Estimation of S-Wave Velocity Profiles at Lima City, Peru Using Microtremor Arrays Paper: Estimation of S-Wave Velocity Profiles at Lima City, Peru Using Microtremor Arrays Selene Quispe∗, Kosuke Chimoto∗, Hiroaki Yamanaka∗, Hernando Tavera∗∗, Fernando Lazares∗∗∗, and Zenon Aguilar∗∗∗ ∗Department of Environmental Science and Technology, Tokyo Institute of Technology 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226-8502, Japan E-mail: selene.q.aa@m.titech.ac.jp ∗∗Geophysical Institute of Peru (IGP), Lima, Peru ∗∗∗Japan Peru Center for Earthquake Engineering and Disaster Mitigation (CISMID) Faculty of Civil Engineering, National University of Engineering, Lima, Peru [Received July 4, 2014; accepted September 20, 2014] Microtremor exploration was performed around seis- mic recording stations at five sites in Lima city, Peru in order to know the site amplification at these sites. The Spatial Autocorrelation (SPAC) method was applied to determine the observed phase velocity dispersion curve, which was subsequently inverted in order to es- timate the 1-D S-wave velocity structure. From these results, the theoretical amplification factor was calcu- lated to evaluate the site effect at each site. S-wave ve- locity profiles at alluvial gravel sites have S-wave veloc- ities ranging from∼500 to∼1500m/s which gradually increase with depth, while Vs profiles at sites located on fine alluvial material such as sand and silt have S- wave velocities that vary between∼200 and∼500 m/s. The site responses of all Vs profiles show relatively high amplification levels at frequencies larger than 3 Hz. The average transfer function was calculated to make a comparison with values within the existing am- plification map of Lima city. These calculations agreed with the proposed site amplification ranges. Keywords: Lima city, microtremors, S-wave velocity profiles, site response, AVs10 1. Introduction The area used for this study, Lima city in, Peru, is sit- uated in a seismically active region, due to the subduc- tion of the Nasca Plate beneath the South American Plate. Numerous earthquakes (Mw > 8.0) have struck the region in the past. The city suffered substantial damage during the last strong ground motion (which took place on Oc- tober 3, 1974 with a moment magnitude of 8.1), most of which was related to local subsurface conditions of the soil [1]. In relation to this, it was considered necessary to undertake a comprehensive study of Lima’s soil dynamic. At present, Lima counts on seismic microzonation maps [2] and an existing soil amplification map [3]. These maps based on geological and geotechnical information, and the H/V peak period from microtremor measure- ments show that Lima city mainly overlies Quaternary al- luvial deposits. However, a dynamic analysis of these de- posits is only in its preliminary stages due to a lack of in- formation relating to the S-wave velocity structure [3, 4], which is a key parameter for use in assessing the potential for sediment amplification. The use of microtremor array measurements to de- termine S-wave velocity distributions in soil deposits is being used more extensive globally because of its ad- vantages of being affordable, non time-consuming, and possible to conduct in urban and crowded areas such as within Lima city. In this study, the spatial autocorrela- tion coefficient method (SPACmethod) introduced by Aki (1957) [5] was used to extract S-wave velocity informa- tion from microtremor measurements. The aim of the present paper is to determine the S-wave velocity structure at five selected sites and to analyze the site effect of each site. It is noteworthy that one of the study sites is located in the La Molina district, an area that has suffered extensive damage in previous earthquakes in relation to the local subsurface conditions [1, 4, 6, 7]. Also included in the study is a site located on the outskirts of Lima city (the Anco´n district), which was chosen because of a lack of existing data. 2. Geological and Geomorphological Aspects Lima city is the capital of Peru, and is the country’s largest and most populated city. Fig. 1a shows the coastal location of Lima city within Peru. The geologic units of Lima city are composed of sedimentary and intrusive rocks from Cretaceous age, in addition to unconsolidated sediments, as shown