Imaging three-dimensional anisotropy with broadband seismometer arrays

Abstract

Broadband arrays dramatically increase the ability of a variety of seismic phases to resolve three-dimensional variations in anisotropic structure. We explore the sensitivity of SKS phases to lateral variations in anisotropy using full waveform synthetic seismograms computed with a parallel pseudospectral method. SKS sensitivity decreases as the depth to the anisotropy or wave period increases, as was documented in earlier studies. However, we also find that the total depth extent of anisotropy, not simply its minimum depth, must be considered when interpreting gradients in splitting. SKS wavefront healing explains the reduction in SKS sensitivity with greater depth to the anisotropy. In models with laterally alternating vertical bands of anisotropy, SKS phases can detect zones of anisotropy whose lateral dimensions are smaller than the SKS sensitivity ranges for a single boundary over the same depths. However, lateral variations in anisotropy at a scale that is less than 40% of the corresponding single boundary sensitivity range typically produce a lack of well-resolved splitting. A shallow layer of strong and rapidly varying anisotropy does not appear capable of perturbing the splitting signal from a deeper layer of coherent anisotropy to the point where the fast direction of the deeper layer is significantly overprinted or no splitting is resolved. Rapid variations in lithospheric anisotropy over slowly varying asthenospheric anisotropy therefore cannot explain regions where weak SKS splitting and strong azimuthal anisotropy in long period surface waves are both observed. With the 70 km station spacing anticipated for USArray, array-based inversion of Rayleigh waves at periods of ∼100 s should be able to resolve isotropic anomalies at scales of ∼175 km, although potential trade-offs between isotropic and anisotropic structure remain even for anomalies with lateral dimensions of 300 km. Joint inversions of multiple phase types hold promise for further improvement in the resolution of three-dimensional anisotropic mantle structure.