Site response in a representative region of Manzanillo, Colima, Mexico, and a comparison between spectra from real records and spectra from normative

Abstract

The October 1995 Mw=8.0 Colima-Jalisco earthquake triggered damage in buildings located in Manzanillo city (Tena, A, 1997) [1]. Great damages were reported in different regions of that city. It was also observed liquefaction in the Manzanillo port. In order to improve the knowledge of the subsoil of Manzanillo, Rayleigh waves were inverted applying spatial autocorrelation (SPAC) to array measurements of microtremors at five representative sites of the city. This procedure allowed us to obtain a shallow S-wave velocity profile at each site. We compute the transfer function for each profile and correlate the results with geology. We calculate theoretical surface records. For this purpose, we used a numerical simulation of wave propagation of an input signal (Green function) through the structure. In this case, records of two great subduction earthquakes were used as Green functions. We use these surface signals to estimate response spectra for each site. On the other hand, design spectra according to the manual of seismic design of the Federal Commission of Electricity 2008 edition (MOC-CFE) [2] were obtained and compared with previous results.We observed important differences between the seismic responses in different regions of Manzanillo. For instance, the response at site 5 (alluvial deposits) for a period of 0.25s is twice the response of site 4 (sand deposits) for the same period. Therefore, the site effects are relevant in Manzanillo.For the majority of the cases we observe a reasonable agreement between the values of the dominant periods computed according to the MOC-CFE [2] normative and those obtained directly from Transfer Function (TF). On the other hand, we found significant differences between design spectra and response spectra obtained from seismic records for the 0.1–0.6s interval. Amplitudes are underestimated for the first ones.