Air Canada approached taxiway at SFO

1123heavy

Amazing that things like this are still occurring. Just goes to show that we can never, ever, be too sure of our business in this game. No matter what technological advancements we make.

Air Canada A320 approaching RWY 28R inadvertently lines up for the parallel taxiway C. There were 4 aircraft lined up on C at the time.

Audio available here from 25:40 onwards: http://archive-server.liveatc.net/ksfo/KSFO-Twr-Jul-08-2017-0630Z.mp3

Transcript:

"AC759: "Tower, just want to confirm. This is Air Canada 759. We see lights on the runway there. Across the runway. Can you confirm are we cleared to land?"

SFO TOWER: "Confirmed cleared to land. Runway 28 Right. There's no one on 2-8 Right but you."

"Where's this guy going? He's on the taxiway," the other pilot said.

SFO TOWER: "Air Canada, go around."

AC759: "Going around. Air Canada 759."

SFO TOWER: Air Canada looks like you were lined up for Charlie there. Fly heading 280. Climb maintain 3,000."

AC759: "Heading 2-8-0, 3,000 Air Canada 759."

UA001 PILOT: "United One, Air Canada just flew directly over us."

SFO TOWER: "Yeah, I saw that.""

My initial suspicions were that perhaps there was an error in the aircraft's IRS, telling the crew they were in line with the RWY when in fact it was the taxiway. The AC guys themselves were the first to notice something was amiss anyway.

This image below shows they were dead on taxiway C. Seems peculiar.

images.jpg

faoiarvok

1123heavy wrote: »

My initial suspicions were that perhaps there was an error in the aircraft's IRS, telling the crew they were in line with the RWY when in fact it was the taxiway. The AC guys themselves were the first to notice something was amiss anyway.

They were on a visual approach, so may not have been referring to nav instruments at all at this stage.

Growler!!!

Absolutely no reason to be doing a visual approach in a busy enviornment like that. Confirmation bias of seeing what they expect to see not what is actually there.

smurfjed

This is scary, why were they doing a visual approach without backup?

duskyjoe

The ILS has just in recent days gone back on the air for Rwy 28R......having said that they had the option of an overlay rnav appt as back up while executing a visual. From memory the approach lights are working fine for 28R but not for 28L......how air Canada lined up for C taxiway is beyond me. Just for info NORCAL tracon fire aircraft at approaches so the Swiss cheese model may have come into play with a rushed approach. Just my 2c

tricky D

I wonder is there any merit in making C slightly crooked like F, though F is considerably shorter so much less likely to confuse.

Also the fire station and United hangar are pretty close to the taxiway and this proximity is fairly apparent while waiting to line up for 28R when visibility is good.

1123heavy

faoiarvok wrote: »

They were on a visual approach, so may not have been referring to nav instruments at all at this stage.

Well yes, but they were shortcut to the visual later on in the approach from what I believe. The FMC would have been programmed for an approach onto 28R regardless of whether they went visual or not late on ...

faoiarvok

1123heavy wrote: »

Well yes, but they were shortcut to the visual later on in the approach from what I believe. The FMC would have been programmed for an approach onto 28R regardless of whether they went visual or not late on ...

Ok, I may have misunderstood that section of the comms, thought they were on finals and thought that would mean they wouldn't be required to monitor nav instruments. Not a pilot, totally open to correction.

Even if they were flying purely visually, you would imagine that different colour lights on taxiways/runways, approach lights, PAPIs, etc should make it pretty clear which one is the runway?! Though I did see it reported elsewhere that 28L was closed, which may have contributed to confusion.

1123heavy

http://www.ctvnews.ca/video?clipId=1165295&playlistId=1.3497833&binId=1.810401&playlistPageNum=1

I think this is an extremely plausible suggestion and it is amazing how all of the facts (offset approach, pilots not used to even being on C and so not aware it is normal for aircraft to be in line with it when landing on 28R) can give a totally different perspective.

ElNino

Good explanation but I am not sure many people will believe what he said about there being no chance that the plane could have landed on the taxiway.

