An empirical analysis of peak horizontal acceleration for the Loma Prieta, California, earthquake of 18 October 1989

Abstract A total of 137 ground-level accelerograms recorded during the 18 October 1989 M S = 7.1 Loma Prieta, California, earthquake were used to study the dependence of peak horizontal acceleration on distance, azimuth, and site geology. According to this analysis: (1) peak accelerations recorded on alluvium at distances of 50 to 79 km from the inferred seismogenic rupture zone exhibited little or no attenuation with distance; (2) peak accelerations recorded on alluvium at seismogenic distances exceeding 50 km were significantly higher than those predicted from attenuation relationships available from the literature; (3) both the amplitude and rate of attenuation of peak accelerations recorded on alluvium exhibited a strong dependence on azimuth, with accelerations recorded at azimuths corresponding to San Francisco and Oakland having 68% higher amplitudes and a lower rate of attenuation than those recorded at more northerly azimuths; and (4) peak accelerations recorded at all azimuths exhibited a strong dependence on site geology, with Bay Mud sites having 88% higher accelerations, soft-rock sites having 15% lower accelerations, and hard-rock sites having 32 per cent lower accelerations than alluvial sites. The dependence of peak acceleration on azimuth is consistent with the geographic distribution of Modified Mercalli intensity. Although it is possible to explain both the low rate of attenuation and the constant level of acceleration at distances of 50 to 79 km by the arrival of critical reflections from the base of the crust (Moho), it is also possible that source directivity and radiation pattern, path-dependent attenuation, and systematic differences in local site conditions could have been partly responsible for these observations.