Joint inversion of Rayleigh wave ellipticity and spatial autocorrelation measurements

Abstract

The local soil structure (i.e. shear and pressure wave velocities) can be obtained by inversion of dispersion curves ranging over a sufficiently large frequency band. However, measurements of such dispersion curves using ambient seismic vibrations require a large number of seismometers and a long measuring time. As a simple alternative, we propose to invert Rayleigh wave ellipticity obtained by ambient seismic noise measurements at a single site using a method based on the random decrement technique (Hobiger et al., 2009). Indeed, the freSeismological Research Letters Volume 81, Number 2 March/April 2010 305 quency-dependency of Rayleigh wave ellipticity is tightly related to the shear wave profile of the soil. However, as different soil structures can result in the same ellipticity curve (e.g. homothetic structures in velocity and thickness), the inversion of ellipticity curves alone is ambiguous. Therefore, additional measurements fixing the shearwave velocity in the superficial layers have to be included into the inversion process. We suggest using a small number of seismic stations to measure spatial autocorrelation curves. In this way, three seismic sensors and one hour of measurements can be sufficient to invert the local soil structure. We will present the method to extract the Rayleigh wave ellipticity curve, show which parts of the ellipticity curve have to be included in the inversion process and demonstrate the benefit of the additional spatial autocorrelation curve measurements. Then, we will present an example application to real noise data collected within the framework of the European NERIES project at well-known European accelerometric sites and the results will be compared to th