Strain localisation during dome-building eruptions

Volcanic landscapes often present advantages for people who inhabit the surrounding areas, but the increasing numbers of people threatened by potential activity increases as these settlements grow. It is thus of vital importance to glean as much information as possible by monitoring active volcanoes (including seismicity, ground deformation, gas flux and temperature changes). Although volcanic behaviour can be difficult to predict, precursory information can often be identified retrospectively (once an eruption begins) to help link antecedent behaviour to eruption attributes. Likewise, eruption relics can be used to identify processes in pre-eruptive magma. Additionally, a huge amount of information may be gathered through experimentation on rock and magma samples. This study combines field and analytical studies of natural samples from Volcán de Colima (Mexico), Mount St. Helens (USA) and Soufrière Hills (Montserrat) with high-temperature magma deformation experiments to investigate the processes involved with magma ascent during dome-building eruptions (Figure S-1). The study of conduit-dwelling magma is of the utmost importance for understanding transitions from effusive to explosive eruptions. Of primary interest is the rheology of highly crystalline magmas that make up the magma column. Rheology is integrally linked to the composition and textural state (porosity, crystallinity) of magma as well as the stress, temperature and strain rate operative during flow. Many studies have investigated the rheology of multi-phase magmas, but in Chapter 2 this is notably linked to the evolution of the physical properties of the magmas; tracing the changes in porosity, permeability, Poisson’s ratio, Young’s modulus during strain dependent magmatic flow. Especially at high strain rates mechanical degradation of the magma samples may supersede magmatic flow and crystal rearrangement as the dominant form of deformation, resulting in lower apparent viscosities than those anticipated from magmatic state. This leads to an evolution of the fracture network to form inhomogeneous distribution of the permeable porous network; with damage zones cutting through areas of densification. In a conduit setting this is analogous to the formation of a dense, impermeable magma plug which would prohibit degassing through the bulk of the magma. Degassing may or may not proceed along conduit margins, and the plug formation could lead to critical overpressures forming in the conduit and result in highly explosive eruption. During the multi-scale process of strain localisation it is also probable that another previously unforeseen character acts upon magma rheology. Chapter 3 details the first documentation of crystal plasticity in experimentally deformed multi-phase magmas. The extent of the crystal plasticity (evidenced by electron backscatter diffraction (EBSD)) increases with increasing stress or strain, and thus remnant crystals may be used as strain markers. Thus it seems that crystal-plastic deformation plays a significant role in strain accommodation under magmatic conditions. Indeed plastic deformation of phenocrysts in conduit magmas may be an important transitional regime between ductile flow and brittle fracture, and a time-space window for such deformation is envisaged during the ascent of all highly-crystalline magmas. This phenomenon would favour strain localisation and shear zone formation at conduit margins (as the crystal-plastic deformation leads the magma toward brittle failure) and ultimately preferentially result in plug flow. During volcanic eruptions, the extrusion of high-temperature, high-viscosity magmatic plugs imposes frictional contact against conduit margins in a manner that may be considered analogous to seismogenic faults. During ascent, the driving forces of the buoyant magma may be superseded by controls along conduit margins; where brittle fracture and sliding can lead to formation of lubricating cataclasite, gouge or pseudotachylyte as described in Chapter 4 at Mount St. Helens. Within volcanic systems, background temperatures are significantly higher than the geotherm permits in other upper-crustal locations, whereas confining pressures are much lower than in high-temperature, lower-crustal settings: thus via their exceptional ambient P-T conditions, volcanic systems represent unique environments for faulting. This can result in the near-equilibrium melting and slow recrystallisation of frictional melt, which hinders the development of signature pseudotachylyte characteristics. Thus frictional melting may be more common than previously thought. Indeed Chapter 5 documents a second occurrence at Soufrière Hills volcano. Here, the formation is linked to repetitive seismic “drumbeats” which occurred during both the eruption at Mount St. Helens and at Soufrière Hills. Strain localisation, brittle rupture, sliding and the formation of shear bands along the conduit margin can have important implications for the dynamics of eruptions. Specifically, the capability of degassing via the permeable porous network may be strongly influenced by the formation of pseudotachylyte, which has almost no porosity. Based on the findings in chapters 4 and 5, a series of high-velocity rotary shear (HVR) experiments were performed. In Chapter 6 the results of these experiments demonstrate the propensity for melting of the andesitic and dacitic material (from Soufrière Hills and Mount St. Helens respectively) at upper conduit stress conditions (