Heavy Wet Snowstorm of October 4, 1987

On October 4, 1987, an unusual early-fall snowstorm affected eastern New York, western Massachusetts, western Connecticut and Vermont. For areas east of the Hudson River in Renssalaer and northern Columbia counties, the heaviest snow occurred between 09 and 17 UTC (5am to 1 pm EDT). Snowfall rates of 3"/hr were common in the areas of heaviest snow. The heavy snow began a little earlier further south. The first batch of heavy snow occurred between 22 UTC 3rd (6 pm) and 10 UTC 4th (3am) over the Catskill Mountains where up to 21" of snow was reported. Another area extended from northeastern Westchester county, NY northward into Renssalaer county NY and then north-northeastward into southwestern Vermont. Even locations near the Hudson River at lower elevations received heavy snow. In fact, 10" of snow fell at 300 ft. elevation about 1 mile east of the Hudson River in Troy. Adjacent areas of western Connecticut and western Massachusetts also received very heavy snow. This is the earliest documented major snowstorm to affect eastern New York in recorded history. The event was particularly devastating since trees at lower elevations (below 2000 ft) were fully leaved. Snow collected on the leaves and the resultant weight brought down many limbs and caused several hundred thousand residents to lose power for up to a week or more. Heavy rain preceded the snow, with some locations receiving 1 to 1.5". Despite the warm ground temperatures and soaked ground, the snow still accumulated rapidly. The often repeated myth that snow cannot accumulate with warm ground temperatures is clearly refuted in this case. In other words, if the snowfall is heavy enough, it will accumulate!

Snow Documentation

After stumbling onto this event during my research of the October 9, 1979 heavy wet snowstorm, I decided that proper documentation needed to be done while the event is still in the memory of area residents. In addition to storm documentation by the National Weather Service and snowfall data from the eastern New York Weather Observing Network (graphical form), strategic phone calls were made to the affected area including town historians, local fire departments, libraries and even local residents. Although the final product is not complete, this has already resulted in a more refined snowfall map for the storm. This is currently in the form of amounts plotted in google maps. Keep in mind that most of these values were maximum snow depth measurements and not snowfall. A large map of snowfall totals with states and counties will also be developed in the near future. Unfortunately, there was a large area in eastern NY, western CT and western MA with very few NWS coop stations that measured snowfall. In fact, there were no coop stations from extreme southeast Dutchess county to Valatie in northwest Columbia county (62 mile stretch). In fact there were no coop stations that reported snowfall across a large part of eastern New York, western Massachussetts and western Connecticut. In addition, snowfall amounts were underestimated at several of the coop sites including Valatie since some of the obervers did not check the snow depth very often. A smoothed version of the snowfall map can be found here.

Prior Meteorological Research

A nice research paper was published in Monthly Weather Review in 1991 by Lance Bosart and Frederick Sanders (Bosart and Sanders 1991). LaPenta (1988) also provided some insight into this rare storm.

Meteorological Setting

A strong shortwave trough progressed southeastward from central Canada into the Great Lakes and Midwest by October 2. This allowed low level cold air from northern Canada to spill into the Great Lakes region. The 500mb trough became neutrally tilted by 12 UTC October 3 over the Middle Atlantic region. The system underwent strong baroclinic development between 00 UTC 4th and 18 UTC 4th and lifted north-northeastward in negative tilt fashion. Strong surface cyclogenesis occurred, with the central pressure of the surface low dropping from 1008 to 990mb. By 13 UTC October 4th the surface cyclone began the occlusion process.

