Another one in the sea at night

With the loss of the Ethiopian Airlines Boeing 737-800 offshore from Beirut on 25 January, the phenomenon of fundamentally serviceable aircraft – and all their passengers – being lost over the sea at night is becoming frightening.

I have put this issue under the spotlight before.

Here’s a list of the main airline losses in this category since 2000. There have been seven such accidents, and together they have killed 976 people:

2010 Ethiopian Airlines 737-800, Mediterranean Sea near Beirut

2009 Yemenia Airbus A310-300, Indian Ocean near the Comoros Islands

2009 Air France A330-200, South Atlantic

2007 Adam Air 737-400, Java Sea near Sulawesi

2006 Armavia A320, Black Sea near Sochi 

2004 Flash Airlines 737-300, Red Sea near Sharm el-Sheikh

2000 Gulf Air A320, Arabian Gulf near Bahrain 

In some of them the cause has been established officially.  Gulf Air, Armavia, Flash and Adam Air were all caused by a combination of total or partial pilot disorientation followed by a failure to control an aircraft that could have been controlled.

Yemenia is known to have hit the sea having stalled at quite low level during an attempted circling approach at night, so that could be put in the loss of control/loss of situational awareness category.

Air France was known to have suffered some specific technical anomalies, but in the absence of new information indicating the the aircraft was physically uncontrollable, the crew should have been able to maintain control but failed to do so.

And now Ethiopian Airlines. It sounds distressingly like several of the others, particularly the Flash Airlines event. In the latter, the captain was carrying out a slow turn over the sea at night, and from what he said to the copilot, he clearly had the “leans”. The aircraft went into a spiral descent and crashed. The copilot could see what was happening but left intervention too late.

We have learned from Beirut air traffic control that the aircraft was on a northerly heading soon after take-off from runway 21. We believe it was heading to a coastal way point north of Beirut, possibly with the intent of crossing Lebanon eastward, bound for Syrian airspace. There it would, presumably, have turned south toward its destination, Addis Ababa.

But the flight, still in its early northward climb, was told to turn left onto 270deg to avoid traffic inbound to land on 16. The aircraft’s left turn continued through 270deg around to 140deg, despite warnings from ATC. The crew did not respond, and radar contact was lost.

There were thunderstorms in the area, and this may turn out to be a contributory factor in what happened. But right now it looks terribly like another case of pilot disorientation over the sea at night.

This is an undeniable phenomenon now, but no-one is recognising it as such. Whenever it is finally recognised, determining more appropriate pilot training would be a priority internationally. Maybe the aviation insurance industry should start lobbying, but it’s a pity they should need to.


11 Responses to Another one in the sea at night

  1. David Nicholas 29 January, 2010 at 1:11 pm #

    Perhaps being hand-flown, with (more than) one eye on the weather radar, and with course changes, seems fertile ground for disorientation. 9000 feet unwinds quickly if control has been lost. We’ll know soon enough – the recorders have been located in deep water but not beyond recovery.
    If you haven’t read it yet, see the Flightglobal link below:

  2. Bungling Fergus 30 January, 2010 at 2:22 pm #

    So what are you suggesting here? If you’re not practiced at instrument flying, turn on the autopilot? Makes sense if that is the only safe option, but surely a current and proficient pilot is desirable? Are we allowing skills to deteriorate to the point that we should not handle the plane? What happens when major electrical fault after take off etc that results in AP/FD/FMC failure with standby power instruments only available to hand fly an instrument approach? I think hand flying absolutely should be practiced at appropriate moments – quiet traffic, VFR, day etc. A major issue to consider is how heavily am I loading up the other guy if I hand fly. I must not task saturate him to the point that he cannot do his primary non flying mission – monitoring! I think the ALPA safety dude is right.

  3. David Nicholas 1 February, 2010 at 12:57 pm #

    Fergus, I’m suggesting no more than we already think we know! Loss of control is becoming the most significant causal factor in accidents, which is not what we might reasonably expect from professional pilots. I learned my flying almost 40 years ago, and airmanship was drummed into everything I did from preflight to postflight. The word (and what it infers) seemed to go out of fashion with the advent of the two-crew glass cockpit, but the bean counters insistence on balancing cost against risk may now be coming to be recognised as the devil’s compromise that it always was.
    There is perhaps another underlying factor – extreme navigational accuracy is now demanded by air traffic control, particularly in busy terminal airspace, which might in some circumstances place excessive demands on a crew’s ability to manage the aircraft in the post-departure phase with one of the pilots hand flying. Small wonder that the autopilot plays such an important part in enabling them to do this. Thus precise flying skills become dulled through lack of regular practice. All the exotic kit in the world does not change the nature of an aircraft, it’s safe flight envelope and it’s propensity to bite the unwary pilot if any one of the margins of that envelope is compromised.

  4. v a fernandez 4 February, 2010 at 12:01 am #

    Lets not pass judgement or irrational speculation till the recorders are recovered

  5. Nevin 5 February, 2010 at 2:01 pm #

    Very sad accidents all of these. If the investigations do reveal it was pilot disorientation then I will also be very surprised. Now Im not an airline pilot so have no experience flying glass so pardon my next question.

