As we pause between runs, I wanted to put forward a hypothesis for you to consider (before anyone says anything - I know the models are not some form of AI). 😊
It seems to me that beyond T144, the models seem to scan the horizon for macro changes in the atmosphere, pick up on one, "fiddle" with it (micro changes) and then either develop it or discard it (which the GFS Op run did with the Atlantic push through a few days ago). The more cross model agreement, the more likely this macro change will be adopted. My hypothesis is this. where the macro change is developed, rather than discarded, it also gains confidence in the outcome and gains momentum, ie, the time-frame to that part of the forecast, arriving in reality, shortens. A recent great example of this is the demise of the robust high currently over with which was - until three days ago - due to sit over or close to our shores until after Christmas, perhaps then sinking south into Europe. Now it is forecast to both decline and to shift quite a long way northwards. Over the last few runs, that macro change has gained cross model support and has gained momentum (it is accelerating) BUT ..... we are still outside of T144. I can see that there is room for one more macro change prior to the forecast for Christmas firming up.
One caveat to all this is that sometimes, a macro change can occur as near as T96 - and we all have memories of those!
Just food for thought.
Jeff
On the East/West Sussex Border
70m ASL