A number of interesting things starting with the Tower accident at JFK. I talked with the F.O. who brought the plane to us in Viracopos, Brazil (more about Viracopos later). He was in a cockpit jumpseat of the airplane that went off the runway. He lives in Miami and was on his way home. It turns out that there is plenty of potential blame to go around.

The first problem was that the control tower gave them runway 4 left for takeoff in spite of there being a direct crosswind. Apparently the crew questioned this, but the tower informed them other aircraft had used that runway without difficulty. However, the guy I talked to said that when they lined up, he didn't notice any recent tracks in the snow.

The second problem was that, in hindsight, the captain should not have accepted a crosswind slippery runway. However, as I said, the tower said no problem. At this point the captain was in a no-win situation. If he refuses the runway, he causes all sorts of delay, and at Kennedy the pressure to not do that is significant. If he elects to go with that runway, he's got the problem with the crosswind and the slippery runway, which of course is what he elected to do...and he got unlucky. Apparently there's also some controversy over how strong the crosswind was. At this point, of course, everyone is trying to cover their own asses.

Some where around 100 knots (that's 115 mph) the airplane started to drift (either right or left, I don't know which, we'll use right for talking purposes). 100 knots is below V1, the rejection speed. The rule is that if there is a problem below V1, you abort the takeoff. However, there's another rule that says to not use reverse if you're skidding sideways because that aggravates the skid. Actually, it's not a rule, it's a law of physics. The captain is really in a no-win situation now. He knows that his brakes won't be much good on the slippery runway, and he can't hold the airplane on the runway. Should he continue and risk going off the runway at much higher speed. or should he abort. He elected to abort. He may have had one other option which is very nonstandard, but which we will talk about in a bit.

Okay, so now he's aborting and feels he has to use reverse (probably very necessary to keep from going off the end). The use of reverse virtually guarantees drifting off the side of the runway, and that is exactly what the airplane does. Now things get more interesting and illustrate the fact that in accidents, it's often a combination of things that produce the final result.

In this case there's a trench off the side of the runway that the New York Port Authority neglected to fill up. The nose gear drops into the trench and gets ripped off. This puts the airplane's nose on the ground, which then causes one of the engines to get ripped off. Had the trench not been there, they might have just wound up with the embarrassment of an airplane off in the snow.

The FAA has a requirement for what is called a runway clear area, and that clear area extends off the side of the runway. I don't know how far, and the distance may be different for different classes of runways. It turns out that the trench in question is right on the line of what would be legal and what would not. So, Tower is of course saying it's in the clear area. The Port Authority of New York is saying it's not.

The captain perhaps had another option. It's rarely talked about and is never practiced in training—the simulators are not programmed to handle it. Once you start skidding (let's say to the left from a right crosswind), the wheels obviously have already broken loose. You don't have any traction. Insofar as the wheels are concerned, you may as well be in the air. Now, what do you do if you want to follow a given course (in this case the runway) with a direct crosswind. You crab into the crosswind. The captain could have used the rudder (he had sufficient speed for it to be effective enough) to crab the aircraft into the crosswind enough to match the crosswind. He then would have been going down the runway at an angle to the centerline, but he would have been able to stay on the runway if he did it just right. Then, when takeoff speed was reached, a normal rotation could have been made.

It would be a risky maneuver. If you didn't do it just right, you would go off the runway anyway—off the downwind side if you didn't crab enough, off the upwind side if you crabbed too much. Except that now you might be leaving the runway at a speed pushing 200 mph instead of the 115 they were at—few people would walk away from that.

What would I have done. I don't know. I and everybody else analyzing it have the luxury of taking all the time we want to think about it. That captain had a couple of seconds at most.

Now to back up to the incident at Amsterdam a couple of weeks ago. They landed long and decided they needed to go around. A go-around decision once you're on the runway is always subject to raised eyebrows. A captain is always going to have to explain to the chief pilot why he did that. Anyway, they initiated a go around. As they lifted off, whoever was flying allowed the right wing to drop when he rotated. The wing tip and the right outboard engine pod scraped the ground. The aircraft then lifted off, and they came around and landed without incident.

