Qatar Airways B77W hit airport approach lights on departure, flight continued

[Gabriel]

Preliminary report out:

http://www.caa.gov.qa/sites/default/files/Preliminary%20report%20QR778%20Miami_v3.pdf

To avoid a post of Gabrielian length, I will let you read the report by yourself. It is short and probably shorter than what my post would be.

I will just say that it is very interesting, involving procedures design and compliance, CRM, unfriendly user interfaces, loss of situational awareness and, yes, fundamental airmanship.

Long time that I didn't see a case with so many factors interacting with each other and more or less equally affecting the outcome.

Interestingly, wrong take-off performance calculation was not a factor, contrary to what Leftseat suspected.

[Leftseat86]

Preliminary report out:

http://www.caa.gov.qa/sites/default/files/Preliminary%20report%20QR778%20Miami_v3.pdf

To avoid a post of Gabrielian length, I will let you read the report by yourself. It is short and probably shorter than what my post would be.

I will just say that it is very interesting, involving procedures design and compliance, CRM, unfriendly user interfaces, loss of situational awareness and, yes, fundamental airmanship.

Long time that I didn't see a case with so many factors interacting with each other and more or less equally affecting the outcome.

Interestingly, wrong take-off performance calculation was not a factor, contrary to what Leftseat suspected.

Well, technically they did use incorrect performance calculations because they had done the calculation for the full length, not the intersection, even though the printed information made it seem like T1 was acceptable.

The operating crew calculated the take-off performance figures for Runway 09, using the On-board Performance
Tool (OPT) contained within the Electronic Flight Bag (EFB) using a take-off weight of 342,000 kg. The OPT
offered the crew only one option for Runway 09, which was ‘09#T1’ as displayed by the OPT. This was understood
by the crews to mean Runway 09 full length, although the performance data had been pre-modified by a temporary
NOTAM. The OPT also displayed the information that intersection departures were not permissible for this
runway. An optimum performance take-off thrust calculation was selected on the OPT which generated an assumed
temperature take-off thrust of 36°C. The commander printed off this information, for his reference, and his first
officer wrote ‘09/(T1)#’ on his personal notepad.

[Evan]

The OPT offered the crew only one option for Runway 09, which was ‘09#T1’ as displayed by the OPT . This was understood by the crews to mean Runway 09 full length, although the performance data had been pre - modified by a temporary NOTAM.

I don't get this. Doesn't 09#T1 mean departure from the T1 intersection? If so, how could the OPT give them "correct" perfomance data when the available runway from T1 is 1000ft shorter than the takeoff distance requirement for the GTW? I mean a 773ER at 330K kg cannot take off in 2,610m at any temperature, no matter what the thrust setting is.


And don't thresholds have big fancy numbers and piano key looking things?

[elaw]

Hopefully the full report will shed some light on that because (as an uneducated outsider) I'd have thought exactly the same thing.

Re the numbers and markings, it sounds like what may have happened is the crew thought that as the result of a NOTAM (ignoring for a moment the fact the takeoff performance sheet said "no NOTAM"), the "full length" of the runway was considered as starting at the T1 intersection. They knew they were at the intersection and therefore did not expect to see runway-end markings.

It seems like one thing that might (or might not) have helped in this situation would be if the performance calc printout showed the runway length. If the sheet actually said "13,016' length", they might have noticed that's the full runway length depicted on the chart and realized they'd have less runway available when departing from an intersection.

[Gabriel]

I don't get this. Doesn't 09#T1 mean departure from the T1 intersection? If so, how could the OPT give them "correct" perfomance data when the available runway from T1 is 1000ft shorter than the takeoff distance requirement for the GTW? I mean a 773ER at 330K kg cannot take off in 2,610m at any temperature, no matter what the thrust setting is.


And don't thresholds have big fancy numbers and piano key looking things?

Make that 1000m shorter, not 1000ft.

That was my first thought. The preliminary report doesn't explain the logic behind the output of the on-board performance calculator, but 09#T1 sounds to "similar" to 09 at intersection T1 to me. However, according to the report, apparently all the crew (including the relief crew) was initially aware that they would be using the full length, but since the time that they run the performance to the time of the actual take off (perhaps 1/2 an hour?) that common "knowledge" seems to have diluted, the Captain (himself not knowing exactly why) made the decision to change to T1, the decision felt increasingly acceptable given the confusing picture in the EFB where T1 seemed to be almost at the end of the runway, given the sight of an airplane landing just in front of them (the landing threshold is displaced), given the 09#T1 printed in the performance calculation report, given the acceptance of the FO to the plan change after seeing that he had written 09#T1 in his notepad, and after the relief crew interpreted the captain's seeking reassurance from them regarding the departure from T1 as giving reassurance that it was ok.

