It is also important to extract potential impacts and threats from individual historical events because at times the impact guidance based on the top 15 analogs may not be representative of the forecast. Therefore it is good practice to compare and contrast the most representative analogs to the forecast in order to learn the subtleties of historical events and find a range of impacts for the upcoming forecast. For example, in the animated image below there are three 4-panel maps, the 60-h forecast from the GFS run mentioned earlier in this post and two similar historical events: 7 December 2003 and 25 October 2005. You can see that all three are similar in strength/position with similar upper-level features (e.g, closed circulation at 500 mb).
The animation of the GFS forecast (below) shows how the system forms off the Carolinas, moves north along the East Coast, and then stalls/weakens off New England.
In early-season winter storms, precipitation forecasting can be a challenge since cold air has not had time to build up across southeastern Canada. However, a cold rain can turn into a heavy wet snow as dynamical cooling can overcome the lack of cold air. The following animation is a 4-panel plot of thermal fields from the GFS forecast and shows that most of the low-level thermal fields are above freezing to the northwest of the surface cyclone.
When analyzing the mass fields, the 7 December 2003 analog is very similar to the GFS forecast. A surface low develops off the NC/VA coast and quickly intensifies as the midlevel and upper-level support approaches. As the cyclone strengthens at 850 mb, the -U wind to the north of the cyclone also strengthens...similar to the GFS forecast.
Although the mass and wind fields develop and transition in a similar fashion to the forecast, an important difference between the two is the presence of a large pool of cold air across southern and southeastern Canada. Being a month later than the upcoming storm can make a huge difference and temperatures with the December 2003 analog are easily 6-10 C colder at 850 mb.
The consequence of this is the difference between a higher elevation snow event and a widespread snowstorm across all the Northeast. In the December 2003 analog, widespread snowfall greater than 12" occurred from New York City northeastward to north of Boston and inland through all of interior New England. Unless the temperature forecasts change dramatically, a high-impact event similar to this one will not occur.
The other analog very similar to the GFS forecast is from 25 October 2005 (below). This is an interesting analog because prior to the analog time of 20051025/1200, a strong upper-level low was moving east along the Ohio Valley and actually absorbed extratropical storm Wilma off the East Coast...very similar to what occurred late in October with Hurricane Sandy. The main difference between Wilma and Sandy was that midlevel blocking was not present across eastern Canada in October 2005 so the extratropical storm accelerated north northeastward...much like some nor'easters behave.
What is interesting is that the October 2005 analog has a very similar low-level thermal profile to the northwest of the surface low as the 60-h GFS forecast. Shown below, temperatures at 850 mb are between 0 and -4 C while for the most part staying above freezing at the surface.
These warm temperatures were offset by strong dynamical cooling as heavy snow occurred in higher elevations from West Virginia to northern Maine.
There were many power outages where the heavy snow fell. The following is an excerpt from NCDC Storm Data on the storm from New York.
An early season snowfall hit parts of upstate New York on October 25th, 2005. Low pressure developed over Ohio and tracked to off the New Jersey Coast where it intensified into a major early season winter storm. Snow fell mainly over the higher elevations of central New York and the Catskills, with mainly rain in the valleys. The rain caused minor flood problems in low-lying areas near Waverly along the Susquehanna River. The hardest hits areas were from the higher terrain of southern and eastern Broome County, north and east through the higher terrain of Chenango, western Delaware, eastern Cortland, southern Madison, Otsego and southern Onondaga counties. Between 6 and 12 inches of heavy wet snow fell above 1400 feet elevation in these areas. Highest amounts were found in southern Madison County and in Chenango County at 1800 feet where more than 12 inches of wet snow fell. Between 2 and 6 inches of snow fell between 1000 and 1400 feet in these areas. The snow came down very hard from the late morning into the early and mid afternoon hours on the 25th. The snow continued heavy at times until the mid to late evening hours on the 25th. The snow rapidly accumulated on roads and caused major traffic problems along interstates 81, 88 and route 17 in southern New York and in the Binghamton area. In addition, the heavy wet snow brought down trees and power lines in these areas. Over 10,000 customers were without power in Broome and Chenango Counties alone. Across the Central Southern Tier, and the Finger Lakes region, the valleys saw rain with the higher terrain, above 1600 feet, seeing only a few inches of snow. In the Catskills, the higher terrain above 1800 feet saw up to 6 inches of snow. The Syracuse and Utica areas saw little snowfall, with the Southern Tug Hill Plateau reporting up to a few inches.
This is just one example of why studying and extracting impacts from historical events that are similar to the upcoming forecast is important. Not all analogs are created equal (i.e., like ensemble members), by analyzing individual analogs poor analogs can be thrown out and the forecaster can hone in on a more accurate range of potential impacts.
The analog run that this blog post was written from is available here.
No comments:
Post a Comment