Comment on “Active Crustal Deformation in the Trans‐Mexican Volcanic Belt as Evidenced by Historical Earthquakes During the Last 450 Years” by G. Suárez et al.
2020; Wiley; Volume: 39; Issue: 6 Linguagem: Inglês
10.1029/2019tc006016
ISSN1944-9194
Autores Tópico(s)Geological and Geochemical Analysis
ResumoThe paper by Suárez et al. (2019) is to be welcomed for drawing fresh attention to the ground-shaking hazard in central Mexico caused by continental earthquakes. These have smaller source and damage areas, a lower magnitude, and greater recurrence times than the more common subduction zone earthquakes but are known to have caused locally moderate to severe damage or were even devastating in historical times. However, some of the key assumptions in Suárez et al. (2019) are misguided, and many of the macroseismic observations underlying their analyses are inaccurate. The source of the CE (Common Era) 1567 Mw 7.2 Ameca earthquake is unlikely to be located on the northern shoulder of the Chapala graben as advocated by Suárez et al. (2019) Apparently unknown to the authors, contemporary sources such as the Relación de Ameca (Acuña, 1988) quantitatively describe the surface rupture of this earthquake along the fault bounding the Ameca half-graben. Of the five locations with very heavy damage (macroseismic intensity of degree 9) known from primary historical sources (Suter, 2015a), which are all located in the Ameca, Zacoalco, and Sayula half-grabens, just one is marked in the related graph by Suárez et al. (2019) (their Figure 3). Accounting for these data points in their inversion would obviously change significantly its outcome, such as shifting the source location to this high-intensity cluster. Moreover, there is no evidence that this earthquake occurred on 27 December 1568 as asserted by Suárez et al. Based on sixteenth century documents, its origin can clearly be pinpointed to 28 December 1567 at dawn (Suter, 2019a). The macroseismic observations in Suárez et al. (2019) for the 14 and 15 April 1611 earthquakes are not based on contemporary documentary sources with exception of the CE 1653 chronicle by Antonio Tello, according to which earthquakes were felt on those days in Zapotlán (now Ciudad Guzmán) (Tello, 1891). There is no information about these earthquakes in Bárcena (1875) as claimed by Suárez et al. (2019) and the destruction of the Franciscan convents in Sayula and Zapotíltic and of the recently rebuilt church in Zapotlán by these earthquakes reported in Munguía Cárdenas (2012) and Pérez Verdía (1951) is not backed up by primary sources. According to the Tello chronicle, Zapotlán was not damaged by the 14 and 15 April 1611 earthquakes, as wrongly described by Suárez et al. (2019) but by the 26 August 1611 earthquake that ruined the church, convent, and many residential buildings (Tello, 1891, p. 770). The destruction caused by the same 26 August 1611 earthquake in Mexico City is described in detail by the indigenous historian and eye witness Chimalpahin Quauhtlehuanitzin (Lockhart et al., 2006; Townsend, 2017). This earthquake was likely caused by a major rupture of the subduction interface in the Michoacán region and requires further documentation and analysis. The 14 and 15 April 1611 events, on the other hand, likely were volcanic tremors and not sourced by faults of the Colima graben as assumed by Suárez et al. (2019). According to the chronicle by Mota Padilla, written in CE 1742, the Fuego de Colima Volcano erupted on 15 April 1611 and spread ashes over a distance of more than 40 Mexican colonial leagues (168 km) (Mota Padilla, 1870, p. 271). Because the only contemporary documentary source for the 14 and 15 April 1611 earthquakes is in reality a felt report for Zapotlán, the epicenter location and 6.4 ± 0.2 magnitude estimate by Suárez et al. (2019) lack evidence and should be discarded. Similarly misleading is the presentation of the 22 and 23 October 1749 earthquakes that were devastating north of Fuego de Colima Volcano, in the northern Colima graben. Suárez et al. (2019) ignore contemporary archival sources, according to which these earthquakes razed Zapotlán where only three residential buildings remained standing (Suter, 2019b). Furthermore, the macroseismic intensities of degree 7 they assigned to Guadalajara and Zacoalco (their Figure 5) based on the 23 March 1875 report in the newspaper El Siglo Diez y Nueve are not supported by historical documents. A building inspection of the Guadalajara cathedral shortly after the 23 October 1749 mainshock could not find any damage to the vaulted roof and the walls (Compendio de los libros de actas del venerable cabildo de la santa iglesia catedral de Guadalajara, libro 11, foja 58; López, 1971). Apparently unknown to Suárez et al. (2019) the newspaper report is a flawed summary of a 1771 document (Auto proveído por la Real Audiencia de Guadalajara; Rivera, 1989, p. 235–236). The original source includes not a word about earthquake damage in Zacoalco. Suárez et al. (2019) neglect considerable recent research into the historical seismicity of the Chapala graben region, where the 2 October 1847 continental earthquake was locally devastating on the northern graben shoulder. It razed the