Did the Air France 447 pilots endanger the passengers by not diverting around but flying straight through the weather system encountered and pictured below ? Or was this a defendable, unproblematic choice ?

scope (judging from some answers, I'm obliged to clarify this): I'm NOT asking why AF447 crashed – this BEYOND the scope of this question, which solely asks about the meteorological conditions. What went wrong, went wrong INSIDE the storm cell. On the contrary, my question is about the decision BEFOREHAND whether or not to enter this particular storm cell pictured below. This question is entirely independent from the outcome (crash) and would have merit even if AF447 had not crashed.

meteorological and other background information:

Other flights on the same route chose to divert:

« Three other flights (IB6024 [A343, Rio De Janeiro GIG-Madrid MAD 12 minutes behind AF-447], LH507 [B744, Sao Paulo GRU-Frankfurt FRA] 20 minutes behind AF-447, AF459 [A332, Sao Paulo GRU-Paris CDG 37 minutes behind AF-447]) tracked the same route to TASIL between FL350 and FL370 around the time of the crash. They all had to avoid storm cells and diverted from the airway between 11 and 80nm. They experienced moderate turbulence. All three flights had problems to establish communication with Dakar. » (source)

The french newspaper "Le Figaro" asked why AF447 flew through the thunderstorm when other flights diverted around it:


« I note Figaro's graphic borrows from the BEA interactive where one could see the routings of other planes that night. However, for reasons known only to the BEA, the BEA no longer references LH507 which preceded AF447 along the UN873 airway by 20 minutes, nor (if I recall correctly) does the interactive include IB6024 which followed AF447 on UN873 by 12 minutes. Instead, the BEA interactive mostly plots flights on a parallel airway over a 100 NM distant. Both the LH and IB deviated off the track because of the weather, and one would think their deviations would be the most relevant to what AF447 did not do. » (source)

The weather encountered by AF447 is analyzed in detail here, here and here.

Weather infrared satellite image Meteosat-9: img

Weather Satellite Infrared Image 02:45Z Jun 1st: img

View of AF447 track using GOES imagery: img

img Figure 12. Probable radar depiction (green/yellow/red shading) based on thermal signatures and conceptual MCS models. Units are arbitrary approximations of radar strength ranging from green (weak) to red (strong). This is based on careful enhancement of cloud temperature information. This is just an assumption of course but this is the best guess based on my own experience and the satellite signatures available.

note: I don't understand why the question gets downvoted. I'm neither claiming nor assuming or guessing that the choice to fly straight through the storm cell was problematic. I'm only asking because I honestly don't know. I'm neither a meteorologist nor a pilot, so forgive me for asking, because I am really curious about this aspect. I'll happily accept either conclusion.

  • 1
    $\begingroup$ Was that "obviously" dangerous, or "prohibited" legally or by Air France guidance? No. Was that more dangerous than calm air: Yes. Bypassing it has a price in dollars and in delay. As for everything in aviation, there is a balance to find, sometimes subjectively, specially in the ITCZ $\endgroup$
    – mins
    Commented Dec 30, 2018 at 12:21
  • 1
    $\begingroup$ You are saying there is a choice between $ and safety ? $\endgroup$
    – summerrain
    Commented Dec 30, 2018 at 18:09
  • 5
    $\begingroup$ Of course. Safety improvement has an exponential cost and cannot be total. You need to stop somewhere. I'll take the example of this. There have been deadly accidents with children inserting things in the holes. Now we have that (we must push on the two protections simultaneously to insert something). This has a price, and does decrease the accident rate. However with better, more expensive systems an engineer could decrease the rate further. This is not required by law, and not done, for economical reasons. $\endgroup$
    – mins
    Commented Dec 30, 2018 at 20:04

3 Answers 3


The pilots did discuss the weather ahead and made some corrections to their course to compensate:

  • 25 minutes before the autopilot disconnected they discussed climbing to a higher altitude, but the temperature was too high for them to climb.
  • They turned on the landing lights and noted they were entering clouds and would be encountering some turbulence.
  • They noted the upcoming turbulence to the cabin crew.
  • They experienced light turbulence with a max of about 1.5 g’s, which is borderline moderate.
  • They slowed from 0.83 Mach to 0.80 Mach, which is the recommended turbulence penetration speed.
  • Two minutes before the autopilot disconnect they increased gain on the weather radar and adjusted course 12° left of their planned track.

They knew the weather was coming and made the adjustments that were available to them to get through it. Nothing indicated to them that it would be more problematic than turbulence. This turned out to be correct, since the most they experienced was light to moderate turbulence.

There is really nothing there to indicate that the weather they were flying through was anything that would be unsafe. There were other aircraft that flew through ahead of them, but they got no reports of dangerous weather. I think the images in your question make it appear worse than it really was. They did discuss the weather, but nothing they said on the CVR indicated that they expected it to be unsafe, just bumpy.

It boils down to opinion as to whether they were justified in flying through, but there really is nothing indicating that it was anything more than just a bumpy ride.


The question seems to overlook the facts of the pitot system icing up.

