Roger Smith winter forecast 2017-2018

[Roger Smith]

I've been reading various long-range forecasts posted on this forum and while I see certain points of agreement with my outlook, I think it's fair to say that I am expecting a rather different outcome towards the end especially. 

My methodology is to find analogues in the data set I have compiled for Toronto and then extend out the analogues from whatever data existed for other parts of North America (as Toronto goes back to 1840, this leaves some gaps in the far west especially). 

There is a theory behind my technique which basically has a foundation in proposed interactions between variations in the Solar System Magnetic Field (modulations of solar wind, basically) and earth's atmosphere. I have attempted to focus on the usual climate signals that most forecasters use in their work but I continue to assume that these signals will be partially if not totally predicted through the technique (in other words, they are built into the analogue sets that the method employs). The one exception to that would be solar variability but I feel that solar activity levels are not that good an indicator year to year, they have more utility in decade-to-decade forecasting. We are clearly in a period of reduced solar activity compared to the 20th century but with long lag times of response as shown with the Dalton minimum, and year-to-year variability also shown in that era, I am not basing any of my forecast on low solar activity. 

The first point that I considered as a cross-check against the index values was that the autumn saw a rather rare pattern of cool September, warm October and cold November(or so it appears). The best analogues for that include (in order of correlation) 1932-33, 1879-80, 1854-55, 1894-95, 1858-59, 2007-08, 1995-96 (yeah baby), 1951-52, 1920-21, 1882-83, 1971-72, 1903-04, 1900-01, 1950-51, 1947-48, 1910-11, 1914-15, and 1871-72. This is not meant to imply that these specific signal analogues are the best analogues overall, but from the overall forecast derived it would appear that 1995-96 and 1971-72 may be leading analogues, cannot be as sure about 1854-55 with no western data. 

Forecast overview

It is expected that anomalous ridging will develop near the Pacific coast with a cut-off low at the base of a sharp trough expected to lie north-south around 95W. The east coast will be in a mild southwest flow much of December and part of January with potential for a rapid collapse of that in favor of an extended central to eastern trough with a depressed jet stream. This will allow the current November cold pattern to fade out over the east in December, perhaps not so extensively that occasional coastal storms bring moderate snowfalls, but eventually a mid-December to early January regime of strong southwest flow from Texas to the Great Lakes will develop and it could be very mild at times east of the storm track with frequent snowstorms of blizzard intensity for the Midwest and plains states. During all of this phase, very cold air will dominate western Canada and the interior U.S. west, sometimes pushing as far south as northwest Mexico. 

Some time around Jan 5-10 this pattern will shift slightly east but will also see a lengthening out of the trough so that it still drops well south into Texas but then has a secondary low height anomaly near Maine and New Brunswick. This will force the storm track to run mostly offshore but with periodic lake cutters in the mix. The very cold air in western Canada will make more and more progress in detached arctic highs that track into the northeast U.S. The intensity of the western Canada cold will create numerous opportunities for snow and freezing rain storms through December and January in cities from Portland to Seattle and Vancouver. California can expect a winter of frequent low elevation rainstorms, mountain snows and periodic freezes between the storms. Even Arizona may not entirely escape the unseasonable western cold and Phoenix may see a bit of snow for the first time in several years, heavy snow is likely in Flagstaff and parts of northern New Mexico.

This expanding trough will bring a stormy January and a very cold February to many parts of the eastern U.S. and Ontario-Quebec, and a milder than average trend in eastern NS and NL. There will be persistent blocking over Greenland and a mild southeast flow into the eastern Canadian arctic at times, remnants of these milder frontal systems will drop south as weak arctic reinforcements across Hudson Bay and northwest Ontario.

There is potential in all of this for one or two large snowstorms to hit the northeast U.S. and the mid-Atlantic states. One energy peak around Jan 2 to 4 may be more of a transitional period and could see the last of the late December regime's mild southwesterly events, but with any luck the transition could be underway enough to make this an east coast snowstorm event. Perhaps more likely is that it would be a Midwest snowstorm followed by the full blast of wintry weather across the Great Lakes into the northeast. Then the best opportunities for east coast snowstorms would appear to be around Jan 16-18 and Jan 28 to 31. By February, I expect the storm track to be offshore so that places like s.e. VA and Delmarva to eastern New England may cash in with their biggest snowfalls of the winter.

Some very cold outbreaks can be expected by February and with an active lake effect month in January snow pack should be heavy in upstate NY and northwest New England which will promote intense radiational cooling in these strong arctic highs. 

