Attempts to take off from or land on taxiways are alarmingly common, including those by Harrison Ford:
Harrison Ford won't face disciplinary action for landing on a taxiway at John Wayne Airport [1]
Serious incident: Finnair A340 attempts takeoff from Hong Kong taxiway [2]
HK Airlines 737 tries to take off from taxiway [3]
Passenger plane lands on the TAXIWAY instead of runway in fourth incident of its kind at Seattle airport [4]
Here is a comment from the NTSB report:
"""
Observations made from the flight deck during the flight test approaches indicated that when the lights were set to the same levels as were encountered by the incident crew, from about DEPOT intersection, the runway 27R centerline lights were not identifiable and the taxiway M centerline lights were more prominent. When established on final, the taxiway signs were more visible than runway 27R edge lights. At about 500 feet above ground level the runway centerline lights were barely visible and it appeared that some lights may have been out. The color of the blue taxiway edge lights became distinguishable at about 500 feet above ground level while on approach.
"""
This is facilitated by the fact that in some airports, taxiways and runways are similar and sometimes even used interchangeably (!), i.e. a runway may be remarked as taxiway and vice versa. So it's not 'how can he be so dumb' moment, in some case you have no way to tell the difference unless you just know.
Yeowch. I was about ready to make a comment about how the airport/approach plates should make something like this impossible, but then I looked up the diagram for SEA:
Basically, you've got a runway, the taxiway of the same length as the runway!, and then two longer runways from left to right. I totally see how someone could make this mistake in bad lighting or weather!
It feels like it could be solved incredibly easily by a row of lights at the beginning/end of both the taxiway and the runway, bright red for taxiway, bright green for the runway. If you are attempting to land on the strip marked red, then you are doing something wrong. In case that the taxiway is legitimately used for landing, change the colours around.
I guess it hasn't been done for "reasons", I'd be quite interested in hearing why.
This already exists. Runway is lit up in white, has a green strip of lights at the close end as you look at it, and instrument rated fields have a chase strobe that points right to the end of the runway. Taxi ways are blue.
Things like this are being tried. SeaTac airport has a taxiway which is large, parallel, and frequently (enough) mistaken for a runway.
> SEA-TAC personnel have been very proactive in trying to reduce the problem, including the installation of an unlit, nonreflective elevated lighted X at the threshold of taxiway Tango; broadcasting a notice on the Automatic Terminal Information Service (ATIS) not to mistake the parallel taxiway for the adjacent runway; posting warnings on aeronautical charts not to mistake the taxiway for runway 16R; and the development of numerous training aids and brochures that explain the problem to transient pilots traveling through SEA-TAC.
My assumption based on nothing in particular is that one (of unlit/lighted) means 'Lights contained within the X illuminating it from within' and the other means 'Lights external to the X illuminating it from without'
Yeah. At 1:18 that's clearly HIRL enabled. There's no ambiguity that's for a runway rather than taxiway. And it's not possible to visually misalign yourself with the taxiway, the whole point of HIRL is a visual aid for precision instrument landings in bad weather for Cat I and II ILS landings. On an instrument landing, it's not legal to go below decision height (200 and 100 feet respectively) unless you meet the visual reference requirement in 91.175 c3, ergo the HIRL lighting is simply unambiguous for a night VFR landing, and makes me wonder if it was even on.
No, you can't. A friend of mine couldn't become a commercial pilot because he has some kind of colourblindness.
However, I remember him trying for some time and even campaigning for colourblind people to be admissible as commercial pilots because for the actual flying colourblindness doesn't seem to matter.
So, perhaps ridiculous as it may sound taking colourblindness into account when proposing such an improvement isn't such a weird idea after all.
Actually, airline pilots (and pilots generally) are not the only people you need to take into account here. Many other potential users of the runways and taxiways would need to be taken into account, including ground staff and maintenance staff. Since I doubt that all of these are ruled out on the basis of colourblindness, it makes sense to construct a system that is impervious to ambiguity due to colourblindness.
Not a stupid question at all. I clearly remember doing the colour blind test during my preliminary medical when I arrived at flying school to get my commercial licence.
The test consists of going through a book with images made up of different coloured dots, and within each image was a coloured number made up of similar dots. We simply had to flip through the book and identify the number within each image.
I was totally smug as I flipped through the book, calling out the numbers and watching the doctor tick off his chart. But towards the end of the booklet, I reached a page, and the image was just a jumble of dots to me. I couldn't see a number. I remember the panic and tightening in my chest as I thought to myself "Oh No, I can't see the number - I MUST be colour blind!! My dream of being a pilot is OVER".
I stared and stared at the page, but in the end had to concede defeat and I quietly told the doc that I couldn't see the number on that page. "Oh", he replied casually. "That's OK, there isn't one on that page!". And he grabbed the booklet and tossed it back on his desk saying "You passed."
Yes, but the certificate comes with a limitation (color signal landing daytime only proscription; or night and color signal landing proscription).
The requirements vary depending on class of medical certificate but in this context it's a class I medical for airline transport pilots (not commercial pilots) so it's complicated.
Also, full on dichromats or monochromats are rare. Usually people are one of three anomalous trichomats. Hence testing is required to determine if it's a problem. The initial test is designed to have a high failure rate, and the exception tests to get the restrictions lifted take more effort.
It's HN.... I get a good laugh out of all these aviation related threads. Understandably your average CS undergrad doesn't have much real aeronautical knowledge but that doesn't stop everyone from blowing up the comments every time the word 'airplane' is uttered.
No you just get a night flying and color signals proscription on your certificate if you're a monochromat or dichromat (full color blindness); if you merely have a color discrimination problem by being an anomalous trichromat, you might pass the test, or fail it but then get an exemption (reversal of the prior failure) by later passing a more involve series of tests.
The absolutely stellar safety record of the airline industry exists because they have repeatedly and consistent pushed past human error as an explanation for a crash.
Understanding that if a human made this mistake, another future human will likely make the exact same mistake, they push to understand _why_ an airline captain that should know made this error. Then try to correct for that.
If every crash was written off as "the captain should have known better", aviation would not be nearly as safe as it is.
Perhaps not all tech companies care to push past human error in each postmortem (or even have a proper formal process at all), but some are known to do just that. Etsy and Google are among the well-documented cases.
> This idea of digging deeper into the circumstance and environment that an engineer found themselves in is called looking for the “Second Story”. In Post-Mortem meetings, we want to find Second Stories to help understand what went wrong.
> Blameless postmortems are a tenet of SRE culture. … Blameless culture originated in the healthcare and avionics industries where mistakes can be fatal. These industries nurture an environment where every "mistake" is seen as an opportunity to strengthen the system. When postmortems shift from allocating blame to investigating the systematic reasons why an individual or team had incomplete or incorrect information, effective prevention plans can be put in place. You can’t "fix" people, but you can fix systems and processes to better support people making the right choices when designing and maintaining complex systems.
— Site Reliability Engineering — Postmortem Culture: Learning from Failure
Amazon is pretty strong on this front too, the most recent public example is the S3 outage and postmortem[1].
>Unfortunately, one of the inputs to the command was entered incorrectly and a larger set of servers was removed than intended.... We have modified this tool to remove capacity more slowly and added safeguards to prevent capacity from being removed when it will take any subsystem below its minimum required capacity level. This will prevent an incorrect input from triggering a similar event in the future.
While I agree generally with that sentiment, there's a key difference.
Airplane pilots are licensed, certified, trained, and regulated. There's a clear floor to who is allowed in the cockpit (barring extreme emergencies, e.g., incapacitation of a pilot).
By contrast, software is made available to pretty much the entire world. And it turns out that two thirds of all adults have "poor", "below poor", or no computer skills at all. Which is to say, the qualifications floor is nonexistent.
If you're designing a one-size-fits-all system, you've got to design for this. The results, I'd argue, are ... not particularly satisfactory.
I'm not saying "don't design for the user in mind", or "don't dismiss user error". But rather, than when your floor is zero, you're going to have a remarkably difficult challenge.
They do, or at least the big-scale places who know what they’re doing do. When Gmail or aws-east-1 or something goes down for hours, nobody is fired: they figure out why it was allowed to happen and change procedures so it can’t happen again.
They do. That's what UX is all about. Not just UI, not just making thing possible or visual, but actually making the experience as smooth and frictionless as possible.
UX is the bane of usability (at least as practiced). E.g. We've had professional "UX" designers seriously contend that "cancel" buttons should be made to not look like buttons. The joke at a certain large company was that if the UX designers had their way every default button would be big and green and every other button would be invisible.
True usability is about helping people make the correct decision, not just being smooth and frictionless. (See Don Norman's classic example of fire escapes that send people into basements to be trapped.)
I'm a daily user of some of the most complicated engineering software on the market. Every single release promises "Easier to use!" but all they do is make the most basic beginner-level actions more prominent and hide all medium to advanced level important functionality behind 3 extra clicks.
The end result is a beginner can do a tutorial exercise in 20 minutes instead of 30, while any true day to day work in the software takes 4 hours instead of 2.
Do you think this tradeoff between easy-to-start-using and power-use-friendly is intrinsic, a scale that designers need to choose where they want their software in it, or is it possible to hit both ends? Does anyone have examples of things they think address both beginners and power users adequately?
True. But as with any other discipline, it's about thinking about the intent and final goal, not just following a rule book containing arbitrary roles.
If someone walks into a glass door once, you can say they weren't paying attention. If more and more people start walking into the same door, there's something wrong with the door.
There was also SIA Flight 006 which crashed while taking off from the wrong (closed) runway at Taiwan[1], which was big local news in Singapore at the time.
Mistaking runway 05R for 05L when both are parallel is more understandable though. It was also at night, in poor weather, and the airport had not blocked off the unusable runway (!) and did not have the gear to track aircraft locations on the ground.
As I understand it, planes are largely automated but the automated components will always defer to the pilots whenever action is taken. I would hope with the amount of training required to do that job that the pilot would be the half of that equation to defer to.
Because the pilot is almost never looking at a map on final approach. He is looking out the window, and in the dawn/dusk light, or with a setting sun in the background or inclement weather etc., a taxiway can look exactly like a runway.
Don't forget that almost all taxiways run parallel to the runway, for about the same length.
(NB: I used to be a commercial pilot.)
In most airports I flew into, the surface material for the taxiways was different from the runway, so sometimes the taxiway could end up looking brighter and more prevalent than the runway under certain lighting conditions. Yep, I have had moments of confusion when on long finals into an unfamiliar airport. Never to the point of actually ending up landing on the taxiway, but the confusion is a thing that pilots have to contend with.
Thank you! That raises the question, (a) why is there no kind of warning system for when you're about to land on the wrong runway (there's a GPS system in the aircraft that knows where you're going, right?) and (b) why is the pilot (or at least co-pilot) never looking at a map on final approach? Is it really that dangerous for anyone to take a quick glance?
Yes, autolanding, or if the pilot was using ILS (Instrument Landing System), then this problem would almost never happen, as they would be aware there was a problem.
However, most pilots eschew automation especially on short finals, and prefer the 'hand fly' the aircraft in so that they can maintain complete control over the aircraft. Not a bad thing, and pilots want to actually fly the plane at crucial times.
When hand flying on final approach, your eyes and senses are 75% outside the aircraft. All you are worried about is your airspeed, your rate of descent and whether the runway is remaining visible at all times. Your eyes are darting from your primary instruments in front of you to the runway outside, and your arms and legs are busy working the throttles, control stick and rudder pedals.
Maps and charts are well stowed away at this point. In fact, once you enter the main airport traffic pattern, they are put away as all the details should be memorised by then. At best, the pilot will have the approach plate on his/her control column. But this just gives the instrument approach paths and radio frequencies etc. and not really a detailed topography of the runways/taxiways.
Ironically, AFTER the plane is on the ground, the non commanding pilot will usually pull out the airport plate which details all the taxiways and routes to gates etc., especially if they are unfamiliar with the airport and have to find their way around.
I have to disagree about the maps and charts being stowed away at this point.
I personally can only think of one 121 certificated carrier in the US that isn't using an EFB setup. These are typically mounted to the side of the pilot and there's no reason at all they can't still be on and open to either the airport diagram or the approach in use. In fact I frequently do this and do reference it on approach to parallel runways and to reference the turn offs for taxi.
As this discussion progressed, the lack of progress in the FAA adopting technology was addressed. It seems to me that the FAA is completely missing the boat on tons of safety to be gained with these EFBs due to the bureaucracy. My carrier was recently certified for EFBs and the vendor has the capability to have GPS geo-referenced charts but because of the complicated certification process the airline chose to get certified without to save money and time. As part of that we also received devices without GPS so now there's no way it's coming until the next cycle of devices. Personally I can't understand why there would need to be additional certification for something GA pilots have been using for years now.
I am going to say you are probably more right than me. I haven't flown in over two decades, so things are bound to be different/better nowadays and procedures and safety improves.
Take my replies here with a grain of salt and as being 'old school'. :)
Fantastic explanation, and the question on my mind the entire time is "but why??"! Surely the copilot can look at a map (when I say map I don't mean a piece of paper -- I mean a digital one with GPS and everything) and see their trajectory while the pilot is landing, right? Like I'm imagining there should be something on the plane that shows them their trajectory and the runway information. If there is, why isn't it telling them when they're going the wrong way? If there isn't, why not? Is the technology simply behind, or is there a safety reason or something?
Ah, I think you are picturing something like a HUD (heads up display) that they have on fighter jets? That would certainly go a long way towards solving the issue, but commercial aircraft don't have a HUD display system (though I believe the 787 has a partial HUD display now).
