So from the ATC recordings it seems like everyone was made aware of the traffic, i.e. the Southwest flight acknowledged there was an approach 3 miles out and FedEx acknowledged the 737 ahead of them. i.e. it's not like the ATC forgot it gave one clearance or the other. So it seems like the Southwest takeoff took longer than expected and ATC lost track of the situation and thus did not revoke clearance to one or the other? This is not me assigning blame just trying to get a sense of the context.
Edit: in the tower comms the Tower asks Southwest to confirm it's on the roll just 16 seconds before the call from FedEx for Southwest to abort takeoff. Incredible the tight margins these guys play with to keep things moving at the airport. People are speculating the tower had no visibility of the runway traffic due to weather.
How close they put the traffic reminds me of "Pushing Tin". Great movie with perhaps more bits of truth to it (perhaps in the past) than we want to know.
Pushing tin was a horrible movie. The whole ATC part was just a writers dream of how to make a gripping experience. General public only gets a wrong idea and anxiety.
> For starters visibility was down to 1/8th of a mile. So tower likely never saw how close this was. Second off apparently Austin does not have ground radar. Regardless of that though this is the second near catastrophic accident we’ve had in two weeks or so here in the US alone.
Once? There was also a recent Cloudberg article about a place radar would've been real handy to have. (Plane and ground thought they were S of Kathmandu when they were N, until they suddenly went silent, due to hitting a mountain.)
That’s a different problem. Regular radar is setup to detect flying aircraft and is very useful to avoid flying into mountains, avoid mid air collisions, etc and it’s not what we’re talking about. That type of Radar has proven necessary on many occasions.
Spending a great deal of money to track aircraft on the ground however is only helpful when one of the aircraft is on the ground. It’s only partially useful when visibility is low, at which point air traffic control can be especially paranoid.
This is like saying "I don't wear my seatbelt because I've never been in an accident". Safety systems isn't about one or the other, it's about having multiple systems and procedures in place to prevent a disaster. Proper procedures are good, and it's a first line of defense, but humans make mistakes and are imperfect in other ways (Such as not being able to see through fog). Proper procedure in the case of near 0 visibility would be to greatly increase separation between planes, something which wasn't done in this case, and something which would slow down traffic a lot more than passengers and companies like.
It’s more like saying school busses don’t need seatbelts because based on a great deal of evidence, they don’t significantly improve school bus safety.
In this case the accident was avoided because aircraft could see each other long before there was a meaningful risk. It wasn’t a mismanaged zero visibility situation but rather proper procedures working.
It would interest me to know what the reasons are for ground radar not being more widespread and how many accidents are preventable by ground radar yearly. My first guess is that incidents where ground radar would've prevented them are so rare that it isn't worth the expense or the additional complexity or additional point of failure, and it's more productive/effective to solve the issue at different points in the system instead.
FR24 is a combination of ground-based radar (both primary and secondary surveillance) and ADS-B data.
TCAS (as I understand it) uses the transponder part of secondary surveillance radar, but with out the bearing/ranging info, as well as ADS-B (if equipped).
TCAS is interesting but it’s not really relevant here. It’s something that is applicable to the en-route portion of a flight, when both planes are supposed to be separated by substantial distances while in the sky.
TCAS is useless in the terminal environment, like while landing or taking off. If it wasn’t inhibited (which it is) it would be correctly screaming that there are about to be a dozen airplanes too close for comfort every single time it got near an airport.
It’s a system for avoiding mid-air collisions not stuff like this.
My understanding is that when an aircraft is < 1000 AGL the TCAS would be at Sensitivity Level 2 (SL2) and they would still get Traffic Advisories (but not a Resolution Advisory); and only once landed and the pilot turns off TCAS would they no longer get messages?
These systems exist (STCA/RIMCAS for the tower, TCAS for the pilots, but not around landing), and for all we know at this point they might have helped avoid a collision. The challenge is balancing giving useful alerts while not feeding the pilots noise at the point that you want them to be absolutely dialed in.
Nothing currently comes close to human intuition for reliably detecting these sorts of collisions. Fundamentally when a plane is on final at an airport that's even remotely busy it's always on a collision course with planes on the runway taking off or landing.
Lots of comments and obviously this seems like a big deal - but it usually isn't.
