The one significant fact to emerge about the Cory Lidle crash is that the other person killed was aboard the airplane with Lidle (rather than in the apartment building or on the ground), and was indeed an experienced flight instructor, or CFI. As mentioned yesterday, the whole effort to understand what went wrong goes in different directions, depending on whether Lidle, a newly minted pilot, was known to have had help in the cockpit. For one thing, the presence of a CFI makes the weather that day seem a less significant factor.
In most small-plane crashes, weather is the heart of the problem. Or to be more precise, the combination of difficult weather (low clouds, fog, icing conditions, thunderstorms, high winds, etc) and the pilot's decision to undertake or continue a flight in those conditions. If a very new pilot with no instrument rating, like Lidle, had been alone in the plane, the drizzle and somewhat low ceilings reported that day might have been distracting enough to have started a chain of bad reactions.
Presumably this would not be so for any experienced flight instructor. The aviation-weather readings, or "METARs," from nearby airports showed the lowest reported ceiling in the vicinity—that is, the bottom of the clouds—at 1500 feet. Lidle's airplane had to stay about 500 feet below that altitude anyway, because of airspace restrictions above the East River. Even if wisps of clouds trailed down, as they often do, and even if conditions over the river were different from those reported by airports even as close as LaGuardia, the presence of a CFI makes it less likely that the crash was because of disorientation in clouds.
James Fallows's Web site, with regularly updated dispatches, and information about his writings and appearances.
So what did happen? Perhaps we'll never know, since the two people involved will never explain what they did and saw. Throughout much of small-plane history this has been the case: since so few crashes were survivable, there were few first-hand witnesses to describe how the problems developed. Precisely because Cirrus's parachutes have saved so many people who would otherwise have died, a significant new corpus of knowledge has emerged from pilots who could describe how things went wrong—and they lived. (This posting, by a Cirrus pilot, describes the history of parachute "saves" and how the Cirrus has become safer than the norm for small airplanes, after a difficult early period.) Obviously this crash is not in that category. In particular, we may never know the exact dynamics between instructor and student—always complicated, since the instructor wants to leave the student in charge as long as possible, but not too long, in case big a mistake occurs—or who had the controls at the end. (My guess is: the instructor.)
But here are two interim explanations—not of what happened in this crash but of predicaments like this one.
The first is the nightmare of the "box canyon." When I first heard about the accident, my mind skipped over the mention of "Upper East Side" and let me imagine that this had happened along the Hudson River, on the west side of Manhattan. I did so because that is the only route I had flown, and because it is so much easier and more "normal" a route for aircraft going around Manhattan. On the west side the river is relatively wide, and the course is direct. You watch like crazy for other aircraft, and you just keep going up or down the river.
On the east, there's no "through" corridor of the same sort. A few miles up the East River, roughly parallel to 85th Street, LaGuardia's airspace begins. (Technically, the floor of the LaGuardia airspace reaches to the surface at that point, so there is no way to go "under" it, as you can further south on the East River or along the Hudson "flyway.") So if an airplane is operating VFR—under Visual Flight Rules, not talking with controllers and watching out for traffic and obstacles itself—it has to turn around and head south. In an emergency, a pilot could keep on going and plow into LaGuardia's space, or he could make an emergency request to LaGuardia for entry. Each would disrupt normal LaGuardia traffic (though not endanger it, since the controllers would see where everyone was on radar), and each would mean that the pilot later had a lot of explaining to do to the FAA and would probably be punished in some form. But as one of the flying cliches goes, "I'd rather go to an FAA hearing than to my own funeral."
Making a 180-degree turn to get out of a "box canyon"—a literal one, in the mountains, or an airspace one, like the situation at the top of the East River—is harder than you might think. The best way to imagine an airplane's performance is to think of a bicycle. Like a bicycle, an airplane has to keep moving in order to stay up. If a bike goes too slowly, it falls over. If an airplane goes too slowly, it falls out of the sky. For airplanes the phenomenon is called "stalling," and it refers not to engine performance but to the fact that the air is not going over the wings quickly enough to provide lift for the plane. But as with a bike, the faster an airplane goes, the more space it needs to complete a U-turn. A bicyclist who is roaring down a hill toward a hairpin curve needs to calculate just how fast he can go to make the turn. Something similar happens with an airplane nearing a forced 180-degree turn, like the one on the East River. The pilot has to figure out the combination of air speed (slower the better, but not slow enough to risk a stall), bank angle (steeper the better—except that at a steeper angle the risk of stalling increases), use of flaps, allowable loss of altitude, and so on to create the so-called "minimum radius turn."
The second most uncomfortable moment I have had while flying was in the Cascades mountain range, near the town of Concrete, Washington. The weather was perfect, the scenery was beautiful, but the walls of the valley containing Concrete were getting closer together than I liked. I had recently trained in the "minimum radius turn" and I made one to get out of the valley—flying over toward one edge to gain maximum turning room, reducing the speed, using the flaps, banking for the turn. In reality, it was not a close call at all—I probably had two miles to spare on either side—but it felt much closer than I would have wanted. The turning space at the top of the East River appears to have been much narrower, and a there was a stiff wind from the east. This would have pushed a plane, as it made its U-turn, toward the opposite "canyon wall," in this case the apartments on the Upper East Side. We don't know what happened to Cory Lidle and his CFI, whose experience was mainly in California. But we know that this kind of situation can be difficult.
(My first most uncomfortable moment? When taking off from a short grass airfield, in Vermont, that had trees at one end—on the day after a heavy overnight rain. I had flown out of the field the day before and cleared the trees by hundreds of feet. The grass, when wet, became an entirely different surface, and I cleared the tree tips by much less than was comfortable. Moral to me: never again a grass strip if there is any kind of obstacle anywhere in sight.)
Here is the other explanation: why Cory Lidle could have said, in the quote now replayed with such bitter irony, that flying was "safe." I think this is what he meant:
In normal circumstances—clear weather, no obstacles, traveling point to point—flying seems and to a large extent is fundamentally safe. The big engines are amazingly reliable. On the rare occasions when they fail, you're taught how to glide the plane in for a landing—or use a parachute, with the Cirrus. There is no traffic whizzing by you six feet away, as on a busy road. The airplane "wants to fly"—it is designed to be stable and stay aloft. Once you learn to land a plane, much like learning to ride a bike, it's a relatively easy and natural process. Landing is the one part of flying an airplane that requires coordination and reflexes like those called on constantly when driving a car. Flying requires awareness of a lot of things at once—weather, navigation, terrain, how the airplane is performing, what are the next three things you have to do. But usually it's not as instantaneously demanding as driving is. If you fail to pay attention for a few seconds in a car, you can die. That is very rarely the case in airplanes—for instance, the last phases of an instrument landing under bad weather conditions. In a car you can very easily die through absolutely no fault of your own, as when a drunk driver runs a red light. That is also rare in aviation. For better or worse, most of what happens is your own doing. That is what he meant, though it's all very sad now.