Commercial aircraft have a HSI (horizontal situation indicator) located front and center on the panel in front of the pilots. This instrument presents about 12 distinct pieces of information to the pilot, but once again, on short finals, the pilots is mainly concerned with (1) airspeed (2) altitude (3) rate of descent on this instrument. All of which are displayed on the left and right edge of the HSI in the form of moving bars. The information in the centre with relation to position and orientation is discarded as unnecessary 'noise' under this extremely high workload.
You are right in that the co-pilot has to work in conjunction with the pilot to ensure safe flight. Indeed, in many cases, it was the co-pilot (or non flying pilot) that alerted the flying pilot as to an imminent danger. But that doesn't always happen either. Take the case mentioned elsewhere here of the worlds worst aviation disaster. In that scenario, the junior co-pilot of the KLM jet notified the Captain (who was one of the the most senior pilot in KLM at the time) TWICE that he thought the PanAm jet was still on the runway, only to be ignored.
Yes, I don't know the types, but any kind of display that can show the relevant information would seem viable in 2017 :-) so it seems the main answer to "why?" is just "because adopting such technology is abysmally slow" rather than because it's inherently a bad idea? (Thank you for all the responses again! They're excellent.) (Edit: just saw your other comment, please feel free to ignore this if it's redundant.)
Yes, because getting ANY type of new technology on airplanes certified by the FAA (as well as the relevant air safety authorities in every other country in the world) is a laboriously long, slow and expensive exercise. It is sad, but the bright side is that is seems to be getting better nowadays as compared to when I was flying decades ago.
Also, iPads have become a vector for bringing new tech into the cockpit, since it's not physically installed in the plane and therefore doesn't require certification. You can pull up charts, set courses and upload to the autopilot.
This is actually kind of terrible. A compromised ipad could either compromise the planes actual vital systems for one, any time 2 systems are in communication a compromise is possible, and for two it could introduce a subtle numerical error.
I can't imagine a scenario where this ought to be allowed.
Sort of, at least in general aviation. Some of the newer avionics have bluetooth interface. For instance, popular GTN-series nav/comm can accept flight plans from Garmin Pilot software running on iPad.
Now, the iPad part is not certified, but the nav/comm (including software) is.
What you are describing is called PFD (Primary Flight Display). HSI is just a bottom part of it, basically combined heading indicator + CDI/VDI + (sometimes) moving map. But anyway, unless you are on instrument approach (or at least have it tuned in on your avionics), the HSI won't show you whether you are lined up with the runway or not.
I don't think a map or chart would necessarily solve the problem. The taxiway runs parallel to the runway, in the same direction, is the same length, and is similar in width.
Similar to how a map or GPS won't stop you from turning down the wrong way on a divided highway. Or how there have been many cases of people following GPS and turning onto railroad tracks that parallel roads. https://www.google.com/search?q=driver+turns+onto+train+trac...
It's not like it happens every day, so I expect there is some significant thought going into how to mistake-proof this particular problem.
I suppose that a GPS in an airplane should always take the precision of the instrument into account. If it can't determine which lane it approaches, perhaps it should show a warning sign, and/or highlight the lane that should be taken.
We can use gps to guide munitions moving at a very high rate of speed to very small targets. I'm guessing missiles and planes have better gps than your phone.
> Surely the copilot can look at a map (when I say map I don't mean a piece of paper -- I mean a digital one with GPS and everything) and see their trajectory while the pilot is landing, right?
It's not like the pilot doesn't know there is a parallel taxiway. Of course there is. They would be (subconsciously anyway) looking for it to complete their mental map. It just went wrong in this case.
You wouldn't even have needed a fancy automated system to realise you were looking at a taxiway in that particular situation. There were four big aircraft on there, with position lights and taxi lights and everything. And apparently visibility was good. It was just a spectacular brain fart.
While the system you're suggesting might give better situational awareness, it might also be a dangerous distraction in other cases -- or be wrong for some reason. So more tools aren't automatically better.
In this case the existing system actually worked flawlessly: another controlling instance (the tower controller) discovered the error, took appropriate action, and all that happened was a 15 minute landing delay and a good story. Probably happens more often than we'd care to imagine.
> Like I'm imagining there should be something on the plane that shows them their trajectory and the runway information. If there is, why isn't it telling them when they're going the wrong way?
Well, there's the ILS (instrument landing system), but pilots like to land manually. IIRC it's recommended these days, just to remain in training. I suppose it's also fun.
Also, ILS might be wrong somehow (interference, technical defects, the pilot accidentally entering the data for the left runway when he was told to land on the right, etc).
So the truth is this was just human error, and that kind of thing just happens. You know the saying: if you make a system idiot-proof, nature invents better idiots.
The copilot is also busy with checks and other tasks, it's not like (s)he's just with crossed arms watching the show
However yes, it's the copilot's job to double check what the pilot says and to warn if anything's not what it should be (like trying to land on a runway)
All you need is an universal (ICAO) visual indication for "this is a taxiway, do not fucking landing here". For example, a stripe of alternating red and blue lights.
I was going to ask about the ILS, which I assumed pilots always used for lining up their approach. I watched a friend of mine who was a pilot, punching in ILS codes on a training simulator and I remembered it myself years ago from Microsoft Flight Simulator 4. (now that I think about it, I think he even mentioned that he didn't often use ILS).
It seems like a pretty critical system for lining up an approach on the right runway. FS4 even had a training mode where it would display the ILS bars on the screen without needing to look at the instruments. You'd think some commercial airliners would at least put this in a heads up display, where it would be totally obvious you're lined up with the taxi way since the bar would be glued to one side or the other. Don't Navy pilots have a similar HUD for just their vertical approach angle?
Yes. I flew actual planes, and loved playing with MS Flight Sim too, and the number of times I WISHED we could have the same sort of heads up information on the real thing was quite funny.
(Also, would have loved SLEW MODE on the real thing, but that is a conversation for another time and dimension) :D
A pilot is required to have at least 3 takeoff and landings in the type of aircraft he is to be operating in the preceding 90 days.
And I say sort of because an autopilot coupled ILS approach down to 100 ft above the runway still counts. This mean that as long as the pilot set up the automation correctly the airplane will have him lined up and on proper vertical path at 100 ft. He can disconnect autopilot and gently set it down.
So one, mostly automated, landing per month is all that is required.
Could be, but there have been many cases of pilots missing an audible warning while under heavy workload too.
Most notably the Adam Air(?) crash in Indonesia, where the crew were so intent on debugging a faulty minor flight computer that they simply didn't hear the the warning klaxon indicating that the autopilot had become disengaged and the plane ended up rolling over, entering a spin and crashed into the sea.
When flying visually (VFR), taxiway lighting and runway lighting are visually very different. Taxiway lighting is almost always blue, whereas runway lighting is white (with any displaced threshold marked with red lights). In daylight, the markings on the runway are unmistakably different.
I'm not instrument rated but my understanding is that with ILS, instrument approaches can handle multiple parallel runways precisely and therefore the pilot has the tools necessary to distinguish between a runway and taxiway when in instrument conditions. Someone please correct me if I'm wrong.
Let the investigation happen, but from the reported facts, this seems to be a pretty clear case of pilot error.
And I think you meant to say "raises the question" rather than "begs the question."
I’m intrigued by the question of what you were hoping to add to the conversation with a Wikipedia link with zero explanatory text.
Did you mean “look, some people think popular usage has won out and beg the question is now a reasonable substitute for raise the question”? Or something else?
Does not something wrong enough times doesn’t make it correct.
For example using “because” without a preposition doesn’t become correct just because Buzzfeed makes it popular.
People make grammar/useage mistakes not because they consciously want to change the language but because they don’t know any differently. It’s quite different than how slang enters a language. Begs vs. raises the question isn’t some kind of colloquial thing, it’s just a misuse. I get things wrong when the language frequently enough, but when it’s pointed out, I don’t continue doing it; in fact, I am happy for the opportunity to increase my precision with the language.
>Does not something wrong enough times doesn’t make it correct.
This isn't true in natural language. Do something enough times -- 'wrong' or 'right' -- and it becomes a part of the language. Like it or not.
Crack an Oxford English Dictionary sometime. Look up your favorite word. Notice that there will be tens of different definitions, including a historical trace of 'first uses' through semantic changes, for the single word.
>People make grammar/useage mistakes not because they consciously want to change the language but because they don’t know any differently
Consciously or not, they are changing the language. Language isn't a fixed thing with unbreakable rules. It is mutable. A language is however people use it; it isn't a thing to know in its entirety before using.
People don't have to consciously want to change the language for it to change. It's not like the differences between modern English and Victorian-era English are all because someone willed it. Usage dictates meaning, and I'd say the general public uses 'begs the question' as 'raises the question', whereas only people interested in philosophy use it as 'presume the answer' - essentially the latter version has become a sort of jargon regarding fallacies.
It's like the word 'irregardless'. As much as 'it isn't a word', it clearly is, because you know what is meant when it's said, and plenty of people say it.
'begs the question' is a bad translation of petitio principii, which is itself a bad translation from a greek text. The general usage of the term is the only correct interpretation. The phrase seems to be a trap for a certain category of pedants.
> Is it really that dangerous for anyone to take a quick glance?
Speaking as a private pilot, it absolutely is. Even in a Cessna you're going to be travelling at least 75 mph. Possible faster if the conditions are gusty. For the last 1/4-1/2 mile at least, your vision needs to be 100% outside of the airplane. Now if you need to do that in a Cessna at 75 mph with no more than 4 people on board, why wouldn't you also need to do in a 787 that lands at ~180 mph with nearly 300 people, or an A380 that lands at ~140 mph with over 500?
> Now if you need to do that in a Cessna at 75 mph with no more than 4 people on board, why wouldn't you also need to do in a 787 that lands at ~180 mph with nearly 300 people, or an A380 that lands at ~140 mph with over 500?
Because you have 1-3 other copilots who also have eyes and could be doing this instead of you...?
It would be like looking at a map of California and expect to be able to get around San Francisco - the aeronautical charts are generally for long haul navigation purposes. There are approach plates for instrument landings, but for landing there you'd be looking at your instruments, which you've dialed into the correct runway, not the plate, in order to land the plane. In this situation, however, I can only imagine the pilot was on visual approach because otherwise the plane wouldn't have been lined up to land on the taxi way.
Why can't they make taxiways more obvious (e.g. paint them bright red)? Seems like it would be a small cost for a lot of insurance against an accident.
See the Tests and Research section of the following NTSB report on a Delta incident where a plane landed on a taxiway at KATL. Note Also that the lights had been recently updated with LEDs from incandescents. Basically, the taxiway centerline was brighter than the runway centerlines and the colors of the lights were only apparent at 500ft.
With the setting sun just right, you wouldn't see the red, just reflected sunlight. There's no guaranteed solution for all cases, and aerospace visuals are some of the most heavily studied of any industry.
Runways are already very visually distinctive. An active runway at night is a river of light that you'd think would be unmistakable, and yet these incidents happen.
They do. They have different lighting systems/colours etc. to differentiate them.
Taxiways generally have green lights along the centerline to guide aircraft rolling on the ground, but runways usually have white (bright) lights along the edges. There are times when even THAT causes confusion, because sometimes the taxiway centreline lights can look like the edge lights of the runway - especially if the light is refracting off a wet windscreen.
So, in the end the problem is generally avoidable, but if you are on finals at a really busy airport after a gruelling 20 hours trans oceanic flight, at night, in the rain, and the tower is trying to hurry you up to expedite traffic, you can make mistakes.
Perhaps this is a dumb question but is there any reason why the taxiway doesn't have HUGE red Xs going down it and the runway with HUGE green arrow going down? Sorta like what they do on bridges where the lanes can shift to allow more cares to travel in one direction but more obvious? They could even angle them so the planes on the ground see them differently or not at all.
Red 'X's on a runway usually denote an inactive or out of service runway.
They ARE actually quite different in terms of lighting, width etc. and 99.9% of the time there is no issue telling them apart. Runways also have extra approach lights etc. at night to really make it clear where you should be pointing the nose.
(NB: Check out the video posted elsewhere in this thread of the cockpit view of a 747 landing at SFO at night. It is pretty clear which is the runway and which is the taxiway).
However, it is those 0.01% of the times where the weather is bad, or you are landing into the sun or the traffic is really congested and you are trying to talk on the radio at the same time etc. where mistakes can be made.
In this particular scenario in the OP, I am assuming that having 4 heavy aircraft lined up on the runway pointing at the landing plane, then 4 sets of bright landing and navigation lights on the taxiway could have looked like runway lighting, and disoriented the Canada Air pilot? Just outside speculation here, but stranger things have happened.
It's like the case of the Air New Zealand CD-10 crash into Mt. Erebus in clear weather. Non pilots were incredulous that a plane could fly into the side of a mountain in clear daylight weather, but the visual trickery played by bright snow on a slope actually fooled ALL pilots on board that they were flying down a valley instead of into an upward slope. [0] (A badly programmed navigation computer didn't help at all.)
A good concept in theory, but bear in mind that the cockpit of a 747 or A380 is something like 2 or 3 stories above the ground, in terms of height.
To be able to see a taxiway light 50m in front of the plane, the light will have to be angled up to an extent that a plane 500 feet in the air a mile away will also be able to see it.
Not putting down your idea at all - anything workable that will improve flight operations should be considered.
Trivia: Speaking of lighting assistance, Did you know one of the best inventions out of Australia is the T-VASIS landing system [0] that is in use at nearly every major airport in the world?? Oh, and we invented the DME too [1], in the aviation world... :)
A good concept in theory, but bear in mind that the cockpit of a 747 or A380 is something like 2 or 3 stories above the ground, in terms of height.
