Read our COVID-19 research and news.

It's in there. Corals collected on islands in the central Pacific reveal that the strength and frequency of the climate phenomenon commonly known as El Niño is highly unpredictable.

Gary Meek/Georgia Tech

El Niños Are Highly Unpredictable

El Niño events are not as predictable as scientists thought. The frequency and strength of the ocean-warming climate phenomenon were more variable during the 20th century than, on average, during the preceding 7000 years, according to new analyses of climate records locked within ancient corals. The finding largely discounts the idea that certain long-term variations in Earth's orbit strongly influence the climate-maker, scientists say.

"This will be a shock for many paleoclimatologists," says Axel Timmermann, a climate scientist at the University of Hawaii, Manoa, in Honolulu who wasn't involved in the new research. "[These findings] are a stark contrast to their ideas."

El Niños are marked by substantially warmer-than-normal sea-surface temperatures along the equatorial Pacific. These events—along with their alter egos La Niñas, which are defined by cooler-than-average sea-surface temperatures in the same region—steer weather patterns across large swaths of the globe, baking some areas while drowning others. Together, these phenomena are called the El Niño-Southern Oscillation (ENSO). In modern times, El Niños occur once every 2 to 7 years; sometimes they're strong and long-lasting, and other times they're brief and mild.

To estimate the strength and frequency of ENSO in eras before weather instruments were widespread, scientists often turn to chronicles such as lake sediments, whose layers provide insight into precipitation trends in the surrounding region, or tree rings, which provide clues about temperature and precipitation during growing seasons. But some of the best climate records are locked in corals, says Kim Cobb, a paleoclimatologist at the Georgia Institute of Technology in Atlanta. That's because the ratios of oxygen isotopes incorporated into their calcium carbonate skeletons as they grow can help scientists decipher past trends in both precipitation and temperature on a month-by-month basis.

In the new study, Cobb and her colleagues looked at long-term variations in the ratio of oxygen-18 and oxygen-16 isotopes in samples of corals collected on Christmas Island and Fanning Island, which are both part of the Line Islands archipelago in the central Pacific. In that region, an El Niño causes both warmer sea-surface temperatures and increased precipitation, each of which results in lower concentrations of oxygen-18 in coral carbonates. "That makes the region a 'sweet spot' for climate reconstructions," Timmermann says.

First, the team used radioactive dating to estimate when each of 17 coral samples—some of which had been tossed onto the islands by strong storms or tsunamis—were alive and growing. Then, it sliced the corals in half and took tiny samples, spaced 1 millimeter apart, and chemically analyzed them. The corals chronicled climate in the region for intervals between 19 and 81 years that were scattered throughout the past 7000 years—an archive that more than triples the amount of climate data previously available from corals, Cobb notes.

According to the corals, long-term variations in ENSO magnitude and frequency of ENSO events during the 20th century were on average much larger than at any time known during the last 7 millennia, researchers reveal online today in Science. Nevertheless, Cobb says, there have been short intervals, particularly in the early 1600s, when ENSO variability was even stronger than it was last century.

It's not clear why 20th century variability in ENSO is so much higher than it was in previous centuries, Cobb says. It could be related to increased concentrations of atmospheric carbon dioxide, but that can't be proven with the limited amount of climate data now available, she notes.

But the new results do poke holes in the notion that long-term variations in the amount of sunlight striking the Northern Hemisphere in summer significantly affect ENSO variability. Previous studies have hinted that ENSO variability would have been reduced about 6000 years ago, but the new data don't reflect such a trend. The team's analyses suggest that if such changes play any role at all, it is a minor one.

"This [study] shows that ENSO is very variable, all over the map, and overturns the idea that ENSO was weaker during the mid-Holocene [about 6000 years ago]," says Julien Emile-Geay, a climate scientist at the University of Southern California in Los Angeles. Nevertheless, he says it's too early to say that ENSO is inherently chaotic and that its variability is totally divorced from external stimuli. Some previous studies have hinted that major volcanic eruptions or changes in solar output may play a role in driving ENSO variability, he notes.

Maybe in the next 10 years, field studies will glean more data from corals and help fill in the blank spots in current climate records, Timmermann says. That, in turn, could help climate scientists identify what factors, if any, drive the mysterious El Niño—factors that could be used to enhance current computer models of global and regional climate.