February 7, 2017: Upcoming Snow – Where, When, and Why?

After a hiatus in snowstorms over the Northeast during the midpoint of the winter, the most significant snowstorm since the late-January sleet/rain nor’easter and the early January coastal low is likely to affect the region on Thursday. The complicating factor, however, is that as recently as 3 days ago, the model guidance depicted virtually no storm system affecting the East Coast. This post analyses the meteorological factors driving this snow event, and addresses some forecast predictability issues associated with such cases.

Current Analysis

Always a crucial step to beginning the forecasting process is considering current and past conditions, in order to get a sense of how the weather pattern is evolving and what key features have a significant role in driving the forecast over the coming days.

The image above shows the water vapor satellite image from 1415 UTC this morning (9:15 AM EST). A trough axis can be inferred over Texas, with a developing cyclone downstream producing a widespread precipitation event across the eastern US. Farther upstream, another cyclone can be seen entering the West Coast, which will play a large role in the forecast for the next few days as will be later discussed.


RTMA analyzed temperatures and MSLP (hPa), valid at 1500 UTC (10 am EST) this morning.

The image above shows the RTMA-derived temperatures and mean sea level pressure valid earlier this morning, showing the developing cyclone over Illinois with a 996-hPa minimum pressure. The warm front associated with this cyclone extends eastward into Washington DC and southern New Jersey, as highlighted by a temperature gradient, with an unusually warm and moist air mass to its south with temperatures in the 60s as far north as Illinois. A strong 1028-hPa high pressure over Canada helped to keep low-level cold air in place through this morning, allowing for sufficient cold air in spite of strong warm air advection aloft to allow for freezing rain inland of NYC and snow over New England, although the precipitation type will ultimately transition to rain everywhere as the cyclone rapidly deepens with a strong southerly low-level jet transporting warm and moist air into the region.

As this cyclone rapidly deepens and progresses into Canada, its associated cold front will move through the region tonight with strong cold air advection in its wake. However, the storm system currently over the western US will quickly progress eastward through the US, bringing about a snow event in the eastern US in a winter that has seen few snowstorms compared to recent years.


Snow on Thursday: Meteorological Synopsis

The forecast for Thursday’s event is one that has seen numerous shifts over the last couple of days, with no low pressure forecast at all as recently as 3 days ago, which will be addressed in more detail later in this post. While details are still somewhat uncertain given the relatively large spread in model guidance for a 2-day forecast, the large-scale dynamics governing the forcing for precipitation are largely agreed upon, giving an indication of the likely scenario that will play out, even though the exact location and intensity of heavy snow is still uncertain.


12z GFS initialization (left) and forecast for Wednesday night (right) for 500-hPa vorticity and geopotential height. Images from Tomer Burg’s Model Page.

Forcing for Ascent:

The GFS initialization above depicts a shear vorticity maximum over the Northeast Pacific Ocean, which will initially progress eastward fairly quickly, embedded within a relatively zonal flow. As another system approaches over the North Pacific and induces ridge amplification downstream, the downstream response of a trough amplification over the central US will lead to the shortwave trough digging southeast and becoming more amplified with time, as can be seen in the right image in the panel above. Strong cyclonic vorticity advection is associated with forcing for ascent, as is indicated in the dashed blue region ahead of the shortwave trough.

The 300-hPa wind forecast above, valid on Thursday morning, shows two jet streaks, one positioned off the coast and another over Southeast Canada. The Northeast US is positioned near the co-location of the left-exit quadrant of the southern jet streak and right-entrance quadrant of the northern jet streak, both regions of upper-level ageostrophic wind divergence which is indicative of upward vertical motion. In conjunction with the 500-hPa vorticity advection, there is plenty of mid to upper level synoptic-scale forcing for ascent, supporting a surface cyclone developing off the coast along with widespread precipitation intensification and development over the Northeast US.

