The Enforcer

I struggled to believe the totally unsupported initial ECM solution, which was to show it smashing into the UK and taking the roof off my house.

In my experience, it is often works the opposite here. The initial predicted snow threat is just on the wrong side of marginal so we get rain or if it is cold enough just misses. This means that the minma are higher than other areas and so the next predicted snow threat to hit the target ends up being rain as well. This can repeat, to lead to places within 50 miles having almost a foot of lying snow and nothing here.

Ouch that's gotta hurt. Maybe the national grid should start a campaign to 'have a glass of milk instead'?

You'd better hope that this Marge Simpson-shaped warning for Wednesday gets extended westwards, otherwise you might be going "d'oh":

L to R: ECM 00z, ECM 12z, GFS 12z, GFS 18z, GEM 12z all variations of a welcome theme at T+120.

More meddling from the Weather God, yesterday evening: "Can't have frontal snow reaching lowland UK - go on a 'wishbone effect' course." This evening:

Eyebrows being raised at GFS 18z, but it seems to be offering a largely 'average' solution. ECM 12z, on the other hand required an extension to the y-axis:

Only to find that the ECM had already been in there for 6 hours.

Another one I'd been pondering: I have been seeing many references to 850s mixing out as cold air reaches the UK, but why would this occur when the air coming in now from the east is below average, enabling lower temperatures to establish over the UK, then the northerly introduces polar air, which is only threatened by mild air if/when this hurricane turns up?

I was thinking earlier that usually we find ourselves under a standard west-east mobile set-up and then the models start running with the idea that extremely cold air is going to head our way in 10 -14 days time. At any time between the 4-10 day range, reality kicks in and the cold is watered down in terms of potency and longevity at best or completely removed at worst. This time the current and short-term set up is a blocked one, but highly likely leading to nothing really that extreme, however one chooses to measure it: 850s, 2m temps, dew points, wind chill or anticipated frontal/convection snowfall etc etc. Against that backdrop, where I have a reasonable expectation in 10 days time to be able to finally wake up to a frost, instead the models are trying to persuade me that a raging hurricane that is going to destroy my house is the more likely mid-range outcome.

This week or next week?

Polar Vortex: "Help. I've got stuck in Eurasia and the UK is about to get an extended cold spell." Jet: "No can do, mate, I'm heading off to Africa, your only hope is if the Weather God can conjure up a hurricane." Weather God: "Two hurricanes, actually."

There was a better accumulation here from that than the entire March 2018 BFTE saga.

Another method for avoiding the white noise is to have a self-imposed cut-off for chart viewing. I have settled on T+120, because that's the outer limit of FAX charts and certain runs like ICON 6z & 18z and UKV, so it allows for comparison across all available models. There is still plenty of variation in detail between T+120 and T+0 (W-E track of LPs being a prime example), without having to contend with overreactions to temporary pattern reversals, massive outliers and bogus hurricanes etc.

T+120 Fax chart minus hurricane.

Latest GFS T+120s including a distant 'hurricane':

I dread to think what trying to follow the CFSR/IMD/FNO output would be like - completely different run every time. Yeah, I've yet to see a sleetman.

I am starting to believe that I will get my first frost of the 2022/23 season. Oh wait.

But then what is the difference between that and the darker green over Southern England?

Could someone please explain what's going on in the 'red box' below? Cheers, TE.

And here's the ECM T+120 variant.

With T+120s like these from GFS it isn't really necessary to go any further.

Those apps are useless at such range, given how inaccurate they are at T+0: This morning T+0 10am forecast: 12% chance of rain. Outcome = rain. Yesterday evening T+0 7pm forecast 2% chance of rain. Outcome = rain.

Using old data is standard for them.

Ooh I like it Another one from Jonny L - 1:16 for that moment when a run expected to be a cold outlier turns out to be a mild outlier.