NTSB update on Asiana 214

[glendalegoon]

Regarding Visual Approaches.


AS with any aspect of flying (even tying your necktie), you have to practice to get proficient and maintain proficiency.

Does visual approach experience in one plane transfer to another. I say: yes.


Can visual approaches be taught? I say: YES.


Can meaningful aids, such as distance vs altitude and ground speed vs descent rate be taught: I say: YES.


How much time would I guess the pilots involved in this accident spent on considering the visual approach? I say: negative ten hours.

[safetypee]

What is airmanship, how is it taught, and what do we mean by ‘up’.

RAT, et al, I have made this point before, but worth repeating.
“Children of the Magenta Line’ is a useful training tool, but one specifically designed for a particular operator with a specific ‘culture’, thus it is important to understand this before use.

A major message is ‘Click – Click’, the need to disengage automation, but there is little guidance as to how a pilot determines when to disengage. There may be operator / culture specific assumptions about this; thus before using this video it is important to understand the operator’s assumptions and think about any others which you might apply.
What might be obvious with hindsight might not be so obvious to a crew in ongoing situations.

There is an important need to provide guidance (tools) to aid pilots identify situations where there is a need to change the mode of operation – both autopilot and A/T; when to change the mode of thinking.
There range of possible situations is endless, thus publishing specifics is very restrictive, perhaps only those which you can think of beforehand. Or issues which you dismiss as ‘no one would do that – I couldn’t’ possible believe anyone would do that’, as in many posts in this thread.
The industry needs to identify generic advice and training to help in this process, this probably involves aspects of Airmanship or how to ‘keep the Blue - Side up’; can we answer these points?
What is airmanship, how is it taught, and what do we mean by ‘up’.

[BOAC]

Firstly, COTML IS an excellent training tool - not for pilots but for those in management and who set the priorities of the airline in training pilots

‘Click – Click’, the need to disengage automation, but there is little guidance as to how a pilot determines when to disengage.

- never mind 'when' - more of a problem is the trepidation in so doing, since
a) They have little idea how to fly manually
b) There is almost certainly an underlying culture or pressure not to do so but to 'use the automatics'
c) Sometimes there is concern that the other pilot wil be 'spooked' by the disconnect

The days of being 'proud' to execute a safe, comfortable visual approach from anywhere around the airfield have gone. Now it is more important to be an 'ace' with the buttons. I have lost count of the times my co-pilots have declined a visual downwind in CAVOK and gone out to 15 miles for an ILS - well, tried to decline, shortly to lose the 'PF' status.

What is airmanship - you DO have some good questions.... how old is that one?
Is it
a) The desire to save one's skin, thereby
b) looking after pax, crew and tin/plastic
c) the desire to constantly strive to learn more and become a better pilot.

for starters? I think one is either born or brought up wirh the mindset but there is huge scope for improving the faculty. 'UP' is easy for airline pilots.

[safetypee]

A continuing willingness to learn and improve.

BOAC,
Beware generalisation - “They have little idea how to fly manually”
Is ‘they’ a reference to the participants in ‘handling’ accidents, or is it a more general industry-wide view.

Given the number of successful everyday operations the industry does not appear to be suffering handling problems; yet a very few individuals, in specific combinations of circumstance, fail to provide the performance expected of them, for which general they have been trained and examined.
The latter is not the best basis for an industry wide ‘campaign’ on handling, but this must not overlook less available evidence of non-accident problems in normal operations, nor the possibility that we have discovered the tip of an iceberg; how might we know.
Re ‘up’, this assumes that your view (or any operator / regulator) of an airline pilot’s knowledge is the same as yours which defines ‘up’; if it is then how was this obtained, can it be shared, and can it be assured that the views will remain in sync.

You and I are from the same school of flying – perhaps the best in its time. I believe that we were taught airmanship, although not aware of this at the time. As a late learner I have come to appreciate the value of the training, but even some 40 years on I cannot describe what ‘it’ is, or the mechanism of training to achieve ‘airmanship’. However, ‘it’ appears to be a combination of aspects human behaviour and a continuing willingness to learn and improve.

[BOAC]

Is ‘they’ a reference to the participants in ‘handling’ accidents,

- that is what this thread is about, - but trending towards

a more general industry-wide view.

which seems to be the landscape we see. Here's a thought for your grey matter should it become idle....regarding our common experiences, has the 'last resort' of ejector seats in military aircraft affected the number of pilot error accidents?

I am a little lost on developing the 'UP' topic now.

[IcePack]

Makes me laugh. Airbus saying more basic flying skill should be trained. It was them that decided to invent an aeroplane with no "feel" static Throttles sorry thrust levers. Even if you use them they have all of 2.5 " of movement. So it would seem that a lot of these "skill" problems fall fairly & squarely at their door.
(40 yrs exp. Started on raw data 4 eng props to the a330)so seen it developing over the years so find it is QED when someone does what caused this accident.

