All recent destructive earthquakes (Armenia 1988, Loma Prieta 1989, Iran 1990,
Philippines 1990, Northridge 1994) provided strong evidence that local site
conditions amplify the seismic ground motions significantly as compared
to those recorded at the bedrock. Accounting for such site effects in building
code, land use planning or design of critical facilities is an important
goal of earthquake hazard mitigation programs. Existing methods for site evaluation
utilize expensive and difficult to obtain downhole measurements as
well as seismograms recorded at the site free surface.
We develop a novel framework for identifying the site topography
(reflectivity) based on multiple seismograms recorded at the free site surface
only. Each seismogram will be considered as a distorted version
of the seismic activity at the base of the sediment site. The distortion is
introduced by the reflectivity of the path between the seismic input and the
recording station. Assuming that the seismic inputs have similar statistical
characteristics, we develop robust schemes to reconstruct the reflectivity of
each path, and subsequently correlate it with the topography of the
corresponding recording location. This approach will provide a new and
economically attractive method for estimating the site characteristics and
will eliminate the need for downhole measurements.
This project has been funded by the National Science Foundation
Collaborators
Dr. Aspa Zerva
Civil and Architectural Engineering Dept., Drexel University
Dr. Pierre-Yves Bard, University of Grenoble, Grenoble France