Morning AFD from Mount Holly. Interesting, excellent, and detailed.
It has been a live-or-die-by-every-model-timestep sort of night here at the office. And what the models give, they also take away, which basically describes the model variability we have seen the past couple of days with the potential winter storm for our region Friday night and Saturday. The 00z suite has made a notable consensus shift westward with the low tracking just off the coast, but the high volatility/variability remains. The 00z GFS is a far-east outlier solution and basically brings little snow to the CWA for the whole event. Meanwhile, the 00z NAM returned the snow to our area, in a large return of departure from the 18z NAM no-show (no-snow) event. The 00z CMC brings a blockbuster storm to the area, with widespread warning criteria south and east of the Fall Line; the 00z ECMWF is only somewhat drier than its 12z predecessor run. The 00z UKMET went sharply west, with meaningful QPF/snow for much of the area.
The model volatility with this system has been something to behold. Such run-to-run spread is typical/expected for these types of events, given the highly complicated interdependent phenomena involved. However, simple analysis of the spread in the National Blend of Models is enlightening; the 04z NBM V4.0 (V4.1) 50th percentile storm total snow for the Philly area was around 1.5 (1) inch(es). The 75th percentile storm total snow was around 8 (10) inches, or roughly an order-of-magnitude difference between the median and the upper quartile. Bottom line here: the storm total snow forecast remains highly uncertain and subject to large changes in subsequent forecasts.
Observation-wise, it will be critical to assess three regions/phenomena as the event unfolds: (1) the strength/depth, orientation, and speed of a northern-stream digging vort max through the Midwest on Friday morning, (2) the orientation and speed of a southern-stream vort max in the southern Plains around this same time, and (3) the low-level response to the phasing trough near/off the Southeast coast Friday night (e.g., the 850-mb heights and winds). The progressive solutions have a more compact and faster northern-stream vort max and a slower southern-stream vort max, which results in upper low development farther east (and generally too far east for our region to see substantive snow); the snowier solutions acquire phasing and neutral to negative tilt of the large- scale trough more quickly (and thus, farther west). Trends in the low-level response are obvious Friday night -- the NAM/CMC (aside from the errant 18z NAM simulation) are positioning the 850-mb low/trough farther southwest. The GFS is much noisier, exhibiting little trend. Notably, the 00z GEFS featured unusually low spread, which makes me wholly suspicious of the deterministic and ensemble output from its suite.
With the model camps making the GFS more and more of an outlier, tonight`s forecast is generally a non-GFS consensus blend. This preserves a considerable amount of continuity to fields of importance such as PoPs, QPF, and snow amounts. The main changes were to sharpen the gradient of snow totals near/northwest of the I- 95 corridor, with 1-3 inch totals northwest of the Fall Line, 3-6 inch totals in the urban corridor and immediately adjacent areas, and 6-12 inch totals roughly from Easton, MD, to New Brunswick, NJ. Again, there is enormous uncertainty with these forecast totals. If the more progressive solutions pan out, very little snow may occur in a large chunk of the area. If the slower/stronger solutions pan out, heavier totals would occur at least to the Fall Line. Continue to monitor the forecasts, as large changes may occur leading up to the event.
