Abstract
Caldera collapse represents severe volcanic hazards for the environment, climate, and human society, but it can also be beneficial as it may contribute to the formation of ore deposits and produce fertile soils. A deeper understanding of mechanical conditions under which caldera collapse can occur is thus of great importance and interest and can be significantly advanced through mathematical modeling.
Following a review of the state-ofthe-art numerical modeling approaches, this contribution takes the advantage of the finite element method (FEM) to develop a general model predicting fracture development above inflating and deflating magma chambers. Dozen cases covering both underpressure and overpressure scenarios and a wide range of possible magma chamber geometries and roof aspect ratios R (roof thickness/chamber diameter), from shallow to deepseated, mid-size and large, tabular and cylindrical, were calculated.
Based on selected 11 representative cases, we demonstrate that pressure evolution inside a magma chamber is manifested by a range of fracturing processes in the host rock, including not only the growth of ring faults, but also propagation of radial and circumferential fractures, magmatic stoping, and cauldron subsidence. The modeling strategy also enabled us to describe the orientation (inward-dipping, vertical, outward-dipping), mode (shear or dilation), and direction (upwards, downwards) of a ring fault initiation and growth.
The modeling shows that, regardless of magma chamber shape and caldera collapse scenario (over- or underpressure), the ring faults are reverse and always initiate at the chamber margin and propagate upwards, except for chambers with a low roof aspect ratio R < 0.25, with ring faults propagating both upwards and downwards. The ring fault orientation also changes with R, typically from moderate to steep.
Faults formed above underpressurized chambers are dominantly outward-dipping or (sub) vertical, whereas those formed above overpressurized chambers are either inward-dipping or (sub)vertical. These changes in the ring fault geometry and orientation also imply a change in the dominant caldera collapse mechanism from downsag for low R through piston for moderate R to cauldron subsidence for high R, where the ring fault does not reach the surface but instead defines an arch-like roof block prone to sink into the chamber.
Furthermore, our modeling approach also identifies highly fractured regions that develop within the chamber roof in some cases and potentially may represent traps for hydrothermal fluids and associated ore deposits. The presented study also confirms the FEM as an excellent tool for predicting the caldera collapse, especially when non-linear behavior and failure of host rock and nearly incompressible fluid behavior of magma are incorporated.