Source processes of small earthquakes, M 1-5: Studies of the San Andreas Fault at Parkfield and Long Valley Caldera, California

An hypothesis that calls for fluid-control of faulting processes is rapidly gaining popularity as supporting evidence from diverse areas of study continues to mount. Fluids are thought to be important in fault creep and in seismic rupture initiation, propagation, and termination. Fluid pressure cycling likely controls the periodic recurrence of characteristic microearthquakes observed at Parkfield. This mechanism, along with seismic failure due to localized high stress concentrations near fault irregularities and the possible seismic failure of intact rock could lead to dilational, or tensional, micro-failures and their coalescence into macroscopic shear faults. These features are predicted by fracture mechanics and crack interaction and are observed in many field studies of faults and joints. This thesis is an investigation of the source processes of small earthquakes recorded on the borehole high resolution networks at Parkfield and at Long Valley, Caldera.; Characteristics of microearthquake occurrence along the San Andreas fault at Parkfield are reviewed for evidence that fault zone fluids play a critical role in slip dynamics there. Space-time development of small, transient earthquake sequences reveals diffusive-like outward spreading. The data are consistent with a model in which some earthquake clusters occur by the cyclic pressurization of fluid within localized patches of the fault zone. The method of empirical Green's function deconvolution is used to isolate the source time function of six earthquakes magnitude 1.1-5 at Parkfield. A suite of coseismic rupture radii and stress drops are calculated for events, showing model error for these parameters. For the two largest events, the moment rate functions at several sites are inverted to obtain the slip distribution and another estimate of the radius and stress drop. Results suggest, but are not definitive, that stress drop increases with moment up to 5{dollar}cdot{dollar}10{dollar}sp{lcub}23{rcub}{dollar} dyne-cm which, if real, can be interpreted as a difference in failure mechanism that takes place at about the magnitude 4.5 level. Lastly, we measure the ratio of radiated S-waves to radiated P-waves and find some evidence for tensional failure or mixed mode rupture in earthquakes at Long Valley Caldera.