Diffuse elastic wavefield within a simple crustal model. Some consequences for low and high frequencies

The reliability of usual assumptions regarding the wavefield composition in applications of the Diffuse Field Approach (DFA) to passive seismic prospecting is investigated. Starting from the more general formulation of the DFA for full wavefield (FW), the contribution of each wave to the horizontal- and vertical-component power spectra at surface are analyzed for a simple elastic waveguide representing the continental crust-upper mantle interface. Special attention is paid to their compositions at low and high frequencies, and the relative powers of each surface wave (SW) type are identified by means of a semianalytical analysis. If body waves are removed from the analysis, the high-frequency horizontal asymptote of the H/V spectral ratio decreases slightly (from 1.33 for FW to around 1.14 for SW) and shows dependence on both the Poisson’s ratio of the crust and the S wave velocity contrast (while FW-H/V asymptote depends on the former only). Experimental tests in a local broadband network provide H/V curves compatible with any of these values in the band 0.2–1Hz, approximately, supporting the applicability of the DFA approximation. Coexistence of multiple SW modes produces distortion in the amplitudes of vertical and radial component Aki’s coherences, in comparison with the usual predictions based on fundamental modes. At high frequencies, this effect consists of a decrement by a constant scaling factor, being very remarkable in the radial case. Effects on the tangential coherence are severe, including a π/4 phase shift, slower decay rate of amplitude versus frequency, and contribution of several velocities for large enough distances.