Fun fact! The A380's flight deck isn't that high off the ground. It's exactly the same height as the A330 and A340, part of an effort to reduce differences and cut down on the amount of new things pilots would have to deal with to train on the type.
This is also why the A380 has such a pronounced "forehead" -- the much lower flight-deck height (compared to a 747) means there's a lot of plane above that.
(and bonus fun fact: the 747's flight deck is way up top not for visibility, but because Boeing anticipated supersonic planes would take over the passenger market; the 747's high flight deck was meant to accommodate converting the plane to cargo use, so that the whole nose could swing open for loading and unloading)
Australia is also responsible for "black box" flight recorders, which are one of those things where it's really hard to appreciate in hindsight why we didn't mandate them much sooner.
And yet, when the Asiana accident happened the complaint was that the pilots were relying on automation too much and didn't know how to properly hand-fly the airplane anymore.
Not all runways are equipped for ILS landings. Sometimes the ILS equipment gets taken out of service for maintenance. And even autoland has its limitations.
Just be sure to make a note of the video title so you can find it again -- copies of it tend to get taken down, since American Airlines got in some trouble after other parts of this training course were blamed for the crash of AA flight 587:
> American Airlines got in some trouble after other parts of this training course were blamed for the crash of AA flight 587
More specifically, the pilots manipulated the rudder on that airplane so aggressively that it sheared off in flight. They were trained to react aggressively in the "Advanced Aircraft Maneuvering Program", which this video is a part of. (There are also related videos that I'm going through now that are more relevant to this particular incident; https://www.youtube.com/watch?v=WfNBmZy1Yuc, etc.)
BTW, if AA is indeed DMCA-ing these videos, that's pretty interesting. Copyright was designed to incentivize the progress of science and the arts -- not to make large corporations look less bad after they unintentionally kill a few hundred people. Oops.
From the sound of things, this particular disaster hasn't happened before, even though it's conceptually easy to imagine how it could. Is that because it's easy for the controllers on the ground to wave somebody off if they're going the wrong way, or is it really because we're just lucky?
Side question: do you think one day commercial pilots aren't necessary anymore, i.e. totally unmanned commercial airlines? If so, how far away in the future are they?
There has been some talk of remote piloting (like drones,) but personally, I want whoever is flying the plane to have some skin in the game. Unmanned - fully autopilot aircraft are certainly possible and I sure Boeing probably has a Skunkworks looking into it, but I’ve seen enough source code over the years that I would have a hard time trusting an autonomous airplane unless they had escape pods. Autonomous airplanes seen to be solving a problem that doesn’t seem critical to solve – given the statistically exceptional safety record of commercial air travel. Pilot error is a very rare cause of airliner crashes.
I think electric airplanes powered by Mr. Fusion mini nuclear reactors would be the future I’d like to see!
(pilot here)
Yes, they are very well shown on maps, but it can be hard to map your diagram to what you're seeing, especially if it's raining or if it's at night, and even more if you're tired.
I never made the mistake, but can see how somebody in bad conditions would.
Thanks! But why do you even have to map your diagram to what you're seeing? Isn't there a GPS system that shows you exactly where you are on the map and where you are going to go in the next few minutes? It doesn't seem like there should be any mental or visual processing to figure out that you're heading toward the wrong runway, right?
Even setting aside the GPS/tablet/etc. question, I think you'd probably be astonished to learn how simple and how old the technology of automatic/instrument landing is.
Instrument landing systems date to the 1920s, a time when flying at night, or above cloud tops, was incredibly dangerous.
Imagine setting up a couple of directional radio antennas at the end of the runway. One broadcasts short beeps with long pauses (beep-----beep-----beep), the other broadcasts long beeps with short pauses (--beeeep--beeeep--). Set up the antennas so that if you're lined up horizontally with the center of the runway, the signals merge into a continuous tone (beeeeeeeeeeeeeeeeeep).
That's more or less how instrument landing systems still work today. Modern ones have more antennas and coordinate you both vertically (on the correct slope to touch down at the start of the runway) and horizontally, but the whole thing is still based on directional radio signals that you pick up from your plane, and a published set of information for each airport (the "approach plate") telling you what frequencies to listen for depending on the runway you want.
Airliners can be programmed to "autoland" based on this information, but it requires some setup and also requires the pilots to be ready to take over at a moment's notice (if, for example, controllers at the airport decide to have the plane land on a different runway).
It's a matter of scale. Even on a modern HSI, the difference in distance between the runway and the parallel taxiway is probably the width of the indicator icon. Very hard to see on instruments. Much easier to see by looking out the windscreen.
I can't tell if my expectations are just too high, but the impression I'm getting from all your answers is that pilots just don't have a dynamically updating (i.e. digital) map. It seems like everything is a bunch of numbers the poor pilots have to crunch or ticks that they have to count correctly or sheets of paper they have to pass to each other with diagrams. Why in the world is the system so ancient in 2017? Google Maps on my phone can show me where I am within 10 meters and where I'm going for the next 10 minutes, and it automatically zooms in and out at turning points... why can't an airplane have a display that shows them where they are and where they're going within the next minute?
Ironically, it is mainly because of reliability. Every system on an aircraft HAS to have double or triple redundancy by law, and a lot of these systems ARE antiquated by usual consumer standards.
A lot of pilots these days have iPads in the cockpit to assist with this, but in nearly all airlines and aircraft types, these are unauthorised devices that are not part of the certification. Generally they are used for non critical things like manuals or checklists, but I know of pilots that use them for situational awareness as well. In an accident though, this may work against them, badly.
Yes, it is the old legal and liability issue. Aircraft systems should be upgraded, but the cost of upgrading an entire fleet is prohibitive for most airlines that are already operating on the smell of an oily rag. New generation aircraft are getting better at it though. Like I mentioned before, I believe the 787 now has a HUD of sorts, which is long overdue.
You know what else Google maps does? It quits working at random times and behaves unpredictably in low precision GPS scenarios.
GPS units certified for aviation are tested to edge cases for usability with predictable and clearly indicated failure modes. This slows down new tech, but it's critical.
Even auto zooming on gmaps seems like a no-brainer, except for the times where the auto zoom into a trivial maneuver (continue onto road) occludes the difficult immediately following procedure (take correct exit out of 4 within 50 meters).
Okay, correction. Airplanes don't have GPS systems that work better at the level of reliability required. "It works most of the time, except it sometimes doesn't" might be sufficient when you can stop at the roadside, but doesn't quite cut it in the air.
Okay so let me get this straight: you're telling me that passenger airplanes currently fly by dead reckoning or stellar navigation or ground-based triangulation or something else? When a pilot doesn't know where he is, he doesn't have any kind of certified GPS-based system handy to tell him?
And moreover, even beyond GPS: how the hell can a plane's navigation system be capable of (and certified for!) landing the monster under autopilot, and yet not be able to tell the pilot its own location within the accuracy of a few feet reliably? I literally do not understand how what you're saying makes any sense.
Automatic landing systems rely on ground support by way of microwave signals from transmitters installed near autoland-capable runways. There's no general position information, and the airborne part of the system can't operate independently of the part on the ground.
Civilian GPS signals are limited in terms of position accuracy, too. I believe the CEP is ten meters. That's not bad, but it's not good enough to land an aircraft. Altitude information from GPS is (IIRC) either nonexistent or severely limited, too. About any aircraft with a glass cockpit is going to support a moving map display, and those do see use - but, as a couple of professional pilots in this thread have noted, during final approach the main concern with regard to aircraft position is relative to the glide slope and the runway, and a moving map has nothing to do with any of that.
On a more general note, it is remarkable to me that so many otherwise apparently sensible software engineers seem to regard aviation as a problem trivially susceptible of perfect solution - every time we here on HN discuss any sort of mishap even peripherally involving any aircraft, the same suggestions invariably arise around automation, various schemes of runway lighting, cockpit lighting, instrument augmentation or replacement, pilot augmentation or replacement...
One would tend to imagine that the history of our attempts to perfect our own craft - including those attempts which we who build and maintain software make anew every day - would give us cause for humility on the subject of fields other than our own, about which in the main we know next to nothing.
Imagine five hundred airline pilots confecting an ad-hoc, long-distance postmortem
of the recent us-east-1 incident! You'll have to imagine it, because it did not happen - and it is rare at best in my experience for pilots, because they are skilled in one highly complex and technical field, to imagine they can speak authoritatively in the context of another. Perhaps we might profit by their example.
> Automatic landing systems rely on ground support by way of microwave signals from transmitters installed near autoland-capable runways. There's no general position information, and the airborne part of the system can't operate independently of the
part on the ground.
Yes, I understand the landing portion has other kinds of support. However, I don't understand how it can not have the positioning information we need, because {see next response below}.
> Civilian GPS signals are limited in terms of position accuracy, too. I believe the CEP is ten meters. That's not bad, but it's not good enough to land an aircraft. Altitude information from GPS is (IIRC) either nonexistent or severely limited, too. About any aircraft with a glass cockpit is going to support a moving map display, and those do see use - but, as a couple of professional pilots in this thread have noted, during final approach the main concern with regard to aircraft position is relative to the glide slope and the runway, and a moving map has nothing to do with any of that.
This seems to contradict with what I read [1]. See this quote, for example:
Generally, the pilot will handle takeoff and then initiate the autopilot to take over for most of the flight. In some newer aircraft models, autopilot systems will even land the plane.
OK, so I don't have much of an expectation for older plane models, but my interpretation of the above is that in newer models, autopilot need not disengage before landing and require manual intervention in order to align the plane with the correct track (barring bad weather or other unusual situations like bird collisions, etc.). This is supported by the fact that they later explicitly mention that autopilot occasionally disengages. So it seems clear in normal situations autopilot has plenty of positioning available to it, enough to find the runway and locate the plane on a display of some sort. That goes pretty clearly against what you said. Am I missing something?
> On a more general note, it is remarkable to me that so many otherwise apparently sensible software engineers seem to regard aviation as a problem trivially susceptible of perfect solution
I hope this isn't referring to me, because I've been trying pretty hard not to give that impression (which I don't have). I never suggested anything about aviation perfection in general. I'm specifically talking about this particular kind of problem, because it seems to me that with current technology, pilot errors of this kind should be completely avoidable using current technology and, basically, non-problems in the first place. (Again: note that I said these types of pilot errors should be avoidable. I didn't say all faulty landings on the wrong tracks should be unavoidable.)
In other words, nowhere did I suggest that I think airplanes can fly or land themselves perfectly in unfavorable conditions. But this was quite a favorable condition for an automated system.
And, as a computer scientist, what seems ridiculous to me is the idea that "Oh, but this system you just described won't work if there is a hurricane and there are turkeys getting rammed through our engines while the pilots are asleep and ATC happens to be on strike in the middle of a nuclear holocaust, so let's "be conservative" and never approve it because it's clearly going to make things worse on average".
> OK, so I don't have much of an expectation for older plane models, but my interpretation of the above is that in newer models, autopilot need not disengage before landing and require manual intervention in order to align the plane with the correct track (barring bad weather or other unusual situations like bird collisions, etc.).
What is your definition of manual intervention? Pilots get updated instructions from the tower - the GPS path of the flight within 10 m is not programmed from the beginning due to weather issues, turbulence, traffic at the terminal ends, etc. The pilots hands may not be on the yoke, but they certainly are sending changes to the flight computer throughout the flight.
> So it seems clear in normal situations autopilot has plenty of positioning available to it, enough to find the runway and locate the plane on a display of some sort. That goes pretty clearly against what you said. Am I missing something?
Autopilot landings require a Cat III approach. There are currently no GPS (also known as GBAS for Ground Based Augmentation System) Cat III approaches approved in the US. The autopilot when landing is not using GPS for final approach. The pilot has pre-programmed in the waypoints labeled in the approach plate for the specific runway approach they've been told to take, and then they will switch over to the ILS approach.
The tone of many comments here (yours too) call out people who actually know how the system works and question them. The person you're trading comments with has said twice that GPS is not responsible for the autolanding portion of the flight. They even gave a summary of how instrument landing works - microwaves transmitted on specific frequencies in a specific pattern. You twice try to refute it. And then in your last paragraph you do exactly what he says the software engineers here do and you take offense to!
> sensible software engineers seem to regard aviation as a problem trivially susceptible of perfect solution
> And, as a computer scientist, what seems ridiculous to me is the idea that "Oh, but this system you just described won't work if there is a hurricane and there are turkeys getting rammed through our engines while the pilots are asleep and ATC happens to be on strike in the middle of a nuclear holocaust, so let's "be conservative" and never approve it because it's clearly going to make things worse on average".
Landing a big aircraft is not the easiest thing in the world, if only because the fate of hundreds of people rest in your hands. If you have a system that is designed to cut out the human component, it has to be all or nothing. As we learned from Asiana 214, when pilots have no actual practice doing something even in the absolute best of conditions and still cause loss of aircraft and life, how are they expected to perform when everything is going against them at the exact moment autolanding fails? Everything about flight training is preparing for the worst case, and practicing your skills over and over and over. Complacency can kill people when the margin of error between life and death is razor thin.
I was thinking of autopilot disengaging, but that's not really important here. I'll go along with what you just said.
> Pilots get updated instructions from the tower - the GPS path of the flight within 10 m is not programmed from the beginning
not in the beginning... so it is later?
> due to weather issues, turbulence, traffic at the terminal ends, etc.
Uhm, "traffic at terminal ends" suggests the real problem is that the correct path is not known in the beginning, or that autopilot might not be able to avoid collision on its own... which is quite a bit different from the plane being unable to locate itself accurately and follow the correct path even if it were known a priori. Are we even discussing the original issue at this point?
