New constraints on Vp/Vs ratios and stress distribution at Etna volcano from anisotropic joint P- and S-wave tomography

The preferential alignment of eruptive fissures, dikes, sills, and (fluid-filled) microcracks with the local stress field - typical for the crust beneath volcanoes - causes the velocity of seismic waves to vary with propagation direction with respect to these aligned fractures (seismic anisotropy). Investigations based on shear wave splitting (Savage et al., 2010; Nardone et al., 2020; among others) and anisotropic P-wave tomography (Lo Bue et al., 2024; Del Piccolo et al., 2025) have demonstrated the presence of strong crustal anisotropy beneath Mt. Etna and other volcanic systems. However, the assumption of an isotropic crust is still common in seismic tomography, while methods for anisotropic tomography are still scarce and mostly limited to P-waves.

The well-established sensitivity of S-waves to anisotropy (shear wave splitting) suggests that extending anisotropic tomography methods to S-wave data may provide valuable new constraints on the anisotropic structure of volcanoes. Moreover, comparison of results of previous synthetic P- and S-wave tomography studies (Vanderbeek and Faccenda, 2021; Vanderbeek et al., 2023) indicates that S-wave data may be better capable of constraining shallow anisotropic heterogeneity. Additionally, Vp/Vs anomalies — often interpreted in terms of rock and fluid properties —exhibit a particularly strong directional dependence (Wang et al., 2012) in anisotropic media. This stresses the importance of coupling P- and S-wave data in anisotropic tomographic inversions.

We aim to show preliminary results of joint P- and S-wave anisotropic tomography, providing new constraints on the stress state and rock fluid properties of the crust beneath Mt. Etna. The method is an extension of the probabilistic anisotropic P-wave tomography method of Del Piccolo et al. (2025). Transdimensional Bayesian Monte Carlo sampling is applied to allow for robust uncertainty estimation, without the need for the subjective choice of damping and smoothing parameters that often limits more common deterministic tomography approaches.