There was another possible contributing factor that he didn't mention which was that 28L was out of service and the lights were turned off. So the pilots may have visually thought that 28R was actually 28L and that they were line up with 28R to its right which was actually the taxiway.

circular flexing

http://avherald.com/h?article=4ab79f58&opt=0

On Jul 11th 2017 the Canadian TSB reported the crew of C-FKCK asked ATC to confirm landing clearance as they were seeing lights when the aircraft was 0.6nm before the runway threshold. The controller was coordinating with another facility when a flight crew of another airliner taxiing on taxiway C queried ATC where AC-759 was going and stated it appeared the aircraft was lined up with taxiway C. AC-759 had already overflown taxiway C by about 0.25nm when ATC instructed the aircraft to go around. 4 aircraft were on taxiway C at the time of the occurrence. It is estimated that AC-759 overflew the first two aircraft by 100 feet, the third by about 200 feet and the last by 300 feet. The closest lateral proximity between AC-759 and one of the aircraft on taxiway C was 29 feet. The NTSB is investigating the occurrence.

1123heavy

circular flexing wrote: »

http://avherald.com/h?article=4ab79f58&opt=0

Close.

plodder

I saw a post somewhere else that made a good point. There wasn't any realistic chance that this flight would have tried to land on top of four planes ready to take off. But what if the taxiway was empty at the time? The pilot wouldn't have asked what the lights were, and might have just continued with the landing, and then who knows what might have happened?

Buffman

Some videos showing night approaches to SFO 28R to give an idea of what AC759 may have seen (minus the waiting aircraft on C) . Taxiway C is the green 'line' to the right of Runway 28R.

https://www.youtube.com/watch?v=rNMtMYUGjnQ

https://www.youtube.com/watch?v=MnzNvhQxu90

plodder wrote: »

There wasn't any realistic chance that this flight would have tried to land on top of four planes ready to take off.

They're estimating 100 ft (30.48 m) vertical clearance of the first two aircraft. To give that some perspective, those two aircraft (B789 & A343) have a height of 55 ft 10 in (17.02 m) and 55 ft 6 in (16.91 m) respectively.

USAir Flight 1493 would be a past example of a landing aircraft colliding with an aircraft waiting for takeoff.

plodder wrote: »

But what if the taxiway was empty at the time? The pilot wouldn't have asked what the lights were, and might have just continued with the landing, and then who knows what might have happened?

Presumably, if he landed on the taxiway centreline and kept it there, other than going through the stop bars (red 'traffic light' for pilots) protecting the intersecting runways 1L/19R and 1R/19L, it would have been a relatively 'normal' landing.

1123heavy

It is very hard to imagine how the two can be confused from the video above. Anyway aircraft landing on taxiways has happened plenty of times before and US carriers are no strangers to it ...

https://www.flightglobal.com/news/articles/continental-757-crew-misidentified-lights-before-taxiway-222721/

After its approach, the Continental 757 touched down on taxiway Z near the intersection of taxiway R. None of the 154 passengers and crew was injured.

http://avherald.com/h?article=42187f22 (Delta B763)

The approach and landing was accelerated, the crew received clearance to land on runway 27R in night conditions but good visibility, however touched down on taxiway M, which runs north of and parallel to runway 27R. The airplane slowed safely.

plodder

Buffman wrote: »

Some videos showing night approaches to SFO 28R to give an idea of what AC759 may have seen (minus the waiting aircraft on C) . Taxiway C is the green 'line' to the right of Runway 28R.

https://www.youtube.com/watch?v=rNMtMYUGjnQ

https://www.youtube.com/watch?v=MnzNvhQxu90

They're estimating 100 ft (30.48 m) vertical clearance of the first two aircraft. To give that some perspective, those two aircraft (B789 & A343) have a height of 55 ft 10 in (17.02 m) and 55 ft 6 in (16.91 m) respectively.

USAir Flight 1493 would be a past example of a landing aircraft colliding with an aircraft waiting for takeoff.



