Mantle convection is an important heat and material transport mechanism between the Earth’s deep interior and the surface. The Earth’s deep mantle is seismically heterogeneous and the lower boundary layer of the large-scale mantle dynamics but remains poorly resolved by seismic observations and poorly understood theoretically. Seismic heterogeneities in the Earth’s deep mantle include two Large Low Shear Velocity Provinces (LLSVPs) surrounded by subducted slabs with faster-than-average seismic velocities and numerous patches of ultra-low-velocity-zones (ULVZs) (Chapter 1). The overarching goal of this dissertation is to understand the origin, nature, and evolution of these structures using numerical geodynamic models (Chapter 2).Chapter 3 studies how to link lowermost mantle flow patterns with lower mantle slab deformation styles. Slab deformational styles in the Earth’s lower mantle are often under-resolved in seismic tomography models. In lab experiments, it was found that slab deformation was controlled by slab strength. The strong seismic anisotropy observed in the lowermost mantle is hypothesized to be induced by slab deformation and may be informative in the lowermost mantle flow patterns. To correlate seismic anisotropy with deformation, the first step is to link deformation with flow patterns. I performed 3D spherical calculations and found that slab strength affected lower mantle slab deformation styles and lowermost mantle flow patterns. However, slab trench length has a trade-off effect on lowermost mantle flow patterns, making it difficult to infer lower mantle slab deformation styles solely from lowermost mantle flow patterns. Chapter 4 investigates if the symmetry of ULVZ cross-sectional shapes implies the viscosity of LLSVPs with respect to the background mantle. The shape of ULVZs is controlled by their properties (e.g., density, viscosity) and their interaction with the LLSVPs and the surrounding mantle. Therefore, ULVZ shapes provide information about the rheological properties of the LLSVPs. In geodynamic models, ULVZs and LLSVPs material usually have the same composition-dependent viscosity as the background mantle. However, this might be unreasonable if they originate from compositionally distinctive materials. I incorporated the composition-dependence of viscosity into calculations and found that the composition-dependence of viscosity affects the viscous coupling among LLSVP materials, ULVZ materials, and the background mantle. The viscous coupling, therefore, controls the symmetry of ULVZ shapes. In Chapter 5, I explored how dense thermochemical piles with increased pile intrinsic viscosity respond to changing convective flow patterns in terms of lateral mobility and the stability of their cross-sectional morphologies. Whether LLSVPs are fixed at their current position for up to a few hundred million years is under debate. In geodynamic models, LLSVPs, if caused by thermochemical piles, appeared as passive structures swept around by subducted slabs. However, other parameters, such as compositional viscosity contrast between piles and the background mantle (as explored in this study), have not been sufficiently explored in these models. Using the method developed in Chapter 4, I perturb upwelling and downwelling convective patterns in models with increased pile intrinsic viscosity and examine the long-term stability of thermochemical piles. I found that piles are mobile even if their compositional viscosity increase by 5,000x, but their cross-sectional morphology are more stable as pile viscosity increases. Chapter 6 concludes these research projects and discusses perspectives for future work.
Read