[alf5071h]

Ice Pack, perhaps a pedantic point, but if non-moving thrust levers (not feel) are a contributory factor then this should only apply with A/T engaged.
It should not be a problem with A/P + A/T engaged, but may require a slightly different scan pattern to balance the loss of visual cue.

If the A/T is used for manual flight then pilots may have to adapt their way of flying; point the aircraft (flight path management) and trust in auto speed, but if speed mode is not engaged … ...
another view considers that the thrust levers set the energy level (fixed values for climb / decent), and that the elevator distributes the energy – the attitude / speed relationship, which might be better understood with manual thrust conrtrol.
IRRC this subject has been discussed elsewhere, where the descriptive ideas originated from work by the eminent Dr Werner Pinsker (RAE).

[safetypee]

Excuse a slight thread drift:- ‘has the 'last resort' of ejector seats in military aircraft affected the number of pilot error accidents?’ I don’t know, but I prefer not to use ‘error’.

A life-long memorable event from training:-
Shortly after ‘first-solo’ with continuing attention on aerobatics and judgement of engine-out landings - altitude, speed, high/low key, geographic position, and speed/wind; the weather pattern changed more conducive to instrument flying. After an introductory flight, the next flight was with an alternative instructor, with an overbearing reputation, and included a GCA. Shortly after stabilising the descent the instructor closed the throttle without further comment; my thoughts immediately focussed on speed, altitude, high/low key, … perhaps overlooking geographic position in deference to altitude, etc. After what appeared to be an age, the instructor said “what now”; I was lost for an answer - wrong mind set – “this is a situation in which you eject – not an engine-out landing”. The GCA was resumed; only then with debrief and extensive self-analysis was the overriding importance of understanding the situation made so clear – the basis of a decision to eject - quality training.

This has remain with me ever since, and particularly the vital importance of understanding the situation to aid selection of an appropriate course of action, which helps reduce exposure to ‘error’, but hinging on timing – the appreciation of time available, and an understanding of the developing situation ahead of time, and furthermore having an optional course of action – an undo button; knowing what is important in specific situations … requiring knowledge and situation awareness, ad infinitum.

Does modern training provide sufficient practice for situation assessment, consideration of courses of action, and appreciation of ‘time’? Are training syllabuses sufficiently flexible to enable learning from error, vice one attempt / demo, which must succeed, else … no more time, money, etc.

“Time is the reef upon which all our mystic ships are wrecked.” ― Noël Coward, Blithe Spirit

[AirRabbit]

Hi safetypee, et al:

You may be correct with your statement that “…in order to improve or just maintain the industry’s high level of safety we have to find alternative methods to complement the existing initiatives.” However, I still believe that we may be able to “improve or maintain the industry’s high level of safety” by understanding more completely – perhaps only more basically – what it is we expect out of the training we currently accomplish. I fully understand the concerns voiced regarding “cultural issues” being at least a contributing factor, if not more basic, to the successful completion of training – but I think that the basic premise of training has to return to the basic premise of training.

I firmly believe that this premise is, or certainly should be, to train a pilot to the degree that he/she is able to recognize correctly what the airplane is currently doing; make an immediate decision as to whether or not what the airplane is currently doing is what is desired by that pilot; and if the airplane is NOT doing what is desired, be able to make an immediate decision as to what must be done to the airplane’s control systems to correct that condition, and execute that decision quickly and accurately. Lastly, during this execution, continually monitor the condition of the airplane with respect to whether or not the resulting airplane condition is correcting in the right or desired direction and magnitude – again, correctly identify the changing airplane condition, determining whether or not that change is what is desired. This process is a continual process – taking place all the time – and in situations that are rapidly changing, requiring more directly focused attention to the airplane condition by the pilot flying.

Of course – all of the above is contingent – exclusively contingent – upon the pilot flying having accurate knowledge of what is expected of the airplane – at all times – and having accurate knowledge of what control applications are appropriate to make adjustments in the airplane condition at any time under any circumstance – and knowing when, where, how, and how much of those proper control applications have to be made. Additionally, having these knowledge aspects is only part of the solution – as the pilot flying must be able to apply that knowledge, using the controls and indications available to him/her while performing the duties of the “pilot flying.”

From here, to ensure that the pilot flying, does, indeed, have this knowledge and ability, the training program must be designed to provide that pilot with the opportunity to experience various scenarios where such recognition, decision making, decision execution, and continual evaluation of each, is provided initially and then sufficiently repeated, throughout the multiple combinations of conditions and situations, to provide that pilot the recognition and reinforcement of actions and what kind of reactions and responses that will occur. Throughout this exposure, the instructor must take careful notice of when, where, and how the student addresses each scenario, judging whether or not the student’s performance should be modified, and if it should be modified, understand why, and then how, that modification can be accomplished. This process has to be completed for each required scenario, with each appropriate modification, and the results of each such exposure.

