You are using an outdated browser.
Please upgrade your browser
and improve your visit to our site.
Skip Navigation

Explaining the Global Warming Hiatus

Grappling with climate-change nuance in a toxic political environment

Getty Images News

Even as scientists asserted an incontrovertible consensus on climate change, a funny thing has happened over the last 15 years: Global warming has slowed down. Since 1998, the warmest year of the twentieth century, temperatures have not kept up with computer models that seemed to project steady warming; they’re perilously close to falling beneath even the lowest projections.

Some people are playing the hiatus as good news: “Apocalypse perhaps a little later,” the Economist put it. But in a political environment where vast swathes of the American right reject even the premise of global warming—and where prominent right-wing pols suggest it’s an enormous fraud—this inconvenient news could easily lead to still more acrimony over the subject. Especially since scientists themselves aren’t entirely sure what the evidence means. If scientific models can’t project the last 15 years, what does that mean for their projections of the next 100?

It might seem like a decade-long warming plateau would cause a crisis for climate science. It hasn’t. Gerald Meehl, a Senior Scientist at the National Center for Atmospheric Research, has seen hiatus periods before. They “occur pretty commonly in the observed records,” and there are climate models showing “a hiatus as long as 15 years.” As a result, Isaac Held, a Senior Research Scientist at NOAA’s Geophysical Fluid Dynamics Laboratory, says “no one has ever expected warming to be continuous, increasing like a straight line.” Those much-cited computer models are composed of numerous simulations that individually account for naturally occurring variability. But, Meehl says, “the averages cancel it out.”

The phenomena that most clearly causes the Earth’s temperature to rise and fall are El Nino and La Nina. During El Nino, heat is brought to the surface of the eastern Pacific, raising global air temperatures. The reverse happens during La Nina. Conveniently, the “hiatus” is said to begin in 1998, when a historic El Nino produced the warmest year of the twentieth century. That starting point amounts to cherry-picking: The 2000s were warmer than the 1990s, and the nine warmest years on record have occurred since 1998.

Skeptical Science

But all this leaves a big question, one that scientists have been trying to answer: If the atmosphere is warming more slowly than projected, where did the heat go? 

There are two ways to create a global-warming hiatus: The heat can go somewhere other than the atmosphere, or there might be less heat in the climate system than scientists predicted.

Global warming estimates usually center on measures of air temperature. But  the climate system also can move heat from the atmosphere to somewhere else, like melting ice.

The most obvious culprit is the ocean, which absorbs 90 percent of the heat added to the climate system. With the oceans holding so much heat, the focus on mean surface temperature as the measure of global warming misses much of the point. Minor shifts between the oceans and the air could keep the planet heating up, even while slowing the pace of atmospheric warming. And that wouldn’t necessarily be good news, since warmer oceans would raise sea levels, change the climate, and hurt the ocean’s ecosystem.

But sea surface temperatures and the upper ocean heat content didn’t increase over the last decade, not by enough to account for what Trenberth called “missing heat”—the heat that greenhouse gas emissions should have trapped in the Earth’s climate system, but couldn’t be found.

Scientists speculated that the heat might be hidden in the deep ocean, beneath 700 meters, where there are not reliable temperature measurements. Without good data, they couldn’t prove heat was going into the deep oceans. So Meehl and his colleagues turned to computer models. Their approach was straightforward: Look at naturally occurring hiatus decades in climate models, and see where the computers put the heat. In the simulations, the deep ocean warmed by 19 percent more during hiatus periods, even as sea surface temperatures cooled in the tropical Pacific, creating a pattern similar to the prolonged La Nina-like pattern of the last decade. They then plugged in the conditions of the last decade and found that  “the models produced roughly 20 percent less warming than the free running models.”

More proof of the deep ocean hypothesis: Over the last decade, more than 3,000 Argo floats—basically, seaborne thermometers—were deployed across the world’s oceans to measure oceanic heat content. The floats can measure temperatures down to 2000 meters. They found that more than 30 percent of ocean warming over the past decade occurred beneath 700 meters—potentially covering most of the missing heat.

But analysis was limited by the absence of historical data: Without knowledge of how much heat was trapped in the deep ocean in the past, scientists couldn’t prove much. Just last month, an intriguing paper in the journal Geophysical Research Letters suggested a way around this problem. Using a model that projects ocean temperatures back to 1961, its trio of authors found that the surge in deep ocean heat intake over the last decade was indeed historic. 

But other scientists think that the heat is missing because it never made into Earth's climate system. The idea that heat might not have made it relates to the concept of “forcing.” The term refers to the forces that add or remove heat from the climate system. The best known example of forcing is the Greenhouse effect, where greenhouse gases in the atmosphere trap heat that might otherwise radiate into space. But there’s negative forcing, too—i.e., other pollutants that reflect energy back into space.

