigation Conference in October 2005; at the International Society of Environmental Forensics’ Emerging Contaminants program in June 2005; and at the National Ground Water Association’s Ground Water and Environmental Law Conference in July 2005, where he addressed the fate, transport, persistence, and toxicology of PFOA and related perfluorochemicals. In April 2009 he presented at the Ground Water Summit and Ground Water Protection Council Spring Meeting on PFOA and PFOS contamination of drinking water arising from the use of aqueous film-forming foam at airports in the United States — a source category that took roughly another decade to reach general recognition. In May 2018 he presented on the science of perfluorinated chemicals and the difficulty of their remediation at Law Seminars International in Seattle.

Peer-reviewed PFAS publication. Rosenfeld, Spaeth, Remy, Byers, Muerth, Hallman, Summers-Evans, and Barker (2023), Perfluoroalkyl substances exposure in firefighters: Sources and implications, published in Environmental Research, addresses occupational PFAS exposure across aqueous film-forming foam, protective equipment, and fire-ground pathways. His broader published record includes multi-decade drinking water exposure reconstruction and cancer risk assessment (Rosenfeld et al., 2024), attic dust and paired blood-lipid studies used as historical exposure archives (Hensley et al., 2007; Feng et al., 2010; Gonzalez et al., 2010), community exposure assessment using dispersion modeling with empirical validation (Chen et al., 2012), The Risks of Hazardous Waste (Rosenfeld & Feng, 2011), and three volumes of the Handbook of Pollution Prevention and Cleaner Production (Cheremisinoff & Rosenfeld, 2009, 2010, 2011).

Categories of PFAS assignment undertaken. Fluoropolymer and fluorochemical manufacturing facility investigations; aqueous film-forming foam release sites at military installations, airports, and fire training grounds; carpet and textile manufacturing districts; wastewater treatment plant and biosolids land application investigations; landfill leachate and landfill gas assessments; firefighter turnout gear and protective equipment exposure investigations; consumer product source studies; PFAS in lead-sheathed telecommunications cable settings; drinking water utility and private well contamination assessments; PFOA multidistrict litigation exposure and risk assessment support; AFFF settlement damage model methodology; Proposition 65 Maximum Allowable Dose Level derivation for PFOA; and rebuttal of opposing exposure, modeling, and risk opinions.

Testimony. His curriculum vitae, current as of July 20, 2026, itemizes 114 listed testimony entries across more than 100 distinct matters in state and federal courts in over 30 states — approximately 120 depositions, 16 trial appearances, and testimony at a contested Daubert hearing. PFAS matters within that record include testimony in the United States District Court for the District of Colorado in Bells et al. v. The 3M Company et al. (depositions March and April 2018), in the United States District Court for the Western District of Wisconsin in Varline et al. v. 3M Company (deposition June 2026), and in the United States District Court for the Eastern District of Missouri in Peeler et al. v. SRG Global Coatings, LLC, 3M Company, and Atotech USA, LLC (depositions May and June 2026).

3. Why PFAS Behaves Differently

Several properties of this chemical class defeat assumptions carried over from conventional contaminants, and misapplying those assumptions is a recurring source of error in PFAS assessments.

Persistence. The carbon-fluorine bond is among the strongest in organic chemistry, and perfluoroalkyl acids resist hydrolysis, photolysis, and biodegradation under environmental conditions (Buck et al., 2011; Kwiatkowski et al., 2020). Natural attenuation, the workhorse assumption of solvent and petroleum remediation, does not apply. Mass released is mass still present, redistributed among media.

Mobility, and the reason it is not simply predicted. Perfluoroalkyl acids are anionic at environmental pH and far more water-soluble than the neutral organics for which conventional partitioning models were built. Sorption correlates with perfluorocarbon chain length and with organic carbon, but also with cation exchange and calcium bridging, so a single organic-carbon partition coefficient poorly predicts retardation (Higgins & Luthy, 2006). Air-water interfacial adsorption provides an additional retention mechanism in the unsaturated zone that conventional retardation calculations omit entirely, and it can substantially delay arrival at the water table (Brusseau, 2018). The practical result is that different PFAS in the same release migrate at different rates, producing chromatographic separation along a plume — which is diagnostically useful and, if ignored, misleading.

