A Gulf AIR GF072 Reading List

Grey-shaded Links are now dead (because of Pprune Server Change-over)
http://www.pprune.org/ubb/NonCGI/Forum1/HTML/009550-10.html  Page 10 of 38 pages Major discussion thread (look for BELGIQUE posts and references to somatogravic pitch-up illusion
http://www.pprune.org/ubb/NonCGI/Forum1/HTML/010349.html  FOUR PAGES Preliminary Results

Look for BELGIQUE Posts and links listed there related to the somatogravic pitch-up illusion

http://www.bahrainairport.com/gf072factualinformation.htm  GF072 factual information download pdf files (need acrobat free download to load/read them)
http://www.pprune.org/cgibin/ultimatebb.cgi?ubb=get_topic&f=1&t=013394  TWO pages link to second page at the bottom of first
www.gulfairco.com,   and    www.bahrainairport.com latest reports published here  
http://www.pprune.org/cgibin/ultimatebb.cgi?ubb=get_topic&f=1&t=013156  ONE PAGE  
http://www.pprune.org/cgibin/ultimatebb.cgi?ubb=get_topic&f=1&t=013942  FOUR PAGES inexperienced flight crew
http://www.pprune.org/cgibin/ultimatebb.cgi?ubb=get_topic&f=1&t=014759  ONE PAGE  
http://www.pprune.org/cgibin/ultimatebb.cgi?ubb=get_topic&f=1&t=012972  ONE PAGE  
http://www.pprune.org/ubb/NonCGI/Forum1/HTML/013394.html  TWO PAGES  
http://www.pprune.org/forums/showthread.php?s=&threadid=9199&highlight=somatogravic What is a Somatogravic Illusion?  
http://www.pprune.org/forums/showthread.php?s=&threadid=66543&highlight=somatogravic The False Climb Illusion  
http://www.pprune.org/forums/showthread.php?s=&threadid=59924&highlight=somatogravic GF072 Probe blames captain 6 pages
http://www.pprune.org/forums/showthread.php?s=&threadid=3510&highlight=somatogravic GF072 Crew very Inexperienced 4 pages
http://www.pprune.org/forums/showthread.php?s=&threadid=1364&highlight=somatogravic GF072 Report Out 2 pages



Aviation Medicine Links


Dr Dougal Watson

emails etc for inquiries about somatogravic illusions

wrote the two pdf files below

http://members.ozemail.com.au/~aupa/PDF_files/DNT.pdf  Dark night takeoff need acrobat reader
http://www.ozemail.com.au/~aupa/PDF_files/A&L.pdf  Approach and landing illusions. need acrobat reader
I came across this reference in the aeromed-list mailing list. The article does a great job of explaining, and illustrating, how lack of visual references on a dark night can easily, and logically, fool both your body and your brain. Intriguing reading. If you ever fly on moonless nights, or think you might, this is good information about staying safer. Laine Tammer, CFI, #2810
Dark Night Takeoffs And The “False Climb” Illusion
By Dougal Watson

Aircraft have been destroyed and many aviators, and their passengers, have died as a result of the ‘false cliimb” illusion. Unlike the approach and landing illusions that rarely result in more misery than a hard landing or a missed approach, this one is a killer. Understanding the mechanisms behind the “false climb” illusion is quite difficult, but is an important first step in avoiding becoming one of its victims.

The false climb illusion is a classic example of the limitations of our senses, especially sight, balance, and touch, during flight. This illusion occurs when our otolith balance organs provide misleading information to the brain and there isn’t enough information from the eyes to correct the error.

How could a healthy, command instrument rated, very experienced pilot fly a perfectly sound Beech King Air into the ground only seconds after taking off into a clear, unlit night sky. He did at Wondai, QLD, several years ago, resulting in his own death and that of four of his five passengers. This accident and many, many others like it have the common features of night time departure, dark sky with no visible horizon, and unlit terrain under the take-off path.

To understand the false climb illusion, which is also known as the somatogravic illusion, we will need first to review the workings of our ear’s balance mechanisms – the otolith organs. There are two otolith organs in each ear – one vertical and one horizontal. Their main function is to provide the brain with information about the position of the head. Each otolith contains small crystals attached to the free ends of tiny sensory hairs which are, in turn, connected to special nerve cells.

When the head is tilted gravity acts on the crystals and causes the sensory hairs to bend. This bending of the sensory hairs stimulates the nerve cells which then send signals to the brain. The brain uses the signals from all four otoliths to compute the position of the head. This process occurs very rapidly, very frequently, and without any conscious effort.

