eduggs

The Hudson Highlands area had an OK winter this year, especially when you factor in a few early season localized light/moderate events and the sustained snowpack. But I drove through parts of Orange and Putnam yesterday and there was a lot of bare ground on southerly, sun-baked aspects and the usual late-winter snowbanks on the sides of roads were largely missing. People were ice fishing out of the lakes, but the vibe was of a low-snow winter. Down in northern NJ this is the 3rd winter in a row essentially without snow, particularly east of the I-287 terrain boundary. It's been a miserable stretch for snow lovers.

Yup. Days and days of fairly steady modeling showing a miss despite an anomalous setup. Absent a few 5-7 days runs with coastal impact, this has been well telegraphed. You can't even hang your hat on any particular fantasy SREF member anymore as of 15z.

Japan has had a relatively warm winter. Despite that, some of the "heavy snow regions" have had a snowy winter... probably partly due to the warm anomalies in the Sea of Japan. Or are you referring to a different year?

NMB9 showing the extreme (weenie) solution.

A quick look at recent SREF members shows the kind of upper level evolution that would be required to get snow into our area from the coastal. The amount of modeled snow is closely correlated with the orientation of the height lines out ahead of the ULL (area highlighted below). In the very few members that show a few inches of snow for our area (from the coastal and not ULL pass), the 500mb height lines are oriented nearly south to north, indicating an ULL taking on a neutral or negative tilt. Note the 12z 12km NAM is on the more extreme end of the spectrum although not the most negatively tilted compared to the full 9z SREF suite. In the majority of members and models that show no snow from the coastal, the trof is positively tilted and height rises in this area are less pronounced. This connection is probably obvious to most. But I still find it somewhat illuminating to look at the individual SREF members since they show variation far outside the likely envelope of solutions.

The 12z RGEM actually made a similar shift to (/toward) the 12z NAM aloft in terms of adjusting the ULL position slightly west and raising heights out ahead. Unfortunately this only translates to a very slight northwestward shift in SLP and precipitation shield. But if you're tracking for the loooongshot chance of a big model bust, then you'd want to see a continued adjustment in this direction. The RGEM also has some snow with a norlun type feature as the ULL passes.

I think you discount the GFS slightly here because of the other guidance, but definitely not entirely. The GFS shows that if some aspect of this evolution is not favorable, the whole thing could fall apart. A miss southeast is still very much on the table. I'll be looking for run to run consistency from the other models in upcoming cycles; minimizing the variability will increase confidence. And of course any improvement in the GFS will be welcomed. But there could easily be occasional misses across guidance for the next few days. It is almost always like that at this range in advance of major storms.

As modeled on the 12z ECM, this is a 12-18hr event, depending on location across our area. Several of the individual members that gain more latitude also provide a longer duration event because we spend more time in the CCB as it pivots over our area.

And it's not like the GFS hasn't bit at all. The 0z was a significant snow event for DCA/Balt to PHL and major NYC to BOS. 6z was still significant for most of the corridor.

I don't have access to the individuals, but from the looks of it I'd estimate at least 25% of them would bring > 1.5" liquid to most of the area. Lots of high impact solutions there!... including several more intense than the operational. But the caveat is that individual ensembles cluster around the parent operational run. So the true chance of >1.5" liquid from this storm is probably less than indicated by the EPS spread.

It's a definite risk with a pinwheeling ULL pushing into confluence. If the complex is shunted east before mid-level lows mature, the magnitude of event collapses pretty quickly.

The low totals from the 12z ECM operational run are due to the SLP not gaining latitude. An ULL (offshoot of PV) forms off the coast and oozes eastward as the system stacks, pulling the whole system eastward. Great position for us, but not great for places further northeast. Many of the EPS individuals gain more latitude and would be great up and down the coast. But a slide east either with a modest event or a complete miss are still very possible.

The 12z ECM is obviously great, but we gotta watch that eastward slide at the end. The relatively low snow totals in SNE are a small red flag. I'd prefer this to tuck into RI and stall then to slide east so quickly. We're good for now but if the eastward pull happens earlier we get the SNE modeled snow. Then again other guidance (e.g., UK, CMC) hammers SNE.

I've heard that. But I don't believe it's true. The GFS OP is run at a higher vertical and horizontal resolution and also uses the observed initial state and boundary conditions. The individual ensemble members are run at a lower resolution with perturbed (adjusted) initial conditions. If you move far enough out in time horizon such that the resolution and data advantage of the OP run is minimized, I suppose the accuracy of any single individual member might approach the accuracy of the OP run. I'm thinking probably out beyond day 10. But at closer ranges the superior resolution and more accurate initial time steps of the OP should make it superior.

The UK is kind of a classic major I-95 nor'easter. Snow takes forever to push northward from DC/Balt to our area and then the precip. shield wrap tighter into the bombing SLP and hammers the immediate coastal areas.

The GEFS mean reflects some timing variability that is spreading out the QPF mean. There are a couple of different very minor h5 shortwaves that some individuals are keying on. Lots of spread on the GEFS even compared to the EPS and GEPS. There are a couple of good hits on the individuals, but not a ton.

The GEFS mean at 500mb is a pretty good match for the 12z CMC through 126hr. It looks good to me. But the evolution of the CMC after this time period is key for delivering a big coastal storm.

Not enough amplitude on the trof on the GFS. Hope it's a blip.

I follow that same area for skiing purposes as well, but I don't fully agree with your analysis. The ICON, GFS, and CMC were clustered fairly close together for several days. Only the ECM showed a stronger primary SLP holding on and sending mixing into NVT and NNY. Yesterday the GFS shifted a bit towards the ECM, but the CMC stayed in the southeast (faster 2ndary development) camp. It's a bit difficult to judge since it changes every cycle, but I would say the ICON has adjusted similarly to the GFS. From NW to SE the order has been ECM, GFS, ICON, CMC... with the NAM now the furthest NW as we approach shorter ranges. I have not noticed the ICON having a "weak" bias relative to the other mid-range models. Certainly it has less of a tendency to amplify minor shortwaves than the NAM. But from what I've seen it's fairly middle of the road in its upper level synoptics and surface reflections. Regardless of its specific biases, it is slightly less accurate than the CMC and UK. It's one of the best mid-range models in the world, but clearly on the 2nd tier.

0z EPS looks great. Lots of left leaning individuals and a robust mean QPF total. GEPS similar.

Great 0z cycle. GFS, CMC, and even ICON are not that far off aloft. Big time threat. I'd ignore the kuchera snow output for now as it can only lead to disappointment.

I think the 12z EPS looks excellent for late next week. I couldn't really draw it up much better at this stage unless I wanted to risk mixing with a direct hit. I guess you could ask for slightly longer duration if you were really greedy. But it has high end potential, only modest (for this time range) spread, and lots of good hits among the individuals.

ICON, EC, GFS, and ~ 25% of ensemble members have offered snowstorm solutions over the past 2 days for late next week. There is clearly some inter-model support here. Very tenuous with the PV evolution though.

CMC would work.

Every storm is unique so each favors a different area. Trying to fit everything into "patterns" and storm types like "gradient" or "SWFE" is way too simplistic. There is just too much variability and randomness.