Precursor transformation. Many products, particularly aqueous film-forming foams, contain polyfluorinated precursors rather than terminal perfluoroalkyl acids. Those precursors transform in the environment into PFOA, PFOS, and related acids over time, which means a source can continue generating regulated compounds decades after release ceased, and that analysis for terminal acids alone understates total PFAS mass (Houtz et al., 2013). The total oxidizable precursor assay was developed to quantify that hidden reservoir (Houtz & Sedlak, 2012).

Toxicokinetics. PFAS bind to serum albumin and liver proteins rather than accumulating in adipose tissue, so body burden tracks serum protein rather than body fat, and lipid-based models are inapplicable. Human elimination half-lives are measured in years, with substantial reported variation, and renal resorption is saturable and nonlinear (Li et al., 2018). Serum concentration therefore integrates years of exposure, which is what makes reverse dose reconstruction feasible.

Ubiquity and the background problem. PFAS are detectable in the serum of the large majority of the U.S. population (Calafat et al., 2007), and low concentrations occur in drinking water and consumer products broadly. There is no natural background — every molecule is anthropogenic — but there is a substantial general-population exposure baseline. Attribution of an individual’s body burden or a water system’s concentration to a specific source therefore requires comparison against reference distributions and affirmative exclusion of alternative pathways rather than mere detection.

Nomenclature discipline. PFAS is a class of thousands of compounds. Opinions must specify which compounds are at issue, distinguish perfluoroalkyl acids from polyfluorinated precursors, distinguish linear from branched isomers, and identify which analytical method was used, because method scope determines what could have been detected. Loose use of “PFAS” to mean “PFOA and PFOS” and rigorous use of the term to mean the full class produce very different conclusions about mass, sources, and exposure.

4. Sources and Source Characterization

Attribution begins with knowing what each source category releases, in what proportions, and by what mechanism.

Fluoropolymer and fluorochemical manufacturing. Historically the largest point sources, releasing to air through process vents and stacks and to water through outfalls, with fugitive releases from handling, drying, and waste management. Air deposition from these facilities has contaminated soil and groundwater at substantial distances downwind (Davis et al., 2007; Shin et al., 2011).

Aqueous film-forming foam. Military installations, airports, refineries, chemical plants, and fire training grounds, where repeated training discharges to unlined ground created concentrated source zones (Anderson et al., 2016; Guelfo & Higgins, 2013). Foam formulation matters: fluorotelomer-based and electrochemical-fluorination-based products yield different compound signatures, and military specification requirements governed which products were purchased and when.

Textiles, carpet, and paper. Carpet and textile finishing districts and paper coating operations applied fluorochemical treatments at industrial scale, discharging to wastewater systems and, through biosolids, to agricultural land.

Metal finishing and chrome plating. PFOS was used as a mist suppressant in hard chrome plating, a source category relevant to plating and coating operations.

Wastewater treatment and biosolids. Treatment plants do not destroy PFAS; they redistribute it between effluent and solids, and precursor transformation during treatment can increase measured terminal acid concentrations across the plant. Land-applied biosolids transfer PFAS to soil, with subsequent leaching to groundwater and uptake into crops and forage (Sepulvado et al., 2011).

Landfills. Leachate from municipal solid waste landfills carries PFAS from consumer products at nationally significant loads (Lang et al., 2017), and landfill gas provides an additional release pathway.

Firefighter protective equipment. Turnout gear moisture barriers and outer shells contain fluoropolymer treatments that shed PFAS to dust and to the wearer, an exposure pathway distinct from foam contact (Peaslee et al., 2020).

Consumer products. Food packaging, cookware, cosmetics, ski wax, water-repellent apparel, and treated furnishings and textiles.

Forensic differentiation. Where sources are co-located, differentiation relies on compound-ratio and homologue-pattern analysis, branched-to-linear isomer ratios that distinguish electrochemical fluorination from telomer chemistry, presence of source-specific markers, precursor-to-terminal-acid relationships, chain-length distribution as a function of transport distance, and the timing of arrival relative to documented operations. Combined with the operational record, these lines of evidence support apportionment among sources rather than a single aggregate conclusion.

5. Air Dispersion and Deposition Modeling of PFAS

The air pathway is the most frequently overlooked route by which PFAS reaches drinking water, and it is a documented one.

