Probabilistic Seismic Tomography with Scattered Waves
Fellowship - CSC - China Scholarship Council
Earth materials are heterogeneous at all length scales ranging from different minerals in a rock sample to geological units in continents. The propagation of seismic wave in such heterogenous media is characterized by complex ray paths and waveforms that document the arrival seismic energy far later than the direct wave arrival. This is called the scattered seismic wavefield. Also the direct waves are influenced the heterogeneity that redistributes the seismic energy in time and space. The scattered seismic waves are a very useful tool for the investigation of material properties like intrinsic attenuation, degree of heterogeneity or even small temporal variations of seismic velocity. Conventional approaches developed for direct waves are not sufficient to deal with scattered waves and have to be extended to cope with the fact that the details of the small scale small scale velocity structure (heterogeneity) are intrinsically unknown. We develop a probabilistic approach to image the spatial distribution of statistical material properties and temporal changes thereof in heterogeneous materials. Our approach is based on Monte-Carlo type solutions of the radiative transfer equation in media with spatially varying properties of the heterogeneity. We implement a simulation routine for the spatio-temporal distribution of the specific seismic intensity to model waveform properties for propagation in heterogeneous media. These will be used to calculate sensitivity kernels for the different material properties and observables like seismogram envelope or phase shifts. In a final stage the sensitivity kernels allow for an inversion of observed waveforms characteristics for the spatial distribution of statistical material properties.
