George Mason University

After several years building water-treatment products, I wanted to understand how contaminant measurements become risk estimates and how those estimates shape treatment and regulation.

The project focused on disinfection byproducts: compounds formed when chlorine or other disinfectants react with organic matter in water. The regulated compounds are only part of the mixture. Some unregulated byproducts occur at very low concentrations but have much higher toxicity values, so whether a method can see them matters.

Every laboratory method has a detection limit. Results below that limit are reported as non-detects, but the compounds may still be present. A risk calculation then has to decide what value, if any, to substitute.

If the most toxic compounds sit near or below the reporting limit, improving the method, or changing the substitution rule, can change the calculated toxicity of the same water sample. That is not merely a laboratory detail; it can affect which treatment appears safer.

Once the method could measure lower, byproducts previously recorded as non-detects became visible, and the arithmetic changed. Across the ten systems, the compounds sitting below a conventional reporting limit accounted for 7% to 41% of the calculated mixture toxicity while making up only 1% to 8% of the total byproduct mass. A laboratory can therefore capture almost all of the mass in a sample and still miss a substantial share of the estimated harm.

The broader lesson was straightforward: a risk estimate is shaped not only by the water, but also by what the analytical method can see and how censored results are handled.