COMPARISON OF THE BEHAVIOR OF SITE FROM STRONG MOTION DATA OF 1985 CENTRAL CHILE EARTHQUAKE (M S =7.8) AND MICROTREMORS MEASUREMENTS

Size: px

Start display at page:

Download "COMPARISON OF THE BEHAVIOR OF SITE FROM STRONG MOTION DATA OF 1985 CENTRAL CHILE EARTHQUAKE (M S =7.8) AND MICROTREMORS MEASUREMENTS"

Stella Preston
5 years ago
Views:

1 Paper No. CBSLR COMPARISON OF THE BEHAVIOR OF SITE FROM STRONG MOTION DATA OF 1985 CENTRAL CHILE EARTHQUAKE (M S =7.8) AND MICROTREMORS MEASUREMENTS Felipe LEYTON 1, Sergio RUIZ 2 ABSTRACT In this study, we compare the estimation of the fundamental vibration frequency of the soil obtained from strong motion data (from the Valparaíso 1985 earthquake, M S =7.8) and microtremors measurements. We estimated the fundamental frequency for 12 sites with different soil conditions, using the spectral ratio of the horizontal component over the vertical. The results indicate that, in cases where the strong motion data presents a clear peak or no peak, the microtremors data agrees. This finding enables the use of microtremors data to estimate the behavior of the soil during a large earthquake. Keywords: soil s fundamental frequency, microtremors, strong-motion INTRODUCTION Chile is one of the most seismically active countries in the world (Lomnitz, 1971; 2004); hence, understanding the dynamic behavior of soils during large earthquakes becomes one of the most important challenges for geotechnical and seismic engineering. For this purpose, Chile has an important advantage with respect to countries with low to moderate seismicity, due to the availability of acceleration records registered in epicentral zone, for example Central Chile 1985 (Ms=7.8), Tarapacá 2005 (M=7.8), Tocopilla 2007 (M=7.7), and recently Maule 2010 (Mw=8.8). This study analyzes the accelerograms obtained during the Valparaíso 1985 earthquake (interplate thrust event, magnitude 7.8, with a rupture area of 200 x 100 km, approximately, Figure 1 (Comte et al., 1986; Barrientos and Kausel, 1990)); mostly motivated by the observations made by Astroza and Monge (1989) and Menéndez (1991) who reported differences from 0.5 to 2 points in Modified Mercalli Intensity (MMI) observed in close-by areas, mostly associated to different site conditions (Astroza and Monge, 1991). In order to understand the different behavior of soils during large earthquakes, this study compares the site effect observed from strong-motion accelerograms (from the Valparaíso 1985 earthquake) with results obtained from microtremors measurements performed in the same places. In order we use the methodology proposed by Ruiz and Saragoni (2009), complementary to the used by Ruiz and Leyton (2010), using the spectral ratio between the horizontal and vertical component of the accelerogram (H/V) (Lermo and Chávez-García, 1993). For the microtremors measurements we use the same H/V spectral ratio (Nakamura, 1989, 2000). The accelerograms registered in epicentral zone during the Valparaíso 1985 earthquake (see details in Saragoni et al., 1986; or Celebi, 1987) have relevant information of the behavior of free vibrations of soils 1 Professor, Department of Civil Engineering, Universidad Diego Portales, felipe.leyton@udp.cl 2 PhD candidate, Department of Geology, Universidad de Chile.

during a strong event, enabling the observations of vibration modes, in particular the determination of the fundamental frequency, using the autocorrelogram and the Fourier spectra (Saragoni and

2 during a strong event, enabling the observations of vibration modes, in particular the determination of the fundamental frequency, using the autocorrelogram and the Fourier spectra (Saragoni and Ruiz, 2004; Ruiz and Saragoni, 2004, 2009). However, it is not always possible to find a strong-motion record available in the specific site; hence, other techniques should be used to estimate the dynamic behavior of soils, being one of the most used the microtremors measurements. Figure 1. Sites where the Valparaíso 1985 earthquake was recorded and microtremors measurements where performed; the dashed area represented the rupture area (modified from Ruiz and Leyton, 2010). The spectral ratio of the horizontal over vertical component (H/V) of microtremors is one of the most widely used techniques these days, initially proposed by Nogoshi and Igarashi (1970, 1971) and extensively popularized by Nakamura (1989, 2000). This technique has been previously applied in Chile to study Santiago and Valparaíso (Pasten, 2007; Bonnefoy-Claudet et al., 2009; Pilz et al., 2009) showing that, in most cases, the fundamental frequency of the soil can be estimated, but in others no clear peak can be observed. Despite the extensive use of the H/V in microtremors, it s still not clear whether the behavior of the soils determined with this technique will be maintained during a large earthquake (Lermo and Chávez-García, 1994).