in Fig. 1b. Cretaceous deposits from the Puente Piedra Group appear to the northwest of the city, while deposits from the Morro Solar Group are ex- posed in the southwest. In addition, rocks from the Casma Group outcrop in many parts of the city, and intrusive rocks are found exposed mainly in the eastern part of the Journal of Disaster ResearchVol.9 No.6, 2014 931 Quispe, S. et al. Quaternary deposit Aeolian deposit Alluvial deposit Marine deposit Cretaceous sedimentary deposits Casma Group: Quilmaná Huarangal Pamplona Formation Morro Solar Group: Marcavilca formation Herradura formation Puente Piedra Group: Ancon Cerro Blanco Formation Ventanilla Formation Cretaceous intrusive rock Santa Rosa Granodiorite Santa Rosa Tonalite-diorite Santa Rosa Tonalite-granodiorite Patap km ANC PUCP CER MAY RIN (b) (a) Lima Pacific Ocean Ecuador Colombia Brazil Bo liv ia Chile Peru Earthquake station Microtremor measurements River Fig. 1. Maps of microtremor exploration sites. (a) Location of Lima city within Per. (b) Geological map of Lima city [9] with locations of earthquake stations (squares) and recorded microtremor measurements (triangles). study area. The geological map shows that the main part of the city lies on Quaternary deposits represented by alluvial, ma- rine, and aeolian materials. The distribution of the allu- vial materials is very wide, and they have been formed by the rapidly flowing Rı´mac and Chillo´n rivers as seen in Fig. 1b. The alluvial deposits extend from the ground’s surface to the rock, and consist mainly of medium dense to very dense coarse gravel and sand with cobbles [1], known lo- cally as Lima conglomerate. The thickness of this mate- rial (to the base rock) is reported to be over 100m [1, 6, 8]. 3. Field Measurements 3.1. Site Selection For this study, target sites for microtremor measure- ments (triangles) were selected in relation to the loca- tion of the seismic recording stations (squares) as shown in Fig. 1b. The strong motion network at Lima city is currently composed of more than 15 stations, and oper- ated by two institutions: the Japan–Peru Center for Earth- quake Engineering, Research, and Disaster Mitigation (CISMID) and the Geophysical Institute of Peru (IGP). Calderon et al. (2012) [4] conducted microtremor explo- rations around a number of recording sites maintained by CISMID, and this study intends to estimate S-wave veloc- ity (Vs) structures for the IGP stations. Fig. 1b shows the locations of the selected sites: PUCP, CER, MAY, RIN and ANC; all of them located on Quaternary alluvial de- posits. Table 1 shows information related to each site’s name, location, geographical coordinates, and geology. 3.2. Array Configuration A circular array configuration was applied for this ob- servation. Fig. 2 shows the schematic layout of the in- stallation. Seven 3-component sensors were placed on the ground’s surface: six were distributed at the vertices of two equilateral triangles inscribed within two circles of different radii, and one was placed at the center [10]. For each site, the size of the array (side length of equilat- eral triangle) varied from small to large. Small arrays (the smallest was 1.5 m) gave information of the near-surface layers, while the large ones characterized the deeper soil layers. Although large arrays with side lengths larger than 300 m were conducted for all the sites, the coherence was so low at a low frequency that it was not possible to use data from large arrays in the analysis for most sites. The size and number of deployed circular arrays that defined the observed phase velocity dispersion curve for each site are presented in Table 1. The test equipment used in this exploration was a GPL- 6A3P portable recording system designed by the Mitu- 932 Journal of Disaster ResearchVol.9 No.6, 2014 Estimation of S-Wave Velocity Profiles at Lima City, Peru Using Microtremor Arrays Table 1. Array information. Site ID Location(District) Latitude (deg) Longitude (deg) Geology Array Min∗ (m) Array Max∗ (m) Number of deployed arrays PUCP San Miguel -12.0734 -77.0796 Quatemary Alluvial 1.5 346.4 4 CER San Borja -12.1040 -76.9992 Quatemary Alluvial 1.5 48.0 3 MAY Ate -12.0549 -76.9441 Quatemary Alluvial 1.5 173.2 4 RIN La Molina -12.0873 -76.9240 Quatemary Alluvial 1.5 48.0 3 ANC Anco´n -11.7767 -77.1510 Quatemary Alluvial 1.5 173.2 4 ∗Side length of equilateral triangle Sensor Fig. 2. Geometry used in the microtremor array. toyo Corporation, which has a flat response in the fre- quency range between 0.25 to 25 Hz for estimating the phase velocity [10]. Each recording lasted between 10 and 60 min, and data were recorded at a sampling rate of 100 samples per second. 