The world is warming faster than we can cut emissions. Volcanoes are already cooling the planet, with particles that reflect sunlight. Maybe we can too.
The past ten years have been the warmest on record, with 15 countries setting national temperature records in 2024 alone. Although growth in global oil demand weakened noticeably in 2024 and Europe and the United States reduced their carbon dioxide emissions by 2.2 percent and 0.5 percent respectively, overall global emissions rose 0.8 percent. Permanent carbon removal techniques, though promising, extract just tens of thousands of tons of carbon dioxide per year – almost nothing relative to the five to ten billion tons needed to stabilize temperatures at 1.5 degrees celsius above preindustrial levels.
Alongside 41.6 gigatonnes of carbon dioxide emissions, we are also heating the atmosphere with 120 million tonnes of methane annually, largely from the agriculture and energy sectors, as well as fluorinated gases and nitrous oxide. Meanwhile, industrial processes and energy production release 106 million tonnes of different air pollutants, which lead to approximately seven million premature deaths per year.
Subscribe for $100 to receive six beautiful 120-page issues per year.
As well as warming the planet with greenhouse gases, we have also historically cooled the planet – unintentionally – by emitting sulfur dioxide and other particulates that reflect sunlight back into the atmosphere. That accidental cooling has helped mask the true extent of global warming. This effect will vanish as cleaner technologies phase out these emissions.
Preventing catastrophic climate change will depend heavily on stopping emissions of greenhouse gases. Most scientific bodies, including the Intergovernmental Panel on Climate Change, agree that this will become easier once we learn how to remove carbon from the atmosphere at scale.
But there may be a third option to help us survive in the short run: deliberately cooling the planet.
A volcano turns the sky hazy
On June 15, 1991, Mount Pinatubo in the Philippines hurled approximately 17 million tonnes of sulfur dioxide into the stratosphere, a part of the atmosphere far above our head (roughly 15 kilometers and up), where there are no clouds and relatively weak winds. The sulfur dioxide quickly oxidized into tiny sulfuric acid droplets, which remained suspended and slowly drifted poleward for years, thanks to the slow-moving current known as the Brewer-Dobson circulation. The size of these droplets – about a hundredth the width of a human hair – turned out to be close to the wavelengths of visible light, making them very efficient at reflecting light back into space, like tiny mirrors.
Over the next eighteen months, land masses cooled by approximately 0.5 degrees celsius. Satellites tracked the spreading haze, and climate models successfully predicted the magnitude of cooling. This gave scientists powerful validation of our physical understanding of climate and suggested that sunlight‑blocking aerosols could offset greenhouse warming, at least temporarily.
Pinatubo was not the first such eruption: El Chichón lofted seven million tonnes of sulfur dioxide and cooled the planet approximately 0.3 degrees in 1982. The 1815 eruption of Mount Tambora, the largest global eruption for at least 1,800 years, led to the ‘Year Without a Summer’ across Europe and North America, with summer temperatures the coldest on record for almost 250 years. Suspended particles of volcanic matter gave the sky a reddish cast for years.
On January 1, 2020, international maritime regulations reduced the maximum sulfur content of shipping fuel from 3.5 percent to 0.5 percent. Many studies have identified this as responsible for part of recent temperature anomalies, including a fifth of 2023’s record highs – the equivalent of two to three years of warming. In the Arctic, this dose on its own is responsible for a tenth of all human-associated warming.
Energy in, energy out
The Earth’s climate is dictated by a simple energy balance: ‘energy in, energy out’. Incoming sunlight warms the planet: 173,000 terawatts continuously hit the Earth each day, about 340 watts per square meter. On average, the atmosphere reflects about 29 percent of incoming energy back into space and absorbs 23 percent. Approximately 48 percent reaches the Earth’s surface, while another 7 percent is reflected by the surface itself.
The remaining energy exists in a complex balance, with energy radiating, evaporating and convecting upwards into the atmosphere. Greenhouse gases warm the planet by absorbing some of the radiation on its way out to space, then reradiating it in all directions, including downwards, back to Earth.
The aerosol we know
Humans emit over 80 million tonnes of sulfur dioxide per year as an industrial byproduct (down from 140 million tonnes in the 1980s), and nature adds more sulfur through volcanoes and other biological activity (with compounds like Carbonyl Sulfide, COS, and Dymethil Sulfide, DMS, which eventually turn into sulfur dioxide as well). We also use at least as many tonnes of sulfur-based compounds in fertilizer. Overall, the ubiquity of the element sulfur in its many forms means that it is among the best-studied substances in atmospheric science.
