This Nor’easter turned out to be perhaps one of our most challenging winter events since the “Tax Day” Nor’easter of 2007. There was virtually every precipitation type and often rapidly changing precipitation types within minutes of one place and another.
Up to ten days prior to the event, the ECWMF model indicated a nor'easter of sorts would impact the northeast. While the track varied a little in later runs, the ECWMF always indicated a storm close to or even along the coast, while the GFS had a much weaker solution and further east. This forecast scenario was similar to what happened in the pre-Thanksgiving Storm with the ECMWF tracking a classic nor'easter and the GFS had it going out to sea up until about 3 days prior to the event. The ECMWF model did better with the evolution and track in the long run with that storm.
This was the second major snowstorm of the season in which the ECMWF caught onto a potential storm much earlier than any other model guidance. The GFS starting began to agree with the ECMWF solution within three days of the event.
The set up for the storm was interesting (and promised this would be a challenging forecast). A strong short wave was forecast to dive south from the polar jet, into the Great Lakes and then sinking further south into the eastern Ohio valley by late Monday. At the same time, a weaker system in the southern stream was forecast to develop off the southeast coast, and track northerly along the Eastern seaboard. To the north, a large high pressure became anchored near Nova Scotia. The position of this high would produce lots of low level moisture from an easterly flow.
All the longer range forecast models had these synoptic players, but up until three days out, the GFS kept the northern trough open and more progressive, allowing the southern stream storm to slide east of our region. The Canadian Model actually had this feature hammering our region before the GFS did, but kept the trough from closing off and never "capturing" the storm so it would exit very quickly.
As it panned out, the upper air low never really "captured" the southern storm. There was actually never complete phasing. As the upper level system proceeded to become a cut off upper level low, due to the blocking high well northeast, the surface storm was forced to slow down as it approached Long Island. None of the models forecasted the exact track correctly. The surface storm tracked up the coast to the east of NYC by late Tuesday. Instead of tracking inland, it actually remained along the coast, and was off the coast of Cape Cod by late Wednesday, then tracked northward into New Hampshire and weakened due to becoming fully occluded.
The failure of the exact track of the storm was part of the challenge with the forecast. The bigger challenge however, was the thermal profile of the forecasted soundings, right "on the fence" in whether we were going to have freezing rain, sleet, snow or just plain rain. Turned out, most areas saw all these types of precipitation, sometimes within minutes of each other!
The depiction of the thermal profile leading up to the event was wildly inconsistent from every source of model guidance. The overall thermal profiles, magnitude and height of the warm nose(s), and resultant precipitation types/transitions changed drastically with each 6-hour run of the NAM and GFS and every 12-hour run of the ECMWF. This made forecasting timing of transitions between precipitation types and resultant snow/ice accumulation forecasts extremely challenging.
The BUFKIT profiles were fairly poor with the heavy snow band during the late afternoon and early evening on 9-10 DEC from the NAM/GFS. The guidance once again had difficulty resolving an isothermal set-up, when latent heat of fusion such as cooling the column down due to the volume of the precipitation and melting of the snowflakes. A mixture of rain and snow became very heavy wet snow with aggregating snowflakes 1-2" in diameter. The BUFKIT profiles at 12Z and 18Z 9 DEC showed a rain to snow transition well after dark for KALB/KGFL.
As it turned out, when precipitation reached our region very early Tuesday, there was enough of a warm nose between the H850 and H925 level to cause precipitation to start mostly a light wintry mix (freezing drizzle and sleet). This resulted in a very thin sheet of ice for many places by the morning drive on Tuesday. Surface temperatures were below freezing around 30 degrees. However, there was down sloping which prevented light freezing rain from occurring east of Albany Tuesday morning on 9 DEC. Eventually, this light freezing rain spread north and east...and transitioned to sleet which justified the 12Z 9 DEC KALY sounding which fit a sleet profile.
