Three-dimensional S-wave velocity model of the Bohemian Massif from Bayesian ambient noise tomography

Abstract We perform two-step surface wave tomography of phase-velocity dispersion curves obtained by ambient noise cross-correlations in the Bohemian Massif. In the first step, the inter-station dispersion curves were inverted for each period (ranging between 4 and 20 s) separately into phase-velocity maps using 2D adjoint method. In the second step, we perform Bayesian inversion of the set of the phase-velocity maps into an S-wave velocity model. To sample the posterior probability density function, the parallel tempering algorithm is employed providing over 1 million models. From the model samples, not only mean model but also its uncertainty is determined to appraise the reliable features. The model is correlated with known main geologic structures of the Bohemian Massif. The uppermost low-velocity anomalies are in agreement with thick sedimentary basins. In deeper parts (4–20 km), the S-wave velocity anomalies correspond, in general, to main tectonic domains of the Bohemian Massif. The exception is a stable low-velocity body in the middle of the high-velocity Moldanubian domain and high-velocity body resembling a promontory of the Moldanubian into the Tepla-Barrandian domain. The most pronounced (high-velocity) anomaly is located beneath the Eger Rift that is a part of a Tertiary rift system across Europe.