The sun itself is a major factor in forcing. Over an average of eleven years, the sun’s energy output ebbs and wanes, subtly influencing earth’s climate. The last solar maximum was in 2000, but a prolonged solar minimum has kept the sun even dimmer than usual. According to Kevin Trenberth of the National Center for Atmospheric Research, lower levels of solar radiation account for 10 to 15 percent of the hiatus.

Explaining the rest is more difficult. Susan Solomon, an MIT professor best known for research on the ozone hole, has focused on stratospheric water vapor and aerosols. Water vapor is a greenhouse gas, and satellite data shows stratospheric water vapor decreasing since 2000—meaning less heat is getting trapped.

Working in the opposite way is an increase in stratospheric aerosols, which counter greenhouse gases and serve as a giant parasol—reflecting energy away from the Earth. But even there, the theory means rethinking former givens. Until recently, scientists believed large volcanic eruptions were the culprit when stratospheric aerosol levels rose. But stratospheric aerosol levels have risen since 2002, even though there hadn’t been a large volcanic eruption since 1991. Some initially attributed the trend to increased coal burning from South and East Asia. The data hints at another explanation: a wave of smaller volcanic eruptions, previously assumed to be too small to contribute to stratospheric aerosol. Ground and satellite-based observations show a correlation between increases in stratospheric aerosol and this decade’s smaller volcanic eruptions—like Monserrat, Ruang in Indonesia, and Manam in Papua New Guinea.

What all of these discoveries hint at is that scientists, at long last, have developed a better understanding of year-to-year climate variations. In a way, you could think of it like the stock market. Watching Wall Street, we see the indices rise and fall, and we know the news that has influenced the swings. Watching annual temperatures, scientists could see the fluctuations but, until recently, knew little about the news--even though they were confident that increased carbon dioxide would ensure a bull market over the longer run. With an updated understanding of deep ocean temperatures and stratospheric aerosols, that has changed. Solomon thinks “we’ve learned a lot about interdecadal variability” as a result of the hiatus.

Piecing together the hiatus puzzle—the competing effects of the oceans, stratospheric water vapor and aerosols, and sunlight on the global climate system—is difficult. Scientists want an exact account of earth’s energy budget or balance, or how much energy enters the earth’s atmosphere from the sun, how much is absorbed by the oceans and atmosphere, and how much is radiates back into space.

Ideally, scientists could use satellites to measure how much energy is getting trapped inside of the climate system. But even the CERES satellite constellation, which is designed to do exactly that, gives estimates that can’t be reconciled with other data.1 And without an absolute measurement, Trenberth and other scientists are forced to measure from the bottom up how much energy is absorbed by the oceans or the air. Scientists have a good idea of how much energy is coming from the sun, and how much is absorbed by the atmosphere and the upper ocean. But here again, the problem of measuring oceans returns. Some scientists are content to rely on the Argo data, while others use climate models to estimate the net-energy balance, including an account of volcanic aerosols and changes in sunlight. Then they use CERES to measure changes over time.

All of which leads to very different estimates of how much heat is getting trapped in the Earth’s climate system. Using computer models, Trenberth estimates that the earth’s net-energy balance is about 50 percent larger than James Hansen, a prominent ex-NASA scientist who relies on Argo data.

The difference matters: If the Earth’s energy balance was actually a third lower than the computer models suggest, it would mean rethinking assumptions of climate science, such as whether aerosols are reflecting even more heat out into space than previously imagined. But Trenberth thinks the problem is Hansen’s data, not the climate science. Argo misses key things—the top ten meters of oceans and the seas around Indonesia, among other “huge patches.”2 After accounting for those differences, Trenberth thinks that his estimate "lines-up pretty well with the values from the ocean data.” Hansen says most of the difference is due to the solar cycle, as his analysis assumed the solar minimum, while Trenberth's estimate is for the entire decade. Adjusting for the solar cycle reduces the difference between the two estimates from 50 percent to 20 percent. 

Meanwhile, Trenberth’s model doesn’t directly account for changes in stratospheric water vapor or aerosols from small volcanoes, since those “effects should be included in the CERES top-of-atmosphere measurements.” Whether his approach accounts for these subtle changes in forcing depends on just “how good the CERES values are.” Solomon approaches the problem from the opposite perspective. She emphasizes that the role of volcanoes is “unambiguous,” supported by “great” satellite measurements, and multiple sources of data. As a result, she argues that stratospheric water vapor might cover roughly “20 percent of the hiatus” along with “30 percent from volcanoes”—the oceans, and whatever else, must cover the outstanding 50 percent.

Nonetheless, the combination of imperfect data, overlapping explanations, and continued uncertainty mean that scientists cannot discount the possibility that they have overestimated the climate’s “sensitivity” to additional greenhouse gas emissions. For Held, the last 10 to 15 years “make it more plausible that the size of climate response to greenhouse gas increase is on the lower side of what models have been projecting over the last 10 or 20 years rather than over the high side.” Held is not alone.

