Gross, S.J., 1993, Models of stress changes induced by earthquakes and their relationships to changes in the spatial distribution of seismicity: University of Colorado, Boulder, Ph.D. dissertation, 171 p.
The earthquake generation process is poorly understood, because it involves the interaction of unknown material properties with physical processes which are just beginning to be understood. It is impossible to directly measure frictional properties of active faults, especially in subduction zones. Observations of earthquake occurrence do provide indirect information about the earthquake generation process, because earthquakes change the stresses acting upon neighboring faults and consequently influence future events. In this work, the stresses caused by $52\ m\sb{b}\approx 4.7$ 'target' earthquakes are modeled and systematically compared to changes in the spatial distribution of mostly smaller surrounding seismicity recorded by the Central Aleutian Seismic Network. The models have been used to evaluate possible fault failure models and background stress states appropriate for a subduction zone. Statistical measures developed to assess the significance of the results compare changes in spatial distribution of seismicity occurring at the times of target earthquakes to changes occurring at other times in the same catalog. The statistics break the surrounding seismicity down into eight distance ranges. The best-fitting model has statistics from four distance ranges which each have less than one chance in 20 of arising randomly, with the most significant distance ranges being those with the largest average numbers of surrounding earthquakes. The most successful model has background stresses caused by bending of the subducting plate and weak faults in all orientations. This stress state suggests the main thrust zone between the subducting and overriding plates is also weak, because otherwise forces from the collision of the two plates would dominate the stress field. The movement of surrounding seismicity from areas of stress decrease to areas of stress increase occurs steadily over a few years at the time of the target earthquake, and decays afterwards. There also is a precursory change which may be evidence of source preparation processes. This new approach to the interpretation of seismicity data provides insights into the earthquake generation process and allows our relatively mature understanding of seismic source modeling to be extended into less understood subjects of earthquake nucleation, medium properties and fault friction.
Theses and Dissertations
