Bek Air flight 2100 Crash After Take Off

10 Jan 20 17 Comments

On the 27th of December 2019, Bek Air flight 2100, a domestic passenger flight with 93 passengers and 5 crew on board, crashed after departure from Kazakhstan’s Almaty International airport. The aircraft, a 25-year-old Fokker 100, broke into pieces after crashing into a building, leading to 12 fatalities, including the captain, and 65 people with severe injuries, 22 of whom are still in critical condition.

I know everyone is thinking about the loss of Ukraine International Airlines flight PS 752 in Tehran but right now there’s just not enough information for a reasonable discussion, so I’m going to stick with my original subject, which happened a few weeks earlier.

The captain was the pilot flying and the first officer pilot monitoring. Both flight crew were very experienced and used to cold and icing conditions. The weather and freezing conditions were not out of the ordinary for the airfield, which has comprehensive de-icing and anti-icing procedures.

According to Bek Air, the F100 had been parked at Almaty for two days and there had been no rainfall during this period. They stated that all necessary procedures were carried out. The stabilizers were de-iced but the wings were not as they were clean.

The F100 was configured correctly with zero flaps for departure on runway 05 right. All engines and systems were working normally as the aircraft rotated but at 18 feet above the runway the aircraft started oscillating, rolling right and then left as it lost height and the tail struck the runway. The F100 then climbed to six feet above the runway and then sank again; the tail striking the runway a second time. At the end of the runway, the aircraft again became airborne, reaching eleven feet above the ground and the captain commanded gear up. The speed dropped again and the nose dropped. The plane landed on its belly and glided on the snow travelling 50-55 kph (27-30 knots) through a perimeter fence and then crashed into a two-story house.

The total flight was thirty-six seconds.

There is a number of CCTV videos from the airport. I’ve fast forwarded this one to the point where you can see the Fokker 100 on the runway on the far right of the frame.

This one gives a slightly better view of the oscillations. At the start, the aircraft lights are just visible at the top middle of the screen:

This one shows the whole sequence including breaking through the perimeter fence. The aircraft lights initially show at the top right, above and left of the time stamp:

More CCTV video showing the aircraft and the crash are available on this YouTube Channel listed as Joella Horn.

The airline state that the wind data retrieved from the black boxes shows wind speeds of 96 knots and that they believe that the aircraft was in the vortex of wake turbulence.

An Airbus A321 had departed from the same runway just under two minutes before.

However, AV Herald have posted an argument refuting this conclusion:

Editorial notes: while on the ground the wind measurement by the aircraft systems does not work with any reliability whatsoever, the aircraft was on the ground at that time, the bank angle changed while already skidding on the belly which can explain these values. According to METARs the wind was blowing from 100 to 140 degrees magnetic between 1 and 2 meters/seconds (2-4 knots). The previous departure departed 112 seconds prior to the Fokker 100. Assuming the wind came from 100 degrees during that entire time and thus at an angle of 50 degrees to the runway heading and was blowing at 1 meter/second, any wake turbulence vortices by the previous departure would have been blown left off the runway center line by a minimum of 85 meters (runway width 45 meters) and would have been left off the runway therefore, assuming a wind from 140 degrees at 2 meters/second the wake vortices would have been blown left off the runway center line by 224 meters. Using the wind data off the enhanced FDR diagram while the aircraft was airborne, the aircraft measured the wind from 155 degrees at 6 knots during the first liftoff, from 161 degrees between 5 and 7 knots during the second liftoff and from 132 degrees at 6 knots during the last liftoff. These values would result in the wake vortices blown left off the runway center line by more than 300 meters.

One of the survivors has stated that he saw ice on the wings and that passengers slipped and fell when disembarking via the wing.

On Monday Bek Air released a diagram of additional FDR parameters which can be viewed on AVHerald as a PDF with an explanation on the main page.


  • A high-speed abort is probably one of the most difficult decisions to get right; looking at the dip in engine N1 on the FDR trace, one of the pilots seems to have wanted to abort after rotation failed (and the plane went to 18° of bank!), but presumably the other pilot overrode that — not a good idea.