Presumably, if he landed on the taxiway centreline and kept it there, other than going through the stop bars (red 'traffic light' for pilots) protecting the intersecting runways 1L/19R and 1R/19L, it would have been a relatively 'normal' landing.

and what if another plane or planes entered the taxiway during the landing?

Buffman

plodder wrote: »

But what if the taxiway was empty at the time? The pilot wouldn't have asked what the lights were, and might have just continued with the landing, and then who knows what might have happened?

plodder wrote: »

and what if another plane or planes entered the taxiway during the landing?

If the taxiway became occupied after landing, the potential opportunity for avoiding action from one or both aircraft may be increased, but you'd still be risking another potential Tenerife.

A poster on another site mentioned a similar near miss in 1988 at Gatwick after a One-Eleven landed on the taxiway. The other aircraft on the taxiway tried to escape onto the grass but got stuck. Luckily the One-Eleven stopped with 190 m to spare.

https://www.gov.uk/aaib-reports/2-1989-bac-1-11-g-aywb-and-boeing-737-ei-biz-12-april-1988

ElNino

The NTSB has issued a new update and the near miss was even closer than anyone thought.

They say that the Air Canada plane flew as low as 59ft over the the first plane on the taxiway which was a 787. The 787's tail is 56ft tall so they came within 3 feet of causing the biggest accident in aviation history.

I can't see the 2 pilots having much of a future in aviation.

View Profile

ElNino wrote:

They say that the Air Canada plane flew as low as 59ft over the the first plane on the taxiway which was a 787. The 787's tail is

So they flew 59ft over the top of the tail then!

1123heavy

View Profile wrote: »

So they flew 59ft over the top of the tail then!

Tail is 56 ft. They reached 59ft

View Profile

Ouch! How the crew didn't visually recognise the aircraft sooner is amazing! Should have been clearly evident a mile out!

Buffman

ElNino wrote: »

The NTSB has issued a new update and the near miss was even closer than anyone thought.

That was close.:eek:

Pretty annoying that the CVR was allowed to be overwritten before it could be accessed.

From looking at that NTSB radar and video data, for once the tabloid melodrama may be justified.

Deep Thought

anyone on here actually flown as PIC into this airport? just wondering that their thoughts are as pilots

thomasj

Looking forward to seeing the episode of air crash investigation on this ....
.

View Profile

Would be a fairly short and boring episode IMHO.

tnegun

1123heavy wrote: »

Tail is 56 ft. They reached 59ft

Wonder what sort of angle would the aircraft of been at? The altitude is measured in the nose right?

thomasj

View Profile wrote:

Would be a fairly short and boring episode IMHO.

You'd be surprised what they could get out of the investigation findings , plus it's a Canadian production so I could imagine based on the level of local interest they'll do an episode

circular flexing

thomasj wrote: »

You'd be surprised what they could get out of the investigation findings , plus it's a Canadian production so I could imagine based on the level of local interest they'll do an episode

Do they do Air Crash Investigation for near misses?

thomasj

circular flexing wrote:

Do they do Air Crash Investigation for near misses?

They have done them for near misses before. Mainly because the investigation part was the more interesting .

thomasj

Obviously there's always more than 1 factor in a crash or nearmiss but I'll be interested to see what they say about ATC. With all modern technology these days it seemed like ATC was relying on the pilots views and vice versa. Surely ATC would have had radar etc That doesn't sound right.

duskyjoe

I am sure Norcal Tracon will get a mention in the NTSB final report. They are the controlling agency of all aircraft into SFO until 10nm from the threshold. SFO approaches are notorious for getting vectored onto final approach above the standard descent profiles to ensure separation on tight parallel approaches plus sequencing ......I wonder with all that in mind, at night and the ILS28R off the air that evening( I think) did all the holes line up on the Swiss cheese model with disorientation for the crew involved? Where they ended up circa 60ft above other aircraft on the taxiway below can only summarised as frightening.

KCAccidental

Saw there in the report that Air Philippines pilot threw on his landing lights as a last ditch effort to signal the Air Canada plane. Must have been some soiled jocks in a couple of cockpits!

smurfjed

The Philippine Airlines crew demonstrated excellent situational awareness, but it was also reported that the AC crew didn't see these lights.

