Cross-validation of ambient-noise and earthquake-based observations of Love- and Rayleigh-wave dispersion

More Info
expand_more

Abstract

Phase velocities derived from ambient-noise (AN) cross-correlation are compared with phase velocities calculated from earthquake data via the two-station method (EQ). The EQ approach involves crosscorrelating recordings of teleseismic earthquakes whose epicenters are approximately on the station-station great circle. The AN and EQ methods can be combined effectively to cover a broad period range from about 4 to 250 s. The comparison is conducted for both Rayleigh and Love waves using over 1000 station pairs in central Europe. The analysis shows that at periods between 20 and 40s, AN-based phase velocities are systematically slower by about 1.5% compared to the EQ measurements. We revise the theory, explaining possible biases from the stationary-phase approximation and from using horizontal components of the noise wavefield. A series of numerical tests of the AN method is conducted, which allow us to evaluate the effects of the variable distance between station array and noise sources, of the illumination pattern, and of horizontal anisotropy and attenuation. We finally determine phase-velocity maps based on AN vs. EQ data; we find that discrepancies between such maps are far weaker than the mapped anomalies, and too small to result in significantly different geological interpretation of the mapped structure.