UNDERSTANDING SITE EFFECTS THROUGH DOWNHOLE ARRAY SEISMOGRAM INVERSION

ASSIMAKI, D., Georgia Institute of Technology, School of Civil and Environmental Engineering, Atlanta, GA 30332, and STEIDL, J.M., Institute for Crustal Studies, University of California, Santa Barbara, CA 93106, dominic.assimaki@ce.gatech.edu, steidl@crustal.ucsb.edu.

In current site response methodologies, elastic impedance profiles are obtained through geotechnical and geophysical testing, while attenuation is either approximated by means of empirical correlations or inferred based on limited laboratory data. At larger strains, soil properties are primarily evaluated through laboratory testing, where the in-situ stress-state and seismic loading cannot be accurately reproduced. As a result, ground motion predictions usually compare poorly with weak motion recordings, a discrepancy even further aggravated for strong ground motion. Downhole instrumentation, which has been increasingly deployed in seismically active areas over the past years, may be a valuable complement to in-situ and laboratory investigation techniques. We here present a seismic waveform inversion algorithm for the estimation of elastic and equivalent linear soil properties using downhole array recordings. Based on a genetic algorithm complemented by a local least-square fit operator, the proposed scheme can estimate efficiently elastic impedance and attenuation profiles, and quantify approximately the extent of nonlinearity exhibited by the material during strong ground motion. Results presented for weak and strong motion data from the Miyagi-Oki Earthquake recorded by the Strong Motion Network Kik-Net illustrate that, providing critical constraints on interpretation methods and information on the soil behavior and site response over a wide range of loading conditions, currently available downhole array data may be used to establish a unified methodology of strong motion site response assessment.