Four Months of Shortages on a Tank That Keeps Passing Its Test
A 20,000-gallon single-wall steel tank installed in 1998 is equipped with an ATG that has 0.1 gph monthly monitoring capability and a current annual certification, plus a single-wall flexible product line with a mechanical line leak detector at the dispenser that passes its annual 3 gph test. Monthly throughput is about 60,000 gallons.
Month-end reconciliation for the last four months shows shortages of 610, 780, 910, and 540 gallons. The monitoring vendor's SIR analysis returns 'inconclusive' each month. The ATG's 0.1 gph monthly test passes every month, the tank's water level reads zero, and all sumps are dry. Two of the site's four dispensers were rebuilt six months ago, and the station changed fuel suppliers three months ago. The owner wants to know whether he has a leak and whether he has to call the state.
1. Does this pattern prove a release? Explain what the data support and what alternative explanations must be eliminated before notifying the regulator.
Read the worked answer
Not proven - but not dismissible. The ATG passing a 0.1 gph monthly test equates to roughly 72 gallons per month. A genuine 900-gallon-per-month loss from that tank is about 1.25 gph and would almost certainly fail the 0.1 gph test and trip an alarm. The data therefore point away from the tank shell and toward the measurement chain, or toward product leaving the system somewhere the tank gauge never sees.
The timing is also telling: the two largest variances exceed the investigation threshold of 1% of throughput plus 130 gallons (600 + 130 = 730 gallons at 60,000 gallons per month), and the shortages begin near the dispenser rebuild and the supplier change.
Eliminate, in order:
- Metering - five-gallon test measure against the ATG's dispensed volume at each dispenser; rebuilt dispensers can have mis-set meters or the wrong pulser ratio.
- Bookkeeping - temperature conversion, delivery posting errors, and water.
- Deliveries - tickets compared with pre- and post-drop tank charts; a new supplier is a new variable.
- Theft or unrecorded use.
An SIR result of 'inconclusive' is not a pass; it means the data set cannot resolve 0.1 gph and the owner must keep investigating. If these explanations are eliminated and the variances persist, you have monitoring results indicating a release, and 40 CFR 280.50 requires notifying the implementing agency within 24 hours.
2. What physical tests and inspections would you order, and in what order, to isolate the source?
Read the worked answer
Order matters, because a tightness test run on a line with a leaking test boot or a mis-set LLD gives false assurance.
- Verify the ATG. Manual gauge against ATG reading, probe height and striker plate, water level, and the annual 0.1 gph certification paperwork.
- Verify dispensing. Five-gallon test measure at every dispenser; compare dispensed volume with the ATG's product change; check meter calibration seals on the two rebuilt dispensers first.
- Verify deliveries. Tickets, tank charts immediately before and after each drop, and temperature corrections.
- Walk the containment. Dispenser sumps, tank sump, and spill buckets - any product, staining, or odor; hydrostatic test anything questionable.
- Component tightness tests. Line tightness at 0.1 gph, LLD at 3 gph and the 0.1 gph line test, and a tank tightness test at 0.1 gph if the tank is what remains.
If the piping has an accessible annulus or interstitial sensors, review their history as an independent cross-check before you take a line out of service. Document each step with dates and results - this chain is what demonstrates to the regulator that the investigation was real.
3. If the tank tightness test fails, what are the immediate obligations?
Read the worked answer
A failed tank tightness test is a confirmed release, and the clock starts immediately.
Within 24 hours: notify the implementing agency (and, in most states, the state spill line), the owner, and the insurer. Decide with the agency whether the tank can remain in service under conditions or must be taken out; that is their call, not the technician's.
Initial abatement: remove free product, prevent migration, and control vapors and ignition sources. Under 40 CFR 280.52 the owner submits an initial abatement report within 20 days, and an initial site characterization is generally due within 45 days of release confirmation, covering soil and groundwater along the likely migration path.
Do not repair-and-forget. The investigation must establish the extent of contamination, and the tank must be repaired or replaced to a standard the agency accepts before it can return to service.
Documentation: failed test data, gauge readings and reconciliation history, the time of notification, the names of everyone contacted, and every abatement action. The shortage history will be part of the record - the regulator will want to know why four months of data did not produce a call before the test failed.