After three frigid nor’easters in less than two weeks, even the most devout prophet of climate change could be forgiven for echoing the sentiment that President Trump tweeted a few months ago, to much ridicule, shortly before one of the coldest New Year’s Eves on record: “Perhaps we could use a little bit of that good old Global Warming.” By this point, most people outside the White House understand that climate is not the same as weather—that climate is the forest and weather is the trees. Yet the global climate system is enormously complex, and there is, in fact, a lively scientific debate in progress about the relationship between human-caused climate change (especially in the Arctic) and the increased frequency of extreme cold-weather events across the United States—blizzards, bomb cyclones, and the dreaded wintry mix.

The debate began in late 2011, at a meeting in San Francisco of the American Geophysical Union, when an atmospheric scientist named Jennifer Francis gave a talk that electrified her colleagues. Perhaps, she suggested, the persistent outbreaks of extreme weather that people were experiencing in the Northern Hemisphere were connected to the colossal loss of sea ice in the Arctic—a region that, with increasing greenhouse-gas emissions, has warmed at double the rate of the rest of the planet. Up until that moment, while scientists knew that the Arctic was changing rapidly—since the nineteen-seventies, sea ice, snow cover, and glaciers had all declined dramatically, and the permafrost was starting to thaw—they did not think it had an influence on weather systems in the midlatitudes. Credit for that went entirely to the tropics.

Francis had first started pondering the idea in 2010, while circumnavigating the globe in a sailboat with her husband and their two children. Throughout her career, she had focussed on how global warming was affecting the Arctic, and after many months staring at the sea she began to wonder how Arctic warming was affecting the global weather system. On her return to New Jersey, where she is a professor at Rutgers University, she and her colleague Stephen Vavrus, a climate modeller at the University of Wisconsin, Madison, set about examining changes in the behavior of the polar jet stream since the early nineties. The jet stream—a continuous gale, as wide across as the Amazon, that blows west to east in a ring around the Arctic, between four and eight miles above Earth’s surface—acts as a border between low-pressure polar air to the north and high-pressure warmer air to the south. Its path and speed exert a powerful influence on weather in the Northern Hemisphere. When the jet stream snakes south over North America during winter, it brings extended periods of cold. When it snakes north, warm weather prevails in northern climes.

Francis had already noted, in 2009, that the jet stream was weakening, especially after summers when Arctic sea ice was at record lows. She proposed that Arctic warming was the culprit, positing that higher temperatures decreased the pressure gradient between the North Pole and the midlatitudes, sapping the jet stream of its energy. Building on that research, Francis and Vavrus found that, in the previous three decades, as the Arctic heated up, the jet stream’s amplitude—the distance it strayed north and south in its journey around the globe—had grown in the fall and winter by about ninety miles. This increased waviness, they suggested, meant that the eastward progression of weather patterns embedded in the jet stream could also be slowing down, causing, among other things, frigid blobs of polar air to linger over the East Coast and heat spells to persist in the West. “The jet stream is loopy, but the loops usually tend to move more quickly, zipping in and out,” Michael Coe, an earth-system scientist at the Woods Hole Research Center, in Massachusetts, told me. What Francis and Vavrus proposed, Coe went on, was “that in a slower-jet-stream world, you get these big, lazy loops that get stuck.”

Last week, in the journal Nature Communications, Francis and two of her colleagues published a paper offering further observational evidence for the hypothesis. Led by Judah Cohen, a visiting scientist at M.I.T. and the director of seasonal forecasting for the climate-risk company Atmospheric and Environmental Research, they looked back at meteorological records from twelve American cities, spanning from 1950 to 2016, and found that, when temperatures were anomalously high in the Arctic, extreme winter weather on the East Coast was between two and four times likelier to occur. The paper, Cohen said, does not attempt to explain why this might be. “Here we are just trying to establish that a warm Arctic really leads to a cold, more extreme winter,” he told me by phone from Boston on Tuesday, as the third nor’easter spun toward the coast of New England. “That was a big enough chunk to chew on.”

The evidence appears persuasive, but Martin Hoerling, a meteorologist at the National Oceanic and Atmospheric Administration’s Earth System Research Laboratory, said that it conflicts with other, equally persuasive findings. In the controlled computer simulations that he and others have run—removing sea ice from the Arctic, for instance, but keeping all other variables the same—conditions at the North Pole have shown a negligible effect on those in the United States. The only exception, Hoerling said, is an over-all reduction in wintertime cold extremes over the northern U.S., precisely the opposite of what Cohen and Francis found. “Some observationalists like to point to a variety of different extreme events and argue for a ‘weight of evidence, Hoerling said. “They may cite a cold wave over the U.S. in early 2018, a heavy rain event in California in 2017, or even several polar-vortex events over the U.S. in recent winters. But these are all consistent with occurrences that one expects in due course within a dynamic atmosphere.” In other words, as Coe put it, “you have a naturally varying system that does crazy things,” and the difficulty is in distinguishing the usual kind of crazy from the truly unprecedented kind. When I ran Hoerling’s criticisms by Francis, she said that removing sea ice in a model simulation does not account for every aspect of Arctic warming. Her hypothesis, she added, has more to do with how long a cold spell ultimately lasts than whether it breaks any temperature records. “Persistence is the whole key,” she said.

Whether the observationalists and the modellers ever agree on the Francis hypothesis—Hoerling said that to prove it would require “many decades” of real-world data—there’s no doubt that this March, at least, has come in like a lion. Meanwhile, the Arctic recently experienced temperatures forty-five degrees warmer than normal. Perhaps the debate, rather than the conclusion, is what’s most useful. “We tend to only see what’s out our own windows,” Francis said. “In fact, the weather in one place is connected to what’s happening all over the world.”