Originally Posted by Man Flex
Also very telling is the autothrust disconnect, the realisation that climb thrust has been inadvertantly applied and then the subsequent reduction in thrust.
The PF didn't want to climb otherwise he would have remained in climb thrust.
What is very telling is that none of the facts do fit with your "proposition".
When autothrust disengage, it doesn't "inadvertantly" apply "CLB" thrust (that's an urban legend) as thrust is frozen at its last setting which is independant of thrust levers position (CLB at cruise). Actual thrust levers position is only providing for autothrust its possible range of operation: when placed on CLB, autothrust may adjust N1% between IDLE and CLB level.
When autothrust is disconnected, a warning is sounding and an ECAM will display "ENG THRUST LOCKED", which means that Thrust is frozen. This should be acknowledged by the pilot, stopping the alarm, by moving the levers to their current settings or anywhere wanted.
In AF447 case, it is exactly what happened and thrust was locked at N1% 84, which is low for such FL and weight (about 94% would be needed at Mach .80). In fact, few seconds prior to disconnection, pilots changed their selected Mach from 0.82 to 0.80 (turbulence penetration speed). Consequently, she was decelerating. Next, thrust levers were not unlocked before about 15 seconds after autothrust disconnection, while she first climbed, then still at 84%.
When levers were unlocked, PF manually applied 100% N1, for about another 15 seconds, before reducing thrust when PNF asked him to go down... but she was still climbing. Shortly after that, stall alarms sounded and he applied TOGA while pulling up, and... she stalled. (it is taken from DFDR tracks and report).