Hindcasting May 2018 Kīlauea Summit Explosions With Atmospheric Remote Sensing, Geophysical Monitoring and 3D Eruptive Plume Simulations

Date: 4/26/2019

Time: 11:30 AM

Room: Puget Sound

Between May 17- 24, 2018, a series of eruptions at the summit of Kīlauea Volcano, HI, produced ash plumes as high as 30 km which were monitored by visual, radar, and infrasonic data. Co-eruptive ground deformation was recorded by nearby broadband seismometers. Eruptions manifested as 40 s long M4-5 seismic impulses, coincident with increased high frequency seismicity and infrasonic signals. 20-30 s later, eruptive plumes surfaced that persisted for tens of minutes, and very-long-period seismic oscillations occurred. After May 24, M5 events continued but with different high-frequency seismicity and no explosive eruptions. Here we invert seismic data for source parameters that we then use to drive 3D multiphase atmospheric plume models in order to seek a self-consistent explanation for geophysical and atmospheric data.

For all events, we invert >15 s period co-eruptive seismic data with quasi-static source models: moment tensors, tensile cracks, ring faults, spherical reservoirs, and combinations of these models. Moment tensors show primarily isotropic inflation with appreciable double-couple and CLVD components. Ground displacements are best fit by either a shallowly dipping 1 km deep crack or a 1.8 km deep sphere. Source mechanisms are relatively consistent until May 24 when energy at depths <1 km and double-couple motion increase, likely from collapse with signifigant slip along ring faults.

Plume simulations forced by seismically-derived time-varying basal conditions can produce plumes with heights and subsurface ascent times consistent with observations. These suggest (1) inflation accompanies collapse of rock overlying a shallow magma reservoir, (2) a significant fraction of displaced magma enters eruptive plumes, and (3) simulations can place bounds on fragmentation depths. We find that ground deformation was sufficiently well recorded to constrain source dynamics, but that available atmospheric data provides suboptimal constraints on eruptive plume dynamics.

Presenting Author: Josh A. Crozier

Authors