VOLcanic conduit processes and their effect on PROjectile eXit dYnamics (VOLPROXY)

Volcanic projectiles are centimeter- to meter-sized clasts – both solid-to-molten rock fragments or lithic eroded from conduits – ejected during explosive volcanic eruptions that follow ballistic trajectories. Despite being ranked as less dangerous than large-scale processes such as pyroclastic density currents (hot avalanches of gas and pyroclasts), volcanic projectiles still represent a constant threat to life and properties in the vicinity of volcanic vents, and frequently cause fatal accidents on volcanoes. Mapping of their size, shape, and location in volcanic deposits can be combined to model possible trajectories of projectiles from the vent to their final position, and to estimate crucial source parameters of the driving eruption, such as ejection velocity and pressure differential at the vent. Moreover, size and spatial distributions of volcanic projectiles from past eruptions, coupled with ballistic modelling of their trajectory, are crucial to forecast their possible impact in future eruptions. The reliability of such models strongly depends on i) the appropriate physical functions and input parameters and ii) observational validations. In this study, we aimed to unravel intra-conduit processes that strongly control the dynamic of volcanic projectiles by combining numerical modelling and novel experimentally-determined source parameter. In particular, the multiphase ASHEE model (Cerminara 2016; Cerminara et al. 2016) suited for testing post-fragmentation conduit dynamics based on a robust shock tube experimental dataset. By exploding mixtures of pumice and dense lithic particles within a specially designed transparent autoclave, and by using a raft of pressure sensors, ultra-high-speed cameras and pre-sieved natural particles, we observed and quantified: i) kinematic data of the particles and of the gas front along the shock tube and outside, ii) pressure decay at 1GHz resolution. By feeding the ASHEE model with these datasets, and using initial and boundary conditions similar to that of the experiment, we defined domains composed by a pressurized shock tube and the outside chamber at ambient conditions, and tested particles particle motion according to a Lagrangian approach, as well as gas flow with a Eulerian approach (a 3D finite-volume numerical solver, compressible). The comparison between data and model yields estimate of the particle kinematic inside the tube, the pressure evolution at the top and the bottom of the tube, and the eruption source parameters at the tube exit.

We designed a series of rapid decompression experiments in which we systematically varied componentry, particle size, and packing arrangement of the initial samples. We also carried empty run experiments, where only the gas phase is decompressed. We used the pressure evolution and high-speed footage of these experiments to i) benchmark the expansion of the gas phase against the expansion of gas and particle mixture, and ii) develop 1D and 3D models of particle acceleration.