Arrow Air flight 1285: icing or explosion?

24 Jul 20 9 Comments

On the 12th of December 1985, a DC-8 Jetliner operating as Arrow Air flight 1285 crashed on departure from Gander, Canada. This was an international charter flight carrying US troops from Cairo, Egypt to their home base of Fort Campbell, Kentucky, with stops to refuel at Cologne in West Germany and at Gander.

Recreation of the Arrow Air DC-8 by Anynobody on Wikimedia commons

The flight departed Cairo on Wednesday at 20:35 and arrived in Cologne five hours later. They departed Cologne on schedule and the six-hour flight to Gander was without incident. They refuelled at Gander and left just over an hour later, at 10:15 UTC but early morning local time; it was still dark.

There were three flight crew (Captain, FO, flight engineer), five cabin crew and 248 passengers, most of them from the 101st airborne division on their way home. The mood must have been jubilant. It was just a few weeks before Christmas and the 101st airborne division had just finished a six-month deployment in Sinai on a peacekeeping mission. The weather at Gander was light freezing drizzle with snow. Another aircraft which departed that same morning requested de-icing but after the flight engineer did a walk-around, the DC-8 crew didn’t seem to think it was necessary.

It should have been a routine flight. But when the DC-8 lifted off from the runway, it seemed to struggle. It was travelling about 167 knots indicated airspeed when it rotated, which seems reasonable enough for a fully loaded aircraft at 500 feet elevation. Witnesses saw the DC-8 struggle to gain altitude. The airspeed increased to 172 knots but then it began to decrease again.

I wonder sometimes if the passengers noticed. Maybe not, although for the flight crew, it must of course have been immediately obvious that something was wrong.

The DC-8 flew low over the highway, close enough that two truckers said the lights lit up the cabs of their trucks. Some said they that the aircraft was in a slight right bank. Three witnesses reported seeing a bright glow from the aircraft; two of them thought it might have been a fire, although the third wondered if it was the runway lights reflecting off the aircraft.

The nose pitched up, likely a last desperate effort to get the aircraft to climb away from the terrain. But the aircraft continued to descend. It crashed into the trees about half a mile past the end of the runway and skidded for 300 yards through the forest before exploding as it struck an unoccupied building. The fuel in the fully-laden wing tanks burnt furiously. There was no chance of escape, no chance of survival.

The investigating body was the Canadian Aviation Safety Board or CASB (the Canadian TSB was formed in 1990, in part in reaction to this case). The CASB had a pre-designated investigation response team, the Go-Team, which headed out to the site immediately. A crash site can quickly become compromised so it is urgent that the investigation get on site as quickly as possible. The Go-team is a collection of experts, ranging from engines to aircraft structure to data collection from the FDR and CVR. Their first job is to secure the wreckage and collect as much information as possible.

Press photograph of firefighters and emergency responders on the scene. Attribution unknown.

When they got to the site, they searched the runway and the highway, looking for any components or debris from the plane, which could help them to isolate a mechanical failure. They also looked for marks on the runway to see if perhaps the tail had touched the runway during take-off. There was nothing.

It was snowing and soon the entire crash site was blanketed with snow. The investigators had to remove key pieces of the wreckage in order to examine it and then return a second time to inspect the area when the weather was clear. On that second visit, they cleared the trees, melted the snow and documented the crash site as best they could. This time they recovered all of the wreckage and moved it to a secure hangar at Gander Airport, laying it out on a grid pattern to match how it was found at the crash site.

Wreckage from Arrow Air flight 1285 in a Gander Airport hangar. Photograph by SSGT Arnold W. Kalmanson – DoD image number DA-ST-87-05218

The Flight Data Recorder on this aircraft was very basic and only recorded four parameters, none of which was useful in this case. It was scheduled to be replaced a few weeks later. The Cockpit Voice Recorder was also recovered but they discovered that it had not recorded anything to do with the flight, or indeed any of the previous legs. It had been out of order for some time but no one had noticed.

