The Final Report into the Flydubai 981 crash raises an interesting dilemma. This relates to the first go around attempt which was aborted following a wind shear alarm. Procedure dictated the pilots' actions, but in his judgment a landing was the better option. In the light of the tragic events that followed, it is worth considering - what if he was right ?
The short answer is probably 'NO'. With the benefit of hindsight, initiating a go around was most likely the correct decision. Support for this conclusion is given in the actions of an Aeroflot fight also seeking to land at Rostov that went around three times before selecting to divert, and in the need to abort Flydubai 981's second attempt. Nevertheless, the question is not so easily dismissed, because no matter the 'correctness' or otherwise of this decision, there were consequences that followed it, and these had a significant influence over the final outcome. It is only if we broaden our perspective beyond the immediate context that led to the go around, and include what happened next, right up to the crash, that we can arrive at a fuller appreciation of the issues at stake.
This shift in our focus is entirely legitimate, because the flight ended in disaster. One catastrophic outcome was successfully avoided during the first attempt at landing, but only to end up with another two hours later. This is not a satisfactory result, to say the least, and it means that we can not take the earlier decision simply at face value. If we want a different outcome should such a situation arise again, and the circumstances at Rostov on Don that night were by no means unique, then we have to look for opportunities to do something different. Furthermore, there is no reason why our search should be limited to the final seconds of the flight when everything spiralled out of control. If such a position could have been prevented from ever arising, then the end result would have been much more favourable.
Evaluating the decision to go around is complicated by the fact that the wind shear warning which triggered the pilot's action was PREDICTIVE (MAK Final Report, p. 113), in other words, NO wind shear had actually occurred. At the time the alarm went off, the pilot felt he had good control of the aircraft and he could see the runway. Even so, the Operations Manual was clear, a go around was prescribed, which the pilot duly carried out by performing the Wind shear Escape Manoeuver (WEM).
A first consequence of this decision was that during the climb out, the pitch angle increased sharply and the PIC responded by pushing the control column forward, later adjusting the trim consistent with this action. The significance of this lies in the experience it generated, which set up the pilot's expectations for the second go around when this occurred later on. Also worth mentioning is that while both pilots had conducted standard go arounds before, the PIC no less than 6 times, none of these had involved a WEM. The pilot came out of this first experience with a sense that his main consideration was to avoid excessive nose pitch, and that he would be both pushing the control forward and adjusting the trim to achieve this effect. These expectations did not serve him well during the second go around, contributing heavily to his loss of control over the aircraft.
A second event confused the picture somewhat. This was an overspeed of the aircraft in response to the pilot's inputs, and as it levelled off on approaching the desired altitude. Following the go around, the PIC conducted something of a debrief with the F/O as to how it had gone, and it was this aspect that was highlighted in the conversation. This meant that the PIC went into his second approach to Rostov with a heightened sensitivity to his airspeed, which was why he aborted the attempt immediately after going through another overspeed, even though on this occasion no wind shear alarm had gone off (p. 145).
The PIC's decision making, in other words, was degraded as a result of his experience of the first go around. While he had no choice but to abort when the wind shear alarm went off, this was by no means his only option on the second approach. Under other circumstances this may well have been the sensible course of action to take, however, in this case we have to factor in the aftermath of the first aborted attempt at landing.
The fact was that the PIC had been 'thrown' mentally by this turn of events, to the extent that it was on his mind constantly for the rest of the flight. The pilot referred to it in conversation no less than three times, including with cabin crew (p. 131), and made clear his personal view that a landing had been achievable. In the HCD Team Processes Framework, this factor appears under 'Decision Making', specifically the capacity of team members to 'let go' a decision that has been taken, either by themselves, another crew member, or as in this case by a technical system, and 'move on' mentally, even though they may not agree with that decision and argued against it. The idea of the training is to equip other team members with the ability to pick up on what is going on and resolve it in time.
If fatigue did play a role in the accident, then most likely it is in relation to this aspect of the situation. An inability to 'move on' is precisely one of the symptoms of fatigue, and at its worst it sets up a state of resentment, a 'resistance to the moment', whenever difficulties are encountered further down the track. It means that an individual responds to setbacks by withdrawing mentally from the task at hand and instead dwells on why they are in that position, how unfair it is, how it should have gone a different way. None of this is helpful.
