The tower had Fed-Ex climb to 3,000 and turn left to 360 while he had Southwest turn to 170. I would expect the Fed-Ex plane to be instructed to turn right to 360.
And why would the tower assign a Cat III approach to the shorter runway? Is it because 18R doesn’t have the high intensity lighting with centerliine lights for Catt III ?
The FedEx guy was pretty casual on the radio. He called for the Southwest to abort and he announced he was “on the go” for the go-around. Maybe that’s standard slang around there and easily understood, maybe not.
18R doesn’t have a Cat II/III approach. Presumably someone in charge of the finances has decided it’s not worthwhile having more than one Cat III runway. In the other direction 36R is only Cat II and requires special approval.
He turned the Southwest airplane slightly to the right. Which is actually the norm coming off that runway. It avoids housing not too much further south. This was almost certainly not a collision avoidance vector; just the normal thing they say to every departure shortly after liftoff.
And he had FedEx initially turn 90 degrees to the left / east onto a crosswind leg, followed a bit later by a further 90 degree left turn to north for a downwind. From a non-landing on Runway 18L, making a left-hand pattern was exactly what you’d expect. The controller was simply giving the standard missed approach / go-around instructions from that runway off their standard script.
Once again I don’t see any of what the controller was doing as collision avoidance; IMO he hadn’t quite figured out what the heck happened yet. At least not the full gravity of the situation. Had it been VMC and he’d been watching them both out his windows IMO this would have unfolded very differently.
That airport used to be Bergstrom AFB, a Cold War B-52 base. The current 18R/36L is the AFB’s original single runway and lacks a bunch of modern stuff. The terminal and runway 18L/36R were all built after USAF moved out and the City of Austin moved in. Only that new runway was fully equipped for very low vis ops, and even that only when landing to the south.
It’s not so much that 18L is short in any absolute sense; 9000 ft is plenty for routine airline ops. It’s just that the other 12,000 runway is such a behemoth. It’s even 300 feet wide, not the civil standard of 150 ft wide. That is a lot of expensive concrete.
It may be about time for KAUS to upgrade the old AFB runway 18R/36L for Cat II/III ops if the volume of traffic keeps climbing. But if they decide to do that, it’ll be another federally funded infrastructure project that would be years in the doing. They’ll need both centerline lights and a new approach light system. Depending on the vintage of the ILS transmitters and antennas those might need to be replaced too. I don’t know of any obstacle- or noise-related reasons they could not make the upgrade. But those other obstacles may be enough to keep it back-burnered for years, if not forever.
Yeah. Interestingly, I was looking at my company iPad and we don’t have the 36R SA Cat I/II approach plate. We’re not authorized. I didn’t know that approach existed until you said so. For anyone interested, we’re talking about this:
https://aeronav.faa.gov/d-tpp/2301/00556I36RSAC1_2.PDF
Although there are a heck of a lot of other such SA approaches we do around the system and I don’t see anything about this one that’d be problematic vs our procedures, capabilities, and training. Apparently our management doesn’t think that one is worth paying for either.
In general low vis in this part of the country is accompanied either by nil winds or a monster thunderstorm raining insanely right over the field. In the latter case nobody is landing and in the former case landing south is good enough regardless of which way the negligible winds might be drifting.
Hmm, our manuals are unclear. The SA approach is in the Jepps, our OpsSpec approves us for SA approaches in the USA but our company airport pages for Austin only list Cat II and III for runway 17L (has it recently changed to 18/36?). No doubt there’s some paragraph buried somewhere that clarifies what we can/can’t do. Not my fleet, so not my concern.
I’m sure you know this but for others:
Yes, the Austin runways were renumbered within the last 1-2 years. The US mapping agency released some new variation maps a couple years ago and that’s set off a bit of a slow-motion epidemic of runway renumbering.
We actually have an issue on our very oldest 737s that the magnetic variation map built into the avionics is now too out of date / incorrect for certain high precision operations. And for some reason related to obsolescent electronics, they can’t / won’t be updated. Whether that’s impossible, or just not cost effective on jets they’d planned hoped to park soon I can’t say.
In other news on the AUS event, it appears now that the various citations about “75 feet” were referring to the minimum above-ground altitude the 767 descended to on their approach before climbing again. Which altitude they reached while they were less than 1/4 mile behind the 737 accelerating on the runway directly in front of them. So that “75 feet” was not the distance of closest approach between the two airplanes. Though they still got way too close together for anyone’s comfort.