Why it matters. Stack and fugitive emissions disperse, deposit to soil and vegetation, and subsequently leach to groundwater. Because PFAS neither degrades nor sorbs strongly, deposited mass eventually reaches the aquifer, producing groundwater contamination at distances from a facility that a discharge-only conceptual model cannot explain. Modeling of a fluoropolymer facility demonstrated that air emissions and subsequent deposition, combined with groundwater transport, accounted for the observed distribution of PFOA in environmental media and in nearby drinking water supplies (Shin et al., 2011; Davis et al., 2007).

Method. Emission rates are reconstructed from production records, mass balance, permit files, stack tests, and reported inventories, resolved to an annual or finer time series across the facility’s operating history. AERMOD is run with representative meteorology, terrain, and building downwash, computing both ambient concentration and wet and dry deposition flux (U.S. EPA, 2004, 2017). Source characterization distinguishes stack releases from fugitive releases, the latter often dominating near-field ground-level concentration because they occur at low elevation with little plume rise. Deposition modeling requires the particle size distribution and the vapor-particle partitioning of the compounds at issue, and treats deposition of the fraction in each phase accordingly.

Validation. AERMOD’s performance for perfluorinated compounds specifically has been evaluated against field data. Barton, Zarzecki, and Russell (2010) compared a 24-hour PFOA air concentration simulation against nine perimeter monitoring locations across nine sampling events, and a one-year deposition simulation against 42 soil and grass samples at 16 locations near a fluoropolymer facility. The model located the maximum concentration correctly 89 percent of the time, and modeled deposition agreed with measured soil and grass concentrations across most of the study area. Where predictions diverged, the tendency was toward conservative overprediction, which the authors attributed to the steady-state assumption of constant hourly wind direction and characterized as a positive result for exposure assessment purposes. This is a more direct validation than exists for many compounds routinely modeled without objection.

The linked pathway. The defensible analysis carries the air model through to the water: modeled deposition flux becomes a soil loading rate, soil loading becomes a leachate source term, and leachate becomes a groundwater source, with the resulting predicted groundwater concentrations compared against monitoring well and private well data. Where that chain reproduces the observed spatial pattern, the air-to-water pathway is established quantitatively rather than asserted.

Further detail on dispersion methodology, meteorological data, receptor design, and model validation appears on the firm’s air dispersion modeling page.

6. Groundwater and Surface Water Fate and Transport Modeling

Source zone characterization. Release location, timing, mass, and formulation are established from operating records, discharge monitoring reports, aerial photography, fire training records, and soil and soil-gas data. For AFFF sites, training frequency, foam concentrate volumes, application rates, and drainage configuration convert operational history into a mass loading estimate.

Unsaturated zone transport. Vadose zone transit is modeled with retardation reflecting chain-length-dependent sorption, cation effects, and air-water interfacial adsorption (Higgins & Luthy, 2006; Brusseau, 2018), and with precursor transformation where the source contained polyfluorinated precursors. Long unsaturated-zone residence times mean a release from decades ago may still be delivering mass to the water table, which bears directly on both liability timing and remedial expectations.

Saturated zone transport. Groundwater flow and transport modeling — analytical where the setting permits, numerical where it does not — is calibrated to measured heads and to the observed plume geometry. Compound-specific retardation produces differential migration, so the leading edge of a plume is enriched in short-chain compounds and the core in longer-chain compounds; matching that pattern is a strong test of a transport model. Matrix diffusion into fine-grained units creates back-diffusion sources that sustain concentrations long after the primary source is controlled.

Capture zones and drinking water wells. Well capture zone analysis determines which wells could have drawn from a given source area, and under which pumping configurations and time periods. For public systems, source water contributions, blending, treatment, and distribution system residence time determine what actually reached a given service area in a given year — the quantity that matters for exposure, as distinct from raw water concentration at a wellhead.

Surface water and sediment. Outfall discharges, stormwater runoff, and atmospheric deposition to watersheds are routed with dilution, sediment interaction, and bioaccumulation into fish, supporting fish consumption exposure pathways.

Biosolids and land application. Where the pathway is land-applied biosolids, PFAS loading is calculated from application rates and residuals concentrations, and leaching to groundwater and uptake into crops and forage modeled from soil concentration (Sepulvado et al., 2011).

7. Exposure Assessment

Drinking water. Historical concentrations are reconstructed by year and service area from monitoring data, transport modeling, and system operating records. Ingestion dose uses age-specific intake rates (ATSDR, 2023; U.S. EPA, 2011). Because PFAS are non-volatile and poorly absorbed dermally, ingestion generally dominates household water exposure — the opposite of the volatile organic case, where showering pathways often rival ingestion. Assessments that carryover volatile-compound assumptions to PFAS overstate the inhalation and dermal contribution.

