You should be on the lookout for the bishop sac. All the signs are there:
e5
kick knight
Ng5/Qh5 playable
Bc1 is ready
If you know all that, you should be fairly careful about castling with black. The h4 move is a nice fortification of the plan. It is not such a big step with the basic knowledge you ought to have.
With that, I'm saying a decent club player could well have seen it.
its the classical bishop sacrifice. i think a lot of people would know
The problem with the first position is that Black is too PASSIVE. I would probably play ...d5, in an attempt to break open the centre and gain some space.
It's a typical attacking pattern (look for 'greek gift'). Without the Be7, the immediate e5 and Bxh7+ should work. With the Be7, you need the extra help of h4 (so that the defence Bxg5 is not available)
Of course, knowing some old games is always handy 
Also look at Serper's always very educative columns for this and much more !
This tactical pattern is likely know to all players above 1900 uscf. Under this I am unsure of though. I believe if given enough time, an 1800 would be able to recognize it.
Its not that hard to see if you're constanly assuming your oppenent is going to sacrifice material, however we usually dont. Generally when Im thinking through specific lines I assume that if the line loses material for my opponent he or she won't play it (unless im in a situation where I'm vulnerable to check and forced moves are a big threat). Its easy to read this thread and say you would have saw that but the average player wouldn't have seen that line coming unless it was used against you before or you just happened to have studied this specific strategy - I know I wouldn't have seen it.
This is really a pattern : Black castled, pawn on e5 - N ready to jump to g5 and queen to h5 - Bxh7+
If you don't know it, then it's very difficult to find. But any experienced player will have met this sacrifice at least once in his chess life (I remember losing to it when I started chess, and never forgot it afterwards).
It's even a famous pattern : there is a whole book devoted to this sacrifice !
After black castles, I'd go ahead with something other than castling king side for white. After some moves to clear the center blacks bishp is positioned for a late attack. A half open half close attakc at blacks castled King would demnd some pawn moves kingside, cutting down blacks defensive position. Either way it has to be a slow positional game, faining your intensions to manipulate the openings you need to take advantage of the early castle posting.
I know the classical bishop sacrifice and looking now I perhaps should have done better with the situation. I suppose I was on opening autopilot - develop the last bishop, castle etc. Still, I don't think many at about my rating and perhaps a little higher, would have genuinely played h4 after Black castled - say in a 30 min game. Maybe e5 immediately, but h4 I'm not so sure.
I know the classical bishop sacrifice and looking now I perhaps should have done better with the situation. I suppose I was on opening autopilot - develop the last bishop, castle etc. Still, I don't think many at about my rating and perhaps a little higher, would have genuinely played h4 after Black castled - say in a 30 min game. Maybe e5 immediately, but h4 I'm not so sure.
This is a second pattern : the classical bishop sacrifice doesn't work with a Be7 - h4 with a rook behind it is the way to improve the attack 
I agree this 2nd pattern h4/Ng5 is not as well-known as the Bxh7 idea, if only because it's more uncommon (people castle as white too...)
Yeah, I don't think this is a rating thing, this is if you've seen / remember the pattern or not.
Seems to me the applicable chess proverb for black is 'Castle if you will or if you must, but not because you can.'
As hicetnunc said it's all about patterns. You could also look at this as a pattern of central control. White has a classic center and black is attempting a hyper-modern sort of defence. A lot of books tell you to control the center but they don't tell you why or what to do with it when you have it. What do you do when you have control of the center? Well if black castles the very first thing you look at is kicking the Knight out and make the h6 pawn a target. It doesn't matter if you can see h4 coming or not. You simply need to know that, as black, you can't afford to let white have the center. Of course now that you know why, it will be a whole lot easier to find the right move. Don't get hung up on the h4 move. It's the center first.
Here is a position taken from Better Chess for Average Players by Tim Harding (p. 162). It is a chapter about thought ordering and candidate moves in particular. Here is the text and position:
Castling is so often the correct thing to do, at an early stage of the game, that it can become a habit with players to castle at the earliest opportunity and without deep thought. Yes this, like other reflex moves, will now and then be wrong. Should black castle? If not, what should he play?
I spent no longer than 10 minutes looking at this and some of the main candidate moves I considered (after black castled) moves like e5, Bg5, Ne5 and Be3. After running through a few continuations in a real game I would have met 0-0 with Be3, developing the bishop to a solid position and then would look to castling and shifting the rooks before thinking of a potential pawn push or something else. There is no way I would have seen the following continuation:
I was wondering after seeing this what sort of level of player would see that continuation in a real game? It seemed hard enough for me when I knew there was a better continuation than just a developing move like Be3. What do you think?