Uranium in well water: what your number means, how to read gross alpha, and what removes it

EPA lists the source of uranium in drinking water as "Erosion of natural deposits," and you can't see, smell or taste it. The Vermont Department of Health says testing "is the only way to know" whether uranium and other radioactive elements are in your water. The number most reports compare against is 30 micrograms per liter (30 µg/L), EPA's maximum contaminant level (MCL) for public water systems. Private wells aren't bound by it, and not every state uses it: Vermont's line is 0.020 mg/L, which is 20 µg/L. In the national USGS survey of domestic wells, 1.74 percent were above 30 µg/L.

Two things make uranium harder to read than most contaminants. The main concern is chemical damage to the kidneys, and agencies describe the cancer risk differently. And many labs screen with a gross alpha radioactivity test in picocuries per liter (pCi/L), which you can only compare with the 15 pCi/L limit after subtracting the uranium. Below is how, with the federal conversion factor.

Where uranium in well water comes from

The CDC's toxic-substances agency (ATSDR) says uranium "is naturally present in nearly all rocks, soils, and air," and that "In most areas of the United States, low levels of uranium are found in the drinking water." Higher levels "may be found in areas with elevated levels of naturally occurring uranium in rocks and soil."

Which rocks? The USGS national study of domestic wells (DeSimone 2009) says uranium is "a common trace element in many rock types, but is particularly enriched in certain types of rocks, such as granites." The highest concentrations turned up "most frequently in basin-fill aquifers in the West and in crystalline-rock aquifers in the Northeast and Colorado." Concentrations were low (under 2 µg/L) or not detected in the Southeast's coastal plain and carbonate-rock aquifers and in east-central glacial aquifers. The report names California's Central Valley, northern New Jersey's Newark Basin and New England's crystalline-rock aquifers as high-uranium settings.

Oxygen and carbonate let it dissolve

USGS describes "the conditions that favor uranium solubility—dissolved oxygen and carbonate alkalinity." In its domestic wells, uranium rose with dissolved oxygen and with pH, which the report calls "consistent with the greater solubility of the oxidized form of uranium" and its greater stability at higher pH. It adds that dissolved uranium "can form complexes with carbonate, phosphate, and other species that increase its solubility."

On oxygen, that's the opposite of one of arsenic's release routes. Citing earlier studies, the USGS Circular says arsenic release from iron oxides "is favored by low dissolved oxygen concentrations and high pH" (see the arsenic page). One result doesn't predict the other; test for both.

What our states say about their own geology

• Washington. The state health department (DOH 331-056, revised May 2018) says there is "a potential for elevated levels of uranium in groundwater in Spokane, Pend Oreille, Stevens, Ferry, and Okanogan counties because of naturally occurring uranium in the rock and soil."
• Vermont. Uranium and radium can get into your water "if your well is drilled into or near natural rock formations that contain radioactive elements."
• Connecticut. The Department of Public Health (DPH), as quoted in a 2019 Office of Legislative Research report, says arsenic, uranium and radon "are naturally occurring in some groundwater in CT," that private wells with high levels "have been found sporadically," and that "Levels can fluctuate over time."
• Maine. Maine CDC: "Some Maine wells have too much arsenic, uranium, radon, or other harmful chemicals." See certified labs in Maine.

How common it is: the national numbers

USGS sampled about 2,100 private domestic wells in 48 states between 1991 and 2004. Uranium was measured in 1,725 of them. In most of them (80 percent), uranium was less than a third of the MCL, ranging from less than 1 to 8.03 µg/L. Above the 30 µg/L MCL: 1.74 percent. Gross alpha and radium were measured in a smaller subset, about 17 percent of the wells, located in about half of the aquifers sampled. Here are the radionuclide results against their benchmarks:

The gross alpha figure is uncorrected: USGS didn't subtract uranium, so it runs higher than a regulatory gross alpha number would. The wells above 15 pCi/L were in the New England crystalline-rock aquifers, several sandstone aquifers and the Rio Grande basin-fill aquifer, and USGS notes that "Most of these aquifers also had relatively high concentrations of uranium." And radium needs a caveat. The detailed report gives 2.28 percent of 438 wells above 5 pCi/L; its shorter summary (Circular 1332) gives 0.8 percent, without explaining the gap. Treat the radium rate as approximate.

