Radon in water — the rule proposed in 1999 and never finalised

By Andrew Bates, FounderPublished 2026-09-11Updated 2026-09-11

Radon in drinking water has no enforceable limit. EPA proposed one in 1999 — 300 pCi/L, or 4,000 if your state ran an indoor-air programme — and never finalised it. Only 1 to 2 percent of indoor radon comes from water. Thirty certified models exist, from two companies.

There is no limit, and the reason is interesting

Radon appears nowhere in EPA’s table of regulated drinking water contaminants.

No maximum contaminant level, no treatment technique, no enforceable standard. Radium and uranium are both in that table. Radon is not.

→ Radium and uranium: four standards, one certified claim

It came close. On 2 November 1999 EPA published a proposed rule for radon-222 at 64 FR 59246. Twenty-seven years later it has never been finalised.

Where the 300 figure comes from

You will see 300 pCi/L quoted as an action level in a lot of places, including university extension material.

It is EPA’s proposed maximum contaminant level from that 1999 rulemaking. It binds nobody and never has.

Anyone quoting it as a standard is quoting a proposal.

The part nobody mentions: there were two numbers

This is the unusual bit, and it explains the whole design.

Proposed figure	Value	Who it applied to
Maximum contaminant level	300 pCi/L	Systems in states with no indoor-air programme
Alternative maximum contaminant level	4,000 pCi/L	Systems whose state ran an approved programme

More than thirteen times apart. And which one applied depended not on the water at all — it depended on whether your state ran an indoor-air radon mitigation programme.

A system in a state with an approved programme could supply water at 4,000 pCi/L lawfully. The identical water in a neighbouring state would have breached the limit.

Why EPA built it that way

Because the water is not where the radon is.

EPA states most radon enters indoor air “from soil under homes and other buildings”, and that “only approximately 1-2 percent comes from drinking water.”

So EPA framed the rule as “a multimedia approach to reducing radon risks in indoor air (where the problem is greatest), while protecting public health from the highest levels of radon in drinking water.”

What EPA asked water systems to do instead

Fix houses. Genuinely.

A multimedia mitigation plan had to address four criteria:

Public involvement in developing the plan
Quantitative goals for existing homes fixed and new homes built radon-resistant
Strategies for achieving those goals
A plan to track and report results

A water utility, earning the right to a higher limit in its own pipes by reducing radon in its customers’ basements. That is house mitigation, not water treatment, and it is why the proposal reads unlike any other drinking water rule.

And the monitoring that never happened

EPA proposed that every community water system using ground water sample quarterly for a year at each entry point, after treatment and storage. It named high-performance aeration as the Best Available Technology.

None of that took effect. No utility owes you a radon figure, and most have never measured one.

→ What EPA says removes what

The certified field is two companies

Radon has an NSF claim, which is more than several regulated contaminants can say.

NSF listings returned 34 rows for radon reduction under NSF/ANSI 53, resolving to 30 distinct models from just 2 companies. Fifteen counter-top, fifteen plumbed in to a separate tap.

And every one of the 30 holds an identical eleven-claim bundle: radon, lead, mercury, asbestos, MTBE, chlordane, toxaphene, microcystin, PFOA, PFOS and VOC.

So this is one product family, not a market. Nobody sells a radon filter; two companies sell a multi-contaminant block that happens to carry the claim.

→ Which claims exist in which product form

The decision framework

1. Test the air first

That is where EPA puts 98 to 99 percent of the exposure. An air test is cheap and widely available, and mitigation is a builder’s job.

2. Only then consider the water

Water-borne radon contributes to indoor air when it is released by showering and washing, but it is the minority route by EPA’s own figure.

3. Do not treat 300 pCi/L as a threshold

It is a proposal from 1999. Nothing turns on crossing it, and no authority will act on it.

4. If you buy, you are buying the bundle

All 30 certified models carry the same ten other claims. Judge it on lead, PFAS and VOC as much as on radon.

→ The 615 filters certified for lead

5. On a public supply, ask rather than assume

No federal monitoring exists, so any figure your utility holds is voluntary.

