Your UST Records Do Not All Expire on the Same Day
Ask three underground storage tank operators how long records have to be kept and you will hear the same number three times: three years. It is the figure everyone remembers. It is wrong more often than it is right.
The federal rule does not have one retention period. It has at least eight, on different clocks. Some are counted rather than dated, so you keep a set number of results and not a span of time. One runs until the next test replaces it. Two never end.
August is when that starts to matter. Half the compliance year is behind you, every clock has been running since January, and nobody has checked them.
Eight Clocks, Not One
Here is where the confusion starts. 40 CFR 280.34 lists the nine records you must keep and sets no retention period at all. Every clock lives in the section that created the record.
Sorted by how long they live:
Twelve months. Release detection results for every tank (40 CFR 280.45(b)) and your walkthrough records (280.36(b)).
Three years. Spill, overfill and containment sump testing (280.35(c)(1)), and the annual release detection operation test (280.45(b)(1)).
Five years from installation. Written performance claims for each release detection method (280.45(a)).
One year after the work. Calibration, maintenance and repair of on-site release detection equipment (280.45(c)).
Until the next test. Your tank tightness result (280.45(b)(2)). Not three years. Until it is replaced.
Counted, not dated. The last two cathodic protection tests (280.31(d)(2)) and the last three 60-day rectifier inspections (280.31(d)(1)). One recent test does not satisfy a rule asking for two.
Until the system closes. Repair records for any component (280.33(g)). They never age out.
For as long as it applies. Compatibility documentation while you store that substance (280.32(c)). Your operator list while those operators are designated (280.245).
An operator on a blanket three-year policy gets it wrong twice. They keep monthly release detection sheets four times longer than asked, and they shred tightness results and repair records that were meant to stay. EPA sets the same list out on page 33 of Musts for USTs.
UST Compliance Tools
Mid-Year UST Records Walk
Work through the records one at a time. Each row tells you what to pull, how far back it has to go, and what it means if you cannot produce it. Mark Gap on a record, or Found it on the walk, and the consequence appears. Nothing you enter leaves your browser, and the finished walk prints clean.
You have answered 0 of 32 records.
The thickest file and the one inspectors open first.
-
One passing release detection result for every tank, for each of the last 12 months.
What this gap means: A missing month is a month you cannot show the tank was monitored. It is the single most cited UST records finding.
-
Release detection records for pressurized piping: the automatic line leak detector plus either the annual line tightness test or monthly monitoring.
What this gap means: Piping is a separate obligation from the tank. Operators who have twelve clean tank months and no piping records are half covered.
-
The annual operation test of your release detection equipment, listing each component tested and any action taken.
What this gap means: This test has been required since October 13, 2018 and is the most missed record on a walk. It covers gauges and controllers, probes and sensors, line leak detectors, and hand-held sampling gear.
-
Your most recent tank tightness test, if you use tightness testing as a method.
What this gap means: Tank tightness results are kept until the next test replaces them, not for a fixed number of years. Throwing one away on a three-year rule leaves you with nothing.
-
The written performance claims for each release detection method you use, and how they were justified.
What this gap means: Without this the method itself is unproven on paper, no matter how many clean monthly results sit behind it.
-
Calibration, maintenance and repair records for on-site release detection equipment.
What this gap means: A repair with no record reads as equipment that was never fixed.
Three-year clocks that are easy to lose track of between visits.
-
Your last spill prevention equipment test, on the three-year cycle.
What this gap means: If the last test is more than three years old, every spill bucket on the site is out of compliance today.
-
Your last overfill prevention equipment inspection, confirming it still activates at the right level.
What this gap means: There is no double-wall alternative for overfill equipment. It gets inspected every three years, always.
-
Your last containment sump test, for sumps used for interstitial monitoring of piping.
What this gap means: The three-year tightness test and the annual visual sump check are different obligations. Passing one does not satisfy the other.
-
For double-walled equipment on the monitoring alternative: documentation that both walls are being monitored.
What this gap means: The alternative only holds while the monitoring runs. If it stops, a test is due within 30 days.
