Frantic calls to 911 from passengers who had evacuated Asiana Flight 214 and saw others lying injured without attention. One of the pilots blinded by a flash of light as the plane was 500 feet from the runway. Questions about whether the Boeing 777’s engine throttle controls were malfunctioning. The captain delays evacuating passengers for 90 seconds after the plane ends its violent crash landing.
As various authorities release fragments from the drama at San Francisco airport last weekend, the picture of what happened gets murkier rather than clearer.
The chairman of the National Transportation Safety Board, Deborah Hersman, is doing her best to follow her avowed policy of transparency as her investigators crawl all over this crash, but making sense of it all is hard. Theories based on scant evidence swing from mechanical problems to pilot error and back again.
To be clear, there are just three facts that are beyond argument: the 777 was flying too low and too slow, and the crew was too late in realizing this condition.
It will probably be a while before we definitively know why. It can take the NTSB years to produce a final report on an accident. For example, we’re still waiting for its verdict on Southwest Airlines Flight 812, where a chunk of the cabin roof tore off a Boeing 737 on April 1, 2011—and that was a nonfatal accident.
In the meantime, though, the most concerning question about this crash is that it is, for sure, another in the one category of accidents that has proved hard to eliminate: “loss of control.” Other once persistently deadly categories have been virtually eliminated by technological breakthroughs—“controlled flight into terrain,” for example, in which an airplane hits a hill or mountain, usually on approach. Or midair collisions, now prevented by automatic proximity warning systems.
In 2011 the European Aviation Safety Seminar identified loss-of-control accidents as “top of the rankings of killer events.” Before the crash of Asiana Flight 214 the last fatal commercial crash in the U.S. was Colgan Air Flight 3407 near Buffalo in 2009, in which 50 people died; the two pilots completely lost control on approach.
Experts make a telling point about these crashes: they nearly always occur on approach and landing, when the need for the most basic flying skills is most acute, where the physical situation most resembles the old days of “seat of the pants” piloting, and where the crew becomes the critical and indispensable human extension of whatever technical wizardry has been installed in their cockpit.
Older pilots have often told me that they worry that cockpit automation has reached the point where pilots who have known nothing more primitive get too relaxed and complacent about flying and are in danger of losing their situational awareness and ability to “read” where the airplane is going. It’s a tough argument, because that same automation is what actually enables airplanes to fly safely in weather and traffic that would once have been too dangerous.
Indeed, the reach of cockpit automation is only going to increase. For example, Honeywell is testing something called “synthetic vision,” which it has patented and which is designed to substitute for the pilot’s own eyes in conditions where weather renders the terrain and the runway virtually invisible.
When the eyes follow the arms, the feet, and the seat of the pants into redundancy, you might wonder, what exactly is left for the pilot to do? The answer: everything. There is simply no substitute for the experience and proficiency of a properly trained pilot. And there’s the rub: properly trained, which includes maintaining alertness and acuity when surrounded by wonderful things that are supposed to work but sometimes don’t.
Oh, and by the way, about that decision by the captain of Flight 214 to delay evacuation until he was alerted to the outbreak of fire: it doesn’t mean he was derelict. The captain is responsible for the decision to evacuate and by what means. He had to get a sense of the condition of his airplane. Was it safe to activate the slides? Or were the slides badly damaged in the crash? (A few were.) He was making a damage assessment.
Just how tricky this decision can get was demonstrated in November 2010, when a Qantas Airbus A380 superjumbo, badly crippled by an engine explosion, made an extremely perilous emergency landing in Singapore. When the A380 came to rest just short of the end of the runway (its badly damaged brakes barely worked), it was leaking highly flammable fuel; the temperature of the brakes, virtually red hot and glowing, had passed 1,000 degrees Celsius; and one engine could not be shut down even when hosed by fire crews.
The A380 had been certified to be capable of evacuating 853 passengers and 20 crew from 8 doors in just 78 seconds—in the dark. The Qantas captain in Singapore had 440 passengers aboard in broad daylight. In spite of the fire risk, he decided that evacuation by slide would cause many injuries, particularly among the elderly and infirm. (The doors on the top deck of an A380 are 24 feet above the ground.)
Singapore airport was slow in providing mobile stairs to allow passengers to walk down, on the right side of the airplane, where the fire risk was less. The captain waited 52 minutes before the stairs were in place. He is judged as having been as right as he was cool.