How we work
About 46 million Americans get their water from a private well, according to EPA estimates from the 2020 Census, and no federal rule requires anyone to test it. Public agencies have tested a lot of wells over the years, but the results are scattered across government databases most people never see. This site puts them in one place, county by county, and explains what they mean.
Who runs Tested Water
Tested Water is published by Rudderberry, LLC, a small Florida company. We gather public well water data and explain it for people who drink from a private well. We aren't a lab, a government agency or a water treatment company.
Every health claim on the site is checked against the agency or university source named on the page. Our editorial standards explain how we pick sources, check facts and correct mistakes.
Everything on a county page comes from lab results and records that public agencies have already published. We check those numbers, count them the same way for every county, and say plainly what they can and can't tell you. You can read how we count and rate the results, see every source we use, and read how we make money.
Every number comes from a public agency
Each county page names its sources, and so does the list of everything we use.
- Lab tests come from the Water Quality Portal, run by the US Geological Survey and the EPA. It gathers results from federal, state and local agencies. We use tests from 2000 onward.
- How many people use wells comes from EPA estimates built from the 2020 Census.
- Well depths come from state drilling records, gathered in the US Groundwater Well Database.
- PFAS results come from EPA tests of town water systems between 2023 and 2025. We don't rank areas by PFAS tests of private wells, because most of those wells were tested after contamination was already suspected.
- Radon zones come from the EPA's radon map.
- Health information comes from the EPA, the CDC and university extension services, and each page links to them.
How we read the numbers
When a county page says "about 3 in 10 tested wells were over the limit," we mean 3 in 10 of the wells tested there had at least one lab result above the limit. We count wells, not tests, so one bad well sampled 20 times counts once. A result the lab couldn't detect counts as under the limit, unless the lab's detection limit was itself above the limit. Then we leave that result out, because it can't tell us either way.
"The limit" is the EPA's limit for public drinking water. Private wells don't have a federal limit, so we use the same numbers. For manganese we use the EPA's health advisory of 0.3 mg/L, for lead the EPA action level of 15 µg/L, and for radon in water the 4,000 pCi/L level the EPA proposed but never finalized. For iron we use 0.3 mg/L, the level where it starts to stain.
On county pages, "widespread" means 15% or more of tested wells were over the limit. "Worth checking" means 5% to 15%, and "rare" means under 5%. We only rate a contaminant when 10 or more wells in the county were tested for it. On the chart, the filled dot is the county and the hollow ring is the whole state. We never call a county "safe," because two wells next door to each other can test very differently.
Many of the wells in these records were tested for research, or because someone suspected a problem. So the numbers show what turns up in the area's groundwater. They can't tell you what's in your well. Only a lab test of your own water can.
Results we think are overstated
Sometimes one study tested a small part of a county, years ago, because it was looking for a problem. Then the county's number says more about that study than about a typical home well. We check who tested each county's wells, when and where, and mark a result "likely overstated" when several of those signs line up. Those results stay on the chart, hatched, with a note saying why.
Ranks, neighbors and dates
Each result on a county page says where the county falls among the other counties in its state with 10 or more tested wells. A result marked likely overstated is never ranked, and we leave those counties out when we rank the others.
The small chart under each result shows the counties that share a border with this one, from the Census Bureau's county adjacency file. A neighbor appears only if 10 or more of its wells were tested for the same thing.
The bar chart of test dates counts the wells tested each year. A well tested in two decades counts in both. Older results show the water as it was then, so they're a starting point, not a verdict on your well.
If something looks wrong
If you spot a mistake, a broken link or a number that doesn't look right, please email us at hello@testedwater.com. We fix errors and note the date we changed the page.