Near-fault strong motion complexity of the 2000 Tottori earthquake (Japan) from a broadband source asperity model

Abstract

The October 6/2000 Tottori earthquake that occurred in central Japan was an intermediate size strike-slip event that produced a very large number of near field strong motion recordings. The large amount of recorded data provides a unique opportunity for investigating a source asperity model of the Tottori earthquake that, combined with a hybrid strong motion simulation technique, is able to reproduce the observed broadband frequency near-fault ground motion. We investigated the optimum source asperity parameters of the Tottori earthquake, by applying a Genetic Algorithm (GA) inversion scheme to optimise the fitting between simulated and observed response spectra and Peak Ground Acceleration (PGA) values. We constrained the initial model of our inversion by using the heterogeneous slip distribution obtained from a kinematic inversion of the source of previous studies. We used all the observed near-fault ground motions (−100 m) from the borehole strong motion network of Japan (KiK-Net), which are little affected by surficial geology (site effects). The calculation of broadband frequency strong ground motion (0.1–10 Hz) is achieved by applying a hybrid technique that combines a deterministic simulation of the wave propagation for the low frequencies and a semi-stochastic modelling approach for the high frequencies. For the simulation of the high frequencies, we introduce a frequency-dependent radiation pattern model that efficiently removes the dependence of the pattern coefficient on the azimuth and take-off angle as the frequency increases. The good agreement between the observed and simulated broadband ground motions shows that our inversion procedure is successful in estimating the optimum asperity parameters of the Tottori earthquake and provides a good test for the strong ground motion simulation technique. The ratio of background stress drop to average asperity stress drop from our inversion is nearly 50%, in agreement with the theoretical asperity model of Das and Kostrov [Das, S., Kostrov, B.V., 1986. Fracture of a single asperity on a finite fault: a model for weak earthquakes? Earthquake Source Mechanics, AGU, pp. 91–96.], and an empirical ratio of asperities to rupture area [Seismol. Res. Lett. 70 (1999) 59–80.]. The simulated radiation pattern is very complex for epicentral distances within half the fault length, but it approaches the radiation of a double-couple point source for larger distances. The rupture velocity and rise time have a significant influence on the Peak Ground Velocity (PGV) distribution around the fault. An increase in rupture velocity produces a similar effect on the ground motion as a reduction in rise time.