OceanStWx

My reaction always seems to be WTF. Shift of tails, ensemble sensitivity, etc. It all sounds made up. But once I did some training on it, they're really powerful tools. There's so much more than the mean. The mean may be a good forecast starting point, but some of these other tools can tell you whether you can honk or head home.

I try but honestly I even have to review the literature every once and a while to remember what things mean. All I know is the days of analyzing a single model run in the extended are long gone (or should be). Just a ton of data to be gleaned from the ensembles now.

That's why this is done with the ensembles more so than the deterministic runs. Because you get a stronger wave through many runs and on average the mean shifts east. But the deterministic could have one weird moose fart that leads to a poor solution locally.

I would have to compare the 12z GEFS forecast to the 00z obs, but that would be my guess as well.

More or less. It just allows them to remain separate longer. Now we are actually getting some sounding data from the open ocean, so it's not a black box this time around. They were ingested into the 00z GFS.

Flatten everything and it probably would head east. But I don't know if a weaker wave coupled with a really strong northern stream digging would help any.

We're looking at the wave in the northeast Pacific. That's the feature that digs into the Southeast. If that can trend stronger (or is stronger than modeled) ensembles favor a more benchmark low pressure. Probably due to the fact that the northern stream wouldn't dominate it as much, phasing would be weaker. But that's what we need to keep it from cutting.

That being said the shortwave will be approaching the coast tomorrow morning, and a stronger wave could start to dominate the pattern. 12z GEFS had a ton of spread in a really nice area for New England snow.

They provide good information. It may not be the information we want to hear on the other hand.

I can't say I'm really surprised given the GFS has been underplaying western heights all day. That's one of the ensemble signals for the cutter scenario vs tracking closer to the benchmark.

Nice shift on the 12.00z EPS. Now looking like a big chunk of the forum (sorry AEMATT) with EPS members forecasting extreme snowfall relative to the model climate. The shift of tails hasn't really moved a lot. So this to me says that there is growing consensus for a decent snowfall for the interior, but the really big snow forecasts remain with the farther west members.

12z EPS showing some interesting things when it comes to the EFI. No shading indicates that less than half of the members forecast an extreme snowfall for the 5 weeks centered on this date. BUT the contours (especially above 1) indicate that the members that do forecast the most snowfall (i.e. 90th percentile of EPS member forecast) those members are fairly extreme.

I noted that in my AFD this morning. Ensembles show spread in two ways, in time and in position. Usually you’ll see those 24 hour QPF lines for every member with a random scattering of events throughout time. However the ensembles right now are all showing the same time window for precip. The ones that don’t are because they are out to sea, not because they don’t think they’ll be a storm. It’s a pretty strong signal, especially this far out.

All depends on the situation. On a radiational cooling night you'll almost always be warmer than the valley, but on a cold advection night like tonight the hills will cool off first.

I'd give Friday some hope through maybe 12z Wednesday when the storm north of Hawaii starts to hit the West Coast. And the 06z GFS shows you how to get it done for Monday. MAUL in the DGZ with max lift in that layer.

I do like the ensemble signal better for the latter half of the weekend (especially the EPS).

There are humans digging through the hundreds of reports and updating that map as fast as they can. They actually take hours to put together if there are a lot of reports. Since plenty of this fell after CoCoRaHS report times, they have to wait to tomorrow anyway for an event total map.

4.1” final here. Sad panda. Got into some good deformation snow for a bit this morning, but it was transient. I figured 4-6” for PWM because of the shape of the coastline on this one, but still hurts to see the waste water treatment plant pushing a foot.

These events, kind of like December 2020, are hard because you can’t broad-brush 20:1 all over the place, but it’s also hard to go much more than 24 hours lead time on that kind of band placement. I personally think too many local NWS offices try and capture the max totals (mine included) rather than it be a “most correct” forecast. But every once and a while one of these sneaks through where the band really does maximize what was available.

I still think the models do a good job of highlighting the banding and max areas if you know where to look for it. Both weenies and Mets get sucked in by pretty QPF and clown maps. In the days when models were coarse, you would have to paint a general area for heavier snow and caveat “locally higher” but now the resolution tries to pin it to your backyard. So when it gets taken away the next run people get upset. That didn’t happen as much on an 80 km grid. If you find the relevant features aloft, QPF be damned.

I think late 2023 or 2024

Learn to love it, because it’s going FV3 core and running to 60 hours and you’re going to lose the NAM and HRRR.

It’s unstable based on the moist adiabatic lapse rate (6.5 C/km). So if the forecast sounding lapse rate is steeper than that it qualifies.

Way more fun than a 96 hour lock.

It’s pretty self explanatory, just a truly unstable layer, usually thin but in winter events you don’t need much. The lift is enhanced in that unstable zone, and so if it’s collocated with the DGZ can really crank snowfall rates.