The ILS was off the air, but why didn't they "follow the magenta line". From my limited exposure of flying with American pilots (yes, I know AC is Canadian), they love visual approaches even when not required, and yet that is the country that appears to have the highest number of landings at wrong airports, is there a correlation?
It also appears that the CVR was overwritten due to the amount of time between the incident and engine shut down, so we may never learn what actually happened in the cockpit.

plodder

duskyjoe wrote: »

I am sure Norcal Tracon will get a mention in the NTSB final report. They are the controlling agency of all aircraft into SFO until 10nm from the threshold. SFO approaches are notorious for getting vectored onto final approach above the standard descent profiles to ensure separation on tight parallel approaches plus sequencing ......I wonder with all that in mind, at night and the ILS28R off the air that evening( I think) did all the holes line up on the Swiss cheese model with disorientation for the crew involved? Where they ended up circa 60ft above other aircraft on the taxiway below can only summarised as frightening.

As a passenger flying in there, I often wonder how they manage it, with three big airports around the bay, and numerous other small airfields.

cuterob

Those AC pilots are getting the sack I'm sure after this

Look in this article at the photo from it, how did they not realise sooner is beyond me

http://abc7news.com/ntsb-say-air-canada-flight-at-sfo-came-just-feet-from-taxiway-/2270588/

plodder

cuterob wrote: »

http://abc7news.com/ntsb-say-air-canada-flight-at-sfo-came-just-feet-from-taxiway-/2270588/

abc7news wrote:

It's a very rare occurrence for pilots to actually lock onto a taxiway at night versus the runway because taxiway lights are blue and runway lights are white," said ABC News Aviation Analyst John Nance.

I wonder if the blue taxi-way lights could have been obscured by the lights off the taxiing aircraft? Also, the parallel runways are very close, which must limit the kind of approach lighting they can use. Though looking at the video above, there is a very bright strip pointing to the centerline of the runway. How could they have missed that?

According to the report, the AirCanda flight was so off course that it hit a blind spot in the airport's radar system for 12 seconds. The radar system should have alerted air traffic controllers the aircraft was not where it was supposed to be.

Blind spot in the radar? That wouldn't fill me with confidence..

smurfjed

Those AC pilots are getting the sack I'm sure after this

Do you think that they should be sacked?

plodder

I wonder did they think that the approach lights for 28R were off and they were looking at the lights for 28L?

Van.Bosch

plodder wrote: »

I wonder did they think that the approach lights for 28R were off and they were looking at the lights for 28L?

I saw that theory elsewhere - seems plausible but I would hope there are better systems than "I usually land on the second set of lights from the left"

plodder

Van.Bosch wrote: »

I saw that theory elsewhere - seems plausible but I would hope there are better systems than "I usually land on the second set of lights from the left"

I think it's more like what would the standard procedure be if the approach lights for the runway you want to land on (at night) are off? I'm sure pilots can tell us, would that be allowed at all? If it's allowed then surely it could lead to this happening and if it's not allowed why were they trying to do it?

1123heavy

Baylee Melodic Windjammer wrote: »

Hopefully the investigators won't have such a narrow minded view and will look at ensuring it doesn't happen again.

You see, as much as you may find the statement in the article unsettling, he isn't really wrong.

SMS is a system whereby if even one of those holes in the cheese can be shut off - it is deemed to have worked. We didn't have an accident - the system worked. Did we come mighty close? Absolutely, however, SMS philosophy now is to look at what happened and to take steps to ensure those particular steps are mitigated to the best of everyone's ability in future. However this does not mean something like this will never happen again, nor can anything ever guarantee that we will never have a particular event again. There are an unlimited amount of holes that can develop in every slice of cheese and at every point on it.

The key thing that SMS aims to achieve is as deep a safety philosophy as possible, however, it will never, ever be 100% accident-free. So telling the public that a particular thing won't ever reoccur would be an unfounded promise. Telling the public that we have implemented a system that ensures that the risks of any particular tragic incident are vastly mitigated is actually the truth.

