Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Clouds’ cooling effect could vanish in a warmer world

Pacific ocean fog fills the coastal valleys on Big Sur, California

The cooling effects of cloud banks could evaporate under high levels of atmospheric carbon dioxide.Credit: Mathew Lodge/Alamy

Low-lying cloud banks off the coast of California, Peru and Namibia are some of the planet’s most effective cooling systems, because they reflect sunlight back into space. But new climate simulations show that increasing carbon dioxide concentrations in the atmosphere could break up these cloud layers and exacerbate future warming.

The findings1, published on 25 February in Nature Geoscience, reveal a previously unknown interaction between clouds and greenhouse gases: about three times the current level of atmospheric carbon dioxide can abruptly disperse clouds. Under a business as usual emissions scenario, this could occur in about a century. A world with fewer clouds, projections indicate, could witness up to 8ºC of warming in addition to that caused by greenhouse gases. Earth’s climate would be similar to conditions 50 million years ago, when crocodiles swam in an ice-free Arctic and palm trees grew as far north as Alaska.

“This is a warning shot about the future,” says study co-author Tapio Schneider, who studies cloud dynamics at the California Institute of Technology in Pasadena. “If we do not reduce emissions, very large and difficult-to-reverse climate changes are possible.”

Breaking up

The clouds that clump together into massive sheets over the oceans are called stratocumulus clouds, and worldwide, they can reflect roughly 4 to 7% of the energy from the Sun. But stratocumulus clouds frustrate climate modellers because they’re hard to replicate in computer programs.

Most researchers simplify small-scale phenomena, including clouds, rain, storms and ice, in order to simulate large-scale processes, such as changes in sea-surface temperature, as accurately as possible. There just isn’t enough computing power to handle realistic scenarios for all weather phenomena at once.

To get a more realistic picture of how clouds might behave under future climate scenarios, Schneider and his team simplified the large-scale processes and tried to model cloud behaviour as accurately as possible.

When the researchers cranked up current carbon dioxide levels from 400 parts per million (p.p.m.) to more than 1,200 p.p.m., the atmosphere warmed up and dense cloud layers started to break into smaller, puffier clouds. This occurred because stratocumulus clouds need to radiate heat into the upper atmosphere to maintain themselves — if the atmosphere gets too warm, the cloud banks disperse.

“It’s not bonkers,” says Andrew Ackerman, a cloud researcher at NASA’s Goddard Institute for Space Studies in New York. “The underlying mechanism is totally plausible.”

Nasty surprise

For decades, clouds have remained the leading source of uncertainty in climate-change predictions, including in the models used by the Intergovernmental Panel on Climate Change, says Matthew Huber, a palaeoclimatologist at Purdue University in West Lafayette, Indiana. This means that many models might be underestimating future climate change.

The model proposed by Schneider and his colleagues has similar issues, says Huber. Although the findings point to a warmer world, there’s still a fair amount of uncertainty in those predictions. Some of the large-scale interactions, including how oceans exchange heat and energy with the atmosphere, were simplified or neglected, he says. This makes it hard to know the precise carbon dioxide levels at which stratocumulus clouds become unstable.

But Schneider and other scientists are trying to address the limits of simulating Earth's atmosphere in computer programs. One approach uses machine learning to teach global climate models to better represent clouds by training them on real-world observations and simulations that detail smaller-scale processes. This could lead to faster and more reliable ways of forecasting future climate.

Whatever these models predict, people need to be ready for significant changes to our climate, says meteorologist Paquita Zuidema at the University of Miami in Florida. These shifts were common in the past, she says, “and Tapio’s work suggests how another one could happen in the future”.

The study serves as a reminder that a warmer world might hold surprises for us, says Huber. “And those surprises are not pleasant.”


Updates & Corrections

  • Correction 26 February 2019: This story has been corrected to reflect the fact that globally, stratocumulus clouds reflect about 4 to 7% of the energy from the sun. They can reflect around 30 to 60% of the sunlight that hits them.


  1. 1.

    Schneider, T. et al. Nature Geosci. (2019).

    Article  Google Scholar 

Download references


Nature Careers


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing


Quick links