→ Testing city water: what your utility already measured

The decision path

Your situation	What decides it	Where that lands you
Worried about radon generally	98 to 99 percent is soil gas	Test indoor air
Air test high	Mitigation is a building fix	Radon contractor, not a filter
Air test fine, worried about water	Water is the minority route	Low priority
On a public supply	No federal monitoring exists	Ask; no figure is owed
On a private well	Nobody monitors it at all	Test if the air result warrants it
Told 300 pCi/L is a limit	It is a 1999 proposal	Treat it as context, not a threshold
Want a certified filter	30 models, two companies	Counter-top or under-sink only
Want whole-house treatment	No point-of-entry claim exists	Nothing certified at that scale

Choose to act if

Your indoor air test came back high. The action that follows is building mitigation, not a water filter.
You already want lead, PFAS and VOC reduction, in which case the certified radon models cover all of those too.

Choose to wait if

You have not tested the air. Testing water first inverts the exposure arithmetic EPA published.

Choose to do nothing if

Your air test is low and you are on a surface-water supply. Radon is a groundwater phenomenon, and the water route is the small one either way.

Mistakes that cost the most money

Treating water when the exposure is air. EPA puts only 1 to 2 percent of indoor radon down to drinking water.

Quoting 300 pCi/L as a standard. It is a proposed figure from a rule that was never finalised.

Expecting your utility to have a number. No federal monitoring requirement exists, so most have never measured it.

Buying a whole-house unit for it. No point-of-entry product holds the radon claim.

Assuming a certified claim implies a regulated risk. Radon has a claim and no limit; several regulated contaminants have a limit and no claim.

Reading the 4,000 figure as a safety level. It was an alternative limit conditional on a state air programme, not a judgement that 4,000 is fine.

Three situations, and what changes

A homeowner who saw radon on a well-water panel. The useful next step is an indoor air test, because that is where the exposure arithmetic points. If the air is fine, the water figure rarely justifies a purchase on its own.

A high indoor air reading in a basement. This is a building problem with a building solution — sub-slab depressurisation and sealing. A water filter addresses at most a couple of percent of the source and does nothing about soil gas entering through the floor.

Someone already shopping for a lead and PFAS filter. The 30 radon-certified models all carry lead, PFOA, PFOS, VOC and seven other claims. If one suits on those grounds, the radon claim arrives free. That is a reasonable way to end up with it, and a poor reason to start.

The objections worth answering

“300 pCi/L is the EPA limit.” It is EPA’s 1999 proposed limit. The rule was never finalised and radon appears nowhere in the regulated contaminant table.

“If it were not dangerous they would not have proposed a rule.” They proposed one because radon is a serious carcinogen. The point of the design was that drinking water is the wrong lever, which is why the alternative limit depended on an air programme.

“My water is treated, so radon is handled.” No federal requirement exists to monitor or treat it, and EPA’s proposed technology, high-performance aeration, is not standard equipment.

“There is no certified filter for unregulated contaminants.” There is one here, from two companies. Certification and regulation are separate systems and neither contains the other.

“A 4,000 pCi/L allowance proves the limit is arbitrary.” It proves EPA valued total risk reduction over a single-medium number. A state fixing houses reduced more radon exposure than the same state treating water to 300.

Who this is for

Households that have seen radon on a water test, or been quoted 300 pCi/L as though it were a standard, and want to know what weight the figure carries.

It is also for anyone deciding where to spend on radon. The answer is almost always the air.

Private wells are unmonitored, but the same order of priority applies.

→ Well water: the six-parameter test panel

What to settle first

Have I tested the indoor air, and do I know that 300 pCi/L is a proposal?

The first decides whether there is a problem worth spending on. The second decides whether the number in front of you means anything.

→ Unregulated contaminants: the candidate list and the monitoring rule → Radium and uranium: four standards, one certified claim → The regulated contaminants with no certified filter → Which claims exist in which product form → How to read your water quality report

The bottom line

Test the air, not the water. EPA states only 1 to 2 percent of indoor radon comes from drinking water, and it designed its 1999 proposal around fixing houses rather than treating supply. That proposal was never finalised, so no utility has a radon figure to meet and none is required to measure one. If water treatment still interests you, the certified field is 30 models from two companies, all point-of-use, and every one of them carries the same ten other claims alongside it.

