Researchers have discovered a simple and surprising control over the depth of a volcano’s magma chamber: how much water it contains. The finding is significant because water fuels the most devastating eruptions—from Vesuvius in 79 C.E. to Mount Pinatubo in 1991.
The work could also help improve models that predict eruptions, which for years have been based on a volcano’s seismic rumblings and records of its past behavior. Now, says Daniel Rasmussen, a volcanologist at the Smithsonian National Museum of Natural History who led the new study, “We’re combining that information with models that consider the fundamental physics of what’s happening underneath a volcano.”
Rasmussen and his colleagues wanted to understand why magma chambers—slushy mixes of molten rock, solid crystals, and gases—lie anywhere between about 1 kilometer and 12 kilometers below the surface of “arc” volcanoes, a common kind of volcano that forms near the boundaries of tectonic plates. When plates of ocean crust slide into the mantle, the layer that makes up most of Earth’s interior, they drag water with them that gets locked away in minerals. This water then fuels the formation of magmas.
As this magma rises through cracks and fissures, it is depressurized. Eventually, the water in the magma is forced out as bubbles of vapor, much like the bubbles in a popped can of carbonated soda. But magma also gets stickier as it loses water, and Rasmussen and his colleagues suspected that it gradually becomes so thick it can rise no farther—at least until a physical disturbance such as an injection of extra magma drives an eruption.
Theoretical models suggest the depth at which magma begins to lose its water depends on its initial water content. A magma body with 1% water by weight would begin to lose water just 1 kilometer below the surface, for example, but for magma with 7% water by weight, loss would begin much deeper, at about 12 kilometers. This means, counterintuitively, that “wetter” magmas—even though they’re initially more fluid—thicken up and stall out at greater depths than “drier” ones. The researchers thought this could explain why magmas occur at different depths.
Testing the idea with real-world volcanoes isn’t easy, however, because a magma’s chemical composition can change during an eruption. The researchers needed a way to directly sample unaltered magma at depth, and to do so they went to the microscopic scale.
As crystals in the mush of a magma chamber grow, irregularities in their structure can trap tiny bits of pure magma. These bits, known as melt inclusions, cool and harden but remain unaltered through an eruption. “A melt inclusion is like a tiny little magma chamber that gets trapped,” Rasmussen says.
Rasmussen’s team calculated the water content from almost 4000 melt inclusions gathered from rocks spewed by 62 arc volcanoes around the world. They then cross-checked their data with a subset of 28 of these volcanoes for which the magma chamber depth is known from geophysical data. The relationship, which the team reports today in Science, was exactly what the models predicted: the deeper the magma, the higher its water content.
“It’s very elegant,” says Michael Stock, a volcanologist at Trinity College Dublin. “They show such a strong correlation with water content it’s almost hard to believe anything else could explain why magmas are where they are in crust.”
The discovery may eventually help improve volcanic forecasting because water and other gases in magma are known to trigger particularly explosive eruptions. “One could say that [deeper magmas] have more fuel for explosive eruptions,” Rasmussen says. However, the explosiveness depends not so much on the abundance of water, but on how it escapes. “If the water has escaped before it gets to the surface, you might get a sticky lava flow,” says Christopher Kilburn, a volcanologist at University College London. “If the bubbles stay inside the magma and come to the surface as a froth, then you can get a major explosive eruption.”
The new finding is an important puzzle piece for understanding volcanoes, but far more work must be done to understand the most explosive eruptions, says Catherine Annen, a geologist at the Czech Academy of Sciences’s Institute of Geophysics. “We can only hope that one day we will have physical models that have predictive power on volcanic behavior,” she says. “We are not there yet.”