More woes at SFO - transposing runway numbers leaves little room for error

[WillFlyForCheese]

“In 2017, a commercial airliner lined up for takeoff at San Francisco International Airport on runway 01 Left, the main departure route.

The pilot accidentally punched 10 Left — a much longer SFO runway — into the cockpit computer, causing the plane to incorrectly calculate the appropriate thrust and wing flap settings.

The pilot’s simple reversing of the number caused the plane to nearly run out of runway, lifting off with only 400 feet left of asphalt, according to a Federal Aviation Administration report obtained by The Chronicle through the Freedom of Information Act.”

More here:

https://www.sfchronicle.com/bayarea/amp/FAA-safety-report-reveals-cases-of-pilots-mixing-14436212.php?__twitter_impression=true

[hans brinker]

“In 2017, a commercial airliner lined up for takeoff at San Francisco International Airport on runway 01 Left, the main departure route.

The pilot accidentally punched 10 Left — a much longer SFO runway — into the cockpit computer, causing the plane to incorrectly calculate the appropriate thrust and wing flap settings.

The pilot’s simple reversing of the number caused the plane to nearly run out of runway, lifting off with only 400 feet left of asphalt, according to a Federal Aviation Administration report obtained by The Chronicle through the Freedom of Information Act.”

More here:

https://www.sfchronicle.com/bayarea/amp/FAA-safety-report-reveals-cases-of-pilots-mixing-14436212.php?__twitter_impression=true


"A typical commercial airliner is traveling at 184 mph at liftoff,"

All I needed to read. Pretty hard to find a well written aviation story.

[DaveReidUK]

"A typical commercial airliner is traveling at 184 mph at liftoff,"

All I needed to read. Pretty hard to find a well written aviation story.

You mean that to most journalists, a knot is what you do with your shoelaces? :O

[hans brinker]

You mean that to most journalists, a knot is what you do with your shoelaces? :O

Pretty much. So frustrating to read a well meant article, so full of elementary mistakes.

[AerocatS2A]

This bit is worse:

​​​​​​​Aviation experts say airliners need to lift off the ground with enough runway left to abort a takeoff — 400 feet isn’t nearly enough and 1,000 feet is too close.

Bollocks.

[Obama57]

Way back when, PanAm wasn't as lucky; they put the approach lights through the bottom of the fuselage while rotating. No real drama until they had to land with wing gear only (B747) on the subsequent return. However, when the aircraft came to a stop, it sat back on its tail and the evacuation went sideways when pax tried to exit the upper deck slide which was not close to the ground.

[Airbubba]

Way back when, PanAm wasn't as lucky; they put the approach lights through the bottom of the fuselage while rotating. No real drama until they had to land with wing gear only (B747) on the subsequent return. However, when the aircraft came to a stop, it sat back on its tail and the evacuation went sideways when pax tried to exit the upper deck slide which was not close to the ground.

More on the Pan Am SFO mishap here:

https://en.wikipedia.org/wiki/Pan_Am_Flight_845

[RogueRivered]

"A typical commercial airliner is traveling at 184 mph at liftoff,"

All I needed to read. Pretty hard to find a well written aviation story.
What's wrong with it? That's about 160 knots. It was nice of the journalist to convert to a unit of measurement that the general public can understand.

[Capn Bloggs]

I regularly get airborne around 184. Griping about 400ft/1000ft is pretty petty too, compared to what actually happened.

[hans brinker]

What's wrong with it? That's about 160 knots. It was nice of the journalist to convert to a unit of measurement that the general public can understand.
V2 speed on just my narrow body will vary from about 120kts to 160kts depending on weight. Don't have a clue for anybody else, so saying "184" for a "typical" speed suggests a level of significance that is totally inappropriate.

[kiwi grey]

V2 speed on just my narrow body will vary from about 120kts to 160kts depending on weight. Don't have a clue for anybody else, so saying "184" for a "typical" speed suggests a level of significance that is totally inappropriate.

I'm afraid this is just another case of innumeracy.
The reporter read "160kts" and maybe knew what that was and decided to convert it to 'reader familiar' units, or more likely didn't know what "160kts" was, converted it for their own edification. They put "160kts as mph" into Google's search box, and they were told "184.124712 miles per hour". They were numerate enough to put that as "184mph", but not numerate enough to ask themselves whether "160kts" is intended to be an exact measurement, and if not whether "185mph" or even "180mph" would be suitable.

You see the same sort of thing even in reputable industry magazines nowadays: for example an aircraft's range is quoted as "7,500nm (13890km)" which is arithmetically accurate but completely bonkers - anybody who thinks a long-range aircraft has its range estimated to +/- 10km shouldn't be reporting on aviation.

Folks, even "aviation journalists", don't understand numbers and trust Google far too much.

[Capn Bloggs]

Who cares Hans? They took off on the "wrong" runway! Don't shoot the messenger.

[farefield]

Could this problem be compounded by the habit of saying, say, one left, instead of zero one left?

