Pontificia Universidad Católica de Chile Pontificia Universidad Católica de Chile
Muñoz A. and Sáez E. (2018)

Numerical estimation of site effects in the city of Arica due to natural soil variability using the spectral elements method

Revista : Bulletin of Earthquake Engineering
Volumen : 16
Número : 1
Páginas : 459-478
Tipo de publicación : ISI Ir a publicación


In recent years, because of a seismic gap of 136 years in the north of Chile, several studies have been carried out with the purpose of characterizing the dynamic properties of the soils of the most populated cities in this area. The purpose of this work is to estimate, by numerical modeling, the seismic amplification of a densely populated area of the city of Arica. Spatial distribution of main soil dynamic properties has been obtained from a detailed geophysical survey, including surface-wave based methods and gravimetry. To estimate the site effects, we solve the wave propagation equation in full heterogeneous media by the Spectral Element Method. This method allows including the topography, the irregular contact between soils and bedrock and heterogeneities of main materials’ properties along the computational domain. The inelastic behavior of the soil has been considered, using stiffness degradation and damping curves. The computations were done using the high performance open-source numerical code SPEED. Results from the distribution of the Peak Ground Acceleration (PGA) are compared against standard 1D horizontally layered modeling using equivalent linear and nonlinear viscoelastic approximations of nonlinear soil behavior. Linear viscoelastic results indicate that the north area of the city of Arica is not especially prone to tridimensional site effects induced by soil’s natural variability and the apparent bedrock geometry. Nevertheless, non-conservative differences of about 20% between 1D v/s 3D analyses were found under inelastic material behavior. In general, the two 3D approaches followed to include soils’ nonlinearity provide similar site amplification estimates.