For Europe, I am expecting a somewhat bland start to the winter followed by an increasing tendency to cold and snow through January then a gradual return to the bland and mild scenario in February. 

The main point of uncertainty, I believe, is how long into January the very mild regime persists before the transition to very cold, some analogues have seen it persisting well into January (for example 1950), but the overall index values show that the trend is sharply downward by January 7th to 10th. The Jan 1996 snowstorm hit a somewhat later energy peak (from full moon and northern max) but despite the 3-4 day forward displacement of that peak this winter, it may produce, odds are that it will be more of an inland event with the Midwest and Great Lakes in a better location to cash in on the thermal gradient expected around then. I would expect the "big storm" of the eastern winter season to be one of the two later January peaks. Energy peak analysis for this winter favors intense storm development at times with some 5-7 day intervals of low energy indicating good potential for strong arctic high pressure formation west-central regions and several days maintaining over the eastern regions. So especially towards the end of winter, I am looking for a classic pattern of coastal storms separated by long cold spells. 

 

[raindancewx]

http://www.breadandbutterscience.com/Weather.pdf

You can get an idea of rare/unusual weather events from that document. If I remember correctly, 1856-57 was an exceptionally cold and harsh winter in the West, probably similar to 1914-15 and 1972-73. Snowed in Mexico City in January 1857 (has happened since, 1888 & 1997 I believe). In 1858-59, the temperature got to -32 in Ft. Crook Maine

I like 1932-33, as it a cat 4 hitting TX and PR - like this year - only year since 1850 for that to happen. Also, a warm AMO era, neutral PDO era, similar solar as you pointed out. I have a regression for Albuquerque that looks at hot days (87F or more) and AMO temps in JAS, and solar conditions, and it works pretty well for winter forecasting.

In the United States during the winter of 1871-72, the temperature at Camp Baker, Montana fell to -53° F (-47.2° C) in December. The temperature at Madison Barracks (Sackets Harbor, New York) fell to -44° F (-42.2° C) in December. The temperature at Fort Ellis (Bozeman, Montana) dropped to -53° F (-47.2° C) in January and February.113

In the United States during the winter of 1882-83, the temperature at Spokane Falls, Washington fell to -28° F in January. The temperature at Lancaster, Wisconsin fell to -41° F in 1883. The temperature at Olympia, Washington fell to 9° F in January. The temperature at Fort Buford (near Williston, North Dakota) fell to -46° F in January and -40° F in February. The temperature at Pike’s Peak, Colorado fell to -37° F in January. The temperature at Fort Washakie, Wyoming fell to -27° F in December. [This is the only U.S. Military Outpost named after an American Indian - Chief Washakie of the Shoshone tribe.] The temperature at Roseburg, Oregon fell to 12° F in January. The temperature at Salt Lake City, Utah dropped to -20° F in January. The temperature in San Francisco, California fell to 36° F in January. The temperature at Red Bluff, California fell to 19° F in January. The temperature at Dubuque, Iowa fell to -26° F in January. The temperature at Fort Grant, Arizona fell to 10° F in January. The temperature at Fort Elliott, Texas fell to -12° F in January. The temperature at Saint Vincent, Minnesota fell to -38° F in February. The temperature at Fort Assiniboine (near Havre, Montana) fell to -47° F in February. The temperature at Fort Benton, Montana fell to -41° F in February. The temperature at North Platte, Nebraska fell to -29° F in February. The temperature at Omaha, Nebraska fell to -24.9° F in February. The temperature at Cheyenne, Wyoming fell to -28° F in February. The temperature at Mount Washington, New Hampshire fell to -34° F in March.113, 126

[raindancewx]

The idea of incredible cold shots into the West has already verified to some degree - places in Mexico are seeing some of their earliest to top-five earliest on record snowfall this week. Places Southwest of Brownsville, TX, at low elevations (<2000 feet) will likely see snow Thursday/Friday. My high is supposed to 37F on 12/7, which is pretty cold here, although it does happen a few times at least each winter.

[Roger Smith]

Major cold intrusions starting already into central regions, looks now like my timing was too gradual on the development of broad troughing by January, and this opens the door to a major east coast snowstorm with that energy peak Jan 1-3. 

This could turn into a classic cold winter pattern for eastern-central regions if it holds. With 1933-34 as one analogue, severe cold was interrupted in January but both December and February had record lows. Perhaps that will be a similar outcome, but looking at this major energy peak Jan 1-3 as an important foundation for the main part of the winter season following.