Looking at this incident, we came mighty close again as mentioned. However, it did not happen. That is a success for the system. Now the system dictates that we look at the events that lead to this close occurrence, and what each organisation can do in order to stop the holes that developed earlier on in the cheese layers.

Buffman

thomasj wrote: »

Surely ATC would have had radar etc

o At 2355:46 PDT, when ACA759 was about 0.7 mile from the landing threshold and about 300 ft above ground level (agl), the flight crew contacted the ATC tower, mentioned seeing lights on the runway, and requested confirmation that the flight was cleared to land.

o As ACA759 approached SFO, at 2355:52 PDT, the airplane flew too far right of course to be observed by the local controller’s ASDE-X/ASSC and was not visible on the ASDE-X/ASSC display for about 12 seconds.

o At 2355:56 PDT, when ACA759 was about 0.3 mile from the landing threshold, the local controller confirmed and recleared ACA759 to land on runway 28R.

Without wanting to be a 'Captain Hindsight', a possible missed opportunity for ATC to call for an earlier go-around here.

plodder wrote: »

I wonder did they think that the approach lights for 28R were off and they were looking at the lights for 28L?

Van.Bosch wrote: »

I saw that theory elsewhere - seems plausible but I would hope there are better systems than "I usually land on the second set of lights from the left"

In postincident interviews, both incident pilots stated that, during their first approach, they believed the lighted runway on their left was 28L and that they were lined up for 28R. They also stated that they did not recall seeing aircraft on taxiway C but that something did not look right to them.

That seems to be their explanation.

smurfjed wrote: »

The Philippine Airlines crew demonstrated excellent situational awareness,

Ye, they did well, from looking at the photos the third plane further down the taxiway (a UAL 787) also switched on its lights.

smurfjed wrote: »

The ILS was off the air, but why didn't they "follow the magenta line". From my limited exposure of flying with American pilots (yes, I know AC is Canadian), they love visual approaches even when not required, and yet that is the country that appears to have the highest number of landings at wrong airports, is there a correlation?

Yep, there seems to be a number of avoidable incidents occuring with visual approaches, taxiway landings, wrong runways, wrong airports. Asiana Airlines Flight 214 was on a visual to 28L when it crashed.

1123heavy wrote: »

SMS is a system whereby if even one of those holes in the cheese can be shut off - it is deemed to have worked. We didn't have an accident - the system worked. Did we come mighty close? Absolutely,.......................

Looking at this incident, we came mighty close again as mentioned. However, it did not happen. That is a success for the system. Now the system dictates that we look at the events that lead to this close occurrence, and what each organisation can do in order to stop the holes that developed earlier on in the cheese layers.

I do get where you're coming from and agree that lessons have to learnt from this, however I think saying the system 'worked' in this case is lessening the seriousness of it. IMO this was beyond close, the system got very lucky.

The A320 went around from 85 ft agl, and reached 59 ft agl before climbing. If one of the aircraft it overflew was an A380 (79 ft high) instead of the lower 787s and A340, I don't think the system would have worked.

plodder

Buffman wrote: »

That seems to be their explanation.

Interesting. There's a simulation of the approach at the link below. Looks like some of the details aren't correct, but it gives a good sense of how they didn't start the go around until they were over the first a/c on the taxiway. So, they descended quite a bit further and the lowest point appears to be between the second and third in line. Scary stuff.

https://www.youtube.com/watch?v=ydUqfhNqUIc

smurfjed

I cant see how this can be an indication that the SMS worked, in fact i would see this as exactly the opposite that the system didn't work and only by fluke the crew elected to to a baulked landing due to a "feeling".

Buffman

plodder wrote: »

Interesting. There's a simulation of the approach at the link below. Looks like some of the details aren't correct, but it gives a good sense of how they didn't start the go around until they were over the first a/c on the taxiway. So, they descended quite a bit further and the lowest point appears to be between the second and third in line. Scary stuff.

https://www.youtube.com/watch?v=ydUqfhNqUIc

Ye, I saw that simulation before and yes, I think it's timings are slightly out in comparison to the NTSB data.

nsa0bupkd3948x

1123heavy wrote: »

You see, as much as you may find the statement in the article unsettling, he isn't really wrong.

