A report from Circle of Blue, dated October 23, 2026, raises a question that treatment operators will recognize from their own sample files. The piece, Circle of Blue's coverage of a study on PFAS contamination in drinking water, takes the position that the problem may be larger than current monitoring shows. Its summary explains one mechanism behind that concern, noting that "When PFAS concentrations in drinking water fall below laboratory reporting limits" the results become difficult to interpret. The summary stops before it explains what follows, so this article does not speculate about the rest of the study. The discussion below concerns the operating problem the summary describes, not the study's findings.
What breakthrough depends on
Granular activated carbon removes PFAS by adsorption, and the replacement interval is usually tied to breakthrough, the point at which the effluent concentration of a target compound rises above a chosen threshold. Breakthrough is observed through sampling. The operator needs a series of measured values to estimate how quickly the bed is loading and when it will cross the line. If many influent or effluent results sit below the reporting limit, the operator has fewer usable points to fit a trend and less basis for predicting the change-out date.
The gap matters because a non-detect is a statement about the method. A result reported below the reporting limit tells the operator that the compound was not quantified above that level by that laboratory method on that day. It does not tell the operator that the bed is far from breakthrough, and it does not tell the operator how much margin remains. Treating a non-detect as zero can push the change-out date later than the bed can safely support.
Two ways to set the schedule
Operators generally choose between a fixed schedule and a sample-triggered schedule. A fixed interval, whether set by time or by bed volume treated, is simple to budget and to order replacement media for. It assumes the influent load is reasonably stable and that the design estimate still holds. A sample-triggered schedule responds to what the bed is actually doing, but it is only as good as the detection record behind it. When the record is sparse or heavily censored, the trigger can fire late, fire inconsistently, or fail to fire at all.
Neither approach is correct in every case. A utility with a long, consistent sampling history and well-characterized influent may do well with sample-triggered changes. A utility with a short history, frequent non-detects, or a recent change in source water may need a more conservative fixed interval until the record improves. The choice is a design decision, and it should be made with the data's limits in view.
Write the terms before the result arrives
The useful step is to decide the rules before a result forces a decision. The operating file should state the effluent threshold and the sample point at which it is measured. Every result should be logged with its reporting limit, method, and sample date, so a later reader can tell a true low value from a value that was never quantified. The file should also say how non-detects are treated for scheduling purposes. Treating them at the reporting limit, or at half that limit, is a stated convention that can be defended. Leaving the treatment unstated is not.
The bed-volume log belongs in the same file as the sample log. Read together, they show whether a sampling gap coincided with a period of heavy load, a source change, or a shift in flow. Without both records, the operator cannot tell whether a clean result reflects a clean bed or simply a missed window.
Where the decision reaches the customer
Residential and small-system owners face the same uncertainty with less equipment and less data. A household that relies on point-of-use treatment should ask what the device is certified to reduce, what media or membrane it uses, and how the replacement schedule was set. Some households will find that a point-of-use reverse osmosis unit fits their plumbing and their risk tolerance better than a carbon cartridge, and local RO specialists can assess that choice against a particular supply and plumbing layout. The cartridge or membrane still needs a replacement trigger, and the household should write it down.
The study may change how much PFAS utilities and regulators believe they are seeing. The carbon bed, however, responds only to what passes through it. Set the threshold, the sample point, the reporting-limit convention, and the change-out trigger in the operating record before the next sample is drawn, so the schedule does not depend on a result that the method cannot fully report.