I'd play d3 or 0-0 in this position with white. But I don't see much fault with a4. Sure, d3 and 0-0 are a bit 'faster' and more principled, but that doesn't remove a4 from consideration.
As Larsen pointed out, pushing rook pawns up two squares can be useful. Black has to make up his mind rightaway now, go 0-0-0 or 0-0 manually via Kf7-Kg8 (the a2-g8 diagonal isn't as deadly without the bishops), he can do both here and in both cases white can send some pawns and pieces after the black king. I've known players who'd keep the king in the center scared to death of going either side with the king. That may be a motivation.

If it's a move white made very fast, you could consider it a waiting move. It grabs space obviously and there are some side-effects that wouldn't come into play. For instance, a rook-lift via a3 (or heaven forbid doubling on the a-file if black 0-0-0) the bishop ending up on b2 or a3 (which may be relevant if black commits to c6) after the b-pawn moves up, outposting the Nc3 on b5 or retreating it for some purpose to a2.
On the other hand, white is giving black some information too, he's going to go kingside, staying in the center isn't a real option and going long is just bad.

Bottom line, a4 isn't bad, but d3 and 0-0 are better.