Crash on Final Approach: Pakistan International Airlines flight 8303

22 May 20 19 Comments

Today (22 May 20202) an Airbus A320-200 operating as domestic flight PK 8303 from Lahore to Karachi crashed on final approach at Karachi, Pakistan. Current reports are that there were between 90 and 99 passengers on board and eight crew. It is not yet known if there are any survivors.

The flight departed Lahore at 13:05 local time for a flight just under two hours. After grounding flights during the pandemic, Pakistan International Airlines had resumed offering limited flights on the 16th of May.

There was no significant weather:
METAR: OPKC 221100Z 23014KT 7000 NSC 35/24 Q1004 NOSIG
TAF: OPKC 220330Z 2206/2312 24010G22KT 6000 NSC BECMG 2219/2221 26010KT 5000 FU SCT025 FM 230400 23010G20KT 6000 NSC

The aircraft crashed into a residential area in Model Colony (a Karachi neighbourhood two miles north-east of the airport) while on final approach to runway 25L. Four or five multi-storey buildings have been reported as damaged and on fire.

Jinnah Garden, Model Colony

If you click on the image above, you’ll see the same image in Google maps; zoom out slightly and you’ll find the airport is easily recognisable.

According to the Aviation Herald (who is generally reliable about sourcing information rather than publishing rumours), the inbound flight had aborted their initial approach on the left hand runway (25L) at Karachi, reporting issues with the extension of the nose landing gear. As they came around parallel to the threshold of runway 25L (left downwind) for the second approach, the flight crew declared an emergency and requested an immediate left turn as they had lost both engines. Tower cleared the aircraft to land on either runway (25 left or right).

The following video shows wreckage from where the aircraft crashed into a residential area and includes disturbing images:

The audio clip posted by has 30 minutes of audio for full context. These radio calls can be heard starting at 09:05.

The following is my transcription but it is hard to make out; corrections welcome.

PK 8303: We are proceeding direct, sir, we have lost engine(s?).
ATC: Confirm you are carrying out belly landing?
PK 8303: [Unintelligible]
ATC: Runway available to land on two five.
PK 8303: Roger
PK 8303: Sir, MAYDAY MAYDAY MAYDAY, Pakistan 8303
ATC: Pakistan 8303, roger Sir. Both runways are available to land.

There were no further radio calls. The Airbus A320 crashed moments later.

This video by Aviation News 24×7 shows the final descent of the aircraft along with those same last interactions with ATC:

One eye-witness claims that the aircraft “suddenly went silent” shortly before impact.

These photographs are said to have been taken after the go-around but before the crash. They show black streaks along the bottom of both engine cowlings. The landing gears are not extended.

One reply to the above tweet points out that RAT has come out, a turbine that automatically deploys in case of power failure in order to generate power for basic avionics.

There are some reports that the aircraft did not manage to abort the landing in time and touched down, scraping the engines, before going around. However, this may simply be people making assumptions based on the report of the landing gear problem and the black streaks showing on the photograph. And more importantly, I could not find any verification that the images in question were actually taken on that same day, let alone after the go-around at Karachi.

EDIT: it’s been pointed out to me that this graph by Flightradar24 strongly implies that the aircraft touched down on that first go-around:

Flightradar24 data; click on the image for more info.

Further images from the crash site are appearing on social media now that the fires are out and as the wreckage is cleared.

It’s hard to imagine a sequence of events that starts with a gear extension issue and ends in a dual flame-out.

There is only one bird in the photograph and the crash took place just five minutes after the go-around, so it is hard to believe that either fuel exhaustion or bird strikes (possible grounds for two engines to fail at the same time) are the cause. Fuel or oil contamination could also have that effect. Another possible reason is a single engine failure and the wrong engine shut down. Right now there isn’t enough information even to make an educated guess.