    With respect to the Teheran crash, the undisputable evidence we have is the location of the wreckage; it looks like the plane turned back towards the airport, and the pilots had their hands full with flying so that transmitting a mayday had to wait. I do believe that one engine failed for some reason (could be mechanical failure, could have been hit by a heat-seeking missile). I am comparing this to Qantas QF-32, an A388 that had an engine failure on Nov 4th 2010, and an engine part punctured the wing and made a large hole in it. If an engine failure were to put a hole like that in a wing tank, the fuel could ignite (“Fuel tank puncture: Qantas A380 ‘lucky’ to escape catastrophe, say reports” was a headline at the time), making for a quite spectacular and deadly fire. We’ll find out more with the passage of time.

  • Mendel has a few valuable comments, we will have to wait and see. Certainly the sequence of events that lead to the Tehran crash is far from certain. Usually, if the wreckage is strewn over a large area this points to a mid air break-up of the aircraft. It is to be hoped that politics will not prevent the FDR and CVR to be handed to experts (I still have an aversion
    to calling them “black boxes”).

    The Fokker 100 crash may well be caused by ice. Water on the wings may freeze and become clear ice, easy to miss during a casual pre-flight “walk around”.
    The wake turbulence caused by a preceding A321 can have influenced the controllability, but unlikely to have been the cause of UNcontrollable behaviour in an aircraft the size of a Fokker F100. Yes, the A321 is a bit bigger but not massively so.
    So far, with the data available at this time, I would put my money on insufficient attention to clear ice on the wings.

    • The politics with respect to the recorders is interesting, to say the least. The Iran won’t trust the US with the data, and the Ukraine won’t trust Russia, so which countries remain that can retrieve data from recorders with heavy fire damage? Australia: “The ATSB’s central office in Canberra includes an audio laboratory and an FDR laboratory. ATSB investigators use these laboratories to perform CVR and FDR readouts for occurrences in Australia and overseas.” The CVR is actually an Australian invention! France’s BEA also does this, e.g. for Ethio­pian Airlines Flight 302: “Ethio­pian investigators delivered the mangled orange flight data recorder to specialists at France’s civil air safety investigation authority, the Bureau d’Enquêtes et d’Analyses, or BEA, which has extensive experience analyzing crashes in Europe and around the world.” (Washington Post)

      Does the Fokker have indicator tufts on the wings? In the SAS 751 accident, passengers reporting these tufts not moving was a clear indication in retrospect that the wings had clear ice on them. Also, since wings don’t ice up symetrically, whoch causes different changes to the wings’ lift left and right, that would be the reason for the airplane to bank once it lost contact with the ground. What would you have done in this situation? Continued to try and lift off since you were past V1, or abort since the plane was unflyable? How would you have decided?

      • I would expect Iran to use the French BEA for the recorders and other required support. The ATSB and AAIB would also be possible but there’s a prior relationship there.

    • ‘Water on the wings may freeze and become clear ice, easy to miss during a casual pre-flight “walk around”.’ Forming really clear ice would take just the right circumstances. (I’ve looked at what it takes to form ice for sculpting.) However, forming ice that isn’t visible to a quick inspection under whatever angle the ramp lights are at could happen very easily. I note that they found ice on the tail but not the wings; I’d think finding ice anywhere would call for an overall deicing, on the grounds that flight can be affected by a small amount of ice — especially if it’s not evenly distributed (as Mendel notes).

  • Not much to say on this one. If there was even a very thin coat of clear ice, that would well explain what we see in the videos. Assuming that the dismissal of wake turbulence is valid, it’s pretty simple. The ice simply wasn’t spotted. It’s a tragedy and perhaps could have been prevented but that remains to be seen as the investigation proceeds.

    What stands out to me is the feeling that the Captain must have felt when the aircraft responded as it did. As Mendel pointed out, an abort at high speed is a tough call. Essentially, it would come down to knowing that a crash was likely and then choosing between an inevitable crash from a runway overrun or the tiny glimmer of hope that you can gain a few knots and snatch victory from the jaws of defeat. It’s easy to understand why many pilots would choose the long shot, even though it’s not a good bet.

