Why such a diffrence in my rating?

http://en.wikipedia.org/wiki/Elo_rating_system

ELO isn't about absolute ability (how would you measure that anyway?), it's about how you measure up against other players you are competing against within the context of which the ELO is calculated (rating pool).

Look at how you compare, in percentile, against other players. Your Online and Live standard time controls are within 6% of each other; the pools of players an and # of games are different across those so it makes sense that the actual rating is different. How you rank with other players, is similar.

In Blitz and bullet you do much worse. My guess is that you blunder a lot more in faster time controls. I know I do better with longer time controls and worse in faster time controls; i.e. quality of play increases with longer time controls. I find that to be true in OTB settings too. I blunder and miss things a lot more in the faster time controls.

Definitely skills in Blitz and long games are quite different. In Blitz, calculation and visualization ablities count for almost nothing, but pattern recognition rules.

So I think it is nothing out of the ordinary that people would have great ratings in Blitz, but low ratings in long games and vice versa. Probably only the really good ones excell in both areas. 

I myself would be best at 40 moves/2 hours, which is just how I learned chess and played it decades ago. The one thing I really love in chess is to calculate variations in my head, even though I encounter problems visualizing positions once I get just a bit tired. I never would use an analysis board in a game and when reading a chess book I deliberately try to use the board for the main lines only (if possible). I totally suck at Blitz (and the tactical trainer with "timer") and even the one "long" game (30 minutes) I played here felt like tremendous time pressure to me.