Initially, it should not be unusual for the instructor to provide input in the form of suggestions or critiques … and corrections when and where necessary. As the number of times each appropriate scenario is addressed, the instructor should be aware of the level of repetition of previous suggestions, critiques, or corrections – which should provide the instructor information regarding the advancement of the student toward satisfactory completion of the training program – and whether or not that student is ready for the evaluation of his/her skills.

It probably shouldn’t come as a surprise to understand that all of the above should be able to be accomplished with the use of manual controls AND with the use of auto-flight controls as may be appropriate for both the airplane configurations and gross weights, the existing weather conditions, the pilot position involved, and the type of operations expected to be flown by the pilot being trained and checked. Additionally, I should acknowledge that what I’ve described here, is a rather extensive amount of training – and the specifics are very likely to be heavily influenced by the background and the experience level of the pilot beginning the training on any specific airplane type, and may be influenced even more by the pilot position to be flown by the respective pilot. The same is likely true of recurring training and the frequency of such recurrent training.

If an outline for the above type of training is systematically developed for each airplane type, for each pilot position involved – throughout a range of potential experience levels – the satisfactory completion of such training – to the extent that satisfactory performance can be assured on the end-of-course proficiency check – there is a good likelihood of that proficiency being able to be demonstrated regardless of cultural background. However, I say this with the acknowledgement that it is at least somewhat likely that the amount, the sequence, and the duration of training may be different – perhaps significantly different – due to the sometimes wide variation of cultural backgrounds. However, with those training issues being recognized upfront and appropriate training courses developed and followed, the resultant competency/proficiency levels should be essentially met and be able to be used properly and efficiently.

Of course, as with any training program, I think it quite logical, if not imperative, that very careful records be kept – and USED – for each pilot going through each training program. One of the reasons for defining and requiring recurrent training is the upfront acknowledgement that periodic review of learned tasks – both mental and psychomotor – be regularly revisited, the tasks practiced, and, if necessary, an old skill renewed. Of course, the frequency may vary with the level of difficulty of the required tasks, and with the current level of experience of each of the respective pilots involved … which, naturally, should be (will be, no doubt) influenced by the level of complexities of the overall operation in which each pilot has been or will be expected to operate. Additionally, there should never be a task accomplished without the very careful consideration of the safety factors involved.

In that vein, there should be included, in each training course, an acknowledgement that not everyone’s abilities will be guaranteed to be identical – to the extent that some conditions (weather, airplane conditions, urgency – up to an including emergencies) should be able to be attempted by all pilots all the time in every set of combinations of factors. Sometimes, acknowledging that attempting to do something unusual or abnormal may not be within a particular pilot’s ability level – and there should be little or no sigma assigned to anyone who makes such an acknowledgement. Operating within one’s own abilities should always be expected – and the highly infrequent occurrence of such a situation should not be the encouragement for someone to attempt something that they should not attempt – unless there is absolutely no other option.

Finally (yes - you can relax soon....), this is not an acknowledgment that what we are doing with training at the moment is what I've described above. We can shut one eye and "squint" the other, read very fast, and conclude that what we do today is like what I've described ... but the fact is, this would not be true. I think that today we are focused on the accomplishment of tasks - to the end of having accomplished the task. The nuances of performing those tasks, the decisions involved, the progressive analysis of progressing through those tasks, and making any appropriate adjustments - are not necessarily the focus - the task completion is the focus.

Of course, that can be adequate in a lot of situations, but it leaves out a lot of what the individual pilot may be needing - which means that pilot is formulating his/her own understanding of what is happening - and, unfortunately, that may or may not be what is really intended. THAT should be at least one of the targets we logically seek to hit in a properly constructed, administered, and completed training program.

[Zionstrat2]

Children of the magenta line- Wow

Rat5-
Never seen this before- What a clear, direct, and convincing presentation. Thanks so much for sharing and I'll be watching for your future recommendations-

[chrisN]

Air Rabbit, what sort of training scenarios, in your ideal world, would you have wished for the Commander of AF447? And for its two FOs? And for the pilots of Asiana? And for the pilots in the Gulf Air Flight 072 which plunged into the sea (Somatogravic illusion)? And for the captain and FO of Airblue Flight 202 which crashed on 28 July 2010 near Islamabad? And the FO on American Airlines 587 (use of rudder in wake turbulence)?

I could go on – Helios oxygen system, Staines Trident, etc. etc. – all involved pilots not doing it right, but in so many different ways.

Is it practical to train for all these in some generic ways, so that people do the right situational analysis, don’t fixate on the wrong problem, and do take the right corrective actions?

[400drvr]

Children of the Magenta

Rat_5, I saw that a few yrs ago and I think it should be required viewing for all airline recurrent ground schools. But sadly we will waste time on things that make the regulators happy.