Serum-based reverse reconstruction. Where measured serum concentrations exist, historical intake and water concentration are back-calculated using a one-compartment toxicokinetic model parameterized with compound-specific half-life and volume of distribution, and accounting for age, sex, and pregnancy and lactation history. This approach is independent of any assumption about water concentration or intake rate, making it a genuine cross-check on the media-based reconstruction. Convergence between the two is the strongest evidentiary position available; divergence must be explained.

Reference-population comparison. Measured serum concentrations are compared against NHANES distributions for the matching age, sex, and survey cycle (Calafat et al., 2007; CDC, 2024). Given general-population background, this comparison is what separates elevated exposure from ordinary exposure, and omitting it leaves an attribution opinion unsupported.

Occupational exposure — firefighters. Exposure is built pathway by pathway: foam contact during training and response, protective equipment as a continuing source (Peaslee et al., 2020), fire-ground combustion products, station dust, and contaminated apparatus and gear storage. Serum studies of firefighters provide benchmarks against which reconstructed exposures can be evaluated (Rotander et al., 2015; Trowbridge et al., 2020; Rosenfeld et al., 2023). Job history, training frequency, gear issuance and laundering practice, and station conditions drive the reconstruction.

Other pathways. Diet, particularly fish from affected water bodies; incidental soil and dust ingestion at impacted properties; home-grown produce and livestock where biosolids or irrigation water were affected; consumer product contact; breast milk transfer to nursing infants, which produces the highest body-weight-normalized doses in the exposure profile; and occupational exposure in manufacturing, plating, textile, and waste handling settings (Sunderland et al., 2019).

8. Risk Assessment

Toxicity value derivation. Agency values for PFAS have changed substantially and continue to diverge among jurisdictions, so the firm derives chemical-specific values from primary study data rather than adopting a single regulatory pronouncement. The sequence follows EPA reference dose and carcinogen methodology: structured point-of-departure selection across the health outcomes with the strongest weight of evidence, preferring medium- and high-confidence human epidemiology over animal studies where the epidemiological database supports dose-response modeling; benchmark dose modeling to a lower confidence limit rather than reliance on NOAELs; pharmacokinetic conversion to human equivalent doses rather than default body-weight scaling, including gestation and lactation modules where developmental endpoints govern; and uncertainty factors assigned on the basis of the actual critical study and documented factor by factor (U.S. EPA, 2002, 2005, 2012).

The endpoint literature supporting these derivations includes hepatic effects (Gallo et al., 2012; Nian et al., 2019), immune effects and vaccine antibody response (Grandjean et al., 2012, 2017; Budtz-Jørgensen & Grandjean, 2018), serum lipids and cardiovascular endpoints (Steenland et al., 2009; Dong et al., 2019), thyroid effects (Lopez-Espinosa et al., 2012), birth weight and developmental outcomes (Darrow et al., 2013; Sagiv et al., 2018; Wikström et al., 2020), kidney and testicular cancer in the C8 population (Barry et al., 2013), renal cell carcinoma (Shearer et al., 2021), and the broader toxicological synthesis (Fenton et al., 2021). The C8 Science Panel probable-link determinations, arising from the Mid-Ohio Valley community studies, remain a significant part of the evidentiary landscape.

The firm’s chemical-specific methodology pages document the derivations compound by compound for PFOA and PFOS, including the selected reference dose, cancer classification and slope factor, the basis for applying or declining age-dependent adjustment factors, and the key uncertainties.

Risk characterization. Non-cancer hazard is the reconstructed dose divided by the reference dose; excess lifetime cancer risk is lifetime average daily dose multiplied by the cancer slope factor, compared against the 10⁻⁶ to 10⁻⁴ range used in regulatory practice.

Cumulative assessment across co-occurring PFAS. Real exposures involve mixtures. Where case-specific data and current science support a shared endpoint and additive toxicity, a cumulative hazard index is calculated across co-occurring compounds. EPA’s 2024 drinking water regulation itself adopts a Hazard Index approach for a defined subset of PFAS, which provides a regulatory anchor for the additive treatment of that subset (U.S. EPA, 2024a). Beyond it, mixtures methodology remains an area requiring explicit case-specific justification rather than assertion, and RoseyFields treats it that way (ATSDR, 2018).

