Long-range forecast for winter 2018-2019 -- large variations likely

[Roger Smith]

This is a summary of my long-range forecast for winter 2018-19. 

The El Nino pattern will be wrestling with an unusually cold eastern Canadian arctic for dominance, and the results will be large variations in most of eastern North America, a fairly typical mild western pattern for an El Nino event, but meeting resistance at times from the outer edges of the cold pool which is expected to be most persistent over northern Ontario and Quebec.

Consideration of some analogue cases assists with timing on the variations. For most of the eastern U.S., I think we'll see one cold spell in December, a much milder interval in the first half of January, then back to much colder in later January and early February. The pattern will then progress to coast-to-coast warmth with an early exit for winter. 

At times, the milder Pacific air masses will flood into central regions and in modified form out to the east coast. But whenever the cold pool intensifies, the mean polar front position will slide south across the Rockies to set up some periods of heavy snowfall there. During those episodes, low pressure will redevelop in north Texas or Oklahoma and then track towards the mid-Atlantic coast. These will be good periods for snowfall in the inland Mid-Atlantic and most of PA, NY and southern to central New England. 

Around late December, I expect that the western ridge will swell up over Alaska and Yukon and allow cut-off low pressure to form in the southwest states. This will lead to an interval of very mild weather in most of the south, southeast and eastern U.S. The storm track will then run from about New Mexico to Michigan to southern Quebec. Much above normal temperatures will accompany this phase of the winter. 

Then this mild regime will collapse as the western ridge subsides slightly and moves inland a bit further. That will lead to a weakening if not the end of the cut-off low and allow the storm track to drop back into the mid-south and east-central states. A period around Dec 21 will be one early opportunity for heavy snowfalls in the east. A second window of opportunity will occur some time in late January or the first week of February. After that, I think winter will quickly come to an end and spring-like warmth will prevail. 

Although some of these variations will cancel each other out, I think on the whole it will be a milder than average winter by the numbers, and snowfall will range from near average in most of the Midwest, Ohio valley and coastal eastern states, to slightly above normal in a zone from Wisconsin to Vermont. 

The results for the northern plains and prairies will be alternating periods of very mild and very cold weather with more snow than normal in some places and potential for one or two severe blizzards, especially around the first half of January. 

Record high temperatures are most likely around Jan 5 to 15. Record cold could be achieved later although I think it may be more like -10 F in places with -20 F daily records sort of a cold response. 

Parts of western Europe may have an unusually cold and snowy winter. 

[raindancewx]

I think you have some good ideas here. I would add, that in the SW, when we see huge amounts of rain (for us) in Aug/Oct, there is usually a wet March in the following year - so I'm starting to look at that now that the rains in the SW will end up above average region wide.

[Roger Smith]

After reading the other outlooks being offered, and considering how they might offer any insights into big picture and details of this forecast, I feel that these forecasts are consistent with my scenario with perhaps some caution suggested on the very mild turn in January being more of a stale Pacific type than a really strong torch ... I happen to remember the winter of 1963-64 in some detail because I started observing (my amateur weather station) then, in southern ON what I remember is a very cold and somewhat snowy December (after a warm autumn), a rapid change to quite mild weather most of January, an east coast snowstorm but no snow of any consequence around Toronto, a very sunny and dry February (record high sunshine amounts), then a rather wintry March and early April, followed by a very warm May. I'm not sure that I like the 1963-64 analogue all that much with the different lead in and set-up in Europe apparently heading in a somewhat different direction. But to some extent, I agree it's a difficult winter to find a really compelling analogue anyway. There may be one way back in the 19th century, 1826-27 caught my attention for high correlations of CET in 1826. But data for North America are quite patchy for that winter. 

Bottom line is to trust the research on the general trend and try to build a scenario from the more obvious large scale building blocks. 

So I continue to think it will be a winter in two parts for the northeast and mid-Atlantic, with good opportunities for snowfall in December and again by late January into February, fading by late February.