Disclaimer - PPL, fly mostly twins under IFR, in Australia (still ICAO, if anything we're more or less a copy of the FAA).
Most heavy jets are covering between 2-3 miles a minute when coming in to land. Usually a 3 mile call is to affirm to ATC that they are a minute away (or two for GA traffic).
Most heavy jets also need anywhere from half a minute to three get airborne depending on size / weight / weather.
On a clear day - you can see an aircraft lining up to depart so you can slow down a bit to give yourself some space. Usually ATC does that for you - asking one to slow down or the other to hurry up.
Sometimes the sequencing gets a bit tight, it's annoying, you just power up and go around. Depending on visibility - you can circle in the circuit area to come back around, or follow the missed approach and line up to fly the approach (ILS, RNP etc) again.
Now the hard thing here is that it was IFR weather - the incoming aircraft got visible with the runway when it was lower, and it didn't look like the departing aircraft was lifting off so they went around. It also means they have to follow the missed approach.
Everyone was (mostly) aware of where everyone was, there was no real breakdown in communication nor in procedure. Departing aircraft knew there was one incoming and thought they had time, incoming knew there was a plane leaving and knew there might not be enough time. ATC is trying to keep the ball rolling for both. It's happened dozens of times for me.
Final note - a Missed Approach procedure specifically takes you to safe place away from departing aircraft, who follow their own Standard Instrument Departures precisely because of this scenario. It's incredibly rare for these two procedures to converge / overlap.
P.s. I've gone around / gone missed for everything from slow aircraft to gliders to a snake to wind shear to kangaroos. Literally the last few things just before landing you check are that your gear is down, you have clearance to land and the runway is clear.
Was a professional pilot for many years before illness.
This post above is not correct - this was very close by all accounts. Also, recall they had something like 1000 RVR on the field. This was very close to a disaster.
Based on a similar incident I witnessed in the `90's, everyone in Austin can expect their evening news to feature a segment on how safe it is to fly in and out of Austin.
Thanks for the clear explanation. That is a great insight.
Based on this I would expect goarounds to be more common but I
barely remember witnessing one during my travel (not a heavy traveller but lets say
about 100-200 flights total). Airport diversions yes once at least.
So how often would you expect a go around in an airliner as a percentage of attempted landings?
This is crazy. Another post I saw shows FlightRadar24 data with the fedex flight at only 74 feet above the runway when they started to climb. Fedex pilots need a medal and probably a new pair of pants. Can't wait for blancolirio to break this one down, very curious how this happened.
If you are the pilot in a situation like this, does the airline say, "Alice, you did good. Take the next month to yourself." Or is it just business as usual, "See you on Monday." kind of deal?
My understanding is pilots at the majors (UA/DA/AA) have fairly liberal no-fault call out policies regardless of at-work incidents.
Your wife asked you for a divorce last night and you can't focus on work? You can call out without consequences.
You just couldn't sleep and don't feel able to fly? Call out.
They understand the importance of the human factor here and would much rather have to reschedule a flight than have a potentially impaired person make a mistake.
Obviously if you're doing this regularly, you're going to have a conversation with management or the chief pilot or someone (I've no idea what the actual reporting structure is) but if you don't abuse the policy, you can call out without consequence.
I don't know specifically about FedEx but I imagine they have something somewhat similar, the FedEx/UPS pilot jobs are also considered top tier (but not nearly as top tier as a major long-haul position) with good benefits.
I do imagine, especially considering how much press this is getting, that the pilots in this case are going to spend the next week or two doing paperwork and maybe some press even. How quickly they return to flight will probably be dictated primarily by how they feel. Fact is as dramatic as this seems to us it's possible the pilots are completely unfazed and just want to get back to work. I think if they wanted a week or three off to recoup that FedEx would make accommodations. Pilots seem to be treated reasonably well at major airlines (ignoring regionals and subsidiaries, where they are cattle) and the possible negative value of negative press (FedEx fires pilot who saved plane from near-miss for asking for time off to recover!) this should be a no-brainer for them.
I have to strongly disagree with your comment about shifting blame.
Pilots and airlines do not seem to shift fault away from themselves but seem to do a phenomenal job of identifying root causes to avoid future incidents.
