Field examples regarding Horizontal-to-Vertical Spectral Ratio measurements as a tool for shallow faulting investigation

In this study, we present the results of 76 Horizontal-to-Vertical Spectral Ratio (HVSR) measurements carried out near three active fault lineaments on the eastern flank of Mt. Etna (Santa Tecla-Linera, Fiandaca and Trecastagni). In this area, earthquakes are very shallow (within 2 km of depth) and produce surface fracturing and deformations around the identified emerging fault zones, even for moderate magnitude (Mw > 3.5), leading to widespread damage to  buildings. Therefore, accurately identifying the zones affected by surface faulting is fundamental for improving territorial management.

The dataset was processed using HVSR technique to obtain the HVSR curves and the related spectra, as well as to extrapolate information on the directional effects of the signals as a function of both frequency and azimuth. In the literature, the HVSR method has been successfully used to detect polarization effects across fault zones: previous studies have shown that horizontal polarization in Mt. Etna is often strong and tends to be perpendicular to the predominant fracture field or has high-angle polarization from the fault strike (Rigano et al. 2008; Di Giulio et al., 2009). We thus applied the wavefield polarization analysis to the ambient noise measurements to investigate the areal pattern of horizontal polarization and to identify any existing spatial variations. Moreover, the polarization angles were also estimated by using the Matlab code POLARGUI (Huailiang Li et al., 2021). This method allowed us to map the polar histograms and display the distributions of polarization azimuths in different frequency bands. The code is based on the decomposition of the eigenvectors and eigenvalues of the covariance matrix of the three ground motion components of Jurkevics (1988).

Lastly, since fault zones may produce fault-zone trapped waves, which may consist primarily of Love-type waves with particle motion parallel to the fault strike (e.g. Lewis & Ben-Zion 2010) or may include Rayleigh-type components with different polarization angles (e.g Ellsworth & Malin 2011), we computed the ellipticity curves obtained with the RayDec method (Hobiger et al., 2009) to isolate the contribution of Rayleigh waves alone.

The ellipticity of the Rayleigh waves was analyzed for all the measurements to identify any differences that might indicate the presence of a surface faulting zone. To emphasize the deviation between the HVSR curves and the ellipticity curves, a residuals analysis was performed based on the Root Mean Squared Error (RMSE).

These results enabled the identification and proposal a new indicator as proxy for the presence of the fault system and/or any associated fracture field, to be integrated with other types of geophysical measurements.

Acknowledgements: The measurements are part of the geophysical acquisitions for the ETNA-FAC project, signed by CNR IGAG with the Regional Department of Sicilian Civil Protection. The project involved: CNR IGAG, INGV, University of Catania, OGS, ISPRA, CNR ISPC and CNR IMAA.