Ambient Noise Tomography of central Europe

We measure ambient-noise cross-correlations at central European seismic stations in the Orfeus, SDS-Net (Switzerland) and INGV (Italy) networks, assembling a large, dense database of Love and Rayleigh-wave, group- and phase-velocity dispersion curves. The cross-correlated signal was recorded over two years, January 2006 to December 2007. To validate our database, we derive 2D group-velocity maps at a suite of periods between 8 and 35s, and compare them with analogous, independent studies of smaller areas contained in our region of interest. Well known geological features like the Alps, the Po Plain and the Ivrea Body are robustly resolved, which is an indication of the reliability of our method and subsequent observations. We additionally test our phase-velocity maps by comparison with results from 2-station correlation of teleseimsic recordings. For each pixel of our phase- and group-velocity maps, we extract velocity as a function of period, defining a local dispersion curve. We next invert the dispersion curve associated with each pixel, to determine a local, vertical profile of seismic velocity in the lithosphere. Inversions are conducted with the nonlinear, neighborhood algorithm method. After the exercise is repeated for each pixel, profiles can be assembled to identify a 3-D model of the European lithosphere and lower crust. Because of the unprecedentedly good coverage achieved via the ambient-noise approach (tomography without earthquakes), our new model has a resolution of the region of interest higher than other existing seismic maps of the area. In view of its good azimuthal coverage, our database should effectively constrain azimuthal anisotropy, with implications on lithosphere tectonics and convection dynamics of the central European uppermost mantle.