Yup. Sometimes a pilot has to make a decision that there is no training for, decides wrong, people die, but still has a job because the investigation revealed there was no way to know how to make the right decision. Interesting case study: https://admiralcloudberg.medium.com/man-vs-runway-the-crash-...
The Wikipedia info box summary is "Runway overrun caused by pilot error". And indeed, the pilots were going too fast with the flaps set wrong. At some point, they indeed faced a difficult decision, but one that was caused by their own, multiple previous errors. Having made that decision incorrectly, they also failed to stop the aircraft as well as they could. From your link:
> In the end, the NTSB identified two separate decisions by the captain which led to the accident: first, his decision not to go around immediately when the plane started to float, which was inconsistent with his training; and second, his decision to finally go around after it was already too late to do so. The float itself was caused by excessive airspeed on touchdown, in combination with an ill-timed gust of wind. In a later revision suggested by the Air Line Pilots Association, the NTSB also added the “non-availability of information about the aircraft’s go-around performance capabilities” as a contributing factor to the captain’s decision.
37 people died and the pilots kept their jobs. Truly amazing job security, much like this other example where the pilots almost killed 1000 people (but didn't in fact kill any): https://news.ycombinator.com/item?id=34395221
> seem to do a phenomenal job of identifying root causes to avoid future incidents.
When you see the regular incidents that do not result in crash, that statement would be obviously regarded as false. Yes, things are mostly reliable because most of the time there are at least several safety features or processes in the way, but most crashes and incidents are still due to human error.
I wonder how often these close calls occur. I was on a flight a few days ago and using an app to track my flight while on the plane. We were at 32,000 ft and according to the app another aircraft was headed straight toward us at 33,000 ft. Just before the two planes overlapped on the app, I looked out my window and saw the other plane go right above us. It was absolutely insane how close it appeared and how astonishing a 1,000+ mph speed difference is. (If you're curious which flights: DL75 and TK1344 on Feb 1, 13:04:20 GMT).
My understanding is that planes passing this close to each other is relatively common, as 1,000 ft is the minimum vertical separation for RVSM airspace, but I was still surprised that the other plane was directly above us and that some sort of optimization algorithm isn't used to maintain horizontal separation as well just in case there is an error in communication on the flight levels.
Airline pilot here: passing another aircraft opposite direction on the same lateral track with only 1000ft of vertical separation is about as normal as it gets. I’d say that it has occurred on almost all of the flights in my career so far. In certain areas of poor ATC coverage we sometimes use SLOP (strategic lateral offset procedures) to add a bit more margin for a height error.
Edit: fun fact, you can actually hear the passing jet from the flight deck just after it has passed, a slight whoosh can be heard. Depending on conditions you can also feel a brief bump when crossing their wake (since the wake descends).
That's really cool! I can only imagine from a pilot's perspective it must look like you're about to collide though; that would be a bit nerve-racking for me.
Each plane was at the exact same height, and given their digital altimeters they were at the exact same height and one planes wing clipped the one going the other way. Only the smaller plane crashed.
Commercial airliners have the Traffic Collision Avoidance System that operates independently of ground based systems or ATC clearance to avoid collisions. Adding more separation would be the opposite of an optimization algorithm in the case of transoceanic flights - it would reduce the number of planes able to fly in the most efficient parts of the jet stream at one time.
What’s interesting though is that you were not flying at 32,000 ft, but rather flight level 320, which actual altitude was determined by a barometric altimeter set to a standard pressure. It may deviate considerably form the actual altitude of 32,000 ft above the mean sea level, but the deviation is dictated by local atmospheric pressure, thus shared by all aircraft in close proximity.
One minor potentially life saving standard protocol is that flights are supposed to stick to certain elevations depending on the direction of travel. Half remembering this, but East flights are to be at odd elevations (31,000; 33,000; etc) while West bound are on evens (32,000; 34,000; etc).
So both interpretations are correct. This reminds me of a comic strip (I think) where they're training someone working at a nuclear reactor and the instructor says, "Just remember: you can never have too much water in a nuclear reactor." The phrase can be interpreted to mean two very different things.
If I change both words of the phrase ("near" to "almost" and "miss" to "missed"), also changing both parts of speech and the overall phrase type in the process, that does shift its meaning, yes.