It appears that FedEx saw the 737 looming out of the murk as they broke out of the hard clouds and began to see the runway ahead. That was quite the nasty surprise I’m sure. So they start their go-around and meanwhile the 737 starts to rotate, all as the 767 is rapidly closing the horizontal distance between them. Although as the 737 continues to accelerate, the horizontal closure rate reduces. FedEx is at about 250 feet & climbing when the 737 starts to break ground just an airplane length-or two ahead of them.
By then the FedEx pilots have almost certainly lost sight of the 737 under the nose and also in the murk they’re climbing back into. Meanwhile, Southwest may not really understand there’s a jet close above and slightly behind them. Nor do they have too many good options to do something other than continue taking off in the standard fashion into the same murk, even assuming they have Yeagertastic awareness of what’s going wrong just above/behind them.
Shortly after Southwest breaks ground the two jets end up almost perfectly stacked on top of one another at the same speed, with 300-ish feet of vertical spacing that’s slowly increasing as FedEx is, at least initially, climbing faster than Southwest. Meanwhile Southwest continues to accelerate and FedEx begins to slide aft relative to Southwest, although they’re still stacked with little (no?) fore/aft spacing. By the time Southwest finishes their liftoff rotation and is established in their normal first thousand feet of rather steep climb, FedEx is now ~700 feet above them. Vertically speaking, they’re pacing each other in the climb, neither diverging nor converging. Meanwhile the horizontal situation is starting to improve as Southwest pulls ahead and begins turning right. It was fortunate the standard maneuver after liftoff at KAUS is to turn away from the extended centerline of the runway, and Southwest had been given that turn as part of their takeoff clearance, so was ready to accomplish it immediately once clear enough from the ground.
Even had the weather been perfectly clear, I conclude Southwest never had an opportunity to see FedEx after they entered the runway while FedEx was on a too-short final. FedEx was always in a place where Southwest could not possibly see them even if they’d known to look and the fog was absent.
From FedEx’s POV …
It’s not uncommon in low-vis landings like this to get pretty target-fixated on the place you’re first hoping to see the relevant target lights, then once they appear aiming at them to land. Gazing deeply into the murk farther ahead looking for obstacles is not really part of the standard play. It’ll be very interesting to learn how and when FedEx first sighted Southwest, how obvious it was, and how they decided on their course of action.
Proceeding counter-factually for a bit …
Had the visibility near the runway threshold been slightly worse, FedEx would have seen Southwest slightly later. And if they’d gone around from that later sighting the vertical situation would have been even more critical, perhaps leading to impact.
Had the visibility been even worse than that, FedEx may not have seen Southwest at all until they were committed to stay on the ground and they’re closing the distance to Southwest ahead. The good news is that by then FedEx would be starting to slow and it’s at least plausible that although they’d both be on the same runway at the same time, slowing FedEx would never catch up to accelerating Southwest. Heck, FedEx might never see Southwest ahead.
Just for fun (while acknowledging it would very much not be fun for anyone involved), another counterfactual - what if the 737 had taken slightly longer to start its takeoff roll - is it possible the 767 could have landed ahead of them (assuming they didn’t see them)? Or would the typical height above the threshold for the 767 have them hitting the tail of the 737 in that scenario? And for further ‘fun’, let’s (nonsensically) assume the 737 crew somehow miss the 767 passing inches above their heads and take off anyway - given enough time between the landing and the takeoff, could they reasonably clear a 767 that was still slowing down at the other end of the runway (let’s assume the 767 is exiting the runway at the last taxiway rather than going for a harder stop). My assumption is they probably could in theory, given the KLM 747 almost cleared the PanAm 747 in Tenerife despite not being up to normal rotation speed.
I’d bet both those hypotheticals end badly.
For the first case …
If the 737 was sitting stationary at the beginning of the runway when the 767 passed overhead they’d almost certainly hit the 737’s tail. And in any visibility good enough to land in they’d have seen the 737 at the same time, or mere instants before, they saw the landing area. Net of their different recognition time for expected and unexpected stimuli.
For the second case …
I don’t see how the landing 767 could slow down gradually enough to stay out in front of the rapidly accelerating 737. Obviously if the 737 waits long enough after the 767 successfully passes overhead, the 767 will have landed, slowed, and taxied clear of the runway before the 737 starts moving. No collision there. But if you assume the 737 starts moving promptly after the 767 goes past, I predict they come together a couple thousand feet down the runway.
As with all these things, there’s a wide window of timing where everything is normal everyday happy rainbows and unicorns. And there’s a narrow window of timing where everything is fire and twisted wreckage and dead people. Which narrow window of disaster is surrounded both slightly before and slightly after by timing that leads to close calls, dirty underwear, and investigations.
