If you think pumping the sky full of chemicals sounds like a weird way to fight climate change, you’re not alone. Solar geoengineering — the idea of injecting aerosols into the high atmosphere to reflect sunlight back into space and make for a cooler planet — is very controversial. And not just because it seems so offbeat.

Although geoengineering is not yet being deployed in the real world, past computer modeling studies have suggested it could produce unintended effects like droughts. Some have worried that it might create new climate inequities, worsening the weather in some regions even as it improves conditions in others. It’s incurred so much backlash that until recently it’s been among scientists, and even today, much less attention is devoted to exploring this strategy than to cutting emissions.

But a study published this week in Nature Climate Change argues that the strategy could be highly successful — it’s all a matter of how much geoengineering we use. Yes, spraying a huge quantity of aerosols aimed at totally eliminating global warming can produce unwanted effects. Yet applying the right “dose” — just enough to cut global warming in half — could do the trick without causing negative side effects, the scientists say.

“The analogy is not perfect, but solar geoengineering is a little like a drug which treats high blood pressure,” said lead author Peter Irvine of Harvard University. “An overdose would be harmful, but a well-chosen dose could reduce your risks.”

In the study, Irvine and his co-authors used a high-resolution computer model to simulate what would happen if we deployed geoengineering with the goal of halving global warming, in a scenario where the carbon dioxide levels in our atmosphere have doubled preindustrial levels. (Currently, we’re at about 1.4 times those earlier levels.) Whereas most previous research only looked at temperature and precipitation, this study also examined other things that matter to people, like water availability.

The results? Geoengineering cooled the planet and reduced the intensity of extreme weather events like hurricanes. Importantly, this held true across the entire globe. There weren’t regional winners and losers, just winners. If anything, the researchers noted, the regions that suffered most from climate change were the most likely to see it reduced.

Critics of geoengineering have worried that although it may benefit the rich, it could harm low-income people, who may be less equipped to cope with unpredicted weather changes if things go awry, and who won’t get as much say in deployment. But David Keith, a senior author on the study and a Harvard physics professor, told me he believes it would be a net benefit for low-income people.

“The poorest people tend to suffer most from climate change because they’re the most vulnerable. Reducing extreme weather benefits the most vulnerable the most. The only reason I’m interested in this is because of that,” he said.

The study has significant limitations
Don’t get too excited just yet. The study — a collaboration between Harvard, MIT, and Princeton — is based on a highly idealized scenario.

The researchers chose to use a scenario where atmospheric CO2 levels have doubled preindustrial levels by the time geoengineering is deployed. “Double” may sound like a lot, but some climate scientists believe our CO2 levels will be woefully higher than that by the middle of this century. And we’re not even close to ready for any large-scale deployment of geoengineering. So by using this scenario, the study may be setting itself up for an unrealistically optimistic result.

It’s also important to note that the study doesn’t actually model what happens when you shoot aerosols into the sky. It models what happens if the sun’s rays are dimmed. Although that’s a fairly common proxy, Rutgers University climate expert Alan Robock objects that it doesn’t precisely capture the impact of spraying aerosols, which could have other effects, like messing with atmospheric circulation.

The team behind the study agrees that modeling aerosols is also important but believes that asking one model to do everything isn’t necessarily the best option, according to Keith. “The climate models treat aerosols pretty badly, so it’s not clear you can trust the results. In our opinion it makes more sense to use this model, and then separately do models that have very good representations of aerosols. It’s like building a bridge from two sides,” he told me.