The divergence between GFS and ECM takes place around t+138 (GFS) / t+144 ECM.
ECM has the UK high migrate northwards to join a 'true' high over Greenland, with a Siberian high then ridging to join up as well, creating a strong block to our north
At 138, GFS also moves the UK high northwards, but also begins to erode the warm uppers over Greenland from the west, and by 156, they're gone (just a nominal surface high remains). As such, the high is 'unanchored' and gets drifted eastwards, before sinking.
That's my very layman's take, anyway.
Originally Posted by: Saint Snow