Skip to main content /TRANSCRIPTS



NTSB Press Briefing

Aired November 15, 2001 - 16:16   ET


JUDY WOODRUFF, CNN ANCHOR: The National Transportation Safety Board is beginning its briefing on the crash on Monday of American Airlines Flight 587. Let's listen in and my apologies to Senator Graham.


MARION BLAKEY, NTSB CHAIRMAN: First, I'd like to tell you all that, as you know, we have been assisting with the families and the victims of this accident. Much of that work is going forward in a variety of ways. It may be helpful to know that the medical examiner has now informed us that he believes all of the human remains have been recovered from the accident site. So that stage of this is complete and the medical examiner is continuing to collect information from the families, but the identification process is going ahead fairly swiftly at this point.

Secondarily, it's probably help to know that we have much of the wreckage now cleared. And we expect that we will have most of that done in the next day or so. As you know, almost all of that wreckage will first be going to Floyd Bennett Field. And, for example, we have moved the pylons from the engines, and they are now at Floyd Bennett, while the engines themselves have gone on for a complete tear-down in Tulsa.

That brings me to the point that many of the critical components will be shipped out of here fairly quickly, some to go down to Washington to the NTSB's own lab, some to the manufacturers, where they have the drawings, the toolings, the expertise where, along with our team, they will be examining those specific components.

We'll tell you a bit more about progress on all of this tonight at a later briefing after our progress meeting from all of the teams who are coming from the field. But I do want to point out that this is really the tip of the iceberg here that you're seeing here in New York. And it will be many months of pulling a lot of this information together before we really will have a complete and clear picture.

I think it's also important to take note of the fact that, as this is all going forward, the FAA and the director general of the BEA, the equivalent of the National Transportation Safety Board in France, are working together and in discussions on the issue of inspections. And from what we understand from them, there is an expectation that they will jointly be issuing an air worthiness directive within the next day. That is as much progress as we have on that, and that, of course, will come from them.

Secondly, there was a discussion, and in one of the last briefings we had about an issue of turbulence with this particular aircraft. I would refer to you the fact that there's a report on that, in fact, on the FAA's web site if some of you have not seen it. This was an incident in '94, where were there were a number of issues as a result of turbulence. And there were are finding that at the same time that that occurred there was a turbulence upset inspection done.

In point of fact, the mechanics, in doing that, did not discovery anything in terms of the aircraft that warranted repair at the time.

Now to turn to the issue of the flight data recorder, as many of you know, we were concerned that we had a damaged recorder and we might not have full data. I'm glad to tell you that we, in fact, do and we are in the process of reading that data out.

What we have for you today is preliminary, and it is not complete, but we do want to provide for you everything that we have that we know to be reliable at this point. And I'll go a little slowly because I know many of you will need to take notes on this.

The end of the flight data recorder occurs when the airplane is at about 2,900 feet, which is about 20 seconds or more before the cockpit voice recorder stopped. This is because they are on different electrical systems. It is very likely that the cockpit voice recorder, in fact, was battery operated, so that would account for why there is a slight difference there.

Acceleration data that we have been able to read out from the flight data recorder is consistent with a wake encounter about 28 seconds before the end of the FDR data.

Airplane attitudes did not change much in the first encounter. Acceleration data is consistent with a wake encounter about eight seconds before the end of the flight data recorder data. The second encounter appears to be similar in intensity to the first one.

During the first part of the eight seconds, the airplane appears to be consistent with flight control inputs. In other words, it's responding normally to the pilots and to other inputs -- control inputs.

During the last eight seconds, there were three side acceleration excursions or 0.3 G, 0.4 G and then an opposite 0.3 G coincident with rudder deflections. This is a significant lateral acceleration.

About 2.5 seconds before the end of the FDR data, the rudder position data becomes unreliable. During the last few seconds of the FDR data, we see the following. The side acceleration increases from about 0.3 G to about 0.8 G. The airplane heading changes at about 10 degrees per second to the left. The bank angle is increasing through 25 degrees, left wing down, with the control wheel to the right.