I'm familiar with the claim that "near hit" ought to make more sense (and with the Carlin sketch that either originated or popularised it). But in practice, people use the phrase "near miss", not "near hit" (unless actively making the George Carlin joke), and they don't use "near miss" to mean what you claim ("almost missed, but didn't" i.e. hit). So this, again, isn't really a claim about what the phrase can mean, but about what you'd logically want it to mean. Such claims are a bit counterproductive at the best of times—language doesn't always follow compositional logic, and sometimes we just have to live with that—but here it's especially odd because the phrase means exactly what the component words say it does: a "near miss" (noun phrase) is a miss (noun) that is near (adjective).
I think Carlin missed the funnier part of "near miss": the construction implies that every time a plane taxis, takes off, cruises, lands, all without hitting anything else, that it merely missed all the other planes! We could look at the planes in the sky and say "that's a miss, that's a miss, look how much that one missed by, how boring" It's only when they miss by a little, when they're very close, that we we get excited and say "wow, what a near miss!"
I've always thought that "cleared to land" instructions are a little too optimistic. Multiple planes can be cleared to land on the same runway, and planes can be cleared for takeoff on that runway at the same time. The logical thing to do would be to only have one outstanding clearance at a time, and require landing aircraft to go around if they're not cleared by, say, the decision height on the approach. (Not sure where this point would be on a Cat III ILS approach, though.)
Doing it "my" way would probably reduce throughput, and given that thousands of aircraft a day are cleared to land while the plane in front of them is still in the air, it must not be too dangerous. But I have always thought it's weird. (I did some googling and apparently they do it "my" way in every country except the US? Figures!)
Only one aircraft is actually allowed to depart/land on a given runway at any one point in time.
You can have clearance to land after the aircraft in front has landed or departed, but broadly two civilian aircraft occupying the same runway is a no-no if either are taking off or landing. Seaplanes is another matter entirely, and of course there are dozens of exceptions.
You can have multiple aircraft cleared for a departure or arrival or approach - but note that this is not a clearance to take off or land.
I've been given clearances to land roughly 5 seconds before my wheel touch and my hands were just about to shove power in to go around. I've been told to go around after my wheels are on the runway too.
We usually know what ATC is doing or waiting for and we can anticipate those calls.
I don't think the US is different from the RoW in that manner - they all need to conform to the ICAO regulations - otherwise we'd have chaos with US pilots navigating China or vice versa.
In the US, clearance to land does not mean that other aircraft won't land, take-off, or cross your runway before you touch down.
> Landing clearance to succeeding aircraft in a landing sequence need not be withheld if you observe the positions of the aircraft and determine that prescribed runway separation will exist when the aircraft crosses the landing threshold. Issue traffic information to the succeeding aircraft if a preceding arrival has not been previously reported and when traffic will be departing prior to their arrival.
My mains had just touched and ATC called for a GA, they obviously saw something I didn't and I was airborne a few seconds later. Circled and landed with no issues.
It could've been anything from metal on the runway to an aircraft not responding or reading back in time while taxing to a priority medevac helicopter.
That's insane. Insane because I'm sitting here typing at the computer at 750 MSL / 200 AGL. If I look to my left, I see AUS approach parallel to the horizon about 3 miles / 5 km away.
PS: Located in one of the most gerrymandered districts in the entire US, spanning Austin to San Antonio with a sliver of connection without constituents spanning I-35 for a distance of 80 mi / 130 km. The district looks like a broken q-tip or a lock pick.
They were following standard procedures here for going around and not improvising. Standard procedure is to climb out runway headings usually and then make some turn. In the same way, the pilots would have seen with their eyes before the minimums threshold that the runway was not clear and made the go-around decision. It's not a great situation but I don't think it was that dangerous ultimately.
Any kind of turn costs lift, and that’s not what you want to spend when you’re trying not to hit something on the ground. As such turns are generally not performed below a certain altitude.
It very subjectively feels like the number of incidents i've seen in post-covid commercial aviation has increased significantly, but, again subjectively, I'll say I don't think ATC is the issue.
In the incident last month when a DAL flight had a rejected takeoff after a UA flight crossed their runway, the UA flight crew was clearly at fault when you listen to the transcript and barring any major new evidence i'm sure the FTC report will agree.
