weatherwiz

This illustrates beautifully. While using OP at this time range isn't particularly great, I do think there is some value in assessing how the OP is handing the overall evolution of the pattern during that time frame. In doing this, you can see there has been a tendency to somewhat compress the heights a bit and there is also some pretty strong vort maxes modeled...these would further help to flatten that flow out a bit. That look on the ensembles screams some sort of storm potential. I think that period through the first week of January is shaping up to be active.

Yeah there really is no reason to panic or anything yet. It's already been established the surface may not be totally reflective of what is going on in the mid-levels. What we can gather from that period, however, is that it could be active. How those pieces fall into place...way too early to worry about that.

The one thing I think everyone does best though is when there is a legit threat or when an event is ongoing...the discussion is generally pretty great and top notch. the occasional joke posts here and there but everyone puts on their game cap when needed.

Well stated, especially the having fun part. I know there are some who don't want to read the nonsense or care for it but some of the troll stuff when the weather is boring is downright hilarious.

It's basically trying will things to happen which look bad. Trying to find ways in which it will work. 99% of forecast models could show a rain storm at D6 and one model showing potential for all snow...all the focus would be on how that one model "could verify" over what the actual situation is I mean...I'm not complaining about this, that's why we're all here and it's the purpose of the board...it's a discussion board lol. But it does I think yield in blending the line of fantasy versus reality at times.

If things unfolded as advertised, I wouldn't be surprised if we had a decent chance of throwing more moisture our way given the southwesterly flow around the stout high across the southeast. However, I think what we would really want to see is a deamplifying wave like you said...or even a strong Arctic cold front plowing southeast across the Plains which would help tap into the Gulf. What would be excellent if there was a connecting the the equatorial PAC...we would be golden I think there

good thing they probably don't have many large trees in the northern Plains or upper-Midwest (or maybe they do...never been).

I disagree with this. There would be 100 posts just showing snow maps from 50 different models and then comparing changes in the snow maps to justify any trends and then the if game on how if a,b,c,d,e,f do this it has a chance.

Was outside last night want to say it was around 9:00 or so and it was very peaceful just looking around and seeing everything covered in snow with Christmas lights...for some reason, this scenery makes it so you don't even feel the cold. All it takes is a couple inches of snow to totally change the mood.

No disagreement there, but (and maybe I am flat out wrong on this) we can't teach AI something that we don't know ourselves. For example, when it comes to physics and mathematics, AI isn't going to teach us or give us a better understanding of how atmospheric physics works and how these processes behave and evolve. Let's look at thunderstorms, for example. There are certain processes which occur during a thunderstorms life cycle that we know happen, however, we don't fully know why certain processes happen the way they do or what is the leading contributor. Tornadogenesis is one...we know the ingredients needed for tornadoes, we know how tornadoes form, but we don't know fully understand why some supercells (which look tornadic based on visual features/radar features) produce tornadoes and others don't...AI isn't going to solve something like that. AI isn't going to tell us this because we don't understand it ourselves and we don't have all the necessary data and measurements to be able to do so. But with the phone example, its learning faster because it has a basis to go on...it understands something because that something is known. There are many meteorological processes which we know exist and understand their existence, but don't fully know the why/how. AI will be a major help though in calculating mathematical/physics calculations much more quickly which will hopefully get us faster model output in the future.

Kind of explained it in a post above but I'll add more. We should be focusing on the resources needed to better what we currently have instead of just adding more tools to the toolbox. Now, if AI will be used to better the initialization/parameterization process, that would be amazing. But at the end of the day we still need better computing technology (which we have, we just need it within the field). Quantum computing is going to go a sizable way I think here. Parameterize better and improve initialization, those two alone will go a great way in forecast model accuracy and hopefully reduce inconsistency. If AI models are just going to add to the list of potential outcomes and increase uncertainty, then what good is it?

Bingo. That's exactly what I could see happening. People in charge of making critical decisions will take AI at face value and that will set the stage for some disastrous decision making. And I'm not even talking in the sense of general public - this could be anything, logistics, supply chain, resource distribution, etc

As much as I don't like AI, that doesn't mean I don't believe there is value in it, however, I don't think its value outweighs the "bad" and that's where I see the problem. In terms of weather, I think AI could have tremendous value in the nowcasting (<6-12) hour window, particularly when it comes to severe weather and flash flooding potential. In terms though of using AI as like medium/extended range (maybe even short range), I see little value - AI models in this range will not tell us what a great forecaster will not already know. The only way we will ever greatly improve forecasting skill in this range is to better understand how the atmosphere and how it evolves and better understanding physics and processes...then it's taken that and quantifying that numerically so computers can process this information. Computer forecast models struggle as you move away from initialization because of compounding error. This was why for a prolonged period of time the euro was by far and away the best, it had superior initialization skill - the euro would have very few error inside of 72 hours which resulted in greater accuracy through 5 to even 7 days. AI isn't going to solve medium-to-long range error or reduce inconsistencies...AI will not do this until we ourselves are better able to understand this and teach the algorithm to do this.

His question was valid given my open disdain about AI weather models lol

Those float around Facebook too...and there's this huge circulation too of videos of like destructive weather or chain reaction accidents. I guess at least there is wording that says, "This is AI generated and only for your entertainment" but it's stupid and dangerous.