ANALYSIS: FSF reviews risks of go-around decision-making

London
Source:
This story is sourced from Flight International
Subscribe today »

Failing to abandon a risky approach when necessary can be disastrous, but many times in the past few years, go-arounds – formerly considered a simple manoeuvre – have themselves ended in disaster.

The Go-around Safety Forum, held by the Flight Safety Foundation, Eurocontrol and the European Regional Airlines Association in Brussels this time last year heard that “one in 10 go-around reports record a potentially hazardous go-around outcome, including exceeded aircraft performance limits or fuel endurance”. Go-arounds occur between one and three times every 1,000 flights.

This has led, since the debate began in about 2010, to a dichotomy: is it less dangerous to land from an unstabilised approach than to carry out a go-around? That is a serious question that needs precise answers, and in 2011, the FSF committed to finding them out, because it is clearly not a matter of a simple yes or no.

asset image

Conventional wisdom says pilots should go around if an approach is not stabilised by 500ft - but in most cases, they do not do so, says the FSF

Rex Features

A “stabilised” approach describes one in which the aircraft is on (or very close to) the desired approach path and glideslope, at the correct speed, and in its landing configuration, and all these parameters must be met at or before reaching a designated point on final approach to the runway. The FSF says that 3.5-4% of all approaches are deemed to be unstable.

Conventional wisdom still says that if an approach is not stabilised by 500ft on final approach (some airlines say 1,000ft), crews should go around, but, according to early results from the FSF’s project, they hardly ever do so. Only on 3% of unstable approaches, says the FSF, do pilots elect to go around, so 97% of the time they decide to ignore standard operating procedures and land despite the unstable parameters, and much of the time they do so with no negative consequences.

The reason for the advice to go around from an unstabilised approach is that a poorly executed approach is deemed to be the usual cause of that most common of all airline accidents: the runway excursion. The foundation comments: “It can be argued, therefore, that the almost complete failure to call go-arounds as a preventive mitigation of the risk of continuing to fly approaches that are unstable constitutes the number one cause of runway excursions.” The FSF’s acting chief operating officer William Bozin said earlier this year: “We have to step back and face the reality that a new construct must be considered to drive pilots toward a go-around when it is indeed absolutely required…while [the basic stable-approach criteria] serve as an excellent guide on how to fly an aircraft with precision on approach, a small deviation from the stringent criteria does not necessarily create a risk worthy of triggering a go-around decision.”

The FSF, still in the process of researching data to identify all the risks associated with that land/go around decision, has begun examining the complex psychology of pilot decision-making specifically at that critical point. Its chosen means of understanding the mental processes is to send a series of carefully framed questions to 2,300 professional pilots. Here is an example: “What is the implicit incentive structure for flying go-arounds – versus continuing the unstable approach – that pilots perceive in their organisation’s culture?”

The state of individual pilot situational awareness on final approach is what the FSF is trying to analyse. It explains: “Put very simply, prior to the pilots’ ability to accurately assess the operational landscape for potential threats and risks to aircraft stability, which would then shape their decision-making around compliance, they must first and foremost be fully aware of the objective world around them.”

That sounds fairly obvious, but it is not that simple. The brief time window in which a land/go-around decision has to be made and carried out is not static – the situation is changing continuously – and fast. The potential exists for a pilot to make a decision to land, then change it to a go-around because of a developing awareness of risk, or because of an external effect such as wind shear or a vehicle driving onto the runway without clearance to do so.

The FSF had previously developed a model for assessing how pilots cope with rapidly changing situations – the dynamic situational awareness model. The team will apply it also to the land/go-around decision-making process when it has gathered all the data. The FSF explains how it uses DSAM: “This study employed DSAM for measuring and interpreting the psychological and social factors that collectively make up situational awareness. Within this model, situational awareness comprises nine distinct but interconnected and seamless sub-aspects of awareness.” The FSF wants to understand “how each of these sub-aspects influences a pilot’s risk assessment and ­decision-making processes, singly and in concert with one another, to remain compliant versus non-compliant, in the face of aircraft instabilities while on approach”.

That is a tall order in an increasingly complex world with larger and more complex aircraft, albeit with generally benign, predictable handling characteristics and impressive power to weight ratios. Back in the 1970s, when far less data was gathered and far less analysis was done, military pilots – especially those flying navy fixed-wing aircraft – assumed that every approach was deemed to be a go-around unless, on short finals, all the parameters were seen to be good for a landing. The go-around forum last year made this observation: “Encouraging pilots to be ‘go-around minded’ is essential for operational safety and an analogy may be drawn with encouragement to be ‘go-minded’ after V1 is passed during take-off.”

Something has to be done, but what? The FSF says: “If our go-around policies were effective even 50% of the time, the industry accident rate would be reduced 10-18%. There is no other single decision, or procedure, beyond calling the go-around according to SOPs that could have as significant an effect in reducing our accident rate.” Having said that, the FSF then adds rather despairingly: “Why, then, is compliance so poor?”

