Dedicated thread on point:
The two planes appeared to came within 75 vertical feet of one another.
A video:
“Simon says, Fed Ex 1432 Heavy clear to land 18L” “Watch for aircraft in your vicinity and report empty runway in sight”
I hope the fact that Southwest was able to depart on time is not overlooked.
Here’s a bit more of the audio:
https://s.broadcastify.com/audio/KAUS-Twr-2023-02-04-1230z.mp3
It doesn’t sound like great weather, the “RVR” numbers (visibility at different points along the runway) given by the tower controller are fairly low and FedEx was doing a Cat III ILS which means they’re expecting to be in the cloud pretty much down to the ground. An odd time to be clearing someone to take-off when you’ve already cleared someone else to land and they’re on a 3 mile final.
Some general thoughts:
-
Austin ATC is somewhat famous for their cowboy “keep-em-movin’” approach to traffic control.
-
Southwest is famous for their “We’re always in a hurry and do everything faster than everyone else and deserve to go first” flying style.
-
KAUS is also a kinda-hub for Southwest where they have lots and lots of departures all day long. It’s also a booming city with massive growth of air service by the other carriers. It’s a hustlin’ kinda place. It might be in Texas, but it (the airport at least, if not the entire city) has a NYC kinda vibe.
-
At that time of morning, the ATC tower is still staffing up for the day shift. Overnight one person is doing clearance delivery, ground, tower, and all the intra-ATC coordination. Mid-day that work is divided between 4 or 5 people. Many airports are short-staffed by FAA. The traffic starts rising in the hour+ pre-dawn before ATC has enough people on-shift to easily keep up. The poor ATC bastard handling that ramp-up is often busier than a one-armed paperhanger. I don’t know that AUS is any worse than many other second-tier airports this way. But it does have this issue.
And the poor solo controller may be nearing the end of their graveyard shift while all the departing planes are fresh newly-awake crews impatient to get going. You can sometimes readily tell by their voice & cadence who’s fresh & full of coffee and who’s just about wasted at the end of their shift. -
The KAUS airport is famous for very thick fog around sunrise after prolonged rain or snowfall the previous day. Dallas (both airports) is just 200 miles away and has similar problems with thick early morning fogs. There has been a LOT of precip over the last few days, much of it frozen. As it thaws then evaporates, that creates ideal conditions to keep the air near the ground saturated.
-
The tower controller was doing all this by radar. They could not see the taxiways, runways, or any of the airplanes. Their head is literally up in the clouds.
From the audio this sounds like the pacing would have been fine (if maybe a tad close) in clear beautiful weather but was mistakenly applied to a low-vis situation where margins need to be much larger.
Southwest took a completely reasonable amount of time to get going under the conditions, but it was a little more time than is typical for them, and that ate all the margin and then some. They did not give off any I’m in a hurry vibe to me. Although given what they were told about Fedex’s position when cleared onto the runway, it was less-than-conservative to have continued versus saying “No thanks, we’ll wait until after the traffic (FedEx) has landed.”
The FedEx pilot’s check-in was a little confusing to me. The good news was he did not sound tired or inattentive. He reports being at “5.4”. Which number I initially took to be an altitude (5,400 feet above sea level, about 5000 feet above the airport) which would place him on about a 15-18 mile final = 5+ minutes from touchdown. From the rest of the timing of the audio, he seems to have meant he was 5.4 miles out from the runway, or about 2 minutes from touchdown. Is this discrepancy because the audio was edited to remove ~3 minutes of dead air and extraneous radio traffic? Or is the audio real-time but the ambiguity in FedEx’s phraseology subconsciously set the controller up to be thinking FedEx would get to the runway about 3 minutes later than he really would? Which discrepancy only came to light waay late in the process? No way to know. Yet.
So that’s some plausible whats. Getting to the deeper whys will be important.
But there is a strong similarity to the JFK close-call last week. Namely one of the participants had a mental model that did not comport with reality. And in following their model things got hairy. Reality intruded into the defective model real late, but not too late. Good thing.
How many people were on the planes?
I don’t immediately have any source for the actual totals, but the Fedex would generally have only the 2 flying pilots aboard.
As to Southwest … Interestingly their website does not have a fleet description page (that I could find) that shows the cabin layouts and capacities of their various flavors of 737. Wiki says the -700s have 143 passenger seats and typically 5 crew = 148 typical capacity while the -800s, and -8MAXes carry 175 passengers + 6 crew = 181 typical capacity.
Both FedEx & Southwest jets have some extra jumpseats aboard to carry additional crewmembers just riding along. But at most that’d be 7-10 people between both aircraft if every such seat was occupied. Which would be a statistical rarity. And of course Southwest might have had a few lap children aboard as well who don’t consume a seat.
Most of Southwest’s fleet is the smaller -700 model, so if I had to place an over/under wager I’d go with about 145 people total between both airplanes.
Interestingly, here’s another vid about a recent minor ramp incident at KAUS.
The user commentary is, as always on Youtube, appallingly ignorant.
But the true fact that the KAUS passenger ramp traffic is uncontrolled by either FAA ATC, the airport authority, or one of the major tenant airlines is another example of AUS being a small / sleepy airport turning into a big / busy one faster than safety features are being added to compensate for the additional traffic.
I was going to call bullshit on the 75 ft separation unless the deck was really low. But a Cat III is an automated landing so they could easily have hit the Southwest plane at 140 mph. That would have been 2 massive balls of fire scattered a couple 1000 feet down the runway.
Not disagreeing with you but that doesn’t sound right for an airport as big as AUS. And that ramp scenario is complete nonsense. The operators of the airport have a legal obligation to ensure it’s safe.
