Clouds have lost reflectivity in recent decades, meaning more sunlight reaches our planet, where it is ultimately absorbed as heat. (Credit: TuiPanthida on Shutterstock)
Climate conundrum: Even cleaning up our air appears to be heating up the Earth.
In A Nutshell
- Marine clouds over the North Atlantic and Northeast Pacific became about 3% less reflective per decade between 2003 and 2022, allowing more sunlight to warm the oceans below.
- The change resulted from stricter air pollution controls that cut sulfur emissions, removing tiny particles that once acted as cloud seeds and made clouds brighter.
- These cleaner but dimmer clouds may help explain why parts of the Pacific and Atlantic have been warming faster than climate models predicted, contributing to marine heatwaves and ecosystem damage.
- Most major climate models significantly underestimated these cloud changes, raising questions about whether current warming predictions for these regions are too conservative.
For decades, reducing air pollution seemed like an obvious win-win for public health and the planet. Cleaner air means fewer respiratory diseases, heart problems, and premature deaths. Paradoxically, however, research reveals an unexpected side effect of cleaning up our air. Cleaner skies may actually be speeding up ocean warming in some of the world’s most critical marine regions.
The culprit? Changes in clouds.
Between 2003 and 2022, marine clouds over the North Atlantic and Northeast Pacific became noticeably less reflective. In this case, the clouds can be likened to giant mirrors in the sky, bouncing sunlight back into space before it can warm the ocean below. When those mirrors get dimmer, more heat gets through.
That’s exactly what’s happening, according to the study. These clouds are now reflecting about 3% less sunlight per decade than they did 20 years ago, allowing extra solar energy to warm the oceans beneath them.
The Pollution Paradox
Here’s where it gets complicated. Stricter air pollution controls in China, the United States, and other industrialized nations have dramatically cut sulfur dioxide emissions from power plants and ships. China alone reduced its sulfur emissions by roughly 16 million metric tons per decade. The U.S. cut about 7 million metric tons per decade over the same period.
These reductions have saved countless lives by improving air quality. But sulfur particles had an unintended benefit: increasing cloud brightness.
When sulfur dioxide from burning coal and oil enters the atmosphere, it creates tiny particles that act like seeds for cloud droplets. The more seeds you have, the more droplets form. And clouds with lots of small droplets are whiter and more reflective than clouds with fewer, larger droplets.
It’s similar to how crushed ice looks whiter than a solid ice block, even though both are made of the same material. The smaller pieces scatter more light.
As pollution controls removed these particles from the air, clouds had fewer seeds to work with. This resulted in fewer droplets, darker clouds, and more sunlight reaching the ocean.
Two Ways Clouds Changed
The changes affected clouds in two distinct ways, both making them less reflective.
First, with fewer particles to seed cloud formation, the clouds that did form had larger droplets. Larger droplets don’t scatter light as effectively as smaller ones, so the clouds became less bright overall. Scientists call this the Twomey effect, named after physicist Sean Twomey who described it in 1959.
Second, those larger droplets fall as rain more quickly. Smaller droplets need time to bump into each other and combine before they’re heavy enough to fall. Larger droplets skip that step, which means clouds don’t last as long. Shorter-lived clouds mean less total time reflecting sunlight. This is known as the Albrecht effect, after meteorologist Bruce Albrecht.
In the research team’s simulations, these two effects together accounted for about 69% of the cloud-related heating in these regions.
The Impact Of Ocean Warming
The North Atlantic and Northeast Pacific aren’t random choices for this study. These regions have been experiencing some of the fastest ocean warming on the planet, with devastating real-world consequences.
The Northeast Pacific has endured persistent marine heatwaves since 2014, with brutal impacts on marine ecosystems and fisheries from California to Alaska.
Scientists knew the water was getting warmer. What they couldn’t fully explain was why it was happening so fast. Natural climate patterns like the Pacific Decadal Oscillation, which normally drives temperature changes in this region, didn’t account for the extremes they were seeing.
The cloud changes may be part of the answer.
