swilliam

Terrible grammar in the tweet - one reason why I rarely read tweets

It says in the text UKMO was tied 4th (i.e behind theses three).

Very clear trend on De bilt ensembles. Steady cool down this week then a notable warm up for a few days. this is then followed by a steady decrease again - remarkable support for this at this range with only a few members going differently (including control).

ECM mean at 240 looks excellent - large negatively tilted ridge in the mid Atlantic. It is probably best to not keep reacting to every detail of the Op runs and talking about upgrades/downgrades etc. It looks like it will be blocked for the foreseeable future and the details past 4/5 days are currently unknown.

Not a great set of ensembles compared to previous runs London somewhere in Scotland - Op definitely a cold outlier - hopefully not a trend -

It is not winter yet - or only just in the outer limits of the models. Expecting flatlines below -8c is fairly optimistic IMO - probably only happened once in 100 years (2010) and then for 3 days at the end of November and 3 days at the beginning of Dec (edit -and that came from ENE not N or NW like we have in prospect).

Yes looking at those even the mean is better than the Op and all the spread is around the size of the ridge.

High anomaly to the NW, low anomaly to the SW and E - looks pretty good to me for an anomaly chart.

Yes you can see the split @ on the 21st

Fantastic pictures - well done.

????? - It is nothing like it all.

Not sure what people are expecting - I think the bar has been set way to high. This is the set for London In FI mean of -2/3, most below 0, several below -5 and a couple below -10. Mean like this virtually no Atlantic - what are you expecting or hoping for.

Jet not quite so strong on GFS at 198 so maybe won't just power over the top this time 12Z 06Z

It means it is different to the last run? So is the UKMO which has phased the two lows associated with the cold pool in Russia this time giving a different orientation to the block - no idea whether this is good or bad - let's see.

Right in the middle - but perhaps a bit slower 144- 192.

Are you serious?

The mean charts at 180 look very good - -6 uppers on the east coast.

I think you are being swayed by the appearance of blue everywhere. If you look at the 850 anomalies you will see that there is roughly even spread of colder and warmer. It is just that warmer in say Siberia or N China (where I happen to be going to for 10 days) is still very cold in terms of 850s. In fact under the large HP there the 850s will not have much influence and at the surface there is very little difference (typical Max -15, Min -25 in Harbin where I am going).

Well I believe SM thinks differently but I will let him explain that for himself.

He is talking about the big snowstorm on the US east coast that the UK model is predicting but the GFS is not - this is what SM is talking about as it has downstream consequences for us.

Huge spread on the dutch ensemble at D15 as expected ranging from -10C to + 7 (nothing that could be described as mild then). Mean about 2/3 C and a reasonable cold cluster as well so still all to play for IMO.

The ECM 240 verifies about 40% of the time according to the official statistics. Can't remember the number for GFS but less than this. As for 384 probably about 0-1% but have seen any statistics.

Well the ECM ensemble mean is similar but slightly better than this morning. You could not say it directly supports the op with the mean ridge further E than the OP which must be one of the most favourable outcomes. As expected a lot of variance round the S tip of Greenland One of the colder options at D10 at De Bilt but not the coldest by any means - so probably still all to play with the OP being a possible if less likely outcome.

This seems to me to be very easy to prove or not. There are detailed verification statistics for all the models at different timescales. If someone would care to go back through these over the years then if there was any influence of any change in the amount of data over Christmas then this would show up as highly correlated dip in model accuracy over 25/26/27 Dec. I am not going to do this as It is up to those who believe this to be true to show this - not for others to prove that there is no correlation I am sure the world weather organisations have already checked this and would have highlighted it if there was a dip in model accuracy over this period

Yes not an outlier with 4 other members showing this - in fact P12 gives the best solution with a few days of cold. Even with some troughs giving some inland snow - not just the wishbone effect Chances are pretty low for this kind of solution but obviously possible - we can but hope.