February 2026 Medium/ Long Range Discussion: Buckle Up!

[Heisy]

and were back.............................till the dr says no

I put more stock into the GFS Ai vs the OP so this gives me a little hope. Been saying since yesterday this is a pretty simple setup. Could all go to crap, but I feel pretty confident with this.

18z gfs backed up the western ridge, I bet it shows a strong event

[stormtracker]

9 minutes ago, Heisy said:

18z gfs ai, this is almost the perfect mid Atlantic track

Where's the precip map?

[T. August]

Just now, stormtracker said:

Where's the precip map?

it’s not as pretty  

[stormtracker]

3 minutes ago, T. August said:

it’s not as pretty  

Figured

[bncho]

1 minute ago, stormtracker said:

Figured

It's still 5-10".

[stormtracker]

1 minute ago, bncho said:

It's still 5-10".

Wait, what?  And he called that not pretty???

[Terpeast]

Verification scores coming in for ai-gfs and its ensembles, ai is doing better than their dynamic counterparts. But that’s not saying much. I’d say its middle of the pack on par with op euro and geps, with ai euro leading the pack. 

[T. August]

1 minute ago, stormtracker said:

Wait, what?  And he called that not pretty???

It precipitates for like 48 hours with no panel getting over .1” lol. Yeah technically it is indeed half an inch of liquid.

[Buddy1987]

After the GFS with this last storm, I will find it hard to believe for quite a long time. Trash model.

[stormtracker]

1 minute ago, T. August said:

It precipitates for like 48 hours with no panel getting over .1” lol. Yeah technically it is indeed half an inch of liquid.

In this famine, I'd take it

[mitchnick]

1 minute ago, T. August said:

It precipitates for like 48 hours with no panel getting over .1” lol. Yeah technically it is indeed half an inch of liquid.

That seems to be an AI thing. I've been noticing that on a number of both AI runs.

[T. August]

3 minutes ago, mitchnick said:

That seems to be an AI thing. I've been noticing that on a number of both AI runs.

Agreed… too much smoothing. Almost ensemble-esque

[Heisy]

I know verbatim the gfs didn’t work out, but it did trend like the Ai, slowing down the wave and shifting ridge W. If this thing is going to come up the coast that’s a good way to start so I’ll take it as a positive overall.

[Heisy]

Agreed… too much smoothing. Almost ensemble-esque

The Ai models shouldn’t really be used for QPF for this reason. H5? Different story.

[high risk]

15 minutes ago, T. August said:

It precipitates for like 48 hours with no panel getting over .1” lol. Yeah technically it is indeed half an inch of liquid.

       AI models don't do particularly well with QPF.    They get a general shape of the field but struggle to resolve amounts and detail.    I'd focus far more on 500 heights and SLP.

[stormy]

The drought returns for the next 15 days according to the latest GFS.

[CAPE]

51 minutes ago, stormtracker said:

Where's the precip map?

Well, lets just say that isn't even close to perfect for the MA. He gets over excited anytime he sees a piece of digging vorticity, regardless of its orientation or location, or whether its sharp or a lobe that wont do anything at the surface.

[Weather Will]

3 minutes ago, CAPE said:

Well, lets just say that isn't even close to perfect for the MA. He gets over excited anytime he sees a piece of digging vorticity, regardless of its orientation or location, or whether its sharp or a lobe that wont do anything at the surface.

Not being critical but asking, would you want an intense lobe back about 200 miles over KY/ TN moving eastward?

[CAPE]

I don't look at the AI models very often, and I don't think I am missing anything very insightful most of the time. Mostly seems like a desperation thing when the physics based models aren't advertising what we want. They aren't quite there yet.

[CAPE]

1 minute ago, Weather Will said:

Not being critical but asking, would you want an intense lobe back about 200 miles over KY/ TN moving eastward?

Not a broad lobe, but a sharp shortwave that is neutral/ going negative, well sure.

[osfan24]

3 hours ago, Fozz said:

That’s because 2013-14 (especially early on) was very frustrating for the immediate metros while those who live N/W kept getting buried.

Hmmm, I don’t remember it like that and I lived right on 95. I remember it being freezing cold and constant overperforming storms and nonstop tracking. The big storm in February was the only disappointing part because areas NW got almost two feet of snow while I got maybe a foot that compacted and melted to like four inches by the end of the storm.

[Climate175]

1 hour ago, stormtracker said:

Wait, what?  And he called that not pretty???

The expectation is to see 20+ inch snow total maps across a wide swath of the DMV in some minds lol. 

[Weather Will]

WB 18Z AI EURO came North with midweek threat compared to 12Z.

 

 

[NorthArlington101]

I don't look at the AI models very often, and I don't think I am missing anything very insightful most of the time. Mostly seems like a desperation thing when the physics based models aren't advertising what we want. They aren't quite there yet.

Might be overdoing the praise for the AIFS in some cases but its pretty darn compelling, imo. At least equal to the EURO in terms of midrange synoptics.

[NorthArlington101]

AIFS much better. Where is @stormtracker to unpanic?

[TSSN+]

12 minutes ago, Weather Will said:

WB 18Z AI EURO came North with midweek threat compared to 12Z.

 

 

Give me 25-40 miles north and call it a day. 

[mitchnick]

So AI jumps north from 12z with Wednesday threat and the plays the disappearing act with the 4-6" event it had for next weekend.

So much for the notion of AI being consistent. 

[SnowenOutThere]

Remember when everyone was calling this threat dead at 12z. Good times. 

Seriously though we gotta see how our current failure plays out to know what this one will do so we got at least two more days of model swinging. 

[frd]

[MillvilleWx]

1 hour ago, high risk said:

       AI models don't do particularly well with QPF.    They get a general shape of the field but struggle to resolve amounts and detail.    I'd focus far more on 500 heights and SLP.

 

34 minutes ago, mitchnick said:

So AI jumps north from 12z with Wednesday threat and the plays the disappearing act with the 4-6" event it had for next weekend.

So much for the notion of AI being consistent. 

AIFS has done particularly well with verification on the lower end of QPF distribution (<= 0.25") and has a good indication of the coverage within those bounds. It still struggles with higher QPF allotments, especially anything with convective backgrounds. Might get the areal extent of the "where" right, but magnitude is going to struggle. This has been the case for the past few years of evaluations. Totally agree with the SLP and 5H outputs being the areas it does the best with. I have noticed it tends to waver outside the D5 window, which is the case with next weeks system as well, but it will slowly move and hold firm between the D3-5 window before usually settling within the bounds of a solution. Atmospheric complexities are still prevalent within the grand scheme and the AI models will still be subject to variability at leads due to the complex variables at hand. They are a useful tool and one that helps aid in medium and short range forecasting.