If a supervolcano burps out a choking cloud of carbon dioxide, even if the effects are deadly and devastating, Earth’s atmosphere will eventually return to normal. Where, then, does all that greenhouse gas end up? 

Earth’s surface, it turns out, conceals a natural air filter. 

Certainly, plants play their part, drawing in carbon dioxide for photosynthesis. But there’s an even larger control mechanism: the very earth itself. Carbon dioxide in the air can weather certain minerals in the ground. In the process, those minerals react with carbon dioxide and pull it from the atmosphere. 

Geologists have long known about this air filter, but they’ve yet to master how it works. Now, scientists have evidence of what controls the process on a global scale: Those minerals weather more quickly if the, well, weather is warm and rainy.

“Everybody wants to understand how the globe works,” says Susan Brantley, a geologist at Pennsylvania State University. Brantley and her colleagues published their evidence in the journal Science today.

Weathering is when rocks and minerals deteriorate under exposure to nature’s elements—water, heat, microorganisms, and plants, to name just a few. (Weathering isn’t erosion, which involves movement, such as blowing wind or flowing water that picks up crumbs of rock and drops them elsewhere.) The authors focused on one specific type of weathering, caused by chemical reactions that involve carbon dioxide.

Even then, this gas doesn’t weather all minerals in the same way. Depending on their chemical composition, some might spit carbon dioxide right back into the atmosphere. Brantley and her colleagues instead studied a group of minerals known as silicate minerals, whose molecules contain silicon and oxygen atoms. Silicate minerals react with carbon dioxide and store it in the ground or, sometimes, in the water. Fortunately, these compounds are plentiful: Oxygen and silicon are the two most common elements in Earth’s crust. 

[Related: Earth has more than 10,000 kinds of minerals. This massive new catalog describes them all.]

The authors wanted to answer one question: How quickly do silicate minerals weather, and how does that attribute change as their surroundings shift? 

The answer is not straightforward. Chemical reactions don’t cause all the world’s weathering; it’s hard for geologists to separate chemical weathering from biological activity or groundwater percolation.

Because of that complication, geologists have found that chemical weathering seems to occur far more slowly in the soil outside than in a controlled laboratory. That’s a problematic discrepancy. In Brantley’s words, “if you can’t even extrapolate from your beaker to a stream outside your lab, how could we ever extrapolate to the globe?”

Fortunately, Brantley and colleagues weren’t the only researchers interested in the problem. They had decades of research, performed at local and regional scales, to pore over. They could look at experiments done in the lab. They could zoom out and look at observations of weathering in parcels of soil. They could zoom out further and find studies examining how weathering worked over entire river systems.

Analyzing that data, they could zoom out even farther and estimate a global trend.

Brantley’s group found that, as the temperature heated up, so did weathering. Likewise, weathering slowed down in the cold. But warmth wasn’t the only player they discovered. If the ground wasn’t in motion—if there was less erosion to move rock or less rainfall to create flowing water—weathering slowed down.

[Related: This diamond holds a never-before-seen mineral that can’t exist on Earth’s surface]

“It’s a very detailed analysis,” says Salvatore Calabrese, an environmental engineer at Texas A&M University who wasn’t an author on the paper.

“It’s able to align the field [and] the lab studies and make this coherent message,” says Bob Hilton, a geologist at Oxford University who also wasn’t an author on the paper.

The finding, according to Brantley, certainly helps geologists trying to look into Earth’s past, back at its long history of volcanic eruptions and swings back to normal. Assuming their models are accurate, geologists could look very far back indeed. For instance, they could examine preserved soil that’s billions of years old and make an educated guess at its atmosphere.

More pertinent to the present, the amount of greenhouse gases belched out by volcanoes is a drop in the bucket of emissions from humans burning fossil fuels. That raises another question: If we know weathering can decarbonize the air, then why can’t we accelerate the process?

As it happens, scientists and engineers are already working on that: an idea they call enhanced weathering. The process, as they envision it, might entail sprinkling a rock crumble across the ocean or over vast tracts of land. If done over large chunks of the world’s farmland, the hope goes, minerals in the rock will make a dent in the world’s carbon dioxide. (Of course, doing this might mean mining rocks from somewhere and potentially exposing people to rock dust.)

It’s a new idea, and for now, it’s largely confined to the laboratory. Some experiments have evaluated how it works in the presence of soil and plants, such as tests by Calabrese and his colleagues on small plots in a tropical forest. “You can take some measurements, but you cannot really look into what will happen over the entire forest,” he says.

That means enhanced weathering proponents face many of the same unknowns as geologists like Brantley. They know what happens in the lab, but they don’t know how this process might interact with real-world soils. And they don’t know whether their observations change over the size of an area.

It means, then, that Brantley’s findings could inform future enhanced weathering research: for instance, pointing its researchers to places with a plentiful water supply. “Maybe it’s a good reference to say, okay, maybe we can do something similar” to make enhanced weathering more efficient, says Calabrese.

For her part, Brantley is less interested in enhanced weathering than in some of the other players behind weathering: namely, living organisms. Life can boost weathering: microbes can manipulate their surrounding minerals. At the same time, living things can slow it down—a tree, for instance, can cut their roots into a rock and stabilize it. 

Hilton agrees that geologists should now study what microbes are doing.

“They’re probably driving part of this temperature response,” says Hilton. “So, understanding how they’re working, how they’re functioning, is really important.”