Counted records, not dated ones. Keep a set number, not a span of years.
-
The results of your last two cathodic protection tests, run by a qualified tester on the three-year cycle.
What this gap means: Two is the number. One test on file, however recent, does not meet the record requirement.
-
The results of your last three 60-day inspections of the impressed current rectifier.
What this gap means: Only for impressed current systems. A rectifier that has been off between inspections means the tank was unprotected for that stretch.
The record that proves someone actually looked.
-
Every-30-day walkthrough records covering spill prevention equipment and release detection equipment.
What this gap means: Spill buckets are a 30-day item, not an annual one. Twelve entries a year is the baseline, and each must say what was checked, whether it was acceptable, and what was done about anything that was not.
-
Annual walkthrough records covering every containment sump and your hand-held release detection gear.
What this gap means: The annual sump check applies to all sumps, whatever release detection method the site uses.
-
Delivery records, if you check spill prevention equipment before each delivery because deliveries come less often than every 30 days.
What this gap means: The per-delivery option only works if the delivery records exist to anchor it.
A list, plus proof behind every name on it.
-
A current list of your designated Class A, Class B and Class C operators, with dates trained and dates they assumed duties.
What this gap means: The list must be current, not historical. A Class C operator who left in March should not still be your named operator in August.
-
Training verification for every name on that list, showing who trained them, when, and the trainer's company, address and phone.
What this gap means: A name on a list with no record behind it is an undocumented operator.
-
Retraining records, if an agency has found the facility out of compliance since your last walk.
What this gap means: Class A and B retraining is due within 30 days of that finding unless you run annual refresher training instead.
Records with no expiry date at all. They outlive the three-year clocks.
-
Repair records for every UST system component.
What this gap means: Repair records are kept until the system is permanently closed or changed in service. They never age out.
-
Compatibility documentation for the substance you actually store today.
What this gap means: Switching to a fuel above 10 percent ethanol or 20 percent biodiesel needs 30 days notice to your agency and a compatibility demonstration first.
-
Your notification and registration form, and a change-of-ownership form if the site changed hands.
What this gap means: New ownership carries a 30-day notification of its own. An acquisition is the classic place this gets missed.
-
The corrosion expert's site analysis, where your system relies on one.
What this gap means: One of the nine records the rule names outright.
-
The whole set above, produced at the site or from one readily available alternative location, on request.
What this gap means: Records that exist but cannot be found quickly fail the test that matters. This is the row most operators get wrong.
None of these leave a trace in the file drawer. Every one is on the federal walkthrough list.
-
Liquid or debris sitting in a spill bucket.
What this finding means: A bucket holding water will not contain the next spill. A bucket that is always bone dry is worth a look too: EPA warns it can mean the bucket is not liquid-tight.
-
An obstruction in the fill pipe.
What this finding means: Found only by looking. It shows up nowhere in your monthly results.
-
A fill cap that is not securely seated on the fill pipe.
What this finding means: The cheapest finding on this list and one of the most common.
-
Visible damage to spill prevention equipment.
What this finding means: A cracked bucket can still pass a three-year test done before the crack appeared.
-
The release detection system running with an alarm or an unusual operating condition on screen.
What this finding means: An alarm nobody acknowledged is worse than a failed test. The equipment did its job and the site did not.
-
Release detection testing records that are not current when you look at them.
What this finding means: Reviewing the records is itself part of the 30-day check, not a separate task.
-
Damage, leaks or standing liquid in a containment sump.
What this finding means: Liquid in a sump is either a leak in or a leak out. EPA also warns that a sensor sitting in water may not alarm at all, because some sensors respond only to petroleum.
-
Gauge sticks and groundwater bailers that are no longer serviceable.
What this finding means: A warped stick produces readings that look valid and are not.
-
A double-walled spill bucket or sump with a leak in the interstitial space.
What this finding means: The interstitial check is part of the walkthrough for double-walled equipment, not a separate service call.
Your gap list
Everything you flagged, records and findings together. Each one is a question an inspector can ask on any ordinary Tuesday.