[Centaurus]

After an introductory flight, the next flight was with an alternative instructor, with an overbearing reputation, and included a GCA. Shortly after stabilising the descent the instructor closed the throttle without further comment; my thoughts immediately focussed on speed, altitude, high/low key, … perhaps overlooking geographic position in deference to altitude, etc. After what appeared to be an age, the instructor said “what now”; I was lost for an answer

Of course you were lost for an answer and that is not your fault. After all, did this bloke first demonstrate what he perceived you should do before giving you a go at embarrassing yourself while he had a quiet sneer under his oxy mask? I bet he didn't. As you said, the instructor had an overbearing reputation. What you probably didn't know then was he was "a smart-arse". These personalities exist both in the military and civilian flying.

Judging by your later comments on how much you learned a lesson from this character, I would say you were probably a victim of Stockholm Syndrome. It was the fault of your imbecile RAF instructor who was presumably tasked to train - repeat - train, not "test" you. Again from your description of the event he displayed piss poor instructional technique by throwing you in the deep end - sink or swim. I look back at my own student days in the Air Force and can only sympathise with those unfortunate students who were scrubbed for perceived lack of flying ability, when in fact most of the time the instructor was the problem - we called them Screaming Skulls behind their backs of course. Passing an instructors course does not necessarily mean that one automatically is transformed into a good instructor. Far from it in fact. The Commanding Officers of these characters in general never have a clue what was going on under their command since no student in the military would dare to complain.

[tdracer]

Something that occurred to me while driving home last night. The car has a six speed manual, which means a fair amount of the time I only have one hand on the wheel.


To the best of my knowledge, SOP (at least on Boeing) is to keep one hand on the throttles and one on the column during approach. I've been on the flight deck for countless approaches during flight test, and many, many simulator sessions, and that's what every pilot has done.


By accounts, the Asiana PF was having to pull back with as much as 80 lbs. force as his speed decayed. I dare say most people would be hard pressed to apply 80 lbs. to the control column with one hand on the throttles...

[Ian W]

Originally Posted by tdracer

By accounts, the Asiana PF was having to pull back with as much as 80 lbs. force as his speed decayed. I dare say most people would be hard pressed to apply 80 lbs. to the control column with one hand on the throttles...

Why would he have a hand on the throttles- they are automatic and never move in an Airbu........ oh

[DozyWannabe]

Originally Posted by tdracer

By accounts, the Asiana PF was having to pull back with as much as 80 lbs. force as his speed decayed.

Hold up - was that T7 envelope protection in action?

Originally Posted by Ian W

Why would he have a hand on the throttles- they are automatic and never move in an Airbu........ oh

You would (in a FBW Airbus) if you were flying the approach with manual thrust of course...

[AirRabbit]

Originally Posted by chrisN

Air Rabbit, what sort of training scenarios, in your ideal world, would you have wished for the Commander of AF447? And for its two FOs? And for the pilots of Asiana? And for the pilots in the Gulf Air Flight 072 which plunged into the sea (Somatogravic illusion)? And for the captain and FO of Airblue Flight 202 which crashed on 28 July 2010 near Islamabad? And the FO on American Airlines 587 (use of rudder in wake turbulence)?

I could go on – Helios oxygen system, Staines Trident, etc. etc. – all involved pilots not doing it right, but in so many different ways.

Is it practical to train for all these in some generic ways, so that people do the right situational analysis, don’t fixate on the wrong problem, and do take the right corrective actions?

Wow, chrisN – I guess you believe that my comments are either “off the mark,” “irrelevant,” “overly simplistic,” or a combination, perhaps all, of the above. It may surprise you to know that my world is no more “ideal” than is your world. It may also surprise you to hear that I’m not a terribly big believer in “training for” a recently occurred accident. Sure, there is probably nothing wrong with demonstrating what a crew may have encountered, or observed, either approaching or through-out a sequence of events that preceded, or were present during, the on-set of the accident or incident. BUT that is not what you asked. I believe that the training that is provided to pilots should provide them with a complete and accurate understanding of what kind of “control” that pilot has over the condition of his/her airplane (and by “condition” I mean attitude, configuration, and the ability to achieve either more or less airspeed) and what kind of limitations may be inherent with any specific condition with respect to the airplane achieving any other condition.

Obviously, the amount of “control” a pilot may have over any of these “conditions” changes with the specific condition in which the airplane is in currently, and what condition is desired by that pilot. This is the reason for including a somewhat comprehensive listing of “in flight tasks,” tasks with which the pilot is usually familiar. As the task unfolds during the training, differing factors come into play – some (but only some) are orchestrated by the instructor – that may well entice the pilot to make adjustments in either the “plan” they have determined mentally – or may cause an overhaul of that plan completely – maybe even abandon the process for which the flight crew was planning and progressing, in favor of a plan more in line with whatever circumstances have most recently been recognized. All of this provides the pilots in training an opportunity to ‘think through’ the circumstances they face … and good instructors often get a lot more training accomplished through strategic altering of the conditions presented to that flight crew.

