Understanding the mechanics of lava dome collapse

A lava dome collapse can lead to the generation of pyroclastic flows and debris avalanches, both of which are hazardous to areas surrounding the volcanic edifice. Although the understanding of lava dome emplacement and pyroclastic flows has improved in recent years, knowledge of the mechanisms that trigger collapse is still limited. In this project, I investigate lava dome collapse in three ways: (1) the implementation of a global, historical database of lava dome collapse events and its statistical analysis; (2) the development of a new numerical model simulating lava dome emplacement, evolution and collapse in the context of volcanic activity; and (3) exploration of temporal changes in geomechanical rock properties across the Soufrière Hills lava dome eruption. First, a numerical model using the discrete element software (Particle Flow Code) is implemented and validated by showing how observed dome growth at Volcán de Colima can be reproduced. Next, the database analysis of the Global Archive of Dome Instabilities (GLADIS) highlights the most common mechanisms of lava dome collapse, and clearly links collapse volume to dome emplacement style. The identified collapse mechanisms (e.g. gravitational failure, switch in extrusion direction, internal gas overpressures, and topography) are incorporated into a new suite of numerical models. Simulations show that dome collapse resulting from gravitational failure and internal gas pressurisation leads to deep-seated rotational failures of large dome volume fractions, whereas topography-controlled collapses involve only superficial rockfalls. Lastly, an investigation into physical and mechanical rock properties of products from the 1995-2010 eruption of Soufrière Hills shows a clear correlation between high porosity (found in later eruptive products) and low compressive and tensile strengths. Laboratory investigations are used to define "strong" and "weak" scenarios which are incorporated in order to calibrate the numerical models, as well as scaling these rock properties to investigate dome stability when rock-mass properties are used. By combining the analysis of a global database, laboratory work, and numerical modelling techniques, this project is a comprehensive study of the mechanisms that initiate lava dome collapse, and shows for the first time the links that exist between collapse volumes and modes, and conditions during lava dome emplacement.