Who this is not for

This is about radon in drinking water on a public supply. It is not about radon in indoor air, which is where EPA puts almost all of the exposure and which has its own testing and mitigation route entirely. It gives no health assessment of radon, names no product as a recommendation, and does not cover private wells beyond noting they are unmonitored.

Questions

Is radon in drinking water regulated?

No. Radon appears nowhere in EPA's table of regulated drinking water contaminants, so there is no maximum contaminant level and no treatment technique. EPA proposed a regulation in 1999 and it was never finalised. Radium and uranium, by contrast, are both regulated.

Where does the 300 pCi/L figure come from?

It is EPA's proposed maximum contaminant level from the 1999 rulemaking at 64 FR 59246. It was never finalised, so it binds nobody. Sources that quote it as an action level are quoting a proposal, not an enforceable standard.

Why were there two proposed numbers?

EPA proposed 300 pCi/L as the maximum contaminant level and 4,000 pCi/L as an alternative. A system could use the higher figure if its state ran an EPA-approved multimedia mitigation programme for indoor air, or if the system ran its own approved programme.

How much of my radon exposure comes from water?

EPA states most radon enters indoor air from soil under homes and other buildings, and that only approximately 1 to 2 percent comes from drinking water. That is why its proposal was built around indoor air mitigation rather than water treatment alone.

Is there a certified filter for radon in water?

Yes, but the field is very small. NSF listings show 34 rows resolving to 30 distinct models from only two companies, split evenly between counter-top and under-sink units. No pitcher, faucet mount or whole-house product holds the claim.

Does my utility test for radon?

Not under any federal requirement. The monitoring EPA proposed in 1999 — quarterly sampling for a year at each entry point for ground-water systems — was part of the rule that was never finalised, so no utility owes you a radon figure.

What did EPA want water systems to do instead?

Fix houses. A multimedia mitigation plan had to set quantitative goals for existing homes fixed and new homes built radon-resistant, involve the public in the plan, set strategies, and track and report results. That is house mitigation, not water treatment.

Should I test my well water for radon?

Nobody monitors a private well unless the owner pays, so there is no figure otherwise. But the air test comes first on the same reasoning: EPA puts almost all indoor radon exposure down to soil gas rather than to water.