And, before you ask, no … there is no “course of training” that exists (to my knowledge), or should be developed, to address “airplane conditions.” But, as the pilot is taught the tasks that make up the training syllabus, the combined instruction on each such task should include the information/facts that allow the accumulation of such knowledge. For example – you referenced somatogyral illusions. As you probably know, but for some who may not, these are illusions involving the semicircular/somatogyral canals of the vestibular system of the ear that occur primarily under conditions of unreliable or unavailable external visual references and result in false sensations of rotation. The results of these kinds of illusions, particularly with new-entry student pilots, include the leans, the graveyard spin and spiral, and the Coriolis illusion. Did anything like this happen in any of the accidents you mentioned? What do the investigations reveal? While I would not absolutely rule-out any such involvement, I think there are much more likely and more identifiable causes for each of these cases.

Here’s a link to more information on this subject: Sensory illusions in aviation - Wikipedia, the free encyclopedia

Actually, most of these kinds of “sensory illusions” have been, certainly should have been, addressed, or at least introduced, during basic flight training – and some even introduced prior to allowing a student to “solo” for the first time.

Additionally, you should also know that there are currently several international “working groups” focusing on just these factors – and there is at least one such working group currently looking at what kind of training should be designed and incorporated for training of the instructors and evaluators. We ARE moving in the right direction.

You ask if it is practical to train for all these situations in some generic way … so that pilots would be able to analyze the situation in which a pilot may find himself/herself, and not identify the wrong problem and thereby take the proper corrective action. Of course, THAT is the ultimate goal of virtually ALL training programs. The best way that I’ve seen to do this very thing, is to employ competently and completely trained instructors to teach a competently and completely constructed training program, using the very best of training aids available. In my book, the most meaningful training aid introduced into the aviation community has been the simulated airplane – initially those devices were … well … as good as they could be AT THAT TIME – however (!) we’ve advanced by several orders of magnitude in the last 50 years! Today, simulation is the best it has ever been … now, we have to be sure that we train our instructors on how to best use that equipment! A part of that training simply has to include a focus on where any specific simulation device is weak with respect to its fidelity to the airplane, in either performance or handling qualities; what tasks must be trained, how they should be trained, how to recognize the difference between competent understanding and indecisive luck, etc., etc.

Additionally, we should recognize that from time to time, we will very likely recognize a necessity to expand the kinds of things that really need to be trained – the following is an example of the most recent change to the requirements in the US, published by the FAA as part of its training regulations for airlines:

§121.423 Pilot: Extended Envelope Training.

(a) Each certificate holder must include in its approved training program, the extended envelope training set forth in this section with respect to each airplane type for each pilot. The extended envelope training required by this section must be performed in a Level C or higher full flight simulator, approved by the Administrator in accordance with §121.407 of this part.

(b) Extended envelope training must include the following maneuvers and procedures:
(1) Manually controlled slow flight;
(2) Manually controlled loss of reliable airspeed;
(3) Manually controlled instrument departure and arrival;
(4) Upset recovery maneuvers; and
(5) Recovery from bounced landing.

(c) Extended envelope training must include instructor-guided hands on experience of recovery from full stall and stick pusher activation, if equipped.

(d) Recurrent training: Within 24 calendar months preceding service as a pilot, each person must satisfactorily complete the extended envelope training described in paragraphs (b)(1) through (4) and (c) of this section. Within 36 calendar months preceding service as a pilot, each person must satisfactorily complete the extended envelope training described in paragraph (b)(5) of this section.

(e) Deviation from use of Level C or higher full flight simulator:
(1) A certificate holder may submit a request to the Administrator for approval of a deviation from the requirements of paragraph (a) of this section to conduct the extended envelope training using an alternative method to meet the learning objectives of this section.
(2) A request for deviation from paragraph (a) of this section must include the following information:
(i) A simulator availability assessment, including hours by specific simulator and location of the simulator, and a simulator shortfall analysis that includes the training that cannot be completed in a Level C or higher full flight simulator; and
(ii) Alternative methods for achieving the learning objectives of this section.
(3) A certificate holder may request an extension of a deviation issued under this section.
(4) Deviations or extensions to deviations will be issued for a period not to exceed 12 months.

(f) Compliance with this section is required no later than March 12, 2019. For the recurrent training required in paragraph (d) of this section, each pilot qualified to serve as second in command or pilot in command in operations under this part on March 12, 2019 must complete the recurrent extended envelope training within 12 calendar months after March 12, 2019.

