high risk

Friendly reminder that any "total snowfall" maps will accumulate sleet as snow.

Exactly. The initial lighter stuff will be either light rain or light white rain, but the heavier reflectivities closer to the front will be brief bursts of heavy snow.

Yes, inside of 36 hours, I'll take the NAM vertical profiles in a warm advection event over any model. But at longer ranges, its synoptic errors (and yeah, people dump on the NAM, but regional models really shouldn't be expected to be great beyond Day 2) can dwarf its legitimate skill with sorting out inversions.

The best place for verification is the EMC Verification site: https://www.emc.ncep.noaa.gov/users/verification You can find metrics for many different parameters for short-range, medium-range, hi-res, globals.... whatever floats 'yer boat. I wouldn't go chasing QPF stats, as there is no current attempt to parse events by type (i.e. winter storm precip vs. convective). If you looked at recent Eastern U.S. stats, they would be based off of winter storms mixed in with heavy thunderstorm events like we had the other evening. And any QPF error can be do with synoptic errors or errors in the microphysical and convective scheme. There are, however, specific stats for snowfall, if you're interested: https://www.emc.ncep.noaa.gov/users/verification/global/gfs/prod/atmos/grid2grid/snowfall but in the medium range, they only compare the GFS with the ECMWF and an older version of the GFS. If we look at past 30 days over the Eastern U.S. this is the frequency bias for 2". Forecast hour is the x-axis, and a score of 1 means that a model predicts 2" as often as it occurs. Values lower than 1 indicate an underproduction by the model. The number of events is likely quite small, so I wouldn't read too much into scores like these, but it's worth poking around on that site.

Worth noting that, given the amount of green on the 18Z GFS p-type maps, the GFS is terrible at maintaining cold air damming. If it's in the mid 20s at 7AM, and the sfc winds stay E/NE all day, areas around DC won't get into the mid 30s.

With the timing now appearing earlier, I think that may have no choice but to close schools early, even with the impacts of the event being uncertain.

Spot on. The NAM Nest simulated reflectivity has what's basically a thin squall line for all areas north of the DC Beltway in the mid to late afternoon. There is even some small sfc-based CAPE in the soundings. It will start as light rain but quickly turn into a burst of snow. My guess is that it whitens the ground in any places that get the higher reflectivity values. Whether it sticks around when it ends will be determined by how quickly the arctic air arrives. For those along the MD/PA border, there is a clear signal for a longer period of snow, so the potential for 1-2" isn't fantasy.

Thanks for posting the NBM, but it's worth noting that the NBM actually does compute SLRs and doesn't just use any sort of simple 10:1 approach. If the thermal profiles and such promote healthy ratios, that will be reflected in the NBM snow totals. It doesn't use the super-aggressive Kuchera approach, but it can still go well over 10:1 where appropriate.

I'm hesitant to buy the HRRR area of a stratiform shield. And I'm tossing anything that isn't a hi-res model, as this is a set up with a couple of bands of showers turning to snow squalls along the cold front, driven by dynamics and surface convergence, and enhanced by some steep low-level lapse rates. The NAM Nest simulation certainly looks plausible:

Let's leave it at "they're different". The QPF uses some addition inputs that aren't great, but they are bias corrected, which helps - they certainly add more spread. The NBM QPF also heavily weights the WPC QPF (prior to forecaster adjustment) beyond Day 2, and that is not hooked up to the NBM winter products.

FWIW, the inputs to the NBM QPF and snowfall are not identical. Definitely a source of confusion.

I’m cautiously optimistic for some thunder this evening, but the soundings don’t excite me for severe potential. That said, wind would certainly be on the table if we can maintain some surface-based instability ahead of whatever convection makes it to our area.

The signal is even stronger in this evening's runs. Almost every CAM has some light precip early Tuesday, and temperatures are just below freezing, except for the HiRes Window FV3 which is known to poorly handle low-level inversions. Given that temperatures will be very cold leading into this event, I would expect surfaces to be very cold, and icing would be a problem even for areas right around 31 or 32. Certainly worth additional tracking.

Also a marginal risk for damaging winds for the eastern parts of the area, per the updated Day 3 SPC outlook

You say that NHC runs its own models which are not prone to big errors; you might be surprised to learn that the primary hurricane models are actually generated by the same U.S. center that generates the models you hate so much. And those hurricane models actually use the same model core as the primary U.S. medium range model that drives you nuts. And those hurricane model forecasts, and NHC, bust far more often that you realize. Criticize models to your heart's content, but your 48 cent takes on what models are and how they're developed and run is tiresome.

We get that you don't like computer forecasting models, but could you please stop spouting pure nonsense about them?

It's a nice little forced line just ahead of the cold front. It's there on the HiResW FV3 as well, but that model is warmer and verbatim has heavy showers that might end as a burst of snow.

I appreciate what you're trying to say here, and you were very much on the right track in saying that a small error at the start can lead to a massive error downstream 10 days later. But you can't design a model like the GFS with "bumpers". You can absolutely mitigate extreme solutions with bias-corrected, calibrated products, but the actual model solutions are what the model integration leads to.

The reason is that, despite what one unnamed poster thinks, models are not driven by analogs and don't have climatological factors. You start with an initial state and integrate the equations forward, and they take the solution wherever they take it.

HRRR has the center of the upper low arriving Thursday afternoon with scattered to numerous heavy showers. This sounding sure suggests that small hail or graupel would be possible in any of the stronger cores.

I see what you did there.

Seems to be some growing consensus that there will be a significant east-west gradient of precip. Folks west of the Potomac are in way better shape than those of us on the east side,

The real push of drier air isn't until Saturday morning. Models disagree about exactly how dry Thursday and Friday will be (some of the CAMs show a fair amount of low clouds, while the GFS is more optimistic. The weekend is then really good, although somewhat higher dew points return Sunday afternoon in advance of the next front. This front passes early Monday, and it's the real deal. 60s for Tuesday on most guidance with lows in the 40s.

Convective elements should be rolling through the DC metro area early this evening. While surface-based instability should be negligible, the low-level shear will be very strong with impressive hodographs. I can't rule out a very brief spin-up, or at least a tornado warning or two necessitated by modest storm rotation, although the chances of reaching the ground are likely small unless the surfaced-based instability is slightly higher than progged.

Simulated reflectivities Friday don't look particularly impressive, but we'll be in a high PW environment for a chunk of the day, so some over performing rainfall totals are certainly on the table.