The 19 September 2017 M 7.1 Puebla-Morelos Earthquake: Spectral Ratios Confirm Mexico City Zoning

One important element of understanding basin response to strong shaking is the analysis of spectral ratios, which may provide information about the dominant frequency of ground motion at specific locations. Spectral ratios computed from accelerations
of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Similar Documents
  The 19 September 2017 M 7.1 Puebla-Morelos Earthquake:Spectral Ratios Confirm Mexico City Zoning by Mehmet Çelebi, Valerie J. Sahakian, * Diego Melgar, and Luis Quintanar  Abstract   One important element of understanding basin response to strong shak-ing is the analysis of spectral ratios, which may provide information about the dom-inant frequency of ground motion at specific locations. Spectral ratios computed from accelerations recorded by strong-motion stations in Mexico City during the main-shock of the 19 September 2017  M  7.1 Puebla-Morelos earthquake reveal predomi-nate periods consistent with those mapped in the 2004 Mexican seismic design code.Furthermore, the predominant periods thus computed validate those studies usingmainshock and aftershock recordings of the handful strong-motion stations that re-corded the 19 September 1985  M  8.1 Michoacán earthquake. Even though the number of stations in each of the zones (zones I, II, IIIa, b, c, and d) is not the same, they stillallow confirmation of site frequencies (periods) attributable to the specific zones (par-ticularly those in zones IIIa, b, c, and d). Spectral ratios are computed with two differ-ent methods: (1) horizontal to horizontal (H/H) ratio of smoothed amplitude spectrum of a horizontal channel in direction X of a station with respect to the smoothed am-plitude spectrum of the horizontal channel in the same X direction of a reference stiff soil (or rock) station and (2) horizontal to vertical (H/V) ratio (or also known as theNakamura method) of both horizontal (H) and vertical (V) channels of the same sta-tion. We show a comparison of the identified frequencies (periods) derived by bothmethods and find they are very similar and in good agreement with those indicated inthe zoning maps of Mexico City in the 2004 seismic design code.Introduction On 19 September 2017, the  M  7.1 Puebla-Morelosearthquake occurred at 18:14:40 GMT (13:14:40 localtime) at epicentral coordinates: latitude 18.40° N and lon-gitude 98.72° Wand depth of 57 km (The National Seismo-logical Service of Mexico [SSN]). The U.S. GeologicalSurvey (USGS) gave the epicentral coordinates as18.5838° N and 98.3993° W and depth as 51 km (see Data and Resources). Geotechnical Extreme Events Reconnais-sance (GEER) describes the earthquake as occurring  “ in a complex region of normal and reverse faults with a regionaltectonic mechanism associated with the subduction of theCocos plate under the North American plate. The epicenter was located 12 km southeast of the city of Axochiapan inthe State of Morelos. As expected, there was no surface ex-pression of the fault rupture reported by any of the recon-naissance teams dispatched to the area  ”  (GeotechnicalExtreme Events Reconnaissance [GEER], 2017). A recent inversion study by Melgar   et al.  (2018) indicates a north-ward dip. Additional information on the seismological as-pects, tectonics, intensity, and ShakeMaps related to thisevent can be found at the USGS website.A Note on 1985 and 2017 EarthquakesItiswellknownthatpreviousearthquakesoccurringatfar distances from Mexico City have caused significant loss of lives and extensive damage within the city. One of the most well known of such events was the 19 September 1985  M  8.1Michoacán earthquake (Anderson  et al. , 1986; Çelebi  et al. ,1987a ,b; Stone  et al. , 1987). At an epicentral distance of  ∼ 400  km, this distant event caused 4287 casualties and5728 buildings to either collapse or sustain heavy damage(Çelebi  et al. , 1987a ,b; Stone  et al. , 1987). Relative locationsofthe1985and 2017 earthquake epicentersand epicentral dis-tances from Mexico City are shown in Figure 1. We note that although the epicenter of the 1985 earthquake was farther away from Mexico City than the 2017 Puebla-Morelos event,it was an order of magnitude larger.One of the main reasons that Mexico City sustain exten-sive damage from earthquakes that srcinate at far distances isthat it is densely built on a filled lakebed (GEER, 2017). *Also at Department of Earth Sciences, University of Oregon, Eugene,Oregon 97403. BSSA Early Edition / 1 Bulletin of the Seismological Society of America, Vol. XX, No. XX, pp.  – ,  –  2018, doi: 10.1785/0120180100 Downloaded from by University of Oregon user on 18 October 2018  Hence,seismic design codesinMexico recognize the site-spe-cific zonation issues in Mexico City. The design codes haveconsidered three zones that realistically represent lakebedareas as riskier. For example, at the time of the 1985 earth-quake, the seismic zoning map used in design codes consistedof only three zones (hills [now zone I], transition [now zoneII], and lake zone [now zone III]) (Fig. 2) (Çelebi  et al. ,1987a ).Thisolderzonationmapdepictsnotonlylimitednum-ber of the current key strong-motion stations but also includeslocations of temporary stations established in 1985 earth-quake-related aftershock studies (Çelebi  et al. , 1987a ) that are not repeated herein.After the 1985 earthquake, several research groupsfocused on analyzing the seismic records, local geologicalstructure, possible models, influence of Trans-Mexican vol-canic belt, and generation and propagation of surface waves,as well as the role of the path between the source and thevalley. A comprehensive literature review of this researchis described in the work of  Flores-Estrella   et al.  (2007).Even though the amplification of seismic waves in theValley of Mexico has been widely explained by local andregional soil properties (Sánchez-Sesma   et al. , 1988; Shapiro et al. , 1997) or by the differences of quality factors of propa-gation of seismic waves at certain frequency-band windows(Iida and Kawase, 2004), the long-lasting duration of ob-served long-period strong motions inside the lake zone is a subject still under discussion and actively researched. Exam-ples of such studies are by Chávez-García and Bard (1994)and Bard and Bouchon (1980). Kawase (2003), by analyzing 2D basin models, concluded that surface wavetrains gener-ated at basin edges sustain rapid decay as they propagate andsuggested discarding this mechanism as a possible explana-tion for the long seismic records. Singh and Ordaz (1993)explained the long-lasting duration of seismic signals in thevalley as caused by regional-scale effects, such as the pres-ence of scatterers around the basin that produce multipathswithin the larger Valley of Mexico. Another explanation isthat the interaction between incident wavefields and localbasin conditions could produce elongation of the signalduration by means of coupling valley resonant frequencieswith dominant periods of seismic waves (Chávez-García andSalazar, 2002).More recently, Cruz-Atienza   et al.  (2016) simulatedwavefield propagation in the Valley of Mexico, showing that deep structure provides conditions for a dominance of waveovertones on ground motion in the lake zone and that thispropagation regime strongly contributes to the elongationof intense shaking at frequencies for which the largest am-plification is observed.In addition, significant studies of natural site period dis-tribution across Mexico City (after the 1985 earthquake) werecarried out by Lermo and Chávez-García (1993, 1994). The best-known result of these studies was a map of period con-tours in the Mexico basin. It is important to note that in thisstudy, we show that spectral ratios computed using accelera-tions recorded at multiple stations in Mexico City during the19 September 2017 earthquake allow identification of sitefrequencies (periods) that agree with those indicated by the2004 seismic design code zoning maps of Mexico City. Figure  1.  Relative locations of the 1985 and 2017 earthquakeswith respect to Mexico City. Coordinates (latitudes and longitudes)of the epicenters of both 1985 and 2017 events are also indicated.The red shaded areas are rupture regions (from  Mendoza and Hart-zell, 1989, and Melgar   et al. , 2018). Figure  2.  