Climate Models Missed the Trend
When the research team, led by atmospheric scientist Knut von Salzen of the University of Washington and Environment and Climate Change Canada, looked at climate models, they found a troubling gap. Most models significantly underestimated how much cloud reflectivity had actually declined.
Of 24 major Earth System Models analyzed, 10 simulated trends consistent with satellite observations. The remaining 14 predicted significantly weaker changes.
This matters because these same models are used to project future warming. If they’re underestimating how sensitive clouds are to pollution changes, they may also be underestimating near-term warming in these regions.
Whether the dimmer clouds contributed to the record ocean temperatures in 2023 and 2024 remains unknown, though it seems plausible given that the climate system responds to such changes over years to decades.
What Happens Next
Emission scenarios project that aerosol concentrations will continue declining over the next few decades as countries pursue cleaner energy and stricter pollution controls. That’s undeniably good for public health, but it also means cloud reflectivity in these ocean regions will likely keep dropping.
To put the scale in perspective: over the study period, rising carbon dioxide added about 0.31 watts of heating per square meter globally. The cloud changes in these two ocean regions contributed about 0.15 watts per square meter to Earth’s overall energy budget—roughly half the CO₂ forcing, despite covering only 14% of the planet’s surface. Locally, within these regions themselves, the cloud effect was even more dramatic.
Nobody is suggesting we should keep polluting to preserve cloud brightness. The health benefits of clean air are too important. Sulfur dioxide and particulate matter cause respiratory disease, heart problems, and premature death. Chinese air quality regulations have saved hundreds of thousands of lives. American clean air standards have delivered enormous public health benefits.
The findings, published in Nature Communications, highlight an uncomfortable reality. Climate change isn’t just about greenhouse gases. The interactions between pollution, clouds, and warming are more complex than we thought.
The Bottom Line
Climate scientists have known for years that air pollution particles have been masking some greenhouse gas warming. What this study shows is that the unmasking effect is happening faster and more dramatically in certain regions than climate models predicted.
As von Salzen and his colleagues put it, these cloud changes are “an unintended consequence of efforts to improve air quality and reduce health risks.” The cleaner air is here to stay, and that’s ultimately a good thing. Nonetheless, understanding these secondary effects is crucial for predicting what comes next for regional ocean temperatures, marine ecosystems, and the communities that depend on them.
Paper Notes
Limitations
The study focused specifically on two ocean regions and a 20-year period, which limits the ability to generalize findings globally. While the improved model reproduced observed trends well, it may overestimate cloud feedbacks compared to other models. Cloud adjustments beyond the Twomey and Albrecht effects weren’t fully incorporated, introducing additional uncertainty. The analysis relied on satellite retrievals of cloud properties, which carry measurement uncertainties, particularly for aerosol optical depth and cloud droplet number concentrations.
Funding and Disclosures
Support for von Salzen, Doherty, and Wood was provided by the University of Washington’s Marine Cloud Brightening Research Program, funded by individual and foundation donors. Support for Akingunola, Cole, Digby, and Sigmond, along with computing resources, came from Environment and Climate Change Canada. Doherty’s contribution was partially funded by the Cooperative Institute for Climate, Ocean, and Ecosystem Studies under NOAA Cooperative Agreement NA20OAR4320271. Gryspeerdt was supported by an Imperial College Junior Research Fellowship and a Royal Society University Research Fellowship. The authors declared no competing interests.
Publication Details
Authors: Knut von Salzen (University of Washington and Canadian Centre for Climate Modelling and Analysis, Environment and Climate Change Canada), Ayodeji Akingunola (Environment and Climate Change Canada), Jason N. S. Cole (Environment and Climate Change Canada), Ruth A. R. Digby (Environment and Climate Change Canada), Sarah J. Doherty (Cooperative Institute for Climate, Ocean and Ecosystem Studies, University of Washington), Luke Fraser-Leach (University of Toronto), Edward Gryspeerdt (Imperial College London), Michael Sigmond (Environment and Climate Change Canada), Robert Wood (University of Washington)
Journal Citation: Nature Communications, Volume 16, Article 9433, Published November 5, 2025 DOI: 10.1038/s41467-025-65127-x