Uranium rarely travels alone. Counting contaminants above one-tenth of their health benchmarks, USGS found mixtures of uranium with either nitrate or arsenic "in about 20 percent of wells." Above the full benchmarks, those mixtures were rare: 0.29 to 0.58 percent of wells.

State data is thin. The Minnesota Department of Health (MDH) says: "Very few of Minnesota's private wells have been tested for radionuclides."

Health: kidneys first, and agencies describe the cancer risk differently

EPA's standards table lists the health effects of uranium above the MCL as "Increased risk of cancer, kidney toxicity." Its health goal (MCLG) for uranium is zero. Read the agencies' longer explanations, though, and the emphasis shifts toward the kidney.

• EPA's rule. In the 2000 final rule (65 FR 76708), EPA calls uranium "a nephrotoxic metal (kidney toxicant), exerting its toxic effects by chemical action mostly in the proximal tubules," and "a less potent nephrotoxin than the classical nephrotoxic metals such as cadmium, lead, and mercury." In people, the tubular damage "is marked by mild proteinuria" (protein in the urine). Animal studies suggest the cells may regenerate after exposure stops, but "We do not know if uranium-induced proteinuria is an indicator of the beginning of an adverse effect or whether it is a reversible effect."
• ATSDR. "Health effects of natural and depleted uranium are due to chemical effects and not to radiation." On cancer: "Neither the National Toxicology Program (NTP), the International Agency for Research on Cancer (IARC) nor the EPA have classified natural uranium or depleted uranium with respect to carcinogenicity."
• Maine CDC (uranium tip sheet, January 2022): people who drink water with too much uranium "are more likely to have changes to their kidneys that make their kidneys work in the wrong way. These changes go away once you stop drinking water with too much uranium and they have not been linked to kidney disease or damage."
• Vermont: "Over a long period of time and at high levels, radium increases the risk of bone cancer and uranium increases the risk of kidney damage."
• Washington: some people who drink water above 30 µg/L "over many years may have an increased risk of kidney damage or getting cancer."

The difference is in what each agency counts. EPA still counts uranium's radioactivity: in the 2000 rule it estimated that 30 µg/L would typically correspond to about 27 pCi/L, with "a lifetime radiogenic cancer risk of slightly less than one in ten thousand." ATSDR attributes the health effects to chemistry. For a well owner the point is the same either way: the concern is years of drinking the water, and Vermont says there are "no immediate health risks or symptoms" to warn you.

On children, ATSDR says effects "are expected to be similar to the effects seen in adults," and on pregnancy, "We do not know whether uranium can cause birth defects in people." If you're pregnant or mixing formula, ask your doctor or health department.

The limits: 30 µg/L, 15 pCi/L, 5 pCi/L, and Vermont's 20

Federal radionuclide limits are in 40 CFR 141.66, and EPA says its national primary standards "apply to public water systems." None of them is legally binding on a private well.

Why 30 and not 20

EPA's 2000 rule says 20 µg/L was its best estimate of the "drinking water equivalent level," the concentration matching its safe daily dose for the kidney, and that it "believes the feasible level is still 20" µg/L. The agency set 30 anyway, using its authority "to set an MCL at a level higher than feasible," because it found the benefits at 20 didn't justify the costs. EPA said the difference in kidney risk between 20 and 30 µg/L "is insignificant" and that 30 is "protective of kidney toxicity and cancer with an adequate margin of safety." Vermont uses 20. Its calculator notes: "Uranium MCL is the Vermont MCL. Federal (EPA) MCL is 0.030 mg/L."

One more unit trap. MDH lists EPA's standard as "20 pCi/L of uranium or less," an activity unit, without saying where 20 comes from. EPA's MCL itself is in µg/L. At the rule's 0.67 pCi/µg factor (next section), 30 µg/L is 20.1 pCi/L. Check which unit your report uses.

Gross alpha: the screening test, and why uranium gets subtracted

USGS calls gross alpha "a screening measurement that is used to determine high levels of alpha-particle radioactivity without identifying the specific radionuclides responsible." In groundwater, USGS says, it comes in most cases from natural uranium-238, uranium-234, radium-226, radium-224, radon-222 and other isotopes.

