Electronic Theses and Dissertations

Kiran Pandey

Date

2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Earth Sciences

Committee Chair

Thomas Goebel

Committee Member

Chris Cramer

Committee Member

Christodoulos Kyriakopoulos

Committee Member

Eunseo Choi

Committee Member

Randel Cox

Abstract

The ability to forecast future earthquakes before they occur to minimize seismic hazards is known as earthquake prediction or forecasting. The main aim of earthquake prediction is to find the expected time of occurrence, the location of the expected rupture, and how big the earthquake will be (i.e., magnitude). Earthquake prediction and the study of the indications prior to major earthquakes (precursory signals) have been key objectives since the beginning of earthquake research. In addition, a better understanding of earthquake nucleation and how faults progress toward failure is very important for seismic hazard assessments. The occurrence of earthquakes is known to be controlled by the interaction between the stress on the fault, fault zone properties, and strain localization. Direct observation of these factors in nature is challenging. To overcome these challenges, laboratory experiments provide an opportunity to study fault weakening, microseismicity distribution, strain localization, the role of stress change, and pore pressure by controlling one or more boundary conditions at a time. Stick slip phenomena are often noticed in laboratory experiments and are considered an important mechanism for shallow earthquakes along preexisting faults. In this dissertation, we study and explore seismic velocity changes and repeating acoustic emissions as precursory signals during fault damage accumulation and strain localization before failure. First, we study the changes in seismic velocity for direct P and coda waves. Coda waves are thought to sample scatterers (e.g., microcracks) in the medium. To infer coda wave velocity changes, we used the coda wave interferometry (CWI) method. We investigate velocity changes in intact and saw-cut samples, with different roughness in Westerly granite subjected to controlled loading conditions. The results demonstrate that the bulk material properties on saw-cut samples have a stronger influence on velocity change than surface roughness. Furthermore, direct wave velocity shows strong anisotropy with increasing differential stress, suggesting preferred microcrack alignment within the deviatoric stress field. Opposite to that, the coda wave shows mostly isotropic behavior, possibly averaging over multiple propagation paths thereby reducing the directional variations. The result of this study highlights the importance of viiseismic velocity, especially coda wave velocity, in monitoring fault damage evolution. Second, we analyze the role of fault surface heterogeneity on strain localization and the occurrence of repeating acoustic emissions in laboratory experiments. We conducted experiments on three types of fault surface: 1) rough fractures, 2) rough saw-cuts, and 3) polished saw-cuts. A time domain cross-correlation method with a cross-correlation coefficient greater than 0.95 was used to identify repeating acoustic emission event families. Our finding suggests that repeaters are observed on and around high-density acoustic emission patches. Furthermore, the frequency and proportion of repeaters relative to total event numbers increase with surface smoothness and pore pressure. The number of repeaters increases as faults approach failure at differential stresses close to fault strength. Finally, the findings on seismic velocity changes and repeater analysis from this dissertation show signals associated with preparatory processes and damage accumulation, which have contributed to our understanding of fault mechanics and the earthquake nucleation process.

Comments

Data is provided by the student.

Library Comment

PDF

Notes

Open access.

Download

Share

COinS