Investigating the spatial variability of ground motions during the 2017 Mw 7.1 Puebla-Mexico City earthquake via idealized simulations of basin effects

Abstract

We use strong motion records of the Mw 7.1 19 September 2017 mainshock and one- and two-dimensional (1D and 2D) simulations at six stations along a linear array across Mexico City, to shed light to the site amplification and ground motion variability observed during the mainshock. The 2017 Puebla-Mexico City earthquake occurred at the edge of the flat slab segment of the subducted Cocos plate beneath central Mexico, rupturing the top half of the plate. Compared to the 1985 subduction zone event, the intraslab Mw 7.1 mainshock was characterized by a much richer high-frequency content. This characteristic transpired into site amplification of the incoming shaking by the sediments of the transition zone, contrary to the 1985 Mw 8.1 Michoácan earthquake that was amplified by the deeper lake sediments. By means of idealized 1D site and 2D basin models, this paper seeks to disentangle site response from basin resonance and basin edge diffraction effects manifesting in the ground motion records. We specifically compare ground surface observations from the 2017 Mw 7.1 event to 1D and 2D analyses in the time and frequency domain; and empirical transfer functions from several earthquakes to numerical transfer functions computed using 1D and 2D models. Comparison of empirical and simulated amplification spectra shows that, although our 2D model does not capture all the peaks and troughs of the frequency response, theory and observations are nonetheless in good agreement across the entire frequency spectrum. And while frequency peaks of the 1D and 2D models are aligned in the range f > 0.2 Hz, the long period (f < 0.01 Hz) match between simulations and observations demonstrates that the 2D model appropriately captures the effects of the largest characteristic lengths of the basin, namely the basin width. In the intermediate frequency range (0.01 Hz < f < 0.2 Hz), our 2D model captures the scattering of the energy by the concave shape of the basin-rock interface via spectral deamplification, albeit not as strongly as is evidenced by the observations. One-dimensional site response on the other hand cannot capture the larger scale features below 0.2 Hz, which were likely relevant to the performance of lifelines and other infrastructure networks.

Additional details