Pick a number: 12,600 barrels . . . 20,000 . . . 21,500 . . . 25,000 . . . 30,000 . . . 40,000 . . . 50,000. Scientists put every one of those numbers in play Thursday as they struggled to come up with a solid estimate of how much oil is gushing each day from the black geyser at the bottom of the Gulf of Mexico.

The one scientific certainty: It's a lot -- and more than some of the same scientists thought just a couple of weeks ago. It's so much that the crews trying to siphon it to the surface are going to need a bigger boat.

Early in the crisis, BP and the federal government repeatedly said that the Deepwater Horizon well was spewing about 5,000 barrels (210,000 gallons) a day into the gulf. But the new estimates, released Thursday by government-appointed scientists, show that the well most likely produces 5,000 barrels before breakfast.

One team that has studied video taken of the leaking riser pipe before it was cut and capped last week has concluded that the well was most likely producing 25,000 to 30,000 barrels a day. If that estimate is on target, and if the flow has been more or less consistent since the April 20 blowout, the hydrocarbon reservoir 2 1/2 miles below the sea floor has gushed five to six times the amount spilled in Alaskan waters in 1989 by the Exxon Valdez.

Put it another way: It's roughly one Valdez spill every week. Nearly two Olympic-size swimming pools of oil every day.

Another scientific team, which analyzed satellite images, has come up with a somewhat more modest estimate of 12,600 to 21,500 barrels a day, just a slight uptick from the team's earlier finding.

Or the flow could be much higher still: A team led by scientists from the Woods Hole Oceanographic Institution has studied the leak with instruments normally used in research on deep-sea hydrothermal vents. Its initial estimate puts the flow at 25,000 to 50,000 barrels a day, said U.S. Geological Survey Director Marcia McNutt, who leads the teams of scientists collectively known as the Flow Rate Technical Group.

"These numbers are all over the board," she said in a conference call with reporters.

In response to a question, McNutt said that 20,000 to 40,000 barrels is the most plausible range, but she emphasized that the findings are preliminary and that the techniques have inherent limitations.

The flow rate is significant on several fronts. First, it gives the government and BP a sense of how much capacity they'll need among surface ships to handle all the oil gushing out of the well and up a pipe to the Discoverer Enterprise drill ship, which is capable of processing about 18,000 barrels a day. Other ships are being added to the effort.
ad_icon

Second, the fines that BP faces for polluting the gulf will be tied to how much oil has leaked.

Third, the higher figures call into question the circumstances that led to the much lower estimates of the spill early in the crisis. On April 28, after having received estimates of the size of the spill from BP and the National Oceanic and Atmospheric Administration, the Coast Guard announced the 5,000-barrels-a-day estimate. Not until May 27 did the flow rate group increase the estimate to 12,000 to 25,000 barrels.

Coast Guard Adm. Thad W. Allen, the national incident commander, rejected the suggestion this week that the government has ever low-balled the flow rate.

"I can guarantee you unequivocally, nobody is low-balling anything that works for me," he told reporters. "And I will never low-ball anything."

An internal BP document marked "confidential" shows that the company on April 27 had arrived at three estimates for the size of the spill: low, best guess and high. The calculations were based on satellite imagery of the slick. The low estimate was 1,063 barrels a day. The best guess was 5,758 barrels. The high estimate was 14,266 barrels.

The next day, the Coast Guard announced the 5,000-barrel estimate.

BP spokesman Andrew Gowers said Thursday by e-mail: "There is no secret about this. It was our contribution to the Unified Command estimate that was published shortly afterwards. We have always said we made a contribution to that estimate but that it was an estimate by the Unified Command."

This remains an inexact science, fraught with guesswork, with the well constantly burping gas amid the dark crude, and water and sediment thrown in to make it all the more complicated. The geyser has been a moving target: The flow may have increased significantly since the failed "top kill" maneuver and the shearing of the riser pipe turned a complex system of leaks into a single, coherent plume. Most of the research findings released Thursday were based on study of the leak prior to the riser cut.

Once the riser was cut, said Steve Wereley at Purdue University, "the flow must go up," because the oil is no longer suppressed by a kink in the pipe. The government has estimated that the riser cut could have increased the flow by 20 percent.

The lowest estimate of 12,600 barrels is clearly no longer plausible, because 15,800 barrels were siphoned Wednesday to the surface and much more oil is still billowing into the gulf from around the cap.

More numbers are coming. The plume team and other members of the flow rate group are preparing another estimate, this one based in part on high-definition video taken by submersibles after the riser was cut and provided to the scientists on Tuesday. "The scientists are actually drowning in high-resolution video," McNutt said.