What we can and can’t say about Arctic warming and U.S. winters

It certainly feels like the northeastern United States is getting snowier.

In the first two weeks of March, three winter storms slammed into the northeast corridor from Washington, D.C., to Boston. Over the last decade, a flurry of extreme winter storms has struck the region, giving birth to clever portmanteau names such as Snowpocalypse (2009), Snowmageddon (2010) and Snowzilla (2016).

So what’s going on? Researchers have previously suggested that extreme weather in the midlatitudes might be linked to climate change’s impacts on the Arctic (SN Online: 12/2/11), particularly the dramatically decreased sea ice cover in the Arctic Ocean. And now, a study published online March 13 in Nature Communications reports a strong correlation between the severe winter weather experienced in the northeastern United States over the last decade and the warming trend in the Arctic.

Two of the study’s authors, climatologist Judah Cohen of the Massachusetts-based climate and weather risk assessment group Atmospheric and Environmental Research and atmospheric scientist Jennifer Francis of Rutgers University, have long been proponents of the hypothesis that the warming Arctic is having profound effects on weather at the midlatitudes (SN: 3/12/15), from severe snowfalls to heat waves. It was Francis who, with a colleague, proposed in 2012 that sea ice loss in the Arctic slows the polar jet stream, a band of air currents flowing above the northern and middle latitudes of Earth. The slowed jet stream would become wavier, with large meanders that might jut deep into the midlatitudes; such waves, the researchers suggested, could allow winter storms to push south and linger.

The idea was compelling, particularly coming as it did on the heels of two massive snowstorms in the northeastern United States. But many scientists were skeptical, suggesting that the underlying atmospheric dynamics — how dwindling sea ice in the far north might shape midlatitude weather — remained uncertain. Researchers have continued to wrangle over this question.

Warm Arctic, cold continent

The latest study arrived in the wake of not two but three winter storms that battered the northeastern United States. Cohen and Francis, with coauthor Karl Pfeiffer of AER, analyzed the frequency of winter storms striking 12 U.S. cities since 1950, including Seattle, Chicago and New York City, and found a persistent correlation with Arctic temperatures, particularly for the Northeast.

To do this, the researchers used two gauges of temperature anomalies. One, called the polar cap geopotential height anomaly, or PCH, is essentially a measure of troughs and ridges in pressure in the upper atmosphere — the high and low pressure zones familiar from television news. The other is the polar cap air temperature anomaly, or PCT, a gauge of temperature. Together, they describe variations in atmospheric heat and moisture, and how those translate into weather patterns.

The study’s finding supports what has been called the warm Arctic–cold continents hypothesis. But there’s still a missing piece, one that the authors acknowledge this observational study wasn’t intended to answer: what is driving those changes. Indeed, Arctic change in the last two decades has been profound — not just dwindling sea ice, but also noticeably increased precipitation, and thus snow cover, over Eurasia. Both factors may be linked to the observed weather patterns.

However, atmospheric teleconnections — long-distance links between climate phenomena, often spanning thousands of kilometers — are an exceedingly tangled web. Many factors related to greenhouse gas–induced climate change may feed into midlatitude weather anomalies. For instance, climate warming can alter the balance of heat between the Arctic and the tropics near Earth’s surface, which in turn can influence the jet stream.

Warm waters in the Gulf of Mexico can also feed extra energy into storms battering the U.S. East Coast, says Richard Alley, a climatologist at Penn State who wasn’t involved in the new study. Storm tracks, the strength and frequency of large climate patterns such as the El Niño Southern Oscillation — greenhouse gases can affect them all, and there just isn’t a lot of data yet, he adds.

Untangling weather’s web

Climate simulations intended to determine the influence of Arctic sea ice loss have not yet found the smoking gun. One recent study suggests that the problem may lie with the computer models themselves, which struggle to fully reproduce real-world conditions.

Another issue is that the conditions set within such models can vary widely. Cohen notes that previous models’ conclusions about Arctic influence on lower-latitude weather have been all over the map. “Whatever your point of view is, there’s a model to support it,” he says. He is a member of a large project getting under way this year, dubbed the Polar Amplification Model Intercomparison Project, that aims to address that issue by coordinating model conditions to try to tease out the Arctic’s impact. The first results from these simulations should be available within a year.

“I’m not sure what will come of it,” Cohen says. “People tend to believe their own models the most.” 

For now, the jury is still out on whether Arctic changes are the main driver behind the winter storms that have battered the northeastern United States in recent years. Yet even without a definitive answer, it may be worth keeping the topic in the news, if only to remind people that human-caused climate change is linked to cold snaps as much as to heat waves. Cohen notes wearily that, “Every time it gets cold, everybody’s making a joke: ‘I wish we had global warming.’”  

Regardless of what climate models find, investigating these long-distance links in weather could also pay off by improving risk prediction and forecasts. “We still might find that success is far off or impossible, but understanding the limits of predictability would also be valuable,” Alley says. “The problem is that, the very value of this research creates a public desire for answers and products before the research has had enough time to generate them with the level of confidence that we all would like. This is the inevitable tension of working on emerging science in an important field.”