eduggs

Rates (precipitation per hour) are actually modeled to be fairly high. Modeled FGEN is significant across several models. But total precipitation is relatively light due to the short duration. Several models including the GFS deliver over 0.3" liquid over 4-6 hours in or near our region. A few models even have greater than 0.5" liquid nearby. That's a fairly significant QPF signal for short-duration frontogenic banding. The forecast difficulty is that the modeled placement of this banded feature (CT, LHV, CNJ) is highly variable. Forecast temperatures range from the upper 20s to mid 30s across our region with NYC roughly 31F to 35F depending on model during precipitation. Wetbulb temperatures are a little lower and modeled surface temperatures dropping during precipitation. Notably the HRRR is warmest, particularly after precipitation ends Saturday afternoon. NYC will probably record a minimum temperature below 33F on Saturday.

Saturday only (10:1)

Don't sleep on Sat. It's basically a mesoscale event. Could be nothing or a fast 3". Pretty significant FGEN values across multiple models. And it could come in multiple bursts predawn or daytime. My guess is CT or the HV is the winner, but a few runs have targeted CNJ to NYC. Could be a fun nowcast with low expectations. Contrary to what some say, this would have no trouble accumulating with 0.1" liquid per hour at 33F. Most will probably miss, but if you have the rates it will accumulate.

It's wild how good the AIGFS looks at 500mb. NYC is again over 0.5 liquid from just the Sunday event and the precip. shield ticked slightly NW.

0.4" liquid in 6 hours showing up on globals and mesos. It's very isolated and variable in placement. But someone could get a surprise with some intense frontogenic banding on Saturday!

Long range HRRR for Saturday.

I'm hoping for more than mood flakes on Sat as I expect that to be the snowier day for me. I think LI and southern NJ has a decent shot of accumulating snow on Sun. But I'm not closing the door on a significant west trend.

Lovely write up. For a change I don't think it's all hype and BS. I seem to remember this feedback loop being enhanced when PVA tracked close to the Gulf states and picked up additional moisture inflow... which is what is forecast by most models to occur.

I love how it's snowier than 12z. Lovely model battle

I'm in a similar situation both this year and in recent years. And I agree about the anecdotal observation. The RRFS is prone to wild swings (i.e. errors), especially beyond 48 hours. But it's still better to see it shift west than east. It's showing a significant snowstorm for NYC east and south on Sunday after an appetizer for some on Saturday.

The RRFS is very expansive with the precip. shield to the NW of the SLP. It's probably too expansive. I'm guessing a lot of that would end up as virga with that surface depiction combined with a weak 700mb low. The RRFS, NAM12km and NAM3km have a frontogenic banding signature on Saturday.

Where I agree is that meteorologists can use local knowledge combined with model output to occasionally outforecast a global model locally, in the short range, and for limited parameters like surface temperature. But forecasters who think they can outforecast a global weather model at the synoptic scale or in the mid-range are deluding themselves. They are susceptible to all sorts of biases that convince them that their gut feelings are superior (confirmation bias, availability heuristic, confidence bias etc).

I don't agree with this at all. I think it leads many people off a cliff. People think their intuition regarding loosely defined concepts like "pattern" is superior to supercomputers developed specifically to model exactly what's possible in the atmosphere. It's pure ego.

Where I agree with you is that if you pre-define a synoptic setup as simple (moist southerly flow into a cold dome), then it is comparatively simple. We do get hung up on semantics. But I also think we oversimplify and broad-brush weather events with terms that are too general.

I disagree with this. ECMWF and NOAA/NWS should be the baseline. AI can use verification to improve modeling beyond the current state of physics-equations-based modeling, which is limited by its programming and unable to quickly iteratively improve.

Thanks. I'm not saying it's an analogue. NYC isn't going to break its all-time snowfall record Sunday night. Yes the synoptics are broadly similar, but the similarity I want to highlight is that 3 days before that event, it was written off as an OTC solution.

The UK is at or below freezing for all except LI during Saturday's snow. It looks nice actually. A proper snowy weekend day. Surface temperatures near the City warm slightly above freezing after the precipitation ends.

I like the EC-AIFS too. I think it has done well this winter. Under 48hrs I'll be looking at the NAM. That's probably an unpopular opinion, but I think it does well with these trofs that touch the Gulf when the height field along the upstream trof flank is questionable. It often signals how much room there is to come NW. It the NAM stays east in the short term, that usually ends it.

I don't believe the AI models use physics equations (mass, momentum, energy etc). But they are indirectly built in based on the training sets they incorporate (e.g., GFS/ECMWF).

It's never simple. Overrunning also gives you sleet, freezing rain, or rain. And wave interactions can always apply regardless of synoptic setup. Threats look simpler in the long range because of the ensemble averaging effect. So if you overlay the height field with the precipitation field, it often looks like basic overrunning with flat mid-levels. By the time you get closer to the threat, the ensemble spread significantly narrows, the amplitude of the waves increase, and you see the real synoptic complexity.

Yeah I think it's about 28km depending on latitude. And the lack of model physics would also contribute to smoothing since it's using purely statistics and averaging.

EPS has been very stable for a full 24 hours. The spread is not that large either. I don't expect a big shift at 12z. But it could still shift tonight or tomorrow. A minor model error in a shortwave in northern Canada today can have a big impact on the height field on Sunday. Trofs that touch the Gulf can morph quickly. I usually look at the NAM for short term (<48 hours) trends with these types of events.

CMC is many many hours of mostly light snow on Saturday and Sunday. I think that would be a glorious outcome. I would prefer that to 10" of snow that falls from 11pm to 5am overnight on a work day.

Thanks for catching my error. I actually meant February 11/12, 2006. This was modeled to be a scraper until 2 or 3 days out. It trended NW viciously in the short term and localized mesoscale banding made for a memorable event in NYC/Westchester, and SWCT.

I still think Sat. has sneaky potential. Models have been moving the area of precipitation around a lot so it's hard to pin down. But up to a localized 0.3" liquid has been popping up across guidance for a few days now over only a 6 hour period. With low expectations and most people looking to Sunday, that could spell a surprise for some. I'd guess somewhere in CT but really anywhere is possible. It will probably be somewhat localized. Coastal areas are less likely Saturday unless it's before sunrise.