They searched for mechanical failing and structural issues, especially related to the engines and the hydraulics. They found no signs of mechanical failure before the impact; however there was so much damage, it was hard to be sure. The instruments were completely destroyed and could not be read and they had no idea which warning lights might have been lit before the aircraft began to impact the trees: one master fire warning bulb was off but it was impossible to tell the status of the second one. One of the engines Fire Warning Light was off but the other three? They couldn’t tell.

A French news agency in Beirut received an phone call on the day of the crash, in which an anonymous caller said that the Islamic Jihad claimed responsibility for bombing the aircraft. Islamic Jihad Organization was a Shia militia active in the 1980s who were demanding the departure of all Americans from Lebanon.

The terrorist group did carry out a number of bombings and other terror actions, including two car bombs at the US Embassy in Beirut. Every other attack that they had claimed was proven to have been caused by them. On the other hand, there was no record of any other aircraft which had been bombed by the group.

Investigators searched for signs of an explosion, including shrapnel which couldn’t be explained by the impact or other signs of a high-energy explosion. They also looked for signs of lacerations and burn marks in the clothing and luggage. The investigators found no clear evidence of a bomb but the wreckage of the aircraft was such that it couldn’t be dismissed either. The cockpit voice recorder might have easily solved this case, if it had been working.

Mortar shells recovered from the wreckage. Photograph by SSgt. Arnold Kalmanson – DoD image number: DA-ST-87-05230

There were mortars found in the wreckage, although it seemed unlikely that one might have gone off accidentally without any of the witnesses noting the explosion. There was evidence in other flights around the same time that military personnel were carrying dangerous equipment in their cargo, such as flares and explosives. It was against regulations, of course, but it happened. So it is possible, even likely, that there were smaller explosive items in the baggage which could have caused a small explosion or a fire without any help from terrorists at all. Again, with no cockpit voice recorder, there’s no way to know what the flight crew experienced in the final moments before they crashed.

At the time, there had been a few crashes which were demonstrated as the result of icing, however pilots still didn’t truly understand the impact of ice on the wings, especially how ice on the leading edge would affect the flight profile. During take-off and climb-out, the aircraft is at a low altitude and a high angle of attack, near maximum thrust and at high drag. It is a challenging phase of flight but in a “clean” aircraft, it’s not an issue. But it only takes a very light layer of ice to push the aircraft to the edge of the flight envelope. Rough ice, which is what was suspected in this case, has an especially large effect.

Especially if the aircraft is overweight.

Arrow Air worked out the weight using standard estimates, as do most commercial passenger flights: you may weigh your luggage when you check in but you rarely have to stand on a scale yourself; airlines have standard figures for passenger weights which they use to estimate the total passenger weight. Some will be over and some will be under but the idea is that it ends up about right.

Arrow Air had a standard “winter weight” of 170 pounds per person, 77 kilos, which included the person, their clothes and five pounds of carry-on baggage. This figure was used by the first flight crew, who flew from Cairo to Cologne, when they were working out the take-off weight. They found that they couldn’t carry all of the cargo and ended up leaving 41 duffle bags behind, might to the consternation of the military personnel.

At Cologne there was no change in passengers or cargo, so all that the second flight crew had to do was modify the fuel weight, which the first officer did. But later, that flight crew said that when they reckoned that the take-off weight was quite a bit heavier than what it said on the load sheet — around 10,000 pounds heavier. They increased the take-off reference speed to account for this, so that they were travelling faster at the point when they rotated (increasing the pitch angle to lift off) which gave them better performance in the initial climb.

But neither of them thought to mention this to the next flight crew, who took over the flight at Gander. In Cologne with a clean aircraft, a bit of extra weight may not have been a big deal and was easily dealt with by lifting off at a faster speed. But in Gander, with ice on the wing and not quite so much runway to play with, it could have caused the aircraft to lift off without being able to climb away safely.

Investigators attempted to recreate the actual weight on board. They concluded that the average weight per passenger was 164 pounds… but that was not counting their uniforms and their weapons. Some also carried extra clothes and souveniers with them on board. In addition to that was the carry-on baggage, which had caused the cabin crew concern when the troops had boarded at Cairo. This “carry-on luggage” had filled the baggage holds of the two 737 which had brought the troops to Cairo. The true average weight of these passengers, including their clothes and equipment and carry-on luggage, was closer to 220 pounds per passenger.

