Near miss with 5 airliners waiting for T/O on taxiway "C" in SFO!

[Zeffy]

Originally Posted by cappt

...This is also the distance where pilots flying a visual have already identified the runway and in their mind have shifted focus from runway identification to landing the A/C on the runway, I.E. airspeed, descent, final configuration...
Without mentally forcing yourself to backup and reconfirm your earlier decisions one can become focused on the task at hand and prior mistakes will go unchallenged.

Good points...

Brings to mind the C-17 wrong airport landing near MacDill AFB.

Once the target landing area had been identified (incorrectly) crew attention was focused solely to flight path and airspeed to the (wrong) TDZ.
Ironically, due to the precise flight path and airspeed guidance inherent in the HUD guidance, they got the beast stopped OK.

[aterpster]

Wasn't that during the daytime? And, I don't believe there were any other airplanes on the runway.

There is a wealth of incidents of pilots landing on the wrong runway and even at the wrong airport. But, I don't recall a case with the instant essentials where a pilot attempted to land on a taxiway occupied by well-lighted other large airplanes.

[Heathrow Harry]

firsttime for everything I guess.........

[framer]

As long as we approach the event with the assumption that these two pilots didn't do it on purpose ( I think we can all agree that is the case), and keep in mind that there have been many incidents of experienced and competent pilots landing, attempting to land, taking off, and attempting to take off on taxiways, then we must accept that the human brain is susceptible to creating the wrong patterns ( as mentioned by humble pax) if the stimulus presented is incomplete or differs from that expected.
What we do next is important, do we change the stimulus presented ( difficult in the real world but synthetic vision etc may help), or do we work on expectations? ( ATC required to transmit phrase " Runway 28R is unlit" on first contact) Or both?
If closely parallel runways had a lit orange circle or cross as part of the 1000ft markers and on a visual it was a requirement to have it in site and have a landing clearance to touch down, would that make a difference? On a PRM it wouldn't matter as the guidance is there.
Thinking out loud.

[OK465]

Bergstrom AFB (later Austin Intl) in the 60s:

On the parallel taxiway,

T
A
X
I
W
A
Y

(SV in the HUD will present and definitively outline the runway selected in the box and, in spite of comments about potential HUD fixation and the ever present danger of an incorrect FMS entry, I personally thought it to be extremely useful under a variety of circmstances. AFAIK it is impossible to enter a taxiway in the box as the destination runway. And then there's RAS....)

[galaxy flyer]

OK465 and Ian,

That's my experience with a SVS HUD, even if you hit fixated, the fixation is taking you to the runway. In the sim, it's pretty easy to land on the SVS 3V4Y7F and there is no taxiway, even the approach light is depicted extended out from the threshold.

[aterpster]

Originally Posted by galaxy flyer

OK465 and Ian,

That's my experience with a SVS HUD, even if you hit fixated, the fixation is taking you to the runway. In the sim, it's pretty easy to land on the SVS 3V4Y7F and there is no taxiway, even the approach light is depicted extended out from the threshold.

Even without SVS and HUD, seems like remaining coupled up to the LNAV/VNAV FMS Bridge 28R Visual, would deliver a flight crew to the correct sight picture.

Then again, perhaps without GPS there could have been a slight map shift.

[peekay4]

Summary of other taxiway landing events I found, for comparison. Some may be similar to this latest incident at SFO.

USA

1983 Continental DC-9 landed on the taxiway at Denver (Stapleton). Frank Lorenzo, who was Continental's Chairman, was a passenger on board

2000 Harbor Air Flight 4506 (Cessna 208 Caravan) landed on Seattle's taxiway T (then called Taxiway C) instead of 16R, after a visual approach

2003 American Airlines MD-80 at Seattle landed on the same taxiway T instead of 16R, after shooting the ILS approach and transitioning to visual at 3,500ft AGL, visibility > 10nm

2004 Air Canada Jazz Dash-8 also landed on Seattle's taxiway T instead of 16R, despite the fact that a big X had been installed to mark the taxiway. Neither crew member noticed that they flew right over the X, and the pilots did not realize they had landed on the taxiway until informed by the tower.

2006 Continental 757 lands at Newark taxiway Z instead of runway 29 after a circle-to-land from the ILS RWY 22L. Night VMC conditions.

2009 Delta 767 from Rio to Atlanta landed on taxiway M instead of 27R in night VMC conditions. Fatigue was a factor as one of the three pilots became ill during the flight and the remaining two pilots had to fly the route without rest. The Captain (PF) had been awake for 22 hours when the incident occurred.

