January 2026 Short/Medium Range Thread

[Wurbus]

Anyone know how accurate the Euro AI model has been in the long range? It's been showing something on the 27-28th for a few runs in row now.

[Carvers Gap]

The 12z GEM and Euro have lots of frozen precip.  The west-east gradient appears to be setting up somewhere between I-40 and southern Kentucky.

[John1122]

I'm going to simply hope the Euro Ai, which has been really great, is correct. The Euro would probably cause me to check out for winter if it happened.

[Matthew70]

7 minutes ago, Carvers Gap said:

The 12z GEM and Euro have lots of frozen precip.  The west-east gradient appears to be setting up somewhere between I-40 and southern Kentucky.

Hopefully it will setup on the TN/AL line.  

[Weatheriscool]

22 minutes ago, Carvers Gap said:

The 12z GEM and Euro have lots of frozen precip.  The west-east gradient appears to be setting up somewhere between I-40 and southern Kentucky.

Dont play

[Wurbus]

There is some agreement with the GEFS and Euro AI on the 27-28th storm

[nrgjeff]

Wow lots of heavy ice forecasts coming out from NWP but thankfully all well out in the extended. We can't have that crap!

They are on different days on different models; some north of here, some right on us. If that happens it'll be a huge slap in the face. Other options range from heavy rain to cold and dry. It's beyond Day 7. Of course we hope for all snow.

[Holston_River_Rambler]

40 minutes ago, Wurbus said:

Anyone know how accurate the Euro AI model has been in the long range? It's been showing something on the 27-28th for a few runs in row now.

For the general idea of the patter even 200 hours out the Euro AI has been pretty darn good:

 

There is a little bit of bouncing around with the northern stream for a few runs, but damn...

I have noticed it has had a little bit of a progressive bias (note the NW trend on this weekend) from 120 hours out:

 

But for basic idea of the pattern, it is looking sharp right now. 

 

 

[TellicoWx]

For those wondering about how Ai is being incorporated by the Euro and GFS, they are being done two separate ways.

Both are hybrids of the base system. Where they completely differ on trains of thought is...Euro is used for post processing and corrects the main models known bias's to increase accuracy and precision. The GFS is being implemented to increase the main models speed and lower computing power to save money moving forward. 

So that's why you hear people such as Spann talking more and more about the Euro Ai and not so much about the GFS Ai. 

As far as accuracy, what little that has been released publicly is so far the Euro Ai is miles ahead of everything, with only Googles DeepMind teams pure Ai system keeping pace.

[Holston_River_Rambler]

May have to issue a virga warning soon:

[TellicoWx]

Quick breakdown of the two

[John1122]

The Tennessee basketball game was the sports equivalent to that Euro run. 

[Carvers Gap]

The Euro AIFS EPS is about as beefy as one could get an ensemble for this area.

[Carvers Gap]

47 minutes ago, nrgjeff said:

Wow lots of heavy ice forecasts coming out from NWP but thankfully all well out in the extended. We can't have that crap!

They are on different days on different models; some north of here, some right on us. If that happens it'll be a huge slap in the face. Other options range from heavy rain to cold and dry. It's beyond Day 7. Of course we hope for all snow.

Yeah, man.  No ice over-running.  I had enough power outage stuff during Helene.  The worst!

But probably a good reason to follow ensembles at this point.  EURO AI EPS and GEFS at 12z are pretty beefy.  I have to imagine individuals within each run are showing some big solutions.  I haven't looked yet.  

[Carvers Gap]

I am probably not as down on the 12z Euro.  I am really just looking to see where the gradient is going to set up.  I think the risk has always been that it would jog north.  However, there will likely be an ebb and flow to this.  Anything can change that gradient at this range.  Again, I tend to think S KY is in a prime spot for the first wave of this.  Just now digging through the 12z stuff after watching the basketball game.

[jaxjagman]

I'd be very careful if you are using the GFS into next weekend of what might happen,this is its last 8 runs,just saying

[John1122]

11 minutes ago, jaxjagman said:

I'd be very careful if you are using the GFS into next weekend of what might happen,this is its last 8 runs,just saying

I just discount it no matter what it shows this season.  I have to assume budget cuts have affected data collection and it's affected American modeling. They've been bad this winter so far. 

[jaxjagman]

57 minutes ago, TellicoWx said:

Quick breakdown of the two

Putting the skill of the GFS is gonna make it any better?

[jaxjagman]

4 minutes ago, John1122 said:

I just discount it no matter what it shows this season.  I have to assume budget cuts have affected data collection and it's affected American modeling. They've been bad this winter so far. 

Oh i agree,the GFS has been trash

[TellicoWx]

9 minutes ago, jaxjagman said:

Putting the skill of the GFS is gonna make it any better?

One is trying to cut cost...less computing power = save $. Other is trying to actually improve the output.

If you run the 500 pattern for last 3 days (starting 10 days out to where it is now), you can see the difference between the two. GFS (and even GEFs) bounces more so to speak. As @Holston_River_Ramblersaid, it's uncanny how little the thing bounces around vs everything else.