Seismic Guided Waves Trapped In the Fault Zone of the Landers, California, Earthquake of 1992 (original) (raw)

1994, Journal of Geophysical …

A mobile seismic array of seven stations was deployed at 11 sites along the fault trace of the M7.4 Landers earthquake of June 28, 1992, with a maximum offset of 1 km from the trace. We found a distinct wave train with a relatively long period following the S waves that shows up only when both the stations and the events are close to the fault trace. This wave train is interpreted as a seismic guided wave trapped in a low-velocity fault zone. To study the distribution of amplitude of the guided waves with distance from the fault trace and also their attenuation with travel distance along the fault zone, we eliminated source and recording site effects by the coda normalization method. The normalized amplitudes of guided waves show a spectral peak at 3-4 Hz, which decays sharply with distance from the fault trace. Spectral amplitudes at high frequencies (8-15 Hz) show an opposite trend, increasing with distance from the fault trace. The normalized amplitudes of guided waves at 3-4 Hz also show a systematic decrease with hypocentral distance along the fault zone, from which we infer an apparent Q of 50. In order to confirm the existence of the guided waves, a dense array of 31 stations was deployed at one of the 11 sites. The resultant records revealed unequivocal evidence for the existence of guided waves associated with the fault zone. By modeling the waveforms as S waves trapped in a low-velocity waveguide sandwiched between two homogeneous half-spaces with velocity V s = 3.0 km/s, we infer a waveguide width of about 180 m, a shear velocity of 2.0-2.2 km/s, and a Q of •50. Hypocenters of aftershocks with clear guided waves show a systematic distribution both laterally and with depth delineating the extent of the low-velocity fault zone in three dimensions. We find that the zone extends to a depth of at least 10 km. This zone apparently continues to the south across the Pinto Mountain fault because guided waves are observed at stations north of the Pinto Mountain fault for earthquakes with epicenters south of it. On the other hand, the zone appears to be discontinuous at the fault bend located about 20 km north of the mainshock epicenter; guided waves were observed for stations and epicenters which are located on the same sides of the fault bend but not for those on the opposite sides. the seismic source was within the fault zone. The characteristic waveform was attributed to a Love wave type mode trapped in a low-velocity, low-Q zone. Similar trapped modes were also identified in some of the borehole seismograms obtained at the San Andreas fault near Parkfield, California [Li et al., 1990]. We believe that the study of fault zone trapped modes is important for several reasons. First, since the characteristics Copyright 1994 by the American Geophysical Union. Paper number 94JB00464. 0148-0227/94/94 J B-00464505.00 frequency of about 10 Hz in the acceleration spectrum in 11,705 11,706 LI ET AL.: FAULT ZONE TRAPPED WAVES, LANDERS EARTHQUAKE accordance with the general appearance of observed strong motion records. Ida [1973] interpreted the cutoff frequency as the source effect and estimated the critical slip in his slip-weakening model to be of the order of 10 cm. Papageorgiou and Aki [1983a, b] attributed the cutoff frequency to the size of cohesive zone (break down zone) acting as a spatial smoothing operator on fault slip. On the other hand, Hanks [1982] attributed fmax primarily to the local recording site effect, while Anderson and Hough [1984] attributed it to a strong effect of attenuation primarily at shallow depths below the recording station on high-frequency spectra. They attribute fmax to the site effect because they found that fmax depends on the site geology: lower for softer sediments and higher for harder rocks. Suet al. [1992], on the other hand, found that the site amplification factor increases monotonically with the decreasing geologic age of site up to 12 Hz, using the data from 132 stations of the U.S. Geological Survey (USGS) seismic network in central California. This means that the site-controlled fmax effect does not, on the average, apply to the frequency range up to 12 Hz. Recently, Kinoshita [1992] demonstrated that the source