4. Estimation of the Dispersion Curve The SPatial Autocorrelation Coefficient (SPAC) method was applied to define the observed dispersion curve of Rayleigh waves from the array data. This technique uses SPAC coefficients in the calculation of phase velocity at different frequency ranges. From the array configuration, the SPAC coefficient was computed using the cross-spectrum between records from the vertical components of the sensors at an equal interstation separation, and with different azimuths. Phase velocity at a certain frequency was estimated by fitting the SPAC coefficients to the Bessel function. Details relating to SPAC analysis can be found in literature [5, 11]. Using the assumption that Rayleighwaves mainly dom- inate vertical motion, vertical records from each sen- sor were analyzed in the processing of data. The mi- crotremor recording data were divided into time segments with lengths of 81.92 sec, and time segments clearly con- taminated by noise were removed. The SPAC coeffi- cient obtained from every segment was averaged to ob- tain the phase velocity for the frequency range of inter- est. Fig. 3 illustrates the SPAC coefficients for one of the microtremor arrays recorded at the RIN site. Because in this study, a seven-sensor configuration was used in the measurements, Fig. 3 shows five SPAC coefficients that -0.5 0 0.5 1 1 10 S P A C c oe ffi ci en t Frequency (Hz) Sensor separation distance (m) 12.0 10.4 6.9 6.0 3.5 RIN Fig. 3. SPAC coefficients as a function of frequency for different sensor separation distances at the RIN site (with a maximum side length of 12 m). CER Fig. 4. Observed dispersion curves obtained from mi- crotremor data using the SPAC technique. correspond to five combinations of sensor separation dis- tances. These observed SPAC coefficients, as a function of frequency for a fixed distance, model the Bessel func- tion at frequencies higher than 9 Hz. Figure 4 shows the observed dispersion curves at the sites. For all sites, the phase velocities were estimated in the frequency range from 4 to 30 Hz, except for at the PUCP site, which had the widest frequency range until 1 Hz (due to the contribution from large arrays with a maximum side length of 346.4 m, exploring deeper the structure of the soil). In terms of the velocity, Fig. 4 Journal of Disaster ResearchVol.9 No.6, 2014 933 Quispe, S. et al. shows that the phase velocity values vary between 200 and 2000 m/s. The PUCP and MAY sites reached the highest velocity values of∼2000m/s at frequencies of∼1 Hz and ∼4 Hz, respectively. The high velocity layer (∼2000 m/s) at the MAY site suggests a shallower basement depth than that at the PUCP site. 5. Estimation of VS Profile SPAC analysis provides the dispersion curve of Rayleigh waves, and this is subsequently inverted (using the Genetic Simulated Annealing Algorithm technique) to determine a 1-D S-wave velocity model at each of the ex- amined sites. The inversion technique was introduced by Yamanaka (2007) [12], and it searches to fit (as much as possible) the observed model,U0 ( fi), with the calculated values of phase velocity for fundamental mode Rayleigh waves,UC ( fi), by using the misfit function /0 j defined as: /0 j = 1 N N ∑ i=0 [U0 ( fi)−UC ( fi)]2 . . . . . . . (1) where N and fi represent the number of the observed data and frequency, respectively. In the calculation of the final optimal Vs model, 10 inversions with 100 genera- tions were conducted using different random numbers. In doing so, good models with a smaller amount of misfit were more likely to survive in the next generation, and poor models were replaced by newly generated ones. The unknown parameters to be determined in the inversion were Vs and thickness. P-wave and density of the layers were fixed. P-wave velocity value was calculated using the equation proposed by Kitsunezaki et al. (1990) [13] that correlates Vs and Vp values, as previously used by Calderon et al. (2012) [4] in the same study area. Den- sity