Enough warming allowed surface temperatures to rise above the freezing mark in the Hudson Valley, turning the precipitation to liquid rain as far north as Glens Falls as well as out through the Mohawk Valley. However, snow occasionally mixed north of the Mohawk river, as well as all of the higher elevations to allow some accumulating snow in some sections with rain in others. Rain, freezing rain, and snow fell across the higher terrain around the Hudson valley with icing up to a quarter of an inch, which seemed to vary from one locality to another. One example, it was raining and 37 degrees at Pittsfield while at the same time, Averill Park was reported as an ice skating rink.
The early portion of the storm featured a down sloping easterly wind (at least a few thousand feet off the ground) which initially prevented heavier precipitation from working into the Capital region. That scenario changed however, by midday Tuesday, as moderate rain worked into the Capital region from the south, with wintry precipitation in the form of freezing rain, sleet, and a little snow across the higher terrain. This was due to the low-level winds become more northerly, eliminating any downslope.
By late in the day, it looked as the Capital region had escaped any snow accumulations. At this point, radar reflectivities were in the 30-45 dBZ range, indicative of moderate rainfall (or in some cases freezing rain and or sleet) with the melting layer just above the ground. Then, a line enhanced reflectivity /red color curve/ values (50+ dBZ values) formed and moved into the Capital Region at the tail end of the evening commute /between 2200-2300 UTC/. Upstream there was no clue as to what was going to happen, as the echoes actually dampened westward in KBGM's area. This was the frontogenetic band which strengthened as it headed into the Capital region. Precipitation was heavy and became "convective". Much like the 4 October 1987 "freak snowstorm", melting snow flakes (latent heat of fusion/ cooled the column (a combination of thermodynamic and dynamic cooling from aloft), virtually pinching off the shallow warm nose. As result, a burst of heavy rainfall, mixed with snow, changed to all snow between 500-600 PM. Huge flakes fell and snowfall rates approached an inch per hour in the valleys, and likely over two inches per hours over the adjoining hill towns. A "surprise" extremely heavy snowfall fell hard enough to produce accumulations on all surfaces despite surface temperatures in the valleys remaining just above freezing! Roads became treacherous and there were myriads of traffic accidents. Like the 4 OCT 1987 storm, the liquid to snow ratio was unusually low, likely close to 4:1! Fortunately with no leaves on the trees, power outages were not nearly as widespread, but there were plenty of them, mainly in the hill towns, but locations like Brunswick and Troy had extensive conifer tree damage. Also, people on Facebook were reporting power outages in Windham County, VT, Saratoga and Washington Counties.
The batch of incredibly heavy wet snow ultimately produced 1-4 inches of snowfall in the valleys of the Capital region, and 5 inches to locally over a foot across much of the hill towns! However, some of the totals exceeding 5" were due to the E/SE upslope during the morning into the afternoon /i.e. Southern Green Mountains, Helderbergs and eastern Catskills/. Accumulations varied immensely from town to town due to the convective nature. Areas south of the Capital District escaped this band, and areas well north and west saw it weaken a lot before it got to them. After the heavy band moved through, a dry slot worked in dramatically reducing precipitation rates. At this point, the clouds lost a lot of their ice crystals, so a switch back to light rain and drizzle or freezing drizzle took place in many areas again!
This heavy snow on Tuesday evening was accompanied by lightning and thunder. Lightning detection network recorded lightning strikes. 00z/Wed sounding shows very steep lapse rates just below a drying mid-level mentioned below. High DBz values partly from melting/sticking together of snowflakes and partly from convection.
It should be noted that the water vapor loop nicely showed the nose of the mid and upper-level drying due to the coupled jet streak aloft moving across Ern NY and Western New England around 00Z 10 December when the "snow bomb" was occurring. The enhanced drying in the water vapor loop was associated with the left exit region of an equatorward approaching jet streak at 250/300 hPa, and the departing right entrance region of the poleward jet streak. There was a pronounced area of upper level divergence over Eastern NY and Western New England.
Interesting at this point through early Wednesday, when high clouds worked over the low ones, the precipitation would switch to snow, and then back to drizzle, or a freezing drizzle in a few places. Surface temperatures in most cases bounced back to the mid 30s in the valleys, hovering around freezing over the higher terrain. This phenomena is possibly a seeder feeder process.