2015 Alaska Airlines 737 became the fourth commercial flight to land at Seattle's infamous taxiway T. Back in 2004 another Alaska 737 also nearly landed on this taxiway, until an FAA inspector riding in the jump seat alerted the pilots on short final.


Worldwide

1988 British Island BAC 1-11 at Gatwick landed on taxiway 2 instead of 08L after a night visual approach. A British Airway 737 was on taxiway 2 at the time, and veered off the taxiway to avoid a potential collision (but got stuck in the mud). The 1-11 stopped 190 meters short of the 737.

1993 Air Malta 737 also landed on Gatwick's taxiway 2 after a night SRA approach. The Air Malta crew thought runway 26R was 26L, and after some discussion convinced themselves that the pavement even further to the right must be 26R.

2006 TAP A320 at Sao Paulo landed on taxiway B instead of 27R. Language may have been a factor as ATC tried to tell them to go around, but in Brazilian Portuguese.

2007 KAL 737 landed on the taxiway at Akita, Japan. Both pilots convinced themselves that the taxiway was the runway -- despite looking at the actual runway -- and concluded that the actual runway must be a "new one under construction". VMC with ceiling at 4,000ft and visibility > 10nm. Interestingly, the Captain (PF) made the approach with a HUD, but on this non-precision approach the runway was not indicated on the HUD's symbology and Captain manually aligned the Flight Path Symbol to the taxiway instead of the runway.

2009 Ryanair 737 landed on the taxiway at Cagliari, Italy. Daytime VMC.

2011 Thomson 737 lands on the taxiway at Paphos, Cyprus.

[galaxy flyer]

Originally Posted by aterpster

Even without SVS and HUD, seems like remaining coupled up to the LNAV/VNAV FMS Bridge 28R Visual, would deliver a flight crew to the correct sight picture.

Then again, perhaps without GPS there could have been a slight map shift.

There's just something wrong about "coupled approach" and "visual". Do we know when the crew disconnected the autopilot, or more accurately, are we sure it was hand flown from earlier in the approach? I know what your saying, but at some point a visual is a hand flown maneuver.

[b1lanc]

Originally Posted by peekay4

Summary of other taxiway landing events I found, for comparison. Some may be similar to this latest incident at SFO.

I get that point and there's probably as many incidents of landing at the wrong airfield. But which one of those crews actually tried to set their aircraft down on top of other aircraft? The alarm bells in the cockpit (and maybe an understaffed tower with a gap in radar coverage) went off far too late.

[pilot9249]

Originally Posted by Ian W

The difference between your claim and mine is that professional crews repeatedly land on taxiways. It is not uncommon. What these incidents have in common is that each crew were convinced they were landing on the runway they were cleared to land on.

The way flight safety works is not to crucify each crew with the hope of 'encourager les autres'; it is to examine each incident and try to find why these landings (and in some cases take offs) on taxiways takes place. Then attempt to prevent that happening. Human factors research shows that attentional (aka cognitive) tunneling is one reason, that is then supported by confirmation bias. The same research shows that once an individual has been shown what can happen in their heads then they are less likely to have the same tunneling. Note that is less likely not impossible.

Trying to pretend that it is just that particular crew will. inevitably, only lead to another crew perhaps actually landing on a taxiway with several widebodies queued on it. Identifying what the reason was for misidentification and finding a method of mitigating it is far more productive; and is the way flight safety has been improved in the past.

You reinforce my point by continuing to pursue the same line relentlessly and regardless what the other person actually said.

While you are busy saying what you always expected to say anyway (and hence land on the taxiway) you miss my point and the runway entirely.

I never said crucify the crew.

What I said is that the human brain is so complex and on margin unpredictable, that we should do everything we can to remove the human brain from the equation, before accepting defeat and accommodating it.

And that history dictates regardless what you and I discuss, over a slightly longer arc, this approach to safety is completely inevitable anyway.

[jugofpropwash]

Originally Posted by EEngr

And color vision at low light intensity suffers as well. Lighting/marking systems need to be based on patterns more than colors particularly at night.

Haven't heard this suggested, but some people's color vision is better than others - perhaps in this case, the pilot's color vision wasn't quite 100%?

Not sure when these pilots started their day, whether they would have been up and/or flying during daylight hours? I don't usually have a problem driving at night, but if I've been driving all day in the sun, then I have trouble seeing after it gets dark.

From a distance, could the lights on the planes have drowned out the green center line lights, and seeing the lights, the assumption was made that those were the runway lights? As someone said, once fixated on the idea that "that's the runway" it could be hard to rethink that.

