The mystery deepens. The latest data from the EgyptAir Flight 990 voice recorder show that someone sitting in the co-pilot's seat uttered a prayer as something went terribly wrong with the flight. As alarms sounded in the background, someone struggled to save the plane, but inexplicably failed.
Pilots and investigators held high hopes for that voice recorder, expecting it to explain the many problems of this crash. Early flight data hinted at some sort of pilot-induced cause, a possible intentional crash. While news yesterday focused on the seemingly normal banter between the pilots as something went wrong with the flight, today's news about the utterance of a prayer before the autopilot disengaged remuddles the waters.
Were the pilots acting together, brilliant saboteurs who played the game all the way to the ground? It's doubtful, when they simply could have disabled the voice recorder, as an apparently suicidal pilot on Indonesian carrier Silkair did before pointing the nose of his 737 straight down in 1997, killing all aboard.
As a pilot myself, I was convinced that the voice recorder would reveal a frantic cockpit battle as one pilot struggled to save his craft while the other fought to nose in. I believed it not because of erroneous newspaper reports about how the recovered control surfaces pointed in opposite directions, which led "experts" to conclude that the pilots must have been wrestling over the yoke.
The reason I believed the battling-pilot theory was because of the disconnection of the autopilot -- the first of many mysteries about this crash. The plane was cruising, set up for a long overseas flight, the autopilot calmly tracking the radio beacons and holding altitude. It's common airline procedure to hand the flight over to autopilot shortly after taking off, and very unusual that someone would turn it off until landing. One reason would be if the plane decompressed, as Payne Stewart's Learjet apparently did. Procedure is for pilots, who have seconds to live at that altitude, to put on oxygen masks, disconnect the autopilot and quickly descend to 10,000 feet.
The early readings of the voice recorder bear the sabotage theory out. If both pilots were in cahoots, or one of them disabled, that would explain both the autopilot disconnect and lack of radio contact, the second of the mysteries surrounding the investigation. If they were working together to solve some sort of mechanical problem, the pilots certainly would have called air controllers.
A commercial pilot on an overseas flight plan can talk to a radar controller at any time by simply pushing a small button on the airplane's control yoke, then speaking into his headset's microphone. It's a one-finger operation, easily performed while flying.
Surely one of the two or three people in that cockpit had a second to push the button and say, "Uh, New York, we've got a little problem here." The plane dived for 40 seconds, each of them surely seeming like minutes.
Admittedly, pilots are trained to fly first, talk second. Yakking to air traffic control isn't going to save a falling plane, which would demand a pilot's full attention. But in 40 seconds of diving and a few of climbing, didn't at least one crew member have a moment to radio in an emergency call? Yes, unless they were somehow prevented from calling.
The third mystery is how and why the engines were shut down. Airplane engines aren't turned off by keys but by cutting off the fuel supply -- which may have happened just before this airplane climbed. Recovering from a steep dive puts tremendous pressure on the airframe, which can actually yank off the wings and tail.
Pilots familiar with the plane say there is no operating handbook procedure for a Boeing 767 recovery from a descent as steep as this one made.
But the first step in any dive recovery is to reduce air speed, which means reducing power. Is it possible that one of the pilots, in a frantic attempt to slow the machine down before pulling the nose up, instead cut the engines completely? The two procedures are difficult to confuse, so it's almost certain that the pilot to turn the power plants off. And then climb? It is possible, unlikely, pilots say, that a Boeing 767 could gain 8,000 feet without power.
Pilots, like everyone else, are bewildered by this crash. Like the non-flying public, we want answers. But pilots want answers that we can learn from. The early clues from Flight 990 pointed toward a hijacker, bomb, insane copilot -- these causes are not as interesting to pilots, because the chances of them happening again are rare, and the chances of a pilot being able to save his or her plane from them even rarer.
Explanations grounded in "pilot error" -- the No. 1 factor in small plane and commuter crashes, No. 2 in commercial (bumping into each other while taxiing is No. 1) -- are instructive. We study them to learn how to avoid re-creating the ill-fated pilot's mistake, living to fly again another day.
Pilots, like surgeons, are supposed to perform perfectly. The public allows for no margins of error. But we're human, and so we make them. Fortunately, most of our errors slip by uneventfully. In small airplanes, we sometimes dial in the wrong frequency for air-traffic control, for example. But that just means we get no answer, then look on the maps for the correct frequency. Pilot error, but nobody dies.
Imagine instead that you're teaching a grammar school class. Terrorists enter the room and threaten to shoot the children unless you remember -- quick -- what's the capital of Paraguay? Hah, I thought so. Boom, they're all dead.
So what do teachers learn from this scenario? To bone up on world capitals, obviously. The problem is, the next group of terrorists to storm a school will demand the square root of 79, and then what good is your geography?
The annoying thing about pilot error is that the term gets applied after the fact, with hindsight being 20/20. In May 1997, a Boeing 767 banged into Newark, N.J., damaging the landing gear (no injuries). The plane encountered wind shear on final approach, and the co-pilot battled to keep the plane on the ground once it slammed in. The NTSB conclusion was pilot error -- specifically an "improper landing flare."
In my mind, the guy was probably lucky to get it down without hurting anyone. Wind shear can flip large planes over or slow them down so suddenly that they stop flying and fall out of the sky. But the fact that he didn't react perfectly -- foreseeing the shear and in which direction it would send his plane -- lays the blame on the co-pilot.
Flying is so complicated and technical that no pilot, no human, could be proficient enough to recover from every single possible failure. That's why it isn't 100 percent safe, why even the most proficient test pilots die. There are parallels to surgery. The doctor has your beating heart in his hand and discovers something unexpectedly wrong with it. He can't put it down and go look up some info in his reference manuals. He has to decide, within seconds, how to solve the problem correctly.
Likewise, a pilot landing on a runway onto which a deer suddenly bolts has only seconds to make the decision: Hit the deer? Swerve to avoid? Attempt to take off again? Each option will almost certainly crash the plane. The cause will be pilot error, or, more specifically, "inability to maintain directional control."
A pilot who swerves to avoid the deer will probably ground loop and smash a wing. Is he a bad pilot? Frankly, I think he's a hero for saving the passengers and the deer. Is there something to be learned from that crash report? Sure, if the pilot and passengers walked away, that's the kind of "error" I'll plan to perform when faced with the same problem.
Which brings me back to the EgyptAir flight. As investigators pore over the black boxes, analyzing every nuance of sound and data, the anxious public waits. Was it a partial electrical failure, rendering the airplane almost uncontrollable? An autopilot problem? Clever sabotage? Was it a bungled recovery from cabin decompression? Not knowing what caused the crash is frustrating for everyone. But for pilots, eager to prevent their own machines from inexorably pointing their noses down into the ocean, the difference between knowing and not knowing can be the difference between life and death.
What all pilots dread is that there won't be a finding, and we'll have nowhere to point our fingers, even though the NTSB is valiantly trying to explain what happened. Charged with investigating every civil aircraft accident, the NTSB maintains a database of crash investigations and results. It makes recommendations based on the findings; for example, disabling the problematic thrust reversers on Boeing 767 aircraft. Pilots also read the reports to learn how to prevent their own potential accidents.
It's these kinds of accidents, and without known causes, which EgyptAir 990 may still turn out to be, that are the scariest. How does any pilot avoid those?