How to read your water quality report without a chemistry degree

The short answer: find the big detections table, ignore the MCLG column, and read three things: the measured level, the range, and the sampling year. If the measured level is a range across multiple plants, your tap sits somewhere inside it, not at the average. Everything else on the page is context or reassurance.

What is a Consumer Confidence Report?

It is the annual water quality report your utility must produce under the Safe Drinking Water Act and deliver to customers by 1 July each year, covering the previous calendar year. It lists contaminants detected in the water supply, compares them to federal limits, and explains any violations. Nearly every community water system in the country publishes one, usually as a PDF.

What do the columns mean?

The detections table is where the information lives, and its column headings are the single biggest barrier to reading it.

MCL, maximum contaminant level. The legally enforceable limit. Exceed it and the utility must act and notify you. It is set with cost and feasibility factored in, so it is a compliance threshold rather than a health target.

MCLG, maximum contaminant level goal. The level at which no known health effects occur, with a safety margin, and with no consideration of cost. For lead and most carcinogens this is zero. The gap between MCLG and MCL is the gap between what is safe and what is achievable, and it is where almost all of the interesting information sits.

Level detected. Sometimes a single number, sometimes an average, sometimes a range. Read the footnote, because which one it is changes the meaning entirely.

Range. The lowest and highest individual samples. If the range is wide, the average is close to meaningless for your specific address.

Violation, yes or no. Almost always no. That is a statement about legal compliance, not about health.

What does a lead 90th percentile actually mean?

Lead is reported differently from everything else, and it confuses almost everyone. The figure you see is a 90th percentile: the utility samples a set of higher-risk homes, sorts the results, and reports the value below which 90 percent of them fell. Nine in ten sampled homes were at or below that number, which means one in ten were above it, sometimes far above.

New York City's 2024 report is a clean illustration: the 90th percentile was 10 ppb, but individual samples ranged from non-detect all the way to 107 ppb, and 16 of 335 sites exceeded the action level. A single citywide number hides a lot of variance, because lead comes from the service line and the plumbing in your building rather than from the treatment plant. Your neighbour's result genuinely does not predict yours.

Also worth knowing: the 15 ppb action level is not a health threshold. It is the point at which a utility must take corrective steps. The health goal for lead is zero.

What is an LRAA, and why does it matter?

For disinfection byproducts like TTHMs and HAA5, reports give a locational running annual average: the average at a specific monitoring point over four quarters. Utilities sometimes print an ordinary system-wide average instead, which is a different and usually smaller number. If a report gives you an average without saying it is the highest LRAA, you cannot compare it fairly to another city's LRAA. We hit this exact problem with two large utilities while building our city pages, and left their disinfection figures blank rather than compare unlike numbers.

Three ways these reports mislead without lying

Averages across the whole system. A utility drawing from several plants may serve very different water to different neighbourhoods, then report one blended figure.

Non-detect without a detection limit. "Not detected" only means "not found above the lab's reporting limit." If that limit is high, a meaningful concentration can hide under it. Good reports state the limit; many do not.

Old data presented as current. Lead sampling runs on a multi-year cycle, and some analytes are monitored at reduced frequency. A 2026 report can legitimately contain a 2022 measurement. Always check the sampling year column.

What should you actually check?

Four things, in order. The lead 90th percentile and whether your building predates 1986. The TTHM figure, and whether it is an LRAA. Whether PFAS appears at all, and if it does not, whether that is because it was not detected or because it was never tested. And the sampling years, so you know how old the picture is.