values were set from 1.8 to 2.5 g/cm3, depending on the soil type. The fundamental mode of Rayleigh waves was assumed in the inversion, as well asVs increases with depth. Table 2 shows an example of the search limits at the RIN site. Figure 5 shows the extent that the calculated dispersion curve (solid line) fits the observed one (open circles) for all the selected sites. All models can sufficiently explain the observed phase velocity in the entire frequency range. The inverted 1-D Vs profiles for all the sites are shown in Fig. 6 and Table 3. The deepest profile was obtained for the PUCP array, with a depth over 280 m (Fig. 6a) over the basement and an S-wave velocity of∼2500m/s, while the CER site (where the bottom layer has a lower veloc- ity than the basement) only reached a depth of ∼50 m (Fig. 6b). The top layers at the PUCP, CER, and MAY sites show S-wave velocities of ∼400 m/s, whereas the RIN and ANC sites show S-wave velocities of ∼200 m/s (Fig. 6b); this is related to the phase velocity at the upper- limit of frequency. The phase velocities for PUCP, CER, and MAY at 30 Hz are relatively larger than those at the RIN and ANC sites (as show in Fig. 5). All the Vs profiles estimated in this study detected the engineering bedrock, withVs larger than 500m/s. A further description of the S- Table 2. Search limits used for the determination of the optimal Vs profile at the RIN site. Search limits Layer Vs (km/s) Thickness (m) Density (g/cm3) 1 [0.20-0.30] [1-15] 1.8 2 [0.30-0.45] [1-15] 1.9 3 [0.45-0.60] [1-15] 2.0 4 [0.60-0.95] [25-55] 2.1 5 [0.95-1.45] - 2.2 wave velocity structure is discussed in the following sec- tion. To support the reliability of the results obtained from the inversion, the horizontal to vertical (H/V ) spectral ra- tio calculated from the observed microtremor data (solid line) was compared with the theoretical ellipticity of the fundamental-mode Rayleigh wave from the inverted 1-D soil profile (broken line), as depicted in Fig. 7. The ob- served H/V spectrum was estimated from the recording data of the sensor placed at the center of the array con- figuration, which was smoothed using a Parzen window with a 0.05 Hz bandwidth. The comparison between the observed H/V and the computed ellipticity shows good agreement in the frequency range within 1 and 10 Hz. The dominant peaks observed in the spectral ratios of observed microtremor data were well modeled by the computed el- lipticities of the Rayleigh waves. 6. Discussion 6.1. Geotechnical Description of S-wave Velocity Structure All the selected sites in the present study are located on alluvial Quaternary deposits (Fig. 1b), but the subsurface condition for each site differs. Lima conglomerate is the predominant material over Lima city [1, 2, 6]. Overlying the conglomerate, shallow layers of unconsolidated mate- rial such as sand, silt, or clay are found, and their thick- nesses range from ∼0.5 m to more than 30 m, depend- ing on the location [1, 2, 6]. CISMID (2005) [6] proposed a soil distribution map of Lima city in order to obtain a more accurate picture of the distribution and properties of the various subsurface soils in Lima. According to this soil classification map [6], the PUCP and CER sites are located on alluvial gravel. The Vs pro- files at the two sites show that the Lima conglomerate ex- tends from near the ground’s surface, with aVs larger than 500 m/s (Fig. 6). The PUCP site proves that Lima con- glomerate, especially in the central part of the city, has a thickness of about 200 m, as previously reported by CIS- MID (2005) using water well records [6]. The S-wave ve- locity of the conglomerate increases gradually with depth from ∼500 to ∼1500 m/s, as shown in Fig. 6a (PUCP Vs profile). In the eastern part of the city, sand and silt deposits 934 Journal of Disaster ResearchVol.9 No.6, 2014 Estimation of S-Wave Velocity Profiles at Lima City, Peru Using Microtremor Arrays 0 400 800 1200 1600 2000 1 10 PUCP P ha se V el . ( m /s ) Frequency (Hz) 0 400 800 1200 1600 2000 1 10 CER Ph as e Ve l. (m /s ) Frequency (Hz) 0 400 800 1200 1600 2000 1 10 MAY P ha se V el . ( m /s ) Frequency (Hz) 0 400 800 1200 1600 2000 1 10 RIN Ph as e Ve l. (m /s ) Frequency (Hz) 0 400 800 1200 1600 2000 1 10 ANC P ha se V el . ( m /s ) Frequency (Hz) Fig. 5. Comparison between the calculated dispersion curves (solid line) for inverted models and the observed ones (circles) for all sites. 