high risk

While we focus on the medium range and an outside shot at a little snow as the rain ends Thursday, there appears to be a very subtle threat for tomorrow midday and afternoon. While the bulk of the rain will fall Wednesday night when temps are plenty warm, the NAM and NAM nest both show the possibility of a few light showers around Wednesday, and temperatures are only slowly going to climb into the mid and upper 30s. More relevant, the ground is insanely cold, so I could see a situation in which light rain is falling at 36 degrees, and the roads and sidewalks become icy due to the skin temps lagging far behind the air temps. I called it a subtle threat, as temperatures will be rising, and there probably won't be much precip around, but an icy surprise for some folks isn't out of the question. EDIT: the nam nest is somewhat on its own with temps staying very cold during the day tomorrow, but I wouldn't write it off

ugh. Yes, I came out with the right "totals" but typed 0.6 where it should have been 0.4. I've edited the original post. People are right that math shouldn't be done at 11pm....

What'a going on, I think (based off the NAM3), is that you have a huge slug of precip occurring during the evening hours - looks like well over an inch between 00z and 06z when the "snow" really piles up. The soundings are really, really iffy in terms of sleet potential, but the model microphysics are allowing for some pingers to be mixed in with the rain. Let's say that we have 1.3" of liquid falling and the model thinks that it's 60% rain and 40% sleet pellets: we know that would not accumulate, but the tallying does 1.3" x 0.4 = 0.52" liquid of "snow (+ sleet) accumulation". TT slaps a 10:1 ratio on that and gets 5.2" of "snow", while the Ferrier method will assign a way lower SLR since it knows that the frozen component is sleet and not snow.

a friendly reminder that when the NAM accumulated snow depth and Ferrier (from the nest) maps show much less than the 10:1 "snow" accumulation map, then the model is not predicting the snowfest that you think it is.....

GFS warms most of DC metro into the upper 30s Saturday afternoon in advance of the precipitation, but we have to remember that the GFS is notoriously terrible at maintaining low-level cold air. I don't think we can discount the NAM idea of the cold air being tough to scour out, especially if we continue to have snow cover.

I realize that the 12z parent NAM got everyone's attention, but it's a common rule that when the NAM parent and nest differ significantly, and the nest is line with other guidance, it shouldn't take anyone more than a half second to toss the parent NAM solution.

interesting how the NAM nest (vs the parent NAM) shut off the precipitation Saturday night for a few hours in southern MD and the EZF area. This, combined with some mixing, puts that area in a minimum for snow totals, while other guidance maximizes in that corridor.

Wrong. The 00/06/12/18z cycles of the operational HRRR are run out to 36 hours.

The GFS medium range scores did drop for sure right after the shutdown started, and a data issue related to the shutdown was suspected. But no such issue has been found.

There are no data issues with the American models. The GFS medium range scores did slip during the end of December / early part of January, but they seem to be recovering.

The differences in the 850 winds Saturday afternoon between the NAM and RGEM explain the differences in the precip shield. Remarkable disagreement.

hi-res guidance has a gusty line of showers on the cold front this evening. Mid-level lapse rates above a low-level inversion will be really impressive.

It was supposed to replace the GFS in early February, but the shutdown is going to ruin that plan.

MD issued - basically no chance of a watch, but we do have some "moderately interesting" convection approaching....

as we enjoy a Friday in the 60s, it looks like a repeat next Friday.

So, no MRGL yet, although based on SPC's Day 1 outlook, they almost pulled the trigger on it. Still think that there is a window for some interesting cells, especially west and northwest of DC during the mid and late afternoon hours, although the forcing mechanism is admittedly unclear. Most guidance also has a forced line of heavy showers along the cold front during the late evening, but the wind fields will have largely relaxed by then.

more for damaging wind, as wind profiles are strong but somewhat unidirectional. Still, there is perhaps enough low-level backing of winds closer to the sfc flow to make it so that an isolated tornado cannot be ruled out. Instability is limited (which prevents a more robust threat), but low-level lapse rates will be good. Forecasts like these intrigue me.....

Seems like a pretty good shot that northern VA, eastern WV, western MD, and south-central PA will get a MRGL risk for severe Friday, with an outside chance of an eventual SLGT.

The isothermal profile surprises me, as the models have shown a pretty significant low-level inversion. That could certainly up the threat of stronger overnight surface gusts. 11pm: 60 at DCA, 57 at BWI, and 55 at IAD.

Definitely think we have some thunder potential overnight, but I haven't seen any forecast sounding with surface-based instability. So, while there will be strong winds just above the surface, it seems like the chances of them mixing down are low. I'm still more intrigued by the late afternoon/early evening time period Friday when we'll have modest but surface-based instability.

NAM nest shows a dramatic warm frontal passage, with low 40s for DC metro at 8pm and near 60 by midnight.

There definitely does appear to be a window of opportunity west and northwest of DC Friday afternoon. Instability is limited, but there is some cape, and lapse rates will steepen as the colder temps aloft arrive. The shear is outstanding, and there will be forcing along the arriving front.

Thunder definitely seems to be on the table for most of the area Thursday night and for points east on Friday. Good elevated instability early Friday:

The mageval.ncep.noaa.gov is the site for parallel versions of models. If you click on GFS, you get the parallel GFS, which is the FV3GFS. If you don't believe me, check out the 12z 'GFS' on mageval and compare it to the TT FV3GFS plots.

go ahead and check for yourself. It's clearly listed as the 18z run on 12/4. It clearly ran, and that's the NCEP site which can plot it directly from the files on the supercomputer. There is clearly some issue with the other sites that have to download the data.