In the end, the so-called scientific consensus on global warming doesn’t look like much like consensus when scientists are struggling to explain the intricacies of the earth’s climate system, or uttering the word “uncertainty” with striking regularity. Nowhere is there more uncertainty than in the clouds. “It’s like cancer,” Held said, referring to the “many, many research problems” posed by the many kinds of clouds, each with their own special properties that might reflect or trap more or less of the sun’s heat. Some progress has been made on clouds, especially with cirrus clouds.

In the current political climate, debates about things like climate change are carried out in broad-brush assertions. The challenge for scientists is that the more they understand the climate system, the more complex it gets, and the harder it gets to model with precision—not to mention making the kinds of sweeping statements the news cycle requires.

Last summer, Mark Maslin and Patrick Austin wrote that climate science’s embrace of “more-complex processes means adding in ‘known unknowns,’ like the rate at which ice falls through clouds, or the rate at which different types of land cover and the oceans absorb carbon dioxide.” Known unknowns create uncertainty and a bigger spread between models. They fear that “this will look like the scientific understanding of climate change is becoming less, rather than more, clear.” That fear seems well-founded, especially as climate skeptics deploy the latest research, like Solomon’s on volcanoes, as evidence that scientists don’t understand what’s going on.

Public doubts about climate change are already increasing, even as scientists warn that the window for forestalling dangerous warming is closing. According to Pew Research, just 45 percent of Americans believe that scientists agree that warming is mostly because of human activities, down from 59 percent in 2006. The recent wave of news and magazine articles about scientists struggling to explain the warming slowdown could prolong or deepen the public’s skepticism.

But the “consensus” never extended to the intricacies of the climate system, just the core belief that additional greenhouse gas emissions would warm the planet. The greenhouse effect is truly undeniable—just consider Venus, where 96.5 percent of the atmosphere is composed of carbon dioxide, and the average surface temperature is more than 860 degrees Fahrenheit. Conversely, without greenhouse gases, Earth’s average surface temperature would be 0 degrees Fahrenheit.

And once you concede the existence of the greenhouse effect, it’s tough to dispute the role of greenhouse gas emissions in warming the planet: Over the last ten years, increased carbon emissions added trapped about as much heat in the climate system as small volcanic eruptions, the solar cycle, and stratospheric water vapor combined to deduct. But since the start of the industrial age, carbon dioxide has added nearly seven times as much positive forcing, and that number will keep growing with additional carbon emissions.3 So here’s what’s clear: Over the longer term, temperatures will increase. As Held puts it, “warming over 100 years isn’t that sensitive to fluctuations.”

“I don’t see how you can argue against it,” Solomon observed after declaring that “carbon dioxide will be king over the long run.” The amount of warming over the last century has not been small, and there “has to be a source, if you believe in basic thermodynamics.” Skeptics point to internal variations—the natural shifts that scientists have struggled to explain over the last decade. But oceanic heat content has also been increasing, ruling out the possibility that atmospheric warming is due to internal variability. To Held, that’s “pretty much a smoking gun.”

The last decade is proof of climate change, not a cause for reflexive skepticism. It was the warmest on record, despite a laundry-list of mitigating factors like prolonged La Nina, a wave of modest volcanic eruptions, and an ebb in solar activity. As those attenuating factors subside, climate scientists anticipate another round of rapid warming.

  1. Trenberth says “there may be some problems” in the CERES data. There are only a few satellites, requiring scientists to undertake “heroic efforts” to iron out irregularities and “biases associated with the diurnal cycle and angular corrections.” As a result the CERES data isn’t well-calibrated—its estimates can’t be reconciled with other data—but it’s “supposed to be good at measuring the change,” even if it can’t measure the absolute energy balance.

  2. Trenberth thinks Hansen is “lowballing things.” Hansen’s estimate uses the Argo floats, but Argo doesn’t “measure the top 10 meters of the ocean, and yet we have good estimates from sea surface temperature analysis that 0-10 meters accounts for .2 w/m^2

  3. Between 1750 and 2005, carbon dioxide contributed +1.8 w/m^2 of forcing to the climate system. Solomon likens these abstract units (w/m^2) to the outdated, warm-to-the-touch, energy inefficient, one watt Christmas lights. An energy imbalance of 1 watt per square meter is the equivalent of one of those warm Christmas lights burning across every square meter of the planet, in perpetuity. That’s the excess heat that’s warming the planet. Over the last ten years, surging carbon emissions increased forcing by another .27 w/m^2. That’s significant, even if it’s not enough to swamp small volcanic eruptions, the solar cycle, or stratospheric water vapor, which combined to deduct approximately -.25 wm^2.