The findings suggests that oil from the spill could continue to emerge months if not years from now, and hundreds if not thousands of miles away.

And the study also provides yet more evidence that the initial official spill estimates were off by at least an order of magnitude.

BP on Thursday finally abandoned its 5,000 barrel (or 210,000 gallons) a day estimate, after finding that a tube inserted into a leaking pipe over the weekend and capturing only a fraction of the spill was itself capturing 5,000 barrels a day -- along with 15 million cubic feet of natural gas.

The National Oceanic and Atmospheric Administration, amazingly enough, appears to be sticking to its own 5,000 barrel a day estimate, which was initially based on the size of the oil slick. But if only a tiny fraction of the spill is actually visible on the surface, then that estimate is obviously very badly off.

McClatchy Newspapers reported Thursday night that BP's low-ball estimate, "which the Obama administration hasn't disputed, could save the company millions of dollars in damages when the financial impact of the spill is resolved in court, legal experts say."

Ten years before BP's well blew up and started disgorging oil and gas, the Department of the Interior's Mineral Management Service, along with 23 oil companies and the Norwegian government conducted a test deep under the Norwegian Sea, releasing nearly 16,000 gallons of diesel oil and then carefully watching what happened to it. (See the the Powerpoint presentation of the test, which was first brought to my attention by seminal.firedoglake.com.)

Only some of that diesel was ever accounted for -- somewhere in the range between 250 and 5,000 gallons. The rest presumably either evaporated, dissolved away -- or, in the form of smaller droplets, was carried far away from the observers. Those droplets "would have been carried much further by residual plume effects, and then would have risen to the surface much more slowly," the study found.

Eric Adams, an environmental engineer at MIT, wrote the final report on the study in 2004.

The controlled release was just over half as deep as the Deepwater Horizon spill, and was, relatively speaking, tiny. Yet the lessons were clear, Adams told HuffPost.

"Not very much of it was recovered at the surface," Adams said. "It's probable that a lot of it did ultimately get to the surface, it just got to the surface so far away it was never accounted for."

As a result, Adams said, "I think you should be prepared for more oil to surface over time."

Adams said he was surprised that federal officials weren't more prepared to deal with a deep-sea leak and its consequences, given how much was known ahead of time. Officials should have been aware that oil released so far below the surface would quickly spread out and become unrecoverable unless they did something about it.

"I would have tried to corral it, I guess," Adams said. "Knowing how ill-behaved the oil could be in an ocean that is not quiescent, I'm really chagrined that their efforts to contain it didn't work."

The initial containment effort amounted to an ill-fated attempt to drop a lid on the well, three weeks after the initial explosion.

"I would have proposed at least careful consideration of some sort of a flexible shroud, or a shower curtain, some sort of flexible device that could be anchored above the leak, to form like a chimney to bring the oil up," Adams said.

"Basically what that's doing is preventing the oil from scattering. It would bring it up into a relatively confined area on the surface, and it would be thick enough it could be easily contained with booms and sucked up into tankers," he said.

"It wouldn't be spreading all over the place."

Huffington Post