While industry and agriculture belched greenhouse gases at an increasing pace through the 20th century, global temperature followed a jagged course, surging for 3 decades starting in 1915, leveling off from the 1950s to the late 1970s, and then resuming its climb. For decades, scientists have chalked up these early swings to the planet’s internal variability—in particular, a climatic pacemaker called the Atlantic Multidecadal Oscillation (AMO), which is characterized by long-term shifts in ocean temperatures. But researchers are increasingly questioning whether the AMO played the dominant role once thought. The oceanic pacemaker seems to be fluttering.
It is now possible to explain the record’s twists and turns almost entirely without the AMO, says Karsten Haustein, a climate scientist at the University of Oxford in the United Kingdom and lead author of a new study published this month in the Journal of Climate. After correcting for the distinct effects of pollution hazes over land and ocean and for flaws in the temperature record, Haustein and his colleagues calculated that the interplay of greenhouse gases and atmospheric pollution almost singlehandedly shaped 20th century climate. “It’s very unlikely there’s this ocean leprechaun that produces cyclicity that we don’t know about,” Haustein says—which means it is also unlikely that a future cool swing in the AMO will blunt the ongoing human-driven warming.
Others aren’t convinced the “leprechaun” is entirely vanquished. “They are probably right in that [the AMO] is not as big a player globally as has sometimes been thought,” says Kevin Trenberth, a climate scientist at the National Center for Atmospheric Research in Boulder, Colorado. “But my guess is that they underestimate its role a bit.”
The AMO arose from observations that sea surface temperatures in the North Atlantic seem to swing from unusually warm to cold and back over some 20 to 60 years; the ancient climate appears to have had similar swings. Researchers theorized that periodic shifts in the conveyor belt of Atlantic Ocean currents drive this variability. But why the conveyor would regularly speed and slow on its own was a mystery, and the evidence for grand regular oscillations has slowly been eroding, says Gabriele Hegerl, a statistical climatologist at the University of Edinburgh. “Those are harder to defend.”
The new skepticism kicked off with work led by Ben Booth, a climate scientist at the Met Office Hadley Centre in Exeter, U.K.. In 2012, he reported in Nature that pollution hazes, or aerosols, began thickening the clouds over the Atlantic in the 1950s, which could have cooled the ocean with little help from an internal oscillation. In the past year, several independent models have yielded similar results. Meanwhile, most global climate models have been unable to reproduce AMO-like oscillations unless researchers include the influence of pollutants, such as soot and sulfates produced by burning fossil fuels, says Amy Clement, a climate scientist at the University of Miami in Florida.
Now, it seems plausible that such human influences, with help from aerosols spewed by volcanic eruptions, drove virtually all 20th century climate change. Haustein and his co-authors tweaked a relatively simple climate model to account for the fact that most pollution originates over land, which heats and cools faster than the ocean—and there’s much more land in the Northern Hemisphere. And they dialed back the cooling effect of volcanic eruptions—a reasonable move, says Booth, who is not affiliated with the study. “We’ve known models respond too strongly to volcanoes.”
The also adjusted the global temperature record to account for a change in how ocean temperatures are measured; during World War II, the British practice of measuring water samples in buckets gave way to systematically warmer U.S. readings of water passing through ships’ intake valves. Past efforts to compensate for that change fell short, Haustein and his team found, so they used data from weather stations on coastlines and islands to correct the record.
As input for the model, the team used greenhouse gas and aerosol records developed for the next U.N. climate report, along with records of historical volcanic eruptions, solar cycles, and El Niño warmings of the Pacific. Comparing the simulated climate with the adjusted temperature record, they found that internal variability could explain only 7% of the record. Instead, soot from industry drove early 20th century warming as it drifted into the Arctic, darkening snow and absorbing sunlight. After World War II, light-reflecting sulfate haze from power plants increased, holding off potential warming from rising greenhouse gases. Then, pollution control arrived during the 1970s, allowing warming to speed ahead.
It’s a compelling portrait, but it could have been substantially different if the team had used other, equally justifiable assumptions about the climate impact of aerosols, Booth says. Trenberth thinks the team’s adjustments had the effect of fitting the model to an uncertain record. “There is considerable wiggle room in just what the actual record is,” he says.
Haustein disputes that the team tailored the model to explain the 20th century warming. “All we did was use available data in the most physically consistent way,” he says. The researchers ran the model from 1500 to 2015, and he says it matches paleoclimate records well, including Europe’s Little Ice Age.
If a grand ocean oscillation isn’t shaping climate, a future ocean cooling is unlikely to buy society time to address global warming. But the demise of the AMO also might make it easier to predict what is in store. “All we’re going to get in the future,” Haustein says, “is what we do.”