And now for something completely different. …
The first flight of a subscale Blended Wing Body (BWB) demonstrator in China:
See here for background on other efforts in that direction:
There’s lots of expectation that along about 2050 or so this will become the standard shape of cargo haulers and at least some passenger airliners too.
In fact I posted here about one such cargo effort just last week:
Thanks for playing. For the second case, I guess I was imagining an intermediate scenario in which the 737 spools up once the landing 767 is about halfway down the runway already, such that by the time the 737 takes off, the 767 is already near the end of the runway. But that’s just a special case of what you already covered, and the question boils down to whether the 737 has climbed to a greater height than the tail of a 767 by the time it reaches the end of the runway. Which is going to depend on a whole bunch of other factors such as surrounding terrain and exactly what takeoff performance has been used. Let’s hope it’s never actually tested.
In terms of vertical clearance, it’s certainly reasonable to assume the 737 is high enough by the end of any runway to physically clear a 767 tail. But not by much for all the reasons you say. It’d certainly be more exciting than you’d like to try for real. And of course the sooner the 737 starts after the 767 passes, the less far down the runway the 767 is by the time they pass or collide.
Later articles have said 123 pax + 5 crew on Southwest = 128 total. No articles are mentioning any extra crew on FedEx, so the 2 working crewmembers only is a good bet. So 130 souls between both aircraft.
2 things. I thought the 737 was given a 170 heading which was 10 degrees to the left.
And where is TCAS in all of this?
I got caught by surprise by that once. I use to print out airport diagrams for the airports I flew to and I’d write on them as I get instructions. When I heard the tower give directions for runways not on the sheet it was a “wait what?” moment. It already had runways with numbers close to each other so it was confusing until I remembered 2 of them were parallel. that sorted it out quickly. It was going to be Left, Right, or Other.
You’re right about the heading. The actual runway azimuth on the current chart is 175. My dyslexia had 170 being a right turn; in fact it’s slightly left. Nevertheless, Southwest did drift right of runway centerline after liftoff. Whether that was drift or deliberate collision avoidance or whatever is unknown just yet.
As to TCAS …
There are no warnings below about 1000 feet AGL. But each airplane’s TCAS display should have shown the other airplane as a generic traffic target with their rough position and their relative altitude. If the pilots were looking at it that would have been a scary thing to watch.
Southwest missed a chance to avoid the problem by not noticing FedEx’s proximity before they entered the runway. And FedEx missed a chance to notice how close Southwest was and to go around a bunch earlier and thereby keep the vertical spacing a lot larger between them. Both crews had plenty of other stuff to do at those times and the low vis only added to both their workloads. So not picking up on the advisory data that TCAS was putting out is probably par for the course, or maybe a bit over par.
Once they each independently realized a screw-up was in progress they might then have tried to use TCAS for deconfliction. That may in fact have been what caused Southwest to drive off to the right; perhaps they self-vectored away from FedEx.
Although all our training about TCAS is that the azimuth and range are highly unreliable and you should not maneuver in response to those indications. This is reinforced in use where we often see passing airplanes at a materially different azimuth than the display shows. Conversely one should maneuver in response to the vertical indications, which are precise and reliable. Which is nice and all, but when you as FedEx are already climbing as aggressively as you can there’s not much more you can do to help improve the delta-altitude data you see. And for you as Southwest, you’re sandwiched between the ground, the unknown obstacles in suburbia just ahead, and FedEx looming above. Trying to split that difference is too hard for Yeager.
My IMO bottom line: Better TCAS awareness could probably have prevented the problem but was probably not much use in resolving it. In all, were that me out there, I’d rather have TCAS than not while fighting this fire. But I may well not have gotten much value from it. Shit happens real fast when we’re off script and ad libbing. More cockpit gizmos are usually unhelpful in that spot.
Since the Fed Ex pilot basically took over the tower’s job he could have called his altittude ever 5 seconds so the Southwest pilot knew where NOT to be.
I’m a little confused at this. I’m certainly aware that runways are named for their compass heading. But I didn’t know that the FAA forces a rename if it drifts slightly.
SJC (San Jose) used to have three parallel runways: 30L, 30R, and 29. 29 has since been closed (it was basically the GA runway, and they’ve been trying to end all GA activity), but why was a 29 allowed to coexist with a 30 if they’re parallel and the FAA is forcing renames?
29 is not parallel to 30. it’s 10 degrees off. Pretty close when you’re 10 miles out.