A gross picture of the site amplifications throughout Japan based on the generalized spectral inversion technique

Abstract:

We have been applying the generalized spectral inversion, originally proposed by Andrews (1986), to the strong motion data in Japan (Nakano et al., BSSA, 2015) in order to separate source, path, and site terms in the frequency range from 0.12 Hz to 15 Hz. This study focuses on the spatial distribution of the horizontal site amplification factors (HSAFs) of the S-wave part (5 to 15 s from the onset of S-wave). Although it is well known that the individual shape of HSAF at a site strongly fluctuates from 1 to 100, it is also well known that the HSAF in a low frequency range at a site inside a sedimentary basin is similar to those of the nearby sites inside the same basin. So we are using various kinds of site proxies such as Vs30 or Z1.0 to represent HSAFs empirically, however, we should see their spatial distribution before delineating the phenomenological correlations for these proxies. Thus, we plot a gross picture of the separated HSAF throughout Japan using GMT (Wessel and Smith, 1998) for 2600 sites of K-NET and KiK-net by NIED, JMA strong motion network, and CEORKA network in Kansai. The figure attached here shows the HSAF amplitude distribution at 0.5 Hz from the seismological bedrock with the S-wave velocity of 3.45 km/s. This figure clearly shows that large sedimentary basins such as the Kanto, Osaka, Nobi, and Kumamoto basins create large amplification inside. In general, the amplification in the east of the Itoigawa-Shizuoka Tectonic line, namely in Kanto, Tohoku, and Hokkaido regions, tend to be higher than those in the west. This should be interpreted in the consequence of the tectonic structural difference, rather than the local site amplification phenomena. This picture also provides other implications for previous studies. Wang & Zhao (PEPI 2005) presented a 0-10 km S-wave velocity anomaly distribution map based on the body-wave tomography (up to 6%), whose minus anomaly is very similar to our low-frequency HSAF amplitude. This velocity anomaly may come from the shallow sedimentary basin, not from the whole crustal structure as an average value down to 10 km. For example, if we assume an average S-wave velocity of 3.2 km/s for top 10 km, the vertical propagation time is 3.125 s, then the 6% velocity reduction yields the delay of 0.2 s, which can be explained by replacing only top 300 m by the layer with the average S-wave velocity of 1 km/s.