Two of the FSF’s questions to pilots are aimed at finding out whether pilots have a different idea from their airline bosses of what constitutes an approach from which a landing attempt would be risky: “Do pilots accept the basic definitions set by their organisations for what defines an unstable condition, as well as the standard operating procedures (SOPs) their organizations have set out to handle them?”; and “Apart from their company’s definitions, beyond what thresholds of instability on key flight parameters do pilots personally define themselves to be in an unstable state that warrants a go-around decision?”

asset image

External effects such as wind shear can influence decision-making in a fast-changing environment

Rex Features

Early analyses from the FSF questions were revealing in that they described the problem in detail without providing an instantly obvious solution. Here are several examples presented at the go-around forum: “Flight crews that continued an unstable approach to a landing compared with crews that decided to go around, scored lower on all nine dynamic situational awareness constructs identified in the DSAM, and demonstrated significantly less discussion about potential threats; pilots who continued approaches unstable, compared with those who go around, are less compliant with checklist use and standard calls. Most pilots do not feel they will be reprimanded for non-compliance with GA; flight crews who continued unstable approaches were more comfortable operating on the margin of the safety envelope, and find little fail-safe in protective crew norms and processes. They will use a convenient, easy justification for non-compliance; most pilots believe their company instability criteria for a go around is unrealistic and their personal thresholds are below 1,000ft for profile and around 500ft for energy management. There is little disincentive for non-compliance, nor incentive for compliance to GA policies.”

Other factors influence the psychological equation. For example there is a long-accepted pilot behavioural malady commonly nicknamed “get-home-itis”, which is normally credited with motivating crews to bust through decision heights on approach in marginal visibility because of their keenness to land. But the same motivation could also encourage them to ignore instability as a risk. Get-home-itis is heightened by fatigue, particularly on the final landing of a duty period, or among pilots operating at a period of circadian low, like the pilots of the UPS Airbus A300 freighter at Birmingham, Alabama (see box).

And another factor – highlighted in the conclusions from the 2013 go-arounds forum – might be this: “Go-around is relatively rare manoeuvre for most commercial pilots. On average, a short-haul pilot may make a go-around once or twice a year and a long-haul pilot may make one every two to three years. This might partially explain pilot reluctance to perform a go-around.” Hence one of the forum’s recommendations was that awareness of all the risks involving the land/go-around decision, and in the performance of go-arounds under different circumstances, should be emphasised not only for pilots but across the board in airline management. Management should have an understanding of the implications of the decision pilots are making before they chart the circumstances under which their pilots are expected to make it. Another precondition for dealing successfully with the problem, says the FSF, is this: “A just culture must prevail if problems in go-around safety are to be sufficiently understood and addressed.”

The relative rarity of real go-arounds, plus the fact that pilots are practically never given recurrent training involving all-engine go-arounds in their simulator sessions, is judged to be a major factor in loss of control during go-around manoeuvres. Because of existing recurrent training regulatory requirements, based on tenets that have not been revised since the 1950s, most go-arounds that crews get in recurrent training take place on an approach with an engine failure. This places its own demands on the pilots, but the climb rate during go-around with an engine failure is quite low, so things happen slowly, which makes things easier if the missed approach pattern is complex. An all-engines go-around delivers a high rate of climb together with a powerful pitch-up moment, so if the missed approach point demands a climbing turn and levelling out at a height not much above the go-around altitude for traffic conflict avoidance reasons, the workload is fierce.

A second FSF experiment conducted using the same survey will investigate the “personally held thresholds for instability that pilots believed would necessitate a go-around call”, given a series of developing scenarios. But the foundation is not stopping at the pilots – it is carrying out a management survey in parallel with the pilot survey and a study of the risks inherent in the go-around itself. Also, air traffic controllers play a big part. They are responsible for providing real-time weather and runway surface contamination updates. It is also ATC’s responsibility to consider pilot workload if they demand changes to the approach pattern or runway in use once the approach has begun. Finally, if it is ATC, rather than the crew, who demands that a go-around should be carried out, it should again consider pilot workload and, if the missed approach procedure pattern is demanding because of terrain or traffic patterns, provide simple, brief instructions. A longer-term objective is to simplify missed approach procedures where possible.

In the end, however, the FSF programme’s recommendations will have three general objectives: to enhance crew dynamic situational awareness in the go-around situation; to refine the go-around policy (by reviewing the stable approach parameters and the stable approach height); and to “minimise the subjectivity of go-around decision-making”. The precise measures that will ensure the delivery of these objectives – and the ultimate objective of safer go-arounds ­– should be published before the end of this year.

And if the go-around measures work as intended, a secondary effect of reducing runway excursions could deliver a dramatic reduction in the overall accident rate.

Recent unstable approaches that should have been go-arounds

According to current industry wisdom, which is now being reviewed for fine tuning, pilots should abandon “unstable” approaches and go around safely. Between 3.5% and 4% of all approaches are estimated to be unstable, yet pilots make a go-around decision in only a tiny fraction of those. What follows are some recent examples of unstable approaches from which a go-around decision ought to have been made but was not, and the results were bad.

5 January 2014 - Vineland Corporation Bombardier CL600 business jet at Aspen County airport, Colorado, USA. It landed long in a high, gusting tailwind, bounced dramatically and rolled inverted, killing two of the three people on board. The landing was the second attempt. The pilot had carried out a go-around from the first.

14 August 2013 - UPS Airbus A300-600F at Birmingham airport, Alabama, USA. The aircraft was carrying out a localiser/distance measuring equipment approach before dawn to a runway with no approach lights. It got too low over rising ground during the approach and hit trees. The aircraft crashed about 1,000m (3,280ft) short of the runway threshold.

6 July 2013 - Asiana Boeing 777 at San Francisco International airport, USA. The crew were carrying out a localiser/DME approach in excellent daylight visibility but were relying on the autothrottle to keep the aircraft’s approach reference speed correct, and were not monitoring the air speed indicator. Because of the flight management mode selected and the fact that the crew had retarded the throttles manually to idle to correct a fast/high situation early on in the approach, the throttles stayed in idle as the approach progressed. The aircraft dipped below the glideslope and, when the crew noticed on short final approach that they needed to go around, the airspeed was just 103kt (191km/h), and the aircraft’s tail hit the sea wall short of the runway threshold, snapping off. The aircraft cartwheeled to a halt and three people were killed.