RU848789

Well the friggin' NBM finally came back to earth once the SREFs tanked, but it's time lagged, so it won't die for at least another model cycle.

I don't usually look at the 10% low and high probability graphics, but this one shows the huge difference in snowfall for the 10% chance that the high scenario plays out - not surprising given how steep the snowfall gradient is from SE to NW, which might be why the NWS is showing the snowfall gradient over the ocean. The 10% low probability is a Blutarsky-esque 0.0" for everyone, as one might guess.

Thanks, but even if one takes the GFS, GEFS and EPS over the last 3 runs for Philly, for example, the average is about 1/2" and those models contribute ~60% to the NBM according to your post with the SREFs having 30% input - if that's the case to get from 1/2" for 60% of the input to the 4.8" seen for Philly on the latest NBM, that implies the SREFs would need to be 12+" for Philly at only 30% of the input, which seems impossible, as the SREF snowfall (10:1) from the last few runs has been in the 2-4" range for Philly, unless that snow is at 30-40:1 ratios (and Kuchera is showing 20:1 inland). But I will say I didn't realize the SREFs were counted so strongly and it at least explains probably half of the NBM numbers for inland locations.

Don - thanks, as always, although I'm assuming you meant to say that's a wholly unrealistic estimate, correct?

Don - you shared the table below of NBM snowfall inputs last week, but I simply can't imagine how the new NBM is more than the last run and how either of them could be right even with the wetter SREF inputs, as I think everything else should be near zero, except for low amounts from the EPS. And if it were just an academic question it might not be a big deal, but the NWS regularly refers to it in their AFDs, so clearly they use it, which I don't get. Any insight?

How does the NBM go up for the Philly-NJ-NYC region after every 12Z model was a whiff (other than the SREFs)?

Your failure is complete...

Oh well, the RGEM didn't move and the GFS got worse...

And a decent jump NW by the ICON. Ok, now I have to watch the rest of the 12Z suite, lol.

Fixed it thanks - read the right one, pasted the wrong one, lol. Same message though.

Just when I thought I was out, they pull me back in...

It's fascinating that the NWS Philly is still hugging the NBM (model blend), below, which is an outrageously snowy outlier right now (as is the NWS-Boston), but not the NWS-NYC, which is favoring the rest of the models showing little to no snow outside of eastern LI. See the discussions below from the NWS-Philly and NYC. I think the NBM can be a good tool, but it is time lagged (it includes current and previous runs of over a dozen models), which is good for not bouncing all over the place, but I don't understand relying on it when it's this much of an outlier.Area Forecast DiscussionNational Weather Service Mount Holly NJ 246 AM EST Thu Jan 29 2026 In terms of impacts to our forecast area, this hinges on the exact track the storm takes which still remains uncertain at this time. Generally speaking, the latest 00z deterministic guidance has our forecast area near the N/W fringe of the storm`s precip (snow) shield. The GFS has shifted slightly farther N/W with the storm track and precip shield compared to the 12z run. It offers light QPF for extreme E/SE NJ and south Delaware. The CMC has moved its QPF sharply S/E and now barely has any QPF for our area. The ECMWF has remained steady with some light totals for NJ/DE. It offers a few inches of snow for DE and 1 to 2 for Eastern NJ. Our present forecast will have some of this in our (ending 12Z Sunday) snow fcst. More snow could fall after 12Z Sunday. A decent amount of uncertainity remains. When compared to Wednesday, the latest (01Z) NBM probabilistic data has remained steady near the shore but trended down for places close to the Delaware Valley. For snow amounts greater than 2 inches (plowable), the range remains around 60-70 percent near the coast to 25-35 percent near the I-95 corridor with lower probabilities N/W of the urban corridor. For 6+ inches, these probs have dropped and are now around 20 percent near I-95 up to 35-50 percent near the coast.Area Forecast DiscussionNational Weather Service New York NY 632 AM EST Thu Jan 29 2026 Regarding the potential for a coastal storm for this upcoming weekend, if any impacts were to occur it appears that the timing would be late Saturday night and into Sunday, and possibly Sunday evening for some eastern sections. Most of the dynamic modeling continues to show the storm being the nearest of misses, or, with the NYC metro not experiencing any snow impacts. The longwave pattern does amplify, but the mean eastern trough also begins to progress east while amplifying. A good rule of thumb is whether the 500 mb vorticity maximum or the 850 mb low gets north / south of Cape Hatteras. Just about all the dynamic and AI modeling has the vort max / 500 mb low going just south of Cape Hatteras. Climatologically that is a miss with respect to big snow for our area. In some cases far eastern areas get brushed. Therefore, at this point in the modeling / forecast process the most likely scenario is the nearest of misses. However, some impacts cannot be completely ruled out, especially further east as there remains some room for the modeling to wobble slightly further west which would bring eastern portions of the region in play for some minor to moderate impacts. Some coastal impacts remain a possibility in terms of minor coastal flooding. Stay tuned to subsequent forecasts for any changes.

And the cheese stands alone. The 6Z EPS ensemble mean is the last model of any I know of (to be fair I didn't check the Tanzanian - that's an inside joke on the weather boards) to still show appreciable snow inland from the coast (due to a few western hits) and just about all other models aren't even showing snow at the coast anymore (not counting Montauk). The CMC and GFS ensembles have finally thrown in the towel also. Yes, we're still 3+ days out so it's still possible for significant shifts west, but that's not the way to bet, especially in the time of AI models, which I don't recall making big shifts inside 3 days (but could be wrong on that).

Interesting to see Lee Goldberg's first "guess" (he said it was a model run he agreed with, but his first forecast will be out tomorrow), which I think is reasonable and WxRisk's first guess which seems very aggressive and based heavily on the 18Z Euro. I'm also still stunned by the NBM being so high - and it has been cut down significantly since 18Z (maybe from the ensembles being still snowier than their Ops). With the 0Z Op models all showing very little snow in our region, except for far E LI (and except for the stubborn Euro still showing an inch for the NJ coast and a few inches for E LI), I wonder if we'll finally see the NBM really come down and DT cut back on his totals (and maybe even Lee). Same thing with the probabilities for 4" from the storm in the NWS graphic below, which seem very hight. Sucks to see so much potential likely being unrealized, but I did just get 19" worth of "snow" (as frozen snow/sleet mass), so it's hard to complain too much. On the other hand, we're still 3.5 days out, and it wouldn't take a monumental shift to at least bring a few inches to 95, so you never know.

Hey @mgerb- curious if you have any feedback from NWS folks or any comments on my long post above outlining my rationale. TIA.