Earthquake Report: Puebla, Mexico Update #1

Well, the responses of people who are in the midst of a deadly disaster have been inspiring, bringing tears to my eyes often. Watching people searching and helping find survivors. This deadly earthquake brings pause to all who are paying attention. May we learn from this disaster with the hopes that others will suffer less from these lessons.

I have been discussing this earthquake with other experts, both online (i.e. the twitterverse, where most convo happens these days) and offline. Many of these experts are presenting their interpretations of this earthquake as it may help us learn about plate tectonics. While many of us are interested in learning these technical details, I can only hope that we seek a similar goal, to reduce future suffering. Plate tectonics is a young science and we have an ultra short observation period (given that the recurrence of earthquakes can be centuries to millenia, it may take centuries or more to fully understand these processes).

Here I present a review of the material that I have seen in the past day and how I interpret these data. The main focus of the poster is a comparison of ground shaking for three earthquakes. Also of interest is the ongoing discussion about how the 2019.09.08 M 8.1 Chiapas Earthquake and this M 7.1 Puebla Earthquake relate to each other. My initial interpretation holds, that the temporal relations between these earthquakes is coincidental (but we now have the analysis to support this interpretation!).

  • There are some reasons why these earthquakes are unrelated.
    1. They are too distant (static triggering is often limited to 1-2 fault lengths from the first earthquake).
    2. The Cocos plate (CP) changes shape between these two earthquakes, so it is complicated. The CP dips at a steep angle in the Chiapas region, while it dips at a shallow angle (about flat in places) further north. The Tehuantepec Ridge (TR) has an age offset and this may affect how the CP behaves differently on either side of the TR (mostly a fracture zone, but I need to look into this more, it may be thickened crust for some reason other than simply due to the fracture zone here).
    3. Dynamic triggering is when faults slip because they have increased stress as seismic waves travel through them. There is some work suggesting that these seismic waves can change the fluid pressures for a transient time period, possibly triggering earthquakes for a period after the seismic waves have already passed. The M 7.1 did not happen while the seismic waves were traveling following the M 8.1, so the M 7.1 is probably not due to dynamic triggering.
  • There is one major reason the ground shaking is amplified in the region of Mexico City. Prior to the arrival of the Spanish, the first peoples here lived at the shores of a large lake. They farmed on floating islands made of reeds and other material. Eventually the lake filled with sediment and turned into land, until the lake was gone. Given that Mexico City has the largest population of any city on Earth, as it was developed, the ground water was probably drained to facilitate the construction of large buildings (but I don’t know as much about this part of the history). I include a video about why water saturated sediments (i.e. sand and mud) can amplify seismic waves and intensify ground shaking.

Below is my interpretive poster for this earthquake

I plot the USGS seismicity from the past month, with color representing depth and diameter representing magnitude (see legend). I also include USGS epicenters from 1917-2017 for magnitudes M ≥ 7.0. I include the USGS fault plane solution for the 1985 earthquake. I also include the USGS moment tensor for the 2017.09.08 M 8.1 earthquake.

  • I placed a moment tensor / focal mechanism legend on the poster. There is more material from the USGS web sites about moment tensors and focal mechanisms (the beach ball symbols). Both moment tensors and focal mechanisms are solutions to seismologic data that reveal two possible interpretations for fault orientation and sense of motion. One must use other information, like the regional tectonics, to interpret which of the two possibilities is more likely.
  • I also include the shaking intensity contours on the map. These use the Modified Mercalli Intensity Scale (MMI; see the legend on the map). This is based upon a computer model estimate of ground motions, different from the “Did You Feel It?” estimate of ground motions that is actually based on real observations. The MMI is a qualitative measure of shaking intensity. More on the MMI scale can be found here and here. This is based upon a computer model estimate of ground motions, different from the “Did You Feel It?” estimate of ground motions that is actually based on real observations.
  • I include the slab contours plotted (Hayes et al., 2012), which are contours that represent the depth to the subduction zone fault. These are mostly based upon seismicity. The depths of the earthquakes have considerable error and do not all occur along the subduction zone faults, so these slab contours are simply the best estimate for the location of the fault.

