First, a big thank you to all contributors for this fascinating and most interesting thread.
Anyway, I'm still puzzled by the (probable) crash site being so close to LKP, so I have used an online turn information calculator to sort a couple numbers, and here is how it comes.
Supposing AF447 was at cruise speed at LKP (around 500 kts), coming back close to this point in less than four minutes supposes a constant turn bank angle of 35° at constant speed.
If I reduce the speed to 410 kts, I still come with a constant turn bank angle of 30°. Any smaller turn bank angle doesn't allow reaching LKP again in a 4 mn interval.
Another hypothesis would be that AF447 reduced her speed considerably, which would allow reaching LKP again (after a 360° turn) in 4 mn using smaller bank angle values. However, I can't see any benefit in performing a 360° turn in the middle of the ocean, and any other evolution (e.g. a 180° turn to escape from bad weather) would involve still sharper bank angles.
As far as I can say, these high bank angle values may suggest that, apart from the final stall/spin/whatever event that ended in the crash itself, AF447 first experienced a massive enough "initial upset" that resulted in a change of direction that cannot be achieved using a standard rate turn. In other words, she may have rolled/tumbled heavily because of weather or anything else.
(Obviously, the speculations above suppose that the crash occurred at 02:14, as suggested by ACARS timeline.)
Assuming the scarce info we have is reliable, does anyone buy this theory ?