Below are listed the specific accidents you referenced along with a very brief description of the accident provided by the investigative authority involved:

July 6, 2013, Asiana Airlines Flight 214
The NTSB concluded that the flight crew’s insufficient monitoring of airspeed indications during the approach resulted from expectancy, increased workload, fatigue, and automation reliance and that the Pilot Flying’s faulty mental model of the airplane’s automation logic led to his inadvertent deactivation of automatic airspeed control. In addition, Asiana’s automation policy emphasized the full use of all automation and did not encourage manual flight during line operations. The flight crew’s mismanagement of the airplane’s vertical profile during the initial approach led to a period of increased workload that reduced the pilot monitoring’s awareness of the pilot flying’s actions around the time of the unintended deactivation of automatic airspeed control. Insufficient flight crew monitoring of airspeed indications during the approach likely resulted from expectancy, increased workload, fatigue, and automation reliance. Furthermore, lack of compliance with SOPs and CRM were cited as additional factors.

23 August 2000, Gulf Air Flight 072
The A320 with 143 passengers and crew on board approached the landing at higher speeds than normal and carried out an unusual low altitude orbit in an attempt to correct the approach. The orbit was unsuccessful and a go around was attempted. While carrying out a turning climb, the aircraft entered a descent at 15 degrees nose down. The aircrew did not respond to repeated GPWS warnings and approximately one minute after starting the go-around the aircraft disappeared from radar screens. There were no survivors. There were 36 children on the aircraft. The accident investigation concluded that the primary cause of the crash was pilot error (including spatial disorientation), with a secondary factor being systemic organizational and oversight issues.

28 July 2010 Airblue Flight 202 crashed near Islamabad
This accident was primarily caused by the aircrew who violated all established procedures for a visual approach for RWY-12 and ignored several calls by ATS Controllers and EGPWS system warnings (21 times in 70 seconds) related to approaching rising terrain. The Captain violated the prescribed Circling Approach procedure for RWY-12; by descending below MDA (i.e., 2,300 feet instead of maintaining 2,510 feet), losing visual contact with the airfield, and resorted to flying a non-standard, self-created PBD based approach, transgressing out of protected airspace of maximum of 4.3 NM into Margallas, and finally collided with the hills. The Captain not only clearly violated the prescribed procedures for a circling approach but also did not at all adhere to FCOM procedures of displaying reaction / response to timely and continuous terrain and pull up warnings (21 times in 70 seconds) – despite these very loud, continuous and executive commands, the Captain failed to register the urgency of the situation and did not respond.

On November 12, 2001, American Airlines Flight
The A300-600 took off immediately after a Japan Airlines Boeing 747-400 on the same runway. It flew into the larger jet's wake, an area of turbulent air. The first officer attempted to stabilize the aircraft with alternating aggressive rudder inputs. The strength of the air flowing against the moving rudder stressed the aircraft's vertical stabilizer, and eventually snapped it off entirely, causing the aircraft to lose control and crash. The NTSB concluded that the enormous stress on the rudder was due to the first officer's "unnecessary and excessive" rudder inputs, and not the wake turbulence caused by the 747. The NTSB further stated "if the first officer had stopped making additional inputs, the aircraft would have stabilized."

Obviously, I don’t have a lot of knowledge of what each of the above crews were exposed to, were required to understand, or what the thought processes where that each may (or may not) have discussed with either their flying partner or their instructors. However, I can say that each investigation clearly noted what, in the respective professional opinion of the investigators involved, was “at play” in each of these terrible tragedies. Do I believe that these tragedies could have been avoided through appropriate training earlier in the respective careers of the pilots involved? … Absolutely – Yes. However, I don’t mean being exposed to training on a task that would have mirrored the conditions into which each of the above airplanes was flown. However, training on operating the airplane, including at least some depth into how the airplane was to be flown, could have made a significant difference. The first three examples seem to have a common thread of having made a decision and on 2 of these 3 examples there was an unchallenged deviation from what was logical, and very likely, deviated from what would have been expected, and in all 3 examples, there was an apparent “closed-minded” attitude of acting upon and completing a specific decision – even to the apparent and complete disregard of the on-board warning systems.

The 4th example, the AA587 accident, the pilot flying actually experienced 2 successive encounters with the vortices generated by the preceding B747. The first of these encounters was handled quite satisfactorily, and, in fact, again notwithstanding my aging brain cells, the first recovery from inadvertent wing-tip-vortex encounter was suitable to have been included in a text book, describing “how to do it correctly.” What happened on the 2nd such encounter allows only speculation. He had literally had just demonstrated the correct recovery process, and now he completely abandoned that process. I have speculated that he had not yet “gathered his wits” from the first encounter when his airplane was yet again struck with yet another, very likely a more “robust,” wing-tip-vortex. I think that took this pilot by surprise. I think he was so surprised that he quickly slipped into what I would characterize as a “panicked response.”

Further, I believe that:
1) This panicked response initially resulted in a “pilot induced oscillation.”
2) I think he either didn’t recognize the PIO or did not know how to recover from one.
3) I think he was not aware of the sensitivity of the controls, particularly of the rudder, and how much rudder he was getting with very little force applied, and
4) I think he was unaware of how little pedal deflection generated full surface deflection.

