eduggs

Other than the usual weird jump by the NAM, so far 0z looks similar to previous cycles. Changes look like noise to me... Still looks like a coating to 3" or so from the south coasts to far northern suburbs with a ragged and unpredictable gradient... The variable and inconsistent geographic distribution of snowfall between models suggests a low probability forecast. Model changes are being driven by very minor differences in vorticity and micro-short-waves. The WRF-NSSL is the weenie model of 0z with 3-5" right through the NYC metro. Unlikely but probably not impossible.

I'm thinking 2" as the average across I-84 - not the upper limit - but with a wide range and maybe a funky regional distribution. I wouldn't expect uniform and consistent precipitation. I would guess the precip. shield is more banded and variable with winners and losers. But it's just a guess really.

Considering just about all guidance has 0.2" liquid or more across NNJ and SENY and a snow-supporting model profile, those WPC snow maps look low. You would seemingly have to go against model consensus to come up with those probabilities. Or maybe they are heavily skewed towards the GEFS. Either way, the lack of granularity across our region makes it only modestly useful.

I think Friday still bears some watching. A few ECM runs and GFS-AI run the precipitation into a wall of CAD. Lots of northerly low level drain. It's a much better threat for New England, but there's still enough time and enough ensemble spread... Pretty low likelihood however outside of some ice across the northern interior or maybe flakes to end.

Warm advection precip especially with a parent low significantly separated from the overrunning can be locally unpredictable. It can be banded with distinct maxes and mins that defy typical geographical snowfall distributions. The inter- and intra-model variability supports this. As usual, this will be an interesting nowcast. With surface temperatures near or just above freezing, snow could accumulate pretty much anywhere. Can't really count anyone out for a coating to about 3". I don't think this event will reveal its character until the last moment. Hopefully when it does early Tue it appears wintry and festive.

The RGEM wasn't further north. The shortwave was actually a hair south. But lighter precipitation and different distribution of precipitation screwed the impression. ICON and RRFS highlighting the potential for a narrow area of heavier banding. Some indications of possible 0.1" liquid per hour. It probably wouldn't last too long as it shifts east, but it could put down a quick couple of inches if we're lucky.

We haven't had any particularly good setups for snow this late fall and early winter and yet there have been several accumulating snow events, particularly for NW areas. It's a little frustrating to not have any bigger events to track, but I must admit December has felt pretty wintry compared to recent years. Maybe that continues early Tue... There are some indications of a pretty good burst of snow with cold antecedent conditions. That would continue the trend of sneaking accumulating snow out of a marginally favorable setup.

The RRFS actually has additional snow showers for the area (esp N) Tue night. Doubtful since there's not much support and the RRFS isn't reliable at the end of its run. But if the tight thermal boundary stays close to our area and we remain on the cold side, there will be chances for snow. Different models at different times have shown snow chances every day from Tue through Fri. It depends on the track of the very minor shortwaves in the longwave flow and how they interact with the thermal boundary. At some point next week I suspect we'll switch to warmth... but it's not certain yet.

RRFS is even further south than 12z. Precip is not quite as robust as 12z but it's basically all snow for our area except maybe ELI. Quick burst then the forcing quickly fizzles. Not much snow for SNE. Good sign to see the south shift but we don't want to lose the signal for heavier banding.

Model guidance brewing some big warmth for late next week. GEFS, GEPS, EPS all have it. Most of the country is modeled to be above freezing on Xmas. Our warmest day might come right after. Still waiting for a positive surprise in the mid-range...

All of the mid-range guidance through day 7 has the thermal boundary too far northeast for us to have much of a shot at snow. Out to day 10 isn't much better. I hope it changes. It's a pretty ugly and repetitive looking longwave setup that would seem to favor New England for snow chances. But it doesn't exactly look warm either after Friday and the real cold is not too far off. But it's sure no fun looking at a huge stagnant ridge over the central US.

Cool to see several reports around or even over an inch of snow from yesterday's snow showers in Morris, Warren, and Sussex counties. That made two days in a row of accumulating snow for the higher elevations. https://forecast.weather.gov/product.php?site=NWS&product=PNS&issuedby=PHI This was pretty well modeled.

Right now the best shot at moderate snow accumulations based on a model consensus would be I-95 east in CNJ and maybe into Long Island.

If the GFS is right, some spots might get more snow from tonight's lake effect bands than from the Sat. night event.

The junction of I-287 and I-80 looks like it got into that band pretty good. I bet it dropped a quick 1/2" plus in that area.

For the past few days, meso models and even globals advertised a finger-like lake effect snow band getting into NNJ this evening. And it has verified as advertised. Count me as impressed.

Yet another snowy evening out there. Everything's dusted up pretty good - maybe 1/4". From radar I suspect parts of western or northern Morris county have accumulated 1/2" - 1" this evening. These minor events help soothe the snowstorm deprivation angst.

I have distant memories of short range models fumbling these types of setups, which enhances the tease. But modeling has improved a lot since the 2000s, and especially the 90s. We don't see big, late shifts very often anymore.

FWIW, big improvement on the 18z RRFS vs. 12z... now generally in line with NAM and RGEM.

Yup. But the magnitude of run to run changes is so small at this time range now that it will be really difficult to see big improvements. The range of outcomes has significantly narrowed.

I think it improved early in the run but then shifted east at the end. The ICON did the same.

Noticable shift ESE with the PV on the 18z NAM. Looks more suppressive. Bummer... was really hoping to get lucky with this one and keep trending it. Although it shouldn't be surprising that the outlier solutions shift towards model consensus.

The 12z CMC is the first run I've seen that fully rounds the base in time to spin up the mid/ULL into Nova Scotia. Too late, yes... but closer. The next step would obviously be for it to happen faster and in time to impact EMA instead of Halifax. ... gotta keep pumping heights out ahead of the trof in Ontario/Quebec.

Regarding trend analysis. You should include hypothesis testing and p-values to assess the statistical significance of the trend. Most random scatterplots will show a non-statistically significant linear trend... so you have to analyze if the trend means anything. By visual inspection, the negative trends are largely due to lower-snow years starting in 2015. It could be argued that this down-cycle explains the trend and will be reversed after the next up-cycle. I do believe warmer winter temperatures are affecting seasonal snowfall. But snow is a tricky parameter to analyze and statistical analysis can be very misleading if not presented faithfully.

GEFS mean keeps ticking a little "wetter." NYC just shy of 0.2 liquid now.