> Autopilot landings require a Cat III approach.
I have no idea what that means. I'm not sure what gave you the impression that I know what that means either.
> The autopilot when landing is not using GPS for final approach.
> The person you're trading comments with has said twice that GPS is not responsible for the autolanding portion of the flight.
> You twice try to refute it. And then in your last paragraph you do exactly what he says the software engineers here do and you take offense to!
I think you didn't read my last comment carefully because, as I already said, I understood this. My problem is whether GPS is accurate/reliable enough to lead the plane to the place where the next system can take over, which to me implies GPS is already certified to be accurate and reliable enough to get the plane near the correct runway. Read it again. There shouldn't be a single sentence there where I "refute" the microwave transmissions or claim GPS is actually used on the final approach.
> If you have a system that is designed to cut out the human component, it has to be all or nothing.
Wha..? Autopilot isn't perfect either, and can disengage in various situations that it can't handle, but they approve it and pilots manage to use it just fine. When it's GPS's turn, suddenly it has to be 100% perfect?
And when did I ever suggest you have to cut out the human component? For goodness's sake all I'm asking for is a little display with a map that shows where the plane thinks it's going. That's "cutting out the human component" to you?
No, they don't already do this. As 'phdp already explained, there are no instrument landing systems installed in the US which include GPS augmentation. So, while the moving map will show the same position information as at any other time when the aircraft's GPS receiver has sufficient signal, it's going to look something like this: http://www.stratomaster.eu/lignes/mgl/photos/enscr4.jpg
As far as "stupid mistake" goes, you sure do seem quick to judge professionals in a highly technical field totally unrelated to your own, and of which you've made clear you are happy to preserve your ignorance. Were I you, I'd hope my own errors meet with a greater extent of charity than that precious little you see fit to mete out. But that's your problem, not mine. Good luck with it.
> As far as "stupid mistake" goes, you sure do seem quick to judge professionals in a highly technical field totally unrelated to your own
The heck? Everyone makes stupid mistakes. Hell, I make more of them than a lot of people I know. That's why there are procedures and checklists and redundancies and automated systems -- to prevent stupid mistakes, evne by the best people. Where was I ever judging the pilot for heaven's sake?! Maybe you could be a little more charitable with how you judge people?
I mean you've already said you don't know what you're talking about ("I have no idea what [Cat III] means.") but you're totally cool with taking a confidently authoritative tone on it anyway, I'm not sure what other conclusion anyone should be expected to draw.
> I mean you've already said you don't know what you're talking about ("I have no idea what [Cat III] means.") but you're totally cool with taking a confidently authoritative tone on it anyway, I'm not sure what other conclusion anyone should be expected to draw.
...Did you come here just to fuel flames and burn people, or are you here to have a legitimate discussion about the subject?
You're totally cool being literally in ad-hominem territory at this point and you're lecturing me about not judging people I never even judged?
Is it really ad hominem to draw uncomplimentary conclusions from your evident disinterest in addressing substantive criticisms of the argument you're advancing? I don't think it is. Perhaps I'm wrong about that - but what else are you giving me to work with?
> Uhm, "traffic at terminal ends" suggests the real problem is that the correct path is not known in the beginning, or that autopilot might not be able to avoid collision on its own... which is quite a bit different from the plane being unable to locate itself accurately and follow the correct path even if it were known a priori. Are we even discussing the original issue at this point?
Your thesis of your statements is that it is ridiculous that there is no moving map for pilots to see on the airplane. You've been told directly by some commenters that there is (tuxer), and others, including me, have said that GPS is not the way to land a plane.
>I have no idea what that means. I'm not sure what gave you the impression that I know what that means either.
I mean this in a positive way, but you should look it up on wikipedia. There are many great articles about the systems in place for instrument approaches and landings, and you may learn something interesting for a quick 15 minute investment.
> My problem is whether GPS is accurate/reliable enough to lead the plane to the place where the next system can take over, which to me implies GPS is already certified to be accurate and reliable enough to get the plane near the correct runway. Read it again.
You call us out for dismissing your idea (And, as a computer scientist, what seems ridiculous to me is the idea that "Oh, but this system you just described won't work..."), which to mean seems to be some sort of GPS en-route to to landing automated system. This idea has been informed by your inaccurate reading of a CNBC article (So it seems clear in normal situations autopilot has plenty of positioning available to it, enough to find the runway and locate the plane on a display of some sort. That goes pretty clearly against what you said. Am I missing something?). The en-route exists but can often change from the time the plane leaves the ground to when it gets within 50-100 miles of the airport, at which time it might be disengaged and switched to VOR based navigation. The landing by GPS does not exist because it is a lot less accurate than the existing systems.
This branch of the thread is about auto landing, and you were talking about how it's absurd GPS is approved for automated flying. Given that GPS is used for en-route navigation and not landing, the assumption was that you were proposing GPS landing systems. That is what my comment is about. And yes, having autopilot disengage at 50 feet above the ground when the pilot is not expecting it can be fatal. I would say that for the landing stage it has to be all or nothing. And you've been suggesting cutting out the human component by proposing an system that goes from take off to landing without manual intervention! Even if there is still a pilot in the cockpit, it still takes a few seconds for the brain to spin up to take over a task from the computer.
It's awesome that as someone outside the aviation industry, you're curious about how it works. Your incredulous tone and sense of superiority (As a computer scientist, as if we should automatically and uncritically defer to your opinion, and I say that as a programmer) is a bit off-putting. If you had just done the basic amount of research you could have come back with a list of questions such as:
- Is GPS used to navigate planes from take off to landing? What else is used?
- Why was the pilot in this incident not using automation to land the plane? How often does this happen?
- What is the procedure for switching between different navigation technologies at different points of the flight? I know that take offs and landings can be manual, but when does autopilot kick in?
- Why is GPS not used to land planes?
- What equipment to pilots use in the cockpit? If they are not the latest and greatest, why?
We could then have had a proper discussion surrounding the need for backwards compatibility, the rolling out of ADS technology, how its absurd we still use voice over radio as a primary way of relaying commands, etc. Pilots and others in the aviation industry generally do believe that a lot of the technology guiding our largest planes is antiquated, but you trivialize it without understanding the full context.
>The tone of many comments here (yours too) call out people who actually know how the system works and question them.
Lack of enough common frame of reference that the ability to communicate is almost totally inhibited from the word go. It reminds me of the Feynman magnets youtube video. The interviewer asks Feynman to explain the attraction repulsion sensation and Feynman admits any correct answer the interviewer will not understand because he doesn't understand any of the basic prerequisites of having a conversation; and any answer he gives the interviewer would understand is cheating him out of the correct answer.
Seriously? From here, it looks as if you're doing your best in refusing to understand. Here's a hint: context. If you're en route, even an approximate position within hundreds of feet is great for navigation; please stow your straw men back where you found them.
In this case, the issue is with landing, where you need a precise position; "probably tens of feet, when it works" is exactly the sort of accuracy that's not good enough. As for "know where it is and land" - that's already there, is called Precision ILS, and uses no elements of GPS whatsoever.
The issue is you're treating this as though (1) cruising can be done on autopilot via GPS because it doesn't need much accuracy, (2) landing can also be done on autopilot because we have more accurate systems, but (3) the transition between the two requires manual intervention because the plane can't reliably locate itself accurately enough to get in the right position for landing systems to take over.
But that simply doesn't seem to be the case (in normal weather, which we had here). See my reply to a sibling comment here for a link. It seems that in normal conditions planes are capable of doing it all in one go, which means they are capable of accurately and reliably locating themselves. Nobody is expecting these to work in extreme conditions like bad weather or bird collisions, but in those conditions autopilot can and does already get disengaged (whether automatically or manually), and this could be done the same way. It's not like we forbid autopilot entirely because it isn't 100% perfect, so why not apply the same standards here?
Nope. (3) The transition between the two requires manual intervention, as those are very different flight modes, using different systems and there's no automagic way to switch between them; the quote from the linked article is an oversimplification akin to "computers can do almost anything - therefore what do we need programmers anyway?"
Oh, and aircraft DO have moving maps etc. now, but these are usually dialled out to show a range of hundreds of miles in cruise, and are not usually adjusted to show distances in mere metres when on final approach.
Once again, because these displays are usually on the centre floor console or located below the HSI, they are outside the 'scan zone' of a pilot on short finals.
In their defense, I would guess that making such a system 100% reliable and unintrusive may be a lot harder than we realize. It's not uncommon for Google Maps to be off by a few blocks for 10 seconds or more, and that could very easily be fatal if you're using it to land an airplane.
Are you suggesting their current GPS systems have that kind of issue though? I'm just asking for an interface to be added, not a entire concept of GPS functionality to be added. I'm assuming the latter is already used. It's not like I was suggesting they use Google Maps.
There is, if you use it. In this case they were doing a visual approach , where you use eyesight in order to line up and land, because you have seen the runway. People do visual approaches in SFO as they are necessary to land 2 aircraft at the same time on the parallel runways ( they're like 30ft too close for parallel approaches on instrument ). It is still recommended to even in a visual approach, turn on the instrument approach systems ( for that exact reason ) but it looks like it is not mandatory for that airline.
If you watch any 28R approach video for SFO on YouTube, you'll be even more surprised. Because there is no way you can mistake the taxiway for the runway. There is a huge array of lights pointing the pilot towards the runway.
I would imagine this would be more of a problem during the day because the lights are not as bright compared to the ambient light.
I would think the best way to get this detected early would be to integrate some logic into the ATC radar. As the plane gets closer to the runway (or taxiway!), it should be easy to tell where the pilot is going by looking at the track and issue an automated warning.
> Audio from the air traffic controller communication archived by a user on LiveATC.net and reviewed by this newspaper organization showed how a the confused Air Canada pilot asks if he’s clear to land on 28R because he sees lights on the runway.
> “There’s no one on 28R but you,” the air controller responds.
> An unidentified voice, presumably another pilot, then chimes in: “Where’s this guy going. He’s on the taxiway.”
> The air controller quickly tells the Air Canada pilot to “go around.” telling the pilot “it looks like you were lined up for Charlie (Taxiway C) there.”
> A United Airlines pilot radios in: “United One, Air Canada flew directly over us.”
> “Yeah, I saw that guys,” the control tower responds.
The "Where's this guy going" unidentified person on the channel prevented the worst accident in aviation history. Great Job! Give that guy a medal and a movie.
It's likely that the controller would've seen it as well after the Air Canada pilot indicated that he saw planes on what he believed was the runway.
Even if not, the pilot would've probably aborted the landing since he saw the other aircrafts. So it was certainly a dangerous situation but I doubt that a crash would've happened without that pilot's comment (it was certainly helpful though).
Controllers in the tower do not have a good visual perspective to see runway alignment. I don't think they would have caught this.
At night, there are many visual cues to see the runway environment - it sounds to me like this complacent pilot ignored many red flags and was about to drive his plane on top of those on the taxiway.
A big heavy airliner can't just stop its decent instantly; it's a good thing that pilot on the ground chimed in when he did.
I betcha the landing pilot won't have a job pretty soon.
When you're descending at a rate suitable for human cargo in good weather there's plenty of room to maneuver.
A C-5 that's trying to drop out of the sky ASAP before anyone can shoot at it is a different story.
The pilot and copilot were clearly in WTF mode since there were planes on what they thought was the runway. If nobody had spoken up it would likely have gone back and fourth between them and the tower a few times ("are you sure the runway is clear?") and in all likelihood they would have caught it. If they didn't the pilots would have likely requested to go around. Backing off of what you're doing and assessing the situation when anything is not quite right is SOP in every facet of aviation.
Radar displays are not going to accurately display a displacement of 150m. They're set up for broader area surveillance.
Edit: here's a good example of a terminal radar display. The two parallel diagonal lines are the extended centerlines of two runways that are about 1570m apart (Athens airport, LGAV).
Yeah. When reading about modern air disasters, you often hear that five different things had to go wrong at once for them to happen. In this case, one or two things went wrong, the rest of them went right, and everything was fine.
This is spot on. The strength and weakness in doing something as complex as safely landing an airliner full of people is all in the capability of the human mind to use technology. Especially in flying, I think (I did this for a career, a whole other career ago) it's easy to have everything going smoothly, then some "minor" thing be out of expectation, but hey things are going smoothly, so dismiss the minor thing because here's this other thing that does look just right, and ... and ... suddenly several mistakes have compounded.
It's sometimes harder when there's nothing impeding you at all to do everything right because the routine-ness of habit takes over. Being busy (but not overwhelmed), while more mentally and physically tiring, is sometimes easier, because you pay attention to the right things at the right time.
This is true in all manner of domains. I've never flown but I can remember a number of times I've made (fortunately relatively inconsequential) mistakes that caused me to face-palm afterwards. Why? Not so much that I made a mistake. But that I could recall clearly a number of observations that I made but dismissed because they ran counter to what I "knew" was the reality of my situation.
"Taxiway", coming to theatres near you soon. Directed by Clint Eastwood. Starring John Malkovich as The Pilot, Tom Hanks as the "Where’s this guy going?" guy, and Dave Bautista as the taxiway.
It's from a Delta flight a few years ago that slid off the runway at LaGuardia in snow. The action starts about 5 minutes into that recording; Tower has just set up another incoming Delta flight to land, and is trying to confirm Delta 1086 has cleared the runway. Delta 1086 is poking its nose out over the water at this point, and the controller can't see it.