    I have seen the flightdecks of these NextGen 737s and what caught my eye were the nice bing chunky screens, specially the PFD with neat little square divided into the classic blue=sky, brown-ground. With such a large display staring at you how do you get into trouble with disorientation. It was harder with the smaller steam gauges but these LCD displays fill your eyes more. Is the fact that it was night combined with the rapid acceleration of the jet that overrode the senses?

    Would be great to hear from a 737NG pilot or of course pilots of any glass airplane.

    Thanks as always for a great blog!

  6. David Learmount 5 February, 2010 at 2:21 pm #


    Disorientation is normally the result of not believing what your instruments tell you because strong messages from your human balance sensors (in the inner ear) tell you that what you see is different from what you feel. If it is night, you cannot use the natural horizon to inform you of which sensation is correct – sight or balance sensors – so there is a strong urge to believe your natural sensors rather than the “artificial horizon” which, your confused state will tell you, could easily be faulty.

    The human balance sensors can easily be fooled by linear accelerations or by slow changes in attitude that take place while you are distracted by another actitivity. The sensation you experience when what you see on the instruments is different from what you expect to see is very powerful, very confusing, and very unsettling. American pilots often call disorientation “vertigo”, because it feels something like that.

  7. Nevin 5 February, 2010 at 8:58 pm #

    Thank you very much Mr.Learmount for explaining that to me!

    I understand what you mean now – that the sensation felt by the human operator is *so* strong so as to override what the artificial horizon is telling you even. I can see why they call the sensation “vertigo”. Perhaps even more intense human-factors type training will help prevent more of these?

    After all you dont seem to hear about too many fighter jets getting into trouble like this and those pilots have to turn and burn agressively – even at night. For example a Eurofighter scrambling to intercept at night – immense thrust! Or perhaps their big HUDs help the pilots a little.

    Also I have been following the development of Synthetic Vision (SV) displays. By seeing an actual 3D representation of whats outside would it not help in countering what your confused senses are telling you? Also pardon me if Im wrong but these displays seem to be mostly on small GA planes (new Cessna 172s etc) and businessjets rather than large commercial airliners. On new-build Airbus & Boeing planes why cant the PFD show something like this?:

    The 787 flight deck looks very advanced although Im not sure if it offers SV.

  8. Paul 17 February, 2010 at 6:48 pm #

    What about the Kenya airways 737-800 accident at Doala, this may qualify as well?

  9. Paul 17 February, 2010 at 6:50 pm #

    *Last comment*

    Sorry this was ot into the sea, but rather swamp or something similar

  10. Eric Groen 11 March, 2010 at 2:51 pm #

    As aerospace physiologist, it occurred to me that the articles about the disorientation accidents speak of the “somatographic illusion”. I would like to point out that the correct way of spelling is “somatogravic illusion”. This may seem insignificant, but the name itself explains that in this situation the pilot errs in the perceived direction of gravity (caused by the fact that without adequate visual feedback our brain confuses sustained linear accelerations with attitude changes). Similarly, the word “somatogyral illusion” refers to false sensations of body rotation, e.g. an important factor in accidents attributed to the graveyard illusion. Whereas these illusions arise from pecularities of our equilibrium organs in the inner ear, there are also many examples of visually induced SD (e.g. misperceived altitude when approaching a runway with different length/width ratio). Although pilots generally refer to SD as “vertigo” or “leans”, this does not reflect the underlying physiological mechanism. It rather reflects their state of confusement once they realize something’s wrong. Unfortunately, a pilot may be disoriented long before (s)he recognizes it, and hence does not feel leans!. Obviously, this type of SD (unrecognized or Type I) is most dangerous and has been the primary cause of CFIT accidents. There is no specific training to counter Type I SD, except perhaps more vigilance in instrument cross-checks, and awareness training to help recognize the circumstances where it typically may occur. In many countries this type of training is included in the syllabus of military pilot training (see link to RTO publication), since SD contributes up to 20-30% of all mishaps in military aircraft.

  11. Daniel C 10 August, 2011 at 2:32 pm #

    re AF447 possible pitot tube failures:


    Add to your list of uncontrolled descents into water at night, AeroPeru flt 603 and Birgenair (both 1996) and also Boeing 757s…where either contamination or maintenance error caused mis-reads/disparities on ASI with consequential Auto-Pilot errors.
    Is there a case for a mechanical back-up for the pitot system where such an error occurs – e.g. like the emergency power generator which, if deployed quickly (and automatically) in the event of a cross-disparity between ASIs, might substitute for the pitot system and restore A-P/Flt Director control leaving the pilots with a system error rather than a control problem – I would argue that both events above also had similar issues associated with your “list” and also add to the casualty list which you enumerate. Rather than disputing your analysis, I believe it needs to be expanded to include pitot/static failures and the consequences thereof.
    From an interested amateur.