[DingerX]

Originally Posted by b1lanc

I get that point and there's probably as many incidents of landing at the wrong airfield. But which one of those crews actually tried to set their aircraft down on top of other aircraft? The alarm bells in the cockpit (and maybe an understaffed tower with a gap in radar coverage) went off far too late.

It's harder to find information on near misses than to see the irrefutable proof rolling out on the taxiway. That said, there have been two Archie awards given in the last four years to Boston controllers sending around aircraft on short final to Taxiway B, in both times, much to the distress of the JetBlue passengers and crew on said taxiway.

And by the way, the human brain wrote the equation. You can't write it out, just shift the brains that enter into it and the conditions they do so.

[Volume]

Everyone has talked about the green centreline lights on the taxiway. How many of them would have been visible, with four large aircraft lined up on top? Is it possible that only fragments of green line would have been discernible?

I have asked before, is SFO having "modern" LED or "traditional" light bulb lighting? LED typically have a significantly smaller angle under which they shine their light. Is green light produced by green LED or is it composed from a combination of blue and yellow? Is the colour that clear from any angle? After all the center lights are designed for the aircraft on ground, not for those on approach. Only the approach lights are supposed to be seen from the 3° ILS path.

[DaveReidUK]

Originally Posted by b1lanc

I get that point and there's probably as many incidents of landing at the wrong airfield. But which one of those crews actually tried to set their aircraft down on top of other aircraft?

Albeit not involving a wrong runway/airfield or taxiway, the crew of US1493 (through no fault of their own) ended up doing exactly that.

[DingerX]

Well, if you include wrong runways, this nearly happened at SFO in VFR one fine afternoon in April, 2015 to an American 767 that, when told change to/clear to land 28L, readback but continued to 28R. A United 320 bailed out to C in a hurry. That crew was at the end of a long trip; just after leaving DFW, they had to return with a medical emergency; two hours later, they left again. That one didn't make the news either.

[aterpster]

Originally Posted by galaxy flyer

There's just something wrong about "coupled approach" and "visual". Do we know when the crew disconnected the autopilot, or more accurately, are we sure it was hand flown from earlier in the approach? I know what your saying, but at some point a visual is a hand flown maneuver.

If you go back through this thread (yeah, right) someone who currently operates into KSFO and who has used this FMS 28R database visual approach stated how nicely the airplane flies the approach, with auto-throttles.

I have the source. It aligns with the Runway at FD101, or about 3.6 miles from the threshold, and on a 3.0 degree VNAV path. Jeppesen codes these types of FMS database visuals to switch from terminal to LNAV/VNAV approach sensitivity (RNP 1.0 to RNP 0.30) at what Jeppesen considers to be the "FAF."

So, when are you supposed to disconnect? I don't know. I hope the carriers provide some type of training for commercial crews to use this non-public FMS visual procedure.

So far as where the visual segment begins, that would have to be at the point the clearance for the visual is issued. And, how close in should this procedure be disconnected from autoflight? 500 feet?

We don't know whether AC was coupled to the procedure.

What concerns me beyond this incident is that at least some portion of this procedure is being flown in IMC as a de facto instrument approach. I suppose the vertical path is above the applicable ATC MVAs until 1,600, which is the lowest MVA over the water east of 28L/R.

[peekay4]

What concerns me beyond this incident is that at least some portion of this procedure is being flown in IMC as a de facto instrument approach.

Not sure what you mean there... when flying a visual approach procedure, one cannot be in IMC by definition, i.e., must remain clear of clouds and maintain visual with the airport environment and/or preceding aircraft.

[testpanel]

IMC vs IFR (again)?

[aterpster]

Originally Posted by peekay4

Not sure what you mean there... when flying a visual approach procedure, one cannot be in IMC by definition, i.e., must remain clear of clouds and maintain visual with the airport environment and/or preceding aircraft.

The weather minimums:

WEATHER MINIMUMS
SFO Ceiling 2100'- VIS 5
OR SFO Ceiling 1000' - VIS 3 With VIS 5 in Eastern Quadrant (030^ Clockwise 120^) and San Mateo AWOS Ceiling 2400' - VIS 5 [San Mateo AWOS info available from SFO Tower. If San Mateo AWOS inop, use San Carlos (KSQL) ceiling of 2400'-VIS 5; San Carlos ATIS on 125.9]

Let's say the weather is 2,800 overcast, visibility 8 miles. ACME AIR checks in passing CEDES and requests the FMS visual. The clearance, "Descend via the track and altitudes of the FMS visual. Expect clearance for the visual passing JANYY."

Why else would they have weather minimums well below the altitude at ARCHI?

Pure speculation on my part. :-)