At around 5:56 a ground vehicle calls in to ask permission to cross the runway and gets it. Then the ground vehicle sees what's happened (Delta 1086 slid off the end of the runway), and this is where it gets interesting, because you see very quickly two important things:
1. Trusting the report: the guy up in the tower does not have a perfect view, and in fact in this incident he can't see what has happened to this Delta plane. He's getting an unsolicited report from a ground vehicle, but he believes it and acts immediately on it. If the report's mistaken, worst outcome is some planes circle a bit more before they finally land. If the report's right, though, the worst outcome is planes trying to land on a crash site.
2. Division of responsibility. From the moment the crash becomes known, it's the job of the emergency/rescue teams to figure out what's happened and deal with it. The guy in the tower probably desperately wants to know more about what's going on, but he's got planes stacked up waiting to land, and his curiosity is going to have to wait. Beyond getting confirmation of a couple unusual orders, Tower just lets the folks on the ground do their job, and sticks to doing his.
Here's a rough annotated transcript (timestamps are from the video), since I know people complain that ATC recordings are hard to make out:
6:08: Delta 1086, Tower? (trying to contact the plane that's just gone off the end of the runway -- Tower still doesn't know what's happened)
Tower, call 100, Runway 13 is closed. (ground vehicle has seen the accident)
Tower, red team to go onto 13.
Tower, you copy? Call 100, Runway 13 is closed.
Call 100, you said Runway 13 is closed? (Tower confirms what he's just heard)
Affirmative, 13 is closed.
Team red, Tower?
Tower, you have an aircraft off the runway.
6:45: Delta 1999, go around!
1999, going around.
6:50: The airport is closed. The airport is closed. We've got a 34. (ground crew saying this is bad enough the airport needs to close)
Call 100, say again?
6:59: Tower, you have an aircraft off 31 on the north vehicle service road, please advise crash rescue, LaGuardia Airport is closed at this time.
7:12: Good afternoon, Tower, Delta 2522's on the ILS for 13 (another Delta flight is lining up to land on the runway where the crash happened* -- they can't hear what's happening below)
Delta 2522, LaGuardia Tower, go around!
Go around, Delta 2522. (pilot confirms that he's going around)
From there on out, it's just Tower giving instructions to other planes on what to do, and a report from the ground to tell the pilot of the crashed plane (if he does get on the radio again) that the plane is leaking fuel from a ruptured wing.
Now: what if the ground vehicle hadn't gone out there? Or hadn't seen the crash in time, or hadn't been able to get a report to the tower in time, or the tower had hesitated a bit more in accepting the report and starting to reroute planes? A similar problem to the SFO incident, only instead of an occupied taxiway it would've been an occupied runway (and one that had already caused one plane to slide off in the snow, and now had emergency vehicles converging on it).
The point is, though, that this isn't necessarily a miracle or a lucky escape -- everybody is involved in ensuring safety, and that's what happened in both cases, providing extra layers of watchfulness which averted much more serious trouble.
* The runway is referred to as both "13" and "31" at different points; runways are numbered according to compass heading, so "13" means it's aligned to a heading of 130 degrees. The dual numbers are because you can take off or land from either end; the other number is always +/- 18, since it's 180 degrees around from the first one. In this case, the designation of the runway is 13 from one end, and 31 (for 310 degrees, 130 + 180) from the other end.
I think the most important thing here is that we stop to think about the gender normativity of the unidentified person's use of the term "guy" in "where's this guy going."
There is a very good technical reason we still use AM.
AM is not prone to capture effect, wherein the loudest signal captures the receiver. With AM, you can hear multiple people talking on top of each other at the same time.
For uncontrolled radio situations where you have a number of unknown people who need to access a radio channel, neither digital nor encryption bring any wanted benefits.
As a note - the audio from this recording is very clear to me - I've heard far worse out of scratchy narrowband FM.. with AM the weaker the signal generally the quieter it is - but not generally that much noisier.
Yup. FM is great for noise rejection. But any signal which is not "the one true signal" counts as noise. So FM interferes with other FM signals mightily and the most powerful one tends to "win". Whereas with AM the signals just add on top of each other.
I've been into amateur radio for ages, presumably that has trained my ear. I routinely copy voice signals which are just a garbled gibberish to innocent bystanders. (Same obviously goes for ATCs and pilots to an even larger degree, being exposed to comms several hours a day)
Yeah, Even with P25 - the same skills apply, but instead of pulling it out of noise, you're relying on your skills of interpolation to suck it out of the noise. Still it's a neat parlor trick :-D
Mmm. Here's the thing, I think we heard a lot of the same arguments before DSC took off for marine radio.
They too have uncontrolled radio, a large number of unknown people who need to communicate with whoever happens to be nearby, they have more powerful transmitters owned by governments that "need" to shout down less powerful ones on transport vessels occasionally.
Now, maritime radio IS a different environment. I'm not suggesting that DSC should just be dropped in as a replacement for AM analogue transceivers on planes, but I _am_ saying that I don't buy the theory that it so happens AM analogue is the right choice and not just the result of inertia.
I've used both marine (FM) and airband (AM) radios. My marine experience was in a relatively uncluttered environment (Cleveland) so take this with a grain of salt but I found that you rarely had people talking over each other in that environment, especially with it's limited propagation. Airband on the other hand is almost always a very high radio traffic environment if you are actively switching between tower, departure and center channels throughout the flight. It becomes imperative to be able to tell when someone stepped on ATC and you didn't get some message.
I believe there are significant technical advantages to AM for airband - I also believe that marine radio is less likely to need those advantages, so narrowband digital is more workable there.
they still use AM because higher powered transmitters (which will generally be the tower) can overpower weaker transmitters (planes). i'm given to understand aviation is enamored of that feature.
> Just imagine the economic damage that a single asshole can do with a high-powered transmitter near an airport.
the pilots could probably just pull their cell phones out and call the tower. there is much worse trouble single assholes could get up to.
I think you'd be very surprised what a lot of power and a shit antenna can do. Obviously won't last when someone comes along with DF equipment, but it really wouldn't be hard to conceal visually.
Visually is the key word here. RF-wise, you may as well put a beacon on the moon and scream "LOOK AT ME" for all the good it'll do when the very, very angry FAA/FCC vans show up.
It is a very Silicon Valley thing to assume that everyone else is an idiot and needs to be disrupted, with not a thought given as to why things are the way they are. Sort of funny to watch it from the side
I assure you that there is a reason AM is used. The reason is that in the case that two stations are transmitting, everyone will be aware of it and the stronger signal may be heard. This is not something that FM can guarantee. And a digital protocol requires a much better signal (where old analog tv gave snowy picture digital TV gives nothing). Not to mention the insane costs that digital radio retofitting would require of every single plane in the sky.
Edit: parent seems to have deleted his post. It previously asked condescendingly why we still use AM for comms in the air.
Digital only require better signal if you want high data rates. With analog signal the human has to do the error correction, i.e. extract the signal from the snowy picture. Digital signals can use error correcting codes to get the same result. The more bits you have for error correction, the better your recovery can be.
With a lot of FEC, you can sort of approach it, but still not quite. Human brain is a lot better at picking out voice out of noise than any algorithm currently known. That is , for example, why your Alexa cannot take your commands in the middle of a 100-person party, but your interlocutor can hear you and respond.
That is not very relevant since humans don't use an error correcting code that is easy to understand for machines when talking to each other. So Alexa's ability to filter your voice out of the noise has little to do with the ability of two machines understanding each other over a noisy channel when they use appropriate coding.
I hate aircraft radios. I'm sure the guys in big planes have better equipment and don't suffer from this as much, but when I fly small planes, it's a complete mess of wildly different volume levels (it's fun to turn up the volume to hear a quiet person, then get blasted by the next transmission), interference, and irrelevant transmissions from a hundred miles away. Don't get me wrong, they get the job done and any change would have to be very carefully considered, but I'm always happy when I get far enough from the airport that I can turn the damned thing off.
I think a digital system could be done much better than what we have. But I'm sure it's not worth the huge effort it would take to design and build.
This is a lack of AGC (automatic gain control) - usually even on the (shortwave and Ham) radios I've seen it on, its switchable, otherwise (in certain conditions, like rapid fading) you can end up with gain pumping, which can sound like the audio is surging.
What it normally does is reduce gain on high strength signals, and increases it on weak ones to give a constant volume level.
I'm honestly surprised the radios in general aviation craft are not so equipped, as its generally a standard part of most AM radios.
Thanks, now I know what to look for next time I go equipment shopping. I'm sure many GA radios have this, but I'm using particularly low-end radios since space and power consumption are more important for what I do. Still won't fix getting an earful about skydivers at an airport a hundred miles away, but it would be an improvement!
Okay, but do military-grade systems use AM like this too? I imagine their needs aren't any less than those of civilians, and I imagine they've gotten their communication systems to work just fine. So what's the issue?
"Military-grade" systems use complicated techniques like frequency-hopping spread spectrum. They're designed to be encrypted and resistant to jamming. The jamming resistance is not something we care about (if someone is jamming the signal you just make them turn off their radio), and the whole point of encryption is to prevent interoperability when you don't have the right key.
So sure, you could pay a bunch of extra money for military features and end up with a product that is even less what you want than AM radio. And then you'd have to retrofit everything with these systems.
AM is wonderful. You put a bunch of people on the same channel and it just works.
> "Military-grade" systems use complicated techniques like frequency-hopping spread spectrum
Note that that technology has been around since World War II. And fun fact, while we're on that topic, this page is worth a read: https://en.wikipedia.org/wiki/Hedy_Lamarr
"The jamming resistance is not something we care about (if someone is jamming the signal you just make them turn off their radio)"
Evie might know this, too, so I think you should be concerned about jamming. Luckily, AM provides jamming resistance, since Evie would have to bring a powerful transmitter to be loud enough to drown out the other signals.
Nevertheless, I think large airports should have fast response teams who can rapidly fix the position of a jammer and silence it, if needed. If Evie could effectively take out, say, the main airports of LA and SF for a few hours with a few strong AM transmitters, I doubt all will end well.
Evie? I honestly have no idea who that is. But the more powerful the jamming, the easier it is to find the source. These days even hobbyists play around with cheap off-the-shelf RF analyzers and directional antennas, so I can't imagine that someone jamming airport signals would be able to evade arrest for very long.
I don't understand what is so wrong with the transmission quality in that MP3. The issues I hear sound like they are simply due to trying to transmit from an extremely noisy environment (the cockpit of an airplane); they are mitigating the pilot's voice being entirely drowned in the background noise by placing the microphone right next to their mouth, but that then causes the sound of their breath to be extremely noticeable. I am not sure how to fix these issues but they don't seem to have anything to do with the usage of AM radio.
I wasn't complaining about the transmission quality, I was just trying to understand the parent comment. The person who complained about the former was someone else.
Sorry; I misunderstood (and to be honest even after my tenth reread am still having a difficult time parsing...) the end of your comment about the "issue" :(.
OK, I'll clarify. Someone suggested non-AM transmission has problems X/Y/Z; I suggested that I expect non-AM transmission has overcome problems X/Y/Z in the military, and therefore I'm not convinced this is the real issue preventing non-AM communication. I was not discussing whether we or not we need non-AM communication in the first place.
Airlines don't need military systems which are used more for confidentiality and OPSEC than other things. In the field, militaries tend to avoid radio communications for various tactical reasons. Submarines are a great example of this.
Military-grade systems may not be any "less" but their needs and priorities may be different (e.g. secrecy). So different technology may be used, without any contradiction to why the current one is good for civilian air traffic.
> It is a very Silicon Valley thing to assume that everyone else is an idiot and needs to be disrupted, with not a thought given as to why things are the way they are.
This kind of thinking has also lead to a ton of success in Silicon Valley. See Tesla for example.
Did Tesla really disrupt the car market? Sure, everyone is now building EVs but that already started before Tesla became successful. They have no market power on the overall car market.
Companies like Uber, Google, Amazon certainly disrupted markets but I wouldn't be so sure about Tesla.
I can't honestly say I ever paid attention, but I just checked all clocks I could find, one flashes 00:00 and the other 0:00. Most if not all clocks I've used had a 12 or 24-hour mode, not sure what happens if you set to 12-hour and unplug it (probably just reset to defaults).
Here's a night approach on 28R at SFO.[1] Same approach during the day.[2] The taxiway is on the right. It's a straight-in approach over the bay. The runway, like all runways at major airports worldwide, has the standardized lighting that makes it very distinctive at night, including the long line of lights out into the bay. This was in clear conditions. WTF? Looking forward to reading the investigation results.
The planes on the taxiway are facing incoming aircraft as they wait for the turn onto the runway and takeoff. So they saw the Air Canada plane coming right at them. That must have been scary.
I agree. Check out the NTSB report[1] for the crash into a neighborhood in Queens[2]. 200-odd pages of very sober, rational (and interesting) discussion of how a pilot ended up snapping the vertical stabilizer off an A300. There are many factors leading up to what ends up being summarized as pilot error.
This kind of analysis is aided by the airline industry having very strict training and procedures that flight and ground crew are required to follow, along with regulations that equipment has to comply with.
With them, you can see, for instance, where a pilot or mechanic did not follow procedure, was not trained adequately, or maybe some equipment was out of spec and wasn't caught in a checklist that should have been followed by the maintenance crew, etc.
Compared to the airline industry, most software engineering is much more of a fly by the seat of your pants phenomenon where anything goes, and there's little standardization, process, or regulations that the entire industry has agreed on and actually follows. Such a Wild West approach definitely has a lot of upsides like speed of development, flexibility, freedom to innovate, and so on. But can have a cost in safety, reproduceability, accountability, and analysis of what actually went wrong or how to fix it next time.
