Acoustic multipole source inversions of volcano infrasound

Abstract

Sources of volcano infrasound involve the atmospheric displacement associated with volcanic eruptions, where characteristic source dimensions are generally confined by the vent. Volcano infrasound sources are typically considered as a monopole which corresponds to the first-order term in the acoustic multipole expansion. However, when the wavelength becomes comparable to the size of the vent, the source may have further complexity which can be described only by higher-order terms, yet such complexity of volcano infrasound source has not been extensively explored. This is mainly due to (1) limited sampling of the acoustic wavefields due to poor network coverage and (2) complex sound propagation near the volcanic edifice which significantly distorts the multipole acoustic wavefields. In this study, we present a linearized waveform inversion technique incorporating numerical Green's functions. A full 3-D Finite-Difference Time-Domain (FDTD) method accelerated with GPU is used to compute accurate Green's functions taking into account volcano topography. The presented method is applied to infrasound data recorded at Sakurajima volcano (Japan) and Tungurahua volcano (Ecuador) and volcano infrasound sources associated with explosive eruptions are characterized in terms of a monopole and dipole. These methods could be applied to chemical explosions as well to determine source characteristics and complexity.