Spectral Comparison of Real Spectra with Site Effects Included vs MOC-2008 Teorical Spectra for Guadalajara City , Mexico

Guadalajara city is the capital of the Mexican federal state of Jalisco. It is located close to the Pacific coast and is subjected to a large seismic risk. We present a seismic response study for some sites along the city. We calculated response spectra of shear-waves by using seismic records of actual earthquakes registered on rock and near the source as Green functions and propagated them trough a flat and horizontal layered media using a linear equivalent method to compare them with the response spectra calculated according to the Federal Commission of Electricity (CFE) seismic design buildings Manual (MOC-2008, 2008) which is widely used as reference on construction projects in Mexico. Our results show that MOC-2008 manual underestimates the spectral amplitudes and frequency band for the response spectra.


Introduction
It is well known that seismic damage distribution is strongly influenced by physical and dynamic properties of the soil.The cyclical load response capacity to earthquakes for example, depends on these properties.Seismic response evaluation is one of the most important problems to solve in seismic engineering.The Analysis of the seismic response of the soil is used to predict the response of the seismic movement in the surface, which in turn is essential to obtain the design spectra to evaluate the risk.Evaluation is especially important in places with potentially liquefaction risk.
Design spectra is a tool that allows us to evaluate the forces to which the structures will be submitted in dependency on its own dynamic characteristics.Guadalajara city occupies the second site in Mexico, on population and economy growth.During the last years, more modern and greater civil structures have appeared at the whole city (Figure 1).
Federal Commission of Electricity (CFE) seismic design buildings Manual [1] has been used as a quasi-official reference for construction since its appearance in 1993.A building regulation update, based on this type of studies is required as the recurrence cycle of earthquakes for this region of México which is about 80 to 100 years, and with the knowledge that the last devastating earthquake occurred on June 1932 [2], and it turns clear that this type of studies should be the bases to carry out a review of the building regulations.
In this article we estimate seismic response at eight sites distributed within the urban zone of Guadalajara city using records from real earthquakes occurred in along the subduction zone assuming one-dimensional propagation in order to test the current regulation.
Seismic records and earthquakes used are shown in Table 1.

Earthquakes and Seismic Records
Site effect due to geological conditions is one of the main factors which contribute to damage distribution during earthquakes.Subsoil impedance contrasts can significantly amplify the shaking level, as well as increase the duration of strong ground motion.The first action to take to prevent damages is to know site effect distribution.Subsoil dynamic properties allow us to predict the site response when it is submitted to earthquakes or dynamic charges.
In order to estimate the seismic response we chose earthquakes occurred along the subduction zone and seismic records (accelerations) from the closest stations installed over hard rock in order to prevent site effect to be used as Green Functions.We selected seismic stations far enough to assure vertical incidence at Guadalajara sites.
We selected three earthquakes: Tecomán, 2003 (Mw 7.5) recorded at Manzanillo station 65 km northwest from the epicenter, Colima, 1995 (Mw 7.3) recorded at  1. PGA column refers to maximum acceleration on the station.Seismic records used as Green functions are shown in Figure 3.As the interest is on SH waves propagation, we used only the horizontal components.

Sites of Study
We selected eight sites distributed within the urban zone of Guadalajara city (Figure 4) where Lazcano Diaz [3] performed a site characterization which included thicknesses, density, shear-wave velocity (Vs) among other parameters.
Table 2 shows some basic parameters at each site: location, shear-wave velocity for the first 30 m (Vs30), dominant period from Federal commission of electricity CFE [1] calculated with PRODISIS [4] using the velocity profile, dominant period from Lazcano Diaz [3] and NEHRP [5] soil classification.In all cases, classification corresponds to rigid soil with shear velocities between 180 to 360 m/s.
As it is important to understand shear-wave propagation obtained in this study, we reproduce shear-wave velocity profiles from Lazcano Diaz [3] in Figure 5.