    I include some inset figures in the poster.

  • In the upper left corner I include a generalized plate tectonic map from Wikimedia Creative Commons here.
  • In the upper right corner are two map insets. The upper one is a map that includes the USGS MMI contours for the M 7.1 earthquake and the lower one is the same for the 1985 M 8.0 earthquake. I have outlined the area of Mexico City with a white dashed line. I created polygons for the higher MMI contours in the region of Mexico City and colored them with respect to these MMI valaues. For the M 7.1 earthquake, MMI VI is shown in yellow and MMI VI.5 is shown in darker yellow. For the 1985 earthquake, MMI VI is shown, but MMI VI.5 is not modeled for Mexico City. The take away: the M 7.1 potentially caused greater ground shaking in the Mexico City region than did the 1985 earthquake.
  • In the lower left corner is a comparison of three ground motion model results from the Instituto De Ingenieria. From left to right are the 1985 M 8.0, the 2017 M 8.1, and the 2017 M 7.1 earthquakes. There are a variety of model results for these earthquakes, but I selected the results shown for a 1 second period (the period of seismic waves) because this is a frequency of seismic waves that multi story buildings can be sensitive to (see educational video about resonance below for more on this). Note that the largest ground motions are from the M 7.1 earthquake. The 1985 was quite deadly and damaging, with between 6,000 and 12,000 deaths. If this M 7.1 earthquake had occurred in 1985, there probably would have been even more damage and a higher casualty number.
  • Above these comparison maps is a figure prepared by Temblor here, a company that helps people learn and prepare to be more resilient given a variety of natural hazards. This figure is the result of numerical modeling of static coulomb stress changes in the lithosphere following the 2017 M 8.1 earthquake. This basically means that regions that are red have an increased stress (an increased likelihood for an earthquake) following the earthquake, while blue represents a lower stress, or likelihood. The change in stress are very very small compared to the overall stress on any tectonic fault. This means that an earthquake may be triggered from this change in stress ONLY IF the fault is already highly strained (i.e. that the fault is about ready to generate an earthquake within a short time period, like a day, month, or year or so). The take away: the M 8.1 earthquake did not increase the stress on faults in the region of the M 7.1 (Temblor suggests the amount of increased stress near the M 7.1 is about the amount of force it takes to snap one’s fingers.


  • Here is my original interpretive poster.


    •

  • As I mentioned the lake basin, here are some figures addressing that.
  • Here is a figure showing the thickness of the lake sediments here (Cruz-Atienza et al., 2016).


    • Topographic setting of Mexico City (MC) and the Valley of Mexico. Color scale corresponds to the basin thickness (i.e., the basin contact with the Oligocene volcanics of the Transmexican Volcanic Belt, TMVB). Stars show the epicenters for the vertical body forces applied at the free surface (green) and the magnitude 3.4 earthquake of December 1, 2014 (red). This figure has been created using the Generic Mapping Tools (GMT) Version 5.3.0, http://gmt.soest.hawaii.edu.

  • This is also from Cruz-Atienza et al. (2016) which shows their modeled seismic waves traveling through the basin.


    • Snapshots of the Green’s function for the vertical body force S6 (see Fig. 1) described by the inset time history with flat spectrum up to 1 Hz. Notice the topographic scattering, the generation and propagation of wave trains at different speeds within the basin, and their multiple diffractions. This figure has been created using the Matlab software Version R2016a, http://www.mathworks.com/.

  • Finally, here is a compilation of their model results showing how the lake basin sediments both amplify the ground motions (upper right panel) and increase their duration (lower right panel). Basically, the lake acts like a bowl of Jello.


