Geodynamical modeling of mantle convection to investigate mantle layering and the cause of lower mantle anisotropy

This dissertation focuses on the use of numerical modeling to better understand mantle convection. In particular it investigates two significant questions in the field of solid Earth geophysics. The first asks whether layered convection models that are suggested from geochemical observations are thermally permissible. The second deals with the causes of lowermost mantle seismic anisotropy to better understand the geochemical and dynamic nature of the lower mantle.; The first work consists of a parameterized convection study that examines the thermal feasibility of both whole mantle and layered mantle models. A large parameter range is explored in this study, and it is found that layered models are difficult to reconcile with geophysical constraints. These models will either produce lower layers that are too hot or upper layers that are much too cool. Only a few parameter combinations provide a layered mantle that satisfies constraints. In contrast, it is found that whole mantle models satisfy constraints under a large parameter range, and are therefore, the most likely mode of cooling.; The second work consists of a line of research that explores the cause of lowermost mantle seismic anisotropy. This anisotropy is observed at the core-mantle boundary in regions of paleosubduction and mantle upwelling. Possible causes of anisotropy include lattice preferred orientation (LPO) and shape preferred orientation (SPO). By combining numerical modeling with mineral physics experiments, it is found that LPO provides the likely cause of the anisotropy observed in regions of paleosubduction. Moreover, it is found that the cause of anisotropy observed in upwelling regions is unconstrained.