weatherwiz

Gotcha, I see where the disconnect here is. When I am talking about data I am also including the growth of the database just due to time. For example, for those who like to look at snowfall breakdowns based on ENSO...this winter will add to the database for La Nina winters. But that's the thing with data...there is alot of it out there and its up to the person to decide which data to use and how they want to use it. But we can't just discount data and tools because there is too much of it. We need to adapt because this is an ever evolving field.

It certainly can and one can but that is where it is up to someone analyzing data to make distinctions. I don't think too much data is a bad thing but it can be bad if data starts becoming misused, misinterpreted, or manipulated. In the sense of weather, alot of just broken down into monthly averages. For example, look at the teleconnections...monthly NAO, monthly AO, monthly PNA, etc. Why is it done this way? I don't know the answer (probably more than one) but I would hedge because it's easier, cleaner, and more organized. But how everything averages up in the monthly...over time that probably leads to a story which isn't an accurate assessment.

Now that we've evolved in the field from a resource perspective with a greater slew of tools, knowledge, and databases I think for research purposes we need to move away from comparing everything to monthly plots and monthly averages. We need to analyze and study these periods of a sub-monthly level (weekly, bi-weekly). I think doing this, we could increase accuracy and confidence substantially. I think we're too accustomed to just throwing in a month or season into composites for quick and easy results. For example, we could take the 5 snowiest (or coldest or warmest) months on record for the month of January and get a quick H5 composite...then we could do it for the 5 least snowiest (or coldest or warmest)...and do the same thing. Now in doing that, we probably would see drastic differences...but that is in the mean. That is not telling us what happened in between and that means more than an averaged pattern, IMO.

Well said, this explains that extremely well. I remember we had these discussions in my Oceanography class but because the polar regions are warming faster than the equatorial region, this is having an impact on the Earth's energy budget. Now, it's important to understand that there is always a surplus of net energy at the equator and a net deficit at the poles - hence the Hadley Cell circulation which transfers energy to the poles and vice versa. However, we're building up more energy in and around the tropics region and I know there are some discussions out there on this and my professor theorized this too but this could be a factor in why we've seen several tropical systems recently explode in certain areas of the tropics.

This is where I disagree because while pattern recognition is extremely important it does have its limits and it doesn't always translate to sensible weather. Similar patterns are not always going to produce similar results. At the end of the day small-scale features, mesoscale factors, storm evolution processes are going to end up having more weight than the overall pattern. But what becomes important within patterns is how the pieces move and how they evolve and those can be independent of the pattern...they're not necessarily driven by the pattern. In this sense, AI will always be behind because it is going to rely heavily on what has happened historically and quite frankly, anyone who forecasts for a living or does so as a general hobby and pays attention to details can easily do this already. And the difference between human and AI here is the human is going to have the ability to start thinking about how some of these smaller-scale evolutions will play out...AI will not. At the end of the day, AI isn't going to tell us anything that we can't figure out ourselves when dealing with the medium and long-range. Can they do it more quickly than we can, maybe but I don't see that as being significant value.

I think AI can and will become extremely useful in the nowcast period which will be critical in severe weather events and when it comes to flash flooding. But I think if anyone has an expectation or hoping that AI is going to eventually lead to perfect forecasts 3-4-5 days out...they are very mistaken and if anyone is selling that idea to make money then that is highway robbery and taking advantage of those who aren't knowledgeable in this field. It would be like someone selling a car to someone who knows little about cars. They can play it up because it looks pretty and say how "great it is" and the person is going to trust the person selling the car because feasibly, they would know more than them.

Watch the outer Cape pull off 4-6'' of snow lol

I do buy a widespread light snow region wide and probably even 2-3'' for some. You still have a potent shortwave coming through and there is plenty of available moisture. H7 looks pretty saturated. I think we're also in the entrance region of ulvl jet streak which should contribute to some good divergence aloft.

Based on the last 4-5 years I would classify anything over 5'' as generational

This is such a beautiful way of putting it.

then puked in Will's car

Yeah those soundings are super tight for ATL. Was even looking that way around DFW though there seems to have been a trend back towards colder.

I feel like its a setup like that which will pan out The ones that get little attention 4 days out and then all of a sudden 2-3 days out you have a nice QPF field showing up. But that's a nice shortwave digging in

I wonder if there is a window Monday night to get some accumulating snow.

They could be in line for a decent freezing rain event but even if it is more sleet versus freezing rain that would be pretty disastrous for them I would think.

Don't they tend to be in a favorable spot for significant icing?

Atlanta shut down for days?

Has nothing to do with the models, the hype is driven by the people who share this stuff around on social media like Halloween Candy. I can't stand these posts, "This isn't a forecast but..." or "This isn't likely to happen but".....

How about some of this instead

Close the thread, close the blinds, see ya in May

that they do but there are instances there value isn't the best. I mean in the case of this event, they aren't telling us anything we don't already know. Sure you can look at it and come to the conclusion that the mean favors less or later phasing and little-to-no hit but even within the ensembles on this, they are yielding just as much spread. they aren't helping to reduce the spread here.

Not even in my dreams do they get hit

I don't even know if ensembles matter much at this point anymore. The only was we're going to get clarity on the southern stream shortwave and whether it is held back or not is going to be sampling in real time or probably not until like with 0z guidance tomorrow night. By then models should have an idea. I get the ENS can at least give you an idea of which way to lean but that is almost irrelevant in this case.

can make our own

This wind today is brutal