Generating ground motions using the Fourier amplitude spectrum

In this paper, a simplified method aimed at generating ground motions records using a Fourier amplitude spectrum model obtained by a response spectrum is presented. The use of white noise with specific conditions on the variance and inter-frequency correlation allows to get a realistic variability of the Fourier amplitude spectrum. Moreover, a two-corner frequency model is defined from empirical ground-motion data and a filter is applied to capture the attenuation at high frequencies. The generated series of stochastic ground motions accurately match the mean value of the target response spectrum. The procedure was tested on an Italian site and results showed excellent matching in terms of mean and dispersion values with either the median obtained from ground motion prediction equations (GMPE) and conditional mean spectra (CMS). The advantage of the proposed methodology is that time-histories can be generated with low computational effort, they do not need scaling or frequency content adjustments and are spectrum compatible with a given target spectrum.