8% of U.S. Water Systems Sit Over a Federal PFAS MCL — Large Systems Hit 15%

Federal drinking-water standards for PFAS are no longer a proposal. The Environmental Protection Agency’s April 2024 National Primary Drinking Water Regulation set enforceable Maximum Contaminant Levels (MCLs) for PFOA and PFOS at 4.0 parts per trillion each, plus limits for three additional PFAS and a Hazard Index mixture. The question that matters for utilities, state primacy agencies, and households is simpler than the rule text: what share of monitored systems already sit above those lines on the national survey that was running while the rule was finalized?

EPA’s Fifth Unregulated Contaminant Monitoring Rule (UCMR 5) is that survey. Systems sampled for 29 PFAS (and lithium) at entry points between 2023 and 2025. The January 2026 Data Summary — covering roughly 95% of expected results, about 9,240 public water systems with full result sets at roughly 24,050 sampling locations — then compared location averages to the April 2024 MCLs for technical assistance. That comparison is not a compliance determination. Compliance will use running annual averages of quarterly compliance monitoring, and published deadlines still point to 2029 (with May 2025 signals that EPA may extend PFOA/PFOS timelines). Still, the averages are the best national preview of how large the “over the MCL” footprint already looks.

The dashboard above opens on the size-tier share of systems over any PFAS NPDWR MCL, then layers contaminant-specific rates, state detection concentration (≥UCMR minimum reporting level), co-occurrence curves, and a size-versus-rate scatter.

The headline is 8.0% — but size matters more than the average

EPA’s Table 4 footer states three unique-system rates for systems with at least one sampling-location average above any of the regulated PFAS MCLs or the Hazard Index:

The 8.0% figure is not a simple average of the three percentages. EPA weights by the national SDWIS inventory of community and non-transient non-community systems so that the census of larger systems does not dominate the national story. The policy punch line is still the size gradient: large systems are roughly twice as likely as small systems in the reporting set to show at least one average above an MCL. That pattern matters for capital planning. Treatment trains, pilot studies, and rate cases concentrate where customer counts — and often industrial or military source footprints — are largest.

PFOA and PFOS carry almost the entire MCL footprint

Break the same Table 4 cut by compound and the hierarchy is stark. Across size tiers, roughly 8.9% of systems with full results have a PFOS average above 4.0 ppt, and about 8.0% do for PFOA. Large-system PFOS and PFOA rates sit near 12% and 11%. By contrast, PFHxS over 10 ppt is about 0.7% overall; the Hazard Index about 0.8%; PFNA about 0.1%; HFPO-DA (GenX) about 0.03% — three systems in the GenX large/medium cells combined in the January cut.

That split is why May 2025 agency language about reconsidering HFPO-DA, PFHxS, PFNA, and the Hazard Index — while keeping PFOA and PFOS as the core federal standards — maps onto occurrence, not just politics. In the UCMR 5 averages published so far, the compliance problem is overwhelmingly a PFOA/PFOS problem. The rarer MCLs still matter for the systems that hit them, but they do not move the national percentage.

Contaminant rows can double-count systems that exceed more than one MCL. That is why the size-tier “any MCL” lines (15.3% / 8.6% / 7.3%) are the right unique-system denominators for the headline, while the compound rows answer “which chemical.”

Occurrence by state is highly skewed — even before the MCL lens

EPA’s printed Data Summary does not publish a state-by-state MCL exceedance table. The UCMR 5 Data Finder can filter averages above MCLs by state, but a transparent national concentration story also needs a consistent occurrence metric. Compilations of the EPA occurrence text files (Plain Water Data’s PFAS Tracker among them) show that any PFAS at or above the UCMR minimum reporting level appears in about 34% of tested systems nationally — far above the 8% MCL-exceedance estimate — and that state detection rates diverge sharply.

Among states with meaningful sample counts, New Jersey (~65%) and Connecticut (~64%) sit at the top of the detection-rate distribution. Massachusetts, South Carolina, and Texas cluster in the low-to-mid 50s. Florida and Pennsylvania land in the mid-40s. At the other end, Mississippi (~4%), Hawaii (~4%), and Arkansas (~6%) show sparse detections in the same federal methods. California’s detection rate (~33%) is near the national average even though absolute detection counts are large because the tested denominator is large.

Two caveats keep this panel honest. First, detection ≥ MRL is not an MCL exceedance. Many systems that “detect” PFAS sit below 4 ppt for PFOA/PFOS. Second, states differ in system-size mix and source-water portfolios; a high detection rate can reflect denser industrial history, AFFF use patterns, or simply more entry points in the sample. The dashboard’s concentration curve still shows the geographic punch: detections among larger tested states accumulate faster than a uniform share of systems would imply.

Mixtures are the rule, not the exception

UCMR 5 also documents co-occurrence. EPA reports that 66% of sampling locations with at least one PFAS ≥ MRL have multiple PFAS ≥ MRL. Table 5’s cumulative counts run from about 6,400 locations with ≥1 unique PFAS down through roughly 3,100 with ≥3 and about 1,100 with ≥6, with a thin tail out to 13 unique PFAS at a single location. Unique PWS counts track the same decay.

For treatment design, mixtures matter as much as single-analyte peaks. Granular activated carbon, ion exchange, and high-pressure membranes behave differently across chain lengths; a system chasing PFOA alone may still need a design envelope for co-occurring shorter-chain acids that appear frequently above MRLs (PFPeA, PFHxA, PFBA, PFBS show up far more often as detections than as MCL exceedances).

What UCMR averages are — and are not — for compliance desks

Three technical details from the Data Summary keep the chart from being misread as a violation list:

1. Ground-water locations in UCMR 5 often contribute two semi-annual results to the average; surface-water locations contribute four quarterlies. Compliance RAAs under the NPDWR require four quarterlies at each sample point.
2. Rounding follows the MCL significant-digit rules (for example, PFOA/PFOS averages ≥ 0.00405 µg/L count as above 4.0 ppt).
3. Small systems serving fewer than 3,300 people appear as a representative sample, not a censusone reason EPA publishes the weighted 8.0% national line.

ASDWA’s implementation materials stress the same operational point state primacy agencies already live with: UCMR results can support initial monitoring conversations and public notice, but the enforceable path runs through compliance sampling, treatment decisions, and — for many systems — state revolving-fund and emerging-contaminant grant queues.

Policy clock: 2029, exemptions, and a narrowing chemical set

The April 2024 rule gave systems until 2029 to meet the MCLs. In May 2025 EPA announced intent to extend PFOA/PFOS compliance deadlines, build a federal exemption framework, and rescind/reconsider the HFPO-DA, PFHxS, PFNA, and Hazard Index pieces. Until that rulemaking finishes — and until courts settle related challenges — the published MCLs remain the legal reference. Occurrence data already say where the capital stress concentrates: large systems, PFOA/PFOS, and states with dense detection.

Caveats worth keeping on the page

The January 2026 cut is still ~95% complete; final fall-2026 files can move counts. Tribal and territorial systems appear in occurrence files with small denominators that swing rates. State detection rankings use a different threshold than MCL comparisons and should not be pasted into a “percent over 4 ppt” sentence. Population served is not the same as system count: a 15% large-system exceedance rate covers far more people than a 7% small-system rate. And “over the MCL on a UCMR average” is an early warning, not a Safe Drinking Water Act violation today.

Sources: EPA UCMR 5 Data Summary (January 2026, EPA 815-S-26-001); EPA UCMR 5 Data Finder and occurrence program pages; EPA PFAS NPDWR materials; ASDWA PFAS resources; Plain Water Data UCMR5 PFAS Tracker state occurrence compilation.