![Thumbnail](/bitstream/handle/20.500.11850/390896/Thesis_2019_Sager.pdf.jpg?sequence=7&isAllowed=y)
Open access
Author
Date
2019Type
- Doctoral Thesis
ETH Bibliography
yes
Altmetrics
Abstract
This thesis deals with the development and application of a novel method for seismic interferometry, referred to as full waveform ambient noise inversion. It incorporates the heterogeneous nature of ambient noise sources in space and frequency, and models the full wave propagation physics. By dropping the principle of Green's function retrieval, the basis of traditional ambient noise tomography, we establish correlation functions as self-consistent observables. A more accurate description of the physics of the problem allows us to extract additional information, to advance our understanding of correlation wavefields and to quantify inherent source-structure trade-offs.
We generalize the theoretical formulation, thus facilitating the application of adjoint techniques, and investigate the prerequisites for its application to real data in a synthetic \mbox{2-D} study. We then follow with a framework for modeling correlations in \mbox{3-D} attenuating media with arbitrary noise source distributions and for the computation of sensitivity kernels for changes in source and structure. In a pilot use case, we look into a commonly encountered difference between the observation of body waves and surface waves, and reveal fundamentally different conditions for their reconstruction. For a final synthesis of all results and a proof of concept we invert a global dataset focusing on the long-period ambient noise field, called the Earth's Hum, for the distribution of ambient noise sources, and for Earth structure. We confirm existing models for the source power-spectral density and recover tomographic features well known from global Earth models. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000390896Publication status
publishedExternal links
Search print copy at ETH Library
Publisher
ETH ZurichOrganisational unit
03971 - Fichtner, Andreas / Fichtner, Andreas
Funding
149143 - Full waveform ambient seismic noise inversion (SNF)
More
Show all metadata
ETH Bibliography
yes
Altmetrics