Congratulations to Lorena Medina Luna who successfully defended her dissertation on January 5, 2015.
Advisor: Eric Hetland
Understanding the stresses that lead to earthquakes is critically important for understanding seismic hazard. Geodetic measurements record surface deformation related to strain accumulation on faults prior to large earthquakes, as well as deformation during and following those earthquakes. However, earthquakes are the response to stress accumulation on faults, and direct measurements of stress are difficult to obtain. Focal mechanisms of moderate to large earthquakes are routinely used to infer crustal stresses, using least square techniques.
In this dissertation, I develop, test, and apply a Bayesian Monte Carlo estimation of crustal stress from earthquake data. I consider both coseismic slip models, which had not been done prior to the work presented here, and focal mechanisms. Coseismic slip models of the 2008 Wenchuan, China, earthquake, and of the 1999 İzmit and Düzce, Turkey, earthquakes are used to estimate the stresses that led to those earthquakes. Additionally, focal mechanisms of Wenchuan earthquake aftershocks and earthquakes in the Sea of Marmara, adjacent to the İzmit earthquake, are used to infer crustal stress. I find that the Wenchuan earthquake was caused by a homogenous stress, and that heterogeneous stresses previously appealed to in order to describe the aftershock focal mechanisms may simply reflect the ambiguity in the interpretation of stress from focal mechanisms. Coseismic slip models from the İzmit and Düzce earthquakes are also consistent with a homogeneous stress along all of the fault segments that slipped in those earthquakes, particularly if the coefficient of fault friction is about 0.2 or less. In the Sea of Marmara, inferred stresses from focal mechanisms indicate variability in stress from the eastern to the western segments of the Main Marmara fault, with a potential stress rotation along the western segment between about 1999 and 2003.