The fact is that, we don't yet understand what 09#T1 means, but they all understood initially that it would be a full-length take-off and the performance calculation displayed in the screen (not the printed one) included "No intersection take-off approved for this runway". So it was not a wrong take-off performance calculation but an wrong execution.

Regarding the piano keys, they knowingly took off from an intersection, so they would not expect seeing the threshold markings.

[Evan]

Make that 1000m shorter, not 1000ft.

Of course. Meters. Typo.

The fact is that, we don't yet understand what 09#T1 means...

No, but I suspect the crew of a 777 certified to fly it should know. This makes me wonder if we might have a non-standard teminology issue here. There is just no way that thing can spit out a performance calculation for the T1 intersection.
And maybe we shouldn't be tossing out the Jeppesen charts just yet. Scrolling things are a stealth factor.

Regarding the piano keys, they knowingly took off from an intersection, so they would not expect seeing the threshold markings.

Ah, right.

[Evan]

I'm thinking 09#T1 was a landing option for 09 in the database that was somehow entered as a takeoff calculation.

I'm also thinking the report may have been monkeyed with by Qatari government censors, due to the presence (in a different font) of this bit of disclaimer:

The sole objective of this investigation is the prevention of future accidents and incidents. It is not the purpose of this investigation to apportion blame or liability. Accordingly, it would be inappropriate for this report to be used to assign fault or blame or determine liability, since neither the investigation nor the reporting process has been undertaken for that purpose. Extracts may be published without specific permission providing that the source is duly acknowledged, the material is reproduced accurately and is not used in a derogatory manner or in a misleading context.

[3WE]

I don't get this. Doesn't 09#T1 mean departure from the T1 intersection? If so, how could the OPT give them "correct" perfomance data when the available runway from T1 is 1000ft shorter than the takeoff distance requirement for the GTW? I mean a 773ER at 330K kg cannot take off in 2,610m at any temperature, no matter what the thrust setting is.


And don't thresholds have big fancy numbers and piano key looking things?

So many acronyms... and yet there are these little square signs off the side of most runways these days with a number on them, and Dios forbid we say the required takeoff distance and then think about what's showing on those signs... nope, too many procedures and acronyms to keep straight - including V-3BS where you can still abort, but not run off the end too bad.

[Gabriel]

This part of the report is really cryptic:

The OPT offered the crew only one option for Runway 09, which was ‘09#T1’ as displayed by the OPT. This was understood by the crews to mean Runway 09 full length, although the performance data had been pre-modified by a temporary NOTAM. [...] The commander printed off this information [...] The printed information contained no reference to the fact intersection departures were not permissible from this runway (Figure 3), and contained the message ‘No NOTAM data found’.

[Gabriel]

And now the fundamental airmanship part:

This was understoodby the crews to mean Runway 09 full length. The OPT also displayed the information that intersection departures were not permissible for thisrunway

Ok, se we have full length as mental picture.

An optimum performance take-off thrust calculation was selected on the OPT which generated an assumedtemperature take-off thrust of 36°C.

[Note: The actual temperature was 29°C.] Ok, this means that, since the full length of the runway was more than needed, they were going to do a reduced-thrust take-off to preserve fuel and engine life and to reduce CO2 and noise emissions. Valid.

As they taxied along S the commander decided that the aircraft could depart from the runway intersection T1. Hecould not recall why he made that decision

Ok, that's not clear, but for some reason the commander suddenly thought it might be a good idea to depart from the intersection. But...

As this was not what relief crew recalled had been briefed, they queried T1.

Ok, that's good CRM. Question the captain if in doubt.

The commander made a hand gestureand said something which he thought was seeking reassurance from the crew that everything was OK.

As a side note, that is still good CRM, or it would be if made in a clear way to avoid misunderstandings like the ones that happened in this case:

the relief crew interpreted the commander’s communicationas him confirming he was content with a T1 departure

So, to summarize, until here:
They interpreted the on-board performance tool information as "use full runway".
The captain later decided, for reasons even he cannot explain, that they could take off from T1.
But the relief crew questions that and the captain seeks reassurance from the crew.

I will go one step too far and assume that in the captain subconscious, somewhere in the quick-access part of his brain, there must have been some tension between the original idea of the full length take-off and the intersection take-off.

So the captain applies take-off thrust (the report doesn't mention it but I assume it was the reduced/optimized take-off thrust) and...

During the take-off roll, asthe aircraft approached V1 the crew became aware that something was not right

And?

and theyrecalled the aircraft entering the alternating red and white runway centreline lights which indicated they hadapproximately 900 m of runway remaining

And?

The commander assessed the speed of the aircraft, the rate of accelerationand the runway remaining and concluded the safest course of action was to continue.

The commander assessed the speed of the aircraft, the rate of accelerationand the runway remaining and concluded the safest course of action was to continue.
Ok, right or wrong, that's valid decision making. So now you decided to continue the take off, so now yes, right?