3 This study compares the responses of the soil observed using strong-motion accelerograms of the Valparaíso 1985 earthquake and microtremors measured at the same sites (Ruiz and Leyton, 2010) (see Figure 1 for the specific locations). The microtremors measurements were performed using 2 kinds of 3- components instruments: a 5-sec Lennartz and a 4.5-Hz GVB, in order to check whether the cheaper short period instruments give similar results as the widely used 5-sec Lennartz. DATA AND METHODOLOGY A total of 25 accelerographic stations registered the Valparaíso 1985 earthquake (Saragoni et al, 1987; Celebi, 1987); all of them located in different types of soils (Araneda and Saragoni, 1994), enabling the observation of different site responses during a large earthquake (Ruiz and Leyton, 2010). In few words, the site response observed in the strong-motion accelerograms can be divided into 2 groups: one where the free vibrations of the soils predominate in the record and others where no particular frequency predominates (Saragoni and Ruiz, 2004; Ruiz and Saragoni, 2004, 2009). An example can be observed in Figure 2, showing the Fourier spectra and the autocorrelogram for the Viña del Mar and Melipilla stations. From this figure we can see the 2 characteristic site response observed in all accelerograms from the Valparaíso 1985 earthquake. Figure 2A is dominated by the soils free vibrations, shown by the predominant frequency observed in the Fourier spectra and autocorrelogram (Saragoni and Ruiz, 2004; Ruiz and Saragoni, 2004, 2009); for this particular case, the fundamental frequency of the soil at Viña del Mar is fs = 1.5 Hz. On the other hand, Figure 2B shows the Fourier spectra and autocorrelogram for the Melipilla record, where no predominant frequency can be observed (Ruiz and Leyton, 2010). In this study, we use the spectral ratio of the horizontal over the vertical component (H/V) from the strong-motion records to estimate the fundamental frequency of the soil during a large earthquake. Following Lermo and Chávez-García (1993), we compute the spectral ratio between the horizontal and vertical components, we smooth it using a homogenous filter in log-scale (Konno and Ohmachi, 1998); the final result is the logarithmic average of the 2 components. The accelerograms are filtered between 0.15 and 0.25 Hz using a 4 th order Butterworth filter, no needing to correct the instrumental response because the instruments corresponds to SMA-1 (Boore and Bommer, 2005). We also performed microtremors measurements in 12 sites (shown in Figure 12), in order to estimate the fundamental frequency of the soil using the spectral ratio H/V. For the measurement, we used 2 kinds of 3-component instruments: a 5-sec Lennartz, widely used in this studies (Bard and SESAME Work Group WP02, 2005) and a 4.5-Hz GVB, with a reliable signal up to 1 Hz, useful in this kinds of measurements (Bindi et al., 2008). At each site we registered microtremors for 30 to 60 minutes, depending on the surrounding human activity.; we used a sampling of 100 Hz in both instruments. We processed both signals in the same way: we used a 60-sec time window to compute the Fourier transform from each component (as recommended by Bard and SESEAM Work Group WP02, 2005), filtered using a Konno- Ohmachi smoothing function (Konno and Ohmachi, 1998), and then taken the spectral ratio of the horizontal over vertical component (H/V). Given the large number of 60-sec windows, we are able to estimate the error at each point (standard deviation), done in log scale; the final result corresponds to the logarithmic average of the 2 horizontal components.

4 Figure 2. A Viña del Mar horizontal components: a) Accelerograms, b) Fourier Spectra, and c) Autocorrelograms of the accelerograms (continuous line), along with the theoretical model (dashed line). B Melipilla horizontal components: a) Accelerograms, b) Fourier Spectra, and c) Autocorrelograms of the accelerograms (continuous line), along with the theoretical model (dashed line). The column between the figures shows the shear wave velocity (Vs) profile, with depth shown in meters at the left and Vs in m/s at the middle (Araneda and Saragoni, 1994). Modified from Ruiz and Leyton (2010). RESULTS Comparison of microtremors spectral ratio of 5-sec Lennartz and 4.5-Hz GVB As a first conclusion, we found that the results obtained with the 5-sec Lennartz and 4.5-Hz GVB are practically the same, in the frequency range of interest (from 0.1 to 10 Hz), as shown in Figures 3 and 4. In these Figures, for those cases were we were able to find the soils fundamental frequency from the autocorrelogram (fs), it is shown below the corresponding name; for those cases that were not possible, nothing is shown. It is worth noting that, in general, the H/V spectral ratio in the Lennartz instrument usually present high amplitudes at low frequencies (lower than 0.1 Hz), probably due to electronic noise. This result probably imposes serious difficulties the interpretation of the results, as shown by Bonnefoy et al. (2009).

