A storage tank roof that sank at a Sinclair Oil refinery earlier this year and caused Wyoming's biggest spill in decades wasn't unprecedented: A storage tank roof also sank at the refinery in 2007.
Leaks in the pontoons holding up the floating roofs occurred before both incidents.
The May 3 spill of 2.73 million gallons of gasoline-grade fluid created a serious explosion risk next to the south-central Wyoming town of Sinclair, population 420. Most of the fluid was recovered soon after the spill, but some continues to be removed from wells at the refinery.
On Jan. 20, 2007, the roof covering a tank of crude oil elsewhere at the refinery sank. As would happen more than two years later, the sinking roof punctured the tank bottom and caused a spill, said refinery manager Mike Bellinger.
The Wyoming Department of Environmental Quality found out about the 2007 incident during an inspection that November. The Associated Press obtained a department report mentioning the incident through a records request.
The 2007 spill was much smaller, involving only a few barrels of oil, Bellinger said. An inspection afterward, he said, revealed that oil had leaked into the roof pontoons through defects in the weld where the pontoons attached to a bulkhead _ a flaw that probably had existed for some time.
"It was probably an original manufacturer defect," Bellinger said.
An inspection six months earlier, he said, hadn't revealed any oil in the pontoons.
That wasn't the case with this year's roof sinking and spill. Some months before the accident, a contractor documented fluid inside two pontoons supporting the tank roof, Bellinger said.
The refinery lost track of the report.
"The records didn't get properly entered into our work order system," he said. "So there was kind of a failure in our management system in the plant and the problem of those two pontoons didn't get addressed quickly enough."
Refinery officials aren't certain that the leaking pontoons by themselves caused this year's incident. Another possibility, Bellinger said, is the tank was filled too high, causing the roof to snag on top of a rail inside the tank.
Either way, the roof sank with enough force to puncture eight or nine holes in the tank bottom, Bellinger said.
The resulting spill may have been Wyoming's largest since a spill at a Casper refinery about 80 years ago, said Bob Breuer, inspection and enforcement manager in the department's Solid and Hazardous Waste Division.
On a 1-to-10 scale, the explosive threat from the Sinclair spill was "right up there, at 8, 9 or 10," he said.
Yet the department apparently wasn't in a position to force Salt Lake City-based Sinclair to inspect its other tank roofs after the 2007 incident. No laws seem to address that possibility, said Glenn Spangler, an air quality engineer for the Department of Environmental Quality.
Spangler looked at the roof that sank in 2007 after the roof and its pontoons were rebuilt and returned to operation. He saw that several hatches to the pontoons were open or ajar and wrote up Sinclair for a violation.
Pontoon hatches are supposed to remain closed except during an inspection of the pontoons.
"If you get liquid in the pontoons because the covers are off, it could have an effect on the roof buoyancy," he said.
Breuer said he's still waiting to receive a formal report from Sinclair on this year's roof collapse. He said the department hasn't ruled out citing Sinclair for the spill.
"We're definitely considering it," he said. "I think it's our job to consider it."