When snow advances quickly across the Eurasian continent during October and covers a large area, Cohen says, it heavily favors a cold winter in the Eastern United States.
I emailed Cohen to ask him about the behavior of the snow cover this fall and what exactly it means for winter. His answers are below, lightly edited for length:
Talk to us about how Eurasian snow cover behaved this fall.
This year the advance of October snow cover in Eurasia (south of 60 degrees north latitude) was the fastest on record going back to 1998.
The snow coverage (for all of Eurasia) was fourth highest on record.
What do you think the snow cover behavior predicts for the winter?
I would say that the predictors that we follow are strongly indicating a colder-than-normal winter.
Our U.S. forecast is posted below (and also available on the National Science Foundation website):
Any sense as to where the harshest winter conditions compared with normal will focus?
First, let me say I think the most impressive cold will be across Eurasia. But here in the U.S., extensive Eurasian snow cover favors colder than normal temperatures in the Eastern U.S. more than it does in the Western U.S.
Based on the recent evolution of the polar vortex, however, I do think that initially the most severe winter weather could be across the Western U.S. Also, La Niña has me thinking that the worst of the winter weather will be focused farther north (in a relative sense). Therefore, I expect the worst of the winter will be across the northeastern quadrant of the U.S.
Siberia has been extremely cold in recent weeks while the North Pole has been at record-warm levels. What do you think this signifies?
Most of the climate community believes that the cold in Siberia or elsewhere around the Northern Hemisphere continents is not attributable to a warm Arctic. I strongly disagree and believe that the two are dynamically linked. I also believe a warm Arctic favors a weak polar vortex. When the polar vortex is weak, cold air usually spills out of the Arctic to lower latitudes.
In the fall, the weak polar vortex results in more severe weather (snow and cold) across Siberia that expands to other regions during the winter, usually first to East Asia but later to Europe or the Eastern U.S.
You’ve noted on Twitter there are signs that the polar vortex might split. When might that happen, and what would that mean for weather in the U.S., and when?
Any disruption to the polar vortex usually precedes an increase in severe winter weather but, in my mind, when the polar vortex splits, this is the single greatest atmospheric phenomena that promotes blockbuster snowstorms in the subsequent weeks.
Recent model runs are no longer predicting a significant polar-vortex split but rather a weaker one in the near term. Regardless of what occurs in the near term, my thinking remains that there will be at least one more significant weakening of the polar vortex this winter and possibly a polar-vortex split.
How much snow do you predict for the D.C. area this winter?
I ran our model and it predicts 23 inches for this winter. [Normal snowfall in Washington is about 15 inches.] Snowfall forecasts are extremely difficult, and I provide the forecast in the spirit of trying to push the envelope and for fun. With that said, I was very pleased with last winter’s forecast. [Cohen predicted 28 inches and the actual total was 22.2 inches]
How confident are you in your outlook?
Every winter has its own unique challenges.
The forecast for a cold winter depends on correctly predicting a significant weakening of the polar vortex typically in midwinter (January).
The polar vortex is currently weak, but I do believe that the success of our forecast may depend on a subsequent polar-vortex weakening later this winter. For status updates, follow my blog and Twitter account.
Your winter outlooks are a bit controversial. For example, a forecaster at the National Weather Service said, “Some of Cohen’s mechanisms seem a little bit challenged” and that your methods aren’t “actionable.” What is your response to this criticism?
For the past seven winters, if you used a simple prediction of a cold winter when October Eurasian snow cover was above normal and a warm winter when October Eurasian snow cover was below normal, the forecast was correct six of those winters. The only winter that was wrong was last winter with the record strong El Niño. That is an impressive track record for climate prediction. But I am happy for the community to be skeptical of our ideas and create a void that I am willing to fill.
Some people have pointed out your forecast for last winter didn’t work out that well and that it shows your method doesn’t take enough factors into account. How do you respond to such criticism?
No forecasting system will be perfect but I would argue that we take into account more factors than other long-range forecasters, including the National Weather Service.
First, we do include El Niño and La Niña (ENSO) in our model. The source of our error last winter wasn’t simply because snow cover. Part of the cooling our model predicted in the Southeast and Mid-Atlantic, which turned out to be wrong, was reflecting the record strong El Niño.
Second, the models used in seasonal forecasts predict ubiquitous and incessant above-normal temperatures. There is no variability in those model forecasts. Our forecasts show much greater regional and interannual variability.
More winter outlook reading:
More from the Capital Weather Gang on Cohen’s work