Whatever the outlook, Merry Christmas to everyone in this wonderful, mad place. 🎄

Indeed Kieran. Something brewing perhaps? Merry Christmas to you and yours. 

It's odd actually, I thought they would but comparing the spaghetti plots to the 00Z GEFS they are not quite as cold. Still some beauties in there though, especially P13!

Weird - that also suggests the 6Z set was colder than the 00Z set but the graphs looked otherwise!

Anyway the 12Z GEFS also has some fun ones including a -16C northerly on P26 🙂 

https://modeles16.meteociel.fr/modeles/gens/runs/2025122412/gens-26-0-204.png 

https://www.wetterzentrale.de/en/topkarten.php?map=1&model=aifs&var=1&run=12&time=216&lid=OP&h=0&mv=0&tr=24#mapref 

I was just looking at that run and thinking 'reckon it's got that nailed' 😂

There's plenty of cheer in the charts still but not a huge amount of agreement between the ops, or in the ens (which I'm assuming will continue for the 12z sets). 

A little bit OT but interesting to see the MetO 10 day trend earlier, with mention of a relatively unsettled NWly with HP somewhere to the SW of the UK as a reasonable chance option for 2nd Jan, along with a colder Northerly set up. Shows it's fairly balanced at the moment and as has been the case for quite a few days, where our HP ends up will have a big impact on the weather on the ground.

Yes at the same time it is trending colder in NE USA/NE Canada! Typical.

And the ECM 12Z operational goes into the freezer too.

Keeping my feet firmly on the ground as its Christmas Eve and we all know the (possibly) mythical lack of data could be affecting the charts somewhat.

Probably a better set up than GFS shows at the same time FWIW. 

Maybe I'm wrong about this, but I thought the old debate about a lack of data affecting the model runs over Christmas had been more or less debunked a few years ago? 

I think so David, but there are still some people that say it's a thing.

Come what may, nice looking charts tonight (GFS 12z excepting). Although the GFS ensembles look to have trended colder! 

But good to see ECM looking decent.

Merry Christmas to you and yours.

There's a massive drop off in flights, so a corresponding drop off in data from them. That's empirically true. 

Googles usual AI search stuff comes up with this:-

 Yes, a lack of flight data does affect weather forecast model accuracy, particularly for short-to-medium range forecasts in the upper atmosphere over the Northern Hemisphere.

Commercial aircraft sensors are a critical source of real-time wind and temperature data (known as Aircraft Meteorological Data Relay, or AMDAR) at various altitudes, which are ingested by numerical weather prediction (NWP) models to estimate the initial state of the atmosphere.

Key Impacts

Significant Data Source: Aircraft observations are a primary data source for global weather prediction centers, second only to satellite data in their impact on forecasts at some centers like the European Centre for Medium-Range Weather Forecasts (ECMWF).

Specific Altitude Effects: The most significant impacts of missing data are observed in the upper troposphere, around the polar jet stream (10–12 km altitude). Forecasts in these regions can be degraded by up to 15% for wind and temperature when aircraft data is removed.

Geographic Sensitivity: The largest impacts are typically seen in the Northern Hemisphere, where air traffic is denser than over the oceans or remote areas.

Forecast Range: The loss of data primarily affects short-range forecasts (up to 24 hours ahead), but statistically significant degradation can be observed in forecasts up to seven days ahead.

Regional vs. Global: While the impact is measurable in specific data denial experiments, some analyses suggest that in real-world scenarios with a full suite of other observations (like satellites and weather balloons), the overall degradation to global forecast skill may be subtle or difficult to detect with simple comparisons due to the variability of atmospheric conditions. However, regional models that rely heavily on local aircraft data are more significantly impacted.

So all in all I'd say it's not been debunked - I'd guess the lack of data affects certain patterns differently. 

Thanks mate, and the same to you and your family plus all members of TWO. 👍

Personally, I myself would be more suspicious of a lack of data being a factor with the models at the moment had today's output been vastly different to what we have witnessed for much of the time in recent days. We have seen some pretty cold runs from a few of the models at times recently along with a number of milder ones, and what takes place as we head into the New Year is still some way from being settled one way or another. We shall see what happens over the days in between Xmas and  New Year when I imagine there will be a greater number of flights.

Will freely admit it's not something I pay much attention to - I don't know at what point the lack of data kicks in for example. I guess broadly you'd expect the models to be fairly similar just not as accurate as usual unless the lack of data leads to a butterfly effect collapsing the pattern or inflating HP unexpectedly etc. 

I use YR.no or DMI.dk. They only use the ECM data beyond 2 days though, and use the Harmomie for the shorter range stuff. And both websites take an hour or two after each run to update. 

Edit, DMI ECM data for Manchester (I think that is where you are based) but you can put in a more specific location. 

https://www.dmi.dk/lokation/show/GB/2643123/Manchester 

I think this is the issue: the idea being used to debunk unfavourable outcomes, rather than the quality of output regardless. 

If today’s output showed a similar swing toward milder bias, you could guarantee lack of data would have been blamed immediately. 

The sharp drop off in flights only really happens tomorrow and Boxing Day, so today’s output will not be affected. For example, today’s 12z runs were initialised based on data fed in this morning.  Even tonight’s 00z runs should still have close to the normal amount of data.

I would think the layoffs by the Trump administration are as likely to cause a drop off in accuracy.