3D transdimensional ambient noise surface wave tomography of the Reykjanes Peninsula – a feasibility study

Abstract

Summary Seismic surface wave tomography is an effective tool for 3D crustal imaging. Conventionally, a two-step inversion algorithm is used to recover a three-dimensional model of surface wave velocity. That is, starting from surface wave dispersion data (frequency-dependent phase velocities), an initial inversion resulting in a series of (2D) maps of frequency-dependent surface-wave velocity is followed by a separate (1D) depth inversion. A single-step 3D non-linear algorithm has recently been proposed in a Bayesian framework. The algorithm involves a reversible jump Markov chain Monte Carlo approach and is referred to transdimensional tomography. Here, we investigate the feasibility of this transdimensional algorithm for the purpose of recovering the 3D surface wave velocity structure below the Reykjanes Peninsula, southwest Iceland. In particular, we investigate this for the specific receiver configuration for which we have obtained year-long recordings of ambient seismic noise. To that end, we designed a number of synthetic tests using receiver-receiver travel times associated with that station configuration. We find that the transdimensional algorithm successfully recovers the 3D velocity structure of the area. In particular, the algorithm successfully adapts its resolution to the density of rays and the level of data noise. Moreover, quantified solution uncertainty makes the result better interpretable.