While these forcing mechanisms support widespread precipitation, there will be other mesoscale factors at play determining exactly where the axis of heaviest snow will set up and how much snow could fall. As the cyclone develops and intensifies, the temperature gradient will become enhanced with stronger warm air advection from the south and cold air advection from the north, resulting in strong mid-level frontogenesis, which is often associated with strong upward vertical motion and heavy snow banding.

Given the current model data, the strongest frontogenesis is likely to be centered near and just north of the tri-state area at the 700-hPa, indicating a potential heavy snow band placement over the tri-state area. The sounding above from the 4km-NAM over northern New Jersey, valid at 10am EST Thursday, shows an isothermal layer near -10C extending up to 600 hPa, overlapped with the region of maximum ascent, which while not perfectly ideal for dendritic growth does suggest the potential for snow ratios above 10:1.

A potential limiting factor for accumulations will be the antecedent warm surface temperatures, considering that daytime highs on Wednesday will peak in the 50s and low 60s. While this will likely result in some of the initial snow melting upon contact with the ground, snowfall rates under the heavy snow axis are likely to be sufficiently strong enough to overcome the warm surface temperatures relatively quickly.

As with many storms, predictability issues exist with the northern extent of the heavy snow axis and accumulation in those regions. As currently suggested by the model guidance, this region is likely to be north of the tri-state area, towards the Albany, NY to central VT/NH corridor, where in spite of a relatively favorable temperature profile for high ratios, weaker forcing for ascent along with subsidence north of the heavy snow banding will likely limit snow growth and may support ratios near or perhaps slightly below 10:1, indicating that the snow accumulation gradient on the north end of the system is likely to be sharper than currently modeled.


Snow Forecast

While the model guidance has yet to settle on an exact track of the low pressure and placement of the snow banding, the synopsis above gives a good idea of the likely evolution of the storm impact, with any changes from here on likely to be associated with the exact position of the snow banding and the snowfall amounts under the band assuming stronger or weaker frontogenesis.

Assuming the typical model biases are at play, with the NAM over-amplified and exaggerated with the northward extent of the snow banding, with the GFS slightly under-amplified and under-estimating the northern snow gradient, along with an approximate accounting for potential snow ratios, the current forecast is for 6 to 12 inches of snow over most of the tri-state area, with lower amounts towards eastern/central Long Island and amounts towards the higher end of this range near and west/north of NYC. Locally higher amounts over 12-14 inches cannot be ruled out under the heavy snow banding.


Predictability Issues

There were some issues with the development of the system over the last several days, with the 18z GFS from 3 days ago, shown above valid on Thursday morning, depicting virtually no low pressure off the East Coast. Even when the GFS did begin to depict this low pressure development, it was highly inconsistent from run to run, and the model guidance is only now beginning to settle towards a consensus with as little as 48 hours of lead time left.

This is not the first time such a scenario has occurred this winter; just last month, the model guidance similarly failed to correctly predict a nor’easter affecting the region on January 7th until just a few days before it occurred, and discrepancies regarding how far inland snow will spread continued up until the day of the storm, when the heavy snow band set up farther north/west than modeled. Similar cases have occurred in the last several years as well, and were associated with a similar upper-level flow consisting of a shortwave trough quickly progressing through the US with a broad ridge downstream.

In this case, the GFS ensembles properly depicted a large spread of scenarios over the last 2-3 days, some which depicted the solution that the model guidance is currently converging towards. In some cases, however, properly identifying the potential for such a short range amplification trend is hindered by the under-dispersiveness of model ensembles, when the majority of the ensembles depict a very similar solution to each other which is ultimately incorrect, and fail to depict a larger spread of possible scenarios which may capture the scenario that does ultimately occur.

This case illustrates the idea that it is important not to just use the current model output verbatim, but to consider trends in the model guidance, as well as account for spread in the ensemble members, and the potential ways the forecast can change if the upper-level flow shifts in any particular way.

Leave a Reply