When we examine the crucial passages in the Final Report that discuss the PIC's 'mental incapacity' at the critical point when the second go around was initiated, it is this picture that emerges. The words of the Report are worth quoting at length,
"Most probably, psychologically, the PIC had never got over the impossibility to perform landing at the destination aerodrome and accepted the need to divert to an alternate aerodrome. Particularly given that after the first go-around he had been sure in mind that he would have managed landing. The confidence in question at certain time kept haunting his mind and he spoke it out even in conversation with the cabin attendant."
"Landing at the destination aerodrome had been the predominant goal for the PIC. Just to achieve the goal two hours were spent in the holding area. The PIC had been certain that he would manage landing under actual circumstances, if the windshear warning did not activate. Most probably, it was the activation of the warning in question that should have been that internal trigger for the PIC, which would have altered the action plan – to perform go-around instead of landing. Instead of the activation of the windshear warning the speed leap caused the decision to go around which under the conditions of turbulence and gusty wind had not been as much clear sign from the PIC’s point of view comparing to the warning activation (taking into account the PIC confidence in his capability to perform landing)."
"As the result, probably, the PIC had been stuck in “a clinch” of two opposite goals (motives): to proceed approach for landing or initiate go-around. Even though the PIC, in compliance with SOP, took the immediate decision to perform go-around, “the clinch” resulted in the disruption of the previous flight mental mode (the approach with landing), whereas the new one (the go-around) had not been formed yet." (p. 157)
The Report states, accurately, that in both cases the decision to initiate go around was consistent with the OM and SOP. In each, however, it is possible to question whether the course of action (COA) laid out in procedure was in fact the best one, or whether proper use of judgment would have resulted in a different outcome. There is a certain irony here, in that it is during the second approach that the case for a judgment call overriding the SOP is the stronger one, when compared with the first attempt at landing, but it was precisely because of the after-effects that flowed from the first failure to land that the PIC's ability to deploy such judgment was significantly impaired.
The problem is a tricky one, there is no easy solution. In part this is because we are dealing with unknowns - if the PIC had persisted with his first, or his second, attempt, what would have been the result ? Even if it did lead to a safe landing, this does not really answer our question, because at this point we are faced with the complication of 'hindsight bias'. Poor decisions can lead to favourable outcomes, in fact statistically speaking, in flying operations, they usually do, due to the redundancy that exists in safety systems, just as good decisions can lead to disaster. In evaluating decisions we have to adopt a forward looking perspective, examining risk, rather than a backward looking one, focusing on how things turned out.
The argument in favour of judgment over procedure flows from the limitations contained in the latter. SOPs are also a form of judgment, they represent a consideration ahead of time of what, under most or normal circumstances, is the correct course of action to take. Procedures are above all a mode of temporality, they PRE-DETERMINE a COA. The design and programming of technical systems reflects the same method, which is as much human as is their opposite, situation based thinking, it is their temporal nature that sets them apart. Situational reasoning is based on the UNIQUE properties of the moment, in all its fullness, whereas a technological approach, of which a reliance on procedures is a part, works by rules that apply in each and every case. On 99 times out of a 100, or even 999 out of a 1,000, these will serve us well, a given series of inputs will generate a given series of outputs, which is what we want - consistency, predictability, reliability. The problem is the remaining 1, those 'non-standard' (p. 149) situations where simply following what is usually the correct procedure, leads to the exact opposite result from what is intended.
Given the potentially catastrophic consequences of error in an aviation context, it is still essential that pilots have the authorisation, and are trained, to step in and override both technical systems and SOPs, taking the full situation into account, including all of the human factors and organisational considerations that inevitably play their part in determining the outcome under conditions such as those that applied at Rostov. This means we must at least entertain the possibility that the PIC in this case COULD have been justified in ignoring the wind shear alert on the first approach, or the over-speed on the second.
Many will be uncomfortable with this idea, it leaves us in a world of ambiguity. This, however, is simply reality as it really is, there are ALWAYS situations that can arise when the exact opposite of what is normally the right thing to do is precisely what is called for. This is why we retain humans in the cockpit, and it is why there will always be an element of uncertainty as to whether we are making a good decision or a bad one, even AFTER the event, when the outcome is known. This is just a fact of life, and seeking to deny or escape it through an over-reliance on technology or procedures is not helpful, as it is their limitations that dictate the need for human input, one that is capable of weighing up ALL aspects of a situation and making a judgment call.
Avoiding Catastrophe's training in 'High Consequence Decision Making' and 'Team Processes' is designed to develop judgment and the ability to navigate successfully through complex situations. Go here to enrol in one of our courses, or ask for more information from editor@avoidingcatastrophe.com