[Airbanda]

Curious that, if miss was as near as report suggests, it's surprising it seems to have only been reported via an anonymous report (US equivalent of CHIRPS?). Appreciate the report messes up explanation of how ASDA/V1 should work but on face consequence could have been serious.

It's similar to several incidents in UK (eg Sunwing at Belfast) where mistakes with temperature, weight or full length v intersection departure have lead to flex-thrust being wrongly set. Several examples in AAIB report/bulletins over the years.

That 01/10 confusion is possible is one thing and worrying; that initial mistake was not detected in cross/check is more so.

[Maninthebar]

I appreciate that airborne systems need a degree of safety, development and security not applicable to domestic devices BUT my car knows:

how heavy it is (and adjusts brake pressure accordingly)
what the exterior temperature is
where it is

It seems to be time that HAL caught up

[AerocatS2A]

Your car doesn't know what the temperature is. Your car knows what the temperature at the sensor is, but that's not the same thing at all, especially if it's been sitting in the sun on the same piece of tarmac for a while.

[Maninthebar]

Your car doesn't know what the temperature is. Your car knows what the temperature at the sensor is, but that's not the same thing at all, especially if it's been sitting in the sun on the same piece of tarmac for a while.

Fair point and perhaps I was being facile, but absolutely soluble at machine level. If it was important to my car then it would probably take the data from some 'nearest weather station'. METAR are already expressed in a form interpretable by a very simple machine. I am expressing lack of understanding as to why this ability is not used in performance calcs, as a check on human performance if nothing else.

[WingSlinger]

You mean that to most journalists, a knot is what you do with your shoelaces? :O
No, they wouldn't know how to tie a knot. They have Velcro.

[Capn Bloggs]

That 01/10 confusion is possible is one thing and worrying; that initial mistake was not detected in cross/check is more so.
NAV display: check. ;)

[Ian W]

Fair point and perhaps I was being facile, but absolutely soluble at machine level. If it was important to my car then it would probably take the data from some 'nearest weather station'. METAR are already expressed in a form interpretable by a very simple machine. I am expressing lack of understanding as to why this ability is not used in performance calcs, as a check on human performance if nothing else.

What HAL could have known was that it was set up for runway 10 but it was lined up on runway 01. That kind of check could be useful as a second level of safety.

[Speed of Sound]

What HAL could have known was that it was set up for runway 10 but it was lined up on runway 01. That kind of check could be useful as a second level of safety.

But this would also flag up a conflict while inputting runway data at any other position other than being lined up on the correct runway. Another software check would be to include an ‘Are you sure?’ prompt, when it detects another runway at a particular airport which could be mistaken for the desired one.

Out of interest, how many 01 and 10 runways exist worldwide?

[DaveReidUK]

Out of interest, how many 01 and 10 runways exist worldwide?

Possibly not that many.

But when you add other potentially confusing combinations (03 and 30, 12 and 21) as well, there are probably a fair few.

[DaveReidUK]

The candidates I've been able to find (discounting runways shorter than 6000'):

Airports with both a 01 and 10: Albany, Presque Isle and Savannah.

With both a 03 and 30 (and therefore also a 12 and 21): Albuquerque, Appleton, Artesia, Tel Aviv/Ben Gurion, Prestwick, Laramie

[Ian W]

But this would also flag up a conflict while inputting runway data at any other position other than being lined up on the correct runway. Another software check would be to include an ‘Are you sure?’ prompt, when it detects another runway at a particular airport which could be mistaken for the desired one.

Out of interest, how many 01 and 10 runways exist worldwide?


If the automation identifies that the aircraft is lined up on a different runway to the one that you have calculated for then it should provide a warning. If you are on a taxiway the alert should not be given if you have entered a runway to taxi to another (this is common at some airports) then you would get the warning and cancel it not a real problem.. If I remember correctly this type of warning was also proposed after the Comair crash taking off from the incorrect runway at Lexington Blue Grass. While confusing numbers don't help they are not necessary, worldwide departures from incorrect runways are unfortunately quite common.

[Lonewolf_50]

That 01/10 confusion is possible is one thing and worrying; that initial mistake was not detected in cross/check is more so. Over dependency on automation would seem to be a contributor here.
And yes, I have when using varying tactical and nav displays now and again fat fingered it and hit the wrong key/button.

[Water pilot]

Perhaps for a cross check the computer should require the pilot to enter the runway length as well. As long as humans are in the loop we will screw this sort of thing up. We are not yet at the point where you just touch "Chicago" on a map and the computer figures out all the rest for you (by talking to the ATC computer and the ground control computer and the corporate computer....)

[Ascend Charlie]

Things will be SO much easier when we hop into our personal Ooober flying car and just talk to the computer to tell it where we want to go. Then we sit back and play with our smut-phones while HAL does the thinking.

[FlightDetent]

But when you add other potentially confusing combinations (03 and 30, 12 and 21) as well, there are probably a fair few. Dave, possibly the pattern is 1 to 10, 2 to 20 and 3 to 30. The leading zero is omitted in the US.