For postulated amounts, I will go with 15-20" DCA, 25-35" IAD and BWI, 35-40" PHL-EWR-NYC-LI, 40-50" coastal sNE and 60-80" inland central New England incl ORH over to Albany, 80-110" in northern New England, also generally near normal amounts of snowfall in the Great Lakes except for a few higher pockets in Wisconsin and south central Michigan, northern Indiana. Snowfall may then trend to somewhat below normal in the central plains and a bit above normal in the northern plains trending to below normal in Montana and southern Alberta.

Best opportunities for winter storms in the northeast will be around Dec 20 to 23, Jan 17 to 21 and Feb 1 to 4. An interval around Jan 3 to 7 may be stormy but too mild in the east, so that may be the big storm event for the Midwest this winter although the shape of the flow may include the Midwest in some of the other east coast peaks. 

[Roger Smith]

If you happen to be wondering if this early intense cold outbreak in the east has some predictive value, the answer is no ... as an independent variable.

I took the coldest readings on Nov 21st from Toronto's 177 years of data that I have stored on an excel file, and averaged out the temperature trends from today's date to the end of the following year. (readings of 8 or more degrees below average)

Within five days the cold signal was (on average) largely extinguished from the data, and all three of the winter months showed a trend that was more or less normal (on average of the 28 cases). Through the remainder of the following years the most notable anomaly was a warm mid April and a cold end to September of the following calendar year. Much of the rest of the trend curve stayed very close to average (indicating random scatter). 

The range of outcomes varied from the super cold January of 1977 to a very mild January in 1950. 

So considered "stand alone" this cold tells us nothing predictive about the winter ahead. 

If you are interested, this is the set of years with the lowest temperatures on Nov 21 (in chronological order) ... since 1841

1850, 1857, 1869, 1872, 1873, 1875, 1879, 1880, 1888, 1895, 1903, 1905, 1911, 1914, 1916, 1929, 1932, 1937, 1949, 1950, 1951, 1969, 1971, 1972, 1976, 1984, 1987, 2008, 2014

These averaged 11.8 degrees (F) below the average of the entire data set for Nov 21. By Nov 28th it was no longer below normal and oscillated around normal for the rest of the winter.

The coldest period that spanned Nov 21 to 26 was in 1880, with record lows on several days and an average more than 20 degrees below normal. 

The coldest December to follow any of these years was 1976. December of 1875 stayed very cold to 20th and flipped to very mild where it stayed all through Jan 1876. (Nov 30th 1875 was the coldest November day in the period of record).

The winter of 1869-70 produced the most snow of any winter at Toronto. However, 1932-33 was relatively snow free. 

There are several winters here we would not want to see repeating (if looking for snow and cold) ... these include 1875-76, 1905-06 (improved towards end), 1929-30 (record warm Feb), 1932-33, 1949-50, 1951-52 (okay in Dec 51), 1972-73 (not much snow), 1987-88 (mild). 

So in other words, don't place too much faith in the "early indicators" theme, means nothing by itself. It can lead to almost any sort of winter overall. 

 

[Roger Smith]

Looks to me that the "warm core of winter" concept has come and almost gone now, both start and finish a bit ahead of my predicted schedule (I had derived Jan 5-15 for the warm core from my analogues), mid-December to the 10th of January would have been a perfect fit. From Dec 13th to Jan 8th (27 days) the mean temperature at DCA has been nearly 8 deg above normal. There will be a 31-day interval from perhaps 11th to 10th that runs 7 deg above, quite a large "monthly" anomaly but the calendar months will both be considerably smaller (December was +3.2).

The warmest part of that was probably back around the Dec 21-22 event that was supposed to have been an east coast winter storm (turned out further inland and too mild). As I concluded in a discussion with some other forecasters elsewhere on this forum, the "torch" was a bit subdued being more of Pacific than Gulf origins, although looking at DCA actual data, it has been a very mild 25 day period that, had it overlapped a month on the calendar, might be causing more chatter in weather circles. Not every warm or cold spell gets equally recalled because of what I call the tyranny of the calendar (for a good example, see late February into early March of 1950, as cold as Jan 1950 was mild, but spread out in parts of two different months).

