This one is not law. No federal section requires you to treat diesel or keep a tank bottom dry, and it belongs on a compliance calendar anyway, because water in a tank is upstream of several failures the regulated items exist to catch. A house practice with a date on it beats a good intention.
How the water gets there
An underground tank is not sealed against water in the way people assume, and there are three ordinary routes.
Condensation. The tank breathes through the vent as product goes out and air comes in. Warm humid air meeting cold steel leaves water on the walls, which runs down into the bottom. A tank kept half full through an autumn has more air, more breathing and more surface to condense on than one kept full.
Ingress. Rain through a spill bucket with a drain valve left open, a fill cap that has not been sealing since somebody cross-threaded it, a cracked riser, a grade-level cover that no longer seats. Every one of those is an item on the monthly inspection, which is not a coincidence.
Delivered in the fuel. Fuel arrives with some dissolved water, and biodiesel blends carry more of it than straight petroleum diesel. Taking a water reading after a drop rather than only before one tells you which loads are bringing it.
What lives at the interface
Water at the bottom of a diesel tank creates a boundary layer, and that layer is where microbial contamination grows. The organisms feed on the hydrocarbon and use the water, so they need both and they sit exactly where the two meet.
What you see first is downstream: filters plugging faster than they used to, fuel darkening, a slimy residue on a filter element. What is happening in the tank is a mat of biomass at the interface and an acidic layer sitting against the steel, which is a corrosion mechanism working from the inside of a tank whose cathodic protection is all about the outside.
Ultra low sulphur diesel and biodiesel blends made this more common than it used to be. Sulphur was doing some incidental antimicrobial work, and removing it removed that.
The reading that matters is the change
Most operators record a water level, look at whether it is high, and move on. The more useful number is the difference from last month.
Half an inch that has sat unchanged for two years is a housekeeping job to schedule. Half an inch that was not there last month is a different conversation, because something changed: a fitting, a cap, a delivery, or the tank. Reading the level as a trend rather than a threshold is what turns a maintenance note into an early warning.
This is also where the housekeeping meets the regulated work. Water arriving faster than condensation explains is one of the things a compromised tank or a failed fitting looks like from the inside, and it is worth investigating before it becomes a suspected release.
Order of operations
Where there is a problem, sequence matters, and getting it backwards is common.
- Remove the water first. Nothing else works while the layer is still there.
- Then treat, if it is warranted. A biocide kills what is growing. It does not remove the dead biomass, which still has to come out and which will plug filters on its way.
- Then find the route in. If water keeps arriving, the treatment is a subscription rather than a fix.
- Then reconsider the interval. A tank that stays dry for a year does not need the same attention as one that does not.
Treating on a fixed schedule while a water level climbs quarter by quarter is paying to manage a symptom, and the invoices make it look like the problem is handled.
Why it earns a place on a compliance calendar
Nothing here is enforceable, so the argument for scheduling it is practical. Water in the bottom of a tank is upstream of microbial growth, internal corrosion, filter failures, dispenser complaints, and readings that drift enough to look like a loss. Every one of those costs more than the check does, and several of them arrive dressed as compliance problems.
A house rule with a date behaves like a requirement without pretending to be one. The distinction is worth keeping visible, so that nobody reading the calendar later assumes a regulator asked for it.
How FastDragon handles it
House practice sits on the same calendar as the regulated items and is labelled as house practice, with no citation, because inventing an authority for it would be worse than leaving it off. It carries its own interval, and switching it off at a site is a normal decision with a reason recorded.
Tank readings and the daily reconciliation live in the same place, so a rising water level and a drifting variance are visible together rather than in two systems that never meet.
Common questions
How does water get into a sealed underground tank?
Three ordinary ways. Condensation on the tank walls as air is drawn in and out with the ullage through the vent, which is worse in a half-empty tank through a cold night. Rain through a spill bucket drain valve, a loose fill cap, or a cracked riser. And in the delivered fuel itself, which is why the tank chart reading after a drop is worth taking.
How much water is too much?
Any measurable water is worth removing, and the number that matters is the change rather than the absolute. An inch that has been there for two years is a maintenance job. An inch that appeared this month is a question about where it came from, and the answer might be a failed fitting or a compromised tank.
What actually grows in diesel?
Microbial contamination lives at the water and fuel interface, feeding on the hydrocarbon and using the water. Biodiesel blends hold more dissolved water than straight petroleum diesel, which gives it more to work with. The result shows up as slime, filter plugging and dark fuel, and in the tank bottom as an acidic layer that can corrode the steel it sits against.
Is this a regulatory requirement?
No, and it should not be sold as one. Checking for water is on Connecticut's monthly inspection form as an item, and the rest of this is housekeeping rather than law. It is on this list because it prevents the failures the regulated items exist to catch.
Does treatment fix a water problem?
No. A biocide kills what is growing; it does not remove the water that let it grow, and the dead biomass still has to come out. Water removal first, then treatment if it is warranted, then finding out how the water got in. Treating on a schedule while ignoring a rising water level is paying to manage a symptom.