Cohen, whose winter outlooks boast a 75 percent accuracy rate over the 15 years he’s issued them, has developed a model for predicting winter conditions that is strongly based on snowfall in Eurasia during October. When snow rapidly piles up in Eurasia at that time, it sets off a chain reaction that he says is linked to cold and often snowy conditions in the eastern U.S. during winter.
And this October, the snow in Eurasia increased quickly for a third straight year.
In many instances, the evolving weather pattern is reflected by a negative phase of the Arctic Oscillation, in which the polar vortex weakens, breaks up, and dives towards the U.S. supporting cold, and snowy conditions.
Cohen, who works for Atmospheric and Environmental Research, graciously shared his winter outlook ideas with us in an interview, provided below, which has been lightly edited.
Capital Weather Gang: Tell us what happened with Eurasian snow this October.
Judah Cohen: Snow cover extent was once again above normal this past October. It was a little less than 2013, which was the fourth highest since 1972, making this October the fifth highest ever observed since 1972.
We also compute the Snow Advance Index, which computes the rate of advance of snow cover at all latitudes below 60 degrees north for the month of October. The SAI was high but somewhat less than recent years of 2009, 2012 and 2014. All in all, it was a solid and consistent snow cover advance.
It is unprecedented to have three years in a row with such high snow cover extent values and, as a teaser, I think that it could portend a snowy Northeast season.
CWG: How do the October Eurasia snowfall trends influence eastern U.S. weather patterns?
JC: Our research has shown when Eurasian October snow cover extent or the Snow Advance Index are above normal, this favors a negative phase of the winter Arctic Oscillation (AO) or weakened polar vortex, most often in January.
I do expect in December and January a weakening of the polar vortex, followed by an extended period where the AO is predominantly in the negative phase.
CWG: Do you agree with the idea that it will be cool in the South and mild to the north (compared to average) as predicted by some models and the National Weather Service?
JC: In principal, I would agree with that general temperature pattern. However, I do believe that if the polar vortex is weak and the AO is negative in mid-winter, the cold in the Eastern U.S. will be more widespread than predicted by all other forecasts that I have seen.
CWG: Do you have specific ideas for the Mid-Atlantic?
JC: Our forecast model is predicting below normal temperatures for the Mid-Atlantic region for this winter and I am comfortable with that.
All forecasters are using the strong El Niño winters of 1982-83 and 1997-98 as analogs or paradigms for this upcoming winter and I don’t see why 2002-03 and 2009-10 couldn’t be analogs or paradigms as well. [Editor’s note: 2002-03 and 2009-10 were extremely snowy winters in the D.C. area].
I see no physical or plausible reason why if El Niño is moderate it necessarily forces a cold winter but, if it’s just a little stronger, it forces a warm winter. Sure both the winters of 1982-83 and 1997-98 were relatively mild but that is only a sample size of two. What justification is there to treat those two winters as there own special category?
CWG: Many forecasts out there call for a mild December for much of the Lower 48. Do you agree?
JC: I have seen this consensus for a warm December and, in the interest of full disclosure, our model is consistent with the consensus and is forecasting a warm December as well.
But I don’t have a lot of confidence in warm December idea. Much of it doesn’t make sense to me.
El Niño is supposed to force an overall mild pattern and the El Niño forcing is strongest in January to March period, so doesn’t it make sense then that December has the best chance of being cold with the rest of the winter mild?
Since the El Niño of 1997-98, I think all of the El Niño winters have been colder later and maybe this has helped shape the similar forecasts for this winter. But I think this is just a coincidence….
Our model is predicting a warm start to winter with a cold finish but that is consistent with the Eurasian-snow forced cold not arriving until sometime in January. As I have written, the snow signal does not usually arrive to the Western Hemisphere before January because it typically travels through the stratosphere to reach the U.S. Therefore, December tends to be more of a wild card month.
CWG: Because this El Niño is so strong, close to a record, could it overwhelm the effects of the Arctic Oscillation — like in 1997-1998 when the AO was negative but it was mild with little snow?
JC: This is a great question and one that I am finding a really tough challenge. Don’t think the 1997-98 scenario isn’t running through my head every day. El Niño did seem to triumph that winter and, of course, a repeat is possible. However, I do believe our climate is sufficiently different that a repeat of 1997-98 is less likely today.
The Arctic is much warmer today than it was just twenty years ago and with much less sea ice not only in extent but especially volume. This is a controversial topic but I do believe and quite strongly that a warmer Arctic is forcing colder winters in the mid-latitudes. In 1997-98, it was cold across much of northern Eurasia and even northern Canada. I think if we were to have a repeat of 1997-98, the much warmer Arctic of today would favor a southward displacement of the cold.
Again, I believe it is a mistake to dismiss more recent El Niños [than 1997-98] as paradigms for this upcoming winter. I believe this winter could have elements of 2002-03 and 2009-10, which were winters where the Arctic had already started to warm relative to the 1980s and 1990s.
But, at the end of the day, I let our model do the balancing and it definitely has a blend of El Niño and a negative AO (the El Niño we are sure about, the negative AO not so much). The model is colder than other winter forecasts I have seen and I think that the observed temperatures could be colder still.
CWG: What does your outlook signify for the California drought?
JC: I think that El Niño skews the odds for increased rainfall for California; my own statistical analysis bears that out. However, I certainly don’t take it as a given. I do think that there are other wild card players that could cancel out El Niño’s influence.
The eastern North Pacific sea surface temperatures are still really warm. There has been much talk that it is contributing to the West Coast ridging and lack of rainfall but I still think that the argument needs to be properly vetted. Also, I do think there is still a tendency for ridging in the North Pacific and if sets up in the wrong place I believe that it will continue to remain on the dry side in California. In weather I just don’t believe there are any sure things.
CWG: So what is your snowfall forecast for the Mid-Atlantic and Northeast?
JC: I am surprising myself by how bullish I am for snowfall for this upcoming winter in the Northeast given 1982-83 and 1997-98 seasons.
As I mentioned earlier, October Eurasian snow cover extent has been at unprecedented high levels the past three years. October snow cover extent is a surprisingly good predictor for snowfall in the Northeast.
I do believe that both Washington and Boston could have yet another above normal snowfall season.
Because of El Niño, I think this favors the largest normalized snowfall departures [differences from normal] in the Mid-Atlantic relative to New England. I feel that the atmospheric circulation favors yet another weakening of the polar vortex this winter and that vortex split is favored over a vortex displacement. In my opinion, a polar vortex split is the single most bullish atmospheric phenomenon that leads to blockbuster snowstorms.
For all those interested, you can follow the developments and forecasts of the polar vortex and AO all winter long at my Arctic Oscillation blog.
More winter outlook reading
More from the Capital Weather Gang on Cohen’s work