0 10 20 30 40 50 60 70 0 500 1000 1500 2000 D ep th (m ) Vs (m/s) 0 50 100 150 200 250 300 0 500 1000 1500 2000 2500 3000 D ep th (m ) Vs (m/s) (a) (b) Fig. 6. Estimated shear-wave velocity profiles. Table 3. Estimated S-wave velocity structures from array observations of microtremors. PUCP CER MAY Vs (m/s) Thickness (m) Vs (m/s) Thickness (m) Vs (m/s) Thickness (m) 362 7.7 364 6.6 474 10.4 596 9.9 695 7.3 888 22.6 884 28.5 911 31.2 1225 31.3 1038 153.8 1396 - 1518 47.9 1503 82.9 2492 - 2412 - RIN ANC Vs (m/s) Thickness (m) Vs (m/s) Thickness (m) 254 6.8 225 5.7 418 2.0 429 16.7 496 11.0 938 16.7 769 44.7 1468 40.2 1431 - 2477 - Journal of Disaster ResearchVol.9 No.6, 2014 935 Quispe, S. et al. 0.1 1 10 0.1 1 10 PUCP H /V s pe ct ra l r at io Frequency (Hz) 0.1 1 10 0.1 1 10 CER H /V s pe ct ra l r at io Frequency (Hz) 0.1 1 10 0.1 1 10 MAY H /V s pe ct ra l r at io Frequency (Hz) 0.1 1 10 0.1 1 10 RIN H /V s pe ct ra l r at io Frequency (Hz) 0.1 1 10 0.1 1 10 ANC H /V s pe ct ra l r at io Frequency (Hz) Fig. 7. Comparison of H/V spectra of microtremors (solid lines) with computed ellipticities of fundamental-mode Rayleigh waves based on the obtained S-wave velocity structure (broken lines). overlie the Lima conglomerate [6], and MAY and RIN array measurements were carried out on these materials. Our results show that this deposit has S-wave velocities ranging within ∼200 and ∼500 m/s (Fig. 6b), which is similar to that reported by Repetto et al. (1974) using a down-hole test [1]. The RIN site is located in the La Molina district, a place where a high concentration of damage has previously occurred during previous earth- quakes, due to the local subsurface conditions [1, 4, 7]. The Vs profile at RIN shows that the thickness of the sand and silt deposits is about 20 m, while the thickness in the Vs structure at MAY is about 10 m. Conglomerate is found underlying this unconsolidated material, with a Vs of be- tween ∼500 to ∼1500 m/s. The MAY profile detected shear wave velocities of ∼2500 m/s at a depth of over 120 m, and this high velocity layer (∼2500 m/s) was also detected in the PUCP profile at a depth of over 280 m (Fig. 6); it is considered that this material may correspond to bedrock. Calderon et al. (2012) [4] reported that the bedrock at Lima city has S-wave velocity values of the order of 3000 m/s. The ANC site is located on the outskirts of Lima city to the north (Fig. 1b). The soil distribution map of Lima city proposed by CISMID (2005) [6] provides scarce in- formation related to the area where the earthquake ob- servation station is located. Nonetheless, geological and geotechnical information indicates that layers of aeolian sand overlie alluvial gravel deposits [6]. The Vs profile at ANC shows the top layer has S-wave velocities between ∼200 m/s and ∼400 m/s, and this overlies thick high ve- locity layers (within ∼1000 m/s and ∼1500 m/s) and a very high velocity layer (∼2500 m/s). 1 10 0.1 1 10 S ite a m pl ifi ca tio n Frequency (Hz) Fig. 8. Theoretical amplification factor for all sites. 6.2. Site Amplification The 1-D Vs profiles obtained from microtremor data (Fig. 6) were used to calculate the site amplification fac- tor (defined as the ratio between the surface motion and the input motion from the bottom layer). In this study, the bottom layer had a shear-wave velocity about 1500 m/s since we detected this velocity layer at all the profiles as shown in Fig. 6. Figure 8 shows the theoretical amplification factor for each profile. Both PUCP and CER sites, which are lo- cated on alluvial gravel, show several peaks at frequen- cies above 3 Hz, corresponding to the site response of the Lima conglomerate. Using strong ground motion data, Quispe et al. (2012) [14] reported that this deposit suffers 936 Journal of Disaster ResearchVol.9 No.6, 2014 Estimation of S-Wave Velocity Profiles at Lima City, Peru Using Microtremor Arrays Table 4. Average transfer function AvTF . Site ID Location (District) AVs30 (m/s) AVs10 (m/s) AvTF Eq. (2) AvTF proposed by Sekiguchi et al. (2013) [3], according to the location of the sites PUCP San Miguel 577.9 397.9 1.1 1.10–1.15 CER San Borja 647.8 434.3 1.0 1.00–1.10 MAY Ate 681.6 474.0 1.0 1.00–1.10 RIN La Molina 447.8 