It is this resultant force that the otolith organs actually sense. The confusion, and the source of the “false climb” illusion, arises because man has evolved in an environment where gravity is the main force that influences our otolith organs. We are used to interpreting signals from the otoliths as indicating the position of our head.

To be continued in next month’s Flyer.

The URL for the PDF file of this article is http://www.ozemail.com.au/~aupa/PDF_files/DNT.pdf

Do You Read Me?

Oct 09, 2000


There were no final words indicating any difficulty to the air traffic controllers (ATC) from Capt. Ehsan Shakeeb or First Officer Khalaf Al Alawi before their airplane, Gulf Air Flight 72, suddenly plunged into the waters off Bahrain (see ASW, Sept. 4). According to an ATC transcript released Oct. 6 by Bahraini officials, the controller did not receive any response from the aircraft after passing a new frequency. "GF-072, do you read me?" he asked. No response. The question was repeated. Again, no response.

Bahraini officials provided the following additional information:

The first missed approach: "The Company (Gulf Air) procedures require that, under visual conditions, the aircraft be established on an approach by 500 ft. in order to complete the landing as planned. In the case of GF072, this was not achieved...including not having the correct landing configuration. The Captain then elected to perform a '360? orbit.' Company procedures require this to be undertaken at a minimum altitude of 1,000 ft. In this case, the actual orbit was initiated at an altitude of 600 ft. Subsequent altitudes during the orbit ranged between 940 and 330 ft. at a computed air speed decreasing from 155 to 135 knots."

The 'Go Around': "On the second approach, as the aircraft was again not established for a landing, the Captain elected to perform a 'go around'...The Company procedures require that this be performed at a nose up pitch attitude of 15?. In the case of GF072, this manoeuvre was performed at a pitch attitude decreasing from 9? nose-up to 5? nose-up with computed airspeed increasing from 134 knots to 194 knots. At a height of approximately 1,100 ft., Flight Data Recorder (FDR) data indicates a nose-down movement of the Captain's side stick. The aircraft's pitch attitude responds with a change in pitch to a maximum of 15 nose-down resulting in the aircraft entering a rapid descent. Subsequent nose-up pitch commands produce a reduction in pitch attitude to 6.7? nose-down at or just prior to impact, which took place at a computed airspeed of approximately 280 knots."

Pilot Errors Highlight Continuing Importance of Basic Flying Skills

Apr 30, 2001

Any assertion that pilot error is decreasing as a major cause of accidents is belied by the recent record of incidents and accidents. In fact, a few prominent cases suggest the need for a renewed emphasis on basic skills and adherence to prescribed procedures.

A recent study by researchers at Johns Hopkins University suggested that pilot error as a factor in accidents was decreasing (see ASW, March 26). However, the database for that study ended with 1996. In the years since, a great many fatal accidents stem largely or in part from pilot error One of the more notable cases may involve the August 23, 2000 fatal crash of a Gulf Air A320 at Bahrain shortly after the decision to go-around after a second failed approach. Bahraini authorities just recently released some 100 pages of factual information surrounding the crash. As one pilot observed, even though it was night, the weather was good and the approach was not complicated. There were minimal distractions from other traffic or ATC commands. Yet the first approach was aborted, then the second, and the crash occurred while the crew was trying to position the airplane for a third approach.

Now consider the details. The 4,000-hour captain was the pilot flying. The 600-hour first officer had 400 hours in the A320. Two points of view relate to the relatively inexperienced first officer. Some believe that because of the excellent stability inherent in fly-by-wire (FBW) aircraft like the A320, these aircraft are easier to fly than conventional aircraft and therefore are safer in the hands of relatively inexperienced pilots. The other point of view is that experience is the pilot's best asset. To this school of thought, it is unrealistic to expect a pilot with 608-hours total flight time to be proficient in all of the tasks expected of an airline first officer (see ASW, April 2).

First landing attempt

On the first failed approach, the airplane was moving too fast, at 300+ knots while descending through 2,000 feet. At 2 NM from the runway threshold, just 600 feet above the ground, the captain said, "We're not going to make it" and elected to perform a 360 deg orbit to the left. This decision, according to sources, was a major error in judgement. Instead of performing a standard go-around, the captain elected to circle around, with the gear remaining down, flaps not adjusted and positive rate of climb not established.

