P&T vs. PRB: Which Remedy Wins?
Pump-and-treat is the default. A permeable reactive barrier costs more upfront but works passively. Deterministic analysis says P&T is cheaper. Monte Carlo reverses the decision.
How Does Pump-and-Treat Work?
The concept is simple: extract contaminated groundwater, run it through granular activated carbon (GAC) filters to strip the PFAS, and discharge the clean water back to the aquifer or a surface outfall. Pump-and-treat is the most common PFAS remedy because it's well-understood, regulators trust it, and the capital costs are manageable.
The problem is time. At 4 parts per trillion, you need to remove essentially all the PFOS — but PFOS sorbs to aquifer solids. As pumping depletes the dissolved phase, sorbed mass slowly re-releases, maintaining a low-concentration trickle that can extend cleanup for decades beyond what the dissolved-phase model predicts. At our site, the sorption coefficient (Kd) ranges from 0.5 to 20 L/kg across the Monte Carlo ensemble — a 40x range that drives enormous uncertainty in cleanup duration.
How Long Does Cleanup Take?
The deterministic model assumes average Kd, average hydraulic conductivity, and a single source concentration. It gives one answer: 45 years to reach 4 ppt across the plume. But the 200-realization ensemble tells a different story. The distribution is heavily right-skewed — the median is 55 years, but the tail extends past 140 years. In 20% of realizations, cleanup takes more than 80 years.
Why? Because Kd is the dominant driver. High sorption means more mass locked in the soil matrix, more rebound, and more decades of pumping. The deterministic model uses a single Kd value. The Monte Carlo samples the full published range. The tail is where the budget lives.
What Does It Actually Cost?
The P5 line (best case) plateaus around $67M — cleanup finishes by year 50 and costs stop accruing. The P50 line reaches $90M. But at P95, costs keep climbing past $120M because the system is still pumping at year 100.
The gap between P50 and P95 is not a modeling artifact. It's the budget risk. If you fund to the median, you're underfunded half the time. If you fund to P95, you're committing $120M+ over a century of operations.
The tail matters enormously. At P95, P&T costs exceed $120M. Budgeting to the median means underfunded half the time.
Is a PRB Better?
A permeable reactive barrier is a trench filled with reactive media (typically GAC or ion exchange resin) installed perpendicular to groundwater flow. The plume flows through it passively — no pumps, no energy, no operators. Capital cost is higher ($50M vs. $30M), but annual O&M is near zero between media replacements every 15 years.
The key advantage: PRB cost is largely insensitive to cleanup duration. Whether the plume takes 50 years or 150 years to exhaust, the barrier just sits there working. P&T, by contrast, accrues $2M/year in operations for as long as the pumps run.
At the 30-year horizon, the difference is small: P&T P50 is $65M vs. PRB at $62M. A project manager looking at median costs might pick either one. But at 100 years, P&T P95 reaches $124M while the PRB stays at $85M. The PRB's fixed-cost structure means it doesn't have a tail — and that's why it wins when you account for uncertainty.
This is the decision that changes with fidelity. Same site, same data, different model, opposite recommendation.
How We Modeled This
Remediation costs from EPA (2021) treatment technology report and ITRC (2023) PFAS guidance. All costs in net present value at a 3% real discount rate. P&T: $30M CAPEX (well field, treatment plant, discharge infrastructure) plus $2M/year O&M (energy, GAC replacement, monitoring, labor). PRB: $50M CAPEX (excavation, reactive media, backfill) plus $20M media replacement every 15 years.
Cleanup duration driven by MODFLOW 6 transport model with Kd sampled from Anderson et al. (2019) range (0.5–20 L/kg, log-normal). 200 realizations. Each realization runs P&T until all monitoring wells read below 4 ppt for 4 consecutive quarters.
Sources: EPA (2021) EPA/600/R-21/164. ITRC (2023) PFAS Technical and Regulatory Guidance. NPV at 3% real discount rate. P&T: $30M CAPEX + $2M/yr. PRB: $50M CAPEX + $20M replacement every 15 yr.
Why Turning Off the Foam Doesn't Help
AFFF use stopped at most bases by 2000. You might expect concentrations to decline after the source is removed. We tested this: 30 years of active source, then complete removal. At 80 years, the difference in well concentration is less than 1%.
The reason: the aquifer itself has become the source. Decades of PFAS sorbed to soil particles throughout the plume body slowly re-release into groundwater, sustaining contamination long after the original foam stopped flowing. With a retardation factor of 9.5, only about 10% of the PFAS mass is in the water at any given time — the other 90% is stuck to soil, slowly desorbing over decades.
This is why PFAS cleanup takes so long. It's not the source zone — it's the sorbed mass throughout the plume body. Pump-and-treat removes dissolved PFAS, but the soil keeps releasing more. Each pump cycle gets a little less, asymptotically approaching the MCL but never quite reaching it within a human planning horizon. The “forever” in “forever chemicals” describes not just the molecule but the remediation timeline.
MODFLOW 6 GWF+GWT. Source: 100 ppb constant concentration for 30 stress periods (years), then CNC boundary removed. Compared to infinite source at 80 years. Difference: <1% in well concentration, 26% reduction in plume area (138,550 m² vs 187,100 m²).