    As accidents go, this one did not claim a lot of lives, but it has seriously injured a number of people and to every victim or the surviving relatives of those whom died, this crash is as major as they get.

    32 years ago, I was living in metro Denver when Continental 1713 crashed, due to icing. Denver has some impressive snowstorms with heavy, wet snow, and this was one of those occasions. I was warm and cozy at home when the phone rang and a friend that worked for Continental called to tell me that there had been a crash and that he had been to the sight. I can’t say that I remember the details, but I do remember that he painted a grim picture. A few weeks later, I saw the wreckage as it was stored in an old hangar at Stapleton.

    The sight was sobering. A once sleek aircraft broken up and twisted. When you consider the strength of the skin panels in a transport category aircraft, it becomes apparent how powerful the forces at work are, in a crash such as this.

    Strangely enough, some years later, I had one more encounter with 1713. I was riding my bicycle through a neighborhood and saw a flatbed trailer stacked with pieces of the wreckage. Apparently, the wreckage had been sold and was now scrap metal and would be recycled.

    Somehow I always felt connected to that particular crash and whenever the subject of de icing comes up, I think back to the lives that were ended and the many other lives which were negatively affected by that crash. It’s very easy to underestimate the effects of ice on an aircraft.

    • One thing that is odd in this case was that the aircraft had been at the airport (not flown) for two days and there had been no rain. So yes, clear ice seems likely but how did it form?

  • It seems to me that previous commentators were on the right track. Yes, undercooled water – possible when the air is clear and there is little or no wind – can accumulate on smooth surfaces and freeze nearly instantly when there is something to disturb the equilibrium. The movement of the aircraft as it taxies to the holding point can do the trick.
    Tufts, as mentioned, are not used in modern jet aircraft to my knowledge. I have not flown the F70 / F100 so I do not know what form of ice detection has been installed. The SN601 “Corvette” had a very simple system, if it can even be called that: The inboard front section of the (initially optional) tip tanks was painted matt black. Ice accumulation would be visible as white. Simple and effective. Modern aircraft have more sophsitcated ice warfning systems, but sometimes “KISS” (Keep It Safe and Simple) is just as effective and far more reliable.
    The Tehran crash, as emerged on the news in the past few hours, has indeed been caused by a mistaken identification of the aircraft as a cruise missile. What form of communication existed between ATC and the military, and how this comunication failed, will no doubt be the subject of intense scrutiny. By whom is another question. Will the Iranian authorities allow the involvement of those of other countries? They may well insist that it is now an internal matter between the government and the military.
    The Boeing Max, no doubt, will be the subject of a few more interesting blogs on this website. There are many aspects to consider, not all immediately apparent. But what does become a distinct possibility, if not a probability is that this aircraft may not be fully back in service before the end of this year. The potential cost to Boeing and their clients may run in billions of $$$ and confidence in FAA oversight will have suffered severely. The prestige of this aviation administration as an independent authority has been badly dented as a result. Which may well result in knock-on effects.

  • In response to Sylvia’s wondering how water could have accumulated on the wings without precipitation: There are plausible explanations, the most probable being the result of condensation.
    The videos show a layer of snow on the airport surface; obviously the air temperature was low. The aircraft had been parked for a prolonged period and may well have been undercooled. Any fuel left in the tanks would have been of the same temperature.
    The departure took place in the dark. Even in winter, the sun may have had sufficient power to raise the upper surface of the wing, maybe just by a few degrees but the temperature will have dropped again after sunset.
    Any moisture in the air may then condensate on the wings and form water, liquid but possibly undercooled.
    Another clue may be the time when the aircraft was fuelled. The fuel stored in underground tanks may well have been of a different temperature, maybe a few degrees higher. Again, this may have caused a difference between the temperature of the wings and the ambient air.
    Of course, I have no concrete data to back this up in relation to this particular incident.

    • If the fuel was warmer, loading it would have warmed up the wings and made condensation less likely. (Temperature below ground tends to vary much less than air temperature, so stored fuel would \tend/ to be warmer than winter air.) OTOH, if the fuel was loaded the previous night (cf advice to not let automobile gas tanks sit empty in winter, to reduce in-tank condensation leaving water in the gasoline), it might have cooled enough to prevent the wings from warming in the sun, or even empty wings might not have warmed on a cloudy day; in either case a flow of warmer, wetter air over the cold wing would condense.

