Phase velocity tomography of surface waves using ambient noise cross correlation and array processing

Continuous recordings of ambient seismic noise across large seismic arrays allows a new type of processing using the cross-correlation technique on broadband data. We propose to apply double beamforming (DBF) to cross correlations to extract a particular wave component of the reconstructed signals. We focus here on the extraction of the surface waves to measure phase velocity variations with great accuracy. DBF acts as a spatial filter between two distant subarrays after cross correlation of the wavefield between each single receiver pair. During the DBF process, horizontal slowness and azimuth are used to select the wavefront on both subarray sides. DBF increases the signal-to-noise ratio, which improves the extraction of the dispersive wave packets. This combination of cross correlation and DBF is used on the Transportable Array (USArray), for the central U.S. region. A standard model of surface wave propagation is constructed from a combination of the DBF and cross correlations at different offsets and for different frequency bands. The perturbation (phase shift) between each beam and the standard model is inverted. High-resolution maps of the phase velocity of Rayleigh and Love waves are then constructed. Finally, the addition of azimuthal information provided by DBF is discussed, to construct curved rays that replace the classical great-circle path assumption.