Wave One-Dimensional Propagation. The Linear Equivalent Method
We propagated through a one-dimension layered media SH waves computed from the horizontal components of   or a damping of rs of any project from a probabilistic po methodology (LEM) [6].Soil shear modulus depends on the deformation resistance of the soil.Damping ratio is associated with dissipative behavior of the soil to cyclic inputs.In a linear equivalent scheme, non-linear behavior of shear modulus and damping ratio are represented by curves as those shown in Figure 6 proposed by many authors ( [7,8], among others).
We used ProShake [9] soft ave propagation in order to add site effect to the original record assuming:  Harmonic waves:

Spectra (5%) and Design Spectra MOC-2008
ng the seismic recor media, we built seismic records with site effect included for each site.We then used them to calculate, the response spectra using Degtra program [10].
Response spectra are usually estimated f 5% because it is representative of the observed damping of reinforced concrete and structural steel.Figures 7  and 8 show the response spectra for a damping of 5% for each site.We show the spectra for both horizontal components using each real record as well as the mean plus standard deviation.
Seismic paramete int of view (the deterministic way is no longer used), requires certain knowledge of seismic activity of the area.The probability of occurrence of a seismic event greater than a reference is given by  where "P" is the probability that an "X earthquake" in building codes.
" event is less than an x event previously defined in a time t.This probability in terms of the project can be defined considering the probability of the usual buildings which is of 10% in 50 years.It is associated with the existence of an earthquake which occurs every 450 years and it is named "rare For the same sites, we calculated the response spectra for each site according to MOC-2008 [1] procedure which is based on probabilistic criteria.Figures 9 and 10, shows mean plus standard deviation response spectra for both horizontal components calculated in this study and the responses for a rock site, for edge state of service and for collapse according to MOC-2008 [1] procedure.We observe the higher amplitudes at sites Av.Patria fortunately we do not have records to analyze in the soud namericana, Torrena, J. del Bosque and Gran Plaza where of around 3.12 g and the lower amplitudes of around 1.2 g at the sites located in downtown, Biblioteca and Rotonda.Amplitudes are higher for response spectra using seismic records and frequency band is shorter for response spectra from MOC-2008 [1].The flat zone in the last case is from about 0.1 to 1.6 sec while significant amplitudes from seismic records extend up to 1.6 seconds.

Co
Our results show th zone of Guadalajar fication of the seismic waves despite NERPH soil classification which corresponds to rigid soil with shear velocities between 180 to 360 m/s.
It is obvious to conclude that the actual regulation must not be applied indiscriminately to every site where there is a construction project but it is necessary to make studies of this kind to characterize the site effect at each site.While MOC-2008 manual i odels, a comparison with real records is needed.This kind of studies should be complemented with other studies such as seismic microzonation.

This research was partially supported by Institutional
Copyright © 2013 SciRes.OJCE

Figure 1 .
Figure 1.Large buildings appear each year in Guadalajara city.

Figure 2 .
Figure 2. Location of Guadalajara city and earthquakes used in this study.

Figure 4 .
Figure 4. Location of sites selected for this study (triangles).
dotted line represents spectra from the Oaxaca 1999 record.Green line is the mean plus standard deviation.(a) Biblioteca site; (b) Rotonda site; (c) J. del bosque site; (d) Torrena site.

Figure 8 .
Figure 8. Response spectra for 5% damping for the eight sites using real records propagating through a layered media.Continuous line represents spectra using the Colima 2003 record, continuous line with dots, the spectra using Colima 1995 record,

Figure 9 .
Figure 9. Response spectra calculated according to MOC-2008 procedure vs response spectra using seismic records.The different curves represent mean plus standard deviation response spectra for both components, response for a rock site, for edge state of service and for collapse.(a) Biblioteca site; (b) Rotonda site, (c) J. del bosque site; (d) Torrena site.

Figure 10 .
Figure 10.Response spectra calculated according to MOC-2008 procedure vs response spectra using seismic records.The different curves represent mean plus standard deviation response spectra for both components, respon for a rock site, for edge state of service and for collapse.(a) Gran Plaza site; (b) Eulogio Parra site; (c) Av.Patria site; (d) U ite.

Table 1 . Main features of earthquakes used in this study.
same station 25 km northeast from the epicenter and Oaxaca, 1999 (Mw 7.5) recorded at LANE station 19 km west from the epicenter.Locations of earthquakes used in this study are shown in Figure2.Basic information from earthquakes used are shown in Table the

Table 2 . Main features of the sites used in this study.
As it is important to understand shear-wave propagation obtained in this study, we reproduce shear-wave velocity profiles from Lazcano Diaz, 2007 in Figure5.