A "black box" can reveal why an airplane crashed or how fast a car was going in the instant before an accident. Yet there are no records of a critical safety test supposedly performed during the fateful hours before the Deepwater Horizon oil rig exploded in the Gulf of Mexico.
They went down with the rig.
While some data were being transmitted to shore for safekeeping right up until the April 20 blast, officials from Transocean, the rig owner, told Congress that the last seven hours of its data are missing and that all written logs were lost in the explosion.
The gap poses a mystery for investigators: What decisions were made — and what warnings might have been ignored? Earlier tests, which suggested that explosive gas was leaking from the mile-deep well, were preserved.
"There is some delay in the replication of our data, so our operational data, our sequence of events ends at 3 o'clock in the afternoon on the 20th," Steven Newman, president and CEO of Transocean Ltd, told a Senate panel. The rig blew up at 10 p.m., killing 11 workers and unleashing a gusher that has spewed millions of gallons of oil into the Gulf.
Leak far worse than thought?
Another uncertainty is just how much oil is spilling. Government officials have put the amount at about 5,000 42-gallon barrels per day. But National Public Radio reported Thursday that a scientific analysis of video of the spill source made available by BP showed that the true figure might be far higher — closer to 70,000 barrels a day.
That would mean far more oil has already spilled into the Gulf than was spilled in the Exxon Valdez tanker accident in Alaska in 1989, when about 250,000 barrels spilled.
NPR cited an analysis by Steve Werely, an associate professor at Purdue University, who used particle image velocimetry. He told NPR that the method was accurate to plus or minus 20 percent, meaning the flow could range between 56,000 barrels and 84,000 barrels day.
NPR also said an analysis by Eugene Chiang, a professor of astrophysics at the University of California, Berkeley, calculated the flow at 20,000 barrels to 100,000 barrels a day.
But The Washington Post said other experts warned against making an estimate using just the video.
Question whether test was done
About the safety test, Houston attorney Tony Buzbee, who represents several rig workers involved in the accident, questioned whether it was even performed.
"I can just tell you that the Halliburton hands were scratching their heads," said Buzbee, whose clients include one of the Halliburton crew members responsible for cementing the well to prepare for moving the drilling rig to another site.
Buzbee said that when Halliburton showed BP PLC and Transocean officials the results of the pressure tests that suggested gas was leaking, the rig workers were put on "standby." BP is the rig operator and leaseholder.
Buzbee said one of his clients told him the "Transocean and BP company people got their heads together," and 40 minutes later gave the green light.
The attorney said the Halliburton crew members were not shown any new test results.
"They said they did their own tests, and they came out OK," he said. "But with the phantom test that Transocean and BP allegedly did, there was no real record or real-time recordation of that test."
Buzbee suggested that BP and Transocean had monetary reasons for ignoring the earlier tests.
"The facts are as they are," he said. "The rig is $500,000 a day. There are bonuses for finishing early."
None of the three companies would comment Thursday on whether any data or test results were purposely not sent to shore, or on exactly who made the final decision to continue the operations that day.
Undersea robots at work
Meanwhile, out in the Gulf, BP settled on its next attempt to cut down on the spill: Undersea robots will try to thread a small tube into the jagged pipe that's leaking on the sea floor. The tube, which will suck crude to a ship on the surface, will be surrounded by a stopper to keep oil from leaking into the water.
BP said it wasn't sure how much of the roughly 210,000 gallons leaking daily would be captured by the improvised device.
If that doesn't work, engineers can still attempt to use a "top hat" box now on the sea floor to cover the leak and siphon the oil to the surface. They also might plug the leak with golf balls and other debris — a "junk shot."
Details of a likely blowout scenario emerged this week for the first time from congressional and administrative hearings. They suggest there were both crew mistakes and equipment breakdowns at key points the day of the explosion.
Rigs like Deepwater Horizon keep a daily drilling report. It is the version of that report given to Congress that cuts off at 3 p.m.
The log confirms that three pressure tests, conducted from the morning to the early afternoon of April 20, indicated unseen underground leakage into the well. But there is no mention of a fourth test that BP and Transocean say was conducted and that they say indicated it was safe to proceed.
In the hours leading up to the explosion, workers finished pumping cement into the exploratory well to bolster and seal it against leaks until a later production phase. After the tests that indicated leakage, workers debated the next step and eventually decided to resume work, for reasons that remain unclear.
At the same time, heavy drilling fluid — or mud — was being pumped out of a pipe rising to the surface from the wellhead, further whittling the well's defenses. It was replaced with lighter seawater in preparation for dropping a final blob of cement into the well as a temporary plug for the pipe.
When underground gas surged up uncontrollably through the well, desperate rig workers tried to cap it with a set of supersized emergency cutoff valves known as a blowout preventer. However, the device was leaking hydraulic fluid and missing at least one battery, and one of its valves had been swapped with a useless testing part.
Problems with the blowout preventer might have been a moot issue had someone not decided to continue offloading the drilling mud.
Newman, the Transocean CEO, told legislators that alarms are monitored on the rig through a vessel management system, or VMS. But he said such records were not transmitted to shore.
"And so the VMS system, along with the logs of the VMS system, would have gone down with the vessel," he said.
"So you have no mirrored backup data device so that that information is recorded at some other location than on the rig itself?" asked an incredulous Rep. Bruce Braley, D-Iowa.
"We do not have real-time off-rig monitoring of what's going on on the vessel," Newman replied.
Part of Halliburton's contract on the Deepwater Horizon was to provide real-time data to officials on shore. That company was able to produce a chart showing events up to two minutes before the explosion. But that document would not be expected to show the key test results.
The chart indicated that shortly before 10 p.m., pressure in the standpipe increased sevenfold to 3,500 pounds per square inch. Halliburton had essentially two minutes' notice that something had gone horribly wrong.
Halliburton monitors temperatures and pressure in offshore wells through sensitive sensors and instruments often capable of transmitting data in real time to officials on the rig and on shore, said Jack Madeley, a consulting safety engineer in College Station, Texas.
"Operators like BP use that information to make sure the well is in the right location," said Madeley, who specializes in forensic investigations of rig accidents. "They need that to make sure the overall procedures for getting it cemented and getting the well secured before they pull the string out and plug up the well are done."
The data help BP and other well operators make crucial decisions about the formations where they are drilling, but it is not always streamed to shore minute by minute, he said.
Rep. Braley said the lack of offsite data storage is something he intends to look into further.
"I'm sure we'll be taking action to follow up with those requirements," he said. "Because it's critical information that would give rise to understanding of what happened and why more wasn't done to shut off the flow of oil and prevent this from happening."
Associated Press writers Jeff Donn in Boston, Chris Kahn in New York, Justin Pritchard in Los Angeles, Garance Burke in Fresno, Calif., and Fred Frommer in Washington also contributed to this report.