    • (a,c) Comparison of average eigenfunctions for the 8 sources with standard deviation bars for both elastic (blue solid) and viscoelastic (red solid) simulations at two representative sites, P1 and P2, and different frequencies. Dashed lines show theoretical eigenfunctions for the vertical component of Rayleigh waves in the model of Figure A1a (Table A1) for the fundamental mode (blue) and the first (red) and second (green) overtones. Normalized peak vertical displacements observed in different boreholes (green dots in Fig. 1) are shown with black circles and error bars (after Shapiro et al., 2001). (b) Fourier spectral amplifications (geometric mean of both horizontal components) at 0.5 Hz with respect to the CUIG site (Fig. 1) averaged for the 8 sources. The black contour corresponds to the 2 s dominant-period. (d) Duration of the strong shaking phase for f < 1 Hz averaged for the 8 sources.

  • Here is an educational animation from IRIS that helps us learn about how different earth materials can lead to different amounts of amplification of seismic waves. Recall that Mexico City is underlain by lake sediments with varying amounts of water (groundwater) in the sediments.

  • Here is an educational video from IRIS that helps us learn about resonant frequency and how buildings can be susceptible to ground motions with particular periodicity, relative to the building size.

  • So, bringing this work as applied to this earthquake, Dr. Jascha Polet prepared this map that shows the outline of the lake and the locations of damaged and collapsed buildings. Note the correlation. Below the map, I include her tweet.


  • Here are some figures that show how the subduction zone varies across the Tehuantepec Ridge. More about this in my initial report, as well as in my reports for the M 8.1 earthquake.
  • This is a figure showing the location of the Tehuantepec Ridge (Quzman-Speziale and Zunia, 2015).


    • Tectonic framework of the Cocos plate convergent margin. Top- General view. Yellow arrows indicate direction and speed (in cm/yr) of plate convergence, calculated from the Euler poles given by DeMets et al. (2010) for CocoeNoam (first three arrows, from left to right), and CocoeCarb (last four arrows). Length of arrow is proportional to speed. Red arrow shows location of the 96 longitude. Box indicates location of lower panel. Bottom- Location of features and places mentioned in text. Triangles indicate volcanoes of the Central American Volcanic Arc (CAVA) with known Holocene eruption (Siebert and Simkin, 2002).

  • Here is another figure, showing seismicity for this region (Quzman-Speziale and Zunia, 2015).


    • Seismicity along the convergent margin. Top: Map view. Blue circles are shallow (z < 60 km) hypocenters; orange, intermediate-depth (60 < z < 100 km); yellow, deep (z > 100 km). Next three panels: Earthquakes as a function of longitude and magnitude for shallow (blue dots), intermediate (orange), and deep (yellow) hypocenters. Numbers indicate number of events on each convergent margin, with average magnitude in parenthesis. Gray line in this and subsequent figures mark the 96 deg longitude.

  • This shows the location of the cross sections. The cross sections show how the CP changes dip along strike (from north to south) (Quzman-Speziale and Zunia, 2015).


    • Location of hypocentral cross-sections. Hypocentral depths are keyed as in previous figures.

  • Here are the cross sections showing the seismicity associated with the downgoing CP (Quzman-Speziale and Zunia, 2015).


    • Hypocentral cross-sections. Depths are color-coded as in previous figures. Dashed lines indicate the 60-km and 100-km depths. Tick marks are at 100-km intervals, as shown on the sections. There is no vertical exaggeration and Earth’s curvature is taken into account. Number of sections refers to location on Fig. 3.

  • This figure shows thrust and normal earthquakes for three ranges of depth (Quzman-Speziale and Zunia, 2015).


    • Earthquake fault-plane solutions from CMT data. a. Shallow (z < 60 km), thrust-faulting mechanisms. b. Intermediate-depth (60 < z < 100 km) thrust-faulting events. c. Deep (z > 100 km), thrust-faulting earthquakes. d. to f. Normal-faulting events, in same layout as for thrust-faulting events.

  • Here are three figures from Tremblor.net, one of which is in the interpretive poster. These are the analyses I was discussing that we needed to see in my initial report. More detailed discussion can be found here.


  • This figure shows that there are not many earthquakes in the region between the M 8.1 and M 7.1 earthquakes. This is supporting evidence that there was not a significant increase in stress in this region (independent negative evidence for static triggering of the M 7.1 from the M 8.1).