But I am reasonably happy with the pattern compared to the forecast, the El Nino has been its usual self in some ways but is clearly meeting with a lot of resistance from the early invigoration of the arctic this past few months, and Quebec has been the favored ground for arctic highs to reside longest. 

We should be moving into a much better pattern now for winter synoptics and so I would be looking at dates around Jan 19-21, Feb 3-5 and Feb 14-19 (by then likely two peaks) as best opportunities, although I'm hoping this Jan 12-13 event works out well for the MA crew.

I still expect at least near-normal snowfalls in many locations despite the very sparse amounts to date. And with a warm March in the forecast, it implies that I am predicting a snowy late January and February.

. 

 

[Roger Smith]

Still thinking that we are entering the best part of the winter for eastern states, one or more of these energy peaks seems likely to deliver good snowstorm events: Feb 2-4, Feb 12-15 and/or Feb 17-20. The predecessor events in the series have tracked mostly inland and delivered to regions like the eastern Great Lakes, interior New England, central PA, Ohio. With a bit of retrograde influence peaking in early to mid February, expect the storm track to loop further south ending up closer to the east coast. I am holding to my earlier prediction that a warm trend would then quickly develop into March and that March-April might be well above normal in temperature in many parts of eastern and central NA. 

[raindancewx]

My analogs had 1/20-2/10 as the coldest period in the Northeast US. The convection in the Indian Ocean now tends to precede warm ups in the US by 10-15 days according to a research paper, and it last showed up before the December warm up. The CFS has a pretty warm second week of February now, possibly in conjunction with this research.

The positive SOI in December (and maybe January?) favor warmth in the SE US to some extent. For March...I could see things getting a bit crazy in the West or Plains, but I agree on a warm up for the East. The El Nino Marches following an El Nino hurricane season with a major hurricane hitting the US Gulf Coast tend to feature severe cold in the Western half of the US...somewhere. Look at March 1942, 1958, 1966, 1970, 2005. Big NAO+ periods in October correlate pretty strongly to a cold Jan/Mar in the West, regardless of ENSO, and that looks fairly solid so far. Low solar El Ninos tend to be warm in the East in March too. I have my analog set for Spring, tentatively, but can't see what they look like yet with the NOAA/ESRL stuff down. I've been pretty happy with my blend of 1953-54, 1976-77, 1986-87, 1994-95, 1994-95, 2006-07 for winter to date in the US.

[Roger Smith]

Well, Feb 1954 was warmest on record in southern Ontario, at least back when I was still living there, might not be any more I suppose. March was near normal and it turned cold for a time in early April. The spring of 1977 saw a rapid reversal of the severe cold of January around 10th of February and it stayed generally near-record warm most of the spring. A similar but subdued pattern repeated in 1987 with the warmth peaking in the summer. I don't remember much of winter 1995 as we were packing to move but I think it was variable more than any one trend. And I was out west by 2007, would have to look that one up. There have been some similar storm tracks to 1991-92 this past winter also (frequently Ohio valley to eastern Lake Ontario). That was not a particularly memorable winter where I was living, all elements fairly close to average IIRC. Where does 1929-30 fit into your thinking? There was a noteworthy warm spell in mid-February of 1930 in the eastern and central regions. I know it was relatively high solar but frankly I tend to discount solar during long active periods (as to being high or low), it's only when you get a prolonged downturn that you see real correlations with temperature anomaly. 

[40/70 Benchmark]

On 10/24/2018 at 4:44 PM, Roger Smith said:

This is a summary of my long-range forecast for winter 2018-19. 

The El Nino pattern will be wrestling with an unusually cold eastern Canadian arctic for dominance, and the results will be large variations in most of eastern North America, a fairly typical mild western pattern for an El Nino event, but meeting resistance at times from the outer edges of the cold pool which is expected to be most persistent over northern Ontario and Quebec.

Consideration of some analogue cases assists with timing on the variations. For most of the eastern U.S., I think we'll see one cold spell in December, a much milder interval in the first half of January, then back to much colder in later January and early February. The pattern will then progress to coast-to-coast warmth with an early exit for winter. 