The low level cold air originated over Hudsons's Bay and adjacent central and northern Canada on October 1. This 850mb chart from 12 UTC 1st shows 0C and colder isotherms in blue and +3C and warmer isotherms in red. At this time, there was a large pocket of cold air with temperatures below -10C over the central and Northern part of Hudson's Bay. The 850mb baroclinic zone associated with this cold airmass was situated over southern Canada and was starting to charge south with the passage of a strong shortwave trough dropping south-southeasward from Canada. The 500mb chart with 850mb cold front overlayed shows that the 850mb front was situated immediately ahead of the upper trough. This chart was analyzed to the best of my ability given the paucity of upper air data. A key upper air site at The Pas, Manitoba did not report at 12 UTC. By 00 UTC 2nd, the southern edge of this cold airmass was charging south into the northern Plains. The 850mb cold front was still just ahead of the intense shortwave trough. By 12 UTC2nd, the cold front stretched from Joplin to Evansville to Cleveland and was still located just out in front of the deep upper trough. By 00 UTC 3rd, the front was moving through western PA and eastern TN and was still out ahead of the strong shortwave trough. The upper trough and front had slowed considerably by this time By 12 UTC 3rd the front stretched from Atlanta to Greensboro to east of Washington, DC to west of Albany, NY and into Quebec and was still located just ahead of the upper trough. By 00 UTC 4th the front had slowed dramatically in the northeast but was still charging eastward off the eastern seaboard. The front stretched from east of Hatteras to New England. A wave was forming off the NJ coast by 00 UTC 4th. The 500mb chart with the 850mb low and front overlayed shows that the upper low was still well removed from the low level circulation. However, by this time the upper trough was neutrally tilted as it approached the Atlantic seaboard, with strong vorticity advection undoubtedly becoming situated over the low level circulation, leading to the beginning stages of baroclinic development. By 12 UT4th the front had swept well off the Middle Atlantic coast, but was still anchored over New England where cyclogenesis was occurring. The 500mb chart with 850mb features overlayed shows that the upper low center was now closed and almost coincident with the 850mb low. This marked the end of the rapid deepening phase of the low level cyclone. The occlusion process started just after 12 UTC 4th. The entire progression 500mb and 850mb features can be found in this loop.

Cold advection obviously played a role in changing the rain to snow in eastern NY and adjacent states. By 00 UTC 4th, the rain had aready changed to heavy wet snow in the Catskill Mountains above 1500 ft. The 00 UTC 850mb chart shows that the 0C line was right through Albany. The 850mb temperature at 00 UTC over the Catskills was probably around -1C. However, the precip did not mix with snow in Albany until 07 UTC and didn't change to snow completely until around 09 UTC. This is because the elevation of Albany is around 200 ft. So the layer of above freezing air was just too deep to get snow at the surface. Locations south of Albany along the Hudson river and just west of the Hudson river experienced northwest downslope winds off the Catskill Mountains throughout the event. This had a warming effect on the low-level thermal profile and prevented these locations from receiving snow. However, east of the Hudson River the low-level flow had an upslope component and the rain changed to heavy snow by 7 UTC east and northeeast of Poughkeepsie and by 09 UTC further north . Even though surface temperatures were coldest where it was snowing heavy and actually warmer upstream, cold advection was still occurring leading up to the event. 950mb wet bulb temperature values were analyzed from 12 UTC 3rd to 12 UTC 4th in order to illustrate this. One can readily see that the low level airmass upstream from Albany was indeed drier. As this drier and cool air advected southeastward, strong evaporational cooling, along with strong low to mid-level ascent, helped cool the low levels and change the precip to snow. At 12 UTC 3rd, the 950mb wet bulb temperature was 7.7C at Albany, +0.6C at Buffalo, -1.0C at Maniwaki, Quebec, -4.5C at Sault Saint Marie and -4.5C at Waskaganish, Quebec. Obviously there was cold advection occuring over upstate NY. By 00 UTC 4th the 950mb wet bulb temperature had dropped to 3.6C at Albany, 0C at Buffalo, and -1.8C at Maniwaki with continued northwest flow. In fact, the northwest flow was in the process of increasing dramatically due to low-level cyclogenesis to the southeast. As a result, the 950mb wet bulb values over eastern NY cooled strongly between 00 UTC and 12 UTC 4th and were down to -1C at Albany, -1.5C at Buffalo, -4.8 at Maniwaki, Quebec and even 0.4 at Atlantic City, NJ. The 950mb level was about 1300 ft above Albany at 12 UTC 4th. Since above freezing layer near the ground must generally be less than 2,000 ft to get snow, it makes sense that the precip was all snow at Albany before 12 UTC. It is interesting to note that the rain changed to snow in the Poughkeepsie area south of Albany first since the cold advection cooled the column in this area before if did further north due the orientation of the low-level isotherms (cold air was sweeping in behind the system). Sanders and Bosart (1991) proposed that melting of falling sowflakes through the above freezing layer near the ground actually cooled the column and was the primary mechanism in changing the precipitation to snow. It is impossible to tell which of the 3 processes (cold/dry advection and subsequent wet-bulbing, lift or melting) played a greater role in changing the precipitation to snow, but I personally believe based on the evidence presented above that the former process was very important.