Charred weapons recovered from the wreckage. Photograph by SSGT Arnold W. Kalmanson, DoD image number DA-ST-87-05231

Also, the cargo weight as logged for the flights to Cairo was about 4,000 pounds heavier than was listed on the load for the departing flight. Even taking into account the 41 duffle bags left behind, it seems that the cargo was some thousand pounds heavier than used for the calculations.

In addition, the basic weight of the aircraft, which should be a set figure that the crew can rely on, was not correct. The aircraft weight didn’t include removeable equipment, which included pillows, blankets, drinking water or even galley equipment. The basic weight was actually another thousand pounds more than had been listed.

Once the investigators had corrected for all of these issues, they found that the take-off weight had been under-estimated by 14,000 pounds.

The ZFW (zero fuel weight) of the aircraft was in reality just four hundred pounds under the maximum authorised weight. Once fueled, the aircraft exceeded the authorised take-off weight by 8,000 pounds.

The Canadian Aviation Safety Board released a final report but the nine board members could not agree on the cause. Five of the board members believed that the rain and freezing temperatures had allowed rough ice to form on the aircraft’s wings.

The Canadian Aviation Safety Board was unable to determine the exact sequence of events which led to this accident. The Board believes, however, that the weight of evidence supports the conclusion that, shortly after lift-off, the aircraft experienced an increase in drag and reduction in lift which resulted in a stall at low altitude from which recovery was not possible. The most probable cause of the stall was determined to be ice contamination on the leading edge and upper surface of the wing. Other possible factors such as a loss of thrust from the number four engine and inappropriate take-off reference speeds may have compounded the effects of the contamination.

However, the other four published a dissenting opinion, arguing that there was no evidence to show that there was ice on the wing and that it was more likely that the aircraft was brought down by an in-flight fire.

In our judgement, the wings of the Arrow Air DC-8 were not contaminated by ice — certainly not enough for ice contamination to be a factor in this accident. The aircraft’s trajectory and performance differed markedly from that which could plausibly result from ice contamination. The aircraft did not stall. Accordingly, we cannot agree — indeed, we categorically disagree — with the majority findings.

The evidence shows that the Arrow Air DC-8 suffered an on-board fire and a massive loss of power before it crashed. But, we could not establish a direct link between the fire and the loss of power. The fire may have been associated with an in-flight detonation from an explosive or incendiary device. Consequential damage to various systems precipitated the crash.

The minority report cites the two witnesses who described bright yellow/orange glow as the aircraft passed over the highway, which the majority report didn’t mention, only quoting the witness who thought it was a reflection of runway lights. Neither report actually documented all three witnesses, only the ones that fitted their conclusion.

A scan of the full report, including the dissenting opinion, can be found here: http://flightopsresearch.org/data/files/arrow1285.pdf.

At the time, he Canadian Aviation Safety Board (CASB) was a new organisation, created in 1984 as an independent accident investigation board to replace Transport Canada’s Aircraft Accident Investigation Branch. It was modelled on the National Transportation Safety Board in the United States. This new board, still a part of Transport Canada, would investigate situations it felt were potentially hazardous and, where appropriate, make safety recommendations. These recommendations were not binding for Transport Canada to respond to.

The split vote in Gander showed the CASB in a bad light, especially when a judge in 1989 stated that he didn’t think either argument had any merit. He determined that neither conclusion was sound. The Canadian Aviation Safety Board lost the faith of the public and, in 1990, the Canadian Transportation Safety Board was formed. It was focused on all types of transport safety (railway, marine and pipeline) and was designed to be truly independent from federal departments and agencies. Its only goal was (and is) to advance transportation safety.

Meanwhile, thirty-five years later, the cause of the crash of Arrow Air flight 1285 is still unknown.

