A global Bayesian radially anisotropic mantle model from transdimensional inversion.

We present the results of a new global Bayesian tomographic inversionfor a radial anisotropic shear wave velocity model of the Earth. Thedevelopment of radial anisotropic shear wave models are useful forunderstanding mantle dynamics. Our model improves on existing work byproviding robust estimates of uncertainties.A key aspect of this new model is the careful selection of surfacewave observations produced by earthquakes from the Harvard CMTcatalog. We automatically process a large set of seismic signalsrecorded for these events by simultaneously inverting Love andRayleigh waveforms up to the fifth overtone. After initial processing,only events where high quality Love and Rayleigh surface waves areobserved are retained for the tomographic inversion. This careful dataselection ensures that inference of radial anisotropy is not biaseddue to variations in spatial coverage of Love and Rayleigh observations.Our tomographic inversion uses a novel Bayesian Trans-dimensional Treeapproach that dynamically adapts the complexity of the solution toincrease spatial resolution where observations provide additionalconstraint. This leads to a robust estimation of both the shear wavemodel and its uncertainties. Furthermore, we demonstrate how ourapproach can be extended to a joint inversion to include body shearwaves to further constrain radial anisotropy of the Earths mantle. [ABSTRACT FROM AUTHOR]