Earth Planets Space, Vol. 58 (No. 12), pp. 1593-1597, 2006LETTER
Kenichi Tsuda and Jamison Steidl
Department of Earth Science and Institute for Crustal Studies, University of California, Santa Barbara, USA
(Received November 29, 2005; Revised January 28, 2006; Accepted February 12, 2006; Online published February 2, 2007)
We estimate nonlinear site response by comparing site response estimates from the 16 August 2005 Mj=7.2 and 26 May 2003 Mj=7.0 Miyagi-Oki earthquakes with site response estimates from aftershocks of the 2003 event. Site response is solved by a spectral inversion technique to separate source, path, and site components. The constraint motion in the inversion is a regional attenuation model derived from fitting the spectra of data recorded at borehole KiK-net stations in the region and a theoretical source spectrum for each event determined using the same borehole stations. Site response is calculated at the surface of the KiK-net and K-NET stations. In general, the average aftershock site response is larger than for the two mainshocks, especially at a higher frequency. When comparing site response with input ground motion level, the predominant frequency and the site response values tend to decrease as the level of input ground motion increases.
Key words: Nonlinear site response, ground motion level, PGA, 2003 and 2005 Miyagi-Oki earthquakes.