Today’s scientists have a range of tools to monitor its effects in near-real time: NASA’s Aura satellite’s Ozone Monitoring Instrument can detect sulfur dioxide emissions from a unique fingerprint in the ultraviolet spectrum. The Stratospheric Aerosol and Gas Experiment-III on the International Space Station looks through the stratosphere to see precisely how much aerosols are dimming it. These come on top of a range of ground-based sensors, lidars – laser radars that create three-dimensional images – and modern weather balloons. Together, these have let us map exactly what happens to emissions when they hit the atmosphere.
Most sulfur dioxide emissions occur in the lower atmosphere. Industrial facilities and marine ships emit sulfur into the troposphere, where it reacts with water and falls out as acid rain. Sulfate, which is produced when sulfur dioxide interacts with oxygen and water in the air, is highly soluble, so this process happens quickly, within days to weeks. Because it doesn’t last long or spread far, this tropospheric sulfate cools inefficiently. Still, the sheer volume of these emissions has produced a significant masking effect, suppressing an estimated 0.4 degrees of warming since the beginning of the industrial revolution.
Unlike this surface pollution, aerosols formed in the stratosphere last 12–18 months. There is no rain in the stratosphere to remove them, and the thin, stable air keeps them suspended long enough to spread around the globe. This makes them far more effective at cooling per tonne of sulfur dioxide. Once released, this gas oxidizes into aerosols over days to weeks, depending on sunlight and humidity. Pinatubo’s 17 million tonnes of sulfur dioxide reduced the energy hitting an average square meter of the earth by about 2.5 watts, or 0.7 percent. This led to about half a degree of cooling. Models generally believe that targeted injection would prove somewhat more efficient, suggesting that approximately 10 million tonnes would be needed to raise the Earth’s overall reflectivity enough to offset about one degree of global warming.
Stratospheric aerosols eventually fall out of the atmosphere thanks to gravity, which is why volcanic eruptions have not cooled the world permanently and why the world became habitable again after the Chicxulub comet hit the earth 66 million years ago and cut solar radiation by half. When an injection stops, the aerosols eventually fall, and solar radiation once again reaches the surface unimpeded, just as global temperatures bounced back after shipping-driven sulfate pollution briefly plummeted during Covid-19.
In total, the world has warmed about 1.2 degrees since preindustrial times and is expected to warm another 1.5 degrees, even with mitigations and carbon removal baked in. The world currently emits 80 million tonnes of sulfur dioxide. If a quarter of that – 20 million tonnes – went into the stratosphere, rather than the troposphere, both the warming since the industrial revolution and all of the future warming we expect would be balanced out.
Why not pump sulfur into the stratosphere right now?
The evidence that sulfur emissions cool the world is overwhelming. It is now uncontroversial that we do this unintentionally, and that, if we wished, we could do it deliberately. What we don’t yet know is exactly how much a given injection would cool the world, how uniform that cooling would be, how long the effects would last, and whether it might produce dangerous interactions with other processes we don’t yet understand.
When Pinatubo erupted, the effects were not localized to the Philippines. The particles it emitted increased stratospheric optical depth (a measure of how much light is blocked by something) by 10 to a 100 times, depending on where you were in the world. The effect was much larger at the equator. In general, however, the movement of aerosols in the stratosphere is governed by established, but not always fully predictable, circulation patterns. In the tropics, air rises and flows poleward through the Brewer-Dobson circulation. As a result, aerosols injected at tropical latitudes tend to spread toward the poles and remain largely confined to the hemisphere of release.
Aerosols reduce the temperature of the world by preventing visible light from reaching the ground. This is not a pure good. Photosynthesis depends on light, as do solar panels (which now produce about seven percent of the world’s electricity). Both farms and solar panels are responsive to changes in air quality due to pollution and eruptions: Sandia National Laboratories reported that their Solar Electric Generating System, an early concentrator solar power plant, was producing around 30 percent below normal capacity during 1992 when Pinatubo’s effects were strongest. Other estimates suggest that Pinatubo and El Chichon’s eruptions affected global maize, soy, rice, and wheat yields.