He recalled as he rotated theaircraft entered the red runway centreline lights (indicating only 300 m of runway remaining.)

AND!?!?!?!?

the aircraft made contact with the Runway 27 approachlights as it got airborne

Summarizing again:

There was some (hidden) tension in the captain's mind regarding the full length vs T1 take-off.
He knew they were taking off with a selected thrust that was less than available.
He knew during the TO roll that there was something wrong.
He noted the airplane entering the red/white lights zone, indicating last 900m of runway
He noted the plane entering the red lights zone, indicating last 300m of runway...

AND HE DID NOT THINK TO MAYBE ADVANCE THE THROTTLES TO MAX THRUST?????

I would not be surprised if a pilot, under these circumstances, with the concern of not lifting-off by the end of the runway, would advance the throttles with enough care and momentarily exceed some engine limit. But not advancing the throttles?

[3WE]

And Gabriel.

Whats wrong with determining:

1. Our takeoff roll should be 9,285.13 feet at this weight, temperature, wind and power setting...

2. We have 9,185.13 feet available from this intersection as determined.by diagrams and the distance remaining signs.

3. Confirm if value 1 and value 2 work together in the best interest of the approach lights.

But no (and again).were focusing on computers and decision trees to reduce CO emissions.and the Boy Scout CRM handbook and cryptic acronyms, and mundane stuff, while glossing over old fashioned important stuff.

[Evan]

And now the fundamental airmanship part...

Gabriel, an ambiguous hand gesture is TERRIBLE CRM. A practiced hand gesture to reinforce an action, like pointing at an A/P setting after setting it, is one thing but a mere hand gesture as a query for a takeoff decision? And the failure of the relief crew to speak up anyway stinks of bad culture to me. Again, I think this is a case of bad culture and non-standard terminologies.

Also, They managed to get off the ground at 330K kg in 2,610m + the overrun. I'm guessing that WAS full throttle. And if you look at the dotted black line indicating "takeoff roll" on the FDR pitch attitude plot, the angle seems to exceed 10° on the ground. AFAIK, the 773ER has a tailstrike pitch attitude of 10° so the PF seems to be giving it all it's got at that point.

And lest we forget, the failure of the crew to request an immediate ground inspection after nearly plowing down the approach lights must be mentioned here. They were quite aware of the POSSIBILITY of some contact. That should be enough to require this.

[elaw]

Whats wrong with determining:

1. Our takeoff roll should be 9,285.13 feet at this weight, temperature, wind and power setting...

2. We have 9,185.13 feet available from this intersection as determined.by diagrams and the distance remaining signs.

3. Confirm if value 1 and value 2 work together in the best interest of the approach lights.

But no (and again).were focusing on computers and decision trees to reduce CO emissions.and the Boy Scout CRM handbook and cryptic acronyms, and mundane stuff, while glossing over old fashioned important stuff.

Dude... seriously?

What you posted *is* a decision tree. And computers are really really good at decision trees if supplied with correct information ("garbage in, garbage out") and used properly.

I'll give you that the UI on the computer may not have been absolutely the best, but to blame this accident on the simple fact the pilots were using a computer IMHO is improper. The computer presented certain information and not other information, and the pilots in their head concluded from that information and other info they'd gathered that 2+2=7 and that "7" was enough distance to take off. Which it probably was, but "4" was not.

[3WE]

Dude... seriously?

...to blame this accident on the simple fact the pilots were using a computer IMHO is improper...

Yes, seriously...

BUT, we're in that gray area of turning it ALL over to the computer and putting our minds on autopilot.

Because the computer is so good, no one looked to see IN OLD FASHIONED TERMS if the available runway matched the needed runway...

Subtle stuff in the realm of whether the computer should watch the humans or vice versa or where is the middle ground.

The pilots were managing the computer and not managing whether the runway length worked.

"Procedures TO THE DETRIMENT OF import basic stuff".

Concur or disconcur?

[elaw]

I sort of agree and sort of disagree. Geez, I should be a politician!

The computer never explicitly said "you have enough runway to take off from your current position". Nor did it ever say "you do not have enough runway to take off from your current position". It was up to the carbon-based life forms to determine whether they had enough runway given the rather confusing info the computer was giving them. But instead of acknowledging the fact the info was confusing and trying to clarify it, they ASSumed things were the way they wanted them to be, firewalled the throttles, and headed off into the wild blue yonder. Among other factors, it seems like one of Gabriel's favorite things, "confirmation bias", may have been a big player here.

So in terms of taking care of "basic stuff", there is little stuff more basic than ensuring there's enough runway in front of you for your aircraft to safely depart. Did the crew attempt to do that? I think the simple answer is "yes", but they made a mistake in judgement in not realizing the info they were presented with could be interpreted several different ways, and not pursuing the correct interpretation.