Seismic tomography has allowed us to image many important and interesting features of the mantle, such as slabs reaching the lower mantle, heterogeneity at the core-mantle boundary, and plumes of many shapes and sizes. Until recently, with a few exceptions, though, most tomographic models focused on isotropic velocity structure. By considering these velocity variations to be primarily thermal effects, isotropic tomography can be considered as a snapshot of the current thermal state of the mantle, which allows an estimate of the density variations that drive mantle convection.

However, most mantle minerals have significant elastic anisotropy, and can only be seen as isotropic seismically due to random orientation of crystals on the scale of seismic wavelengths. It is possible, at least in some regimes of the mantle, to orient the materials to produce macroscopically observable anisotropy. If we observe this on a global scale, it provides an opportunity to directly link the currently observed seismic models with their deformation history. To this end we have developed a 3D radially anisotropic model of S velocity in the whole mantle. This model can be linked to mantle flow in several regions, from the uppermost to the lowermost mantle.