The idea of intentionally injecting aerosols into the stratosphere to avert more warming isn’t new. The concept dates back to the early twentieth century. In the 1970s, Soviet scientist Mikhail Budyko proposed injecting aerosols to counteract global warming. In 2006, American atmospheric chemist and Nobel Prize winner Paul Crutzen suggested this was worth exploring as the potential solution to what he termed ‘the policy conundrum’: carbon dioxide emissions not decreasing, aerosol not decreasing, and warming continuing unabated.
If we wanted to cool the planet deliberately, then we would likely want to focus the effects of such efforts on the regions where the costs of heat outweigh the benefits of the sun – especially the hottest countries and the melting poles. These are the places where extra warming carries the steepest costs.
There are also known unknowns: sulfate aerosols can accelerate ozone loss by acting as reactive surfaces where otherwise slow or unlikely chemical reactions would take place. After Pinatubo, scientists observed a slight drop in ozone levels for two years at the poles.
Scientists now have pretty compelling theories about how to cool the planet deliberately, while mitigating some of these risks. They have pretty good models for how long it would last and how sulfate aerosols would interact with other chemicals. But since deliberate sunlight reflection has never been tried at any scale, many are rightly nervous about jumping straight to ‘geoengineering’ the globe deliberately.
What would a SRM clinical trial actually look like?
We don’t have to jump straight into full-scale deployment. And when it comes to testing risky new interventions, we already have a proven playbook: the modern clinical trial, which, though often slow and expensive, does work to prove that drugs are safe. After a US pharmaceutical company accidentally killed more than 100 people with a poisonous antibiotic, Congress passed the 1938 Food, Drug, and Cosmetic Act, obligating drug companies to prove their drugs were safe before they could be sold to consumers. The 1962 Kefauver-Harris Amendment added a requirement that drugs be proven to work, and in doing so codified the three-phase trial framework that drug companies use worldwide today.
Sunlight reflection research could proceed in a similar way. Rather than immediately releasing millions of tonnes of aerosols in the sky, the trials would start small, tracking results, and scaling only as uncertainties shrink. This would reduce the scientific unknowns, but also create transparency, align incentives, and build trust in the results.
Phase one: small-scale outdoor tests
Phase one would involve releasing approximately ten tonnes of sulfur dioxide (for comparison, South Africa’s Matimba coal power plant emits over 270,000 tonnes per year, or around 750 tonnes a day) in one location, one kilometer or more above the tropopause, and carefully measuring what happens to it with aircraft and ground-based instruments. This amount would be far too small to affect climate, but this would allow researchers to study how aerosols form, disperse and persist at this altitude in real-world conditions.
Some scientists have proposed using different aerosol substances or particles, such as calcite, alumina and novel engineered particles. These might have theoretical advantages over sulfur, such as neutralizing acidic reactions and reducing ozone loss. However, a clinical trial should start with the more familiar sulfate, as these alternatives introduce a new set of unknown unknowns. Unlike alternatives, sulfur dioxide’s chemical pathways and climate impacts are well known.
Phase two: mid-scale dispersal and regional observation
If the sulfur dioxide behaves the way we expect it to in these tests, the second phase experiment would be to release ten or a hundred tonnes. This would still be orders of magnitude smaller than a small volcanic eruption like Ruang in Indonesia, which instantly injected 250,000 tonnes of sulfur dioxide in 2024, and which had no measurable impact on the global climate.
This phase would make it easier to measure crucial unknowns, such as aerosol lifetime, how it moves between hemispheres, and its impact on regional climate indicators like surface temperature, rain patterns, and cloud cover.
Phase three: sustained cooling trials
After researchers worldwide have studied the data, governments would need to decide if they were prepared to move towards initial deployment.
This third phase, which would resemble the post-licensure Phase IV in medicine (the monitoring that occurs after a drug has been approved for public use), would test the effects of continuous, low-level injection. This might rise to cooling 0.1 degrees celsius over five years, significantly slower than our current rate of warming. It would require coordinated, near-daily dispersal of significant quantities of sulfur dioxide, amounting to as much as a million tonnes per year. This would still be far from the peak sulfate emissions of the 1980s.
This would answer crucial operational questions like whether we can predictably cool the warmest regions of the globe we want to cool without hurting solar power, crop yields, and things like tourism in more temperate regions, and also determining whether ozone depletion stayed tolerably low. This phase would also begin to test the scalability of infrastructure and the ability of global systems to respond adaptively to observed effects.