Comparison of spectral ration of accelerograms and microtremors In Figures 5 and 6 we show the comparison of the H/V spectral ratio for the accelerograms and microtremors using the 4.

5 Comparison of spectral ration of accelerograms and microtremors In Figures 5 and 6 we show the comparison of the H/V spectral ratio for the accelerograms and microtremors using the 4.5-Hz GVB, at the 12 sites studied in this work. The H/V ratio of the accelerograms corresponds to the logarithmic ratio of the 2 horizontal components (dashed lines in Figures 5 and 6); the H/V ratio for microtremors considers the logarithmic average of all the 60-sec windows (continuous line in Figures 5 and 6), with its corresponding standard deviation (gray area in Figures 5 and 6). Again, for those cases were we were able of defining a fundamental frequency from the autocorrelogram (fs), it is shown below the corresponding name; for those cases that were not possible, nothing is shown. Next to each graph we show a column with the shear wave propagation velocity (numbers inside the column, in m/s) and the corresponding depths (in m) at its side. Figure 3. Comparison of the results obtained with the 5-sec Lennartz (dashed lines) and the 4.5- Hz GVB (continuous lines), with its corresponding standard deviation (gray area). The columns next to each figure shows the shear wave velocity (Vs) profile, with depth shown in meters at the left and Vs in m/s at the middle. For those cases that we were able to compute the fundamental frequency using the autocorrelogram (fs), we presented below the corresponding name.

From Figures 5 and 6 we can see that, sites with H/V accelerograms showing a clear peak, such as Llolleo, Viña del Mar, Iloca, Ventanas, and Cauquenes, the H/V of microtremors also presents a clear

6 From Figures 5 and 6 we can see that, sites with H/V accelerograms showing a clear peak, such as Llolleo, Viña del Mar, Iloca, Ventanas, and Cauquenes, the H/V of microtremors also presents a clear peak; more over, for all of these cases, the autocorrelogram was able to estimate a fundamental frequency close to the one shown by either method. On the other hand, were no clear peak could be seen in the H/V of accelerograms (as in Zapallar, Valparaíso UTFSM), the microtremors present no peak, as well as the autocorrelograms. Special cases are: (i) Melipilla and San Felipe, where the spectral ratio of the accelerogram show a clear peak (or a series of them), but the microtremors and autocorrelograms did not, (ii) Llayllay and La Ligua, where the spectral ratio of the accelerogram show a series of peak, that might (Llayllay) or migh not (La Ligua) agree with the value estimated from microtremors and autocorrelograms, and (iii) Hualañe, where the spectral ratio of the microtremors did not show a peak, but the accelerogram and autocorrelogram did show a clear fundamental frequency with similar value. Figure 4. See Figure 3 caption for details.

Figure 5. Comparison of the spectral ratio H/V of accelerograms (dashed lines) and microtremors (continuous lines), with its corresponding standard deviation (gray area).

7 Figure 5. Comparison of the spectral ratio H/V of accelerograms (dashed lines) and microtremors (continuous lines), with its corresponding standard deviation (gray area). The columns next to each figure shows the shear wave velocity (Vs) profile, with depth shown in meters at the left and Vs in m/s at the middle. For those cases that we were able to compute the fundamental frequency using the autocorrelogram (fs), we presented below the corresponding name. The results we present here are complementary to those shown by Ruiz and Leyton (2010) that used the average Fourier spectra to estimate the fundamental frequency from the accelerograms. This last method enabled a better comparison for Melipilla, San Felipe, La Ligua, and Llayllay; probably due to the fact that the accelerograms do not fullfil all the requirements to adequately perform the H/V spectral ratio. Finally, as shown by Ruiz and Leyton (2010), the fundamental frequency found by the H/V spectral ratio of microtremors is consistently higher than the one observed from the accelerograms (e.g. Ventanas); however, this point is hard to observe in the log-scale used in Figures 5 and 6, the usual way of presenting these kinds of results.

