Valley Earthquake: More Evidence for an Asperity

The well-recorded strong ground motion data for the 23:19 aftershock of the 15 October 1979 Imperial Valley earthquake provide a good opportunity to study the high-frequency source characteristics and the path effects at near-source distances. The best-firing point source model has a strike-slip mechanism, N40°W, which is nearly identical to the main event. The estimated stress drop is extremely high, roughly 500 bars, with a triangular time history (0.1, 0.1 sec) but with a moment of 1.0 x 1024 dyne-cm. A double-source model found by inversion fits the highfrequency data better but indicates complex faulting: the first source (with strike = N319°E, dip = 42°NE, and slip angle = 165 °) has a moment of 0.7 x 1024 dynecm, the second source (with stdke = N324°E, dip = 82°SW, and slip angle = 181 °) lies about 0.5 km to the north and has a seismic moment twice that of the first source. Source dimensions appear very small for this amount of energy release. Many of the anomalous behaviors observed at certain stations for the main event are, also, present in the aftershock data. These features are examined in terms of path effects. INTRODUCTION In recent years, significant progress has been made in understanding long-period strong ground motions (displacements) produced by some of the larger California earthquakes~ Assuming a distributed shear dislocation along planar surfaces embedded in a layered velocity structure provides an adequate set of parameters for modeling observations in a deterministic sense. Unfortunately, such models appear inadequate to explain the higher frequency signals (velocities or accelerations) which apparently are not consistent with the radiation patterns predicted from the longer period studie s . What causes this breakdow n of the deterministic approach at high frequencies is not clear. Crustal scattering (inappropriate Green's function), microsource complexity (irregular dislocation elements), and perhaps macrosource complexity (interference of waves produced by various portions of the fault) are possible reasons. The latter cause of complexity can be largely eliminated by selecting a small event where the dislocation is confined to a localized region. We propose to study such an event and to address the first two possibilities, namely crustal scattering and irregular source elements. The event modeled in this study is an aftershock of the 15 October 1979 Imperial Valley earthquake, that was very well recorded by the Imperial Valley strong-motion array at all azimuths and at distances ranging from 8 to 26 km. Furthermore, we have the rare opportunity to compare P waveforms with S waveforms which allows the estimation of different apparent attenuation coefficients within the sedimentary basin. THE 23:19 AFTERSHOCK The 15 October 23:19 aftershock ( M L = 5.0) occurred about 2~ rain after the main shock, and unfortunately most of the far-field data are overwhelmed by the coda of * Present address: Schlumberger-Doll Research, Ridgefield, Connecticut 06877. 689 690 HSUI-LIN LIU AND DONALD V. HELMBERGER the main event. The only available information for this aftershock is from the strong-ground motion recordings. Using these data as well as some low-gain array data, the epicenter was located at 32°46,00'N and 115°26.48'W and at a depth of 9.5 km (Brady and Porcella, 1983; Carl Johnson, personal communication, USGS), which is on the Imperial fault near Highway 8 and close to the zone of high energy release for the main shock (Hartzell and Helmberger, 1982). AR8 ~ 7 44 ,,~.~sw ,,~.~o'w A R ~ ~ _ . . . . . . . ~ t / ~ . , . . ~3.oo,