Depending on who made the 75 foot claim and how they figured it out measured between what and where there are a lot of variables. I’d bet the only data available to armchair investigators right now is the FlightAware baro altitude trace from the 767’s ADS-B out. Which itself has a granularity of IIRC 100 feet.
A 737 tail is 41 feet tall. On a 767-sized go-around, you’ll typically descend 40-60 feet from “decision made” until you bottom out & start climbing again. At approach / flare / go-around attitude, the difference between radar altitude measured below the main gear and baro altitude sensed up near the nose is ~30 feet.
A go-around decision made at e.g. 200 feet could easily have put them near enough to 75 clearnace at the moment of closest approach. Not that 200 feet is a magic number, just illustrating how what looks from a distance like a large margin may not have been once all the tolerances are taken into account.
It’s pretty clear they had a pure overfly, with the 767 passing directly over the still-accelerating 737 on the ground. Probably a darn good thing the lower airplane was a ground-loving 737, not an RJ or A319 that can accelerate to its much lower liftoff speed much quicker / earlier and might have rotated and broken ground right up into the 767 just above.
Ramp at my home airport is uncontrolled as well. In our case it’s not busy enough to be a problem.
If a 767 flies over you at 100 feet it will be noticed.
'Zactly. I fly through lots and lots of uncontrolled ramps. And many controlled ones too.
The cutoff for when it’s busy enough, or the taxi paths are constricted enough, or the sightlines are restricted enough that a ramp control function is past nice-to-have and is now need-to-have are necessarily vague. And once somebody decides it’s need-to-have, then they need to persuade all the other stakeholders of that and then they all have to decide how it’s to be paid for. Once funding is aligned, then facilities must be designed & built, procedures worked out, staffers hired, training performed, an opening date selected and the user community informed in advance, etc. All that’s a multi-year process.
Which process may well already be well in train at KAUS. Or may not even be on anyone’s radar. Yet. All I personally can say with certainty is that as of today, the ramps both cargo & pax, are uncontrolled.
As to legal obligation for safety, anyone can be sued for anything. But the idea there’s some bit of federal or Texas law or regulation that requires KAUS to already have ramp control while they do not is ludicrous. So far, IMO/IME nobody has been killed out on the KAUS ramp. There’ve certainly been some needs to suddenly stop taxiing due to conflicts; the vid I cited is just one of many such events. I’ve certainly heard of wingtips getting whacked or nearly so late at night when airplanes are towed hither and thither and nearly every square inch of that ramp is full of empty darkened sleeping jets, with more still arriving.
All of this is a sign of too-small and too-slow-moving government-supplied infrastructure being overloaded by businesses eager to supply the product the customers are desperately clamoring to buy. AUS’ terminal facilities have been repeatedly expanded and no sooner is a new section opened after 5 years’ planning and work than it too is overwhelmed the next morning.
Good reconstruction vids are starting to come out on the KAUS FDX/SWA near miss. It was a lot worse & closer than I thought. This seems like a well-detailed treatment:
The main surprise to me was that SWA got airborne and was climbing rapidly just as FDX was passing overhead in their go-around. Had they collided this would have been a zero survivors event.
I was originally under the impression they’d passed much earlier in the takeoff run while SWA was totally ground-bound. Which collision at least would have afforded some hope of some people getting out alive.
Would a commercial pilot ever say that? I had it drilled into me for my ASEL license that you don’t say as much as “fifty-four hundred feet”, let alone something like “five point four”. It is always and ever “five thousand, four hundred feet” (or “fife thousand”). But maybe commercial pilots aren’t quite that consistent.
There is less perfection of radio procedure than there would ideally be. Saying altitudes and flight levels in non-standard ways is probably the largest area of non-compliance. You hear things like “Passing flight level two two point five for two three zero”. Or as in this flight’s example altitudes said as “[thousands] point [hundreds]”, e.g. “Passing twelve point five for twelve thousand.” The fact many jet altimeters, whether glass cockpit tape instruments or round dials display altitudes digitally with visual contrast between the thousands digits and the hundreds digits doesn’t help.
I personally strive pretty hard not to do that. As a general rule military & ex-military folks are real strict about this stuff and the non-military are less so. Ultimately you end up migrating towards doing what you hear your peers doing. If you fly with folks with loose radio procedure, yours will loosen too unless you’re making a conscious effort to push back.
Maybe I just want the kids off my lawn, but there does seem to be a general lessening of strict procedure that is accelerating in recent years. Although thankfully the drawling good ol’ boy of the 1960s & 1970s is 99% gone.
My micro-issue with the FDX pilot was not that he said “5.4”, but rather that he didn’t say “5.4 miles” or “5.4 thousand feet”, he just said “5.4”.
The AIM tries to prescribe that each type of number be said in a different fashion to reduce the opportunities for confusion. But that’s not 100% effective. But during e.g. approach manuevering we often get clearances like “Descend to two thousand three hundred, turn left heading two two zero, slow to two one zero.” You now need to dial “230”, “220”, and “210” into 3 similar knobs 3-4 inches from each other and get the right number dialed into the right knob. Mistakes happen. Usually caught by the other pilot at the time they’re made. But not always.
It’s busy fast-paced work and it’s done right enough tens or hundreds of thousands of times every day. That the collective effort falls short of perfection is expected. Just not too far short of perfection at any one place and time.
All right, thanks. I didn’t use my license for much beyond the local airport, which had a pretty high proportion of GA students. So it’s likely that most people (including the instructors) were still making a specific effort to follow good radio procedure.
It’s not CB slang land out there; you do hear a lot of careful workmanship on the radio. But you also hear some bad habits, and the occasional cringe-worthy goofs.