I’m leaving this open for discussion and am happy for updates in the comments as more information is made available. However, I will moderate any commentary referring to “third world aviation” and/or unsupported assumptions about the pilots’ religion (as yet unknown) as a contributing factor. Let’s try to focus on the specifics and not use stereotypes to support early accusations of pilot error.

Updated article (posted after the preliminary report) is here:


  • As you say, it feels like we can’t do more than make semi-educated guesses at this stage. I assume the flight recorders should hopefully shed some light on things? Having said that, I really hope it isn’t low fuel loading due to the global economic conditions.

  • Good write up so early after the event. The CCTV you posted today in the comments seems to show the aircraft with wheels down. RIP all.

  • At first blush, I’d say they grounded the engines enough to inhale FOD on that first pass, then they extended the mains to avoid that happening again, but that’s just a SWAG, because that’s all we’ve got to go on at this point.

    The investigating board has its work cut out for them on this one. They’re going to have to correlate altitude with engine health and the voice recorder and see what that says.

  • Sylvia was very quick with this tragic accident. My sympathies are with the families of the victims.
    From the little bit of information that is available at the early stage it is impossible to make much sense of the cause. Not everything would play on my old laptop and the audio is very noisy. Sylvia did a good job attempting a bit of a transcript.
    From what I can gather: the most likely cause if a serious mechanical or electronic malfunction. The crew, from what I could hear, sounded composed and professional.
    So I do not want to speculate too much at this point. Did the gear fail to fully extend? The level of automation of the A320 makes it virtually impossible to land without the gear at least selected down. Assuming of course that the engines are running. Did the landing gear fold on touch-down, causing the crew to abandon the landing?
    What is the importance of the “scrapes” on the runway? Did this cause damage to both eninges to the extent that they both failed? Fuel starvation is not high on the list of possibilities, nor does it look like another case of “pilot error”.
    As time goes by more details will become available.

  • I’m unconvinced that the FlightRadar graph shows the plane actually touching the runway. The resolution (roughly 1cm per 5000 feet) is too low to be really precise, and the line clearly stops short of 0 AMSL; given that Karachi Airport is only 100 feet AMSL (per Wikipedia) a scrape-and-go can’t be assumed. (“Scrapes” on the runway are standard artifacts of making tires go from 0 to 100-200mph in no time; I vaguely recall reading that ?DFW? has to clear rubber remains off the ends of their runways every day.)Taking FOD (as Gene suggests) seems possibly, but barely; how much force would be needed to lift something off the ground and into an engine on a plane that is in landing (nose-high) attitude, and what are the chances of that happening on both engines? Wikipedia lists this as Karachi’s first notable accident involving a major modern airplane built in the West; there was a hijacking, and an Ilyushin crash, and some planes from ~60 years ago, but no indication that the airport is careless about keeping the runways clear — cf Air France Flight 4590, although in that case the debris blew a tire rather than being lifted up into an engine, AND the debris was big enough to find, AND it was dropped by an airplane that had left only 5 minutes before. (Wikipedia also reports that the dropped part was neither made nor installed to standard — and the installation was done in Houston, not abroad.) It will be interesting to hear about this in a year or so, when the investigators have had time to look for clues and dismiss rumors.

  • The Plane Spotters pix appear to show a completely clean wing. This seems unlikely at an early stage of an emergency go around. It would also be unusual for the crew to continue an approach after a landing gear malfunction, when normal emergency actions would require terminating the approach from the moment the malfunction was seen.

    The cowling marks are also puzzling – any impact severe enough to cause these would probably lead to rapid engine failure.

    I agree that there is much more to be learnt about this tragedy.

  • Yes I agree. I was reacting to incomplete information, not improved by an impending computer HD breakdown.
    What may be the case: the aircraft had been grounded due to the covid-19 outbreak. Pakistan has a hot and humid climate. What is not out of the question is that there may have been corrosion in electrical circuitry and possibly water in the fuel. The latter is not all that probable, but I have had personal experience with an airplane that had been sitting outside over winter. The engine stopped immediately after take-off even though we had assumed that we had drained the tanks properly.
    The difference was: I was in a Piper Super Cub, the runway was about 1500 metres long. I still landed and stopped on the runway.