Category: Accident Reports,

9 Comments

  • I vaguely remember the news reports about this sad crash.
    Sylvia seems to have all angles covered, but like in the official reports, it is very difficult to come to an absolute conclusion. It also would appear that the investigating authorities were eager to close the books. Lacking a functioning CVR and an- even then – outdated FDR, there would have been insufficient firm clues that enabled the Canadian Investigation Board to solve this case.
    What do I think was the cause?
    Of course, I will have to start with the caveat that I know no more than what is written in Sylvia’s account, so it is guesswork.
    – The aircraft is on a long distance flight from Cairo with a refuelling stop at Gander. It is early morning (LMT), cold and there is a crew change.
    – Under those circumstances the possibility that the incoming crew had an early wake-up and were not totally alert cannot be ruled out .
    – The aircraft carried military personnel with lots of luggage and equipment, not all accounted for on the weight and balance sheet.
    – Soldiers carried some “trophies” in the form of weapons and ammunition that had not been included in the calculated payload.
    – There were discrepancies in the zero fuel weight.
    – The aircraft may have been substantially over its maximum take-off weight.
    – The reports mention a take-off in conditions of freezing drizzle and snow.
    – The aircraft had not been de-iced. Under the circumstances, even in those days, de-icing should (and probably would) have been mandatory.
    – The runway was limiting, especially as the crew increased the take-off speed. It seems that they were aware of a very heavy aircraft, unfortunately not to which extent the aircraft was over weight.

    So the odds were already stacked against this flight. Some safety margins were compromised without anything going wrong.

    My guess is that the aircraft suffered an engine failure or loss of power.
    Maybe a compressor stall? Who knows.
    Under normal conditions an airliner must be able to continue its take-off if an engine fails after V1.

    But the conditions were not normal:
    A crew that may still have been rubbing the sleep out of their eyes. No, I am not suggesting that they were unfit to fly. But an early take-off is no guarantee that the crew go to bed early, nor that they had a good night’s rest. Yes, conjecture, I admit.
    The wing surface may have been contaminated with ice accretion. Maybe not to the extent to prevent the aircraft from becoming safely airborne under normal cicrumstances, BUT
    The aircraft also seems to have been substantially over its maximum take-off weight.
    It still might have made it, BUT
    If an engine had malfunctioned, even a loss of power could have tipped the balance.

    So I think that there were more factors, all culminating in the aircraft having been unable to reach and maintain sufficient airspeed to climb The crash would have been unavoidable.

    Again, I must repeat that this is no more than guesswork.

  • As Rudy says, there are almost too many factors to consider in this tragic sequence of events, leading to so much loss of life. In some accidents, the cause is a combination of just one or two. In my research into it some time ago now, I came across this comment in a TV series called Unsolved Mysteries. The Gander accident uniquely qualifies.
    About halfway down the comments is a contribution from “[email protected]” which throws some light on the maintenance and crew mindset at Arrow Air. Unfortunately it raises more questions than it answers – and leaves you unsettled as to the fate of its passengers. The author was in no doubt about the DC 8’s final fate.

    https://unsolved.com/gallery/airliner-crash-at-gander/#comment-7024

  • John, I read into the accident report, and my impression from that was that the pre-owned aircraft was somewhat lacking in maintenance, so that may well have been a factor.

    Rudy, the accident report does state that the crew (who boarded on Cologne at 2 am) were probably fatigued. The captain had checked into the hotel at noon and was awake from 7pm onwards.

    The accident report says on page 55 that runway length was not the limiting factor: with that weight, they should have had a ground roll of 6700ft and be at 35ft AGL after 7800ft; 200 ft later if an engine failed at rotation. They had 9900ft of runway (would have been 10200, but the 300ft of runway 22 north of the runway 13/31 intersection hadn’t been snow plowed), so that should have been ample. But the report also says that, based on the FDR data, they appeared to have lifted off later than that, after 8000 ft of takeoff roll (p.18), and their speed of 172 kts a few seconds later should have allowed them to keep climbing. There appear to have been some issues with engine 4 not running at full speed, and engine 2 possibly having been throttled down to 1.34 EPR (p.25).

    This seems to have bern a poorly maintained aircraft piloted by a fatigued crew with incorrect weight on the load sheets flown in icing conditions: lots of small factors that, by themselves, might not have mattered much, but in sum they crashed the plane.