It's detailed, but well-written and the quality of the analysis is extraordinary.
For a sample, take a glance at the debris maps on page 45, followed up with amazing results on page 75. If you like understated drama, start at the comm transcript on page 42. ("Lock the doors.")
English is not my native language, but shouldn't the headline have read "SFO near miss would have triggered aviation disaster"? "Might" seems to indicate that something else happened afterwards as a possible result of the near miss
Not a native speaker either (though I've been fluent for decades) and I find the title horribly misleading as well. "Might have triggered" implies the disaster happened, or at least that it's not known whether a disaster happened.
"Would have triggered" implies neither the disaster nor the near-miss happened, so I think the most unambiguous phrasing would have been "could have triggered".
OTOH I've seen native speakers use headlines like "Why this works?" (which by my understanding of the language shouldn't have a question mark), so IDK.
the problem is that while Could is technically preferable to Might in this situation for exactly this reason, the proper distinguishing between the two conditions has long been vague.
People will often use Could have when they mean might have.
To be accurately and unambiguously understood by everyone it should be SFO near miss could have triggered aviation disaster but didn't. Since the completely unambiguous phrasing is not very elegant then Could have would be preferable, but with the understanding that for some people it will still have some ambiguity that needs resolving by actually reading the article.
"Might" seems to indicate that something else happened afterwards as a possible result of the near miss
In this case it doesn't, the headline can be interpreted in slightly different ways but is not wrong (or really that ambiguous) as written. Standard headline-speak for 'something happened that was a possible cause for something else that also happened' is 'may have caused' like here:
The moral of this story for me is: be that "another pilot." To be clear, "another pilot" of another aircraft. Not as clear as it could be just like the title of this article is ambiguous.
The moral of this story for me is: call out immediately if you see something off. He's the real hero. Even if the ATC controller immediately saw the plane being misaligned at the same time - that feedback confirming another set of eyes on something that is off couldn't have hurt. All 1000 people on the ground needed that feedback. Always speak up in situations like this.
I don't mean to be a downer, but you could certainly cause an accident by calling out an observation which was faulty. I agree with the sentiment that you should speak up when you see something amiss, but let's not kid ourselves that, like every decision, it comes with inherent risk.
I wonder how many people have died throughout history due to embarrassment, or (similarly) going against your instinct. The number exists but is unavailable to us.
- Not having that dark mole near your privates checked out.
- Not wanting to turn the car around.
- Not walking away when those shady guys walked in.
Korean airlines decades ago had a huge issue with copilots not speaking up to captains when concerned. While not embarrassment - there was a lack of speaking up. IIRC they forced cultural change in order to make it less of an issue.
Right, but a problem with humans is that they tend to seriously over-estimate their capability. So we need to be statistics led on this
Given the choice between "One time in a million a human flies the aeroplane into the ground and everybody dies" and "One time in 10 million the computer flies the aeroplane into the ground and everybody dies" we ought to be hard-headed and take the ten times fewer deaths, but we prefer to say "Ah, but _my_ pilots are better than average".
Now, the reality gets very complicated, because we ask computers and humans to fly in different conditions. A pilot will refuse to try to land an aeroplane into thick fog, the Cat III ALS is happy to try this because it sees through fog. On the other hand when external circumstances are trying, the pilots may choose not to attempt ALS at all. But despite this complexity, we do need to accept that sometimes a smaller risk of automation error is preferable to the risk of human error even if _neither_ is perfect.
We are not in disagreement. My reply was to "why not automate everything?" and indeed the answer is "because sometimes humans are better," not "do not automate anything" :)
In the early 1960s, a pilot mistook a WW2 airfield for Heathrow, and landed his 707 on it, barely stopping before the end of the runway.
The runway being too short to lift a 707, mechanics stripped everything out of it they could to reduce the weight - seats, interiors, etc. They put barely enough gas in it to hop over to Heathrow, and managed to get it there safely.
I've lived 22 years of my life in the UK and 12 in North America and never heard of this before so I don't think it can be particularly common. I've heard of being 'laid off', 'fired', 'let go', 'terminated', 'made redundant', 'purged', but never this. Based on this link https://en.wikipedia.org/wiki/Cashiering it might be a military way of saying it so perhaps the poster has some background in that field.
I have never heard the usage of this word - I assumed the commenter was inventing a probably-understandable, slightly humorous usage on the fly which I've found to be a fairly common habit among ultra-intelligent people.
A few years back when SFO was closed after the Asiana crash, a lot of big planes got re-routed to Oakland, then once SFO was open again they made the hop across the Bay.
Before crying pilot error, we must all read Sydney Dekker's A Field Giude to Understading "Human Error" (and fully appreciate why he uses those quotes). Don't immediately assign blame to the sharp end. Take a look at the blunt one first. Most likely not a pilot error. Assigning blame is a very human need, but assigning it to the most visible and accessible part is almost always wrong.
I second this. This book has quite a boring cover, and quite boring "clip art" diagrams, but the message is astonishingly useful and can be applied in all situations, particulary in software development.
In hindsight it's easy to say something like "the red light was flashing to indicate the widget wasn't working. The pilot didn't do anything about it, he's obviously negiligent". At the time though the pilot may be dealing with 100s of stimuli and he doesn't know that this single red light is "the one" he should especially be paying attention too.
There's a quote about assigning blame - "Looking for scape goats is like peeing your pants, in the short term it feels pretty good but in the long term it makes you look pretty stupid."
Well, yes -- but the intent of making this deeper assessment is to get at the root of why the pilot may have made the error.
Let's say we do some deep dive assessment and we find that there's several contributions (all contrived for discussion): (1) during summer, sun sets directly behind runway 28R, (2) lighting system activates at 1 hr before sunset, (3) lighting system was refreshed with LED bulbs this year, (4) taxiway C is 30% wider than median taxiway in US airports of size similar to SFO, (5) the A320 (used by Air Canada 759) has reclining seats in the cockpit and this pilot was shorter than the prior pilot for this plane.
If you had an assessment like that you could reasonably take action on some of these without waiting for a fatal accident. The action wouldn't even necessarily have to be to remove/replace/alter these things, it could even be to commission a study to see the wider impact of LED lighting or reclining seats or something. Changing the runway orientation is a very large expense, but constraining 28R use during the critical sunset period is a little less so.
If this pilot made the error, it stands to reason that other pilots may make the same mistake. If we consider a near miss as seriously as we consider a fatal accident we can still learn great things. A near miss is likely only a failure of (N - 1) elements out of the critical N required for a fatal accident.
The point is that it doesn't help to know that it was a pilot error, because a pilot is human, and humans in the same situation might make the same mistake - and the main goal is to prevent accidents, not to have accidents happening all the time and saying "it was again human error".
It boils down to:
1 - Was the "error" intentional? If so, how to prevent someone from doing it again?
2 - Was it a real error, with no intention of any kind? If so, how to prevent that someone makes it again?
That's exactly the attitude that leads to more pilot error ;)
But seriously, it's important to understand how it made sense to the pilot from his perspective at that point in time. Because he believed he was aiming for the runway (I'm gonna give the benefit of the doubt and assume he wasn't distracted with online farming).
So understanding why it made sense to him can lead to actions to prevent this in the future.
Versus chalking it up to "pilot error", in which we assign blame and don't take action to prevent this kind of mixup in the future, thus practically ensuring it happens again. Because if it made sense to him, it'll make sense to someone else too. In fact the top comment says this type of mixup is rather common!
The tense/mood is wrong. It implies it could have been the cause of something that did happen rather than it having had a possibility of causing something that did not happen.
Text of article has correct tense: "In what one aviation expert called a near-miss of what _could have been_ the largest aviation disaster ever..."
Might have triggered: SFO near miss happened, either it's unknown whether the aviation disaster happened (but there are suspicions it did), or it's unknown whether the two events were related (but there are suspicions they are).
Would have triggered: SFO near miss did not happen, but if it had happened, it would have triggered the aviation disaster.
Could have triggered: SFO near miss did happen, but there was no aviation disaster, though there could easily have been one if it hadn't been actively averted.
> Could and might are often, but not always, interchangeable. He might have come / studied = 'We don’t know whether has come / studied or not', but you could substitute could. If you want to indicate ability or permission, however, you need He could have come / studied = ‘He had the ability to come / study.'
> It's possible that he came from a different culture would be used in considering a historical figure who didn’t seem to fit into his environment. If you were talking about a current situation, you’d say It's possible that he comes from a different culture. It’s hard to think of any circumstances in which It's possible that he come from a different culture would be used, even if you regard ‘come’ as subjunctive.
> English modal verbs are both important and subtle. You're unlikely to gain a full understanding of their use in exchanges such as this and if, as I assume, Noah, you are a non-native speaker of English you really need the help of a qualified English teacher.
A different kind of error... I was returning from Las Vegas in the middle of the day and the tower cleared us for departure on 9 and another plane on 27. We had taxied out and then the pilot pulled over, turned around and waited for the other plane to depart. He told us what had happened - there was a bit of frustration in his voice. Imagine pulling up and seeing another plane sitting at the opposite end of the runway ready to go. (it may not have been 9 and 27 I don't know which pair it was) Earlier waiting in the terminal I had seen a different plane go around, but didn't know why. Apparently there was a noob in the tower that day. This is why you look out the window and communicate.
Possible explanation for why this happened: it was night, and the parallel runway 28L was closed and therefore unlit. The pilot may have mistaken 28R for 28L and hence the taxiway for 28R. This comes nowhere near excusing this mistake (there is no excuse for a screwup of this magnitude) but it makes it a little more understandable.
I wonder just how likely this was to end in disaster. It feels overstated. The pilot in question seemed to think something was wrong, he just hadn't figured it out yet. I imagine he would have seen the aircraft on the taxiway in time to go around on his own if he hadn't been warned off.
I'm having trouble figuring out the timeline. The recording in the article makes it sound like this all happened in a matter of seconds, but it's edited down to the highlights so that's misleading. LiveATC has an archived recording of the event (http://archive-server.liveatc.net/ksfo/KSFO-Twr2-Jul-08-2017..., relevant part starts at about 14:45) but even those appear to have silent parts edited out. (That recording covers a 30 minute period but is only about 18 minutes long.) In the archived recording, about 40 seconds elapse between the plane being told to go around and the "flew directly over us" call, but I don't know how much silence was edited out in between.
Certainly this shouldn't have happened, but I wonder just how bad it actually was.
People "could" run their cars off of bridges every day, but they don't because they can see, and because roads have signs warning them of curves.
This sounds like a story of how well the aviation system works more than anything. The pilot is in constant communication with the tower. The system worked as intended here and he went around.
I'm pretty sure SFO always has functioning ILS, otherwise the airport would have to close every time there was even mildly bad weather. Each runway has its own system, though, so an individual runway's ILS can be turned off.
If there was fog, they wouldn't have been using 28L.
In a foggy situation they would've used ILS or even autoland. Looking at instruments instead of out of the window would've prevented this much earlier.
Because when you're coming in on runway 17L at DFW, all nice and programmed into your autoland, and suddenly the controller tells you to switch to landing on 17C, you have to just fly the plane. If you muck about trying to reprogram your autoland you'll be halfway to Oklahoma before you're set up again. See the video I linked in another comment in this thread for more details and examples of why complete automation is undesirable.
There were incidents in the past where pilots didn't have enough experience flying the aircraft manually. For that reason, most airlines prefer if pilots land manually as often as possible so that they can do it in case the system fails. If you always land automatically, pilots lose situational awareness.
Comment like that would put me off hiring you for any sort of mission- or safety-critical project. The approach for Asiana 214 looked fine until it wasn't; that was less than 4 years ago. Maybe you should stare at the video of that crash for a while and consider the large consequences that can ensue from a small error of judgement made a little too late.
One of my hobbies is shooting fire effects at big music festivals. We're shooting huge flames in front of 10s of thousands of people. I think that is pretty safety-critical.
We have a very rigid communication protocol that I designed based on my experience as a pilot and all the time I've spent flying with my father, who is a very experienced pilot.
The communication protocol and operational checklists are there specifically to catch human errors.
For instance: "snoop" (which is a sort of soapy water used for checking gas lines for leaks) every gas line any time it is changed. Snooping something, finding a leak, and then either reseating the gas fitting or shutting off its supply is the system working properly.
You could absolutely write an article that said "multiple gas leaks found on fire effects system", omitting the part where they were immediately found and fixed as part of the operation of the system.
I also wrote the ignition and control systems for a fire effect that we use, which has fail-safes built into every stage of its operation. That doesn't mean that failure modes don't ever occur (for instance: dead battery on the controller, out of range events, etc), it means that I designed around them so that they occur safely.
So I'm sorry that you wouldn't hire me for safety-critical systems. I'm curious what safety critical system you design or manage, because the idea that you either don't think errors will ever occur, or don't know what they are and how to design around them, is perplexing.
It's your 'miss is as good as amile' attitude that I don't agree with. The people who are pointing this out as a problem aren't clueless passengers who happened to be in the airport and were surprised to see something they didn't understand, but aviation safety professionals.
your basic claim here seems to be that journalists took an unremarkable everyday incident and wrote it up in sensational style to make it look bad. I don't think that's the case.
I would say that the approach for Asiana 214 looked bad from the moment it started, as the pilot felt uncomfortable performing a visual approach. Something has gone terribly wrong with training standards if a qualified airline pilot feels uncomfortable performing a visual approach in good weather into an airport with no tricky features.
Asiana wasn't a small error of judgment, it was a fundamental failure of training and over-reliance on automation which resulted in a pilot who didn't really know how to land.
