Efficient Estimation of the Evolutionary Spectra of Earthquate Signals Based on Wavelet Analysis

William Marsh Rice University, Houston TX

Investigators

Abstract

The research program examines novel procedures of spectral estimation related to the processing of seismic data. Wavelet analysis is employed to describe the time-dependent characteristics and localized events present in earthquake records. The wavelet transform is increasingly being used in a variety of applications. Various wavelets families are examined for use in seismic spectrum estimation problems. The suitability of the various methods for estimating the cross spectra of two accelerograms is investigated, and the spatial variability of seismic records examined. The method developed will be applied in processing a small but representative number of records from important earthquakes, such as the Kocaeli seismic event in Turkey, and the Northridge seismic event in California. The evolutionary spectral content of seismic records, obtained via the wavelet method, has potential application to the safety of critical constructed facilities.