Proposition 65. Where OEHHA has not adopted a safe harbor level for a listed PFAS compound, a No Significant Risk Level or Maximum Allowable Dose Level can be derived using the methodology in OEHHA’s regulations, and RoseyFields has performed such a derivation for PFOA. The comparison is between reasonably anticipated daily intake from the product or setting at issue and the derived safe harbor level (OEHHA, 2015; Cal. Code Regs. tit. 27).

Medical monitoring and damage models. Risk output supports medical monitoring analyses, class certification and exposure-differentiation questions, property and water-treatment damage models, and settlement allocation methodologies keyed to exposure and dose.

9. Standard of Care and the State of Knowledge

“Standard of care” carries two distinct meanings in PFAS litigation, and conflating them weakens both. The technical expert’s role is to establish the factual and scientific record — what was knowable, what was known, what the prevailing practice was — while the legal duty is for the court and the trier of fact.

9.1 Corporate and industry state of knowledge

This analysis constructs a documented chronology of what was known about PFAS toxicity, persistence, bioaccumulation, and environmental mobility, and when, drawn from the primary sources: internal corporate research, toxicology studies, worker medical surveillance and serum monitoring programs, environmental sampling conducted by manufacturers, submissions to EPA — including the administrative record compiled in EPA’s PFOA docket — trade association materials, product stewardship program documents, safety data sheets across successive revisions, product literature and marketing claims, and the contemporaneous open scientific literature.

The chronology is then compared against what was communicated and to whom: to workers, to downstream industrial customers, to fire departments and other end users, to water utilities, to communities, and to regulators. Where a gap exists between internal knowledge and external communication, the analysis identifies the gap, its date, and the specific documents establishing it. Reporting obligations under the Toxic Substances Control Act, including the substantial risk provision at 15 U.S.C. § 2607(e), provide one objective benchmark against which corporate conduct can be measured, as do the voluntary stewardship commitments manufacturers themselves adopted.

The analysis is symmetrical by design. The same chronology that can establish early knowledge can also establish that a particular fact was not reasonably knowable at a particular time, that a specific downstream user could not have known what its supplier knew, or that a defendant’s conduct was consistent with prevailing practice and regulatory expectation in the relevant period. RoseyFields performs this work from both directions, and the methodology is the same either way: the record governs the conclusion.

9.2 Professional and technical standard of practice

The second sense concerns whether the technical work performed at a site — by an operator, a consultant, a water utility, or a public agency — met the standard of practice for its time.

The analysis compares what was done against the applicable benchmarks then in effect: EPA analytical methods and detection capability available at the time, since a system cannot be faulted for failing to detect at concentrations below the achievable reporting limits of the period; investigation and monitoring guidance current at the time, including quality assurance project plan requirements and sampling protocols; military specifications governing foam procurement and use; consensus standards governing protective equipment selection, care, and maintenance, including the NFPA standards applicable to firefighter protective ensembles and their maintenance; occupational exposure control practice and hazard communication requirements; and drinking water monitoring and notification obligations, including the unregulated contaminant monitoring rules and subsequent regulatory requirements. Comprehensive technical and regulatory guidance for PFAS has been compiled by the Interstate Technology and Regulatory Council (ITRC, 2023) and provides a widely used reference for current practice.

Common questions in this category include whether a facility’s investigation scope was adequate to the known source configuration; whether monitoring was designed to detect the compounds actually released, using methods capable of the necessary sensitivity; whether sampling and quality control were adequate to support the conclusions drawn, given PFAS-specific contamination risks; whether responses to detections were timely and proportionate; whether notification to affected parties occurred when it should have; whether treatment selection and design were appropriate; and whether a consultant’s conceptual site model accounted for pathways — particularly air deposition and precursor transformation — that the science of the period recognized.

A recurring and important distinction is between what could not have been detected and what was not looked for. Analytical capability for PFAS at low nanogram-per-liter concentrations developed over time, and it is not a criticism of a 1990s investigation that it did not report a concentration no laboratory of the period could measure. It is a different matter where the compounds were known to be present, the methods existed, and the analysis was not performed.

10. Regulatory Framework

Regulatory context matters for standard of care, for benchmark comparison, and for damages.