    I can also understand why the families of the victims are looking for some “CIA operation” that was covered up: it’s very distressing to accept that your loved ones died from simple negligence: the accident board did propose a force of nature (icing) or a an onboard fire (unproven) as causes, which is charitable, but ultimately unprovable.

  • I’m a bit surprised by Sylvia’s observation that “pilots still didn’t truly understand the impact of ice on the wings”; this crash happened almost 4 years after the Air Florida 90 crash and over 3 years after the NTSB report blaming the failure to pay enough attention to the iced wings. (It’s also over 10 years after my instrument instructor, who made a living flying 727’s for Delta, pointed to the slight distortion on the front edge of the wing struts of our C172 and said ~”You don’t want that.”) I get that commercial pilots can be conservative about listening to new ideas, but I wonder whether this was a marginal pilot who was an especially poor learner when it came to subtle hazards that had shown up since he got his ATP.

    I’m amused by the dissenters who claimed that ice couldn’t have brought the aircraft down; this sounds like most of the cases Sylvia has brought up, where a number of problems (as enumerated by Mendel, with the additional possibility of the #3 engine failing outright) had to happen simultaneously in order to cause the crash, so their statement is technically true but interferes with getting to a useful conclusion.

    John: that’s a fascinatingly detailed and clear-eyed comment amid all the would-be Illuminati. The crash didn’t need bombs going off (although I have to wonder about the passengers who thought they could put ?live? ammo on a passenger plane); all it needed was chains of people going along to get along. (It’s tempting to say that this an especially severe problem in the US armed forces, but I wouldn’t assume any other country’s armed forces are better and there are plenty of examples in civilian life.) The claim that records were sealed does not surprise me — that was during the Reagan administration, when the US military tended not to be monitored — but I doubt we’ll ever learn for sure. (minor note — the commenter was jDT40u, not jTD40u. The link looks right but didn’t jump to the comment, which is why I saw all the hare-brained responses.)

    • CHip, I meant to respond to this earlier; I wasn’t clear in my statement. Although the issues of icing were already much better known, it still took some time to be taken seriously, especially if the pilots couldn’t see the distortion that your instructor was able to point at. Air Florida 90 was blamed on an inexperienced crew who had departed with visible snow and ice on the wings. At the time, a thin layer of ice contamination was not seen as critical; in fact, one of the dissenting investigators said that it was not possible for a thin layer of ice to bring down the aircraft.

      So it would have been better for me to write that pilots still didn’t truly understand the impact of a thin layer of ice on the wings, which came up again in 1989 when Air Ontario flight 1363 crashed. It was that case that finally put the spotlight on training and ensuring that pilots truly understood the impact of ice on the wings.

