A PFAS result is only the start of a treatment decision. Granular activated carbon and ion exchange resin are sized against a concentration, a flow, and a contact time, and the change-out date follows from those three inputs. When the concentration figure is uncertain, the date is uncertain too, and a sizing spreadsheet will look more precise than the data beneath it.
Circle of Blue's coverage dated October 23, 2026 of a study suggesting PFAS contamination in drinking water may be underestimated puts the sampling question at the center. The reporting issue it raises turns on what happens when concentrations "fall below laboratory reporting limits." For a plant designer, the consequence is less a headline than an input: the influent value used in the breakthrough estimate.
What sets the change-out date
A carbon bed does not fail on a calendar. Contaminants move through the bed as a front, and the effluent begins to show them once the adsorption capacity nearest the outlet is used up. This is breakthrough. Its timing depends on influent concentration, flow rate, empty-bed contact time, and competing organic matter that occupies the same surface. Higher influent concentration shortens the run, so an underestimated input produces an optimistic run length. The error is not visible in the output until the bed begins to pass compounds it was expected to hold.
Chain length matters as well. Shorter-chain compounds generally move through activated carbon sooner than longer-chain ones, so a bed sized on a long-chain profile can break through earlier than planned when shorter compounds are present. Ion exchange resins respond differently. Their selectivity depends on the specific resin and the water chemistry, which is why a vendor's general performance claim should be tested against the utility's own water before it is written into a schedule.
Build the input before the schedule
Before setting a change-out interval, record the influent values used, the reporting limit for each target compound, and how any non-detect result was entered. Results recorded as zero, as a fraction of the limit, or excluded entirely each move the modeled concentration, and the schedule shifts with them. Write the rule down before the first run so the interval does not drift when a new result arrives. Keep the sampling dates, the lab method, and any change in either on the same sheet. A method change can alter the reported number even when the water has not changed, and the model should show that change rather than absorb it silently.
Watch the effluent, not only the calendar
Where a plant runs vessels in lead-lag service, place a sampling point between the first and second vessel, and set a trigger for early change-out rather than relying only on the projected date. Where a single vessel is in service, the effluent sample is the only check on the model. Treat the first breakthrough reading as a calibration event. Compare it with the modeled curve, then revise the influent input and the interval together, and record the reason for each revision.
Procurement follows from the same logic. A specification that names the target compounds, the reporting limit the laboratory must achieve, and the sampling frequency gives a supplier something concrete to design against. Buyers working with independent regional dealers such as Jones Air & Water can ask for the breakthrough assumptions in writing, including the empty-bed contact time and the influent value used, so the projected run can be checked against the utility's own records rather than accepted as a quoted figure.
What the study does not settle
An underestimate claim describes a possible gap in the sampling frame. It does not supply a new measured concentration for any particular plant, and it does not tell a utility how far its own reported figures sit from actual levels. Until a utility's results or a review of its method show that gap, the more defensible course is to carry a range into the treatment model rather than a single point value. Uncertainty in the input should appear in the output as a wider change-out window, not as one date presented with confidence.
The practical test is simple. If the influent value were raised by a margin the team considers plausible, would the change-out date move, and would anyone notice before the effluent did? If the answer is yes, the schedule needs a trigger tied to measured breakthrough, and the calendar date should be treated as a planning estimate rather than a commitment.