Yarmy

The high-res ICON-EU 15z (yes, there is such a thing) is advertising an organised band of snow to head SW from Lincs/EA through the Midlands on Sat Night/Sun Mon (this is not from the main band of frontal snow on the low). So a surprise for some on Sunday morning?

Of course, they all are, and they won't pick up any detail of the showers coming off the North sea. But it doesn't change the fact that by pulling the low southeastwards, the models are reducing the area of the UK over which the northern flank snow extends on Sunday.

Yes, plenty of interest there and then for the Midlands. Undoubtedly there will be showers of the North Sea too, but it's a long way from the runs of just a couple of days ago.

It's a poor run all round, as Sunday is rapidly becoming a non-event for the vast majority of the country.

Yes, as the low track keeps trending ever more southeastwards from run to run.

1mm precipitation = 1cm snow (approximately)

That's Ryan Maue's weather business, isn't it? I'm not sure whether he's applying the Kuchera method to the raw ECM output, or that's what the model itself is predicting? 14.4 inches seems like...a lot.

Not with sub -10hPa temps and sub-zero dew points. Wet bulb temp, Theta-e, etc are all conducive to snow as well in coastal areas (eastern coastal areas, anyway). I understand everyone's caution though. Been burned so many times, and this one is so close.

Blimey, if these numbers are even close, Norfolk will come to a standstill. So pretty much normal.

Wash streamer at T93 on the ICON: It will change of course, but interesting to watch these features pop up.

Yes, see: WWW.METEOCIEL.FR Meteociel propose le modèle deterministe de ECMWF CEP ...and the model thread at the time:

Hard to say, but looking at the ECM06z T72, it looks like the low to the East of Scotland does track SE thereafter at least. The important point is that there are big differences within 72 hours between the ECM 0z and ECM 6z, so anyone hoping for closure after the 12z runs is deluding themselves.

Why bin the GFS for the Para when the former is verifying better in the 5-6 day range currently? These stats jump around all the time. A week or so ago, the Para was verifying better (slightly). There's no objective reason to have any more confidence in it than the old one.

With a slider dessert for the southwest.

Such fine margins though. Southern Germany had balmy +12C 850hPa temps on the 0z. On the 12z, they get -16C. Different timeframes, but still...

Where are you getting the ECM 6z data from? I thought it was hidden behind paywalls.

Actually improved since I last checked. But it's still a case of the ECM then everything else: Day 6 Day 8 I'm not sure any of this really matters in the current situation though because, as you say, it's on a knife edge. Slight variations in the extent of the med ridge have big implications for the cold advection into the UK.

You've got it the wrong way round. UKMO isn't using Meteociel, Meteociel is downloading and publishing UKMO data.

All the major weather centres (UK Met Office, ECMWF, NCEP, etc) publish their model output in a standardised format, usually either GRIB or NetCDF. Some of this data is free, some of it costs money (a *lot* of money). Websites like Meteociel, Wetterzentrale, and Netweather download that data and then render it into charts for plebs like us to view. There are many, many other websites that do the same, but charge for it via subscriptions. Meteociel and Wetterzentrale are popular here because they are free to use, and it's possible to compare model output from the various weather centres in a single chart format.

I wouldn't advise going for a swim in it. It's only a couple of degrees above average. Any Easterly flow over that would be a snow machine for southern Scandinavia.

I've been watching the para stats for a while. These are my observations, so far: 1. It is consistently beating the old GFS up to day 6. 2. It is matching (and sometimes beating) the UKMO and GEM up to day 6. 3. It falls off a cliff after that. As you point out, it's dreadful past day 8. 4. It's still miles behind the ECM. As can be seen, it currently second to the ECM at T144.

I've been looking for some public verification stats on the GFSP (the next version of the GFS set to go live in Feb 2021), but they are non-existent as far as I can tell. Ryan Maue, however, was underwhelmed back in the summer: