Case Study: PFAS Tracing

This case study illustrates the simulation of a PFAS plume in groundwater, highlighting the unique characteristics of this family: no significant degradation within the considered timeframes, highly mobile transport for short-chain compounds, and the critical distinction between precursors and terminal compounds. The practical objective is to understand how a legacy release continues to impact a modern well field and why measured concentrations fluctuate over time even without changes to the apparent source.

Context

The site is a former fire-training area that utilized AFFF (Aqueous Film-Forming Foams) containing PFAS between the 1980s and 2005. Operations have ceased, and source zone soil was partially excavated in 2010. A drinking water well field located approximately 1.2 km downgradient provides water to a local community. Recent analyses show the presence of several PFAS, with concentrations approaching the regulatory limit for the "PFAS 20" sum defined by the European Directive.

Three questions drive the simulation: what are the projected concentrations at the well field over the next 30 years, what role do precursors remaining in the source zone play in future plume feeding, and will the PFAS 20 sum exceed regulatory thresholds?

Scene Setup

The model footprint covers 1.6 km along the flow axis and 600 meters transversely. The modeled depth reaches 25 meters to include the shallow aquifer exploited by the well field. The stratigraphy is simplified but representative of typical field-testing contexts.

Scene Element	Configuration
Horizontal Footprint	1600 × 600 m, aligned with the flow axis
Depth	25 m, referenced to ground level
Calculation Grid	160 × 60 × 50 = 480,000 cells
Unsaturated Soil (0 to 3 m)	Sandy loam, moderate organic matter
Shallow Aquifer (3 to 18 m)	Coarse sand, porosity 0.28, K = 8e-4 m/s
Semi-permeable Layer (18 to 22 m)	Silty sand, porosity 0.15, K = 1e-6 m/s
Deep Aquifer (22 to 25 m)	Coarse sand, porosity 0.30, K = 1e-3 m/s
Sampling Points	5 piezometers and 3 extraction wells

Target Compounds and Parameters

The simulation simultaneously tracks five compounds representative of typical AFFF chemistry: a major fluorotelomer precursor (8:2 FTS), two long-chain terminal compounds (PFOS and PFOA) considered conservative, and two highly mobile short-chain compounds (PFHxS and PFBA). Each compound has its own Kd and transport trajectory.

Compound	Role in Plume	Kd (L/kg) in Sand	Retardation Factor R
8:2 FTS	Precursor, gradually feeds PFOA	1.5	Approx. 12
PFOS	Terminal C8 sulfonate, highly persistent	2.5	Approx. 20
PFOA	Terminal C8 carboxylate, persistent	0.8	Approx. 7
PFHxS	Terminal C6 sulfonate, mobile	0.3	Approx. 3
PFBA	Terminal C4 carboxylate, highly mobile	0.05	Approx. 1.4

The transformation of the 8:2 FTS precursor into terminal compounds follows slow kinetics: the selected half-life is 10 to 20 years, depending on local microbial conditions. Consequently, this transformation only becomes significant at the decadal scale, which is exactly the timeframe of interest for this case.

Choosing to simulate five individual compounds rather than a single "Total PFAS" metric is what allows us to answer the initial question. The leading edge of the plume is dominated by short-chain compounds (PFBA, PFHxS), while the source zone remains rich in long-chain compounds (PFOS) and precursors. Using an aggregated sum would have masked this spatial separation, which is critical for interpretation.

Source

The source is modeled as a 200 × 80 meter surface area, with an active flux from 1980 to 2005, followed by a decreasing residual emission after the 2010 partial excavation. Inventories of foam volumes applied during operations allow for a total injected mass estimate in the range of several hundred kilograms of cumulative PFAS. The distribution between compounds reflects typical AFFF compositions of that era.

Period	Source Intensity
1980 to 2005	Nominal flux, near-continuous active emission
2005 to 2010	No new releases, desorption from soil
2010 to Present	Residual emission reduced by a factor of 5 post-excavation
Prospective Scenario	Slow exponential decay of the residual source

Results

Following a 45-year simulation (1980 to 2025) and a subsequent 30-year projection, the plume exhibits a highly differentiated spatial structure based on the compound tracked. Short-chain compounds reached the well field several years ago and account for the concentrations currently measured at the field boundaries. Long-chain compounds are still largely trapped near the source zone due to adsorption. The 8:2 FTS precursor migrates slowly but, by transforming into PFOA over time, contributes to a delayed PFOA influx that has not yet reached its peak.

Compound	Current Concentration at Well Field	Predicted Peak	Peak Year
PFBA	40 ng/L	60 ng/L	Approx. 2030
PFHxS	15 ng/L	30 ng/L	Approx. 2040
PFOA	8 ng/L	25 ng/L	Approx. 2055
PFOS	4 ng/L	15 ng/L	Approx. 2065
PFAS 20 Sum (Est.)	70 ng/L	110 ng/L	Approx. 2045

Interpretation

The model accurately reproduces current measurements at intermediate piezometers and well field pumps, providing reasonable confidence in the medium-term projection. The most significant qualitative finding is that the cumulative concentration will continue to rise for several more decades, despite the active source having stopped 20 years ago. This inertia is a direct consequence of long-chain adsorption and slow precursor transformation: what reaches the well field today largely reflects emissions from 20 to 40 years ago, and what will arrive in 20 years reflects, in part, what was emitted up until 2005.

The PFAS 20 sum threshold is exceeded in the projection around 2040, reaching its peak before slowly declining. This projection is sufficient to justify anticipatory action: the question is no longer *if* the well field water will be affected, but *when* and for *how long*. Standard options (activated carbon treatment, well field reconfiguration, or searching for alternative resources) can be sized based on this quantified diagnostic.

Projected values carry significant uncertainty margins, particularly regarding precursor transformation kinetics and residual soil desorption. "Slower desorption" and "faster transformation" variants should be explored in the multi-hypothesis comparison module to bound the prediction.

Extensions

"Deeper excavation" variant: simulate the effect of a second intervention on the source soil to quantify the expected concentration gain at the well field.
"Hydraulic barrier" variant: introduce a pumping well between the source and the well field to extract the plume before it reaches the resource.
"Active carbon at intake" variant: skip groundwater remediation but model treatment at the well field head, quantifying the cumulative mass to be treated over 30 years.
TOP Assay scenario: integrate an estimate of unidentified precursor mass and test prediction sensitivity to this unknown variable.

Learn More

PFAS: Panel, precursors, and emerging isotopy.
Hydrogeological Parameters: The role of flow in the rapid transport of short-chain PFAS.
Multi-Hypothesis: Exploring the variants mentioned above.