Ground Amplification from Earthquakes

Abstract originally published by the American Geophysical Union

Poster presented at 1997 Fall Meeting of the AGU

Site Response for Seattle and Source Paramenters of Earthquakes in the Puget Sound Region; A. Frankel (US Geological Survey), D. Carver, E. Cranswick, M. Meremonte, T. Bice, and D. Overturf; AGU EOS Transactions, v. 78, no 46, p. F433.

 

We have operated 14 portable digital seeismographs in urban Seattle since May 1996. The primary purpose of this deployment is to determine the site response of various surficial geologic units by recording the ambient seismicity, but the data area also useful for evaluating earthquake source characteristics. We have recorded about 20 local earthquakes (M 2.0 - 4.9) to date, including aftershocks of the May 1996 Duvall earthquake, a M 3.5 event south of Seattle in Feb. 1997, and the M 4.9 Bremerton earthquake of June 1997.

We used spectral ratios with respect to a reference rock site to determine the relative site response. Site amplifications as great as a factor of ten in some frequency bands are observed relative to the rock site. We find the highest amplification for sites on artificial fill, such as Harbor Island. A strong resonance at about 2 Hz is observed for our site near the Kingdome. Substantial amplification is observed for sites on Pleistocene deposits of the upland portions of Seattle.

Spectral ratios were calculated for colocated pairs of mainshocks and aftershocks such as the Feb. 1997 events and the Bremerton sequence. These spectral ratios were used to determine the source corner frequencies of these events. Based on these source spectra, we claculated the site response specra relative to a half space for the sites, including the reference rock site. Source corner frequencies of other earthquakes were found by dividing their spectra with those of earthquakes with known source corner frequencies.

We detemined low stress drops of about 1-10 bars for events with magnitudes from 2.0 to 3.5. We found a stress drop of approximatelly 30 bars for the Bremerton mainshock, but the apparent corner frequencies of this evernt are affected by directivity. Spectral ratios of coloated events are often fit best with a steep spectral falloff of w- exp (-3). This steep apparent falloff may be caused by directivity and radiation pattern effects.