Above: Sea ice is seen from NASA's Operation IceBridge research aircraft along the Upper Baffin Bay coast on March 27, 2017. (Mario Tama/Getty Images)
A number of research papers in the past decade-plus have explored the notion that sea ice loss in the Arctic is affecting midlatitude weather patterns in both summer and winter. As more and more research delves into this area, there’s been robust debate over the nature of the proposed links—and now a new study is extending the latitudinal stakes.
A paper published in the Proceedings of the National Academies of Science on January 27 argues that sea ice loss in the Arctic may be related to El Niño events in the central Pacific Ocean. The PNAS study—by physicist Charles Kennel, a former director of the Scripps Institution of Oceanography, and Scripps analyst/programmer Elena Yulaeva—explores the correlations between September sea ice extent and the evolution of El Niño events in the Pacific over the following months.
Kennel and Yulaeva analyzed historical data to identify which atmospheric phenomena are correlated with Arctic ice loss. Among these were trade winds in the equatorial central Pacific, which typically weaken during El Niño and strengthen during La Niña. Kennel and Yulaeva found a distinct change after 1998, focusing on two trends: the intensification of Arctic sea ice loss and a westward shift in El Niño events.
By comparing conditions for 1980-1998 vs. 1999-2015, they contend that rising air above open water in the Arctic in late summer may be triggering planetary atmospheric waves—broad weather features a few miles high—that propagate to the tropical Pacific by December, a time of year when El Niño and La Niña events often intensify.
El Niño events that peak in the central Pacific, which are sometimes nicknamed El Niño Modoki, tend to be less reliable in producing California rainfall as opposed to “classic” eastern-Pacific events, which typically bring large-scale updrafts and jet-stream energy closer to the state.
According to a Scripps news release, the PNAS study marks the first time that researchers have looked at the Arctic and central Pacific in this context. “There’s now a network of consistent correlations,” Kennel said in the news release.
In their study, Kennel and Yulaeva wrote: "These results add to the evidence that loss of Arctic sea ice is having a major impact on climatic variability around the world.” They add that “two limitations of technique prevent this paper’s results from being conclusive”—one involving the loss of statistical power by dividing the data set into small time periods, and the other involving the inability to test the model by using the model itself.
The Arctic and the tropics: which is pushing which?
The PNAS study might be seen as the first sign that Arctic sea ice loss—a clear outgrowth of long-term human-driven climate change—is affecting the tropical atmosphere. However, several researchers argue that it’s more likely that the vast stores of heat in the tropical Pacific are influencing the Arctic, rather than vice versa.
“I think they likely have it backwards, based on published evidence,” said climate scientist Gerald Meehl (National Center for Atmospheric Research, or NCAR).
Meehl and NCAR colleague Kevin Trenberth noted that the study did not directly analyze rising motion (convection) over the Arctic, instead using the extent of Arctic sea ice loss as a proxy. “There is no way that heating anomalies in the Arctic are large enough or stationary enough to force planetary waves,” Trenberth said in an email.
Also not ready to accept the Arctic-to-tropic thesis is James Screen (University of Exeter), who’s been active for years in studying midlatitude-Arctic relationships.
“Although the conceptual model proposed by Kennel is plausible,” Screen said, “one cannot ignore two possible alternative hypotheses: there is no causal link between Arctic sea ice and the tropics, and the trends have coincided in recent decades by chance, or the causality is opposite to suggested by Kennel et al, and the tropics are forcing the Arctic.”
Among the research arguing for a tropic-to-Arctic route is a new paper led by James Warner, a graduate student of Screen’s. “Recent dramatic changes in Arctic sea ice due to climate change have been linked to changes in weather patterns across the Northern Hemisphere. Many studies have proposed such links, but correlation does not necessarily imply causality,” Warner and colleagues said in the paper.
In fact, their study suggests, the observed correlation between late-autumn ice in the Barents and Kara seas and the North Atlantic Oscillation can be explained by internal atmospheric variability, and an observed link between Barents-Kara ice and the strength of the North Pacific’s Aleutian low is being driven by a third element: sea surface temperatures over the tropical Pacific.
“It's plausible there is a connection between Arctic sea ice and the tropics—and indeed, there is some modeling evidence to support this,” Screen said. He points to a 2015 paper led by Clara Deser and a 2018 paper led by Kun Wang, both from NCAR. Both papers found that the effects of Arctic sea ice loss may propagate through oceanic dynamics to produce a variety of impacts at lower latitudes over the course of decades.
