January 7, 2017 – Snow Forecast for Today

The second snow event to affect the region in as many days is currently unfolding over Virginia, where a rare blizzard warning is in effect along with one of the largest snow accumulation forecasts in recent years. Despite substantial forecast uncertainty over the last several days, there is a much clearer consensus now that at least some snow will affect the tri-state area.

This post discusses some of the factors behind the uncertainty in the forecast, and the latest analysis on the meteorological factors contributing to the snow forecast.


Current Analysis:

HRRR initialized 500-hPa vorticity at 0600 UTC (1am EST). The key feature to note is the shortwave trough over NW Mississippi.

It is always important in the forecasting process to not only consider the forecast model guidance, but to factor in current and past observations as well. The image above shows initialized 500-hPa vorticity and geopotential heights, with a shortwave trough positioned near NW Mississippi. This trough originated from a cutoff upper-level low over the Northeast Pacific Ocean several days ago which interacted with a northern stream trough over the Midwest and progressed southeast through the United States.

Initial uncertainty from several days ago was related to how much the cutoff low would remain detached from the northern stream upon re-entry into the United States; earlier runs such as the GFS depicted much greater separation, which in turn allowed the trough sufficient room to further amplify, leading to stronger forcing for ascent downstream and a resulting major cyclone producing heavy snow centered near the NYC tri-state area.

While the former scenario is off the table, the latest model guidance have trended towards a more amplified shortwave trough again, bringing accumulating snow back into the picture for the tri-state area. A preliminary subjective analysis of this trend appears to be the result of mostly subtle shifts in the modeled depiction of mid-upper level features, such as the strength of the ridge over the Atlantic which was under-estimated, as well as the amplitude of a departing trough over the Northeast and the position of an upper-level low over Canada. These subtle shifts resulted in a gradual trend towards a slightly more amplified system, and a more northerly track component once off the coast, thus spreading snow farther north than previously depicted.

 


Forecast Analysis

Forecast snow accumulation map for this event, issued at 3 AM this morning. An updated map may be posted on Twitter in the morning depending on any significant changes.

 

The current forecast continues to have an aspect of uncertainty to it, with the amplitude of the system influencing the start and end times of snow, the snow intensity, and accumulations. The overall idea at this time is for the highest probability of snow to occur during the early-mid afternoon hours, with starting time around the late morning and ending time around the evening.

Forecast accumulations, with lower than average confidence for a 1-day outlook, are less than an inch in interior NJ/NY, 1-2 inches in the north/west suburbs of NYC, 2-4 inches in NYC into western LI and SW CT, and 4-8 inches in SE CT & eastern Long Island, perhaps locally higher.

The next sections discuss the meteorological factors behind the forecast above and the areas of uncertainty.

 


Forcing for ascent

When forecasting precipitation intensity and amounts, factors contributing to upward vertical motion need to be considered, as strong ascent along with sufficient moisture results in condensation and thus precipitation. Precipitation amounts can be thought of as a function of precipitation rates multiplied by duration (P = RD; Doswell et al., 1996), and as duration of snow will be relatively uniform throughout the region, the axis of heaviest snow amounts will be associated with the highest snowfall rates lasting for a sufficiently long duration.

4k-NAM 300-hPa winds valid at 0900 UTC (4am EST).

In this case, there is widespread synoptic forcing for ascent across the region. The image above shows 300-hPa winds, with a powerful 170+ kt jet streak positioned over New England. The region from about Virginia into New Jersey is located near the right-entrance quadrant of the jet streak, with ageostrophic divergence indicative of ascent. In conjunction with strong cyclonic vorticity advection downstream of the trough, there is widespread mid-upper level forcing for ascent across the region, including the tri-state area.

HRRR forecast 700-hPa frontogenesis, valid at 1500 UTC (10am EST).

 

Lower-level forcing mechanisms, however, show a somewhat different picture. The 700-hPa trough remains relatively flat initially, only becoming more substantially amplified once it has passed the area. Mid-level frontogenesis, or the enhancement of the temperature gradient which is known to contribute to ascent and enhanced snowfall banding, is mostly maximized over southeast Virginia into Delaware and southern New Jersey, although the HRRR depicts a band of weaker mid-level frontogenesis spreading into the tri-state area on Saturday morning, which may provide sufficient lift for snow to spread into NYC and parts of northern New Jersey.

Given the forcing mechanisms in place, there is high confidence for the axis of heaviest snowfall banding to extend from southeast Virginia into southern New Jersey, possibly extending into eastern Long Island and SE Massachusetts. With this axis of strong upward vertical motion in place, there is often an axis of subsidence, or sinking air, inland of the banding, which with this scenario would be more likely to occur towards the I-95 corridor from DC into NYC. Given the inland subsidence, as well as dry air being advected from the north, a sharp snow gradient is likely north of the heavy banding, a feature which model-derived quantitative precipitation forecast (QPF) often struggles to fully resolve, especially with low-resolution models such as the GFS, ECM and CMC.

 

Snow Ratios: The next aspect to consider once creating a QPF forecast is the snow to liquid ratios. An average ratio is 10:1 (10 inches of snow melting to 1 inch of water), although these can vary significantly both spatially and temporally within a single storm system. While there is some correlation between surface temperatures and snowfall ratios, an important factor to consider is the dendritic growth zone (DGZ), or the layer in which temperatures range from -12 to -18 degrees Celsius. A deep, saturated DGZ layer overlapped with strong ascent is conducive for formation of dendrites, which have a low density and are thus efficient in accumulating at a ratio higher than 10:1.

Model-derived soundings depict a deep DGZ layer over the heavy snow banding, which combined with the previously noted strong ascent supports ratios higher than 10:1. While strong wind will be somewhat of a limiting factor, ratios of at least 14:1 to 18:1 are likely under the heaviest banding, which may contribute to snow accumulations near or locally over a foot of snow in Southeast VA into Cape Cod. Inland of the banding, despite a similarly deep DGZ layer, subsidence inland of the heavy banding may limit snow growth and result in lower ratios close to 10:1. This gradient in snow ratios implies that the gradient in snow accumulations will be larger than the precipitation gradient, which itself will likely be larger than the QPF gradient as previously noted. For these reasons, the snow forecast depicts a stronger snow gradient than most of the model forecasts.

 

Forecast Bust Scenario: No forecast is perfect. There is always some room for error, even in cases of high confidence, which is where incorporating probabilistic-based forecasts can help in cases of low-predictability but potentially high-impact events.

In this case, the main uncertainty is with the amplitude of the trough as it approaches the area, which will influence how far north/west the snow will spread. The current forecast is close to most of the model guidance, albeit with a sharper and slightly farther east snow gradient, although it is noteworthy to mention that the latest high-resolution models, particularly the HRRR, have been slightly backing off on the amplitude again. Although the snow map depicts 2″ of snow in NYC, there is a high likelihood of this total ending up anywhere between 1″ and 4″.

Additional short-term updates will be posted on the NYC Area Weather twitter account.

Leave a Reply