Zoning map of Mexico City that was in effect in 1985(Çelebi  et al. , 1987a ). Solid lines are major avenues, and thick solidlines are boundaries of zones. Stations included in the study aremarked with circles. 2  M. Çelebi, V. J. Sahakian, D. Melgar, and L. Quintanar  BSSA Early Edition Downloaded from by University of Oregon user on 18 October 2018  Naturally, during the past three decades since the 1985event, Mexico City seismic zonation maps have evolved. Thecurrent seismic code further divides the lake zone (now zoneIII) into four subzones. Figure 3a  shows the current seismiczoning map of Mexico City that has been in effect since 2004(Mexican Seismic Design Code, 2004). Four subzones (a, b,c, and d) on the lake zone (zone III) are depicted in this map.Figure 3b shows a colored zoning map modified from the2004 Mexican seismic design code. It includes some of thestations in different zones that recorded the 19 September 2017  M  7.1 earthquake. Station CUP5, used in this study, isin close proximity to Universidad Nacional Autonoma deMexico (UNAM) station that was the major reference stationused in studies conducted after the 19 September 1985  M  8.1Michoacán earthquake (Çelebi  et al. , 1987a ).Recently, Arroyo  et al.  (2013) evaluated the change indominant periods in the lake zone of Mexico City. Theystud-ied the changes produced by ground subsidence through theuse of site amplification factors and proposed an updatedmap for inclusion in the revisions to the 2004 seismic designcode of Mexico. GEER (2017) cites this study and a UNAM-GEER study after the 2017 earthquake that confirmed thefindings of  Arroyo  et al.  (2013).Accelerations recorded in Mexico City during both the1985 and 2017 earthquakes were not large (mostly  < 0 : 25 g ).Table 1 shows a comparison of peak accelerations recordedin some of the stations that existed in 1985 as well as those Table 1 Representative Number of Peak Accelerations during the1985 and 2017 Mainshocks 1985 Peak Acceleration ( g )*2017 Peak Acceleration ( g ) † Zone North – South East  – West North – South East  – West  UNAM I 0.03 0.035 0.046 ‡ 0.055 ‡ CUP5 I  – –  0.05 0.06TACY I 0.03 0.03 0.06 0.06VIV III 0.049 0.024JC84 II  – –  0.22 0.21CH84 II  – –  0.15 0.23SI53 IIIa   – –  0.13 0.18SCT IIIb 0.098 0.168SCT2 IIIb 0.09 0.09Zone designations are according to the current zoning map in the 2004seismicdesign code. UNAM, Universidad NacionalAutonomade Mexico.*NIST report (Stone  et al. , 1987). † GEER report (2017). ‡ Center for Engineering Strong Motion Data (see Data and Resources). Figure  3.  (a) Seismic zoning map in effect in 2017 in Mexico City (adopted from the Mexican seismic design code of 2004). Comparedwth the 1985 zoning map, this map is far more detailed (zones denoted as zona; escala gráfica denotes the map scale). (b) 2017 zoning mapmodified from the 2004 Mexican seismic design code. Here, we use color to emphasize the different seismic zones and include some of thestations that recorded the 19 September 2017  M  7.1 earthquake. Station CUP5 used in this article is in close proximity to UniversidadNacional Autonoma de Mexico (UNAM) station that was the major reference station used in studies after the 19 September 1985 M  8.1 Michoacán earthquake (Çelebi  et al. , 1987a ). The inset in (b) depicts location of Mexico City within the Map of Mexico. The 19 September 2017 M 7.1 Puebla-Morelos Earthquake  3BSSA Early Edition Downloaded from by University of Oregon user on 18 October 2018  during the recent 2017 event. In 1985, because there were a limited number of strong-motion stations, some larger peak accelerations probably occurred but were not recorded.Nonetheless, considering the reported damages caused bythe 1985 or 1987 events, the level of peak accelerationswas not the main cause of damage. As strongly evidencedby previous studies of the 1985 event, the frequency content and resulting resonating amplifications of ground motionswere the main culprits leading to structural damage (Ander-son  et al. , 1986; Çelebi  et al. , 1987a ; Stone  et al. , 1987).Relevant to this study and for comparison later, ampli-fications of motions in the lake zone (the dubbed culprit during the 1985 event) are best displayed by the now well-known acceleration time history and corresponding responsespectra plots. In 1985, the strong-motion stations that existedwere UNAM and TAC in the hills zone; VIVin the transitionzone; and SCT, CDAO, and TLA in the lake zone. Anderson et al.  (1986) described the surficial geologic formations of these stations as very soft soil (clay) for SCTand CDAO bothin the lake zone, soft soil for VIVat the transition zone, andhard soil for TAC and rock (basalt) for UNAM, both in thehills zone. The other stations that recorded the 2017 event arelisted in Table 1.The 1985 recorded acceleration data from SCT and UNAM are integral in de-scribing the main culprit (frequency con-tent and amplification) responsible for the extensive structural damage in thisevent. As such, these stations best symbol-ize amplification in the Mexico City lakezone. Sample 1985 event acceleration re-cords and corresponding response spectra from SCT (1985 ID), UNAM, VIV, andCDA are shown in Figure 4 (adopted from Anderson  et al. , 1986). Design spectra from the code in effect in 1985 have beensuperimposed on the response spectra for comparison (Anderson  et al. , 1986). Thesetime histories clearly depict the amplifiedmotions (e.g., at SCT when compared toUNAM). The SCT response spectra inFigure 4 clearly define a 2-s (0.5-Hz) reso-nating site period (frequency) (Mena   et al. ,1985). It is fair to say that after the1985 event, the response spectrum of 1985SCT accelerations became a symbol of theamplification and resonating lakebed siteperiod. Furthermore, station SCT (of 1985butcurrentlyknowntobethesameasSCT2station) is also important because it wasclose to the (Secretaria de Comunicacionesy Transportes — Ministry of Telecommuni-cation and Transportation) building that was severely damaged in 1985. Upper floors of the building collapsed; two photosof the collapsed top floors of the SCT building are seen in Fig-ure 5. This is notable because these motions were recordedfrom an event that srcinated  ∼ 400  km away (Fig. 1). It hasbeen reported (but not confirmed) that the SCT building sus-tained extensive damage during the 2017 event as well and isunder consideration to be razed.After the 1985 mainshock, the USGS, in collaborationwith Mexican scientists, deployed temporary data loggers at existing strong-motion stations and additional new tempo-rary stations shown in Figure 2 (USA, SFO, and TLA). Inan earlier study, spectral ratios with respect to UNAM stationfor strong and weak motions are provided by Çelebi  et al. (1987a) and are not repeated herein. Those spectral ratiosconfirmed the significant amplification of motions and theresonating periods (frequencies) that are also seen in the re-sponse spectra of Figure 4 (Anderson  et al. , 1986). As statedbefore, these factors caused extensive damage and loss of lifeand property (Stone  et al. , 1987).Purpose of This StudyIn this study, we analyze the spectral ratios computedfrom strong-motion data recorded by several stations inMexico City during the 19 September 2017  M  7.1 Puebla- Figure  4.  (a) Acceleration time histories for the 1985 mainshock at different loca-tions in Mexico City compared with UNAM station. (b) Response spectra of the accel-erations also depict the site amplifications. Design response spectrum in effect in 1985 issuperimposed as dashed lines on each spectrum to facilitate comparison. (Adapted andredrawn with permission from the author, J. Anderson [9 April 2018] and reprinted withpermission from American Association for the Advancement of Science [AAAS].) Figure  5.  Photos ofthe damaged Secretaria de Comunicaciones yTransportes (SCT,Ministry of Telecommunications and Transportation) building after the 1985 earth-quake. It was severely damaged with several collapsed floors (see USGS PhotographicLibrary website in Data and Resources). 