If two-thirds of all fatal airline accidents involved a flightcrew-related primary causal factor, there must be a message here about crew training.

That data is taken direct from the UK Civil Aviation Authority's Global Fatal Accident Review 1997-2006, and it encompasses jet, turboprop and business jet operations. There were 283 fatal accidents in the period.

The report notes that the single most frequently identified primary causal factor was an "omission of action/inappropriate action" by the crew. A look at the statistics from another angle makes it clear that an accident's primary factor is rarely the only causal component in any given event, but - whichever way the facts are examined - human factors dominate. For example, the review calculates that in 42% of all fatal airline accidents there was at least one aircraft-related causal factor, but a human causal factor was present in 75% of them.

Training and simulation
 © CTC Aviation Group

Specific data about the most common types of causal human failure in fatal accidents gives a useful pointer to where training is failing, and whether the types of error are changing with time. The CAA review says the four most common causal (but not necessarily primary) factors are "omission of action/inappropriate action" (38% of fatal accidents), flight handling (29%), lack of positional awareness (27%) and poor professional judgement/airmanship (see box P35).


These factors were also the most common in the CAA's previous 10-year global review, but the number of accidents caused by lack of positional awareness - which is the cause of controlled flight into terrain accidents - has fallen, probably because more of the world fleet has been fitted with terrain awareness and warning systems (TAWS) since then.

Accident consequences that are on the increase include loss of control (LOC), frequently associated with flight handling factors. The review cites LOC as a consequence in 110 flights that ended in a fatal accident, 47 of them following technical failure, 47 following non-technical failure, eight following icing and the final eight for unknown reasons. Failure of crew resource management is cited as a causal factor in 20 of the LOC accidents that followed a non-technical failure.

Training inadequacy is actually cited as a circumstantial factor in 11% of the fatal accidents studied, because the investigators found direct evidence of a low quality or incomplete training regime, but pilot performance is the ultimate indicator of training effectiveness even if the accident investigator does not examine the crew's training background.


The fact that, in nearly 32% of the accidents, low visibility or lack of external visual reference prevailed, and that the pilots were clearly unable to deal with it, shows that the crews were poorly prepared for their tasks. Another indication of inadequate training and line discipline standards is that poor crew resource management was a factor in nearly 30% of all fatal accidents.

Clearly, initial, type rating and recurrent training for pilots leaves a lot to be desired in many parts of the world with such a high proportion of serious accidents caused by human factors. One point that emerges from the CAA's study is that the various identified categories of causal human error rank in much the same order in all regions, with human factors as a whole dominant everywhere except Africa. There, human factors is displaced from the top of the causal league table by engine failures followed by handling failures, reflecting the number of poorly maintained Soviet-era aircraft still deployed in many of the poorer states. The fact that the pilots clearly cannot always cope with the aircraft failures they face suggests they were not prepared for their operational environment by their training.

 The most common error types and their consequences

*The data is from the UK Civil Aviation Authority's Global Fatal Accident Review 1997-2006:

  • Omission of action/inappropriate action: fatal accidents caused by this most often involve "crew continuing approach below decision or minimum descent height without visual reference; or failure to carry out a go-around when one would be appropriate; or omitting to set the correct configuration for take-off". Out of 63 fatal accidents in which this error category was judged the primary causal factor, only five involved deliberate non-adherence to procedures, suggesting the remaining 58 accidents resulted from non-intentional omissions or actions, which in turn brings into question pilot training and/or procedural discipline on the line.
  • Flight handling: fatal accidents resulting from flight-handling errors most frequently involve "inadequate speed, pitch attitude and/or directional control, often following engine failure, resulting in the aircraft stalling."
  • Lack of positional awareness: the fatal consequence is controlled flight into terrain.
  • Poor professional judgement/airmanship: the CAA does not provide examples of the most common cases.
Meanwhile, it seems likely that Africa would have just as high a proportion of human factors accidents as the rest of the world if it were not for the fact that so many of its aircraft are so poorly maintained that they statistically trump operational human error as the primary cause.

There are many training organisations that would like to provide more comprehensive pilot training for today's pilots flying today's aeroplanes if money and the regulatory system would allow them to do so. Chris Clarke, chairman of UK-based pilot training organisation the CTC Aviation Group, says his client airlines do not want pilots trained simply to meet regulatory minima, and if a carrier were to approach CTC with that requirement he would suggest they look elsewhere for a training provider. The implication is clearly that training to legal minima is not good enough, and most airlines know that.

Training alone is not sufficient to maintain safe line operations, Clarke says. Clear airline expectations of high standards in line operations are essential as well, he says. If line discipline is slack, he says, even well trained pilots can adopt bad practices.

Despite the advent of an entirely new generation of highly automated aeroplanes from all the manufacturers since the early 1980s, Clarke says the training requirements to meet regulations have hardly changed. Mostly this is acceptable, especially for type rating training for already experienced pilots, he says. But training requirements for inexperienced pilots who are going for a type rating could do with some "updating", he says, as could recurrent training for pilots at all experience levels.

The kind of training flexibility available under an advanced/alternative training and qualification programme (ATQP) is the ideal way of approaching this, Clarke believes, because it potentially addresses the training needs of individual pilots and individual airlines, which are not all the same. Clarke would like to see European regulators taking a more imaginative and positive approach to ATQ programmes.