Also, because the certification of any aircraft includes a maximum rudder deflection, held at maximum deflection, and then released to neutral. Also, if, instead of releasing the control to the neutral position, the control is moved to the opposite side at maximum deflection – the loads on the structure go up dramatically. And IF the maximum reversal deflection is achieved while the structure is moving in the opposite direction, those dramatically increased loads are again dramatically increased. And IF, on top of all of this, the reversal is repeated several times, and includes reversals while the structure is moving the other way … well, you get the picture; we are definitely in an uncharted area.

My question here would be, how does a pilot, having the training, background, and experience of this particular pilot, become apparently unaware of these limitations … or … did his panic-induced actions preclude him from recognizing and/or understanding what would help and what would exacerbate the problem?

Apparently, some airlines train to use the rudder. Others train to NOT use the rudder - remember the UA B747 departure out of SFO? That pilot used the rudder, and dependence on the rudder was proper and very likely resulted in the happy ending that ensued. So, which is correct … use it or don’t use it? Airplane designers put a rudder on airplanes for a reason. Clearly, it can be used for other applications, both to the benefit and to the detriment of the safety of flight. Wouldn’t it be nice if ALL pilots were trained on the proper use of all the available flight controls – and how their use can be a detriment to the reason they are in training?

[chrisN]

Air Rabbit, thanks for your comprehensive reply.

Apologies for leading you to think I might have been believing that your comments are either “off the mark,” “irrelevant,” “overly simplistic,” or a combination, perhaps all, of the above. I meant no such thing, and was taking them at face value.

My interest lay in my perception that now many of the old causes of airliner accidents have been largely eradicated by technology, what is left is a much lower incidence of diverse ways of human failing. (By contrast, in my world of gliding, our major fatal and serious accident causes – at least in the UK – are much the same from one year to another, because we keep doing the same things the same way, and rarely see significant improvements in technology or human behaviour. One notable exception was a big reduction in the last few years in launch-failure-related stall/spin accidents, following a big effort in training including post solo recurrent training. But all too often I have seen evidence of the pilot either acting wrongly when faced with a problem he/she recognised, or failing to evaluate correctly a problem and becoming fixated on the wrong solution. It seemed to me that the AF447 FO did that too.)

I wish I knew how to improve the human weak link – in my world and yours. I have a perception that good test pilots coolly, calmly and methodically evaluate situations and mostly come out with a good outcome. I don’t suggest that all pilots are put through test pilot training, and anyway I think they are selected from the best of the best. Many/most professional pilots are selected as the best of the rest. We ordinary mortals at the bottom of the aviation pyramid of abilities are least equipped to do well in unusual circumstances – the evidence is that all too many fail at the expected events, let alone at the unusual.

So, my genuinely intended question was about whether some form of training can help pilots cope better with very unusual, even unique, events.

(By the way, I am no paragon of virtue – I have had accidents, thankfully without injury, by departing from practices I was trained in. I have also experienced an unusual, and not trained for, event – some form of turbulence which led me to think at first that the glider had developed a mechanical fault – which I did not know how best to handle.)

[AirRabbit]

Hi chrisN

Many thanks for your clarification – and I appreciate, very much, how you took my response … and, also, I admit a twinge of “jealousy” in that gliding is something I’ve watched in amazed interest for a quite a large number of years – I actually worked with one of the primer glider instructors in the US for quite a number of years – and have never flown one – I guess because I’m afraid I’d screw up the landing and need to “go around.”

Today, I think that a good share of the accidents/incidents in the airline industry could easily be attributed to some aspect of “technology” – or at least the misunderstood way in which it should be operated, or the way(s) in which it provides the information that it is supposed to provide – and, unfortunately, the culprit that is growing most in its contributions to accidents or incidents is the over dependence on automation to save one’s backside. Personally, I think that kind of “over dependence” on automation played a major role in the accidents of both Asiana Flight 214 and Gulf Air Flight 072.

Of course having ALL the facts of any accident or incident is paramount in making any decisions about the cause or causes. But there are some issues that almost jump off the page into your face. In these 2 accidents, it is quite clear that no one in the cockpit was paying any attention, or certainly not enough attention, to what the automatic systems were doing – or more appropriately NOT doing. What is surprising to me is how long 2, supposedly competent, aviators sat right there, with all the controls available on the airplane and consciously decided to “leave the flying to us.”

I’ve spent a good portion of my career involved in training and simulation issues – and it is clear, at least it’s clear to me, that there was something very definitely lacking in the training provided to the pilots in these 2 cases. I don’t think there was any specific task left out of the training provided to the pilots in question – but certainly there was a significant lacking in what was trained in at least a portion of the tasks that were addressed throughout the training they all received. My guess is that if one could go back and watch a video of the training that these pilots actually received – somewhere along the line an astute instructor could very likely point out when and where the omission was made and how it was made. Unfortunately, there is a good probability that the same omission was made, time after time after time.