4  M. Çelebi, V. J. Sahakian, D. Melgar, and L. Quintanar  BSSA Early Edition Downloaded from by University of Oregon user on 18 October 2018  Morelos earthquake. With these data, we aim to identify thepredominant frequencies at select sites. When applicable, wecompare these frequencies and spectral ratios with observedpredominant frequencies from the 1985 earthquake. We alsocompare the frequencies and spectral ratios we compute withthe current site periods (frequencies) from the seismic zoningmap of Mexico City (Mexican Seismic Design Code, 2004).The scope of the article does not include tectonics, seismic-ity, earthquake damage reconnaissance, or assessment. 2017 Earthquake Data, Sources, and Organization Figure 6 shows a Google Earth map with the locations of the stations in Mexico City that recorded the 2017 Puebla-Morelos earthquake. These stations are operated by theInstitute of Engineering ’ s (IINGEN) strong ground motionnetwork of the seismic instrumentation unit within theEngineering Seismology Laboratory of the UNAM andCentro de Instrumentación y Registro Sísmico (CIRES).In this figure, we show only stations from which data wereavailable to us and are used in this study. It should be notedthat there may be more stations that recorded the earthquake.Figure 7 is a map showing predominant periods in thedifferentzones (digitized by authors using a similar map intheMexican seismic design code of 2004). We will use this mapto compare site periods from the code (obtained by interpo-lation of this map) with those site periods from strong-motiondata of the 2017 earthquake computed by spectral ratios.In Table 2, identification of stations, their coordinates,site classes, and particulars (srcinal record length, number of points [npts], and sampling period [ Δ t ], which is1/sampling rate). As shown in Table 2, there is not a commonrecord length or   Δ t  for all of the dataset. It was noted that even for a few stations, the record lengths varied. Therefore,both the lengths of the records and Δ t  were standardized for all data used. This was necessary to obtain an identical lengthfor each channel of each station and to achieve identical fre-quency vectors and sampling frequency ( Δ  f  ). Hence, wechose a record length of 260 s and  Δ t  of 0.01 s for all data.We accomplished this by carefully padding short (records < 260  s) or chopping excess parts of records longer than260 s. No portion of the actual earthquake time series data was truncated in the process.Because the srcinal lengths of the records are neither standard nor necessarily long, we did not discuss the nowwell-known long-duration records generated by deep basinand soft surface layers. However, such characteristics havebeen studied in detail by Kawase and Aki (1989) andFlores-Estrella   et al.  (2007), who presented a comprehensivereview of studies on the Mexico City basin and rightly statedthat the 1985 earthquake pointed to the importance of basin-site effects and that the 1985 earthquake records revealedlarge amplifications and long durations in the lake zone of Mexico City. Iida and Kawase (2004) reach similar conclu-sions by studying records from a borehole. Lermo andChávez-García (1993, 1994) showed that microtremors show good reliability compared with those using strong motions.Other notable studies on these and similar topics related tohorizontal to vertical (H/V) studies with microtremors andearthquakes include but are not limited to Kawase  et al. (2011, 2014), and Chávez-García   et al.  (1995). Figure  6.  Google Earth Map of Mexico City with strong-mo-tion stations used in this study, that recorded the earthquake on 19September 2017 (yellow circles with station numbers labeled). (In-set) Stations SCT, CUP5, and UNAM are highlighted in red. Thewhite box on main map shows the area in the inset. Figure  7.  Map showing predominant periods in the different zones (digitized by authors using the zoning map from the Mexicanseismic design code of 2004). This map is used in this article to com-pareperiodsfromthecodewiththosefromstrong-motiondataof2017earthquake.Stations includedinthisstudyareshownasyellowcircles. The 19 September 2017 M 7.1 Puebla-Morelos Earthquake  5BSSA Early Edition Downloaded from by University of Oregon user on 18 October 2018
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!