Federally, EPA established the first National Primary Drinking Water Regulation for PFAS in April 2024, setting enforceable maximum contaminant levels of 4.0 nanograms per liter for PFOA and PFOS, 10 nanograms per liter for PFHxS, PFNA, and HFPO-DA, and a Hazard Index approach for mixtures of a defined subset, with associated monitoring and public notification requirements (U.S. EPA, 2024a). In May 2024 EPA designated PFOA and PFOS as hazardous substances under CERCLA, with consequences for liability, reporting, and cost recovery (U.S. EPA, 2024b). These followed a sequence of non-enforceable benchmarks — the 2016 lifetime health advisories at 70 nanograms per liter combined, and the substantially lower interim advisories issued in 2022 — and nationwide occurrence monitoring under the third and fifth Unregulated Contaminant Monitoring Rules. Significant new use rules under TSCA restrict resumed manufacture of legacy compounds.

Internationally, PFOS was listed under the Stockholm Convention in 2009, PFOA in 2019, and PFHxS in 2022, and OECD hazard assessment work on PFOS dates to 2002.

State standards preceded and in many cases remain more stringent than the federal MCLs, and they differ substantially among jurisdictions in both numeric value and compound coverage. Because litigated exposure typically spans periods governed by different benchmarks — or by none — the applicable comparison value must be identified by period rather than by reference to current standards alone.

Biosolids and land application, industrial pretreatment, and effluent limitation frameworks continue to develop, and this area of regulation is changing quickly enough that the operative requirements should be confirmed for the specific period and jurisdiction at issue.

11. Sampling, Analysis, and Data Quality

PFAS data quality problems are distinctive enough to warrant separate attention, because a risk or attribution opinion built on unreliable data fails regardless of the sophistication of the modeling.

Method scope determines findings. EPA Method 537.1 and Method 533 apply to drinking water and cover defined analyte lists; Method 1633 addresses non-potable water, soil, sediment, biosolids, and tissue with a broader list. A compound outside a method’s scope will not be reported, and its absence from a data table is not evidence of its absence from the sample. Comparing datasets generated by different methods, or across different reporting limits and periods, requires explicit reconciliation.

Precursors and total oxidizable precursor analysis. Targeted analysis for terminal perfluoroalkyl acids omits the polyfluorinated precursor reservoir that continues to generate them. The total oxidizable precursor assay quantifies that reservoir and is often decisive at AFFF sites, where precursor mass can exceed measured terminal acid mass substantially (Houtz & Sedlak, 2012; Houtz et al., 2013).

Field and laboratory contamination. PFAS-containing materials are ubiquitous in ordinary sampling equipment and personal gear — fluoropolymer tubing and tape, waterproofed clothing and boots, certain field notebooks and adhesives, some decontamination products. Sampling protocols must exclude them, and field, equipment, and trip blanks are essential rather than optional. Laboratory background contributions and instrument carryover require attention at the low nanogram-per-liter concentrations at issue.

Matrix effects, branched isomers, and quantitation. Matrix suppression, the treatment of branched versus linear isomers, calibration standard selection, and isotope dilution practice all affect reported concentrations, and quantitation conventions differ among laboratories in ways that matter when datasets are pooled.

Non-detects and statistics. Censored data are handled with Kaplan-Meier or regression-on-order-statistics methods rather than by deletion or naive substitution, and exposure point concentrations are derived as upper confidence limits on the mean using distribution-appropriate estimators (U.S. EPA, 2015).

Validation. Full data validation against method QC criteria, and reconciliation of multi-consultant, multi-decade datasets, precede any use of the data for exposure or risk purposes.

12. Daubert and Litigation Considerations

Federal Rule of Evidence 702, as amended effective December 1, 2023, requires the proponent to show more likely than not that testimony rests on sufficient facts or data, is the product of reliable principles and methods, and reflects a reliable application of those principles to the facts. Daubert (1993), Joiner (1997), and Kumho Tire (1999) govern, with the Reference Manual on Scientific Evidence (Federal Judicial Center & National Research Council, 2011) the standard judicial reference. Methodologies are evaluated under Daubert and analogous standards.

The methods applied here are the established ones: AERMOD under Appendix W, with published validation specifically for perfluorinated compound dispersion and deposition (Barton et al., 2010); standard groundwater flow and transport modeling; RAGS and ATSDR exposure and dose methodology; EPA reference dose and carcinogen risk assessment methodology for toxicity value derivation; and documentary and chronological analysis for state-of-knowledge opinions.