QUESTION: (OFF-MIKE) too fast. BLAKEY: Too fast, still? I'm sorry.


BLAKEY: Slower than that?

QUESTION: The whole thing.

BLAKEY: I'm told I speak slowly, so therefore I thought maybe I was.


BLAKEY: Let me give it to you from the top then. All right.

The end of the FDR data occurs when the airplane is at about 2,900 feet, which is about 20 or more seconds before the cockpit voice recorder stopped. Acceleration data is consistent with a wake encounter about 28 seconds before the end of the FDR data. Airplane attitudes did not change much in the first encounter.

QUESTION: In the first encounter of this wake?

BLAKEY: Correct.

Acceleration data is consistent with a wake encounter about eight seconds before the end of the FDR data. The second encounter appears to be similar in intensity to the first. During the first part of the eight seconds, the airplane appears to be consistent with flight control inputs. In other words, it's responding normally, essentially, to the pilot's input and input of other control factors.

QUESTION: The second encounter or (OFF-MIKE)

BLAKEY: The first part of the eight seconds.

And I'm sorry. One of the things -- bear with us, as I say, this is preliminary and we have not been able to do the kind of truly exact work on this data that we will do and we'd like. So when we're saying first part, second part...

QUESTION: But are you saying the first part of the second encounter? OK. That's what I'm just trying to understand. All right. Thank you.


BLAKEY: Of the last eight second. In other words, we're talking now about the last eight seconds on the FDR.

QUESTION: And it's responding normally.

BLAKEY: Yes. Responding normally to the pilots. It appears to be the inputs -- the plane is reacting normally.

OK, during that same last eight seconds, that's what we're talking about here, there were side acceleration excursions of about 0.3 G, 0.4 G, and an opposite 0.3 G coincident with rudder deflections. And I just point...

QUESTION: That was the plane was moving and the rudder was moving, as well.

BLAKEY: Exactly.

QUESTION: Did you say coincident?

BLAKEY: Tom, check me on this, but that is my understanding.

QUESTION: Were all three coincident with rudder deflections or just the last point going in the opposite direction?


QUESTION: All three. Thank you.

QUESTION: Does coincident mean apparently caused by?


BLAKEY: No. That's the reason we use the word coincident.


And again, we're asking that you all understand that some of these things we are, obviously, looking into, we need to be able to mine this data much further before we can tell you some of those answers.

But the point, I think, to take away from this is that this is a very significant lateral acceleration here we're talking about.

QUESTION: In other words, you measured two separate parameters and they both went the same ways at the same time.

BLAKEY: No. Let's go back to that a minute, because I'd like to have a little bit more explanation than this. But let me just take everybody down through the fabric, because I only have a couple more points here to be made.

We're now at 2.5 seconds before the FDR data runs out here.

BLAKEY: At this point, the rudder position data become unreliable. Then, during the final few seconds -- and I'm not being precise about 2.5 -- I'm just saying the last few seconds here of the FDR data, the side acceleration increases from 0.3 G to about 0.8 G. We've already commented that 0.3 G and 0.4 G is a very significant side acceleration already. Now we're at 0.8 G.

The airplane heading changes at about 10 degrees per second to the left. The bank angle is increasing through 25 degrees, left wing down, the control wheel is to the right.

QUESTION: Can you repeat that?

BLAKEY: Happy to. The bank angle...

QUESTION: Go back to the heading changes.

BLAKEY: The airplane heading changes at about 10 degrees per second to the left. The bank angle is increasing through 25 degrees, left wing down, while the control wheel is to the right.

I see everyone doing this, and a minute I'll ask Tom to help you by doing this, OK?

QUESTION: The wheel is going that way, the wing is going left.


QUESTION: So that would indicate the plane was cross-controlled?

BLAKEY: I've got two more points here, and I'd like to -- let us discuss it with you, all right?

The final two points: the pitch attitude drops to 30 degrees, nose down, while the vertical G-load increases to over 2 Gs.


