EGU 2011: An outdoor volcano laboratory; lots of questions, and a few answers

Surface activity at Mount Stromboli, the volcano on the small island of Stromboli off the north coast of Sicily, has been recorded in detail for a thousand years. The volcano’s long-lasting eruptive state consists of explosions from the summit that recur at intervals of minutes to hours and regular emissions of gas that last about 100 years. Occasionally, the steady state is punctuated by large eruptions or lava flows.

The wealth of data on Stromboli, past and present, provided ample material for a discussion at the general assembly of the European Geosciences Union, which was held earlier this month in Vienna.

As volcano models are developed further, questions about the small-scale processes that drive the large-scale phenomena observed at Stromboli focus on ever more specific causes and effects. Frances Beckett and Fred Witham of the University of Bristol’s department of Earth sciences want to know why the gas content measured coming out of the volcano amounts to ‘a lot more than what should be there,’ based on subsurface melt volatile content.

In the prelude to a volcanic eruption, magma that contains dissolved volatiles is stored in a deep surface reservoir. The magma may already be supersaturated, or it may become supersaturated as it decompresses on its way through a conduit to the surface. Bubbles nucleate and grow as the magma rises. Bubble growth drives the magma to accelerate and potentially fragment into separate pieces. Fragmented magma leads to an explosive eruption; nonfragmented magma leads to an effusive lava flow. In the case of Stromboli, gas coalesces into slugs that burst near the top of the conduit and drive the magma upward, ejecting tephra and incandescent cinder.

Beckett cited an ongoing, nonexplosive sulfur dioxide output of 200 tons per day from the Stromboli volcano. The magma melt, however, contains just 0.28% sulfur.¹ Said Beckett, ’50 000 tons per day of melt needs to degas to account for this output, but there is no significant [volume] eruption of magma'—which is why Stromboli’s persistent degassing is such a puzzle

‘We know that magma must come to the surface to degas, but then where does it go if it is not erupted?’ she asked. The answer she presented at the EGU meeting involves flows driven by density differences: Buoyant gas-rich magma ascends and degasses near the surface, becomes less buoyant as a result, and then descends to the subsurface reservoir.

Beckett looks to golden syrup to find out how this exchange flow might behave. Her setup consists of two tanks, one above the other, connected by a vertical pipe. By adding water to the syrup, she can alter its density and viscosity. At the beginning of an experimental run, the top tank contains dense syrup, while the bottom tank contains less dense syrup.

Within the pipe, two flow regimes occur: for viscosity ratios less than 100, a side-by-side flow occurs in which both fluids are in contact with the pipe walls and share a single interface. For viscosity ratios greater than 100, the buoyant fluid occupies a cylindrical core and the denser fluid flows downward in an annulus.

Beckett’s model shows that transition between flow regimes occurs at 200 MPa when the melt viscosity is set by Strombolian melt composition and crystallinity. Numbers in hand, she calculates the conduit radius and magma volume flux at Stromboli. ‘But we don’t understand why different flows occur at different viscosity ratios,’ said Beckett. Nor does she yet know what may happen when an ongoing exchange flow is disturbed—say, by a gas slug that drives a volume of magma to erupt.

Witham had hoped to use Beckett’s study to constrain the flux through the conduit in his model, which predicts the time-dependent fluxes and compositions of volcanic gases emitted from Stromboli. He determines magma crystal content based on changes in pressure and volatile content, then calculates how much volatile material gets exsolved, the material’s composition, and the resulting magma rheology. ‘One main difficulty,’ he explained, ‘is that the up-welling and down-welling magmas can potentially mix, and certainly expand dramatically due to conduit convection.’

To deal with that difficulty, Witham assumes that what goes up must come down, minus the exsolved gas, and that at some pressure the magma reaches its eutectic point—that is, the composition at which it solidifies (well, becomes extremely viscous). This magma plug acts as a permeable lid, below which a ‘conveyor belt’ or exchange flow occurs.

Beckett and Witham use observations of gas emissions and melt inclusions at Stromboli to support a theory of conduit dynamics in a volcanic system. But they don’t look at deep surface processes or what happens once the volatile-rich magma fragments and solidifies into rock on its way out of the vent.

Michael Manga, a geophysicist in the department of Earth and planetary science at the University of California, Berkeley, and recipient of the EGU 2011 Robert Wilhelm Bunsen Medal, noted at the EGU meeting that ‘Stromboli erupts more gas than liquid, and water separates from the melt on the way to the surface.’ To explain the eruptive activity, the vertical conduit must be coupled to a deep magma-filled chamber.²

Manga also explained how fast and how far pyroclastic flows in explosive eruptions travel once they’ve escaped the volcano vent, based on the generation of ash particles through particle interactions, steam generation when flows enter the ocean, and the role of boundary conditions.

‘But the motivation is not [to understand] every tiny detail,’ Manga added. Rather, it’s explaining processes that cause certain outcomes. And while hazard mitigation is not a huge driving factor for volcanic process research, a reliable model of coupling between surface measurements and deep surface processes could provide a way of predicting what might happen in future eruptions.

Rachel Berkowitz

References

• 1. M. Burton et al., Science 317, 227 (2007) .
• 2. H. M. Gonnermann, M. Manga, Annu. Rev. Fluid Mech. 39, 321 (2007) .