A low gross alpha result can rule things out. For public systems, federal rules (40 CFR 141.26(a)(5)) allow a gross alpha result to stand in for the radium-226 measurement if it "does not exceed 5 pCi/l," and for the uranium measurement if it "does not exceed 15 pCi/l." Vermont's private-well version: "Results under 5 picoCuries per liter (pCi/L) are not considered a health risk." Above 5, "further calculations or testing is needed."

The subtraction

The 15 pCi/L limit excludes uranium, which has its own limit by mass, so you take out the part of a raw result that came from uranium. EPA's 2024 training for state regulators (Radionuclides Rule Training) lays it out: convert the uranium mass to activity, subtract it from gross alpha, and use the "net" result. Vermont calls this net figure adjusted gross alpha (AGA).

The conversion factor is fixed in rule. 40 CFR 141.25, footnote 12: "If uranium (U) is determined by mass, a 0.67 pCi/μg of uranium conversion factor must be used," based on "the 1:1 activity ratio of U-234 and U-238 that is characteristic of naturally occurring uranium." Vermont's spreadsheet calculator uses the same factor:

The factor is a convention, not a measurement of your water. EPA's 2000 rule says the ratios actually observed in drinking water "range from 0.67 up to 1.5 pCi/µg," with a best-estimate average of 0.9. EPA's 2024 training says the 0.67 factor "is based on a conservative assumption and is not intended to be accurate, but instead to provide a conservative screening-level value." Our reading: using the low end of that range subtracts the least uranium, which keeps the adjusted number on the high side.

Radon is excluded from the 15 pCi/L limit too, and in the USGS study "almost all radon is removed from the water sample during the preparation steps for sample analysis." A gross alpha result says little about radon.

Results often arrive in pieces. At the Vermont Health Department lab, gross alpha comes back "within 2 weeks" and uranium, in the inorganic chemicals kit, in 2 to 4 weeks. Vermont says to wait for the uranium result before judging the gross alpha risk.

How often to test: each state's own advice

States schedule uranium testing very differently. Here is what five of the states this site covers say:

Maine wants a repeat every three to five years, Vermont every five, Connecticut ideally every five, and Washington one test in five counties. Since Connecticut says levels "can fluctuate over time," our suggestion (not a rule) on granite or crystalline bedrock with a first result well above zero is Maine's three-to-five-year interval.

Ordering the test

Ask for uranium by name, reported in µg/L, and add gross alpha if the lab offers it: that's the screen that points to radium. Find a state-certified lab with the uranium lab finder (or certified labs in Vermont), and see which tests your well needs for the rest of the panel. Get the sample container and instructions from the lab. Sample an untreated tap to learn what's in the well; the guide to testing well water covers collection. If you already treat, test raw and treated water. Vermont says to "Re-test your drinking water for gross alpha radiation and/or uranium after any treatment system is installed."

Reading your result: a decision table

Convert first: mg/L × 1,000 = µg/L, and 1 µg/L is 1 part per billion (ppb). Every threshold below comes from the agency named. You can also check numbers with the results checker.

If your report also has gross alpha, work out the adjusted figure (formula above) and use Vermont's tiers:

Vermont adds a point worth knowing even below the lines: "Because gross alpha radiation causes cancer, any exposure to it will increase your risk of getting cancer," so you may choose to treat "even if it is under the drinking water standard."

Treatment: uranium and radium carry opposite charges

Vermont explains that gross alpha emitters enter groundwater "as positive ions called cations (for example, radium 226 and 228), negative ions called anions (for example, uranium), or as radiation with no charge." Vermont says a softener, which swaps positive ions, "will not lower the levels of uranium," and anion exchange, which swaps negative ions, "will not lower the level of radium."

For public systems, EPA's best available technologies for uranium are "Ion exchange, reverse osmosis, lime softening, coagulation/filtration," and for radium "Ion exchange, reverse osmosis, lime softening" (40 CFR 141.66(g)). For households, state health departments describe the three in the table. MDH puts it generally: "ion exchange water softeners and reverse osmosis systems, can reduce the levels of some radionuclides."