At times, the milder Pacific air masses will flood into central regions and in modified form out to the east coast. But whenever the cold pool intensifies, the mean polar front position will slide south across the Rockies to set up some periods of heavy snowfall there. During those episodes, low pressure will redevelop in north Texas or Oklahoma and then track towards the mid-Atlantic coast. These will be good periods for snowfall in the inland Mid-Atlantic and most of PA, NY and southern to central New England. 

Around late December, I expect that the western ridge will swell up over Alaska and Yukon and allow cut-off low pressure to form in the southwest states. This will lead to an interval of very mild weather in most of the south, southeast and eastern U.S. The storm track will then run from about New Mexico to Michigan to southern Quebec. Much above normal temperatures will accompany this phase of the winter. 

Then this mild regime will collapse as the western ridge subsides slightly and moves inland a bit further. That will lead to a weakening if not the end of the cut-off low and allow the storm track to drop back into the mid-south and east-central states. A period around Dec 21 will be one early opportunity for heavy snowfalls in the east. A second window of opportunity will occur some time in late January or the first week of February. After that, I think winter will quickly come to an end and spring-like warmth will prevail. 

Although some of these variations will cancel each other out, I think on the whole it will be a milder than average winter by the numbers, and snowfall will range from near average in most of the Midwest, Ohio valley and coastal eastern states, to slightly above normal in a zone from Wisconsin to Vermont. 

The results for the northern plains and prairies will be alternating periods of very mild and very cold weather with more snow than normal in some places and potential for one or two severe blizzards, especially around the first half of January. 

Record high temperatures are most likely around Jan 5 to 15. Record cold could be achieved later although I think it may be more like -10 F in places with -20 F daily records sort of a cold response. 

Parts of western Europe may have an unusually cold and snowy winter. 

Not a bad call.

[Roger Smith]

On 2/10/2019 at 8:58 PM, 40/70 Benchmark said:

Not a bad call.

Thanks, I don't feel terrible about this forecast by any means but there are things about it that have derailed and spun into the ditch. The problem for all of us who made a long-range forecast is trying to make sense of the chaotic pattern, I think when I was saying "variable" I was thinking of variations from type to type of regime but not on the frequency scale we have seen, with all the rather localized anomalies. 

Consider for example at my own location, we had a lot of very bland, dry weather that you expect to see here in an El Nino winter, a few episodes of snow and cold, then this past week has been much, much colder and now we have 15" of snow from a very minor event in the grand scheme of things that just parked over our valley and let go. 

So what kind of winter has this been here? Just about any call made by any forecaster would seem right at certain times and wrong at others. 

I have the feeling there's a lot of that happening all over North America. Yes, it was rather like El Nino to see a fair amount of rain in CA, AZ and into the Gulf states, but I don't think it was quite the classic El Nino. And also, the snow drought (now hopefully ending) in the northeast coastal regions was a fairly confined zone that just happened to park over half the population (and two thirds of this forum's membership), so a similar glitch over a less populated region might have gone unremarked upon altogether. 

So while some parts of this forecast have done alright, I am not very confident about the end game as described. I won't make excuses about the SSW event because my philosophy is that events like that should be buried in the research methodology as consequences, so it would be disingenuous to claim that it had derailed the forecast in any way. What's more likely true is that the El Nino has weakened to such an extent that I don't think there's a lot of heat energy available to build strong west coastal ridges that will feed mild to warm Pacific air east now, once we get to the end of this current blocking episode. The patterns in the past week have been so singularity-oriented (jet stream south of Hawaii, 20" snow in Seattle after nothing to end of January, Pebble Beach covered with sleet), that I am scrambling a bit to adjust thinking to this rather challenging new pattern and basically, I think what's most likely to happen now is some sort of huge distorted southward push of linked troughs with cut off highs forming at mid-high latitudes, and where did I see that before? 

1993. Hmmm. Also some similarity in current pattern to Feb 1888. Hmmm. 

The plot thickens. I honestly believe we're going to see a historic major storm of some kind in the next month and more likely in March than late February. Even with that, it may not be cold all the time either. A mixed up synoptic soup is on the menu.