Linking gas and particle ejection dynamics to boundary conditions in scaled shock-tube experiments

Predicting the onset, style and duration of explosive volcanic eruptions remains a great challenge. While the fundamental underlying processes are thought to be known, a clear correlation between eruptive features observable above Earth's surface and conditions and properties in the immediate subsurface is far from complete. Furthermore, the highly dynamic nature and inaccessibility of explosive events means that progress in the field investigation of such events remains slow. Scaled experimental investigations represent an opportunity to study individual volcanic processes separately and, despite their highly dynamic nature, to quantify them systematically. Here, impulsively generated vertical gas-particle jets were generated using rapid decompression shock-tube experiments. The angular deviation from the vertical, defined as the "spreading angle", has been quantified for gas and particles on both sides of the jets at different time steps using high-speed video analysis. The experimental variables investigated are 1) vent geometry, 2) tube length, 3) particle load, 4) particle size, and 5) temperature. Immediately prior to the first above-vent observations, gas expansion accommodates the initial gas overpressure. All experimental jets inevitably start with a particle-free gas phase (gas-only), which is typically clearly visible due to expansion-induced cooling and condensation. We record that the gas spreading angle is directly influenced by 1) vent geometry and 2) the duration of the initial gas-only phase. After some delay, whose length depends on the experimental conditions, the jet incorporates particles becoming a gas-particle jet. Below we quantify how our experimental conditions affect the temporal evolution of these two phases (gas-only and gas-particle) of each jet. As expected, the gas spreading angle is always at least as large as the particle spreading angle. The latter is positively correlated with particle load and negatively correlated with particle size. Such empirical experimentally derived relationships between the observable features of the gas-particle jets and known initial conditions can serve as input for the parameterisation of equivalent observations at active volcanoes, alleviating the circumstances where an a priori knowledge of magma textures and ascent rate, temperature and gas overpressure and/or the geometry of the shallow plumbing system is typically chronically lacking. The generation of experimental parameterisations raises the possibility that detailed field investigations on gas-particle jets at frequently erupting volcanoes might be used for elucidating subsurface parameters and their temporal variability, with all the implications that may have for better defining hazard assessment.

The online version contains supplementary material available at 10.1007/s00445-021-01473-0.