The captain is in trouble on this one regardless of whether he or the f.o. was at the controls. The FAA will question his judgement and also accuse him of failing to control the aircraft.

In theory, pilots at the controls of 747s are supposed to be experienced enough that they will never touchdown at a point on the runway beyond which they will be able to stop. Necessary go-around decisions should always be made before touchdown, and nobody will seriously question a go-around made before touchdown. They may say that they thought it unnecessary, but the rule is that if you're uncomfortable with what you see just prior to touchdown, go around. So it's the flying pilot's call. If the f.o. is flying, the captain can also override him. That's as it should be.

By the way, the touchdown zone of a runway is the first 3,000 feet. Ideally a 747 will touchdown between 1,000 and 2,000 feet from the end. If you touchdown less than 1,000 feet from the end, that's considered cutting it a little close, especially since it indicates that you were a little below the glideslope—a real no-no in the 747. The glideslope (the radio signal that allows you to follow a constant three degree, usually, angle to the runway) intersects the runway 1,000 feet from the end. Thus, if you're on the glideslope, you will always touchdown beyond 1,000 feet. A touchdown between 2,000 and 3,000 feet is okay, just not perfect.

Anything beyond 3,000 feet tends to be questionable. However, many pilots on long runways routinely touchdown beyond 3,000. They hold it off trying for a real greaser. It's no real problem if, say, the runway is 12,000 feet long and dry. I like to follow consistent habit patterns, so my rule is to always be down within the touchdown zone. The one time I took an airplane from an f.o. when I was a captain was on a long landing at Narita (Tokyo). This f.o. had been a captain at another airline. He was an excellent pilot, a better stick than I really (the term “stick”refers to a pilot's ability to maneuver the aircraft in the air—it originates in the fact that aircraft used to have sticks rather than today's control columns). He always tried for greasers. To get them, he'd carry ten knots extra speed into the flair and then float, and float, and float...easing it down. Of course, when he did get down, he had to often really pile on the reverse and the brakes to get it stopped. One day at Narita (13,000 foot runway), he still wasn't down at mid length, in other words, we had six to seven thousand feet of runway behind us. That did it, I pushed the control column forward, forcing it on. We had a little talk later. I told him plainly but nicely that when he was with me, get it down within the touchdown zone. Greasers were nice, but I was tired of overheating brakes and raised blood pressure.

Back to Amsterdam, the next problem is that of their having let the wing drop. It's not unusual for pilots new to the 747 to have a little difficulty with dutch roll in the approach and takeoff configurations. Dutch roll is the term applied to alternate movement about the longitudinal axis of the airplane—rocking back and forth left to right. When the flaps are down, the airplane is not as stable about the longitudinal axis (the longitudinal axis is a line from the nose of the airplane to the tail of the airplane) as it is when the flaps are up. The problem for the pilot is that he may not notice that a wing has dropped a little. He has few cues to alert him when he's on the takeoff roll or the last few feet of the landing. In both those situations, his attention is outside the airplane. He has little time to glance at the artificial horizon inside the airplane, and that's the only real indication he has that a wing is down. Of course, he finally does realize a wing is down, and he then corrects. Possibly, and especially if he's new, he may overcorrect. It's the classic situation. The new guy doesn't realize the need for a correction until he has to make a large correction. Large corrections are harder to make than small corrections. I must have said that last statement to students a thousand times when I was instructing.

There's another problem if you're on the takeoff roll, which they were since they had touched down. If you have a crosswind, you should hold a little aileron into the crosswind. That means that if the crosswind is from the right, you hold a little right aileron. When you lift off, you take that out. What happens if you forget to take it out. The right wing will drop.

Or, alternately, what happens if you don't hold aileron into that same right crosswind. You risk the wind getting “under” the right wing and lifting it, which of course puts the left wing down. I don't know yet whether they were contending with a crosswind at Amsterdam.