Provisional CET for 1st-10th was right up against the record high for the period (1898): https://twitter.com/peacockreports/status/1436777893766569988
Actual is probably two or three tenths lower, mind.
Still, this is proving to be a staggeringly warm month. Far more so than the models were indicating as of August's end. Good of GW to apply penalties for entries after 1st. Even on 1st there was arguably some advantage so I wish I'd held on now!
Anyway, a curious feature of this month has been GFS periodically producing a batch of runs featuring extraordinary warm spells. The 06z of today went full late Sep 2011 on us, while the 12z is more laidback but, through sheer persistence of very warm days and mild nights, gives an even higher CET estimate to 27th (dashed line on the 2nd graph in the tweet).
Usually I'd toss them aside as crazed ravings, but here's the thing - GFS spotted the big plume of early this week ahead of all the other models. Not consistently, by any means, but then, it's not being consistent this time around either.
So maybe it's onto something.
There is one key difference, though. Before this week's hot spell, ECM was bringing changeable westerlies. Currently, it has an increasingly cool easterly (from a warm starting position in the middle of next week). So, if GFS' last success was bias-related, it might not enjoy such good fortune this time.
I have a bit of a conflict of mind on the matter. On the one hand, I could do with the CET coming right down to my prediction for obvious reasons. On the other, I enjoy exceptional warmth a great deal when it happens in spring or autumn, especially March-April or September-October. I'm also a big fan of impressive monthly stats. The warmest September for over 100 years, let alone on record, would be amazing to reflect on.