Abstract

The coda method has been utilized to determine the seismic response at various sites on the island of Hawaii using data from the Hawaiian Volcano Observatory (HVO) seismic network. Site amplification factors are calculated using spectral ratios of S-wave coda amplitude measured for many events and lapse times between 33 and 59 sec for frequencies between 1.5 and 15.0 Hz. We find that these site amplification factors vary by factors of as much as 12 for a given frequency due to differences in absorption and/or impedance properties beneath the site.

Similar to results from Japan, central California, and Long Valley, the coda site amplification factors showed strong correlation with station residuals for magnitude determined by Johnson (unpublished results) for the HVO network between 1.5 and 6.0 Hz. Johnson (unpublished results) made an independent determination of event magnitudes and station site corrections by measuring averaged windowed amplitudes of the S wave and its coda using a method previously tested with data from the southern California seismic network (Johnson, 1979). Station calibrations were simultaneously determined and included the effects of instrument gain and site effect. After correcting Johnson's (unpublished results) results for the instrument gain provided by HVO, we obtain correlations similar to the previous works, with the best correlation appearing at 3.0 Hz.

Despite the complex subsurface structures on the island of Hawaii, the frequency dependence of our coda site amplification factors displays a remarkable regional trend. In some cases, plausible causes related to the underlying geologic structures can explain the similar frequency dependence; in other cases, the similarity cannot be explained or may be coincidental. Similar frequency dependences among stations accompanied by a large variability in the absolute level of the site effect may be due to inaccurate instrument calibrations.

First Page Preview

First page PDF preview
You do not currently have access to this article.