Ambient Noise and Earthquake Surface Wave Phase Velocity Tomography of the South American Lithosphere

The development of new automated techniques to measure earthquake-based interstation phase velocities allows for more detailed imaging of the South American Lithosphere. We calculated Rayleigh-wave phase velocities using the earthquake records from 1022 broadband seismic stations (South America, Antarctica and the Caribbean) operated between 1990 and 2020. A total of 1.069.259 earthquakes were selected with the following criteria: (1) events aligned within 10° of the 2-station great circle path; (2) a linearly increasing minimum magnitude between 4 and 6 Mw as a function of the epicentral distance; (3) maximum magnitude of 8 Mw; and (4) epicentral distances between 2.5° and 30°.
We used surface wave fundamental mode dispersion curves calculated automatically using a new implementation of the 2-station cross-correlation method and a number of strict quality criteria, which include the selection of the phase velocity curves based on a 3D background model, curve smoothness and width of the considered frequency range. Following this process, we obtained 46.763 broad-band dispersion measurements between 4 and 315 s. Finally, the single-event dispersion curves were averaged for each interstation pair with associated error estimates and further quality control comparing results from both directions, evaluating smoothness and the standard deviation of the resulting dispersion curve.
The dispersion curves were simultaneously inverted for isotropic and anisotropic (2ψ and 4ψ) phase-velocity maps parameterized on a triangular grid with a knot spacing of 30 km.
The isotropic phase velocity maps at periods of 15 and 30 s indicate around 8% high-velocity perturbations in the regions of: (1) cratonic blocks of the South American platform (Brazilian Shield, São Francisco and Rio Apá cratons); and (2) the basement of the Pantanal basin (a 500 m thick sedimentary basin) possibly related to a high-velocity lower crust. For those periods, we also observed between -8 to -4% low-velocity perturbations associated with the Andean Mountain range root below the Altiplano Boliviano region (central Andes), the forelands of the Andes. The Paraná, Chaco and Parecis intracratonic basins also have lower velocities with relation to the neighboring cratonic areas.
At periods of 60 and 100 s, we observed around 4% high-velocity perturbation associated with the deep roots of the oldest region of the Amazonian craton (eastern area of the Brazilian Shield) and the São Francisco craton. Azimuthal anisotropy is laterally and vertically variable within the South American lithosphere. At longer periods, fast directions are pointing to asthenospheric flow guided by LAB topography, for example below the Pantanal basin (central-west Brazil).
For the next steps, we plan to jointly invert for a isotropic and anisotropic 3D model using earthquake and ambient noise dispersion curves with Rayleigh and Love waves.