Relocation of Earthquakes that Occurred Beneath Parkfield Region of California using VELHYPO

Abstract

When only direct-waves are inadequate, different combinations of both direct- and refracted-waves are used to determine hypocentral parameters of local earthquakes. Compared with only direct-waves, the addition of refracted-waves tends to increase errors of hypocentral parameters. This is true especially when the velocity structure of the region under investigation is not known. We use an accurate and fast algorithm to determine the hypocentral parameters based on an optimum 1-D velocity model yielding a minimum misfit error in media with multiple refracted-waves as first arrivals. Using the algorithm, we relocated earthquakes that occurred beneath the Parkfield region of California during the period between January 2000 and December 2000. The shallow part of the velocity model used in the study area has high velocity contrast among layers. Therefore, refracted-waves are recorded as first-arrival phases at most stations around the region for events of shallow focal depths (less than 3km). The study area covers the transition between the creeping segment of San Andreas Fault (SAF) to the northwest and the locked segment to the southeast. Accuracy test for the algorithm was conducted using some reference models. We compare the inverted hypocentral parameters with those determined by previous studies using HYPOINVERSE and the HypoDD. The epicentral distributions of the three methods show a close relationship with the regional fault distribution. However, compared to HypoDD, the epicentral distribution of HYPOINVERSE shifted slightly along the fault lines while the epicentral distribution of VELHYPO moved further in the same direction. The result of VELHYPO suggests that the dip of San Andreas Fault is 6-10 o SW while the results of HPOINVERSE and HypoDD suggest that the dip of San Andreas Fault is nearly vertical and/or 3-7 o NE. This study supports the result of previous studies that San Andreas Fault is non-