Can a state of the art atmospheric general circulation model reproduce recent NAO related variability at the air-sea interface?