Revealing the Fine Structures of the Lithosphere Asthenosphere Boundary

Earth's near surface layer is made of relatively strong materials and often referred to as tectonic plates (or the lithosphere). Below this layer is a softer layer called the asthenosphere. The transition from strong lithosphere to weak asthenosphere is caused by temperature. However, recent high-resolution seismological observations suggest that the transition from the lithosphere to the asthenosphere cannot be attributed to the temperature alone: the change in seismic wave velocity at the boundary is too sharp and too large to be attributed solely to the gradual increase in temperature. There have been hot debates on what causes the observed sharp transition between two layers. In this dissertation, I synthesize insights from seismological observation, in particular receiver functions, with experiments and theory of anelasticity caused by grain-boundary sliding, to test and assess candidate models for the oceanic lithosphere asthenosphere boundary (LAB) as well as the Mid-lithospheric discontinues (MLD) observed in stable continental regions. I conduct new statistical analysis of the results of mineral physics experiments, providing a description of the uncertainties in the parameters of an elastically accommodated grain boundary sliding model (EAGBS). I extend the EAGBS model originally proposed by Karato (2012) to describe and explain the seismological signatures at the oceanic LAB, showing that this hypothesis suitably explains both the oceanic LAB and MLD in the continents within the limits of uncertainties in both mineral physics studies and seismological models.;I then describe new supporting evidence for the specific predictions of the EAGBS model using novel seismological technique and data from stations in the oceanic regions. High-resolution receiver function (RF) stacking techniques can provide robust characterization of the age-dependence in the sharpness, depth, and anisotropic fabric within the normal oceanic LAB and underneath 'anomalous' Pacific Ocean Island. I show how frequency-dependent receiver functions can help discriminate between different velocity models -- partial melting and EAGBS.;The first test region is focused on normal oceanic setting, using seventeen ocean-bottom and two borehole seismometers located in the West-Philippine sea and the Northwestern Pacific Ocean. In the normal oceans, seismological analysis shows age-dependence in depth and sharpness. A synthesis using the EAGBS model suggests that the simple model of cooling of oceanic lithosphere sufficiently explains the seismological LAB without the need for widespread partial melting. In this model, the velocity drop at the LAB is caused by the onset of EAGBS. The onset of EAGBS is defined by the conditions at which the characteristic frequency of EAGBS agrees with the seismic frequency. In young oceanic upper mantle, this condition is met when temperature reaches ∼1300 K. In contrast, in an old oceanic upper mantle (>70 Ma), this conditions is met when the water content jumps at ∼70 km. Consequently, the LAB depth in the young oceanic regions is age-dependent and not sharp, while the LAB depth in the old oceanic regions is nearly independent of age and sharp. Larger compliance and higher velocities above this critical temperature is the stiffer lithosphere, while the lower compliance and lower velocities is within the weaker asthenosphere. These are predictions that bear out in our inferences from seismological data.;Seismological observations on oceanic islands are different. I observe multiple layering both in the crust and at the LAB, with a change in anisotropy across these layers. Within the crust, harmonic decomposition of receiver functions allows the inference on orientation of the symmetry axis. I show that there is a statistical correlation between the direction of this symmetry axis and plate drag. I hypothesize that the generation of anisotropic gradients are related to fractures in the crust, which precede melting and lead to the generation of the observed fabric. A contrast with normal oceans lends further evidence that this hypothesized origin for crustal anisotropy is a distinct feature of ocean-islands, separate from the formation of normal ocean crust/plates. Anisotropic structures are also observed at the LAB and may be due to changes in plate motion through time, hotspot activity or both.