The asthenosphere, a weak layer beneath the rigid lithosphere, plays a fundamental role in the operation of plate tectonics and mantle convection. While this layer is often characterized by low seismic velocity and high seismic attenuation, the global structure of the asthenosphere remains poorly understood. In this dissertation, twelve years of SS precursors reflected off the top and bottom of the asthenosphere, namely, the LAB and the 220-km discontinuity, are processed to investigate the boundaries of the asthenosphere at a global scale. Finite-frequency sensitivities are used in tomography to account for wave diffraction effects that cannot be modeled in global ray-theoretical tomography. Strong SS precursors reflected off the LAB and the 220-km discontinuity are observed across the global oceans and continents. In oceanic regions, the LAB is characterized by a large velocity drop of about 12.5%, which can be explained by 1.5%-2% partial melt in the oceanic asthenosphere. The depth of the Lithosphere Asthenosphere Boundary is about 120 km, and its average depth is independent of seafloor age. This observation supports the existence of a constant-thickness plate in the global oceans. The base of the asthenosphere is imaged at a depth of about 250 km in both oceanic and continental areas, with a velocity jump of about ∼ 7% across the interface. This finding suggests that the asthenosphere in oceanic and continental regions share the same defining mechanism. The depth perturbations of the oceanic 220-km discontinuity roughly follow the seafloor age contours. The 220-km topography is smoother beneath slower-spreading seafloors while it becomes rougher beneath faster-spreading seafloors. In addition, the roughness of the 220-km discontinuity increases rapidly with spreading rate at slow spreading seafloors, whereas the increase in roughness is much slower at fast spreading seafloors. This observation indicates that the thermal and compositional structures of seafloors formed at spreading centers may have a long-lasting impact on asthenospheric convections. In continental regions, a broad correlation is observed between the 220-km discontinuity depth structure and surface tectonics. For example, the 220-km discontinuity depth is shallower along the southern border of the Eurasian plate as well as the Pacific subduction zones. However, there is no apparent correlation between 3-D seismic wavespeed in the upper mantle and the depths of the 220-km discontinuity, indicating that secular cooling has minimum impact on the base of the asthenosphere. / Doctor of Philosophy / In classic plate tectonic theory, the outermost shell of the Earth consists of a small number of rigid plates (lithosphere) moving horizontally on the mechanically weak asthenosphere. In the classic half space cooling (HSC) model, the lithosphere is formed by gradual cooling of the hot mantle. Therefore, the thickness of the plate depends on the age of the seafloor. The problem with the HSC model is that bathymetry and heat flow measurements at old seafloors do not follow its predicted age dependence. A modified theory, called plate cooling model, can better explain those geophysical observations by assuming additional heat at the base of an oceanic plate with a constant thickness of about 125 km. However, such a constant-thickness plate has not been observed in seismology. In this thesis, the asthenosphere boundaries are imaged using a global dataset of seismic waves reflected off the Earth's internal boundaries. Strong reflections from the top of the asthenosphere are observed across all major oceans. The amplitudes of the SS precursors can be explained by 1.5%-2% of partial melt in the asthenosphere. The average boundary depths are independent of seafloor age, and this observation supports the existence of a constant-thickness plate in the global oceans with a complex origin. The 220-km discontinuity, also called the Lehmann Discontinuity, was incorporated in the Preliminary Reference Earth Model in the 1980's to represent the base of the asthenosphere. However, the presence and nature of this boundary have remained controversial, particularly in the oceanic regions. In contrast to many studies which suggest the 220-km discontinuity does not exist in the global oceans, SS precursors reflected from this interface are observed across the oceanic regions in this thesis. Furthermore, there is a positive correlation between the topography of the 220-km discontinuity and seafloor spreading rate. Specifically, the 220-km discontinuity is smoother beneath slower-spreading seafloors and much rougher beneath faster-spreading seafloors. In addition, the roughness increases faster at slowerspreading seafloors while much more gradual at faster-spreading seafloors. This indicates a close connection between seafloor spreading and mantle convections in the asthenosphere, and seafloors have permanent memories of their birth places. Different melting processes at slow and fast spreading centers produce seafloors with different physical and chemical properties, modulating convections in the asthenosphere and ultimately shaping the topography of the 220-km discontinuity. Reflections from the 220-km discontinuity are also observed across the global continental regions. In addition, the 220-km discontinuity beneath the continents is comparable to that under oceanic regions in terms of their average depth (∼ 250 km) and velocity contrast across the discontinuity (∼ 7%). In continental regions, there is a general connection between the 220-km depth structure and plate tectonics. For example, the boundary is shallower along the southern border of the Eurasian plate from the Mediterranean region to East Asia where mountain belts were formed as a result of collision between the Eurasian plate and the Nubian, Arabian and Indian plates. Depth perturbations of the 220-km discontinuity are also observed along the Pacific subduction zones including the Cascadia Subduction Zone, Peru-Chile Trench and Japan-Kuril Kamchatka Trench. In addition, depth anomalies are mapped in the interior of continents, for example, along the foothills of high topography in the interior of the Eurasian plate, which may be controlled by far-field convection associated with the convergent processes at the plate boundaries.