Why global warming is accelerating and what it means for the future

Temperatures over the past three years have been even higher than expected, which has provoked a debate among scientists. Almost everyone agrees that global warming has accelerated. But some scientists say it is increasing even more than climate models show, while others claim that the increase in temperatures is due to natural fluctuations that will soon disappear.

Depending on who is right, we may have even less time than we thought to avoid or adapt to catastrophic consequences.

“Ultimately, this is a question of how bad climate change is going to get,” says Zeke Hausfather of the nonprofit Berkeley Earth in California.

Earth warmed at a steady rate of about 0.18 °C per decade until the 2010s, when observed temperatures appeared to begin rising slightly faster.

Then 2023 became the warmest year on record by a margin of 0.17 °C, more than expected even with a slight acceleration in warming in the 2010s. Deadly floods hit Libya, unprecedented cyclones hit Mozambique and Mexico, and record wildfires set fire to cities in Canada, Chile, Greece and Hawaii.

That year, James Hansen of Columbia University, New York—who famously told the US Congress in 1988 that humans, not natural fluctuations, were warming the planet—published a much-discussed paper with colleagues claiming that the rate of warming had accelerated to about 0.32°C per decade after 2010.

They attributed this mainly to the “Faustian bargain” humanity made with aerosol pollution from burning fossil fuels. Sulfur aerosols reflect sunlight back into space, and they also help form reflective clouds. For decades, these aerosols and clouds have masked some of the warming from carbon dioxide emissions.

Now that the world is reducing aerosol pollution, which kills millions of people every year, this hidden warming is being exposed and climate change is accelerating, the paper claimed.

China, the world’s largest emitter of fossil fuels, launched a “war on pollution” when it hosted the Beijing Olympics in 2008. Since then, it has reduced sulfur aerosol emissions by at least 75 percent.

At the same time, the International Maritime Organization has cracked down hard on sulfur emissions from ships. The air is less dirty over oceans than over land, so this reduction in aerosols from ships can result in very few clouds, and ship tracks – lines of clouds that typically follow vessels – have become smaller.

As a result of both actions, global sulfur dioxide emissions have fallen by 40 percent since the mid-2000s. “The atmosphere is cleaner, so more solar radiation comes in,” says Samantha Burgess of the EU’s Copernicus Climate Change Service.

That appeared to be underlined when 2024 was even warmer than 2023, exceeding 1.5°C above pre-industrial levels for the first time – a step towards missing the more ambitious Paris Agreement target. Temperatures remained almost as high in 2025, when heat waves killed thousands in Europe and cyclones ravaged Southeast Asia and Jamaica.

But while most scientists agree that reductions in our aerosol emissions have accelerated global warming, they disagree about how much. The rate of 0.32°C per decade found by Hansen and his colleagues was higher than the 0.24°C rate estimated by the UN’s Intergovernmental Panel on Climate Change and the average of 0.29°C given by the latest generation of climate models.

A major complication is that natural fluctuations have also affected the planet’s temperature. The first was the arrival in 2020 of a particularly strong maximum in the approximately 11-year solar cycle, when sunspots and magnetic activity began to increase the amount of sunlight reaching Earth.

Then, in 2022, the eruption of a massive underwater volcano near Tonga in the South Pacific shot 146 million tons of water vapor, a greenhouse gas, into the stratosphere. At the same time, it ejected sulfur aerosols that cooled the atmosphere somewhat.

Finally, 2023 and 2024 saw a strong El Niño, a natural climate pattern in which weakened trade winds allow a mass of warm water to slosh back over the central and eastern Pacific, inflating global temperatures.

To determine how much global warming is accelerating, scientists must estimate and exclude the impact of this natural variability, and then try to fit a curve to a trend in observed temperatures that is only beginning to emerge. Less natural variation would mean more acceleration, and vice versa.

Earlier this month, a statistical analysis by Stefan Rahmstorf of the University of Potsdam, Germany, and US statistician Grant Foster found that global warming has jumped to about 0.36°C per decade since 2o14.

But researchers such as Michael Mann of the University of Pennsylvania have argued that Rahmstorf, Hansen and others have overestimated the impact of aerosols and underestimated the impact of natural fluctuations. He says that he and his colleagues are working on a study that has only found modest progress since the 1990s.

“Recent warmth, which is strongly influenced by the 2023-2024 El Niño event, is entirely consistent with standard climate model simulations,” says Mann. “There is no need to invoke any ad hoc mechanisms, including a supposed acceleration over the past decade.”

But it is possible that unexpected climate feedback loops have also contributed to the recent warmth. The biggest wild card is clouds, which are too small and scattered to be well represented by climate models.

A study last year by Helge Goessling of the Alfred Wegener Institute in Germany and his colleagues attributed about 0.2°C of the nearly 1.5°C warming in 2023 to a decline in low-lying clouds. While some of this cloud loss was due to the reduction in sulfur pollution, some of it may be due to an “emerging low-cloud feedback”, the researchers said.

Traditionally, cold, moist air above the subtropical ocean and the warm, dry air above have formed what is known as a temperature inversion, keeping the layers separated. But if climate change increasingly warms the cold air and breaks down the inversion, the dry air can sink and reduce moisture and thus clouds, says Goessling.

“The more you warm, the more you’ll be able to dissolve your low-level clouds,” he says. “It’s most likely to be low-cloud feedback.”

If the acceleration can mainly be attributed to sulfur reductions, climate change will probably slow down again in the coming decades, when there is no longer sulfur pollution to cut. But if climate feedback loops have been unleashed, it could potentially continue to accelerate.

This would mean that we have underestimated the climate sensitivity, or the amount of warming that will occur from a given increase in atmospheric CO2.

“The worst-case scenario would be that this comes from a cloud feedback that the models don’t predict, and would mean that we actually have a much more sensitive climate than some of the models predict,” says Brian Soden of the University of Miami, Florida.

The world is on track for 2.7°C of warming this century under current policies. However, these estimates have an uncertainty of approximately plus or minus 1 °C. More acceleration could mean that the Earth is headed for something closer to 3.7 °C and humanity will have to cut carbon emissions even more to avoid devastating impacts.

“3.7 °C … makes some regions uninhabitable,” says Hausfather. “2.7°C would still be bad, but many more regions could potentially adapt to it.”

However, fossil fuel emissions are also increasing, a trend that governments must reverse if they are to have any hope of limiting the impacts, says Burgess.

“Global warming is a bit faster, but we’ve also lost time because we haven’t seen the ambitious measures to decarbonize our society,” she says.