bncho

Why is that?

Well then, FWIW, the Google WeatherNEXT showed a very Euro-AI like solution

Do any of yall have any idea what the Google WeatherNEXT is? Apparently its verification scores are up there with the Euro and Euro AI, but I've never heard about it until joining SouthernWX.

This winter storm will be a pivotal moment in modeling history. It's the AI models (apparently Google's AI, WeatherNEXT, showed an Euro AI solution after being slightly south) vs the physics models. This storm could change our perspective of current modeling—we may crown a new king.

It's on WxBELL.

I'm too young to be looking at porn...

There are two possible outcomes for this event: 1. A major ice storm affects southern CAD zones, and DC and points north get a major snowstorm. (Euro AI, UKMET, GFS AI, Euro) 2. A major snowstorm affects southern CAD zones, and DC and points north receive little to nothing. (GFS, CMC) What do we know of each model? The Euro AI is often locked onto an event inside D5 and is often deadly inside that range, but can be locked onto an incorrect solution, only to fold 48 hours before the event (credit to @high risk). It also often has a QPF bias. The Euro often tends to overamp low pressure systems, but it also tends to hold back energy in the Baja. It also often has a QPF bias. The GFS tends to show more suppressive, progressive solutions, and often tends to verify NW. The CMC has a major cold bias, and tends to overdo Arctic HPs. It also, generally, has a suppressive bias, but it is less prominent than the GFS. The UKMET tends to struggle on thermals, often being too warm. It also tends to overamp systems. Since the GFS AI is new, I don't know much about it. Although more operational models (that I'm aware of; I have no idea what Google's model outputs) support the northern solution, ensembles have been trending colder and snowier towards the southern solution over the last 24-36 hours. This is why I am skeptical to declare victory for the northern solution just yet. It would be easy to go with the Euro AI, just because that has been the best performing operational this winter inside D5. However, a major caveat to this is that the upcoming HP may be so anomalous that it hasn't received enough training data to output the correct solution, just due to its inexperience (credit to @WxUSAF). This should be a good test for the AI Euro. Now, it's too early to declare anything, but I am leaning towards the northern solution. 1050 HPs are extremely rare and often overdone in setups like these (credit @Terpeast), and I don't think it'll be able to suppress it enough. However, if a ultra-strong HP does pan out, the Euro AI may have a greater probability of being wrong. Stay tuned...

FWIW, the Euro AI was almost dead on at D5. It took the Euro longer to get there for yesterday's event. Tonight's runs will be 5 days out. If the Euro AI continues to show similar solutions I see very little reason why not to believe it.

Well IMO he's partially right, the -NAO obviously helps but it's mainly because a 1050 high squashes everything

EPS went crazy!

18z only runs to 144,s orry.

btw the 20:1 ratios are real according to mark mcgarbage

Guys, if it snows, it's going to stick around and become glacier for at the very LEAST a week. Models are literally projecting high temps barely getting into the 20s after Jan 25

20:1 ratios baby

So I may have underestimated the Euro, we actually DO get a foot. Ji is happy now.

0z Suite Scoreboard >4" vs <4" scoreboard (Euro = 5, Euro AI = 5, GFS AI = 4, GFS = 3, CMC = 2, UKMET = 1, ICON = 1) DC: 11 - 10, more than 4" wins Baltimore: 14 - 7, more than 4" wins Stay tuned for 12z! Good night!

It's also snowing in the teens. Wow.

Thanks for the PBP, Randy!

Yeah, based off precip maps DC gets ~8", Baltimore ~6". SE VA is jack zone with 12+". Would take instantly.

Exactly! Just based off those panels it might be 8" for DC, 6" for Baltimore, and boy would I be satisfied with that.

when we're talking about 21 hour storms as "slow" then you know we have potential

Can't tell whether it's better, I wanna say yes?

The Euro AI spits in the GFS's face, incredible consistency

A cave ain't happening today, y'all.

we should put signficance ratings on the models for an accurate scoreboard (e.g. the euro is 5 pts, gfs 3, ai euro 5 pts, etc.)