The flight in November where a LATAM flight collided with a fire truck was also not the fault of ATC, thought we might never know exactly what caused the runway incursion there because the driver (and everyone else) in the lead vehicle was killed.
How does “at fault” really apply here? Shouldn’t these be blameless post mortems that identify what needs to change for everyone to avoid an incident like that in the future? Not sure assigning blame to a single crew really helps in moving forward unless there’s a pattern of repeated negligence by some individual or group.
Being at fault doesn't mean you're being blamed. It would be really stupid to have a postmortem if nobody is indicated as having done something wrong.
A pilot may have made an error and been at-fault, but blamelessness means that you focus on preventing the issues that led to them making the error. The point is to acknowledge that humans make errors and to fix the conditions that allowed the human error to take place to begin with. The point is not to avoid identifying who made the error.
Of course, gross negligence is still deserving of punishment, especially when lives are at stake.
> It would be really stupid to have a postmortem if nobody is indicated as having done something wrong.
This sounds wrong to me. If the processes in place allow an incident when everyone is properly following the process, that seems like the most important time for a postmortem, since the existing process clearly is insufficient and must be adjusted.
But if someone wasn't following the process, and you avoid stating that they made an error for the sake of blamelessness, you're avoiding talking about how the incident occurred. That doesn't mean the process doesn't need fixing.
"Nobody done wrong" happens in aviation accident investigations.
I remember one about a collision between a Bundeswehr fighter jet and a glider over the Black Forest. The detailed report found that both aircraft were where they were allowed to be, the fighter pilot had little chance to see the glider, and it was an unfortunate event.
A blameless postmortem can apply in some circumstances, especially when there are multiple confounding factors but in the case of the JFK runway incursion it really doesn't. There were many confounding factors but they were all sins committed by the UA crew.
The UA crew basically got lost on the ground at JFK and drove onto an active runway without clearance and then argued with ATC about it. They were clearly negligent. ATC gave them clear instructions, there was clear and correct signage, they mistook a runway for a taxiway.
> The UA crew basically got lost on the ground at JFK
> there was clear and correct signage, they mistook a runway for a taxiway.
This would contradictory claims to me. I have 0 experience in aviation, but just running a general post mortem, I would say for whatever reason they got lost and the sins of the UA crew don’t matter. Unless the UA crew was doing this deliberately, either JFK signage isn’t actually as clear as you claim or JFK’s runways are so complicated that they need dedicated training. Them arguing with ATC is also strange and I’d like to better understand why the crew thought they needed to do that and resolve that underlying problem.
Basically, devoid of malicious intent, what are situational changes earlier that would have prevented the sins from occurring in the first place. For what it’s worth, that’s the way self driving cars are developing. People worry about the trolly problem when I’m fact software just tries really hard to avoid getting into that situation in the first place.
To be fair, the one that happened at JFK isn't looking like it was ATCs fault, but rather a possible pilot deviation. Maybe ATC could have caught it earlier? I don' think a report has been produced yet.
But in general there is some concern in the industry due to the COVID induced reduced number of flights in the last years, which led to less "day to day work/practice" for both ATC and crew. Simulator trainings don't always necessarily cut it.
I think ATC is overworked and there hasn't been as many advances in technology in that sector as well as under training for years. Honestly surprising that planes don't collide more often, the world around us relies on airplanes for logistics but everything about the entire aviation field from pilots to mechanics is a battle of never having enough people.
"quiet quitting" is just starting to be used as a substitute for "the lack of our workers ability to feed themselves and focus on their work is starting to impact job performance"
> On 1 December 2003 at 6am, a major disaster in the stack was narrowly avoided. An air traffic controller was blamed by a later inquiry for misdirecting traffic when he ordered a United Airlines Boeing 777 into a level of the Bovingdon Hold (or stack) already occupied by a similar British Airways plane. The two planes, carrying 500 passengers, flew within 600 vertical feet (180 m) of each other.[4]
Edit: in the tower comms the Tower asks Southwest to confirm it's on the roll just 16 seconds before the call from FedEx for Southwest to abort takeoff. Incredible the tight margins these guys play with to keep things moving at the airport. People are speculating the tower had no visibility of the runway traffic due to weather.