Using stress shadows to invert for changes in local stress field

When a large earthquake occurs, stresses in the crust are redistributed creating regions that experience an increase in stress while others experience a stress decrease which are called stress shadows. In many studies, these stress shadows are said to contain less seismic activity than the average background rate, and so correlations are made between lack of seismicity or a decrease in seismicity rate and the stress shadow locations and magnitudes (the amount of decrease of stress). In this study the opposite procedure is applied: We use seismicity rate changes to determine information about the stress changes due to a large magnitude earthquake, as well as its effect on the stress field itself. We use the Pattern Informatics method to examine the changes in seismicity rate, as it is an objective measure of the rate changes with respect to the regional background rate. The results from this analysis are then used to invert for, with a genetic algorithm, parameters that define the stress field such as the principal stress orientations, the coefficient of friction, and the calculation depth. The modelled stress data is calculated using Coulomb stress change theory and the Coulomb 3 program, and it is trying to produce the same size and location of stress shadows as seen in the seismicity rate change data. Four different Californian earthquakes were chosen in order to determine their effect on the local stress field: (1) 1987 Superstition Hills (2) 1989 Loma Prieta (3) 1992 Landers and (4) 1994 Northridge. In order to find out the effect that each of the parameters have on the modelled results, we performed a Monte Carlo simulation to find the errors associated with each, and a sensitivity analysis to determine the magnitude of change that each one produces. We hope with this new information of the changes incurred due to a large magnitude earthquake occurrence, that modelling of earthquakes can be advanced, and our understanding of their mechanics enhanced.