      Simpler example: blow on a cold window and watch it fog up; humidity in the breath condenses when it hits the much colder surface. The same can happen on a wider but subtler scale, depending on where the winds are blowing from.

  • Chip,
    It depends on when the fuel was loaded, and how much time before the aircraft departed. There are potentially many variables: Cold air does not hold a lot of moisture, but if the air just on top of the wing became a bit warmer it could have absorbed a bit of water in vapour form. This, in turn, may have condensated again if the air – and the fuel in the tanks – cooled. Another factor is the actual air temperature, and did it maybe rise above freezing during the day and dropped below zero as night fell?
    This is something that requires a lot more knowledge of physics than I learned from the meteorologist who was our lecturer when I did my ATPL, and that was many years ago
    There are so many more possibilities, for instance something loosely implied by Chip: maybe during the period when the aircraft was parked other aircraft started up, and blew warm air over the F100? I seem to recall that something like that caused water on the wings to freeze over on a B737 which plunged into the frozen Potomac during take-off. Have to read up on that again !

  • Chip’s reply could also lead to yet another theory:
    Again, it all depends on air temperature, were there significant variations, how much time had elapsed between the aircraft had been refueled and the departure. Also the source of the fuel could have played a role: Was it from an underground reservoir or a bowser?
    Jet fuel does not freeze like water. It becomes kinda lumpy. But it contains water which of course can freeze. The companies that provide the service are in virtually all cases very careful to ensure that the quantity of water in aviation fuel will be kept to an absloute minimum. The bowsers have clear filter bowls that clearly show water, they are drained and the crew usually will be presented with a sample drawn and tested for water. This is done by putting a strip in the sample, the colour will change if water is present.
    But: suppose that there WAS water contamination? The engines would have been started but perhaps not sufficiently long enough for the fuel heaters to do their work? The engine power would fluctuate, even surge. It could have caused assymetric power (hence the wobble as evident on the videos) and enough alarm to try to abort the take-off at (too) high speed?
    I am now totally “winging” it, but if this were to be the case, of course investigation would not reveal any defect. Maybe the maintenance recorders would show engine irregularities.
    This accident is a bit of a mystery.

  • There was 84% humidity at the time. I’m expecting dewfall at night, I regularly have to clear the rime ice from my car’s windshield after nightfall (or in the morning) even with no rain, since dew precipitates out of the air as it cools. The pilot chose to de-ice the stabilizers only, and took off 22 minutes after the de-icing, which is probably stretching the holdover time. With passengers reporting that they slipped on the wing while evacuating the plane, ice on the wings looks like a strong contender for the cause of the accident.

    The problem of water in jet fuel is mainly that of the water freezing, and the ice clogging up fuel lines, filters, or pumps. The first engine N1 drop was synchronous for both engines, so I’m guessing that was the FO, but the next N1 drop at second 68 happened only on engine 1, and that might have been caused by something like that. (Or the aircraft hitting the fence?)
    Jet engines don’t have a problem with burning water. As a sidenote, hydracarbons (fuel) burn to heat+H2O+CO2, so jet exhaust always contains water (contrails are actual clouds!); I’d expect passing exhaust over a frozen wing might lead to more ice buildup until the wing heats up?

    The Potomac crash was Air Florida Flight 90 in 1982, 38 years and 4 days ago, and it reads like a collection of about every cold weather mistake you can make, short of forgetting to de-ice the wings at all.

  • New official information has come to light:
    — the co-pilot did attempt to abort the take-off, but the captain overruled him
    — Fokker requires pilots to do an ice check touching the wings in several places
    — video footage shows that Bek Air crews don’t usually do walk-arounds or wing checks
    — Bek Air crews don’t have documented winter training
    — Bek Air takes the serial number plates off some of their engines, and the airworthiness of their fleet is poor

    Apparently the Kasakh Aviation Administration audited Bek Air immediately after the crash, and suspended their Air Operator’s Certificate. While this is the right action, it’s tragic that it took a fatal accident for that to happen.

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.