A global Bayesian seismic shear-wave velocity model of the upper-mantle using seismic hum

Imaging the anisotropic shear-velocity structure of the lithosphere-asthenosphere system is key to understanding the Earth’s internal dynamics and the mechanics of plate tectonics, whose nature and working mechanism remain debated across the geosciences. Global tomography robustly resolves long-wavelength shear-velocity structure, but the uneven distribution of earthquake sources and receivers reduces the capacity of current methods to resolve short-wavelength and shallow structure. Ambient noise offers an independent dataset that provides complementary path coverage for investigating Earth’s structure and enhances our understanding of the physical nature of the lithosphere-asthenosphere system by improving resolution in regions that remain insufficiently resolved.

In this work, we use the Earth’s hum at 30-270 s periods to produce a global probabilistic model of the shear-wave velocity of the upper mantle and its uncertainties. We extract empirical Green’s functions from 1989-2004 continuous records at 389 broadband stations using the wavelet-phase cross-correlation and time-scale phase-weighted stacking. Then, frequency-time analysis and the spectral method yield 55615 R1 and 23467 R2 group-velocity, and 56539 R1 phase-velocity Rayleigh-wave dispersion curves, primarily on new paths. Finally, we solve the inversion problem using the two-step method, employing probabilistic continuous inverse theory to construct phase- and group-velocity maps in the regionalization step, and transdimensional inference for the depth inversion.

The phase and group velocity maps we obtain compare well with velocity maps derived from earthquakes. Similar velocity anomalies are observed at all periods, including the cratons, the African rift system and the Pacific belt, and the mid-ocean ridges. Given the strong complementarity between the ambient-noise and earthquake datasets, and the fact that the model derived from ambient noise alone is already accurate, a joint inversion has the potential to enhance the imaging of the lithosphere-asthenosphere system in global mantle models.