8 Figure 6. See Figure 5 caption for details. CONCLUSIONS AND FINAL COMMENTS The results we presented here agree with those presented by Ruiz and Leyton (2010): in those sites where the response of the soil during a large earthquake is controlled by its fundamental free vibration mode, this mode can be estimated using the spectral ratio of the horizontal over vertical component (H/V); on the other hand, for those cases that during a large earthquake no clear frequency predominates, the H/V spectral ratio does not present a clear peak. However, the fundamental frequencies observed during a large earthquake are consistently lower to those estimated from microtremors, probably due to the non-linear response of the soil, this aspect requires further studies in the future. In general, we found that the accelerograms H/V spectral ratio present a more complicated results compared to microtremors due to the fact that, during large earthquakes, the soils not only respond with their free vibration modes, but also present forced response, produced by the seismic waves coming from the source (Ruiz and Saragoni, 2009).

9 Another interesting result that we obtained in this study is the fact that the 4.5-Hz GVB presents the same results that the 5-sec Lennartz, in the frequency range of interest (0.1 to 10 Hz), but the second kind of instrument usualy shows an anomalous large peaks at lower frequencies. Hence, we recomend the use of short period instruments (such as the GVB) for microtremors measurements, as shown in previous studies (Bindi et al., 2008). In the results we found in this study, there seems to be a correlation between the observation of a fundamenal frequencies and lower Vs; nevertheless, La Ligua is a clear counter example with low Vs and no apparent fundamental frequency. As mentioned before, we found that the H/V spectral ratio of accelerograms might not be the best procedure to estimate the site s fundamental frequency, while the autocorrelogram presented by Ruiz and Saragoni (2009) could be a more plausible technique. We believe that this result might be produced by the fact that accelograms at epicentral distances are strongly influenced by the seismic source and/or due to amplifications observed in the vertical component (Ruiz and Saragoni, 2009); however, this issue requires further studies. The results we present in this study are useful in seismic hazard and microzonation studies because they enable the use of microtremors to estimate the behavior of the soil during a large earthquake; nevertheless, it is important to note that the seismic response of the soil is only part of the seismic demand during a large earthquake, while the seismic waves coming from the different asperities are responsible for a large amount of the observed damage (Ruiz et al, 2010). AKNOWLEDGEMENTS We would like to thank Ricardo León for his invaluable help during the fieldwork and the Chilean Seismological Service for providing us the instruments used during this study. This paper funded by Fondecyt and Millennium Nucleus in Seismology and Seismotectonics. Figures were made using GMT (Wessel and Smith, 1991). REFERENCES Araneda, C. and Saragoni, G.R. (1994) Project of geological survey of strong motion site in Central Chile. Report for: Kajima Institute of Construction Technology of Tokyo, Department of Civil Engineering, University of Chile,. Astroza, M. and Monge, J. (1989) Aumento de intensidades según las características geológicas del los suelos de fundación, sismo del 3 de marzo de 1985, Anales de la Universidad de Chile, Universidad de Chile,. Astroza, M. and Monge, J. (1991) Seismic microzones in the city of Santiago. Relation damage-geological unit, in Proceedings of the Fourth International Conference on Seismic Zonation, Vol. 3, Stanford, CA, USA, August 25 29, pp , Earthquake Engineering Research Institute, Stanford, CA, USA. Barrientos S.E. and Kausel, E. (1990) Rupture process of the 1985 central Chile earthquake, Revista de Geofísica 46, (In Spanish). Bard, P.-Y. & SESAME-Team (2005) Guidelines for the implementation of the H/V spectral ratio technique on ambient vibrations measurements, processing and interpretations, SESAME European research project EVG1-CT , deliverable D23.12, available at

