Imaging azimuthal anisotropy in the alpine crust from ambient noise beamforming.

Noise cross-correlations provide a good azimuthal coverage, limited only by the distribution of noise sources and the layout of the stations used. It is therefore a promising method to constrain azimuthal anisotropy. As noise cross-correlations consist mainly of surface waves, they are especially sensitive to the crust and provide good depth constraints, as opposed to SKS-splitting data that are more sensitive to the upper mantle. We use the AlpArray network as well as stations from permanent networks all across Europe to perform time-domain beamforming on noise cross-correlations. The extent and density of the AlpArray network allows us to obtain reliable measurements all across the Alps. We divide the area in smaller zones using all stations outside the zone as sources and all stations inside as a sub-array for beamforming. This allows us to estimate the quality of our measurements in a region where strong lateral heterogeneities make measurements challenging, by estimating the magnitude of bias due to heterogeneities using the cos(theta) amplitude and evaluating uncertainties with bootstrap. This way, we measure Rayleigh wave azimuthal anisotropy in several period bands between 15 s and 60 s period. Inversion of dispersion curves in specific areas allows us to constrain the depth of the observed anisotropy. The results are broadly similar to results from SKS-splitting as they are generally parallel to the mountain belt. However, we observe lower anisotropy at short periods (40 seconds and less) in the Alps themselves than in surrounding regions. We also observe several structures in the crust that are not observed with SKS-splitting data. The most striking is a strong and spatially coherent NE-oriented anisotropy to the NW of the Alps that is possibly related to Variscan inheritance (at 40 seconds and less, in the upper and lower crust).  In the Northern Apennines, we observe anisotropy perpendicular to the belt at 30 s period (middle crust) that correlates well with an area of strong radial anisotropy recently observed by Alder et al (in review) at 30 km depth.