Joshua

Love how GFS 12z is the cold outlier. About time we had this:

I wasn't going to put it like that, but let's just say: there's plenty of room for improvement? ?

Worth looking at this fantastically in-depth weather forecast from Gavin Partridge released this morning. If the GFS fails I will put it in the shredder, stick the pieces back together one by one and then claim I have a more accurate GFS model than before. It's standards continue to slip* away in winter forecasting terms. *I think wash is a better word to put there, no ice with a GFS slice! Also nice to see nearly 300 signed in browsers earlier on, a couple of hours later, just 27! Browser numbers just as hard to predict as the weather past a week ahead, or even less.

One comment only, in two halves. When, as now, it's GFS vs the European models it's often the GFS which is right. However, there's one notable exception to this general rule. An easterly set up. As now. /—————————————————————/ Actually I have to disagree with the first statement. More often than not, European models tend to accurately forecast something before the GFS understands the pattern and joins in. GFS did very well (In My Opinion) in Autumn but has slid down the cliff edge since mid December, even when it’s been consistent. IMO, an accurate model predicting something beyond a week should explore with the possible scenarios before firming up on the general pattern. Even the Met Office is cautiously preparing to announce a brutal cold setup, I think the GFS will need reforming after this! ?

Unfortunately, due to a data transmission issue we're currently not receiving the data for the Met-Office global model. We're hoping the data will be available again soon. It seems they are currently offline. Never mind - they only show 72, 96 and 120 hours ahead as far as I can see.