PFAS-specific reliability disputes cluster in identifiable places. Background and attribution — treating any detection as proof of exposure to the defendant’s source without reference-population comparison or exclusion of alternative pathways. Compound conflation — opinions that fail to specify which compounds are at issue, or that extrapolate PFOA findings to structurally different PFAS without justification. Toxicity value selection — adopting one jurisdiction’s value without acknowledging the divergence among agencies, or extrapolating a value derived for one compound to another. Mixtures additivity asserted rather than justified beyond the subset for which a regulatory basis exists. Pathway omission — most commonly the air deposition pathway, and precursor transformation as a continuing source. Carryover assumptions — applying volatile-compound showering and inhalation methodology, or organic-carbon-only retardation, to compounds that behave differently. Data quality — reliance on datasets with inadequate blanks, mismatched method scopes, or unreconciled reporting limits. Anachronism in standard-of-care opinions — measuring 1980s or 1990s conduct against detection capability, guidance, or regulatory standards that did not exist at the time, which is the single most common defect in this category and cuts against plaintiffs and defendants alike.

Two structural boundaries are stated explicitly in the work product. First, exposure and risk quantification is not a specific causation opinion; connecting a reconstructed dose to an individual’s disease requires additional reasoning about dose-response, latency, and competing causes, supplied by an appropriately qualified expert. Second, a technical state-of-knowledge chronology establishes what was known and communicated; it does not establish a legal duty or its breach, and it is not offered for that purpose.

Rebuttal work applies the same framework in reverse, examining source characterization, model configuration and calibration, retardation and transformation assumptions, exposure parameter sourcing, toxicity value derivation, data quality, and period-appropriateness of standard-of-care benchmarks, and re-running the analyses to test how much of an opposing conclusion survives its own assumptions.

13. Representative PFAS Matters

Consistent with RoseyFields’ confidentiality practice, the following are described without client, counsel, facility, or location identifiers.

Fluoropolymer manufacturing facility — combined air and water pathway. Historical PFOA emissions were reconstructed from production and inventory records and modeled in AERMOD for both ambient concentration and wet and dry deposition across the surrounding area. Modeled deposition was converted to soil loading and carried forward as a groundwater source term, and predicted concentrations were compared against soil sampling, monitoring well data, and the distribution of impacted private wells — establishing an air-deposition pathway to soil and groundwater distinct from direct discharge.

AFFF release site at a military installation. Foam use history was reconstructed from training records, procurement documents, and site drainage configuration to develop mass loading estimates by source area. Unsaturated and saturated zone transport was modeled with chain-length-dependent retardation and precursor transformation, and well capture zone analysis identified which drinking water wells could have drawn from each source area under documented pumping configurations and periods.

Firefighter occupational exposure and turnout gear. Exposure was characterized across foam contact during training and response, protective equipment as a continuing source, fire-ground exposure, and station dust, using job history, training frequency, and gear issuance and maintenance practice. Measured serum concentrations were used to back-calculate historical intake with a compound-specific toxicokinetic model and compared against firefighter cohort studies and NHANES reference distributions. Related published work appears in Rosenfeld et al. (2023).

Textile and carpet district — wastewater and biosolids pathway. PFAS loading from finishing operations to a municipal treatment system was characterized, with attention to precursor transformation across the plant, and land application of biosolids evaluated as a secondary source through soil loading, leaching to groundwater, and crop and forage uptake. Affected private wells and public supply sources were evaluated against modeled and monitored concentrations.

Drinking water system exposure and risk assessment. Historical concentrations were reconstructed by service area and year from monitoring records, system operating and well pumping history, blending, and transport modeling. Age-specific ingestion doses were developed, hazard quotients and cumulative hazard indices calculated across co-occurring PFAS, and excess lifetime cancer risks estimated using chemical-specific toxicity values derived from the primary epidemiological literature.

Proposition 65 safe harbor derivation. For PFOA, a Maximum Allowable Dose Level was derived from the primary toxicological literature following the methodology in OEHHA’s regulations, and reasonably anticipated daily intake for the exposure scenario at issue was estimated and compared against the derived level.

14. Ten Frequently Asked Questions

1. Can air emissions really contaminate groundwater miles away? Yes, and it is documented. Because PFAS neither degrades nor sorbs strongly, deposited mass migrates through soil to the water table rather than attenuating. Fate and transport modeling of a fluoropolymer facility showed that air emissions and subsequent deposition, combined with groundwater transport, accounted for the observed distribution of PFOA in environmental media and drinking water supplies (Davis et al., 2007; Shin et al., 2011). Omitting the air pathway is among the most consequential errors in PFAS source attribution.

