Cornell Study Finds Two Mount Etna Eruptions Followed Wildly Different Paths

A Cornell-led research team has reconstructed two major eruptions at Mount Etna in Italy and found that magma took very different paths to the surface at very different speeds. The findings, published in the journal Geochemistry, Geophysics, Geosystems, could help improve models used to estimate the risks of future eruptions.

The team, led by Esteban Gazel, the Charles N. Mellowes Professor in Cornell's Department of Earth and Atmospheric Sciences, used a Raman spectroscopy technique it pioneered to examine crystals formed inside magma and measure microscopic gas bubbles trapped within them, about 1 to 10% the thickness of a human hair. Measuring the carbon dioxide density in these bubbles let the researchers calculate pressure, and then depth.

First author Maxim Gavrilenko, a former Cornell postdoctoral researcher, led the analysis; collaborators Terry Plank of Columbia University and Bruce Houghton of the University of Hawaii, Manoa, traveled to Etna to collect samples. One eruption, in 122 B.C., was mafic — low-viscosity magma rich in magnesium and iron — and Plinian, the most explosive eruption class, named after Pliny the Elder, who described Mount Vesuvius erupting in 79 A.D.

That eruption's magma began rising from about 22 kilometers deep, then moved slowly and stalled at a shallower 2 to 5 kilometers, remaining there for several weeks and gradually releasing gas before erupting. A second eruption, the Fall Stratified event, dated to nearly 4,000 years ago, saw magma rise rapidly from 24 to 30 kilometers deep and erupt within hours, associated with much higher carbon dioxide levels.

Gazel said Etna is one of the few volcanoes where carbon dioxide and water compete to control an eruption: a higher threshold of carbon dioxide drives a fast, deep eruption, while a higher threshold of water keeps the process controlled at shallow levels. His team is applying the same technique to volcanoes in Chile, Hawaii and other regions to build data needed for physical models of eruption risk.