    • My guess, based on what I’ve read generally, is that furloughed aircraft have not yet been put back in service; there have been reports in the last few days that reservations for travel months away have been picking up, but I haven’t seen reports that airlines are restoring scheduled flights that were dropped due to lack of demand.
      I also note that your incident happened on takeoff, almost as soon as the plane changed attitude; here the plane had gone through an entire flight. I could see corrosion (in the unlikely case that the aircraft had been parked for some time) not being damaging until after climbing to altitude and coming down, but water in the tanks would have been at the low point throughout the flight and so should have manifested early if it existed.

      • This aircraft was definitely furloughed; it had only made 8 flights since March.
        “On May 23rd 2020 the airline’s engineering department reported the aircraft had last been checked on Mar 21st 2020 and had flown 8 sectors since, the last sector before the accident flight was on May 21st 2020.”

  • “On May 23rd 2020 Karachi Airport reported based on CAA inspection report that the runway inspection revealed scrape marks of the left engine start 4500 feet down the runway, the right engine scrape marks begin 5500 feet down the runway. About 6000-7000 feet past the runway threshold the scrape marks end.”

    Thankfully, it’s quite rare nowadays for aircraft to crash in dense housing. The incident in Iraq comes to mind, but that aircraft disintegrated in the air, so it didn’t do that much damage on the ground.
    And sometimes the Hudson river is there to save you…

  • My HD failed. I will try to keep up.
    My son may be able to fix it. I am not good at the email by mobile phone.

  • The Flight Data Recorder and the Cockpit Voice Recorder have been found, and the BEA will read them out. BEA is the “French NTSB”, and they’re involved because the aircraft was made (assembled) in France.
    So there’s likely going to be a good report on exactly what happened that led to the crash.

  • Pakistan have promised a preliminary report to be released on the 22nd of June, so I’ll come back to this in a few weeks.

  • I just watched the analysis of the approach and descent on the blancolirio youtube channel at , and it reminded me of the M2-F2 accident: fell like a stone, and touched down with the gear not extended. Obviously there is no excuse for piloting an Airbus like a wingless body!

  • The preliminary report is out at , and it sheds more light on the accident:

    (e) “Karachi Approach” inquired “confirm track mile comfortable for descend” and later advised to take an orbit, so that the aircraft can be adjusted on the required descend profile. No orbit was executed and the effort to intercept the glide slope and localizer (of ILS) was continued. The FDR indicated action of lowering of the landing gears at 7221 ft at around 10.5 Nautical Miles from Runway 25L.

    (f) “Karachi Approach” advised repeatedly (twice to discontinue the approach and once cautioned) about excessive height. Landing approach was not discontinued. However, FDR shows action of raising of the landing gears at 1740 ft followed by retraction of the speed brakes (at a distance slightly less than 05 nautical miles from the runway 25L). At this time, the aircraft had intercepted the localizer as well as the glide slope. Flaps 1 were selected at 243 knots IAS, the landing gears and speed brakes were retracted. Over-speed and EGPWS warnings were then triggered.

    They were coming in high and fast on their initial approach, they could have simply flown an orbit (a circle) to slow down and descend, but they didn’t. They extended the speed brakes and the landing gear to slow the aircraft down, and when they had captured the glide slope, they not only retracted the speed brakes, but the landing gear as well, leading to the engines scraping the runway.

    There were some issues with ATC, radio communication was with Approach only, who cleared the landing with “Aerodrome control” via telephone.

    I didn’t see any information about maintenance, although the airport had been Covid-grounded until May 5th.

    It stands to reason that flying a proper stabilized approach in the first place would have prevented this accident.

Post a comment:

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.