That’s Harlem for ya. Didn’t need the NAM to tell you that was going to happen. I grew up in NW Rockland county where it was typically 5-8-10 degrees colder during storms and snowed twice as much annually despite being only 25-30 miles away from downtown (as the crow flies) That’s just the climo of that area. 287/87 has always been the fall line in the NYC metro. It’s a different world N and NW of the NYS thruway. I can recall at least 7-8 storms where nyc rained and my area saw over a foot. The Bronx always did better than Manhattan which did better than Brooklyn, queens, and the island, etc. 24-36” of pure snow in downtown NYC is extremely rare for a reason.
Your point about the NAM’s performance is spot on. Inside of 24-36, the hi res has been cash money. I’m not sure how many storms people need to live through until they grasp the NAMs dominance in the short range and that it is literal garbage outside of 60. I said it to @WinterWxLuvr last night…. If the 3k NAM shows a near identical low track and high positioning to Globals within 24 hours but shows the warm Nose 50 miles further north, you best be paying attention - its probably right. IMO, the NAM shouldn’t even be looked at until within 48. It will save people a ton of confusion about the NAM showing a snowstorm in Ohio / Buffalo 5 days out and the whiplash it does to reality within 48.
I understand many are disappointed and rightfully so, but Models overall did quite well with this. Euro, GFS, PARA, EPS, UKMET, etc all pretty did pretty well in regards to the overall pattern evolution and track a few days out, which is what globals are used for - and in the short term, the mesos (RAP, NAM twins, HRW) took it home on the nitty gritty details. That’s why using a mixture of model consensus and local climo is so important when making a forecast and why Mets don’t just choose a “reliable model” and make a forecast off of it verbatim. I know it’s tempting for folks to compare ground truth to kuchera snow maps and see a bust during what’s been a very tough winter for many, but if you analyze models based on their strong suits, they actually did pretty well here once we got within 3-4 days. In a general sense, it is a shame that in the year 2021, models are still so far apart on the details but what can we do.