eduggs

I'll be watching obs from VA late tonight with interest. The NAM/RRFS bring sleet rapidly through S and C VA overnight while the colder models (RGEM/HRRR/RAP) keep it snow. That should give us a good heads up regarding how quickly we are likely to mix or change over on Sunday.

As I mentioned earlier, the NAM doesn't forecast snow. It forecasts QPF and temperature across multiple vertical levels. Other organizations have developed algorithms to translate that data into accumulated snowfall. And in marginal cases where a warm nose is very close to freezing (like Little Rock this morning), it's possible those algorithms incorrectly count snow as sleet. When QPF is significant, this graphical error can add up quickly. In these cases, the fault lies with the algorithms, not the model itself. We would need this morning's sounding from Little Rock airport to try to diagnose any model errors. But based on the NAM forecast sounding from the past 4 runs at Little Rock, I suspect this was the case. Unfortunately the NAM is not showing a marginal sounding for NYC tomorrow afternoon. But NWNJ and the LHV could benefit from a similar clown map error.

Careful about extrapolating a borderline snow sounding in Arkansas to our area. On one hand, we could also benefit from snow and sleet mixed being treated as non-snow on the clown maps. Sometimes rimed flakes aggregate into parachutes that accumulate fairly well (~12:1). Unfortunately the NAM- and RRPS-modeled soundings have a much more pronounced warm layer for our area tomorrow afternoon than for Little Rock this morning. It's several degrees C above freezing in the warmest layer. We need that to be completely wrong. In the Little Rock case the shallow warm layer is near freezing.

That raises a good point about 3rd party vendor snowfall graphics. Weather models don't forecast snowfall explicitly. They generate QPF and temperature at numerous vertical pressure levels at each horizontal grid point. 3rd party algorithms translate that information into a ptype forecast. But sometimes more than one ptype is present at the same time. Snowfall can be under-counted in cases where the warm layer is shallow and not-pronounced because the algorithm may lump it all as sleet when it's actually a mix of ptypes or rimed snowflakes. See the Little Rock NAM sounding below.

The steady precip. should be done by midnight if not sooner, so theoretically there's plenty of time to clear snow and treat roads to open schools and businesses Monday morning. But just based on the hype I expect most things to be closed on Monday.

The lead in forecast evolution for tomorrow's storm reminds me a lot of the December 2020 across our area (differences of course especially SNE and mid-Atl). The big snows for that one slipped away too but the impact was still significant locally.

FWIW I'm pretty sure the 10:1 counts sleet as snow on the Pivotal graphics for the RRFS. In this case the Kuchera algorithm (while not great) is better for visualizing snow accumulations.

It also shows half the QPF as sleet (or rain - SNJ).

Synoptically it tracks with the GFS. So usually too cold. Too bad though because it's a snowy solution... right into Mon. morning.

God that's an ugly sounding for NYC. I wonder if the warm layer being between 700mb and 850mb makes it less likely to show up on 3rd party graphics for other models... maybe other models don't generate and distribute temperature data in between the more common pressure heights.

This reminds me a lot of December 16/17 2020. I don't think the public freakout was as over-the-top for that one... but then again, people didn't have the apple weather app yet I don't think.

We go through this so often. And people cover their eyes and ears beforehand and pretend it didn't happen afterwards. When the NAM shows sleet it sleets. And the extent is usually north of where the 3rd party vendor graphics show it.

What a sweet looking storm for the ALB area. Looking like 24-30 hours of snow up that way. 1"+ liquid, no mixing, great ratios. Those types of long duration, all-snow, big QPF events don't happen very often. Should be a great event for the HV too, but the wraparound shuts off a little quicker than further north.

The RRFS is snowier than the NAM for sure, but the RRFS also shifted northward with the heaviest snow and the thermals. 12z had better ratios and was a mid-HV down to near NYC jack. 18z has it up near ALB with heavy snows to the NY border. Obviously it will shift around somewhat but it's disappointing not to see a clear colder/south trend at 18z so far.

But haven't we heard this dozens of times before marginal events?... but it's almost always correct with mix lines being further north than the consensus of other guidance.

From the clown maps it looks like the City south gets more QPF as sleet than as snow.

The NAM is irritating and somehow predictable. Meaningful QPF lost to sleet and then pronounced dryslot with little lingering wraparound. It's also totally believable. We really need to see a meaningful shift south in some guidance tonight - not just holds - to give us room to think the sleet line won't zoom north Sunday afternoon.

Me too. Was still hoping for clear improvement and more breathing room.

The southern stream wave in TX sharpened up a bit again this run

If you cycle the past 6 runs or so of the NAM you can see the phase getting slightly sloppier over time. The northern stream is little by little shearing off the top of the trof instead of phasing back into the southern stream. It's still a good phase of energy and moisture, just not as extreme as a day or two ago. That trend is really helping keep the primary SLP and the bulk of precipitation south of the NY-Canada border... I'm really hoping it helps just enough to stave off too much sleet and reduce the severity of the dryslot. On the flip side, it might lead to a reduction in total QPF.

The 18z HRRR is actually not that far off from the 12z GFS in terms of northern extent of precip. shield and snow - mix line. It appears to be flattening the height field a little bit more like the GFS although not quite as much. It's definitely south of the 12z RGEM positions.

The HRRR is noticably south of the 12z NAM in terms of mix line from KY to the mid-Atl. It looks about as good as I figured it could at the end of its run.

No noteworthy changes between the 18z and 12z HRRR through 41 hours outside typical noise. 850mb heights look a tiny bit flatter.

No concerns on the 18z HRRR through 29hrs. It's a touch flatter, colder, and slower down south... exactly what I was hoping to see. But it's obviously not impacting us yet. Sure the HRRR is less reliable at range, but its 6-hr trends are sometimes useful.