The 4 September 1981 Santa Barbara Island, California, earthquake: Interpretation of strong motion data

Abstract

Abstract Seven timed strong motion accelerograms were recorded between the distances of 45 and 80 km from the Santa Barbara Island earthquake. From the accelerograms, the average estimate for ML is 6.0, compared to an average of ML = 5.5 from more distant recordings. When ML is estimated from accelerograph data using the distance correction to the ML scale proposed by Luco (1982), an average of 5.3 is obtained. This suggests that the correction slightly overcompensates in this magnitude and distance range, in which Luco lacked data, at least for this earthquake. S-wave spectra, from three accelerograms which recorded the complete S wave, imply a seismic moment of 1.5 × 1024 dyne-cm and, by the method of Brune (1970), a source radius of 2 km and a stress drop of 90 bars. If the displacement spectrum is assumed to be approximated by two asymptotes, f0 at frequencies less than the corner frequency and f−2 at frequencies greater than the corner frequency, the level of the f−2 asymptote can be constrained by rms acceleration (Hanks, 1979). This leads to a smaller estimated source radius and a higher stress drop. In the process of matching these spectra, an average value of Q can be inferred to be consistent with the spectra at high frequencies. This value is between 300 and 500 for the assumed spectral shapes. One accelerograph, located on a ridge, recorded an anomalously high amplitude ringing at 2 Hz on the transverse component only. Subsurface geological structure is the most likely cause of the amplification.