Whole house or one tap?

Maine CDC describes both. A point-of-use system under the kitchen sink "is good if you are worried about drinking and cooking." A point-of-entry system treats all the water in the house: "You may want this kind of system if there are young children in your house." Since Maine says bathing and dishwashing are fine even above 130 µg/L, a drinking-water tap is enough for many households. Vermont adds that other problems may need fixing first: water "may need to be treated for other water quality issues — such as hardness, iron, manganese and pH level — before it's treated for radium or uranium."

Ion exchange makes a waste stream. EPA's footnote for small systems warns that "The regeneration solution contains high concentrations of the contaminant ions." Ask the installer where it goes.

Check what the certification actually covers

Vermont points to NSF/ANSI 58 for RO. On NSF's public listings, many Standard 58 systems and Standard 44 softeners carry a "Radium 226/228 Reduction" claim. When we searched the Standard 58, 53 and 44 listings on November 1, 2026, we found no uranium reduction claim. That doesn't mean RO fails on uranium (Vermont and Washington both recommend it). It means the certificate alone doesn't prove uranium removal. Ask the seller for uranium test data, and test the treated tap after installation.

Uranium and radon: same rock, different test

Radon "is part of the decay series of uranium," USGS notes, and its levels in groundwater are controlled "in large part, by the presence of uranium-bearing rocks." In the domestic-well study, radon and uranium rose together in crystalline-rock and sandstone aquifers, tied to "uranium-bearing minerals in granites, pegmatites, and their derivative metamorphic rocks." A high uranium result is a good reason to check radon; Vermont recommends testing indoor air for it.

The exposures differ. Vermont says alpha radiation "does not pass through the skin" but "can be harmful if you eat, drink or breathe in" something containing it; for uranium in a well, that means drinking and cooking. Radon is a gas: Washington says it escapes into indoor air during showering and washing, and that radon entering through soil "is usually a much larger risk than radon entering through the water supply." See the radon page.

Worked example: a Vermont-style report

Illustrative numbers. Your report from an untreated tap shows Gross alpha: 14 pCi/L and Uranium: 0.025 mg/L.

1. Units. 0.025 mg/L × 1,000 = 25 µg/L. That's under EPA's 30 and under Maine's 30 line, but above Vermont's 20.
2. Adjust the gross alpha. 0.67 × 25 = 16.75 pCi/L from uranium, so adjusted gross alpha = 14 − 16.75 = −2.75 pCi/L. A small negative number happens with the conservative factor; Vermont's calculator only flags results at or below −10 for a recheck.
3. Read the tables. Adjusted gross alpha is 5 or less; uranium is above 0.02 mg/L. That's the "above 20, under 30" row. Vermont's calculator gives this exact case: "Consider treatment of uranium (e.g. with reverse osmosis). Retest again in five years."
4. Pick a technology. A kitchen RO unit covers drinking and cooking. A whole-house anion exchange system also fits Vermont's description: it's "most effective if the adjusted gross alpha is less than 15 pCi/L and uranium is above 0.02 mg/L." A water softener would not help, because it doesn't remove uranium.
5. Verify. Test the treated tap after installation, then raw and treated water every five years. In Maine the same well would be under the 30 µg/L line, with a retest every three to five years.

The reverse case shows why the subtraction matters. Say the report reads Gross alpha: 22 pCi/L and Uranium: 0.008 mg/L (8 µg/L). Uranium is below every line on this page. Adjusted gross alpha = 22 − (0.67 × 8) = 22 − 5.36 = 16.64 pCi/L, which is above 15. Vermont says an adjusted figure above 5 "is likely caused by radium" and recommends treating above 15. The fix is the opposite one: a softener (best, Vermont says, when adjusted gross alpha is above 5 and uranium under 0.01 mg/L), or RO, or a radium-226/228 test first.

Next step

If your well has never been tested for uranium, order the test, especially on granite or other crystalline bedrock or in northeastern Washington. Connecticut asks for at least one test (and requires one for new wells since October 1, 2022), Maine for one every three to five years, Vermont for gross alpha and uranium every five. Keep every report. Find a certified lab with the uranium lab finder.

Questions people ask