On the use of physics-based earthquake simulations to further understand the behavior of ground motions and constrain ground motion prediction equations
Abstract/Contents
- Abstract
- Understanding the processes that control the behavior of earthquake ground motions, typically categorized into source, path (i.e. wave propagation), and site effects, is a critical element of seismic hazard assessment. Despite tremendous progress in knowledge of these processes over the last 100 years, the complex nature of earthquakes presents the seismological and engineering communities with uncertainty that hampers our ability to predict, and therefore prepare, for the shaking that human infrastructure faces in the areas surrounding potential earthquake locations. Often described using ground motion prediction equations (GMPEs), which are models that attempt to predict ground motion intensity measures as functions of parameters including, but not limited to source-to-site distance, earthquake magnitude, and style of faulting, ground motions exhibit a high degree of variability that translates directly into uncertainty in seismic hazard assessments. Therefore, it is of great importance to human safety and the minimization of economic losses due to earthquakes that GMPEs are as accurate as possible. The development of GMPEs typically relies on fitting catalogs of observed ground motions, which is limited in many areas by factors such as earthquake frequency (especially large magnitude events), lack of recording instruments near earthquake sources (further complicated because the precise timing and locations of earthquakes is currently unpredictable), and the short history of observation systems, even in areas such as Japan and California, USA. Additionally, wastewater injection related to oil and gas extraction operations has given rise to seismicity in areas such as the United States (e.g. Oklahoma, Kansas, Arkansas, Texas), Canada, and the Netherlands, that have not been historically considered areas of high seismic concern. In areas experiencing such ''induced'' seismicity, ground motions are especially poorly understood due to the convolution of the aforementioned factors and the neoteric nature of the seismic hazard. In this thesis, I aim to aid in the ongoing effort of understanding the behavior of earthquake ground motions and constraining GMPEs by using earthquake simulations and resulting synthetic ground motion data. I perform simulations in 2D and 3D to address questions about general controls on ground motions, such as material structure heterogeneity and fault geometry, and how ground motion simulations can be structured, validated, and integrated with recorded data such that synthetic data can be employed to improve existing GMPEs and aid in the development of GMPEs for targeted areas. In the first part of the thesis, I use 2D simulations to explore the effect of material property heterogeneity and nonplanar fault geometries affect rupture dynamics and ground motions. I find that material heterogeneity plays a less significant role in rupture dynamics and low-frequency ground motion intensity measures such as peak ground velocity than does the geometry of faults, but can have a substantial impact on high-frequency ground accelerations. I also derive a metric using the parameters of the von Karman distribution, which is a commonly employed model for generation of random perturbations in material properties, to describe the behavior of ground motions in various manifestations of heterogeneous media as a result of wave scattering. The rest of the thesis utilizes 3D simulations and outlines a procedure to generate synthetic ground motion data that respects the characteristics of recorded ground motion data (albeit a limited number of near-source recordings) in the Oklahoma/Kansas target region, an area experiencing recent and dramatic increases in earthquakes due to wastewater injection. Using both point moment tensor sources to simulate small events and finite-area, dynamic rupture models for moderate and large earthquakes, I perform simulations of earthquakes that includes information on the properties of the local material structure and earthquakes that have been recorded in the area with the goal of producing realistic, but synthetic ground motion data tailored specifically to the target region. Once the synthetic ground motion data is determined to be realistic, I combine it with the recorded data from the target region to produce a GMPE for several ground motion intensity measures for earthquakes Mw 3.0 - 5.8 for the Oklahoma/Kansas region. The development of a GMPE for a such a small target region is rare considering that current GMPEs typically attempt to predict ground motions for areas such as the entire Central and Eastern US, but will lead to a decrease in uncertainty in seismic hazard assessment as more regions are targeted for localized GMPE development. Most importantly, this procedure can be ported to other target areas and is not specific to Oklahoma/Kansas.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2017 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Bydlon, Samuel Adam |
|---|---|
| Associated with | Stanford University, Department of Geophysics. |
| Primary advisor | Dunham, Eric |
| Thesis advisor | Dunham, Eric |
| Thesis advisor | Beroza, Gregory C. (Gregory Christian) |
| Thesis advisor | Ellsworth, William L |
| Advisor | Beroza, Gregory C. (Gregory Christian) |
| Advisor | Ellsworth, William L |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Samuel Adam Bydlon. |
|---|---|
| Note | Submitted to the Department of Geophysics. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2017. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2017 by Samuel Adam Bydlon
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...