The crash of PIA Flight 8303 - very strange

Miscellaneous and Personal Stuff I Must Share
1

On May 22 Pakistan International Airlines Flight 8303 (an Airbus A320) from Lahore to Karachi crashed in a crowded urban area just short of runway 25L. 97 of the 99 people on board were killed. (It’s remarkable that 2 survived, and that apparently no one on the ground was injured.)

There’s lots of commentary online, and lots of YouTube clips discussing this. All information is obviously preliminary, and analysis contains much speculation. But some facts seem to have emerged that point toward some very strange actions by the pilots.

The facts:
Weather was fine - clearly not a factor.

During the initial approach, the plane was much higher than normal - ATC at the airport several times pointed this out to the pilots, who replied they were “comfortable”.

The landing gear was not extended during the initial landing attempt.

The pilots did not notify ATC that they were attempting to land gear-up.

Without landing gear, the attempt at landing resulted in scraping the plane’s engines on the runway. Marks on the runway consistent with this were found.

The pilots then elected to add power and go around for another landing attempt.

During the go-around, the plane was observed (and videoed) with landing gear down.

The plane was unable to climb to the assigned altitude (3000’).

It then lost power in both engines, making it unable to reach the runway.
The speculation:
The fact that the gear was not down during the first landing attempt may have to do with the fact that the plane was using much higher than normal airspeeds (to get rid of the excess altitude) and the A320 will refuse to extend landing gear at speeds that make this unsafe.

Despite various alarms and indications, the pilots were probably not aware their gear was not down. If they had been, before landing they’d surely have spent some time trying to get it down, and failing that would have notified ATC.

The fact that the gear was extended during the second attempt argues there was nothing wrong with the landing gear, and points toward the excess speed issue.

The loss of engine power was likely due to damage the engines suffered while scraping along the runway during the first landing attempt. If oil lines were damaged, the engines would have delivered power for a short time, then failed - as was apparently the case here.

2

I thought that there were interlocked indicators for landing gear - no way could they have missed that.

Also when attempting to land there were not have been any automatic flare since the aircraft was not configured for landing - so the attitude would have been wrong, aircraft would not have slowed as expected.

3

I wonder why they tried to descend as fast as they did? Pressurization issue? Fire? Other in-flight emergency?

Task saturation from that, or smoke from a cabin fire, could explain them forgetting to lower the gear, and otherwise configure the aircraft for landing, as well as ignoring a ground proximity warning system that was likely screaming at them.

I also thought the rule of thumb for an inadvertent wheels-up landing was to continue with it, and not try a go-around? There is video of some GA aircraft managing to accelerate and rotate to takeoff with the gear up, but that’s likely the exception.

Tragic. Hopefully the ATP community can learn from this.

4

Is it possible that there were other alarms going off, too; enough that the gear-up warning got lost in the noise? Or could there have been other alarms earlier in the flight that caused the pilots to disable or mute the audible warnings?

I have to say, scraping the engines on the runway and getting back in the air sounds like a pretty impressive feat of flying and engineering. It sounds like the wrong choice, but I wouldn’t have even thought it was possible.

5

We’ll know when the data recorders are recovered, I guess.

Another guess, until someone like Richard Pearse or LSLGuy shows up, is that the pilots hit the ‘Go Around’ setting on the auto throttles, and the airplane took over from there. There’s a picture of the doomed plane at airliners.net, post touch and go, showing a deployed ram air turbine and two heavily oil stained engines, so the pilots would have had to have done more than just told the autopilot to go around, but that’s my guess for how they got it back off the runway.

6

The simple answer would be that they needed to descend quickly in order to make their planned landing. They made no call about any problem on board.

The transmission that includes “We are comfortable” strongly suggests that the pilots were not aware of any problem.

7

This looks to be incorrect - recent reports say that 11 people on the ground were injured.

8

My cousins ex-husband was on the flight and tragically did not survive. Met him several times and he was an honest and honorable man. Taken too soon along with many.

9

It really is a mystery. It looks like extreme incompetence on part of the pilots, but the captain had 24 years flight experience and no prior events and the sequence is mystifying.
One claim is that the full PPE that all pilots and crew have to wear now hindered their work. Maybe they did not hear the alarm under those layers.
PIA service has gone to the dogs last few years but their training and maintenance has always been of a high caliber and they actually train pilots/maintain aircraft for other airlines as well.