294.3 1.2 1.15–1.20 ANC Anco´n 414.6 282.8 1.2 - from high amplification at frequencies larger than 4 Hz. The PUCP site shows a small bump between 1 and 2 Hz, which is also identified in the observedH/V spectral ratio of the microtremor data (Fig. 7). This peak corresponds to the contribution of the deep structure. The PUCP site is located in the center of the city, where the Lima con- glomerate is thicker than in other places, as previously mentioned. MAY and RIN sites lie on layers of sand and silt. Both sites show peaks at frequencies larger than 4 Hz (Fig. 8), which represents the resonance between the top layer and the alluvial deposit. The RIN site shows a higher am- plification than the MAY site due to its softer subsurface conditions. The site response at the RIN site also depicts a peak at a frequency of 3 Hz, which represents the first res- onant mode at this site. Such a peak was also detected by Stephenson et al. (2009) [7] when analyzing the velocity response spectra for earthquake data. The ANC site shows a prominent peak at a frequency between 4 and 5 Hz, which represents the first resonant mode related to the strong velocity contrast between the top layers (Vs ranging from ∼200 and ∼400 m/s) and the high velocity layers (Vs ranging from ∼1000 m/s and ∼1500 m/s). 6.3. Amplification Map for Lima City Sekiguchi et al. (2013) [3] developed an amplification map for Lima city based on the correlation between the average S-wave velocity for the top 10 m of soils (AVs10) and the Average Transfer Function (AvTF) of S-waves. The proposed correlation is shown in the following equa- tion: AvTF = 122 AVs10 +0.76 . . . . . . . . . (2) Table 4 shows the value of (AVs10) and the calcu- lated (AvTF) for each site using Eq. (2). The calculated (AvTF) for all the sites (except the ANC site) falls into the given range proposed by Sekiguchi et al. (2013) [3] as shown in Table 4. Although the amplification map does not include the area where the ANC site is located, (AVs10) and (AvTF) were calculated in order to use the information in future updating of the map. The average shear-wave velocities for the first 30 m (AVs30) in the profile are also shown in Table 4, as this value is adopted as an international standard for soil classification. 7. Conclusions Microtremor measurements were carried out in order to estimate the S-wave velocity profiles at five selected sites, based on the location of strong motion stations. The PUCP and CER sites are located on Lima conglomerate, which extends from the ground’s surface to the rock be- neath. The Vs profiles show that this deposit has S-wave velocities ranging from∼500 m/s to∼1500 m/s, as previ- ously reported [1, 4]. At the central part of Lima city, the thickness of the Lima conglomerate is larger than 200 m, as proved by the Vs profile at PUCP (the site response of Lima conglomerate is characterized by a high amplifica- tion for frequencies larger than 3 Hz). MAY and RIN sites are situated on sand and silt deposits overlying the Lima conglomerate, and the Vs of this unconsolidated deposit is between ∼200 and ∼500 m/s. The surface condition at RIN site is softer than at the MAY site, and therefore the site response at the RIN site is higher in terms of the amplification. The RIN site is located in the La Molina district, where excessive earthquake damage has previ- ously occurred [1, 4, 7]. The Vs profile of RIN shows a distribution of the soil layers to a depth of up to ∼70 m. In this work, the Vs profile of ANC (located on the city’s outskirts) was estimated up to ∼80 m depth. The site re- sponse here shows a predominant peak at 4 Hz, due to the strong velocity contrast between the top layers and the high velocity layers. The S-wave velocity for the top 10 m of soils (AVs10) and the Average Transfer Function (AvTF) [3] were de- termined to compare with the Lima amplification map [3], and the calculated AvTF shows agreement with the previ- ously proposed AvTF ranges. Acknowledgements We are very grateful to the assistant researchers from CISMID who helped us conduct the array measurements. We would also like to express our sincere gratitude to Prof. Santa Cruz and Prof. Villa Garcia from Pontifical Catholic University of Peru for their cooperation in gaining access to this private university, and for enabling the measurements