BLAKEY: Two Gs. And it's over 2 Gs by the end.

And remember, as I said before, the flight data recorder does quit about 20 seconds before the cockpit voice recorder. So we're continuing to read out information from the cockpit voice recorder which will carry us a little bit further than that.

QUESTION: Twenty seconds more on the CVR, is that what you're saying?

BLAKEY: Exactly.

QUESTION: Can you tell us what happened...


BLAKEY: I will make one more point please, because this has come up a number of times today and I feel that I should say this. I would point out that the investigation continues, on the basis of all the information we have to this point, to point to an accident. We're continuing to work with the FBI looking at all the evidence, and we will continue to consider all questions as to whether or not there is any evidence of criminal activity.

But to date, we do not see it in any of the information we have brought in since we last briefed you. So it continues to point to an accident at this point. But as I say, we're going to continue to look at this question as time goes on.

All right, now on that perhaps we could take questions. QUESTION: A lay person, if one of you could describe, I have no idea of what you just said. Is the plane...



TOM HAUETER, NTSB ASSISTANT DIRECTOR: When you take a look, and describe this first, it's left wing is down. The heading is changing to the left and the nose is also nose down.

QUESTION: And those G forces, does that mean it's being really hammered? I mean, what?

HAUETER: Well, 2 Gs mean -- you're sitting in your seat, you're feeling 1 G right now. So it would be like twice your weight in the seat.

QUESTION: The side acceleration Gs, explain in layman's terms what that means.

HAUETER: The side acceleration is simply that. You're being pushed to the side. In this case, we say 0.3 G, that'd be like a third of your weight pushing you to the side.

QUESTION: Would that be something that would create the condition known as flutter in the vertical fin?

HAUETER: I wouldn't speculate on that now.

QUESTION: Coincident rudder movements: can you explain that? Is the rudder moving in a way that produces the G or against that -- is the rudder being used to recover against that?

HAUETER: We're still examining the data.

QUESTION: Can you tell us what the rudder deflection is? Is that measured on the FDR?

HAUETER: Rudder deflection is measured. I do not have that with me. But, yes, rudder deflection is measured.

QUESTION: And I think this is pedal rudder, this is not that -- we spoke before about trim.

HAUETER: This is rudder position.

QUESTION: We're talking about pedal inputs not trim inputs?

HAUETER: We're talking about rudder position.

BLAKEY: Some of this we -- now, bear with us, again.

HAUETER: We're talking about the actual position of the rudder. And obviously, there are several things that can move the rudder, but what we're providing is rudder position. QUESTION: But you said "coincidental."

QUESTION: I mean, you know which way the rudder turned; to go back to the New York Times question, is this against this force, or is it with this force?

HAUETER: We're still looking at all the data. We have not fully validated all of our directions to make sure of that.

QUESTION: Does it actually remain consistent with, as opposed to coincident?


QUESTION: When the rudder data becomes unreliable, is that consistent with when it appears the rear section of the wing, kind of, came off?

HAUETER: We haven't done any correlation of all the timing yet.

QUESTION: What do the voice recordings tell you that coincide with the data? Are you getting any information that's useful there?

HAUETER: Once again, we haven't done any correlations between the FDR and the CVR yet.

QUESTION: What occurs, though, in the last 20 seconds of the CVR period? Do you know...

HAUETER: We reported that to you several days ago.

QUESTION: But we forgot that.


QUESTION: Was it right, right, left, or left, left, right?

BLAKEY: What we hear on the CVR...


BLAKEY: I can't be actually precise with you about the last 20 seconds. What I just wanted to remind you is that there was a significant rattle twice during the end of the CVR. And one of the things we're going to be doing, of course, is mapping the FDR data and the CVR data together so that we can then bring this together much more precisely, but we have not been able to do that yet.

QUESTION: I'm trying to understand this. Is there a correlation with the wake encounter and acceleration?


QUESTION: Can you explain it as clear as possible?

HAUETER: I think as we mentioned, there's an acceleration about 28 seconds prior to the end of the FDR that is consistent with a wake encounter. There is also another...


