Editorial Board

Does Your Airline Pilot Know How to Escape From a Stall?

These days, there’s a good reason fear of flying is classified as a phobia: Fatal airline accidents are exceedingly rare.

They could be made rarer. When planes do crash, the No. 1 cause is pilot loss of control. From 2001 to 2010, 1,756 people died in 20 such disasters. About half of these accidents involved aerodynamic stall, which is when the wings become so sharply angled into the wind that they no longer generate lift and the plane becomes dead weight. If stall-related accidents could be avoided, airplane fatalities might drop by as much as a quarter.

The just-released report on Air France Flight 447, which plunged into the ocean off Brazil in 2009, killing all 228 people aboard, points the way forward. In it, France’s Bureau of Enquiry and Analysis for Civil Aviation concluded that current training poorly prepares pilots for coping with aerodynamic stall.

As the Federal Aviation Administration works on new training rules for U.S. airlines, due next year, it should pay particular attention to three training flaws cited in the French report, all of which can be applied to American aviation.

The French report concluded, first, that commercial pilots, who receive some stall training at low altitude, don’t get it at high altitude, which is where the problems started on the Air France flight. The thin air high up leaves pilots a narrow safety margin in which to conduct a stall recovery: Aircraft have to fly fast enough to stay aloft yet slow enough to remain structurally sound.

Pilots of small planes practice diagnosing and recovering from stalls using actual aircraft, but it’s too risky and expensive for the major airlines to do this in their planes, which can cost as much as a few hundred million dollars apiece. Thus, high-altitude conditions need to be programmed into scenarios on the flight simulators that airline pilots use to train.

Second, the report found that pilots aren’t thoroughly trained to deal with sudden developments. Flight 447’s pilots were confounded when they lost airspeed readings because of a malfunctioning part; one of the pilots inadvertently put the plane into the stall that doomed it. French investigators wrote that training scenarios are too familiar to pilots and don’t sufficiently surprise them or test them with harrowing situations. In light of this observation, American aviation officials would do well to scrutinize the simulations used by U.S. airlines to make sure they’re challenging and ever-changing.

Finally, simulators themselves need to be improved so they can better replicate stall conditions. This was also the conclusion of the U.S. National Transportation Safety Board’s examination of the 2009 crash of a Buffalo, New York-bound Pinnacle Airlines Corp. Colgan Air flight, in which 50 people died. Today’s devices can train commercial pilots how to avoid getting into stalls but not how to recover from them.

For decades, the U.S. military has operated excellent simulations for its aircraft, having collected vast data on how the planes behave when they stall. Comparable information has not been developed for the commercial aircraft fleet. According to NASA research, though, it can be, using wind-tunnel testing and our current knowledge of aerodynamics.

It’s clear that more training would help. The safety board report on the 2009 Buffalo crash pointed out that although pilots are taught to drop a plane’s nose during a stall, they have done the opposite in one fatal accident after another.

and innovations addressing other major crash causes -- including wind shear, the icing of wings and the inability of pilots to see unlit mountaintops at night -- have produced a drastic decline in airline fatalities over the past decade. In fact, an American child to become president as to die on any single flight. With the right reforms, aerodynamic stall might be added to the list of aviation problems largely conquered.

    To contact the senior editor responsible for Bloomberg View’s editorials: David Shipley at davidshipley@bloomberg.net .

    Before it's here, it's on the Bloomberg Terminal.