FV3 Waiting In The WingsAdded : Thursday at 15:15 A pretty big week in the weather world next week as we say goodbye to the old GFS ► 14 and welcome in a brand new upgraded model which is a total overhaul of the physics. The model is called FV3 at the moment and works on the world being a finite cube (which it is) as opposed to the old style square grids (which it isn't).
Plenty of work for lots of weather sites around the world, many of whom use the GFS ► as a backbone model for their forecasts. We don't, but we do incorporate some of the data into our modelling as well as providing all the charts for the GFS.
The switch happens next Wednesday for the 12z run when the FV3 model will come through as the new GFS. There has plenty of talk in the met community about the cold bias of the FV3, but recent upgrades have made this less of a concern.
We've been following both models now for nearly a year and we're all ready here for the switch, but from a forecasting point of view, is this a better model or not?
If you take a look at the GFS ► forecast for a week from now :-
You can see high pressure to the North and low pressure to the East of the UK. Here is the FV3 forecast for the same time :-
High pressure to the North, but low pressure across Scandinavia. This may seem like a trivial difference, but this synoptic draws in cooler air from the Northeast later next week and into next weekend, which coincidentally the ECMWF ► goes with too :-
The FV3 is getting better. In our view, it's better than the current GFS ► in the short to medium term, but in the longer range it still produces plenty of "noise" and strange charts which we tend to scoff at.
But, the FV3 is looking a little more like the ECMWF ► which is a good thing. The question, is whether improvements and upgrades over the next few years to the FV3 can take it near the accuracy of the ECMWF? That's another question for another month.. or year.
METEOROLOGIST : MARSH |