According to the Flight Data Recorder (FDR) readout, the turn began at an altitude of 564 feet above ground level, or AGL (Gulf Air's policy in effect at the time required a minimum altitude of 1,000 ft. AGL for such a maneuver). During this turn, the altitude varied from 965 ft. AGL to a scant 332 ft. above the ground. In other words, from the altitude at entry into the turn, the airplane rose 400 feet and descended 232 feet. That is a 600-ft. variance in a simple turn. The inaccurate maintenance of height during the turn may be explained by the fact that the crew was flying visually in relation to the bright lights of the city of Manama and the runway lights. In the latter part of the orbit they lost sight of the bright lights and suddenly were pointed into the inky blackness of the Persian Gulf.

Second attempt aborted

In any event, coming out of the orbit, the airplane was not properly configured for a second approach. The captain announced, "We overshot it." Another left turn was begun. At local time 19:28:57 the engines were at take off go around (TOGA) thrust. Flaps were still down. Instead of applying the requisite 15deg nose up, the FDR indicates that the captain applied just 5deg nose-up pitch. As a consequence, the airplane at TOGA thrust would accelerate rapidly.

This rapid acceleration can lead to a phenomenon known as false climb or "pitch up" illusion. The pilot's brain misinterprets the stimuli, perceiving incorrectly that the airplane is climbing. To cite from an excellent paper by Dr. Dougal Watson, "Because there is no horizon or ground lights to be seen.... Consciously or subconsciously the pilot applies judicious forward stick to 'correct' his perceived nose-high attitude. This compounds the problem by allowing the aircraft to accelerate even more causing a worsening of the illusion...the pilot rapidly 'corrects' his way into the ground." The sinister false climb illusion, Watson notes, can affect both pilots simultaneously (see ASW, Sept. 11, 2000).

At time 19:29:41 and an altitude of 1,054 ft. AGL, the aural master warning sounded, indicating a flap overspeed condition. The so-called "barber pole" would appear in the speed portion of the Primary Flight Display (PFD). At the time the overspeed warning occurred, the airspeed was 191 knots. The maximum speed for the "flaps three" configuration is 185 knots.

Two seconds later, the captain pushed forward on the side-stick controller, holding it in that position for the next 11 seconds. This action put the airplane into a -15deg nose-down attitude.

The factual reports do not analyze these control inputs, but an experienced A320 pilot described the pitch inputs, from an insufficient +5deg nose-up to a -15deg nose-down, as absolutely the worst response at such a low altitude. About 15 seconds later the airplane struck the Persian Gulf waters at a point about three miles northeast of Bahrain International Airport. All 143 aboard were killed.

Why did the captain push the stick forward to drop the nose? With an overspeed, the correct response would have been to pull back on the sidestick, increase pitch, hence angle of attack, and thereby decrease the airplane's speed. At fixed TOGA thrust, speed is controlled by pitch.

Incorrect control inputs

To increase pitch would have meant moving into that descending barber pole. By dropping the nose, in a direction away from the barber pole, the pilot actually increased the airspeed during the overspeed condition. The pilot flying pushed when he should have pulled.

As one industry official observed, "We teach people to fly away from the red." Indeed, the response when the needle on an automobile tachometer goes into the red is to back off on the accelerator. However, this situation required the pilot to fly into the red in order to bleed off excess airspeed. To be sure, the attitude indicator clearly told the story. Even with the evident confusion in that cockpit, it is difficult to believe that a seasoned pilot would pay less attention to the primary attitude instrument while attempting, incorrectly, to deal with a slow-moving "barber pole" warning on an adjacent speed tape and its insistent "ding - ding - ding" aural warning. If so, perhaps deep reflection may be in order on whether or not urgent aural warnings can promote deeply irrational actions.

In those critical seconds, the Gulf Air captain, possibly experiencing the somatogravic pitch up illusion, utterly lost situational awareness. When the ground proximity warning system sounded its "Whoop, whoop, pull up!" alarm, the FDR readout indicates that he did not apply full aft sidestick. And throughout the final moments, according to the factual reports, the first officer's sidestick did not move.

If the basic go-around standard operating procedure (SOP) had been executed correctly in the first place, it is likely that the captain, his crew and passengers would not have ended up dead in the water.

Even so, an airline pilot should be able to execute a simple 360deg turn, and to react correctly to an overspeed warning. But there is more involved here than "basic flying skills." Loss of situational awareness and an apparent failure to properly execute a missed approach/go-around seem to have played their compounding roles. The seminal question: Why did this crew not adhere to approved SOP when it failed to properly fly a "piece of cake" missed approach? As one pilot observed trenchantly, "Transport category airplanes are NOT flown by looking out the window. They're flown 'by the numbers' - proper airspeed, proper flap position, proper engine thrust, proper altitude, etc. These numbers are NOT outside the window. They are inside the cockpit - on the PFD, on the Upper ECAM (Electronic Centralized Aircraft Monitoring), etc. And, given all that, adherence to SOP is absolutely essential, including a proper crew briefing."

