WASHINGTON - Instrument data from Comair Flight 3272 that crashed south of Detroit in January indicates the plane stalled on its descent in a snowstorm and then spun hopelessly out of control.
While some aviation experts said Wednesday that pilot error may have been a factor in the stall, others said it was too early in the investigation to determine a cause for the worst crash in Comair history.
Stalling does not mean the engines stopped operating, but that the plane slowed so much its wings lost lift needed to keep the craft in the air.
Transcripts of cockpit voice recordings released to the public Wednesday captured little of the crew's thoughts during the last few seconds of the Jan. 9 crash that killed the three crew members and 26 passengers.
About 30 seconds after the Embraer EMB-12ORT began to pitch and roll, it slammed nose-first into a snow-covered field 18 miles from Detroit Metro Airport. The plane had departed at 2:51 p.m. from Cincinnati-Northern Kentucky International Airport.
The cockpit microphone picked up several "unidentified thumps" and "whirring noises" and then a surge in power.
In the final minute, registered as 3:54 p.m., a series of warnings went off. The pilot and co-pilot responded with "oh" and "oh, (expletive deleted by the NTSB)." Fourteen seconds later, after more warning sounds, both cockpit crew members shouted and the tape ended.
The NTSB did not reveal what the pilots said as their final words.
The data recorder indicates a relatively normal descent until 3:54 p.m., when the plane - on automatic pilot - dropped to 4,000 feet in preparation for landing. At 3:54 p.m. and eight seconds, the plane stalled and began its fatal roll.
Chuck Eastlake, a nationally known aviation expert and an aeronautical engineering professor at Embry-Riddle Aeronautical University in Daytona Beach, Fla., said the data and voice recorder information show problems began during that crucial descent in altitude.
"It's a set of flight conditions that is normal but doesn't leave (the pilot) a lot of room for error," Mr. Eastlake said.
He said the NTSB records show the pilot did not appear to increase power immediately to the engines once reaching 4,000 feet. Power would have been needed quickly to keep the plane from rolling.
Mr. Eastlake said the data showed a "fairly long delay."
Though he said it has yet to be determined whether pilot error caused the crash, he said the lack of power to the engines "strikes me as untypically slow response."
Once the plane began to roll, it was at such a low altitude that correcting the roll became impossible.
Meghan Glynn, Cincinnati Comair spokeswoman, said such a theory is not supported by the documents NTSB released so far.
"Any theory out there right now is speculation and is inconclusive," she said. "We continue to cooperate with NTSB investigators."
Michael Barr, director of aviation and surface transportation safety programs at the University of Southern California, was more cautious than Mr. Eastlake in his assessment of the data.
"Now you know what happened: It stalled," he said. "Now you have to wait for the (NTSB) board to answer why it happened."
A final NTSB report stating conclusions about the cause of the crash is not expected until the latter part of the year.
As a matter of course, NTSB investigators and officials do not speculate on causes of a crash prior to that final determination.
Transcripts of the cockpit recordings and information from the flight data recorder released Wednesday were contained in hundreds of pages on the crash compiled by NTSB investigators over several months.
Among other highlights of the NTSB's preliminary reports:
The wing flaps were not lowered. Pilots sometimes lower flaps to keep their plane in the air.
Neither engine showed signs of fire or ruptures before impact.
Numerous witnesses reported what they perceived as irregular engine noises shortly before the crash.
Weather information indicated possible icing conditions.
Some in-flight hazardous weather information that the Federal Aviation Administration said should have been broadcast was not.
At 20 seconds before the crash, the captain, Dann Carlsen of Grant County, Ky., says, "Looks like your low-speed indicator." And then he says, "Power."
Four seconds later, other warning sounds are heard.
For the most part, however, the cockpit recordings tend to show a crew that was in a good - and sometimes joking - mood during most of the 63-minute flight.
At one time, 1st Officer Kenneth Reece of Fort Wright, Ky., is heard joking that an air controller appears to have "short-term memory disorder" for repeating an instruction.
Numerous witnesses to the crash commented on the plane's engine noises shortly before it went down.
"All witnesses, except for one, reported that the engine sound was steady, or gradually increasing in frequency or volume prior to the sound of the explosion," NTSB documents said.
One witness reported that the plane was banking in a clockwise rotation until it abruptly stopped flying and pitched to an "extreme nose-down angle," NTSB investigators said.
COMPLETE CRASH COVERAGE
PROFILES OF VICTIMS
