Years before the tragedy of Flight 3407, two airline crashes blamed on icing set the stage for an experiment.
In a 2002 study, researchers sought to determine how well new airline pilots could handle an "upset airplane," including airplanes upset by ice.
The answer: Not well at all.
Forty relatively new pilots with various levels of training were tested in the skies over Western New York. The in-flight simulator gave them eight possibilities, including two modeled after the crashes of turboprop airplanes affected by ice -- in 1994 in Roselawn, Ind., and in 1997 in Detroit.
Fewer than half of the 40 pilots were able to regain control on their own. Those pilots later talked about how inadequately they had been trained in recovering an airplane from an ice-induced stall and told researchers that they felt unprepared.
Their training had stressed protecting altitude during a catastrophe in flight. But an airplane stalled by icing needs speed, even by sacrificing altitude.
The results provided more evidence that the threat of icing demands myriad solutions, not only with de-icing equipment and warnings against the use of autopilot but with deeper levels of training.
While the Federal Aviation Administration has responded in some ways, people within the industry cite the crash of Continental Connection Flight 3407 as proof that federal regulators must do more.
Capt. Marvin D. Renslow had more than 3,000 hours of flight experience but 110 hours at the controls of a Bombardier Dash 8 Q400 turboprop, the airplane that plummeted into a Clarence Center home after a stall Feb. 12. The crash killed 50 people.
Veteran pilots have said that even the greatest aviators could not have rescued Flight 3407 since the problems began only 1,600 feet above the ground. Still, investigators with the National Transportation Safety Board are examining the "human factors" to determine whether Renslow or First Officer Rebecca Lynne Shaw tried to pull up the aircraft's nose to stabilize the plane and preserve altitude.
Several articles on ice-induced stalls suggest that pilots in similar situations should fight their instincts to raise the nose and instead adjust the angle downward to gain speed.
That lesson was highlighted by the 2002 research conducted by Veridian Corp., as Western New York's Calspan was known then, for the National Aeronautics and Space Administration. The newer pilots who recovered the airplanes after simulated ice-induced stalls did so largely because they flew at greater speed.
The results have been quoted in airline industry articles, including "New Airline Pilots May Not Receive Sufficient Training to Cope with Airplane Upsets," published in 2003 in the Flight Safety Digest.
"The appropriate actions must be highly learned skilled responses that can be executed quickly," the authors said. "Under current airline training regimens, pilots rarely have opportunities to practice the appropriate recovery procedures."
The lesson of sacrificing altitude to recover a plane affected by icing has not been emphasized enough -- in the simulator or in the Federal Aviation Administration certification standards, wrote John P. Dow Sr., formerly the senior icing specialist for the FAA.
"Pilots of turboprop airplanes should be taught that they might need to trade some altitude for airspeed if the airplane stalls in flight during icing conditions," Dow said in a 2005 article for Flight Safety Digest.
The FAA has not been deaf to issues about icing. It issued a safety alert in November 2006 urging pilots who see ice building on their planes to disconnect the autopilot every five minutes.
The National Transportation Safety Board, however, recommends that the autopilot be permanently turned off as soon as a plane encounters icing.
Also, in its practical test standards for airline transport pilots, the FAA has added the expectation that pilots be able to recognize and deal with the effects of icing in all phases of flight.
But the standards still contain the language that some experts have found ambiguous and might limit pilots from sacrificing altitude to come out of an ice-induced stall. The FAA expects an airline transport pilot to be able to recover an aircraft with a "minimal loss of altitude."
Jim Hall, a former NTSB chairman, has been criticizing the FAA for not implementing the safety board recommendations that followed past crashes involving icing, particularly those in Roselawn and Detroit. Hall said the FAA needs to do more to combat icing in several areas, including with training and certification.
Veridian's research, in which most newer airline pilots failed to stabilize a plane stalled by icing, "speaks volumes to the need for a pilot-certification process that requires that there are standard procedures to recover aircraft in stall situations," he said.
The researchers wrote that they selected 40 newly hired airline pilots without military flight experience because they represent the majority of future airline pilots. Newly hired pilots are more commonly used by regional airlines, such as Colgan Air, based in Manassas, Va.
Colgan, which employed Renslow and Shaw, was operating the Continental Connection flight through its owner, Pinnacle Airlines.
On its Web site, Colgan says its training deals with such emergencies as icing and stalls.
Colgan Air said that Renslow and Shaw reviewed the Winter Ice Bulletin in November in preparation for this winter season. The bulletin provided the latest information on flying the Q400 in icing conditions, and it supplemented the training and experience the captain and first officer already had with the Q400.
But in November, Renslow still had not been certified to fly the Dash 8. He still was completing his 172 hours of training on the airplane.
He did not declare on a networking Web site that he had finished his training until Dec. 8.
News Columnist Donn Esmonde contributed to this report.