Worse yet, I can imagine a little light reflection on the screen obscuring the readability between 1L and 10; add a fair dose of presbyopia and you get my drift. What about 1l and 11, not nice.

Which SW were they using?

[Capn Bloggs]

But this would also flag up a conflict while inputting runway data at any other position other than being lined up on the correct runway.
No, just add the runway check to the throttles/config check. When the throttles go up, system does a check of the actual position verses FMS-loaded runway.

As I said before: just add a check of the ND when you arm NAV before takeoff. It will stand out like dog's balls if you are not using the runway/SID you put in the box, or check the runway on the ND as you start rolling.

[Uplinker]

Could this problem be compounded by the habit of saying, say, one left, instead of zero one left?

Yes !

Calling and thinking about a runway by only its non-zero digit might lead to confusion - doing so gets one into the habit of ignoring the zero and just concentrating on the other number. This is dangerous because you then lose the crucial distinction between, say, 01 and 10.

Voicing the zero adds safety because it reinforces the image of the number in the mind as one looks at the MCDU or NAV display - of both the digits instead of just the non-zero digit.

Pilots in certain places in the world almost routinely abbreviate RT comms and use non standard phrases - presumably to sound slick? - but in my experience this practice often adds confusion, and the irony is that non standard messages have to be repeated much more often than in airspace where standard phraseology is more fully adhered to.

[DaveReidUK]

Calling and thinking about a runway by only its non-zero digit might lead to confusion - doing so gets one into the habit of ignoring the zero and just concentrating on the other number. This is dangerous because you then lose the crucial distinction between, say, 01 and 10.
Voicing the zero adds safety because it reinforces the image of the number in the mind as one looks at the MCDU or NAV display - of both the digits instead of just the non-zero digit.
Yes, it's compounded by the fact that, as mentioned above, long-standing US practice is not to paint the leading zero on the threshold marking, in contravention of ICAO SARPS (Annex 14), which specifies that runway heading markings should always include two digits.

[aterpster]

"A typical commercial airliner is traveling at 184 mph at liftoff,"

All I needed to read. Pretty hard to find a well written aviation story.
That's 160 knots.

[OldLurker]

The candidates I've been able to find (discounting runways shorter than 6000'):

Airports with both a 01 and 10: Albany, Presque Isle and Savannah.

With both a 03 and 30 (and therefore also a 12 and 21): Albuquerque, Appleton, Artesia, Tel Aviv/Ben Gurion, Prestwick, Laramie02 and 20 are reciprocal, so there are plenty of those (also 13/31).

[DaveReidUK]

02 and 20 are reciprocal, so there are plenty of those (also 13/31).

Well yes, that's why I didn't include those.

Unlike the other examples, confusing 02 with 20 or 13 with 31 isn't typically going to make much difference to the TORA, though the resulting wind might come as a bit of a surprise. :O

[Speed of Sound]

Well yes, that's why I didn't include those.

Unlike the other examples, confusing 02 with 20 or 13 with 31 isn't typically going to make much difference to the TORA, though the resulting wind might come as a bit of a surprise. :O

Unless of course they are parallel runways of differing length, although the ‘L’, ‘C’, or ‘R’ should then be an added clue

[FlightDetent]

2R and 20R, the letter does not really give a strong warning.

[hans brinker]

That's 160 knots.

And there is no typical airliner liftoff speed.

[Check Airman]

And there is no typical airliner liftoff speed.

I’m pretty sure the journalist was just trying to give the reader some general idea of the speed involved. “How fast are we going at takeoff/landing?” is a fairly common question. 160kt is a reasonable enough answer without getting into a lesson on runway performance.

Regarding the incident, my airline has had crews misread performance data for 01 and 10 at SFO. I assume it’s got to do with the way our brain processes information. You’re much more likely to confuse 01 and 10 than 08 and 18.

Similar situation with 13/31. Even more so since it’s the same bit of pavement, and your mental picture of the airport may actually add to the confusion. “Did ATC say 13L or 31L?”

Always have to look out for that at JFK, for example.

[hans brinker]

I’m pretty sure the journalist was just trying to give the reader some general idea of the speed involved. “How fast are we going at takeoff/landing?” is a fairly common question. 160kt is a reasonable enough answer without getting into a lesson on runway performance.

Regarding the incident, my airline has had crews misread performance data for 01 and 10 at SFO. I assume it’s got to do with the way our brain processes information. You’re much more likely to confuse 01 and 10 than 08 and 18.

Similar situation with 13/31. Even more so since it’s the same bit of pavement, and your mental picture of the airport may actually add to the confusion. “Did ATC say 13L or 31L?”

Always have to look out for that at JFK, for example.

I am sure you're right, it just pushed a few of my buttons. Saying a typical speed is 160kts means somewhere between 155 and 165. Still way too narrow of a gap. Saying 184mph (to me) means somewhere between 183 and and 185 mph, obviously too narrow of an band. We now confirm the physical runway against the FMS at least twice, I am sure there will still be some events where people get it wrong (hope it's not me...).