  • This figure shows their modeling of the subduction zone in the region of the M 8.1 earthquake. I queried whether the megathrust had an increased stress following the M 8.1 earthquake. Part of the megathrust here ruptured in 1902, but the rest of the “Tehuantepec Gap” does not have an historic record (since ~1600 AD). Note how the megathrust is mostly blue, suggesting a lower likelihood of rupture. There is a narrow band of increased stress (in red). This model uses the finite fault model from Dr. Gavin Hayes (USGS).


    •

  • As far as the likelihood of dynamic triggering (increased stress on faults while seismic waves are travelling through them), here is an analysis that helps us visualize this. This analysis (Pollitz et al., 2012) shows regions of increased dynamic stress following the 2012 Wharton Basin earthquakes. The lower spheres show seismicity for a time period following the earthquakes and note how they align with the red areas, areas of increased dynamic stress.


    • The 2012 M = 8.6 mainshock and M = 8.2 aftershock fault ruptures and maps of strain duration tstrain at a threshold value of 0.1 microstrain. a, Inferred fault ruptures of the 11 April 2012 M = 8.6 east Indian Ocean earthquake and an M = 8.2 aftershock that occurred 2 h later. Superimposed are the first 20 d of M > 4.5 aftershocks of 0–100-km depth. These earthquakes probably ruptured a complex set of subparallel and conjugate faults with the indicated sense of motion (arrows). Parts of the rupture areas of the 2004 M = 9.2 and 2005 M = 8.7 Nias earthquakes on the Sunda megathrust are indicated. b, c, Global maps of tstrain (colour scale). Superimposed are the epicentres of M>5.5 events that occurred during the 6 d preceding the mainshock (2 epicentres) and following the mainshock (24 epicentres, 16 of which are remote, that is, .1,500km from the mainshock). Focal mechanisms of six post-mainshock events with near-vertical strike-slip mechanisms (plunge of neutral axis, >60 deg) are indicated with red beachballs. The 9:00:09 11 April 2012 M = 5.5 event (in the western Aleutian Islands) occurred 21 min 33 s after the mainshock between the direct P- and S-wave arrivals from the mainshock; all others are delayed by hours to days. The focal mechanism of the mainshock is plotted at its epicentre.

  • Here is the comparison I put together for the ground motion modeling presented in the poster above.


    •

  • Here is a really cool video that shows the seismic record of Hurricane Maria and the M 7.1 earthquake are recorded by seismometers (prepared by . The top panel shows the seismograph. The middle panel shows a spectrogram of these seismic data (showing the frequency content of the seismic waves). The lower panel shows the position of the Hurricane and M 8.1 earthquake epicenter (they should have shown the M 7.1, but that is not important. The audio is a conversion of the seismic data into sound. Here is the 1 MB mp4 file for downloading. This was prepared by Zhigang Peng from Georgia Tech for the station IU.SJG — San Juan, Puerto Rico. This is posted on the IRIS special event page. note: the hurricane and this earthquake are NOT RELATED!

References:

  • Benz, H.M., Dart, R.L., Villaseñor, Antonio, Hayes, G.P., Tarr, A.C., Furlong, K.P., and Rhea, Susan, 2011 a. Seismicity of the Earth 1900–2010 Mexico and vicinity: U.S. Geological Survey Open-File Report 2010–1083-F, scale 1:8,000,000.
  • Benz, H.M., Tarr, A.C., Hayes, G.P., Villaseñor, Antonio, Furlong, K.P., Dart, R.L., and Rhea, Susan, 2011 b. Seismicity of the Earth 1900–2010 Caribbean plate and vicinity: U.S. Geological Survey Open-File Report 2010–1083-A, scale 1:8,000,000.
  • Cruz-Atienza et al., 2016. Long Duration of Ground Motion in the Paradigmatic Valley of Mexico in Scientific Reports, v. 6, DOI: 10.1038/srep38807
  • Franco, A., C. Lasserre H. Lyon-Caen V. Kostoglodov E. Molina M. Guzman-Speziale D. Monterosso V. Robles C. Figueroa W. Amaya E. Barrier L. Chiquin S. Moran O. Flores J. Romero J. A. Santiago M. Manea V. C. Manea, 2012. Fault kinematics in northern Central America and coupling along the subduction interface of the Cocos Plate, from GPS data in Chiapas (Mexico), Guatemala and El Salvador in Geophysical Journal International., v. 189, no. 3, p. 1223-1236. DOI: https://doi.org/10.1111/j.1365-246X.2012.05390.x
  • Franco, S.I., Kostoglodov, V., Larson, K.M., Manea, V.C>, Manea, M., and Santiago, J.A., 2005. Propagation of the 2001–2002 silent earthquake and interplate coupling in the Oaxaca subduction zone, Mexico in Earth Planets Space, v. 57., p. 973-985.
  • Garcia-Casco, A., Projenza, J.A., Iturralde-Vinent, M.A., 2011. Subduction Zones of the Caribbean: the sedimentary, magmatic, metamorphic and ore-deposit records UNESCO/iugs igcp Project 546 Subduction Zones of the Caribbean in Geologica Acta, v. 9, no., 3-4, p. 217-224
  • Gérault, M., Husson, L., Miller, M.S., and Humphreys, E.D., 2015. Flat-slab subduction, topography, and mantle dynamics in southwestern Mexico in Tectonics, v. 34, p. 1892-1909, doi:10.1002/2015TC003908.
  • Quzman-Speziale, M. and Zunia, F.R., 2015. Differences and similarities in the Cocos-North America and Cocos-Caribbean convergence, as revealed by seismic moment tensors in Journal of South American Earth Sciences, http://dx.doi.org/10.1016/j.jsames.2015.10.002
  • Hayes, G. P., D. J. Wald, and R. L. Johnson, 2012. Slab1.0: A three-dimensional model of global subduction zone geometries, J. Geophys. Res., 117, B01302, doi:10.1029/2011JB008524.
  • Lay et al., 2011. Outer trench-slope faulting and the 2011 Mw 9.0 off the Pacific coast of Tohoku Earthquake in Earth Planets Space, v. 63, p. 713-718.
  • Manea, M., and Manea, V.C., 2014. On the origin of El Chichón volcano and subduction of Tehuantepec Ridge: A geodynamical perspective in JGVR, v. 175, p. 459-471.
  • Mann, P., 2007, Overview of the tectonic history of northern Central America, in Mann, P., ed., Geologic and tectonic development of the Caribbean plate boundary in northern Central America: Geological Society of America Special Paper 428, p. 1–19, doi: 10.1130/2007.2428(01). For
  • McCann, W.R., Nishenko S.P., Sykes, L.R., and Krause, J., 1979. Seismic Gaps and Plate Tectonics” Seismic Potential for Major Boundaries in Pageoph, v. 117
  • Pérez-Campos, Z., Kim, Y., Husker, A., Davis, P.M. ,Clayton, R.W., Iglesias,k A., Pacheco, J.F., Singh, S.K., Manea, V.C., and Gurnis, M., 2008. Horizontal subduction and truncation of the Cocos Plate beneath central Mexico in GRL, v. 35, doi:10.1029/2008GL035127
  • Polltz, F.F., Stein, R.S., Sevigen, V., Burgmann, R., 2012. The 11 April 2012 east Indian Ocean earthquake triggered large aftershocks worldwide in Nature, v. 000, doi:10.1038/nature11504
  • Symithe, S., E. Calais, J. B. de Chabalier, R. Robertson, and M. Higgins, 2015. Current block motions and strain accumulation on active faults in the Caribbean in J. Geophys. Res. Solid Earth, v. 120, p. 3748–3774, doi:10.1002/2014JB011779.

1 thought on “Earthquake Report: Puebla, Mexico Update #1”

  1. Pingback: Deadly earthquake in the Puebla region, Mexico - at least 226 fatalities - September 19, 2017

Leave a Reply

Your email address will not be published. Required fields are marked *