While we currently observe the stratosphere with a few satellites, many of these lack the required resolution, are close to retirement, and not slated to be replaced. This phase of the trial would require a new, more detailed observation system: one that likely includes novel satellite instruments and high-altitude platforms such as stratospheric balloons, solar-powered gliders, or long-endurance drones capable of operating in the upper stratosphere for extended periods.
While some of these tools already exist, many are still in prototype form or were not designed with aerosol injection in mind. Building a robust detection system from scratch or by upgrading existing assets will be part of the challenge. It would need to detect small-scale interventions and distinguish them from background signals like new sources of emissions or volcanic eruptions.
Scaling up
Who, then, should be responsible for running and funding such a trial? In practice, a national government with strong scientific capacity and existing aerospace infrastructure could do so with relative ease. The United States, for example, could allocate on the order of $100 million — a rounding error in federal climate or defense budgets — to mount a serious phase one trial.
Legally, there is no binding international treaty that prohibits small-scale stratospheric aerosol experiments. The main frameworks in play are ‘soft law’ norms, such as the Convention on Biological Diversity’s non-binding moratorium and the London Convention’s restrictions on ocean fertilization, which do not cover the stratosphere. The Montreal Protocol is legally binding, but it regulates only substances that directly deplete ozone, not sulfur dioxide (which does so only indirectly), so it is relevant but does not prohibit such research.
The more significant constraint is political legitimacy: without broad international agreement, even modest real-world tests are likely to be controversial. Because of that,most people working on this think that funding should be pooled and managed multilaterally, rather than by a single country alone.
If, at any of the previous scales, something worrying were to appear, the trial could be stopped, without concerns about an abrupt termination. But if the trial were a success, the next step would be to scale it up.
Full deployment of stratospheric sulfur dioxide would rely on a fleet of specialized aircraft. Keeping temperatures at a steady 1.5 degrees above preindustrial levels would require over 20 million tonnes of sulfur dioxide annually by the end of the century, delivered by over 300 planes, about as many as FedEx’s entire global fleet, flying year-round to deploy it. Airplanes are likely the best option because they can carry more: a balloon-based approach would require over five million annual flights, given balloons’ payload capacities of less than 4,000 kilograms each.
To maximize cooling per unit of sulfur dioxide, scientists believe the ideal approach would be to inject aerosols at tropical and mid-latitudes, at altitudes of around 22 kilometers, where aerosol lifetimes may be longer and cooling more effective than at lower altitudes. At the moment, there are only five aircraft models that are capable of reaching these heights with substantial payloads. These are all repurposed bombers used for research and owned by NASA.
The tropopause – the boundary between the troposphere and stratosphere – is significantly lower at high latitudes. Existing commercial jets such as a Boeing 777 could be retrofitted to reach the necessary injection altitudes there, typically around 9–13 kilometers. This approach could still deliver meaningful cooling and could help to offset the amplified warming we’re seeing at the poles, though the politics of cooling countries with cold climates might be more difficult than cooling those already suffering from the worst effects of climate change.
The cost of all this would be about $70 billion each year, according to one speculative estimate, or between $2,000 and $3,000 per ton of injected sulfur dioxide. This would be a lot of money. But it is significantly less than the damage caused by the Los Angeles wildfires this year, just one example of a wave of the extreme weather consequences we are already experiencing due to global warming, a wave that is likely to continue for decades as emissions slowly ramp downwards.
While initial high-level international discussions have been contentious in wealthy countries, the global dialogue is picking up. The Degrees Initiative now funds research into the impacts of sunlight reflection in 22 developing countries; African researchers have launched the Africa Climate Intervention Research Hub; and Chinese institutions are investing in sunlight reflection studies. We, the authors, have set up an organization called Reflective, which aims to build tools that simplify some of this research and help translate scientific findings into decision-relevant insights. Our hope is that this makes it easier for people to understand the research that is going on and understand what’s at stake.
A complement to emissions reductions, not a substitute
There is no substitute for cutting carbon dioxide emissions. But even if we hit the most ambitious timelines for doing this, low-lying nations could still face catastrophic flooding, equatorial regions may see escalating heat stress, and many parts of the world could still suffer sharp declines in crop yields and freshwater availability. If there are tools that could help us to safely alleviate these effects of climate change, we have a responsibility to explore them.
We are still only at the initial stage of performing safe, controlled experiments to see if any of this can really work. But if it can, it could be a way to prepare ourselves for the worst.