10 Bindi, D., Marzorati, S., Parolai, S., Strollo, A., and Jäkel, K.H (2008) Empirical spectral ratios estimated in two deep sedimentary basins using microseisms recorded by short-period seismometers. Geophys. J. Int., doi: /j X x Bonnefoy-Claudet, S., Baize, S., Bonilla, L.F., Berge-Thierry, C., Pasten, C., Campos, J., Volant, P., and Verdugo, R. (2009) Site effect evaluation in the basin of Santiago de Chile using ambient noise measurements. Geophys. J. Int., 176(3), Boore, D. and Boomer, J. (2005) Processing of strong-motion accelerograms: needs, options and consequences. Soil Dynamics and Earthquake Engineering, 25, Celebi, M. (1987) Processed Chile earthquake records of 3 March 1985 and aftershocks. USGS open file report. Comte, D., A. Eisenberg, E. Lorca, M. Pardo, L. Ponce, R., Saragoni, S. K. Singk and Suarez, G. (1986) The great 1985 Central Chile earthquake: A repeat of previous great earthquakes in the region?. Science, 299, Konno, K. and Ohmachi, T. (1998) Ground-motion characteristics estimated from spectral ratio between horizontal and vertical components of microtremor, Bull. Seism. Soc. Am., 88, Lermo, J. and Chavez-Garcia, F. (1993) Site effect evaluation using spectral ratios with only one station, Bull. Seism. Soc. Am., 83, Lermo, J. and Chávez-García. (1994) Are microtremors useful in site response evaluation?. Bull. Seism. Soc. Am., 84(5), Lomnitz, C. (1971) Major earthquakes and tsunamis in Chile during the period 1535 to Geolg. Rudschau., 59, Lomnitz, C. (2004) Major Earthquakes of Chile: A Historical Survey, Seism. Res. Let., 75, Nakamura, Y. (1989) A method for dynamic characteristics estimation of subsurface using microtremor on the ground surface. Quarterly report of railway technical research institute (RTRI), 30(1), Nakamura, Y. (2000) Clear identification of fundamental idea of Nakamura s technique and its applications. Proc 12th Word Conference Earthquake Engineering. Paper N Nogoshi, M. and Igarashi, T. (1970) On the propagation characteristics of microtremor. J. Seism. Soc. Japan., 23, (in Japanese with English abstract). Nogoshi, M. and Igarashi, T. (1971) On the amplitude characteristics of microtremor (Part 2), Jour. Seism. Soc. Japan, 24, pp (in Japanese with English abstract). Pasten, C. (2007) Respuesta sísmica de la cuenca de Santiago. Tesis de Magíster (Inédito). Departamento de Ingeniería Civil, Universidad de Chile. Pilz, M., Parolai, S., Leyton, F., Campos, J., and Zschau, J. (2009) A comparison of site response techniques using earthquake data and ambient seismic noise analysis in the large urban areas of Santiago de Chile. Geophys. J. Int., 178, Ruiz, S. and Saragoni, G. R. (2004) Medida experimental de la respuesta dinámica 1D de suelos empleando los autocorrelogramas de los acelerogramas del terremoto Ms = 7.8 de Chile de Proc. 5to Congreso Chileno de Geotecnia,. Ruiz, S. and Saragoni, G. R. (2009). Free Vibration of Soils during Large Earthquakes. Soil Dynamic and Earthquake Engineering, Vol 29, pp Ruiz, S. and Leyton, F. (2010) Comparison of fundamental soil frequency using strong motion data of 1985 Central Chile earthquake (ms=7.8) and H/V microvibrations measurements. Soil Dynamic and Earthquake Engineering, under review. Ruiz, S., Kausel, E., Campos, J., Saragoni, G. R., and Madariaga, R. (2010). Identification of Earthquake Asperities Along Chilean Subduction Zone Using Strong Motion. Pure Appl. Geophy, doi: /s x. Saragoni, G.R., Gonzalez, P., and Fresard, M. (1986). Análisis de los acelerogramas del terremoto del 3 de marzo de El sismo del 3 de marzo de 1985-Chile. Editor J. Monge, University of Chile, Acero Comercial S.A. Ed. Universitaria,. (In Spanish).

11 Saragoni, G. R., y Ruiz, S. (2004) Modelos Elásticos 1D para la respuesta dinámica de suelos obtenidas del análisis de acelerogramas de sismos de Chile Central. Proc. 5to Congreso Chileno de Geotécnia, Universidad de Chile,. Wessel, P. and W.H.F. Smith (1991) Free software helps map and display data. EOS Trans. AGU, 72, 441.

TWO PEAKS RESPONSE SPECTRA (2PRS) FOR SUBDUCTION EARTHQUAKES CONSIDERING SOIL AND SOURCE EFFECTS

TWO PEAKS RESPONSE SPECTRA (2PRS) FOR SUBDUCTION EARTHQUAKES CONSIDERING SOIL AND SOURCE EFFECTS The 4 th October -7, 8, Beijing, China TWO PEAKS RESPONSE SPECTRA (PRS) FOR SUBDUCTION EARTHQUAKES CONSIDERING SOIL AND SOURCE EFFECTS ABSTRACT : Sergio Ruiz and G. Rodolfo Saragoni Civil Engineer. Graduate