Start by Counting
The first pass takes an afternoon and needs no expertise. Pull the release detection file and count. January to now, that is one passing result per tank per month, with a walkthrough entry for each of those months.
There is a published bar to count against. EPA's guide for inspectors tells them an owner needs records for the two most recent consecutive months, and for 10 of the last 12. Two missing months in a row fails even when the year looks healthy.
Then count the piping, which is a separate obligation and the one most often missed. Twelve clean tank months with no piping records is half a file.
You are not judging whether the results were good. You are finding the months that are not there.
The Record Almost Nobody Has
Since October 13, 2018, release detection equipment has needed an annual operation test under 40 CFR 280.40(a)(3). It covers tank gauges and controllers, probes and sensors, automatic line leak detectors, vacuum pumps and pressure gauges, and hand-held sampling gear. Results are a three-year record under 280.45(b)(1).
This is the most missed item on a mid-year walk. A site with a spotless monthly file often has nothing here at all. The monthly result proves the equipment reported something. It does not prove anyone confirmed it still reports correctly.
"Available" Is a Requirement
40 CFR 280.34(c) says records must be at the UST site and immediately available for inspection. The alternative is a readily available second location, with the records provided on request. Both halves of that are conditions.
A box at a regional office, behind someone on vacation, does not meet it. Neither does a file split across a shared drive, three email threads and a cabinet at the store. If your mid-year walk turns into a scavenger hunt, you have already found the finding.
It shows up nationally. EPA's mid-year FY2026 measures put walkthrough compliance at 83.2%, and EPA's definition of that measure covers record retention as well as the walk. Roughly one facility in six falls short on one or the other.
Then Walk the Site
Everything above is paper. Now go outside, because one whole class of problem leaves no trace in any file.
You do not have to invent that list. The federal walkthrough rule is EPA's own answer to what paper cannot show you. Under 40 CFR 280.36, every 30 days someone checks the spill prevention equipment for damage. They clear liquid and debris from the bucket, clear the fill pipe, and confirm the fill cap is seated. The same check confirms the release detection system is running clean, with no alarms and nothing unusual on screen. Once a year the list adds every containment sump and the hand-held gear.
Read those as findings rather than chores and the point lands. A spill bucket half full of water passes any records review ever written. A cracked bucket sits comfortably inside a three-year test done before the crack appeared. An alarm nobody acknowledged is worse than a failed test, because the equipment did its job and the site did not.
Carry one warning outside with you, from EPA's sumps and spill buckets manual. A sensor sitting in a sump full of water may not alarm at all. Some sensors respond only to petroleum.
Note the interval too. Spill buckets are a 30-day item, not an annual one. Only sites taking deliveries less often than that may check before each delivery instead.
The Federal Rule Is a Floor
Every figure here is the federal minimum. States running their own UST programs must be at least as strict, and 40 CFR 281.12(a)(3) lets them be stricter. Many are.
California makes this article's argument for us. Under 23 CCR § 2631, a certified designated operator inspects each site every 30 days. That inspection must review the alarm history, the testing and maintenance records, and employee training records before it is signed off. California has written the records walk into the physical walk. Maine comes at it from the other side, with a weekly documented walkthrough by the Class A or B operator.
We covered the testing-cycle divergence in Annual or Triennial? The Answer Depends on the Test and the State. Check your own agency's rule before trusting any date here.
The Bottom Line
Run two audits at mid-year, not one. The file drawer proves what happened. The walk shows what is happening. Each catches what the other structurally cannot.
The work is not hard. It is counting, then looking. What makes it hard in December is that nobody did it in August.
Stop tracking eight clocks from memory.
PASS Harmonics holds release detection results, testing cycles, walkthrough records, operator designations and permit dates in one place, each on the retention clock the rule actually gives it. When the mid-year count comes around, a missing month is already visible instead of waiting to be found.
Working out what your states require on top of the federal floor? Ask us.
More from this month
Get these in your inbox: subscribe to PASSing Notes, our monthly newsletter.