Gulf Air has since beefed up its policies, procedures and training (although the crash investigation is not likely to be completed until later this year). Nevertheless, as one source said, this accident should never have happened - and it may well provide a peephole into a more generalized erosion of piloting skills. (For the Gulf Air accident factual reports, see http://www.bahrainairport.com/gf072factualinformation.htm. For more on pitch-up illusion, see http://www.ozemail.com.au/~aupa/PDF_files/DNT.pdf )

Partial Listing of Recent Accidents Involving Human Error
05 Nov 00 CDG 747 Cameroon

Overran - severe damage
31 Oct 00 Taipei 747-400 SIA

SQ006 (took off on wrong runway in poor vis)
05 Oct 00 Reynosa DC-9 AeroMexico

Skidded off runway / destroyed
23 Aug 00 Bahrain 757 Gulfair

Apparent pilot disorientation after go-round
17 Jul 00 Patna 737 Alliance Air

Stalled during gear-down steep turn onto short finals in reduced visibility
12 Jul 00 Vienna A310 Hapag Lloyd

Gear-down transit / believed FMC fuel figures / ran out of fuel
25 May 00 CDG France SD360 Streamline
Entered runway without clearance and collided with 727 taking off
19 Apr 00 Davao 737 Air Philippines

Crashed during circling approach to land on reciprocal (in poor visibility)
05 Mar 00 Burbank 737 SWA

ATC kept high and pilot tried unsuccessfully to burn off height / overran onto highway

The accident can be put down to:

a. inexperience  b. poor pilot discipline (and training)  c. loss of situational awareness and spatial disorientation in the form of the somatogravic pitch-up illusion.

It is worth noting that the approach to Manama has claimed many victims because of the dark waters of the Gulf. Because of this deadly phenomenon a night visual approach, without target height gates at DME distances can easily lead to exactly the situation that the GF072 crew found themselves in. Contributory is the fact that Saudi ATC normally keep inbound aircraft high until they have overflown Dhahran - so they end up making a plummeting descent when landing to the South at Manama.  An ILS should be available on both ends of the main instrument runway and an ILS approach mandated at night. Orbiting in an environment that offers bright visual references (of Bahrein itself) for only 1/3rd of the turn (and the inky blackness of the Gulf's dark-pool for the rest plus the climbing overshoot) - was always going to be a bad idea.

 Gulf Air has taken some drastic steps to recover its reputation - some appropriate and some quite inappropriate. It has hired consultants and initiated new safety and training programs. However, in what I would call a typical Arab red-herring tangent, they have included the significant expatriate pilot population in a scape-goating exercise in which they psychologically vet pilots quite astringently for job suitability. This has led to at least one very experienced British Airbus captain being grounded pending disposition - and much unwarranted stress for the remainder (yet to undergo their own inquisitions). It is also typical of the Arab psyche (in my considerable experience of that) that they should be using an accident such as GF072 for evening up old scores and eliminating their internal critics..

The pitch-up illusion is a very powerful illusion and requires knowledge combined with very strict personal discipline in order to escape its cloying clutch upon the mind's ability to think clearly. When it happens suddenly on a dark night, the reaction is normally instinctive and in completely the wrong sense. The error is reinforced by the physiological response to the phenomenon and the outcome is nearly always fatal. It remains the answer to many such "unexplained" accidents - particularly on dark night take-offs. However Derek Piggott, (the famous UK Gliding Instructor) wrote a lengthy article which became a booklet upon the subject - and threw new light upon it. He noted many accidents where glider-pilots on winch launch would arc-over and fly into the ground in bright visual conditions - simply because of the abject confusion wrought in their vestibular system by the heavy acceleration of a steep winch-launch. The trials done by GulfAir's investigators are claimed to have more or less eliminated this as a cause - however like many illusions they are triggered by a particular set of conditions and aren't easily replicated on demand. The USN has chronicled many many instances of it - particularly on night carrier catapult launches.

The more worrying aspect about the pitch-up illusion is that it can happen to both pilots simultaneously - and that would possibly explain why the albeit inexperienced GF072 second officer did not intervene. However at his level of experience and being in awe of the captain, CRM was probably far from his mind. If the captain was "behind the aircraft", the inexperienced F/O was most probably at that stage simply "along for the ride", puzzled but unaware of why what he saw and felt - was happening. That's understandable because it is not able to be replicated in a simulator. Most pilots experience it for the first time just before their untimely demise.

To Hot Off the Press


Click here for a free download of Adobe Acrobat Reader