Joint body- and surface-wave probabilistic transdimensional tomography of upper mantle seismic anisotropy

Body- and surface-wave seismic data provide complementary sensitivities to the anisotropic elastic structure of the Earth, and the potential constraints of a simultaneous inversion would extend significantly beyond those of the individual phases. However, joint body- and surface-wave anisotropic imaging remains limited, mainly because of the high nonlinearity of the problem and the different inversion methods traditionally adopted for body- and surface-wave phases. Here, we implement a nonlinear transdimensional stochastic solver based on the reversible-jump Markov chain Monte Carlo (RJMCMC) algorithm to simultaneously invert P-, S- and Rayleigh-wave data. By sampling irregularly meshed anisotropic velocity models for the upper mantle, with different mesh configurations and complexities adaptable to the heterogeneous data constraints, we populate an ensemble of variable solutions describing the data within the uncertainties. The method is validated using independent synthetic seismograms simulated with SPECFEM3D Globe in an anisotropic upper mantle plume model. We show how the different sensitivities of the data translate into different constraints on upper mantle seismic structure, and we analyze metrics to quantitatively assess uncertainties in the inferred solutions.