HAUETER: Acceleration vertical, lateral, horizontal. All three modes, all three axes. About eight seconds prior to the end of the FDR, there is another group of accelerations or acceleration that's consistent with a wake encounter.

QUESTION: When you say acceleration, you mean the pilot increased his engine speed to try to get out of that? Or that the speed...

BLAKEY: The speed of the aircraft.

HAUETER: No, not speed. It's the G-loads vertically, laterally and longitudinally.

QUESTION: Can I follow up? About the amount of Gs that you're talking about here, are these excessive forces that the plane would have encountered? Or are they similar to what I might encounter when I'm experiencing turbulence out in normal flight?

HAUETER: They are slightly higher than you would see. Normally, 0.1 Gs lateral is something you see; going up to 0.3, 0.4 is something not that usual.

QUESTION: So what you're saying here is that these are significant forces. Are you at a point where you can say these are significant enough forces that they compare with accidents that have been caused by wake turbulence, or that they are significant enough to rip a tail off a plane?

HAUETER: I wouldn't say at this time.


BLAKEY: Yes. I think that -- you know, one of the things we have to do is analyze this against other wake turbulence information we have. We haven't done it yet

QUESTION: The event that you think is the wake turbulence is not the significant 0.3 G lateral motion; those are separate, right?

HAUETER: The 0.3 happened slightly after that acceleration.

QUESTION: What is the measurement of the wake encounter lateral acceleration?

HAUETER: There's no standard. It depends on what wake you encounter.

QUESTION: No. I mean, on the flight data recorder, you had the recordings of the lateral acceleration in the second encounter. In the first encounter, with the wake turbulence, what were the figures for that?

HAUETER: I don't have those here.

QUESTION: But they were less than these subsequent figures that you called very significant.

HAUETER: That would be the indication. But I don't have those with me.

QUESTION: If I got it right, it's gone -- the lateral Gs, it's gone 0.3, 0.4, so, you know, up to a maximum of a half a G in one direction. Then it swings off. Is it over to 0.3. So that's only -- it's still less than 1 G over in the swing. Finally, it gets up to 0.8. I mean, all told you're not putting together more than 1 G, a little bit more than 1 G in those swings until the very end when it goes up to a couple of G, and it sounds like he's already upset there.

So you say it's very significant, but isn't that -- that's enough to give you a good solid jolt. It's not more than that, I don't think. I don't think it's not terribly uncommon, is it?

HAUETER: Lateral Gs higher than, like I said, 0.1, 0.2 is uncommon.

QUESTION: Could it do physical damage to an aircraft? (CROSSTALK)

QUESTION: Two questions, if I can: Just to clarify, I think you answered Eric's (ph) question, but I want to make sure I understand, the time where the high lateral Gs were occurring in that last eight seconds in the FDR, that was after the second wake encounter; is that true?

HAUETER: It's very close and appears to come right after the encounter.

QUESTION: Right after. All right.

And would it be -- were the wake encounters consistent with airplane flying through one and then the second wake vortices of the other jet? Is that what that would be consistent with?

HAUETER: That would be consistent. Obviously, an airplane puts out two vortices.

QUESTION: Can you tell from the data whether the pilots responded any differently to the two wake encounters?

HAUETER: I can't. We haven't looked that far in the data yet.

QUESTION: Would it be fair to say that the plane had a -- that the wake encounter had a dramatic impact on what happened to that (inaudible)?

HAUETER: I wouldn't say that at this time.

QUESTION: At what point can you say that?


HAUETER: I think it's going to take a lot more work and examination.

HAUETER: We've got a lot of data here. We have to verify a lot of data, which has to match all the data with what we have. It's going to take quite some time.

QUESTION: Can we clarify this one last time? You have an initial wake encounter which you have no measurements for the lateral acceleration.

HAUETER: We do have measurements for it, yes.

QUESTION: We haven't been given them.

HAUETER: That's correct. I don't have them here.

