Seismic displacements near a fault

Abstract

Ground motion at only 80 meters from the San Andreas fault was recorded during the Parkfield earthquake of June 28, 1966, by the accelerographs of the U.S. Coast and Geodetic Survey. Seismic displacements obtained from the accelerograms are compared with synthetic seismograms calculated for a moving dislocation model. An excellent agreement was obtained between the observed and theoretical seismograms. The dislocation is the only important source parameter that controls the seismic motion at such short distances. Other parameters such as the fault length and depth show negligible effect of the motion. The rupture velocity is a significant but not dominant factor. From the comparison of observed and theoretical seismograms, it was concluded that the fault dislocation was about 60 cm. This value is an order of magnitude greater than the fault offset observed at the surface. This discrepancy was attributed to the decoupling of a thin surface layer (probably less than 100 meters from the surface). The fault depth was estimated by combining the estimated dislocation with the seismic moment obtained from long-period surface waves. The fault depth within the basement rock was found to be about 3 km, which is significantly smaller than the depth (15 km) of the aftershock zone. This discrepancy was attributed to the effect of increasing friction with depth which momentarily prevented downward extension of rupture during the main shock.