Sources — 10 claims

Federal Register — National Primary Drinking Water Regulations; Radon-222; Proposed Rule, 64 FR 59246 (2 November 1999), RIN 2040-AA94
EPA states most radon enters indoor air from soil under homes and other buildings, and that only approximately 1 to 2 percent comes from drinking water. Its 1999 proposal was framed as a multimedia approach to reducing radon risks in indoor air, where EPA said the problem is greatest, while protecting public health from the highest levels of radon in drinking water. (captured 2026-09-11)
Limits: An exposure share, not a statement that water-borne radon is harmless. It explains why EPA built the proposal around indoor air mitigation rather than around water treatment alone, and why a household worried about radon should test the air first.
EPA — National Primary Drinking Water Regulations, full regulated contaminant table
Radon appears nowhere in EPA's National Primary Drinking Water Regulations table of regulated contaminants on 2026-09-11, so there is no maximum contaminant level, no treatment technique and no enforceable standard for radon in drinking water. The 1999 proposed rule was never finalised. (captured 2026-09-11)
Limits: Established by searching EPA's current regulated contaminant table for radon and finding no entry, with radium and uranium each returning exactly one entry as positive controls, so the search works and the absence is real. EPA separately states the gross alpha standard covers all alphas but does not include radon.
Federal Register — National Primary Drinking Water Regulations; Radon-222; Proposed Rule, 64 FR 59246 (2 November 1999), RIN 2040-AA94
EPA proposed in 1999 a maximum contaminant level for radon-222 of 300 pCi/L and an alternative maximum contaminant level of 4,000 pCi/L. A community water system could comply with the higher figure if its state ran an EPA-approved multimedia mitigation programme for indoor air, or if the system developed its own state-approved programme. The 300 pCi/L figure applied to systems in states without such a programme that chose not to run one. (captured 2026-09-11)
Limits: Proposed figures from 1999 that were never finalised. Neither binds any utility. The structure is unusual and worth recording: two limits differing by a factor of more than thirteen, with which one applied depending on whether the state ran an indoor air programme rather than on the water itself.
Federal Register — National Primary Drinking Water Regulations; Radon-222; Proposed Rule, 64 FR 59246 (2 November 1999), RIN 2040-AA94
Under the 1999 proposal a multimedia mitigation programme plan had to address four criteria: public involvement in developing the plan, quantitative goals for existing homes fixed and new homes built radon-resistant, strategies for achieving those goals, and a plan to track and report results. (captured 2026-09-11)
Limits: A proposed obligation that never took effect. It is recorded because it shows what EPA was asking of water systems: fixing houses rather than treating water, for a contaminant that mostly does not arrive through the tap.
Federal Register — National Primary Drinking Water Regulations; Radon-222; Proposed Rule, 64 FR 59246 (2 November 1999), RIN 2040-AA94
EPA proposed high-performance aeration as the best available technology for radon for all systems, with small system compliance technologies added for systems serving 10,000 people or fewer, and proposed that every community water system using ground water monitor radon quarterly for one year at each entry point to the distribution system, after treatment and storage. (captured 2026-09-11)
Limits: Proposed and never finalised, so no utility monitors radon in drinking water under federal requirement today. Aeration at plant scale has no household equivalent, and the technology EPA named is not the technology NSF certifies against.
NSF Official Listings — Drinking Water Treatment Units, NSF/ANSI 53 radon reduction claim search (listings current as of Thursday, September 10, 2026, 12:15 a.m. Eastern Time)
NSF listings returned 34 rows for radon reduction under NSF/ANSI 53 on 2026-09-10, resolving to 30 distinct models from only 2 companies — 15 counter-top and 15 plumbed in to a separate tap. Every one of the 30 models holds an identical set of 11 claims, so the radon claim exists on a single product family rather than across a market. (captured 2026-09-11)
Limits: A certified claim for a contaminant with no enforceable limit. Certification records tested performance, and says nothing about whether treating water is the useful response when EPA puts only 1 to 2 percent of indoor radon exposure down to drinking water.
Penn State Extension — How to Interpret a Water Analysis Report, by Paul D. Robillard, William E. Sharpe and Bryan R. Swistock (updated 2025-10-13)
The action level for radon in water cited by Penn State Extension is 300 picocuries per litre. (captured 2026-09-09)
Limits: Radon in water is a separate problem from radon in air and is usually treated by aeration or granular activated carbon rather than by a general-purpose filter.
EPA — Radionuclides Rule: A Quick Reference Guide, EPA 816-F-01-003 (June 2001)
EPA states the gross alpha particle standard of 15 pCi/L covers all alphas but does not include radon and uranium, and that a total of 168 individual beta particle and photon emitters may be used to calculate compliance with the 4 mrem per year standard. (captured 2026-09-11)
Limits: Radon in drinking water has no enforceable federal standard at all. EPA published a proposed radon rule, and it remains proposed, so a utility has nothing to comply with on radon.
EPA — Drinking Water Health Advisories for Cyanotoxins
EPA states that under the Safe Drinking Water Act it may publish Health Advisories for contaminants not subject to any national primary drinking water regulation, that these describe concentrations at which adverse health effects are not anticipated over specific exposure durations such as one day, ten days, several years and a lifetime, that they contain a margin of safety, and that they serve as informal technical guidance to assist officials and water system managers. (captured 2026-09-11)
Limits: A health advisory binds nobody. No utility is judged against one, no violation follows from exceeding one, and no notification duty attaches. It is the same class of figure as the manganese health advisory already recorded on this site.
EPA — Contaminant Candidate List 5 (CCL 5), page last updated 2026-02-27
EPA defines the Contaminant Candidate List as contaminants that are not subject to any proposed or promulgated national primary drinking water regulation but are known or anticipated to occur in public water systems, and which may require future regulation. CCL 5, published 14 November 2022, contains 66 chemicals, three chemical groups — PFAS, cyanotoxins and disinfection byproducts — and 12 microbes. (captured 2026-09-11)
Limits: A candidate list, not a regulation and not a warning. Listing means EPA considers a contaminant a priority for regulatory decision-making and information collection. It sets no limit, obliges no utility to do anything, and implies nothing about concentration in any particular supply.