Surface waves in the western Taiwan coastal plain from an aftershock of the 1999 Chi-Chi, Taiwan, earthquake

Significant surface waves were recorded in the western coastal plain (WCP) of Taiwan during the 1999 Chi-Chi, Taiwan, earthquake and its series of aftershocks. We study in detail the surface waves produced by one aftershock (20 September 1999, 18hr 03m 41.16sec, M 6.2) in this paper. We take the Chelungpu-Chukou fault to be the eastern edge of the WCP because it marks a distinct lateral contrast in seismic wave velocities in the upper few kilometers of the surface. For many records from stations within the WCP, body waves and surface waves separate well in both the time domain and the period domain. Long-period (e.g., >2 sec) ground motions in the plain are dominated by surface waves. Significant prograde Rayleigh wave particle motions were observed in the WCP. The observed peak ground velocities are about 3-5 times larger than standard predictions in the central and western part of the plain. Observed response spectra at 3 sec, 4 sec, and 5 sec at the center of the plain can be 15 times larger than standard predictions and 10 times larger than the predictions of Joyner (2000) based on surface wave data from the Los Angeles basin. The strong surface waves were probably generated at the boundary of the WCP and then propagated toward the west, largely along radial directions relative to the epicenter. The geometry of the boundary may have had a slight effect on propagation directions of surface waves. Group velocities of fundamental mode Rayleigh and Love waves are estimated using the multiple filter analysis (MFA) technique and are refined with phase matched filtering (PMF). Group velocities of fundamental mode surface waves range from about 0.7 km/sec to 1.5 km/sec for the phases at periods from 3 sec to 10 sec. One important observation from this study is that the strongest surface waves were recorded in the center of the plain. The specific location of the strongest motions depends largely on the period of surface waves rather than on specific site conditions or plain structures. Accordingly, we conjecture that surface waves could be generated in a wide area close to boundaries of low-velocity sedimentary wave guides. In the case studied in this article the area can be as wide as 30 km (from the Chelungpu fault to the center of the plain). Surface waves converted by P and S waves at different locations would overlap each other and add constructively along their propagation paths. As a result, the surface waves would get stronger and stronger. Beyond a certain distance to the boundary, no more surface waves would be generated. Consequently, no more local surface waves would be superimposed into the invasive surface waves, and the surface waves would tend to decay in amplitude with distance.