Comment Several management failures contributed (Score 1) 43
1. The pitot tubes were known to ice up under extreme conditions. Airbus decided they ought to be replaced but didn't require this only suggested it, very much downplaying the urgency.
2. Air France seeing that this was not a mandatory change decided not to replace the pitot tubes until the next major service of each plane. Several other airlines thought this could be a serious safety issue and replaced them as soon as possible.
3. There were many flights from South America to Europe that night. A severe storm over the mid-Atlantic had been forecast. All planes except one decided to fly around it. AF447 was the only flight to try to fly over it. When flying very high over a major storm the plane is likey to encounter rain at very low temperatures - hence the risk with a dodgy pitot tube. It is known that Air France management was at the time very keen that pilots choose the cheapest route whenever possible: did they put pressure on this captain to not divert and therefore use extra fuel? I don't know as nobody seems to have enquired into this.
4. I note that the captain also chose to take a nap just when the plane, flown by two not very experienced pilots, was about to enter nnd try to fly over this major storm,
Then there is the question of the design of the autopilot system. These and the pitot tubes are duplicated but redundancy only helps if there are random failures. Here the failure was a design fault, and not surprisingly both pitot tubes failed in the same way a few seconds apart. But then the autopilot dropped out because the pitot tubes provide the air-speed measurement. There are obviously other ways of estimating air speed as GPS and radar will give you the ground speed and the velocity of the winds is known approximately from met forecasts. It seems to me a design fault in autopilots that they simply give up if they don't have all their sensors working. The assumption is, I suppose, that a competent human pilot is always there to take over. A faulty assumption in this case.