2. Is AERMOD reliable for PFAS specifically? It has been validated for a perfluorinated compound against field measurements of both air concentration and deposition, with the maximum concentration located correctly 89 percent of the time and modeled deposition agreeing with 42 soil and grass samples across most of a study area (Barton et al., 2010). Where predictions diverged the model tended to overpredict conservatively. That is a more direct validation record than exists for many compounds modeled without objection.

3. Everyone has PFAS in their blood — how can exposure be attributed to one source? By comparison and exclusion, not by detection alone. Measured serum concentrations are compared against NHANES distributions for the matching age, sex, and survey cycle (CDC, 2024); alternative pathways are evaluated and excluded where the evidence permits; and compound patterns, isomer ratios, and precursor relationships are examined for source-diagnostic signatures. An attribution opinion resting on detection alone is not supportable, and the general-population baseline is precisely why.

4. Why do different PFAS show up at different distances from the same source? Because retardation depends on perfluorocarbon chain length, on cation and organic carbon interactions, and on air-water interfacial adsorption in the unsaturated zone (Higgins & Luthy, 2006; Brusseau, 2018). Shorter-chain compounds migrate faster, so plume leading edges are enriched in them and cores in longer-chain compounds. That pattern is diagnostically valuable and, if a transport model reproduces it, a meaningful check on the model.

5. What are precursors and why do they matter? Many products, particularly AFFF, contain polyfluorinated compounds that transform environmentally into PFOA, PFOS, and related acids. A source can therefore keep generating regulated compounds decades after the release stopped, and analysis for terminal acids alone understates total PFAS mass. The total oxidizable precursor assay quantifies the reservoir (Houtz & Sedlak, 2012; Houtz et al., 2013).

6. Which toxicity values should be used when agencies disagree? The divergence should be disclosed and its effect on the result quantified rather than resolved silently. RoseyFields derives chemical-specific reference doses and cancer slope factors from primary study data using benchmark dose modeling, pharmacokinetic dose conversion, and documented uncertainty factors, so the value is traceable to the underlying studies rather than to a single regulatory pronouncement (U.S. EPA, 2002, 2005, 2012).

7. Can hazard quotients for different PFAS simply be added? For a defined subset, EPA’s 2024 drinking water regulation adopts a Hazard Index approach, which provides a regulatory basis for additive treatment of those compounds (U.S. EPA, 2024a). Beyond that subset, additivity requires case-specific justification based on shared endpoints and current mixtures science, and should be presented as a reasoned assumption rather than as settled methodology (ATSDR, 2018).

8. Can a company be faulted for not detecting PFAS in the 1990s? That depends on what was achievable and what was known. Analytical capability at low nanogram-per-liter concentrations developed over time, and it is not a valid criticism that an investigation failed to report a concentration no laboratory of the period could measure. The different question — whether compounds known to be present were not analyzed for using methods that did exist — is a fair one. Measuring historical conduct against current capability and current standards is the most common defect in standard-of-care opinions in this field.

9. What does a state-of-knowledge analysis actually consist of? A documented chronology assembled from internal corporate research and toxicology, worker medical and serum monitoring records, company environmental sampling, regulatory submissions including EPA’s PFOA administrative record, trade association materials, successive safety data sheet revisions, product literature, and the contemporaneous open literature — compared against what was communicated, to whom, and when. It establishes the factual and scientific record. It does not establish legal duty, and is not offered for that purpose.

10. What is needed to begin, and what is produced? Useful materials include facility operating, production, and purchasing records; air permits, emission inventories, and stack tests; discharge monitoring reports and outfall data; AFFF procurement, formulation, and training records; all environmental analytical data with laboratory reports, QC, and method documentation; water system operating, pumping, blending, and monitoring records; biosolids application records; serum or biomonitoring results; residential and employment history for the individuals at issue; and, for standard-of-care work, internal documents, safety data sheets, product literature, and regulatory correspondence. The deliverable is a report documenting source characterization, model configuration and calibration, exposure reconstruction, toxicity value derivation, risk characterization with cumulative assessment, data quality evaluation, uncertainty and sensitivity analysis, and — where retained for it — a sourced state-of-knowledge chronology, with model files, databases, statistical output, and calculation spreadsheets produced so the analysis can be independently reproduced.

References