That's the systemic problem that led to the error; I worded it poorly but the 'small error' I'm referring to is how the pilot came in just a little too slow and low. It's easy to think of even worse things a pilot might do when panicking, eg breaking off the approach only to stall and land on the tail, for example.
What I'm getting at is that a small error can have big consequences, and that this near-miss last week by the Air Canada plane is not something that should be dismissed as a non-story just because it was spotted and rectified in time. A collision avoided at the last minute is still indicative of a problem. If the Asiana pilot had stuck the landing that day but still felt uncomfortable, their training and operation culture at that firm would still have been problematic, they would just have gotten away with it for a little longer.
I see what you mean, and I agree. I'm sure the Asiana pilot (and many of his colleagues) had lots of uncomfortable but successful visual approaches before that. This event should definitely be investigated to identify any improvements that could be made. The tone of reporting seems to be going well beyond the severity of what actually happened, but it's not nothing.
The pilot not only had his plane headed "over a cliff" (lined up to land on the taxiway), he would have done so had nobody else caught the error. Causing massive loss of life in the process. This is officially a capital-P Problem.
Near as I can tell HIRL could not have been on, they were not following another aicraft to land, and the runway and taxiway lighting must've been sufficiently low that the taxi lights (low intensity version of a landing light) on the queued up airplanes on the taxiway, made it look like the taxiway was the runway. Pilot fatigue, and experience at this airport also are questions.
All runways have high intensity runway lighting (HIRL) and 28R has touchdown zone and centerline lighting (TDZ/CL). Runway lights are white, taxiway lights are blue. If you see these elements, there's no way to get confused. So my assumption is the pilots, neither of them, saw this distinction.
HIRL is typically off for visual landings even at night. That's questionable because night conditions are reduced visibility situations and in many other countries night flying is considered as operating under instrument rules, but not in the U.S. You do not need instrument rated aircraft or pilot certification. For a long time I've though low intensity HIRL should be enabled briefy in the case of visual night landings, where an aircraft is not following behind another, at the time "runway in sight" verbal verification happens between ATC and pilot.
Without knowing the cause but if I had to guess this looks like pilot error. At least statistically that the leading cause of crashes.
I am surprised pilots still manually land planes. Is the auto-landing feature not implemented well enough? But then it's relied upon in low visibility. So it has to work, they why isn't it used more often?
From what I understand pilots still practice landing in good conditions so they aren't completely out of practice if something goes wrong and they have to manually land the plane.
The point is that it doesn't help to know that it was a pilot error, because a pilot is human, and humans in the same situation might make the same mistake - and the main goal is to prevent accidents, not to have accidents happening all the time and saying "it was again human error".
It boils down to:
1 - Was the "error" intentional? If so, how to prevent someone from doing it again?
2 - Was it a real error, with no intention of any kind? If so, how to prevent that someone makes it again?
> doesn't help to know that it was a pilot error, ... and the main goal is to prevent accidents,
You can't prevent accidents if you don't understand their cause. Saying well we don't know what caused it but we'll just tighten all the screws a bit more and send the pilots for more general training. It just doesn't work that way.
Everything is analyzed and tracked to discover the exact root cause. It can often take many years and then new laws or guidelines are usually issued based on findings.
Too soon to tell what the problem was. Possibly there was some other problem in the aircraft that was taking some pilot attention. Give this some time.
Yes, the pilot made a mistake, but no, the cause of the accident isn't pilot error. From the point of view of preventing similar accidents in future, the cause of the accident is whatever caused the pilot to make the mistake.
AFAIK (not that I follow the issue closely) the problem of radio interference that ended the last-chance attempt to prevent the Tenerife crash has not been addressed [1]. If so, then it may be very fortunate that only one person called out that the landing airplane had lined up its approach on the taxiway, and not, for example, the crews of every airplane on the taxiway, simultaneously.
TL;DR: At Tenerife, both the Pan-Am crew and the tower realized that the KLM aircraft had started its take-off roll, and both tried to warn its crew at the same time, but the resulting radio interference made the messages unintelligible. The author states that a technical solution is feasible and relatively easily implementable.
Incidentally, I heard a story on KQED (SF Bay Area public radio) today that mentioned a potential clue. There are two parallel runways on this heading -- however -- the left runway is closed for repairs and therefore is currently unlit. If the pilot didn't remember this (it would have been included in his briefings and approach charts for the flight, but he may not have internalized it), he would likely have been looking for two parallel runways and would have lined up on the right one, which in this case would have been the taxiway...
I'd like to suggest that if you are still interested in learning more about what happened, you should look for a video from "VASAviation" on youtube. I'm sure his subscribers have asked him already for analysis and he's working on the video.
The channel focuses on aviation comms channel.
I find it informative because the youtube channel provides detailed voice/video/photo/analysis of incidents (actual/close-calls) involving planes/passengers taxing/landing/taking-off in/around airports.
I wonder why on 35R they wouldn’t have the taxiway to the left of the runway. Then the “right” is always the runway. Same for the left. Basically have parallel taxiways on the opposite side of the R/L designation of the runway. So at SFO, the parallel taxiways would be inside the two runways.
However, approach lighting is pretty clear, but at dusk, I agree with another comment that it can be rather hard to distinguish depending on angles. I think that approach would be landing into setting sun, so that could have some bearing.
SFO is said to be a somewhat abnormal/difficult approach [0]. This could cause stress and fractured attention in the pilot.
I wouldn't be surprised to learn that there was some other complicating factor. Planes don't usually crash (or nearly crash) without several safety factors being breached.
When the Asiana flight 214 crash landed in SFO in 2013, there was a blog post from a pilot who claimed to be a Lufthansa Air pilot.
He wrote that because of city of San Francisco's unusually stringent noise mitigation rule for airliners flying over the city to land at the SFO airport, an airliner has to make a very steep descent from an altitude higher than normal. So instead of bringing down an airliner gradually, a pilot has to get down to the airport in a hurry, requiring much more stringent management of energy/speed. It sounds simple for a plane that's flying straight in, but for a pilot at end of a 10+ hr flight, it's not always so simple. At least I think that's what the purported Lufthansa pilot claimed.
that's only in the (quite rare) case that the wind is coming from the south and the planes are coming in over Daly City .... normally planes land in the other direction and (like the Asiana flight) come in over the Bay.
They do however take off over a lot of homes in Daly City and are required to fly a flight plan that seems to postpone a lot of their climb until they are out over the Pacific
In Seattle, there have been taxiway landings in 2005 and 2015 because the taxiway looks similar to the runways. Presumably the ILS system should align them with the correct runway, but I guess that in these cases, ILS is not being used. Even it it were, the runways are the same heading and just a few hundred yards apart, so the mistake would not be realized until short final.
Seattle-Tacoma has three parallel runways with unequal horizontal spacing, it's easy to mistake that parallel taxiway between the center and right (north approach) as the right runway.
I've only done it in flight simulators but it's really difficult to tell from a distance.
I'm kinda surprised manufactures don't have full HUDs in commercial airliners yet. For airports that support ILS, it seems like using it should be mandatory, with a HUD providing a visual guide to ensure the pilot is lined up with the correct runway.
I mean sure, you could still have mistakes (pilot types in the wrong runway in a parallel runway airport and lands on the wrong one), but it'd be another check/safety system in place.
I'm not a pilot, or real-world ATC, but I was one a _virtual_ air traffic controller, on VATSIM[1]. I was a member of the virtual Indianapolis ARTCC[2], and have worked positions from Dispatch up to & including Center.
Listening to the LiveATC clip, the important bit is the phrase "visual 28 right", meaning the Air Canada pilot had been assigned a visual approach to runway 28 Right. The clip sounds like it's from Tower, so you're not hearing the point where the pilot was assigned that approach (that would've happened with Approach).
A visual approach is pretty simple: The controller sets you on a heading & altitude. Then, the controller confirms that you have the runway/field in sight, as well as any other air traffic that may potentially pose a conflict. Once the pilot verbally confirms those things are in sight ("traffic in sight" / "field in sight"), the approach controller will clear the pilot for the "visual runway XXX approach", also issuing appropriate warnings regarding wake turbulence.
Once the visual approach has been cleared, and the pilot has acknowledged, it is now the pilot's responsibility to navigate the plane to the runway, while keeping the plane clear of the traffic that has been pointed out to them.
If the pilot were assigned an ILS, LOC (localizer), LDA (localizer-type directional aid), or RNAV (area navigation) approach, that approach would have included a lateral component, which would have ensured that the plane was lined up with the runway, and not the taxiway.
So, why do a visual approach, when you have so many other options? Mainly because it allows more planes to be put into the same space. If a plane is on a different approach, or is enroute, or is being vectored by the controller ("turn right heading 350"), the controller is responsible for maintaining proper lateral (3 or 5 miles) and vertical (500, 1000, or 2000 feet) separation from other traffic. But traffic cleared on a visual approach does not have that restriction. Visual approaches are an important part of maintaining the high flow rates that Class B airports (SFO, ORD, etc.) want.
This comment's gotten pretty long, so I'll post another one with another possibility as to what may have happened! I'll leave you with the FlightAware radar trail, where (if you zoom in on SFO) you can see the go-around taking place.
So, I should clarify something: A visual approach is pretty simple from the controller's point of view; it's more work for the pilot. But, the pilot is always the one in command of the plane: If the pilot is unwilling to accept a particular approach, they just need to say so. It may lead to delays, but controllers will accommodate.
I originally said that the pilot had been assigned a visual approach to runway 28 Right. Looking at the charts for SFO[1], and the FlightAware trace[2], I have an alternative: The pilot may have been flying the LDA PRM runway 28 Right approach[3].
(You should open up [3] to look at while reading the rest of this comment.)
As has been mentioned, SFO is a bit weird: The two main runways are close enough together that it's not possible for you to have normal simultaneous approaches. For example, have a look at the diagram for Cincinnati/Northern Kentucky International (CVG)[7], another Class B field. Now, compare that to SFO[8]. At CVG, the parallel runways are far-enough apart that simultaneous approaches are possible, but with SFO, that's not possible. So, the FAA came up with the PRM (perimeter) approach.
Essentially, the plane approaching runway 28 Left is on an ILS approach. The plane on approach to Runway 28 Right is on an approach that is kindof like an ILS approach, but the plane isn't heading towards the runway, the plane is heading to a point to the right of the runway. Notice on [3], the point labelled "CFFKC". That's a fix (a point in lateral space), marking the point that you're being guided to. Note how the fix doesn't line up with the runway; that's why this approach is an "LDA (Localizer-type Directional Aid) with Glideslope" approach.
As per the detailed instructions (in [4], and point 4 in [6]), pilots on the LDA PRM runway 28 Right approach switch to a visual approach after the DARNE waypoint (3.4 nautical miles from the runway end). For most of the approach, they are monitoring a second frequency, which is only used if planes get too close together. But, once you pass DARNE, you no longer need to listen on that frequency. And although tower might have a radar view of surrounding traffic, it's not their primary tool (their eyes are).
If the Air Canada pilot was flying the LDA PRM RWY 28R approach, I think the final sidestep (from the LDA to the runway) may have been missed, or not completely executed, causing the pilot to line up with what he thought was the runway, but what was actually the taxiway.
Wow, that's a lot of detail and analysis. That does make a lot of sense. I know it will be a while for the FAA report to come out, but if it does turn out to be what you've stated, you can totally say you called it.
By the way, for anyone still interested a month or so from now, check out NASA's Aviation Safety Reporting System site [1]. This is a place where pilots, air traffic controllers, and other aviation personnel can go to anonymously report things that they were involved in. Reports are confidential [2] (they're anonymized before publication), and reports may not be used for enforcement purposes [3].
I expect that the pilots and controllers involved will have already submitted ASRS reports of the incident, and they'll eventually pop up on the site.
I think many of the readers here would be interested in virtual ATC, as many parts of the brain: Keeping many things in order, looking multiple steps ahead at what's going to happen (it takes time for a plane to turn), and communicating with your "customers".
The only downside is, you shouldn't sign in to control when you're on call 8-)
Aren't there lights that have to line up if you're on the right course for the runway like with nautic harbors?
Or warning lights that are visible when you're not aligned correctly?
How will an autonomous system handle this issue? Will it figure out the light colors of runways vs. taxiways or will it rely close geolocation capabilities?
> Off-Topic: Most stories about politics, or crime, or sports, unless they're evidence of some interesting new phenomenon. Videos of pratfalls or disasters, or cute animal pictures. If they'd cover it on TV news, it's probably off-topic. [1]
Is it just me or is this blatantly off-topic? Or is anything major happening in the bay area automatically on-topic for Hacker News?
There's something about air traffic that is very HackerNews. Search for "air traffic" to see what I mean. Maybe it's something about coordinating big complex systems, or maybe radio chatter is nerdy, I don't know, but it does make sense to me.
Maybe that's it then, still find it odd how the news that something almost happened is related to HN. I would understand it if the investigation showed a system failure or something similar.
That audio warning is only available on the newest planes, but even if it is available, if the ILS is off/not tuned in, there is no way for the plane to know it's not lined up.
I can understand wanting to hand-fly the plane, but ILS should always be on just as a reference/check.
I also think the linked article is making much to big a deal about this:
* The pilot checked in because he saw lights on the runway, he was probably very close to aborting the landing anyway.
* We don't know how far away the plane was when the go-around command was given.
* People overestimate how easy it is to see the runway/airport at night. You only really get to see real resolution when you're pretty close.
I am just a passenger, but this looks very over-blown. A pilot aligned with the taxiway, that's bad. But no pilot would ever land on a runway (or taxiway) with 3 planes on it. Just search the Aviation Herald for "runway incursion". And indeed, he spotted them, communicated, went around.