And the reason I can say this a bit more forcefully is that the guy sitting in the right seat of the Asiana flight was supposed to have been a former instructor. That tells me that what he was seeing is what he would have expected to see in the simulator or that someone routinely provided a subtle “hint” as to what that pilot should do and when. However, even then, this “former” instructor simply sat there and let the other guy wait on the automatics to “kick in.” At some point, the instructors I know would have had a “not-beyond-this-point” clearly identified, at least to themselves, and had the pilot flying not taken control manually or made the appropriate correction to the automatics, when that point was reached, they each would have taken control of the airplane, and, depending on the point they each decided for themselves, would have manually controlled the airplane either through the landing or initiated and flown a missed approach/go around.

So it’s not necessarily a matter of some specific addition to any specific training program – it’s more of an assurance of what instructors are to teach, how they are to teach it, how they make corrections (when appropriate) and how they determine that each student has reached a point where that student can be legitimately defined as “competent.” It comes in stages, of course – and periodically throughout the remainder of any yet-to-be-accomplished training aspects, a good instructor will either boldly, or surreptitiously, slip in something to verify that the student really has “learned” what he/she was supposed to have learned. It's not "rocket science" - but neither is it something that can be "plugged in" or "unplugged." It is a cultural thing, based on sufficient and adequate training of the instructing staff.

You described a “pyramid” of aviation abilities … of course there certainly is a “pyramid” of sorts, with the most proficient aviators at the top. However, the difference I see is that the pyramid has to have a predetermined base level – and each and every airline pilot (at least for my discussion point) MUST be able to be at or above that base level of abilities or competency - that is a combined level of satisfactory knowledge and personal performance on each and every task that each pilot may potentially be asked to accomplish. Sure, there may be others who are better in some areas – but everyone in that pyramid should be basically competent in each and every task – broken down into each control use, each cross-check accomplished, each trim adjustment, each and every everything, including what passes as acceptable performance – must be able to be demonstrated at that minimum level of competency – or that pilot should do something else for a living.

And the reason I can say that, and say it unapologetically, is that, someday, one of my family members may be on THAT pilot’s airplane.

[FL XXX]

Well said! AirRabbit.

Let's not forget that the ones that really should be held responsible are the Airline CEO's!! These people ultimately make the decision how much manual raw data flying is allowed on the line. They all look at this at a money perspective. Most of them have absolutely no faith in their flight crews. If they would have it absolutely 100% their way, then no "pilot" would be allowed to touch any of the controls on the flight deck. They would do everything automatically, from T.O to Landing. They play a statistical game, eg, complete statistical risk management. They put their faith/MONEY in the reliability statistics of automatic systems! If a pilot is used to regular manual raw data flying regardless of which airplane he/she flies, then it becomes a routine experience and manual flying is not considered an extra workload since it all comes natural. You automatically look at all your instruments in your scan, you feel as a one package with the airplane, it all comes natural. Anything that would be out of the ordinary will be picked up immediately. This Asiana crash and AF 447 are a direct result of Airline CEO's policies! If an airline pilot is no longer capable of manually flying an airplane as a result of Airline automation policies, it is not the fault of the pilot. He/she is restricted to how much manual flying is allowed. A lot of airline pilots incl. myself were used to manual flying and we had a high level of manual flight skills because we didn't sometimes have not even a flight director!! Forget about an autopilot. Manual flight wasn't a workload! It came all natural, because you were used to it. Then all of a sudden you get a job with another airline with strong automation policies and all of a sudden you're been told that raw data flying is a safety risk and 'dangerous'. You think what the heck are these people thinking! But you keep quiet in order to not jeopardize your job, before you know it since you are not been given the opportunity to learn to fly your new airplane your manual flying you had slowly become a thing of the past. You operate your airplane everyday but you actually have never flown it.

Between the Airline CEO's and their MONEY!! and the Airplane manufacturers wanting their MONEY!! They design their new airplanes to the tastes of the Airline CEO's which means as automatically as possible. Does all this automation really contributes to safety as the CEO's all think? Yes, the airplanes have become a lot safer over time but their automatic systems make them to complicated sometimes for human brain. Especially when you combine this with fatigue. KEEP THESE AIRPLANES SIMPLE PLEASE!!! For me a basic mode non-precision approach is easier than a managed/VNAV mode approach. Too many automation options to think about, it takes your brain power away especially when you're tired.

My solution: we have arrived beyond a point of no return, trying to retrain pilots how to manually fly an airplane will require an onslaught on the training resources of airline training departments which most airlines don't have and mostly are unwilling to fund since it is a lot cheaper to play the statitical game. Yes, pilots make mistakes, but no one talks about how many times the pilot has saved the day and has prevented accidents because the automation messed up.

My message to the Airline CEO's: It's mostly your fault if your pilots seriously lack manual flying skills because you restrict them with your policies. Retrain them in raw data flying, as soon as they have shown to have recovered sufficient raw data manual flying skills, give them the freedom to regularly manual fly on the line. Taking fatigue issues into consideration.