SMS is a system whereby if even one of those holes in the cheese can be shut off - it is deemed to have worked. We didn't have an accident - the system worked.

Just because there wasn't an accident doesn't mean the system worked.

To me, a near miss being avoided as a result of a TCAS warning is evidence the system works.

An accident being avoided by non standard actions such as turning on landing lights to try and get the attention of a landing aircraft, or a pilot asking on ATC "Where's this guy going? He's on the taxiway" is not proof the system works.

Van.Bosch

plodder wrote: »

Interesting. There's a simulation of the approach at the link below. Looks like some of the details aren't correct, but it gives a good sense of how they didn't start the go around until they were over the first a/c on the taxiway. So, they descended quite a bit further and the lowest point appears to be between the second and third in line. Scary stuff.

https://www.youtube.com/watch?v=ydUqfhNqUIc

If the lowest point was between 2nd and 3rd, imagine being in the cockpit of the 3rd plane.

1123heavy

smurfjed wrote: »

I cant see how this can be an indication that the SMS worked, in fact i would see this as exactly the opposite that the system didn't work and only by fluke the crew elected to to a baulked landing due to a "feeling".

That's not quite what happened, far from. A balked landing is when the aircraft touches down and then lifts off again. This aircraft, whilst coming close, did not touch the ground and he wouldn't have even been in the flare stage at 60ft AGL.

Just because there wasn't an accident doesn't mean the system worked.

To me, a near miss being avoided as a result of a TCAS warning is evidence the system works.

An accident being avoided by non standard actions such as turning on landing lights to try and get the attention of a landing aircraft, or a pilot asking on ATC "Where's this guy going? He's on the taxiway" is not proof the system works.

You do not have access to what was happening on the flight deck and what systems were in place that prevented this.

I assume you are aware that an aspect of SMS may consist of something as simple as instructing crew that if they feel anything is off whatsoever they either query it or just go around. If you listen to the recording, the AC pilots themselves were the first ones to pick up something may have been wrong and so they queried it and took the necessary action to GA before being told to do so, and they copped something was wrong and prepared for a GA even before the United pilots said anything.

If that procedure was what saved the day, then the system worked. You may not like how close it came (no one does), but this is aviation. We have to learn from it and see how we can do our best to not get that close again.

The truth is that nobody here has any access to what was said, what procedures were in place by the airline that may have prevented this, even the crew themselves (they're receiving a fair beating but people are forgetting they were the first ones to cop their mistake).

Whilst it is sort of de rigeur for most forums to speculate and be judge and jury as well as the prosecuters when things like this happen, the fact of the matter is that we know nothing. So to be saying with such conviction the system did not work, when in reality we have no access as to exactly what it was that may have blocked that final slice of cheese is a bit baseless.

Wait until the report and then point fingers.

nsa0bupkd3948x

1123heavy wrote: »

Wait until the report and then point fingers.

I have not pointed any fingers and I haven't said categorically the system didn't work. My point was that just because an accident didn't happen, doesn't mean we can conclude the system worked.

I think we both agree there are obviously lessons to be learnt from this event.

I believe you cannot state the system is safe as Zwingle did just because there wasn't an accident. That comes across as very unprofessional. I don't think the AC pilots necessarily should never fly again after all pilots are humans and are fallible but we need to learn from their experience and not just say there was no accident, nothing to see here.

Buffman

Van.Bosch wrote: »

If the lowest point was between 2nd and 3rd, imagine being in the cockpit of the 3rd plane.

From the NTSB data, it appears that it may have been number 2 in line (PAL115) which got the 'worst' view.

Baylee Melodic Windjammer wrote: »

I don't think the AC pilots necessarily should never fly again after all pilots are humans and are fallible

Agreed, 'human factors' will likely be a major part of the investigation.

circular flexing

Some other interesting points

- Control tower was understaffed, one controller was on duty, standard is two.
- Aircraft was lined up for taxiway C for approx. 3nm, would this have been caught if there were two controllers?