Along with other trainers, CTC would also like to see a more flexible approach taken by regulators to the use of the latest smart, low-cost training devices for all procedural and technical training, with the use of expensive full-motion flight simulators reserved purely for training and testing handling skills. One of the changes since the 1980s, says Clarke, has been the advance in the training value available from low-cost flight training devices with high-fidelity visual systems, but the opportunity to take maximum advantage of these has been pointlessly limited, he believes.

Meanwhile, the assumed benefits of modern simulation in conferring handling and manipulative skills is being challenged. NASA and the US Department of Transportation studies have suggested that pure handling skills learned in them "do not transfer" to real aeroplanes. In the light of that doubt, Clarke's observation about the different value of simulator training for experienced and inexperienced crews sounds logical.

He says it is all very well for experienced pilots to do zero flight time type ratings in full-flight simulators, but it is not enough for inexperienced pilots doing their type rating. In other words, the ones who can probably fly the new aeroplane already do not need additional handling skills, and those who do need additional skills cannot learn them in a simulator - they have to do base training at the airline as well.

The amount of old fashioned airline base training in real aeroplanes has been cut to almost nothing at most carriers, but for inexperienced pilots there remains a regulatory requirement for half a dozen touch-and-goes in the circuit. This may not be enough to confer the skills necessary for more difficult - but absolutely essential - handling tasks such as crosswind landings and approaches in windshear conditions.


Full-flight simulators are not good at this, says Filip Van Biervliet of Sabena Flight Academy - Development, who has developed a system for changing the logic of existing full-flight simulator motion systems to provide sensory feedback that feels more like the real aeroplane, and makes reactions to pilot control inputs feel "real" rather than generating a series of lurches that encourage pilot-induced oscillation.

As for base training and how attitudes to it have changed as simulation has become more capable, Van Biervliet says that in the past 25 years training is much changed. He directly relates the changes in training to the changes in the type of incidents airlines experience now.

Sabena's base training following type rating used to consist of 9h flying involving 40 touch-and-goes, 11 go-arounds, and seven full-stop landings. Now airlines accept 40min base training involving six touch-and-goes.

Pilot training
 © CTC Aviation Group

Van Biervliet argues that if the industry is going to rely on simulation for inculcating handling skills as well as systems knowledge and crew co-ordination in pilots, simulators need to start imitating real aeroplanes more closely than they do now. The biggest single advance in simulation over the last 25 years, he argues, is the visual system; the rest is much the same as it was.

Aircraft have changed, becoming more automated, one of the results of which is that pilots get less on-the-job practice at manual handling. This is scarcely ideal when base training has reduced so much. Van Biervliet explains his view of the shift in the nature of the airline pilot's job, and in the training system that ostensibly prepares him/her for it. Pilots, he says, are progressively less like pilots in the traditional sense, and have become "aircraft operators".

The emphasis on flying skills has been replaced by an emphasis on the skilled operation of the flight management system - "hard skills have been replaced by soft skills", as Van Biervliet puts it. He defines crew resource management and the disciplined use of standard operating procedures as soft skills, and manual flying skills, the ability to maintain situational awareness with raw data, and airmanship as "hard skills".

Van Biervliet also sees a shift "from understanding to applying". Manual handling, he observes, is always done with the aid of the flight director, and there is no training requirement in managing the flight using raw navigation data. Training, he says, is to "proficiency", whereas it used to be to "fluency". Maybe "fluency" is what CTC's Clarke means by training pilots beyond regulatory compliance, because today's pilot licensing regulations define what tasks pilots must be able to perform to earn a licence, but they do not specify fluency.

Van Biervliet also takes lessons from the UK CAA's Global Fatal Accident Review, extracting two figures from it to make his point: 66% of primary causal factors are aircrew-related, and an additional 14% are flight-handling related.

Knowing that there is no way that airlines would go back to a base training regime similar to that of the 1970s and 1980s, Van Biervliet - a fully qualified airline training captain and an aeronautical engineer - decided to work out what is wrong with the motion systems used with today's full flight simulators, from the point of view of the confusing messages they send to the sensory organs of pilots who "fly" them.


If he could do that, and then apply corrections to the motion system algorithms, he reasoned, maybe simulation could be used effectively to train pilots in handling skills such as crosswind landings - skills that would actually transfer to the aircraft.

Since simulators cannot provide crews with real, sustained accelerations, particularly lateral accelerations combined with roll, the task a simulation engineer has to perform is to find the best possible way of making the pilot's senses and balance mechanisms perceive the motion as being like that of the real aircraft.

With his LM² system, Van Biervliet claims to have done that. Certainly, when Flight International "flew" a Boeing 737 simulator with its motion modified by LM², the assessment was that it made handling the simulator feel like handling the real aircraft.


Those who manage airline pilots face conflicting demands: the marketplace requires the highest personal and team performance, but also the lowest cost. Identifying and setting priorities is difficult. Flight International's Crew Management Conference 2009 addresses the issues that airline operations and training departments face in a constantly changing environment, enabling them to track best practice and review options with peers from all over the world. The event will be in London on 30 November-1 December.

The global training message is that, although airline safety failures in North America and the rest of the world are all dominated by crew-related human shortcomings as the single biggest causal factor, there remains a massive regional difference in accident rates. North America is the safest region - safer than Africa by a factor of 30 according to the UK CAA study.

It seems fair to assume that some of that performance difference is accounted for by differences in training standards, but even so, none of the regions' training systems has been able to drive human factors out of its top place in the accident causal league table.

Source: Flight International