The Wednesday morning drive was much better than the previous evening drive. Most roads were just wet.
The next part of the storm began late Wednesday morning into the early afternoon. As mentioned by Wednesday morning the surface storm, not all that strong around 990 MB, was off the coast of Cape Cod. Meanwhile the upper level storm, associated with the northern stream energy was pinching off just to the south of the Capital region. By Midday Wednesday, jet energy associated with another short wave, and low pressure off of Cape Cod, produce a trowal like band of snow that developed across the Berkshires, and initially extended northwestward into the Saratoga region, just skimming northern Albany county. Precipitation, which has been intermittent light snow or drizzle, changed back to all moderate snow (with some riming of the flakes). Ground truth indicated snowfall rates increased to an inch per hour in this band as it expanded southward into the Capital region and out to the Mohawk Valley as once again impacted Wednesday's evening drive home.
Tuesday evening the band displayed most of the characteristics of James Kenyon's Mesoscale snow bands similar to what is described in research from CSTAR IV. It met most of the characteristics of a pivoting band based on Kenyon’s conceptual model. With the pivot point in or near the Capital District it further enhanced the snowfall rates to 1 to 2 inches/hour peaking at the midpoint of the evening rush hour. In addition the NAM showed along the X/Z cross section perpendicular to the thermal gradient, a classic tilted 2D Frontogenesis signature just below maximum omega, and impressive -EPV, strongly indicative of a strong mesoscale banding event (similar to research by Novak during CSTAR I).
Special weather statements were issued to cover this band. A late low end Winter Storm Warning was hoisted across the Capital region, northern Berkshire County and the Lake George Saratoga region. A second winter warning was issued for the Greens of southern Vermont (the first had been cancelled).
This band last into the early evening hours, finally lifting further northwest and weakening as it did so. Still snow fell most of the night in the Adirondacks, and this is when they got the brunt of their snowfall. This band produced anywhere from an additional 4 to 8 inches of snowfall from the northern Catskills, Mohawk Valley, Capital region, northern Berkshires and southern Vermont. A little more periodic light snow fell across most areas overnight.
By Thursday, we were done with the worst of the storm. However, the upper level low was still parked right over the Capital region, still producing deformation bands of light to moderate snowfall that rotated mainly north and west of the upper air low, including the Capital regions and points to the north and west. The surface storm was ill defined at this point, and located in New Hampshire. Maximum snowfall rates were generally half an inch or less.
This was a multi hazard storm. While freezing rain and snow were the biggest challenges, we had hydrology issues too. Rainfall from the previous weekend has swelled rivers a bit in the southern portions of our HSA. Mostly liquid rain fell over southern area, 1-3 inches, enough to produce some minor flooding on the Still River at Brookfield. Stevenson dam was only a few inches short of flood. A flood warning was issued but just fell short of it. Many other points got close to bank full but did not exceed them on the Housatonic.
The greatest ice accumulation was reported over the Berkshires with about a third of an inch followed by the heavy wet snow resulting in some power outages. Snow fell across areas mainly north of the Mid Hudson valley, as well as the Catskills. The greatest two-day snowfall totals were reported in Elka Park Greene County and Duanesburg in Schenectady both with two feet. Many other elevated areas reported well over a foot of snowfall. Snowfall amounts were generally in the 6-12 inch range across valley areas from the Lake George Saratoga region south to the Capital region and westward through the Mohawk Valley. Snowfall amounts dropped off quite a bit south of the Capital region...generally averaging 3 to 7 inches in Columbia County...and coating to two inches in Dutchess County. The hills of Litchfield County saw 1-4 inches. Savoy in Berkshire County reported 17 inches, but most other areas saw considerably lighter amounts in the 3-8 inch range.
It was stressed prior to the onset of this storm that this event would be a "lower than normal confidence forecast" due to the many the potential precipitation types and how the two systems played out. This was emphasized again and again in our many Area Forecast Discussions and Conference calls.
