A review of analogue and numerical modelling in volcanology

Abstract. Modelling has been used in the study of volcanic systems for more than one hundred years, building upon the approach first described by Sir James Hall in 1815. Informed by observations of volcanological phenomenon in nature, including eye-witness accounts of eruptions, geophysical or geodetic monitoring of active volcanoes and geological analysis of ancient deposits, analogue and numerical models have been used to describe and quantify volcanic and magmatic processes that span orders of magnitudes of time and space. We review the use of analogue and numerical modelling in volcanological research, focusing on sub-surface and eruptive processes including the accretion and evolution of magma chambers, the propagation of sheet intrusions, the development of volcanic flows (lava flows, pyroclastic density currents and lahars), volcanic plume formation and ash dispersal. When first introduced into volcanology, analogue experiments and numerical simulations marked a transition in approach from broadly qualitative to increasingly quantitative research. These methods are now widely used in volcanology to describe the physical and chemical behaviours that govern volcanic and magmatic systems. Creating simplified depictions of highly dynamical systems enables volcanologists to simulate and potentially predict the nature and impact of future eruptions. These tools have provided significant insights into many aspects of the volcanic plumbing system and eruptive processes. The largest scientific advances in volcanology have come from a multidisciplinary approach, applying developments in diverse fields such as Engineering and Computer Science to study magmatic and volcanic phenomenon. A global effort in the integration of analogue and numerical volcano modelling is now required to tackle key problems in volcanology, and points towards the importance of benchmarking exercises and the need for protocols to be developed so that models are routinely tested against real world data.