Tank tightness testing was once the ordinary way a UST proved it was sound. Monthly monitoring replaced it for most sites, and the question that remains is whether a tank with a working gauge on it owes a tightness test as well. Usually not. The cases where the answer flips are worth knowing, because they are the ones where somebody is about to spend money either way.
The standard: 0.1 gallons per hour, with five caveats
40 CFR 280.43(c) sets what the test has to achieve. It must be capable of detecting a 0.1 gallon per hour leak rate from any portion of the tank that routinely contains product.
Then it names the things the test has to account for while doing it: thermal expansion or contraction of the product, vapor pockets, tank deformation, evaporation or condensation, and the location of the water table.
That list is the reason a tightness test is a scheduled event with a contractor. A tenth of a gallon an hour is a small enough signal that ordinary physics drowns it. Fuel delivered cold into a warm tank expands. A tank with a slight bow holds a different volume at different levels. Groundwater above the tank bottom changes the pressure across the shell. A method that ignores any of those produces a number, and the number means nothing.
Why the gauge gets a looser number
An automatic tank gauge is held to a different threshold. 280.43(d)(1) requires the automatic product level monitor test to detect a 0.2 gallon per hour leak rate. Twice as coarse as the tightness test.
The reason is frequency. The gauge runs every month, twelve times a year, every year. The tightness test resolved finer and historically ran once every few years. Repetition buys back the sensitivity: a 0.15 gallon per hour loss the gauge cannot resolve in one month tends to declare itself over several, and a slow loss large enough to matter does not stay hidden through a dozen consecutive tests.
Knowing which number belongs to which method is also how you read a contractor's quote honestly. A vendor offering a 0.2 gph method to satisfy a requirement written at 0.1 is offering the wrong test.
The clause that has run out
Tightness testing survives in the rule as an alternative for older tanks, and under the federal rule the window has now closed on all of them.
280.41(a)(1) requires tanks installed on or before April 11, 2016 to be monitored at least every 30 days, with an exception at (a)(1)(i): systems meeting the performance standards in 280.20 or 280.21 and the monthly inventory control requirements of 280.43(a) or (b) could use tank tightness testing at least every 5 years, until 10 years after the tank was installed.
Do the arithmetic: the newest tank it could cover went in on April 11, 2016, so its ten years ran out on April 11, 2026. For tanks installed after April 11, 2016, 280.41(a)(2) requires monitoring at least every 30 days with no tightness testing alternative offered.
So under the federal rule, the honest answer to "do I still need a tank tightness test" is that the monthly method is the requirement and the tightness test is not part of the routine calendar, unless your state adds one.
When it is still the right test
The requirement fading from the periodic calendar moves the test from a scheduled item to a triggered one.
- After a repair. 280.33(d) requires a repaired tank to be tightness tested within 30 days of the repair, unless it is internally inspected, the repaired part is monitored monthly, or the implementing agency approves another test.
- After a suspected release. When an unexplained loss shows up, 280.50 and 280.52 can require a report within 24 hours and a system test within 7 days, and a tightness test is one of the tests 280.52 names. The monthly method that produced the suspicion is not going to resolve it.
- Buying or selling. A precise result on a known date is exactly what a buyer, a lender or an environmental consultant wants, and it is cheap next to the liability it addresses.
- After a gap in monitoring. A console that was out of service for months leaves a hole in the record. A tightness test gives a clean baseline to start again from.
Connecticut's periodic testing table makes the same distinction in its notes: the table covers periodic testing, and says plainly it is not intended to cover initial testing, testing after a repair, or additional testing required after a suspected release or an unexpected loss or gain.
The trap in the word tightness
A tank tightness test covers the tank. Piping runs on its own requirement with its own numbers, in 280.41(b) and 280.44, and a line tightness test is a different test on a different schedule.
Two tests with nearly the same name, one for the vessel and one for the pipe. A record that says only "tightness test, passed, May" has not said which, and that ambiguity is worth removing while the person who ordered the work still remembers.
How FastDragon handles it
Tank tightness sits on the calendar only where the site profile puts it there, which for most sites means it is switched off with the reason recorded. A site relying on monthly monitoring shows the monitoring, and the tightness item shows why it is quiet.
Where a test is triggered, it is logged as a completed record against the tank with its report attached, so the history reads correctly years later when somebody is asked what the tank was doing before a sale. Tank and line tests are separate requirements with separate records, so neither can be filed as the other.
Common questions
I have an automatic tank gauge. Do I still need a tank tightness test?
Not as a routine federal requirement, because the gauge running monthly is the release detection method and the tightness test was the alternative. That alternative has run out: 40 CFR 280.41(a)(1)(i) let systems meeting the performance standards in 280.20 or 280.21, plus monthly inventory control, use tank tightness testing every 5 years until 10 years after the tank was installed. It covered only tanks installed on or before April 11, 2016, so the last window closed on April 11, 2026. Tanks installed after April 11, 2016 are monitored at least every 30 days with no tightness testing alternative. A repair or a suspected release can still require one, and your state can ask for more.
How sensitive does a tank tightness test have to be?
It has to detect a 0.1 gallon per hour leak rate from any portion of the tank that routinely contains product, under 280.43(c). The rule also names five conditions the test has to account for while doing it: thermal expansion or contraction of the product, vapor pockets, tank deformation, evaporation or condensation, and the location of the water table.
Why is the ATG threshold 0.2 and the tightness test 0.1?
Because they run at different frequencies. 280.43(d) sets the automatic product level monitor at 0.2 gallons per hour, and it runs every month. The tightness test resolves twice as fine and historically ran once every few years. A coarser test done twelve times a year and a finer test done occasionally are two ways of covering the same risk.
Does Connecticut require a tank tightness test?
Its periodic testing table carries a tank tightness test column referencing RCSA 22a-449(d)-111(b)(2)(F), and it reads N/A for most tank categories with one exception. Read the table for your own tank category, and remember the table is about periodic testing. Initial testing and testing after a repair fall outside it.
When would I choose a tightness test today?
After a repair or a suspected release, where 280.33(d) and 280.52 can require one; when a tank is being bought or sold; or when a monitoring method has been out of service long enough that you want a clean baseline. Those are situations where somebody wants a precise answer about the tank on a specific date, which is what the test provides.
Sources: 40 CFR 280.41, 280.43, 280.33, 280.50 and 280.52; state programs under 40 CFR 281.12; Connecticut DEEP periodic testing table. Checked October 2026.