Abstract

Plate tectonics, and its familiar dynamic consequences including earthquakes, volcanoes, and even the surface topography, are intrinsically linked to convection in the silicate mantle below. Mantle convection is a complex thermochemical process by which hot material rises from the deep Earth to the surface via upwellings like plumes, and cold material is returned to the deep via subduction. Tomographic images of seismic velocity provide a snapshot of the thermochemical state of the mantle at the present day, but do not directly constrain dynamic processes such as deformation. Seismic anisotropy, the variation of seismic wave speed with direction, emerges as a result of the long-wavelength ordering of smaller features (such as crystals, fractures, or melt inclusions) and thus can give information about processes such as deformation and flow. Recent advances in seismic tomography have provided a global picture of anisotropy throughout the mantle, but quantitatively interpreting these for mantle flow requires models. The NERC -funded MC2 project has focused on building a large suite of mantle circulation models (MCMs) – convection simulations constrained by models of the Earth’s plate motion over the last 1 billion years – exploring a range of different parameters of mantle convection. It is providing a framework to compare these models to a broad spectrum of observations (seismic, geodynamic, geochemical, and geomagnetic). In this talk, I will outline how we are using these models to predict the seismic anisotropy resulting from the flow in the upper mantle in these models and comparing it to tomographic models of radial anisotropy. These comparisons demonstrate the influence of parameters including the radial viscosity profile of the shallow mantle and the core-mantle boundary temperature on the resulting anisotropy. The models we produce show a consistent discrepancy with the tomographic images at around 100 km below mid-ocean ridges, suggesting that the anisotropy observed here for the Earth cannot be explained by solely by the lattice preferred orientation of olivine. The most plausible alternative explanation for these signatures is the presence of deep melt below the ridges.