10

It’s maddening, but this kind of thing does happen. In the Trans Asia Airways crash in Taiwan a few years back, one of the plane’s engines failed on climbout - after which the very experienced cockpit crew cuts power on their one remaining engine, and then flies the plane into a stall. :smack:

11

The initial report is still up and is a very interesting read:
Initial Report PIA PK8303

At 500 ft, the FDR indicates: landing gear retracted, slat/flap configuration 3, airspeed 220 knots IAS, descent rate 2000 ft/min. According to the FDR and CVR recordings several warnings and alerts such as over-speed, landing gear not down and ground proximity alerts were disregarded. The landing was undertaken with landing gears retracted. The aircraft touched the runway surface on its engines. Flight crew applied reverse engine power and initiated a braking action. Both engines scrubbed the runway at various locations causing damage to both of them. Figures hereunder show selected screenshots of security / CCTV cameras footages of the aircraft engines touching the runway and showing sparks due to scrubbing, along with marks on the runway.

So, the pilots slammed the aircraft into the ground, deployed the thrust reversers as they skidded along, and still decided their best option was to go around. There were some arguments initially that the flight crew had recognized they were in trouble and tried to go around before hitting, but I think the accident report makes clear they flew the aircraft all the way into landing without a clue. The go around was not initiated until after they started the braking sequence on the ground.

I have to admit to being highly impressed with the engine power available to shove the airplane back into the air under those conditions.

However, it seems quite probable that if the crew had left the plane on they ground most, if not all, of the occupants would likely have survived.

I also like that the investigation notes the tower “observed the scrubbing of engines with the runway but did not convey this abnormality to the aircraft.” What can you say when a couple of guys ignored every alert, checklist, and what have you in the book to crash an airplane?

I can’t wait for the release of the CVR. The conversation, or lack thereof, between the flight crew will be interesting. Except for the guys driving, it seems like everybody involved (including the airplane) realized the approach was out of control. What was going on in their minds?

12

[Part 1 of 3 due to post length limits.]

[Yes, I’m back, at least intermittently / temporarily. No assurances though; much as I love and miss y’all, my IRL remains kinda busy.]

Wow. Just Wow.

For an insight into “What was going on in their minds?”, which is a very insightful question, read this article: https://www.nytimes.com/2019/09/18/magazine/boeing-737-max-crashes.html. (It may be behind the NYT’s login-required-wall, but a site registration is free, no obligation, & doesn’t seem to trigger spam.)

That article is explicitly about the 737 MAX accidents, but ultimately it’s about craptacular-by-design piloting at craptacular-by-design airlines. And what that portends for the future of international aviation regardless of the brand of airplane that hits the ground all too often.

Our Richard Pearse flies the A320. I don’t and never have although I have copies of my employer’s A320 manuals which I reviewed before posting here. For this post I have not read the entire report txtumbleweed has cited, merely the contents of this thread. So all my thoughts are highly conjectural.

Those manuals tell me that, barring malfunctions, as they were approaching under the parameters txtumbleweed has cited, the gear would have been down or coming down had they lowered the gear control; they were well slower than the speed limit the automation would have enforced against lowering the gear.

As well an audio alarm & warning messages about “no landing gear” would have been happening for some 15 seconds before that cited moment, and the alarm would have continued for the next 45-ish seconds through that initial touchdown / impact and the start of the go-around. So the circumstantial evidence is they simply never thought to lower the gear.

What were they thinking? IMO they weren’t. They were simply rotely pushing buttons and expecting HAL to make it all work out. HAL’s good, but even Airbus hasn’t repealed the laws of physics.

Descending from altitude to touchdown is a hundred-mile long process. If for whatever reason you start down late and therefore are too high for your distance to the runway (or equivalently, too close for your altitude) there is some ability to catch up some percentage of that along the way. Every jet pilot has done that a little bit at one time or another due to ATC problems, weather, radio issues, traffic, or inattention to some navigation programming detail. Or simply due to flubbing some mental arithmetic. In some areas of the world due to traffic, it’s a known trap that happens to most of the airplanes traversing that area. It’s just one more non-optimality of the larger environment that the pilots have to manage.

But if you mess up bad enough that there isn’t enough remaining forward distance to correct the amount of vertical path error you have, there is no alternative except to alter your lateral route to extend the distance to the runway. Which in turn involves breaking out of your “everything’s routine” mindset (=”mental rut”) and 'fessing up to ATC that you’ve screwed up and need special handholding to get down properly.

13

[Part 2 of 3 due to post length limits.]

Unfortunately, things happen fast at 400+ knots. If in addition to the airplane getting behind its descent profile the pilot is getting behind the airplane’s state, he/she can be sitting there just unaware of the growing disconnects between normal ops, the actual aircraft state, and his/her complacent / erroneous mental model of the aircraft’s state. And once reality intrudes so hard that the mental model can no longer tenably be imagined as accurate, chaos (physical and mental) ensues as the pilot tries to wrench the model to conform to something utterly unexpected happening in just seconds.

Ultimately, good pilots fly their mental model. And then operate the aircraft to conform its behavior to the model. While simultaneously continuously cross checking all aspects of the model against true reality.

Done well this is very powerful since the model can be well ahead of what’s going on in real time. And the two pilots’ models can be actively talked about and compared to the reality shown on the instrument panels & out the window. Appropriate to the MAX accidents this also allows any discrepancies in the instruments themselves to be detected and worked through.

