| Popis: |
This thesis is directed at understanding dynamics of the Earth's mantle. I adopt multidisciplinary approaches toward the problem: geodynamical and seismological. My approach in geodynamics is directed at understanding the relationship between large scale surface observables (geoid, topography, plate motions) and mantle rheology and convection of the present-day Earth. In chapter 2, I do best-fit correlations of shallow mantle structure with various tectonic features and remove them to generate what we call "residual tomography." In chapter 3, I show that the pattern, spectrum and amplitude of the "residual topography" are consistent with shallow origin of the "Earth surface dynamic topography;" the very long wavelength geoid and topography (l = 2 - 3) are successfully explained by density models inferred from the "residual tomography," assuming layered mantle convection stratified at the "920 km seismic discontinuity." In chapter 4, I develop a new method to calculate mantle flow in the spherical coordinates with lateral variation of viscosity. The viscosity contrast between continental and oceanic regions is identified to have dominating effects on both the observed poloidal/toroidal ratio and pattern of toroidal motions at long wavelengths. I show convection models with lateral variation of viscosity are capable of producing long wavelength plate motions observed in plate tectonics. My approach in seismology is focused on exploring fine structures near the core mantle boundary and developing new techniques for computing synthetic seismograms. I discuss the method development and strategies to explore fine structures near the core-mantle boundary region in the following chapters. In chapter 5, I develop a hybrid method which can handle the seismic wave propagation in heterogeneous regions at large distances. The hybrid method is a combination of analytical and numerical methods, with numerical methods applied in heterogeneous regions only and analytical methods outside. In chapter 6, I discuss wave propagation of SKS and SPdKS phases through ultra-low velocity zones near the core-mantle boundary and constrain the general structures of ultra low velocity zones near the core-mantle boundary under Fiji subduction zone and Iceland. The long period SKS-SPdKS data are explained by ultra low velocity zones with P velocity reduction of 10% and horizontal length scales of about 250 km and height of about 40 km. S velocity reduction of 30% is consistent with the data, although the trade-offs between S velocity reduction and height of the structure exist. In chapter 7, I discuss wave propagation of PKP and its precursors and constrain the detailed structures of the ultra low velocity zones near the core-mantle boundary from observed broadband PKP precursors. The observed long period precursors are explained by the existence of ultra low velocity zones with P velocity reduction of at least 7% and horizontal length scales of 100-300 km and height of about 60-80 km, whereas short period precursors suggest that the structures have smooth edges and structures with smaller scale are adjacent to these large Gaussian-shaped structures. These fine structures may be indicatives of vigorous small-scale convection or the instabilities of the bottom thermal boundary layer of the mantle. |
