A report from Circle of Blue, dated October 23, 2026, takes up a question that every utility with a PFAS file eventually faces: whether the results already on record understate what customers receive. The piece, which originally appeared in Inside Climate News, raises the possibility that PFAS contamination in drinking water may be underestimated. Its summary opens with the point that "When PFAS concentrations in drinking water fall below laboratory reporting limits" the reading becomes a difficult one to interpret. The summary available here does not show the full method, so the claims should be checked against the complete Circle of Blue report on PFAS contamination in drinking water before any of them enter a board packet.
A reporting limit describes the lab, not the location
A reporting limit tells a reader what a laboratory can state about a sample. It does not say where the sample came from, when it was drawn, or what the system was doing at the time. Many utility files hold a set of results drawn from a limited number of points, on a limited calendar, under a single operating configuration. A clean summary of those results inherits all three of those limits along with the floor of the method.
Entry point and tap answer different questions
Sampling at an entry point describes water leaving the source or treatment train under the conditions the utility was running that day. Distribution sampling describes water as it moves through mains, storage, and service lines. Tap sampling, where it is done, includes premise plumbing and household fixtures. A result from one location does not carry over to another without a stated reason. Whether premise materials or stagnation could affect a PFAS reading in a particular system is an open question for that utility's own data and method, not a general assumption to adopt in either direction.
Timing and source changes shape the result
A single sampling event captures one operating state. If a utility blends sources, switches between surface water and groundwater, or changes its treatment train by season, the same point can yield different results at different times. Two events in the same season from the same location say less about annual conditions than the count suggests. The useful question is what the source mix was on each sampling date, and whether the utility recorded that mix at the time rather than reconstructing it later from memory.
What to assemble before a coverage claim is made
Before a file is cited as evidence of coverage, a utility or its consultant can assemble four items from its own records. The first is a list of sample site identifiers, each mapped as entry point, distribution, or tap. The second is the sampling date set against the source and blend record for that date. The third is the laboratory reporting limit attached to each event, since limits can differ between events and between laboratories. The fourth is the number of events at each point. Gaps in any of these should be recorded as gaps rather than filled by inference. A map that shows where no samples were drawn is as informative as one that shows where they were.
Wording for the board packet
A statement such as "no detections" needs a location and a date attached to it. A statement that results represent the system needs the sampling plan that supports the claim. If that plan was built around a fixed schedule rather than around the operating conditions the utility actually runs, the summary should say so in plain terms. Staff who draft the packet should also note which results predate any change in source, treatment, or method, because older data may not describe the system as it now operates.
The Circle of Blue report is a prompt to run this check, not a settlement of how any particular system should read its own record. The question for the next sampling round is not only what the laboratory reports but where each sample was drawn and what the system was doing when it was drawn. Answering that question in writing before the results arrive is cheaper than reconstructing it after a board or a regulator asks.