Crappy pilots have no mental model at all. They’re just pushing buttons in a rote sequence and assuming the outcome follows established routine like it did yesterday.

Notso hotso pilots are somewhere in the middle. As long as things stay simple enough and the airplane is automated enough, and ops don’t get too far off the normal plan, having a weak or absent model doesn’t spell immediate doom. But it is a very brittle state; once something goes wrong enough what happens next is down to luck.

Even a good pilot can become temporarily crappy if he/she lets their attention wander at an inopportune time. Again things happen fast and at critical moments the horse can be a long ways out of the barn before you “wake up” from a 20 second daydream or completely legit distraction. The skill comes in being able to recognize and recover from these momentary losses of model coherence and accuracy. And being professional enough and disciplined enough to always consciously take active measures to prevent “daydreams” at inopportune moments and to manage and prioritize the inevitable distractions.

If the airplane is on profile and the pilot is a little behind, or the airplane is a little behind profile but the pilot is on-point, the situation is readily resolvable and in fact happens briefly on substantially every flight by every carrier every day. Nothing is ever done perfectly from gate to gate.

14

[Part 3 of 3 due to post length limits.]

But once both the airplane is behind the profile and the pilot(s) are behind the airplane, that’s often too behinder to catch up. Good pilots notice the snowballing fucked-uppedness (WARNING: technical aviation term) well before ground impact and buy time to regroup somehow by going around or entering holding or whatever.

Really bad pilots simply become bore-sighted on pointing the airplane at the end of the runway & being somewhere between oblivious and bewildered as the snowball gets worse. I read these PIA guys as oblivious, not bewildered.

That bad state will persist until the accident sequence starts at which point they experience a mental reboot and suddenly do something else. Which something else (like going around in this case) is itself an unthinking almost autonomous knee-jerk that often as not is one of the wronger choices on offer.

I agree these guys would almost certainly have killed fewer people had they not gone around. On the other hand, had the engines not been damaged as much and had stayed running normally a few more minutes they might, might have been able to finish this flight with no bloodshed and just a bunch of scraped metal.

Some folks have 24 years of experience. Other folks have 6 months of experience repeated 48 times. Aviator, airman, and airplane operator are three very different ways of approaching the training and the day-to-day execution of the work. The latter type is simply an accident waiting for the first opportunity to arise. They are increasingly common.

That is not good news.

15

Huh. Thought this bump would be about 40% of Pakistani pilots having fake licenses

https://airwaysmag.com/industry/pakistan-fake-pilot-licenses/

16

I saw that. The preliminary report only notes that the pilots’ records are in custody and being reviewed. I wonder what they have found.

A scary thought, how long have the aviation authorities over there known this was going on? Those numbers in Airways Magazine as well as these in this Associated Press report seem awfully specific for an investigation to have only been started after this accident.

17

I just read the pilot’s license is in question. Which forced a credential review across the airline. 150 of 435 pilots, NOT qualified! Faked licenses, expired, and cheated on the tests. This kind of corruption likely extends into the control tower as well. Turns out, the whole airline safety over site thing is just an old boys club for retired, but wholly unqualified generals!

I’m thinking this could depress ridership on PIA.

18

Why am I not surprised at the level of corruption in Pakistan? It seems to me to be a thing at all levels - in the UK we have to take a classroom based book test as part of our driving courses, we have had serious issues where people were taking the exam on behalf of others for a greasing of the palm - after a number of incidents this came to light so now proof of identity prior to taking any part of the driving exams must be provided, and this must include photographic i.d such as passport or the fairly recent photo i.d driving licence. When I used to teach food hygiene or health & safety it was common practice for candidates to pass the exam and sell their certificates to food safety outlets, which has lately been noted by safety inspectors, but this is hugely prevalent among Pakistani origin people, I have never seen it with other cultures - I’m sure there must be some malpractice elsewhere but it does appear to be a culturally and socially acceptable practice from that community - maybe its just my own perception bias.

19

My heart almost stopped when I saw the title, because I thought PIA was Pacific Island Aviation, which flew between islands in the Marianas (Guam, Rota, Tinian, and Saipan) when we lived there many years ago. Company has been out of business almost as long as we’ve been away from there, but I haven’t kept up.

20

Thank you for the very informative series of posts. It’s good to see you here again, LSLGuy.

I can’t add anything to the thread except another voice that’s bewildered at how things could have gotten so out of control, both with the flight and with the culture of aviation professionalism at PIA.

A third of the pilots having faked credentials is boggling. I hope the fraud is only for comparatively ‘minor’ paperwork deficiencies like the medical certificate is six months out of date and got backdated to be in compliance (but was still completed satisfactorily). Were they completely gundecking tests like, had the pilot completed a check ride or not?