eduggs

This is also why the hilltops look to get a lot more ice than the valleys and lowlands. The cold drains in at low levels, but off the deck. So 1500ft could drop to, say 27F, while down at 400ft is at a marginal 32F and not really accreting (meanwhile 5000ft could be 40F). In a more typical CAD - ice scenario, the sheltered valleys hold on longer than the hilltops. This upcoming modeled scenario is not the prototypical ice event.

It looks like elevation will be key for areas that get ice. Pretty steep low level lapse rates on the model soundings. Near freezing temps. in low elevation areas will probably limit accretion. But up 1000ft+ with surface temps. in the upper 20s will start to become a problem. Actually for big problems you'd probably want to be up above 1500 or 2000ft at least in SNE.

There's really pronounced upper level divergence south of SNE starting around Monday on most guidance. That definitely seems to match the developing surface reflection and cold northerly surface drain. Too bad there's not much mid level redevelopment in the same area to more fully cool the column and wrap in some more moisture.

Well the squalls always hit some areas more than others. But overall the front and snow shower activity was well modeled since last weekend.

The mid-level charts never looked good for snow locally. Trof axis and mid-level lows too far N&W. But many cases in the past have shown that a burst of overrunning precip. into a marginal airmass can produce over a short duration. We obviously did not produce in those critical early hours. I don't know how you can know for sure which way things will break when you are along a sharp thermal gradient. The outcome is so sensitive to minor differences. I guess nobody should have been so quick to discount the GFS. It seems like maybe the practice of discarding outlier model solutions might lead to suboptimal statistical forecast verification. I'm not sure how much snow fell up north, but it consistently showed mixing to ALB with snow to the Canadian border and ski country. It's too bad model soundings aren't publically available for the ECMWF hi-res.

Sleet and freezing rain mixed in Putnam County. There's some pretty good icing in the hills. Main roads were plastered with salt before and during the precip. so they are just wet. The GFS or even a model consensus worked out pretty well up here. What threw people off was the shift towards an initial burst of snow over the last 24-36 hours and the very high EPS snow probs. Soundings consistently showed a 3hr window or less to accumulate snow. After that all guidance showed freezing rain with some sleet. This can't really be a model bust if the outcome was within the model envelope. Forecasters hedged towards the ECM on snow accum. and the GFS was more right this time. It was a minor forecasting bust but one that would have been very difficult to avoid.

Radar looks moist with the overrunning out ahead. And it looks relatively suppressed too. The mid-level precip. is also clearly and distinctly visible in KS and MO. There's probably going to be a quick burst of snow for many tonight. But this continues to look really icy locally. Maybe some paste and parachutes initially with the WAA but the powdery snow is confined to BGM, ALB, and SVT.

I would say somewhere near Morristown with an approx. north-south gradient. But I would guess it ends up a bit further north. However, the initial burst of precip. has been trending more impressive. That could do some work late tonight. And it could result in less linear snow - sleet - freezing - rain - rain boundaries and more banding dependent.

Euro shifted a bit further than that with mid-level temps. Shifts the sleet and freezing rain lines a good deal north on Tuesday. Still a good front end thump for those who can avoid sleet.

There could be some good snow squalls on Wednesday evening. The hi-res models are picking it up nicely. Even the globals show some QPF for the area with the arctic frontal passage despite local downsloping. I'm thinking there could be isolated 1-2" accumulations. Upslope regions further up the taconics and also NNW NJ and NEPA are preferred but some squalls should make it to the coast. Obviously snow lovers want big storms not isolated squalls. But a ground whitening is always nice around Christmas time.

I agree on the ice threat. 800mb temps still warm to about 5C right up through the mid-Hudson valley. The 0z NAM increased the QPF with the first 6 hours of the main WAA push so there's a little snow thump there. But after that mid-level warmth wins out like previous runs. NW NJ keeps showing up as a focus area for icing potential. But I think lots of areas from NEPA to SENY through SCT are at risk for some moderate ice.

You are right. People only seem to remember snow maps. And snow accumulation and depth aren't even directly modeled geophysical parameters. The ECMWF models literally dozens of parameters across the entire globe and people only consider one derived parameter from an individual run for a very localized area when evaluating model performance.

It looks like it's showing freezing rain as snow too.

I think the global medium range models have performed pretty well so far this winter. Maybe I'm just old enough to remember when it was really a crapshoot beyond 3 days. I can really see the improvements from 10 and 20 years ago. These days people get upset when the thermal gradients shift 100 miles at 5 days out. For a global model that's like a millimeter across a football field.

It's sad that anyone even knows who Joe Bastardi is. He does a disservice to the meteorological field. He was exposed as a hypster fraud 20 years ago.

The 18z NAM looks to keep any rain limited to southern Westchester and southern Bergen on south. It has maybe an inch or two of snow and sleet and then a few tenths of ice for the immediate N&W burbs. Right on the border of a legitimately wintry event. I'm not sure what those snowmaps above are showing. Definitely overdone.

The best wintry event between now and Christmas might end up being the arctic front on Wednesday.

The 0z and 06z Euro were cold and suppressed. The 6z ICON had a similar look but slightly warmer and wetter. The global and regional Canadian appear to be in the middle of the model envelope in terms of temps. but also relatively wet (high QPF). Snow and sleet down to near NYC switching transitioning to a significant ice event in NNJ and SENY. The GFS and NAM are both warmer and show minimal frozen precip. in the immediate NYC area. But both show ice for the northern burbs. The suppressed Euro is the one thing that gives me pause. It looks weird considering the mid-level features but you can't discount it. Right now I think a compromise is the best way to go. Minimal snow for the area with a light to moderate ice event north and northwest of the City.

Goose are you referring mostly to the initial WAA push when you describe models as being south? Yes that initial features has shifted south for a day or two, but the follow up main shortwave is still modeled to push warmth pretty far north on most guidance. The GFS, RGEM, and UK look pretty warm and wet to me with the main system. Are you anticipating less of a lull between the initial push and the main batch than shown?

RGEM looks wet and warm to me... at least compared to the EC and 0z NAM. It looks like ALB might mix on that run. Unless I'm looking at an old run...

The GFS has snow showers on Mon. night and then ice to rain on Tue. It has been fairly consistent with this evolution although it has shifted south slightly over the past day or two. It seems reasonable to me. I hope it's wrong but I really don't like the 500mb setup. Even if the cold wins out I feel like we get sleet down here.

There's no reason NCEP should have better weather models than Germany. Maybe 30 years ago but not anymore. All countries with national meteorological agencies want to use their own data. It's interesting to see how respect for different models differs depending on where in the world you live.

But the initial WAA surface south of us occurs with the main shortwave back in Kansas. And the shortwave is still pretty sharp as it traverses the ohio valley on Tue., which could easily spike mid and low level temps as the main batch of moisture arrives. Ice is definitely likely considering the antecedent airmass, but this could also easily be a brief ice to rain scenario. The 500mb chart is far from ideal for extended frozen precip. in my opinion... except maybe far inland. But I'm certainly hoping this south trend is real.

Just a general model comment: IMO the only global, medium range model that is clearly superior to the German ICON model is the ECMWF. In terms of 500mb verification, the latest FV3 incarnation of the GFS is probably also superior. But compared to the others: Canadian, UK, Japanese, Korean, French... the German model is as reliable or better. I have much more confidence in the ICON for the extended short range period than the NAM lately.

It's actually a pretty sophisticated model and the German water and meteorological authorities rely on it heavily to manage their water resources in the Bavarian Alps. It also seems to be performing well (locally) so far this winter.