The BEA's final report, released at a news conference on 5 July 2012, concluded that the aircraft crashed after temporary inconsistencies between the airspeed measurements – likely due to the aircraft's pitot tubes being obstructed by ice crystals – caused the autopilot to disconnect, after which the crew reacted incorrectly and ultimately caused the aircraft to enter an aerodynamic stall, from which it did not recover.

It seems to me that you are playing up the weather aspect too much.

Weather conditions in the mid-Atlantic were normal for the time of year, and included a broad band of thunderstorms along the Intertropical Convergence Zone (ITCZ). A meteorological analysis of the area surrounding the flight path showed a mesoscale convective system extending to an altitude of around 50,000 feet (15,000 m) above the Atlantic Ocean before Flight 447 disappeared. During its final hour, Flight 447 encountered areas of light turbulence.

Commercial air transport crews routinely encounter this type of storm in this area. With the aircraft under the control of its automated systems, one of the main tasks occupying the cockpit crew was that of monitoring the progress of the flight through the ITCZ, using the on-board weather radar to avoid areas of significant turbulence.[82] Twelve other flights shared more or less the same route that Flight 447 was using at the time of the accident.

It seems to me that the younger pilots didn't do their job very well, allowing the aircraft nose to climb too high and stalling the airplane.

  • $\begingroup$ re: "you are playing up the weather aspect too much." I'm not "playing up" the weather aspect, I am focusing the question on this aspect, which – I hope you'll agree – is legitimate, right? I'm not playing up anything. You are misinterpreting the question by reading something into it, which just is not there. There is nothing to interpret – I'm just posing a question. I am not implying the answer to be "YES" at all. In fact, I don't know one way or the other. If I did, I wouldn't ask the question. You might want to rethink that personal comment, as it has no relevance for the answer. $\endgroup$
    – summerrain
    Commented Dec 30, 2018 at 19:02
  • $\begingroup$ The main effect of the weather aspect was to ramp up the stress level of the inexperienced flight crew. The actual phenomenon that caused the airspeed fluctuations, thought to be ice crystals in the pitot lines, wasn't related to the weather itself and incidents have happened in "clear" air. What killed the crew was inadequate basic instrument flight skills and poor situational awareness. The right seater was in pure panic mode the entire ride down, and the left seater was too mentally saturated to think to override the right seater's panic mode full aft control inputs. $\endgroup$
    – John K
    Commented Dec 30, 2018 at 19:32
  • $\begingroup$ There's an old saying from the helicopter world... "don't just do something, sit there". It means don't overreact. A well trained IFR pilot, when he/she sees some indication go bonkers when cruising along, will "just sit there" and evaluate. Cross check other displays. If the airspeed jumps but the attitude didn't change and you didn't feel any vertical motions, just leave things alone because you are probably still in level flight. They reacted and got themselves disoriented instead of assessing, and the capt didn't intervene soon enough. I'dve dragged the left seater out of his seat. $\endgroup$
    – John K
    Commented Dec 30, 2018 at 19:43
  • $\begingroup$ @JohnK I’m not defending his flying skills, but he couldn’t “just sit there.” The autopilot disconnected and it started banking. He had to hand fly. Why a qualified pilot was unable to hand fly and maintain straight and level flight for a minute or so while they worked the problem is the crux of the situation. $\endgroup$
    – TomMcW
    Commented Dec 30, 2018 at 21:41
  • $\begingroup$ Good point, but still, if the AP kicks off and the airplane is banking, do the absolute minimum; level the wings, pitch attitude, assess. All 3 airspeeds increasing, but all 3 pitch attitudes show level flight and no noise buildup, no altitude change? They needed to be able to filter that information, but instead they quickly became disoriented. If there is too much conflicting information, you have to ignore it all and drill in on something basic you can trust, like standby attitude. Part of the problem is the type courses are as short as possible and can't possibly cover all scenarios. $\endgroup$
    – John K
    Commented Dec 30, 2018 at 22:02

It depends on what the crew was seeing on their own weather radar. If they were seeing something like what the satellite imagery was depicting, a giant mega-cell, I would say yes, they were foolish to blunder through. But they may have been seeing something more manageable or benign looking, such as individual cells with gaps, on their own radar.

Plus the seats were being occupied by the FO on the right and a fairly inexperienced relief pilot on the left, and their ability to interpret radar returns may have been limited. One important factor here is the fact that attenuation of the signal from heavy rain can create black spaces beyond what appears to be the main area of the storm, which can entice pilots to that area thinking it is clear, when actually the black space is the real center of the cell. Experienced pilots learn to interpret the returns keeping that effect in mind.

447 was a confluence of bad luck, bad decision making and some bad flying, with the key decision maker unable to directly intervene. A pretty rare combination.

  • $\begingroup$ What "bad luck" ? $\endgroup$
    – summerrain
    Commented Dec 30, 2018 at 18:13
  • $\begingroup$ Capt being away when they had to make some key decisions. $\endgroup$
    – John K
    Commented Dec 30, 2018 at 19:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .