A phenomenological model for precursor volcanic eruptions

Intense explosions of relatively short duration frequently precede large explosive and effusive volcanic eruptions—by as much as weeks to months in the case of very viscous magmas1,2,3,4,5,6. In some cases, such pre-eruption activity has served as a sufficient warning to those living in the vicinity to evacuate and avoid calamity1. Precursor events seem to be related to the formation of a magma pathway to the surface, but their precise interpretation is a long-standing puzzle. It has been inferred from theoretical studies that exsolution of volatiles might create an almost separate gas pocket at the tip of a propagating dyke7,8,9. Here we explain the role that such a process may have, using a laboratory study of the transient propagation of a liquid-filled crack with a gas pocket at its tip that grows with time. We show that once the gas pocket acquires sufficient buoyancy to overcome the fracture resistance of the host solid the dynamics of the gas pocket, rather than those of the liquid, determine the velocity of the crack tip. Furthermore, we find that the gas can ultimately separate from the liquid. We propose that fast-moving, gas-rich pockets reaching the surface ahead of the main liquid-filled fissure could be the origin of many precursor eruptions.