A PSHA for Mexico City based solely in Fourier-based GMM of the response spectra

A key component of any probabilistic seismic hazard analysis (PSHA) is the capability of predicting ground motion intensities for future earthquakes with different characteristics and occurrence frequencies. Ground motion models (GMMs) have been the preferred tools to predict, in a probabilistic manner, the size of the ground motion as a function of the earthquake's rupture parameters (i.e., location, depth, orientation of the rupture plane, rupture area, and rupture shape, among others). For this purpose, two different approaches have been used in PSHA practice. The first one, widely used and with hundreds of available models, are semi-empirical derivations of GMMs for response spectral values. The second one, which is less used in professional practice, requires the development of the predicting model for the Fourier amplitude spectrum (FAS) and the duration of the strong motion, to convert them into response spectral values through random vibration theory (RVT). The second approach, although computationally complex, should be preferred when site-effects are considered in the PSHA since these can be incorporated in a more efficient and theoretically correct way. Also, this second approach allows estimating hazard levels for different damping ratios and other hazard intensity measures such as peak ground velocity (PGV) and peak ground displacement (PGD) straightforward. This paper shows the methodology and results of a PSHA for Mexico City, a place with well-known existence and relevance of site-effects, carried out exclusively using Fourier-based GMMs, developed to be representative for the different types of earthquakes that contribute to earthquake hazard in the city (i.e., interface, intraslab, and crustal). The strategy to estimate ground motions in Mexico City consists in obtaining FAS-based ground motion models for a reference firm ground station and multiplying the computed FAS by an empirical transfer function that describes the frequency-dependent amplification factors between the reference station and the soft site of interest.