Aviation safety margins are so wide that this does not qualify as a near-miss.
> Aviation safety margins are so wide that this does not qualify as a near-miss.
I think you have this kind of backwards here. If you were going in for an amputation, and they only prepped the wrong leg, would that not be a near miss?
Rare incidents are hard to reason about and measure because they are so rare; but there are things that lead up to rare events - like being on approach to the wrong runway, that are less rare, and so they are easy to measure.
After you have plans and procedures to stop the actual rare disasters, reducing near misses is the most effective way to stop disasters.
> But no pilot would ever land on a runway (or taxiway) with 3 planes on it.
> After you have plans and procedures to stop the actual rare disasters, reducing near misses is the most effective way to stop disasters.
Oh, full +1. This is a bad incident that should result in an investigation and maybe safety recommendations.
My point is that aviation has been at it for so long, that a "bad incident" is very far removed from what the common interpretation of "near miss could have triggered greatest disaster" (#1 on HN) would suggest to a random observer.
(And I think the author knew it, and went for clickbait.)
Maybe, the author went for the clickbait, but only 584 people have to die for this to be the worst airline disaster in history. With potentially five planes involved, if he had crashed, I think that'd be highly likely in this situation.
Tenerife was very different. Here, the aircraft were at the beginning of the runway and clearly visible. In Tenerife, aircraft were at different part of the runway and couldn't see each other until it was too late. That's not comparable.
We know from the recording that the pilot saw planes on the taxiway and asked the tower about it. He would've never landed there after the tower confirmed that the runway should be empty.
I'm a private pilot who has flown into SFO once, and flies over SFO on a regular basis (I'm based at Palo Alto airport). This is not being overblown. If you listen to the audio tape, the tower orders the go-around at the 33 second mark, and United 1 reports that "Air Canada flew directly over us" 16 seconds later. That is a breathtakingly close call.
Safety margins are wide when pilots don't screw up. But this one did. Royally.
> Safety margins are wide when pilots don't screw up. But this one did. Royally.
And the safety margin was wide enough to absorb that screw up.
This is _why_ we have wide safety margins. So that one screw up doesn't absorb all of it. People will always screw up, and policy accounts for that fact.
Can you tell me what the second safety margin that kicked in here and saved them was? Seems to me that it was another pilot noticing them and asking "Where's this guy goin'?" which is not something I'd consider to be a safety margin kicking in... it sounds like sheer dumb luck, where the last margin had already failed.
* Conditions were clear. The pilot landing would likely have seen the aircraft on the ground in plenty of time to go around.
* Conditions were clear. The pilot landing would have an increasing probability of noticing their mistake from lights on the runway and taxiway as they got closer to landing.
* After being alerted that the pilot landing could see lights on what they thought was the runway, the ATC would likely have paid closer attention to the various tools available to them to see the location and heading of the aircraft.
And finally, the one factor that happened to be the first of these to happen:
* Multiple pilots looking directly toward the oncoming aircraft, one of whom alerted ATC.
Just because this happened first doesn't mean the other things failed or wouldn't have happened.
> And finally, the one factor that happened to be the first of these to happen:
> * Multiple pilots looking directly toward the oncoming aircraft, one of whom alerted ATC.
I don't get this.
You're an engineer/pilot/whomever filing for {the current system/pilot/plane/whatever-needs-approval} to get approved. The power-that-be (e.g. FAA) asks you what your safety margins are. You cite "other pilots looking directly at my aircraft in the 30-second span of my mistake" as one of your "safety margins"? ...what?
I'm inclined to agree with you. Airports tending to be full of hyper-aware pilots who are communicating on the same radio frequencies as you shouldn't really be considered a safety margin, but it can still prevent accidents sometimes.
Incidentally, Aviation Herald reports[1] that "AC-759 had already overflown taxiway C by about 0.25nm when ATC instructed the aircraft to go around" and further that "it is estimated that AC-759 overflew the first two aircraft by 100 feet".
100 feet isn't much, and I imagine this was utterly terrifying for the pilots lined up on the taxiway, but 0.25nm past the beginning of taxiway C is basically touchdown point on the runway (had they been lined up with it), and they were still 100 feet above the ground. That indicates pretty strongly that the pilot of AC-759 had noticed their mistake and was taking evasive action already, before ATC instructed them to go around.
Unfortunately we don't have enough information here to know for sure.
For example, the approach controller may have noticed that a plane they recently handed off to Tower was not yet lined up. The controller would then call up to Tower to make Tower aware of the issue.
Or, Tower may have a screen showing them a radar view of the approach, similar to Approach's screen. They may have noticed it there.
But, the nature of aviation radio is such that you use one frequency to transmit and receive. That means you can't speak over anyone else; you have to wait for them to stop talking before you can start. And if you mess up, there's no easy way for you to know (except when someone else on frequency says "blocked").
considering that jet turbines take 5 seconds to spool up to go-around and 8 to full thrust at landing airspeeds (...not entirely sure if that's correct - that's the best i could google), the margin was about 10 seconds. i'm inclined to agree that it should be a terrifying incident that ended in a successful landing.
The recording is edited for highlights. The LiveATC archive has a ~40 second interval between "go around" and "flew directly over us" and even that archive is edited to remove silence, so the real interval may have been even longer.
Can you post/link the unedited version? I can't find it.
Notwithstanding the timing on the tape, the AC jet could not have been very far away because the fact that he was lined up with the taxiway was first noted by a plane on the ground. That would not have been evident unless the plane was very close.
> But no pilot would ever land on a runway (or taxiway) with 3 planes on it.
Just like no pilot would ever crash into the seawall short of 28L in broad daylight with unlimited visibility. In that case the pilots knew what was about to happen seconds before impact and they were powerless to stop it. I could easily see the same thing happening here. Airplanes are big cumbersome objects with a ton of inertia.
Though air traffic was not heavy at Los Angeles airport, the local controller was distracted as Flight 1493 was on final approach by a series of abnormalities, including an aircraft that inadvertently switched off the tower frequency and a misplaced flight progress strip, which resulted in the SkyWest flight being told to taxi into takeoff position while the USAir flight was landing on the same runway without the Metroliner ever being given a takeoff clearance.
Upon landing, the 737 collided with the twin-engine turboprop, continued down the runway with the turboprop crushed beneath it, exited the runway, and caught fire. All 12 people aboard the smaller plane were killed, as well as an eventual total of 23 out of the 89 passengers on the Boeing.
Assuming the liveatc audio in the article has not been edited, it sounds like there's about 10 seconds between ATC giving go around and the United pilot saying they flew right over. With instructions given to Air Canada in between so it might have been seconds earlier.
Based on that, it was seconds away from a collision. It seems like a near miss when considering aviation safety margins.
I would assume it was edited. He only joins the frequency about 30 seconds before it happens. That means tower only took control once he was 60 seconds before landing. I don't know the procedures at SFO but that seems too short.
It has been edited. The LiveATC archive has about 40 seconds between "go around" and "flew directly over us." The LiveATC archive appears to be edited to remove silence, so it may have been even longer.
What's missing is, when (relative to radio traffic) did the pilot begin the go-around maneuver. Did they begin doing so before being told to go around, or did they only start to do so after?
If the pilot initiates a go around they need to announce it to the ATC. It is usually a drastic full power pull up that will certainly disrupt nearby air traffic.
So the pilot probably did not start the go around before being told so.
Sorry, I have to partially disagree. The FAA is pretty clear on this: The priorities (in order) are Aviate, Navigate, and Communicate[1].
The pilot is ultimately the person responsible for the safe operation of the aircraft. That will occasionally mean performing a maneuver before telling anyone.
In response to this, procedures are designed so that pilots can safely abort an approach without posing a danger to themselves (i.e. running into terrain) or other aircraft.
For example, let's look at the LDA PRM runway 28 Right approach [2] into SFO, which the pilot may have been flying. If you have to abort the approach prior to the DARNE waypoint, you make a climbing right turn to heading 030 and 3000 feet, then you proceed directly to the Oakland VOR, and then you hold as published.
When Approach clears you for this approach, the controller is also clearing you to fly the missed approach, if needed.
If you have to go missed-approach after DARNE, the instructions are similar, but are documented in [3] instead of in [2]. But note the "unless otherwise instructed by ATC". In this case (see [4]), Tower placed the pilot onto runway heading, likely with a climb to avoid the hills. Approach would then have taken over, and brought the pilot back around.
This isn't quite a regular "runway incursion". That term is usually reserved for a taxiing aircraft (or maybe a service vehicle) which didn't stop in time at the threshold and went a bit beyond it.
This is more like runway "confusion". A term I made up, not sure if there is such an official name for it.
Right, I meant that if pilots sort-of-routinely spot runway incursions, they surely won't land on 3 wide-body jets sitting there.
This is why you want human pilots, instead of computers, not the opposite. A human can apply common sense and decide not to land on three planes after all even if it looks like a cleared runway.
> This is why you want human pilots, instead of computers
I actually believe the opposite is true. Systems like ACARS have managed to avoid the prevalence of air to air incidents that used to be almost an epidemic in the 60's and 70's.
Autolanding technology has brought planes home safely where a pilot would have had to go around and find an alternate aerodrome.
Sorry yes (Duh!) TCAS, not ACARS! Thanks for the correction. I haven't flown in a LONG while now, and sometimes the acronyms get muddled in my brain. :)
Ehh, the fact that he was cleared to land and the taxing aircraft could tell he was over the taxiway means it was probably less than 1 minute from landing. That's a pretty close call imo.
>Surely the pilot would notice at least one of the planes he'd be inevitably crashing into well before he started to land
In theory yes, but the risk is that the plane you crash into is not yet on the taxiway or that poor visibility or tunneled focus prevents you from noticing it.
You don't need to be moving quickly to have a significant accident. The deadliest aviation accident in history (583 fatalities) resulted when two planes collided on a taxiway (edit: actually on the runway) in dense fog [1].
Edit: as correctly pointed out below, the actual collision took place on the runway at significant speed.
That was just one of seven distinct mistakes that lead to the crash. The PanAm pilot missed his turnoff, the controllers were busy watching a football match and missed the radio cues that could have alerted them to the fact there was an impending problem etc. etc.
At night, and if he was fatigued or on instruments for a long time before that, the disorientation could have easily lead to such an error.
LOTS of planes accidentally line up on the taxiway if visibility is reduced and cockpit workload is high. Heck, aircraft have even taken off on non active runways at the most advanced airport in the world with disastrous results at night. [0]
Inclement weather on approach at the end of a long flight can also lead to crew pressing home while under confusion, as in the case of the Korean airlines accident that landed short. [1]
More recently, at SFO too, the Asiana B777 crash was something that should have been almost impossible in a modern jetlines, but over reliance on automation and lazy piloting lead to that disaster.
I expect you are correct, but I also note that it has happened before. One could have said this just before the Asiana flight flew into the ground as well. "What pilot would miss the fact that they were way too low?" Or just before Harrison Ford flew over a plane on the Taxiway and then landed on that taxiway.
Fundamentally it takes several things to fail all at once for an air disaster to actually transpire, but it does happen.
The Asiana problem is that some asian pilots are encouraged NOT to fly manually ever.
There is a pprune forum post of an american instructor explaining how a visual circuit flown in manual was the worst "emergency" he could set in a simulator. Up to the point that he failed one of the chef pilots in Asiana and they wanted to fire him.
This seems to be a cultural thing, they ace the theory (really know all the manuals by heart), but are not allowed to think outside the box at all.
This airmiss is important, but could only happen flying in manual and visual conditions.
It's hard to ignore the sigthing of other planes in that conditions, even if you are confused about which strip you are aligned with.
In my opinion it's much more dangerous taxing in a complex airport with fog. There is a real problem of runway incursions, that could be solved with better systems.
Then this specific type of incident would likely not have happened.
The pilot here was flying visually (which I expect either means a visual approach, or else the LDA PRM Runway 28 Right approach).
If the airport was fogged out, the pilot would be on an ILS or RNAV (area navigation) approach. The former uses ground-based signals to provide precise lateral and vertical navigation; the latter uses either GPS or multiple ground-based aids to accomplish the same thing.
The downside to ILS, RNAV, etc., is that you can't fit planes as close together while still being safe. That's why you get delays when major airports have reduced visibility: The flight scheduled are "tuned" for maximum planes per runway per minute, which is only possible when skies are clear and winds are (relatively) calm, so that you can use visual approaches.
Theoretically - if the plane had landed, how many planes would it have taken out? It obviously wouldn't have been pretty, but I doubt the AirCanada would have reached the fourth plane, or maybe even the third.
I always wondered what about SFO makes it so much more dangerous than the other airports in the area? It seems like they have a potential disaster every couple years.
Paint the runway and the taxiway in different colors and also use different colors for the light signals that illuminate them at night. Blue/white is rather confusing. Use clearly distinguishable colors such as red/blue or orange/blue or magenta/yellow.
Another comment mentioned green vs white. Even though it's practically impossible to get an aviation license with deuteranopia that just seems like it's asking for trouble. White light is rarely pure white and depending on the weather conditions I have a hard time believing the contrast is unambiguous.
Heck, maybe aviation should take a page from basic accessibility and figure out that information should be conveyed more than one way. Don't rely on color alone (and no, spacing of the rows of lights isn't unambiguous either, especially when looking at varying distances and angles).
[2] https://news.aviation-safety.net/2010/12/03/serious-incident...
[3] https://www.flightglobal.com/news/articles/hk-airlines-tries...
[4] http://www.dailymail.co.uk/travel/travel_news/article-337864...