Imaging the upper crust in the eastern Pyrenees with ambient seismic noise

We present here a 3D shear-velocity regional model of the eastern Pyrenees centered at the Cerdanya Basin using seismic ambient noise to image structure of the Iberian and Eurasian uppermost crust. In the context of the SANIMS project, we deployed a network of 24 broadband seismic stations that complement existing permanent stations from the CA, ES and FR networks and previous temporal deployments from the OROGEN project, and aside, a high-density short-period seismic network focused on the study the Cerdanya Basin. For the regional study, we use the broadband dataset to compute symmetric cross-correlations of the seismic noise using the wavelet phase cross-correlation and the time-scale phase weighted stack, then estimate the phase and group velocity of Rayleigh waves between 1.5 – 7 s and 1.5 – 5 s, respectively. Afterwards, we build a 3D S-wave velocity model from path-average velocities in two steps: regionalization and pointwise depth inversion. In the regionalization step, we build velocity maps from the dispersion curves measured using the fast marching method on the forward problem and a hybrid l₁ – l₂ norm criterion on the inversion to enforce robustness to outliers. In the depth inversion step, we apply transdimensional inference to explore for S-wave velocity profiles with an unknown number of layers of constant velocity and the noise variance of the Rayleigh phase and group velocity maps using a least-square misfit function. The best models show a top low-velocity layer 2 – 3 km thick followed by distinct velocity profiles to the North and to the South, corresponding to the expected differences between the Iberian and Eurasian plates. To the South we observe two layers with a boundary at 6 – 7 km depth and velocities of about 3.2 and 3.5 km/s respectively, while to the North velocities are generally lower, increase much less with depth and there is no clear boundary.