Extracting P-primary transmission and reflection impulse responses from teleseismic data

An approach is presented in this dissertation to extract P-primary transmission and reflection responses from teleseismic data. Three component teleseismic data are recorded under the transmission geometry. Instead of using the forward scattered P-to-S converted waves (P transmission conversions) as most existing methods do, I am aiming at directly using the teleseismic P component data. To recover receiver-side lithospheric/upper mantle structures from teleseismic P wave data, it is necessary to extract accurate P-impulse response estimates. The brand new concepts developed by this research are to construct the reflection response from the teleseismic data and remove the free surface effect for teleseismic transmission response and the free surface multiple for teleseismic reflection response. This new theory is based on the inverse scattering series for transmission response to remove the free surface effect.; The approach can be summarized as follows. Rather than attempting to estimate the complicated source wavelet itself, I use the Multitaper method to estimate the power spectrum of the source wavelet. Then, a spectrum deconvolution method is introduced to estimate the power spectrum of the teleseismic transmission response. Two types of P-impulse response estimates can be constructed from their common power spectrum: one is a minimum phased P-transmission response constructed by the Kolmogoroff method; the other is the P-reflection response constructed by one-way wavefield correlation-type reciprocity. I remove the free surface multiple by using inverse scattering series for both the P-transmission and reflection responses. The final outputs are P-primary transmission and reflection impulse response estimates. Synthetic tests show the stability and effectiveness of this approach and it has great potential to provide high resolution imaging results with less assumption of the subsurface. Lithospheric structures are successfully recovered by this approach from a real data set recorded by the RISTRA seismic network.; In addition, two separate studies are presented in this dissertation: one is the inversion of the earth properties from elastic forward scattered teleseismic waves by using the inverse scattering series; the other is a general mapping tool for computational seismology as a three dimensional geographical curvilinear grid library.