  • We must bear in mind that this was a real crash and real people lost their lives. Yet, I enjoy trying to make sense of what Sylvia presents without bothering to look up accident reports or other facts and data.
    Was the runway critical? Apparently not. But the days when I had to deal with concepts like “critical runway”, “balanced field” and other perfomance related jargon are now far in my past and I am too lazy to go and look it up.
    Crew fatigue: I did not check up on that aspect, but I tried to form a picture in my mind. Taking over an aircraft from another crew, early in the morning on a dark, cold apron. Sometimes this can be after a night in a noisy hotel. I was once in a nearly forgotten time working for a company that exchanged crews at Paris CDG. We stayed in the Ibis Hotel smack in the central area of the airport. Many people arriving late from incoming flights, checking in to catch some sleep before taking the train home, scurrying through the corridors, loud talking, tired children crying, doors slamming. Others getting up very early, just when we could have had another half hour’s sleep. And the walls were paper thin, we literally could hear the conversations in other rooms. Eventially all crews got together and told the company that we would refuse to fly unless we were assigned another hotel. Which the company did, we became guests at the Millennium Hotel in Roissy.
    After a wait in the lobby, still yawning, we would be collected by the crew van, and getting issued with the flight documents. This could be in the ops room or at the aircraft. The incoming crew was always in a hurry, so when taking over it often was difficult to extract information about the status of the aircraft. Some items that occurred duing the preceding flight and that, according to the minimum equimpent list, could be deferred to be corrected “at the first suitable maintenance stop” were not always entered in the maintenance log, so we had to rely on the accuracy of verbal reporting. Our crews were professional enough but even so, a misunderstanding is always possible.
    Next the hand-over from the mechanics and waiting for the “red cap”, the ops man with the final documents and the release. All this whilst trying to get our mind “in gear”.
    So I do have a certain sympathy with crews who have an early morning hand-over. But when reading Sylvia’s blog, two things that really stood out for me were:
    – The aircraft had arrived after a long flight. That means that it is not only possible, but even highly probable, that it still was very cold, maybe undercooled. The fuel still in the tanks would also have been below the freezing point. The temperature of the fuel in the bowser may also have been low. Weather conditions were freezing drizzle and snow. Under those circumstances there is absolutely no doubt that de-icing would have been necessary. At that point we would have taken out the aircraft operations manual (AOM) to verify what the specifics were, and what the consistency and temperature of the de-icing fluid should be, extract a hold-over time (the period between commencement of the de-icing procedure and the actual take-off) and determine whether the take-off would even be allowed. A take-off in freezing rain carries restrictions and if not met (e.g. the type of fluid available) may even mean that take-off is prohibited.
    – With an aircraft over its MTOW and in icing conditions, a loss of power could have compromised the performance to the point where the aircraft would have been unable to climb to a safe alitude. As the reports suggest: it failed to clear obstacles (trees). The reports of fluctuating airspeed would support this theory: the crew tried to gain speed in order to climb but as soon as they did, the speed dropped again as soon as they tried to climb away.
    Of course, we do not know for certain and it is unlikely that his crash will ever be solved.

    • Sounds like somebody at your company wasn’t thinking about why some hotels are cheap. The instant I saw “Ibis” I knew what the story was going to be, having stayed in the one next to Heathrow; it was clean and a good place to recover after being a passenger on a jammed overnight flight (the first day after the probably-more-expensive Easter season), but it was also clearly cheap. Good on your people for standing together and getting safer accommodations.

      The comment that John pointed to suggests that the crew might not have done the AOM-based calculations even if they’d been fully awake; their employer sounds so shady that shoddy procedures (or none at all) would be routine. (That sloppiness could have made the too-hasty handovers you described in your previous comment even worse.) That was also a time when employees in general (not just macho ATPs) were expected to suck up odd hours, before studies made clear just how much a lack of sleep cost alertness. A particular failing was putting people on shifts that rotated earlier, e.g. some time on the 0800-1600 shift followed by time on the 2400-0800 shift; since most people’s natural cycles will run more than 24 hours long if they’re left alone, moving to an earlier shift was especially stressful.

  • The Millennium Hotel was the scene of a funny incident. Totlly unrelated to this week’s blog but I think that you will enjoy it:
    We were flying air cargo. FedEx had a large base at Charles-de-Gaulle. These operations all took place by night.
    And so, one very early morning we – the crews of maybe 5 or 6 aircraft – were waiting in the lobby for our crew transport, rubbing our eyes when the door of one of the elevators opened.
    As we looked in the direction of the sound, expecting to see another crew, instead to our utter surprise out came a very attractive young woman. She was stark naked.
    Apparently, having spent most of the night in someone else’s room she had decided to quickly nip out in the corridor and into her own room.
    But she had locked herself out and probably decided that it was better not to attract attention to herself by banging on doors. Instead, not expecting anyone to be around at that time, took the elevator to get a spare key from reception. Only: instead of a discreet receptionist she was unexpectedly confronted with 10 or 12 pilots who were suddenly very wide awake.
    To her credit, her reaction to the situation was superb. In the first shock she covered her breasts and lower body, but then she dropped her arms and walked by as if she was the Queen, dressed in an evening gown. The receptionist issued another key and she walked back to the elevator as if being naked in a hotel lobby was the most natural thing to do.
    Well, needless to say, all those pilots were not yawning any more.

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.