QUESTION: Then we have a second episode which is a second encounter with a wake, and that's when you also have coincidentally the rudder movements and the measurements that you gave us.

HAUETER: No, the rudder movement I believe happens after that second encounter.

QUESTION: Well, what about these figures -- 0.3, 0.4 and 0.3? Is that during the encounter or just after it?

HAUETER: Just slightly after.

QUESTION: After the second encounter.

HAUETER: Yes. There are eight seconds here, and we haven't provided a great deal of timing. We're still working on it.

QUESTION: And, Tom, the rudder movements come after the second encounter, correct?

HAUETER: That's correct.

QUESTION: What sort of G measurements did you get in the Atlantic City tests?

HAUETER: I haven't looked at those to compare the data yet.

QUESTION: Can you tell whether the rudder movements were pilot- induced? Or were they...

HAUETER: We'll have to be working on that. QUESTION: You're not giving us the measurements on the G forces in the first wake encounter because they're not significant or because you haven't gotten them?

QUESTION: Or the second one?

HAUETER: We have not given the G loads from both. I don't have them with me yet. We are still working on that. But we're saying that they're consistent with a vortex encounter.

WOODRUFF: We are listening to National Transportation Safety Board officials, Chairwoman Marion Blakey and a board member there with her talking about the encounter that the American Airlines plane had, that Flight 587 had with the wake of an airplane in front of it, a Japan Airlines flight that took off about two minutes ahead of it.

And joining me now is aviation expert and aviation reporter Jim McKenna.

Jim, you are obviously much more expert on what they are talking about. What do you take away from these descriptions, this data that Marion Blakey was reading?

JIM MCKENNA, AVIATION ANALYST: Well, it's good, solid, very interesting data. Let's try and go through it a little bit to understand it.

What they are saying, if I recall, the length of the cockpit voice recorder tape was about one minute, 40 seconds. The flight data recorder tape lasted about 120 seconds. So, about 92 seconds into the flight, this airplane encountered some turbulence that they think is consistent with the wake from the 747 that was ahead of it.

The first encounter with the wake turbulence appears to be relatively benign, because they indicate that there is no change in the attitude of the airplane. A severe wake encounter would typically get some roll in the airplane and possibly a pitch up and pitch down of the nose. That doesn't appear to happen. Twenty second later, there is another encounter that they say is consistent with encountering wake turbulence.

WOODRUFF: Now, this is after the plane has accelerated. Is that right?

MCKENNA: The acceleration refers to the pushing of the airplane side to side and back and forth. So it's a scientific use of the term acceleration, as opposed to what we would think of.

WOODRUFF: It doesn't mean speeding up on purpose.

MCKENNA: It doesn't mean speeding up, right.


MCKENNA: And in the second -- the second encounter appears to be more severe. And I'll explain why that may be significant in a second. But in the second encounter, you get some significant pushing of the airplane sideways, what the investigator said was very significant: two pushes in one direction at least twice, or maybe -- at least twice the force of what you would normally feel in turbulence in an airplane -- and then a push of about the same magnitude in the opposite direction.

Interestingly, after those accelerations sideways, the flight data recorder stops tracking the movement of the rudder on the airplane. The rudder is the device that steers the airplane from side to side. And they say that the rudder indications become unreliable at that point.

And shortly after that encounter, we start to see a severe roll in the airplane, greater than 25 degrees, and rapid, very rapid turning of the airplane to the left, 10 degrees per second, which means they would go through two full circles in a minute at that speed. That's...

WOODRUFF: So this is pointing in the direction, Jim -- we are going to have to wrap up here -- but this is pointing in the direction of something that they are going to want to look very closely at.

MCKENNA: Very closely at.

WOODRUFF: And that is whether this wake turbulence caused enough shaking, putting this plane in a position where it was just not controllable by the pilots.

MCKENNA: That's true, yes.

WOODRUFF: All right, Jim McKenna, aviation reporter and expert joining us today, thank you very much.

MCKENNA: Thank you.

WOODRUFF: Appreciate it.




Back to the top