still think its mixed here for us even on colder euro; the BL is getting warmer and warmer. wasn't the mean like 40+F at 2m?
-
Hello, please take a minute to check out our awesome content, contributed by the wonderful members of our community. We hope you'll add your own thoughts and opinions by making a free account!
Jan. 17-18, 2026 SE Winter Weather Threat
- Thread starter RBR71
- Start date
Watching the 12z roll in now on stormvista. @bouncycorn has greater capabilities than me for this. I am not a model whiz.
Something I will note about whatever it shows at 12z (or any z) is that it is a pretty coarse, 64 member ensemble. It is good, but I am not going to be trusting its 2m temps verbatim. I don't trust its snow maps either necessarily, i just don't know if it can resolve a borderline dynamic cooling scenario like this. Maybe it can, we'll see
Awesome, thank you!
Yeah not looking for specifics but a better look at the footprint which it and the Euro AI seem to be similar. Will be a great test of how it handles 2m temps in this scenario and if it's off, since it's AI, one would think it would "learn" and resolve that going forward.
65% to 20% in one run
A remarkably similar northwestern jump to the AIFS from the Weathernext at 12z. New on top
I should note that the 850s do progress eastward with the ensuing frames. Just trying to show the QPF jump, I know the thermos are bleh in general
I should note that the 850s do progress eastward with the ensuing frames. Just trying to show the QPF jump, I know the thermos are bleh in general
rburrel2
Member
The AI models are great at 5h, but I haven't seen any data saying they are the best in the world at predicting overrunning/WAA precip totals. Just keep that in mind.
Both can be true... AI models nailed 5h... and we have more precip than they've been showing from overrunning.
Both can be true... AI models nailed 5h... and we have more precip than they've been showing from overrunning.
Webberweather53
Meteorologist
Starting to look more amped like the GFS from yesterday tho not as cold obviously
That's the right take here I think. And they can still shift further inland with the rest of them. The ECMWF is the stickler right now
The AI models are great at precip. In fact, the purple line on this error chart "FGN" is the Weathernext2 model. Beats IFS HRES and IFS-EPS.
(lower is better)
Mahomeless
Member
Still absolutely love where we sit in east central AL.....the 50-75 mile ticks NW on the Euro will continue......this is exactly where I want to be at this point.
A nice increase on the snow map for the 12z WxNext. Now explicitly showing a 1-2" swath along a similar path to several EPS member
It’s not Sunday evening yet, still time for this to go anywhere!
iwantsouthernsnow123
Member
Seeing trends that we seen in 2017.. Models that were more amped trend less amped and models less amped trends more amped. FFC nws makes it very clear this system will not be figured out until LAST minute
rburrel2
Member
light overrunning precip at mid-lattidudes? Or all precip in the Northern Hemisphere?
We will find out for this setup I guess. Though I think we've already sort of found out with the big NW jump they just had in qpf. (undetermined if there's more adjusting left to do).
Always something to keep in mind with verification scores, too (which tend to be at h5). Something can be the best at h5 and not necessarily be the top dog translating that to the surface (and vice versa). Of course, all else being equal, being right in the mid / upper atmosphere should translate to being right at the surface, but all else is not necessarily equal.
Also January 28-29, 2014.
ha, dec 17 and jan 14. meet in the middle, perhaps?
The big jump NW by the two major AI models is interesting. I think trend wise those are the models to pay attention to imo. They are not infallible but the other models seem to be a cycle or 2 behind in terms of trends. If these trends keep up in next couple cycles it’s possible we get something closer to amped gfs solution maybe less extreme though. Columbia and CLT squeak out a couple inches of snow maybe? I’m in Hickory area so need big NW trend back this way. Obviously more NW equals warm nose issues pretty quickly.
Sent from my iPhone using Tapatalk
Webberweather53
Meteorologist
Well predicting precipitation in general across the board is not the same as predicting precipitation in specific waa/overrunning setups like this one. These broad brushed model skill analyses do not tell us anything about how they handle specific situations as these skill scores and biases are likely state dependent to some degree.
My hunch is if you re-ran these models for the overrunning events we’ve had since the mid-20th century, they would still undervalue the waa aloft at this range
Thus far, I am not impressed with how the AI models are handling this setup. They’re really just correcting to the more amped GFS/GEFS solutions from yesterday which a lot of people scoffed at because they were “bad” models.
These suites thus far have been struggling just as much as most other models, tho I think they are an improvement over the traditional euro which is stubbornly (& likely wrongly) too far south & east.
rburrel2
Member
15z SREF did trend towards the model conensus on precip... but still super bullish and NW with the precip shield.
rburrel2
Member
15z SREF snow mean. Gives a good indication of where the best accumulations could occur, imo.
@bouncycorn this is already like a hundred mile+ jump from the previous WeatherNext solution... earlier you were saying that this kind of jump just doesn't happen within 80 hours on WeatherNext/AIFS -- is this event maybe the kind of outlier scenario that these models can struggle with and this is just as surprising to you? Or does this fit with your thinking still?
Just seems totally not what I expected based on your earlier comments.
WEATHERBOYROY
Member
Anyone got the current local of the energy for Saturday/Sunday?
anc.py
Member
In all fairness, you do have to give credit that they did stick to their guns for as long as they did, and even their precipitation banding moved only by perhaps 25 miles to the NW.
With all of that said though, just because models are superior track-record wise does not mean it will perform to the best. ALL must be considered unless it's a blatant throwaway model run. The GFS (and definitely its ensemble counterpart) definitely had a DA or overcorrection issue at 00z which is what led to the outlier amped run last night, as all other runs remain similar. Just got to keep in mind what is good towards forecastability and the bad.
I like seeing those totals well north into north Alabama as well. Encouraging
iwantsouthernsnow123
Member
Yeah I definitely agree there. On top of this, EPS has been shifting north as well so I don't think it's too much longer until Euro OP folds.
anc.py
Member
Being in Macon, I figure we'll be alright if the ticks still continue. Temps shouldnt be the worst in the world; better than the coastal situation at the moment.
The AI model suites also appear to have a very smoothed out temperature gradient which doesn't exactly indicate it's handling something smaller-scale well.
Precip contours should start moving north
.1 becomes .2
.2 becomes .4
.4 becomes .8
Act like yall have been here before
.1 becomes .2
.2 becomes .4
.4 becomes .8
Act like yall have been here before
NCHighCountryWX
Member
- Joined
- Dec 28, 2016
- Messages
- 611
- Reaction score
- 1,682
WPC official guidance updated
anc.py
Member
I believe it's due to winter weather threats overall relying on many other factors that aren't modeled or represented/measured nearly as well. Hidden small-scale air masses may tip the scales in some people's (dis)favor while severe weather often requires just "few" (relatively) things.
First WPC map. 
accu35
Member
- Joined
- Jan 5, 2017
- Messages
- 8,922
- Reaction score
- 11,400
As of now, this is a Deep South storm, but with this NW trend happening and especially now with the Euro then I wouldn’t rule out snow in areas further north. Let’s wait and see.
Bigedd09
Member