The more controversial idea has been an atmospheric pathway—the notion of sea-ice loss affecting the midlatitudes and tropics through an atmospheric rather than an oceanic route.
Carbon Brief posted an excellent 2019 overview of research into Arctic-midlatitude climate connections. The post includes comments from Screen as well as Jennifer Francis, the Rutgers University researcher (now a senior scientist at the Woods Hole Research Center) who’s been among the leaders in delving into Arctic-midlatitude links.
Since 2012, Francis and colleagues have asserted in a series of papers that a weakened pole-to-equator temperature contrast fostered by Arctic warming is leading to a more variable jet stream and “weather weirding.” In a 2018 paper, Francis and colleagues found evidence at the surface and at upper levels for more-frequent long-duration weather features across North America, which could lead to enhanced impacts by their sheer persistence.
In an email, Francis noted two other recent studies that suggest a possible influence of Arctic sea ice on the tropics. A 2017 paper led by Ivana Cvijanovic (Lawrence Livermore National Laboratory) found model evidence that over a period of decades, Arctic sea ice loss may reorganize the planet’s energy budget, leading to reduced convection in the tropical Pacific and less precipitation in California. And a 2020 study led by Shangfeng Chen (Center for Monsoon System Research) outlines an atmospheric wave train from higher to lower latitudes that may explain why unusually high concentrations of sea ice in the Greenland-Barents Seas are correlated with El Niño–like conditions in the tropical Pacific a year later.
“While the tropics have historically been the dog that wagged the extratropical tail, the fact that three recent studies found a significant Arctic influence on low latitudes means that the huge changes in the Arctic may now wield enough muscle to wag the mighty tropics,” said Francis.
Meanwhile, Judah Cohen (Atmospheric and Environmental Research) has been issuing winter forecasts for more than a decade based on the idea that autumn sea ice loss in the Barents and Kara Seas leads to a more extensive early snowpack and a stronger, northwestward-displaced surface high in Siberia. In Cohen’s chain of events, this is followed by disruptions in low-level flow that propagate upward, a more variable stratospheric polar vortex, and a higher risk of midlatitude cold intrusions by winter. (See Cohen’s forecast for 2019-20.)
“I agree the direction of the wave train [from the Arctic to the Pacific tropics] is in question,” said Cohen in an email, referring to the new paper. He questions whether convection in the Arctic, which is likely less frequent and much shallower than typical tropical thunderstorms, could affect the high-level jet stream in a sustained way.
“The system is so noisy, and I believe it is a challenge to conclusively identify a signal that travels back and forth across the Northern Hemisphere and over many months and not involve the stratosphere in any way.”
Another potential mechanism at work is a resonance at the jet-stream level in midlatitudes—again supported by a weaker temperature gradient—that may be feeding into persistent, locked-in-place weather features.
Building on this line of research, launched in 2013 by Vladimir Petoukhov (Potsdam Institute for Climate Impact Research), Michael Mann (Pennsylvania State University) and colleagues published high-profile papers in 2017 and 2018 on what they termed quasi-resonant amplification. Based on model results in their 2018 paper, such events could increase by as much as 50% if carbon emissions continue to grow through the century.
In an email, Mann said he is not yet convinced that there is an atmospheric pathway leading from Arctic sea ice all the way to the tropical Pacific. Because the new study does not replicate and extend the proposed concepts within a global model, he noted, it’s too soon to know whether the Arctic-to-tropic chain of events is physically valid.
Mann called the concepts in the new PNAS paper “intriguing ideas but quite speculative, and mostly only conceptual and empirical in nature at this point. I would like to see more compelling evidence that these mechanisms can be seen in state-of-the-art coupled climate models before drawing confident inferences regarding the sorts of tropical Pacific/Arctic interrelationships argued for in the article.”
In an email, Charles Kennel said: “We too suspect that tropical events can create teleconnections to the Arctic, and that is one of the things we wanted to look at next. As far as whether Arctic events get to the equator, many people believe they are not strong enough, so we looked for evidence one way or another. Our conclusion is that we looked for the evidence the way we did and we cannot rule out an Arctic-tropic connection.”
Brian Donegan (weather.com) contributed to this report.