2. What was learned from this event:
The Operational European Weather Forecast Model (ECMWF) continues to be the "model" of choice in the 3-6 day time frame. However, the GFS model in my opinion is usually just as accurate or maybe a little more in the 1-3 day time frame. Usually in the 0-24 hour range the meso-scale models are of course the most accurate but not always. HJ With this storm there was not one model that outshined the other in the near term.
The SREFs and GEFS were woefully poor during this event. The plumes kept flip-flopping run to run, even during the storm.
The entire event showed how poor the guidance can be in terms of consistency. The individual models were not consistent for a consensus, but also run to run model inconsistencies occurred with the guidance (i.e. NAM, GFS, ECMWF, etc). The thermal profiles varied a lot too.
When thermal profiles are very close to 0°C and guidance flip-flops between precipitation types, we should expect a wide variety of precipitation types. We should also expect the unexpected, and convey our uncertainty in the forecast. Also, we should make sure we have an extra person or two on shift to handle frequent updates and phone calls.
Mixed precipitation events are extremely difficult at best to resolve in the gridded database and if we updated the grids for every spotter report we received, the forecast would never get out or would be outdated in minutes. Graphics derived from our forecasts for briefings and social media posts can only provide so much detail so the worded descriptions are most important. How do we convey complicated changeable weather in the current world of texting, sound bytes and abbreviated language?
Part of the winter weather advisory was canceled at 415 pm TUE for the immediate Capital District, then north-central Taconics, and southern Berkshires since most of the precipitation was rain or a rain/snow mix. In hindsight, this advisory could have been kept up for part of this area, if the 2-6 inches of heavy wet snow was anticipated. SPS statements were used for the heavy wet snow and thunder snow situation initially. Once snow tallies started to approach 3 inches or so, it was decided to place these some of the canceled advisory areas back into an advisory /immediate Capital District and northern Taconics until 6 am Wednesday. These locations had 1 to 5 inches of snowfall.
"Blocked lows" are much different storms than the typical Nor'easter. They usually hang around longer than a traditional snowstorm due to being cut off from the main flow, may have features such as trowals which can bring warmer air aloft in from the east or north. They storms are usually a lower confidence event as compared to other storms. When seeing a blocked low in the guidance, it's important to convey a greater uncertainty to the EMs/public and to discuss how the forecast is subject to change. Late Feb 2010 is another good example of the wacky potential of blocked lows. It's also important to note that blocked lows can bring very heavy snow to high terrain areas, but often are featured by strong gradients over short distances and can be difficult to forecast.
The total liquid precip of this event was actually very impressive. Any winter event with over 2.50"+ liquid is absolutely insane. It'd be interesting to look back at liquid content in multi-hazard cool season events, but this one was certainly one of the bigger ones in the last 10 years. Hugh mentioned the tax storm; I think the liquid content reminded me of the the 2008 ice storm in terms of extreme cool-season precip. The extreme amount of precip was hinted that by the NAM and SREF, but only very very close to the event start time. With those models bouncing around so much, it was tough to have confidence that it would actually occur
A daily precipitation record was broken at the Albany International Airport on December 9th with 1.60 inches; old record was 0.97 inches set in 1973.
On December 10th, a daily snowfall record was broken at the Albany International Airport with 6.1 inches replacing the old record of 4.0 inches set back in 1903.
The final snowfall total of 10.9 inches placed it just outside the Top 10 Greatest December Snowstorms (#10 had 13.3 December 24-25, 1978).
Above: Loops of heights, wind barbs and observations at a) 8500 hPa and b) 500 hPa.
Above: Heights, wind barbs and isotachs (shaded) at 300 hPa valis 12Z 10 December.
Above: Loop of MSLP.
Above: Loop of soundings from Albany, NY (KALB).
Above: Sounding from 00Z 10 December from Albany, NY (KALB) proximate to thundersnow.
Above: Radar reflectivity with cloud to ground lightning overlayed from KENX at 2328Z 9 December.
Above: Final snowfall map.