of microtremors. This study was supported by the SATREPS project “Enhancement of Earthquake and Tsunami Mitigation Technology in Peru.” Journal of Disaster ResearchVol.9 No.6, 2014 937 Quispe, S. et al. References: [1] P. Repetto, I. Arango, and H. B. Seed, “Influence of site char- acteristics on building damage during the October 3, 1974 Lima earthquake,” Report-Earthquake Engineering Research Center, Col- lege of Engineering, University of California, Berkeley, California, EERC 80/41, 1980. [2] Z. Aguilar, “Seismic Microzonation of Lima,” Japan-Peru Work- shop on Earthquake Disaster Mitigation, Japan-Peru Center for Earthquake Engineering and Disaster Mitigation (CISMID), Fac- ulty of Civil Engineering, National University of Engineering, Lima, Peru, 2005. [3] T. Sekiguchi, D. Calderon, S. Nakai, Z. Aguilar, and F. Lazares, “Evaluation of Surface Soil Amplification for Wide areas in Lima, Peru,” Journal of Disaster Research, Vol.8 No.2, pp. 259-265, 2013. [4] D. Calderon, T. Sekiguchi, S. Nakai, Z. Aguilar, and F. Lazares, “Study of Soil Amplification Based on Microtremor and Seismic Records in Lima Peru,” Journal of Japan Association for Earthquake Engineering, Vol.12, No.2, pp. 1-20, 2012. [5] K. Aki, “Space and Time Spectra of Stationary Stochastic Waves, with Special Reference to Microtremors,” Bulletin Earthquake Re- search Institute Tokyo University, Vol.25, pp. 415-57, 1957. [6] CISMID, “Study of the Vulnerability and Seismic Risk in 42 dis- tricts of Lima and Callao,” Japan–Peru Center for Earthquake En- gineering Research and Disaster Mitigation, National University of Engineering, Lima, Peru, 2005 (in Spanish). [7] W. R. Stephenson, R. A. Benites, and P. N. Davenport, “Localized coherent response of the La Molina basin (Lima, Peru) to earth- quakes, and future approaches suggested by Parkway basin (New Zealand) experience,” Solid Dynamics and Earthquake Engineer- ing, Vol.29, No.10, pp. 1347-1357, 2009. [8] J. P. Le Roux, C. Tavares, and F. Alayza, “Sedimentology of the Rı´mac-Chillo´n alluvial fan at Lima, Peru, as related to Plio- Pleistocene sea-level changes, glacial cycles and tectonics,” Journal of South American Earth Sciences, Vol.13, pp. 499-510, 2000. [9] A. Martı´nez and F. Porturas, “Planos Geote´cnicos para Lima, Peru´, Ana´lisis y Visio´n en Ingenierı´a Sı´smica,” Pontificia Universidad Cato´lica del Peru´, Lima, Peru´, 1975 (in Spanish). [10] R. N. Grutas and H. Yamanaka, “Shallow shear-wave velocity pro- files and site response characteristics from microtremor array mea- surements in Metro Manila, the Philippines,” Exploration Geo- physics, CSIRO Publishing, 2012. [11] H. Okada, “The microtremor survey method,” Geophysical Mono- graph Series No. 12, Society of Exploration Geophysicists, 2003. [12] H. Yamanaka, “Inversion of surface-wave phase velocity using hy- brid heuristic search method,” Butsuri Tansa (Geophysical explo- ration), Vol.60, pp. 265-75, 2007 (in Japanese with English ab- stract). [13] C. Kitsunezaki, Y. Kobayashi, T. Ikawa, M. Horike, T. Saito, T. Kurota, K. Yamabe, and K. Okuzumi, “Estimation of P- and S-wave velocities in deep soil deposits for evaluating ground vibrations in earthquake,” Journal of the Japan Society for Natural Disaster Sci- ence, Vol.9, No.3, pp. 1-17, 1990. [14] S. Quispe, H. Yamanaka, Z. Aguilar, F. Lazares, and H. Tavera, “Preliminary Analysis for Evaluation of Local Site Effects in Lima city, Peru from GroundMotion Data by Using the Spectral Inversion Method,” Journal of Disaster Research, Vol.8 No.2, pp. 243-251, 2013. Name: Selene Quispe Affiliation: Ph.D. Candidate, Department of Environmen- tal Science and Technology, Tokyo Institute of Technology Address: 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226-8502, Japan Brief Career: 2007 State Engineer, National University of Engineering, Peru 2010 State Engineer, National Graduate Institute for Policy Studies, Japan 2012 Student in Doctoral Course, Tokyo Institute of Technology Selected Publications: • S. Quispe, H. Yamanaka, Z. Aguilar, F. Lazares, and H. Tavera, “Preliminary Analysis for Evaluation of Local Site Effects in Lima city, Peru from Ground Motion Data by Using the Spectral Inversion Method,” Journal of Disaster Research, Vol.8 No.2, pp. 243-251, 2013. 938 Journal of Disaster ResearchVol.9 No.6, 2014 View publication stats