Source: Dissertation Abstracts International, Volume: 79-05(E), Section: B.

Adviser: Jeffrey Park.

In this work, Earth's 3D lateral heterogeneities are investigated using normal modes observations. The topics that are covered include observations of quasi-Love waves and their implications for the deformation pattern in Tibet, and a new singlet stacking method and its applications in mode coupling observations and mode attenuation measurements.

In the first part, we will use quasi-Love wave observations to infer mantle anisotropy patterns around Tibet. Based on the quasi-Love wave observations that we identified using stations in Tibet and earthquakes around the globe, we are able to pinpoint the locations of the scatterers that are responsible for the observed quasi-Love waves, mainly through the arrival time of the quasi-Love waves. The frequency content of the quasi-Love waves. which centers around 10 mHz, indicates that the depth of anisotropy should be located between 100km to 200km. By comparing our results with crustal deformation on Earth's surface, we find that the geographical distribution of these scatterers matches with crustal deformation boundaries quite well, particularly in SE Tibet and the Sayan Mountain area. The good correspondence between crustal and mantle deformation at these locations essentially establishes a vertically coherent boundary condition for the continent collision in Central Asia. Such a vertically coherent boundary condition has important implications on the deformation style in Tibet, suggesting that crust and mantle deformation in Tibet should overall be coupled, even with the presence of a weak lower crust.

In the second part, we will discuss a novel epicentral singlet stacking method for Earth's normal modes and its applications in singlet coupling observations and multiplet attenuation measurements. First a benchmark for the epicentral singlet stacking method is performed using synthetic data. The results of the benchmark suggest that epicentral singlet stacking should lead to concentrated energy in the ∣m∣ ≤ 2 singlets and that the method is robust in terms of epicenter uncertainties. As we apply the method on real data, stacking results show significant energy leakage to the ∣ m∣ > 2 singlets. Further, we also observe that singlet attenuation rates can vary significantly even within the same multiplet. All these observations are strong evidences of mode coupling for Earth's free oscillations, which are hard to obtain otherwise, particularly on the singlet level. In quantifying the energy leakage to the ∣m∣ > 2 singlets, we find that the leakage should be mostly due to mode coupling. and that higher-order source mechanisms should only play a minor role, if at all.

The epicentral singlet stacking method is further applied to the measurements of multiplet attenuation rates. The use of epicentral singlet stacking method helps minimize the impact from background noise and singlet interference biases, which are likely present in most of the past measurements. To measure the attenuation rates of the multiplets, we first measure the attenuation rates of the singlets using the time-lapse method, and then aggregate singlets only with high S/N ratios. This way we are able to get accurate and reliable estimates of multiplet attenuation rates for modes 0S 21 through 0S30. While our measurements are systematically higher than the QM1 model, which is a normal-mode based 1D attenuation model, they are however in good agreement with the surface-wave based attenuation model. QL6. We believe our results are more accurate than past modal attenuation measurements, due to significantly reduced biases from background noise and singlet interference. In particular, we find that past normal-mode based multiplet attenuation measurements fail to account for the singlet interference effect due to varying singlet attenuation rates, which can easily explain the longstanding 10% to 15% discrepancy between normal-mode based and surface-wave based attenuation measurements.