villages of Poncitlán and Ocotlán in the state of Jalisco, where at least 58 persons perished (Suter, 2018). The macroseismic observations for this historical event and the elevated background seismicity indicate that the Chapala graben is active and poses a major ground-shaking hazard to the nearby metropolitan areas of Ocotlán and Guadalajara. Suárez et al. (2019) also failed to take into account the historical seismicity of the Morelia region where the seismic hazard is high because of the seismically active Morelia normal fault that dips beneath this urban area with a population of more than 900,000 habitants. Several slip events on this fault can be inferred from paleoseismicity studies. The latest of them must have occurred after CE 1290–1435, during the postclassic period or possibly during early novohispanic times (Suter, 2016), and recent activity of this fault was recorded instrumentally by Singh et al. (2012). As for the 19 June 1858 earthquake, the authors fail to acknowledge the careful macroseismic study of this great earthquake by Molina del Villar (2001, 2004), which provides macroseismic data practically identical to theirs. I consider it unlikely that this was a continental earthquake as interpreted by Suárez et al. (2019) and prefer the interpretation by Singh et al. (1996) of this event as having an intraslab normal faulting source, close to the southern margin of the Trans-Mexican Volcanic Belt (hereafter, TMVB). A continental normal faulting earthquake of a 7.6 ± 0.3 magnitude with focus within the belt, as suggested by the authors, would have an average throw >2 m and a rupture length >100 km at the surface (Stirling et al., 2013; Suter, 2015b) and would thus have to extend over an array of several fault segments, for which there is no evidence. It is highly unlikely that such an enormous surface rupture would have gone unreported in CE 1858 in this densely populated region. Furthermore, a surface rupture of such dimensions should still be easily detectable in this semiarid region, which has been among the geologically most field studied in Mexico over the last 30 years (Ferrari et al., 2012, 2018). In fact, the epicenter resulting from the inversion by Suárez et al. (2019) is not located within the array of east-west normal faults deforming the TMVB but south of it (their Figure 7). The 19 September 2017 Mw 7.1 intraslab earthquake is a reminder that such earthquakes can have a severe societal impact within the TMVB (Singh et al., 2018). As for the 1875 San Cristóbal de la Barranca continental earthquake, the statements by Suárez et al. (2019) that the town of San Cristóbal was moved in the 1960s and the epicentral region was transformed due to the construction of a concrete dam on the Santiago River is outright fabricated and clearly shows that the authors lack a first-hand knowledge of the study area. San Cristóbal de la Barranca is still at the same location where it was during the 1875 earthquake, and the Santiago River is not dammed in that region. As for the 7.0 ± 0.2 magnitude of this earthquake resulting from their inversion, this number is almost certainly far too high. The detailed contemporary field survey by Iglesias et al. (1877) did not find any surface rupture as would be expected for an extensional continental earthquake of this magnitude (dePolo, 1994). Based on the felt and damage areas inferred from Figure 6 in Suárez et al. (2019) and calibrated magnitude-isoseismal area relationships of shallow normal fault earthquakes in the TMVB (Suter et al., 1996), the magnitude of this earthquake is more likely to be in the 5.5 to 6.0 range. The epicentral region of the 1887 Pinal de Amoles earthquake is interpreted by Suárez et al. (2019) to be part of the TMVB, whereas according to Zúñiga et al. (2017, their Figure 4), this region belongs to their Burgos basin seismotectonic province. Both these interpretations lack a basic understanding of the continental tectonics of north-central Mexico. The Burgos basin is a foreland basin of the Paleogene Coahuila fold belt (Eguiluz, 2011; Pérez-Cruz, 1993) near the border to Texas, and borehole breakouts indicate there a northwest-southeast orientation of the current least horizontal stress (Suter, 1987). The epicentral region of the Pinal de Amoles earthquake, on the other hand, is part of the tectonically distinctly different southern Basin and Range Province, which is characterized by active east-west oriented extension (Suter, 1991). A north-south striking normal fault northwest of Pinal de Amoles with a morphologically pronounced fault scarp, 45 km long, likely was the source of the 1887 event (Suter et al., 1996, their Figure 2). An earthquake sequence in 2010–2011, located just south of the trace of this fault, yielded focal mechanisms indicating a north-south striking normal fault as source (Clemente-Chávez et al., 2013). The northern margin of the TMVB is located ~60 km south of Pinal de Amoles, where north striking normal faults of the Basin and Range Province are intersected by east-west striking normal faults of the Aljibes half-graben (Suter et al., 1995). This structural configuration can be explained by horizontal migration of the boundary between the two stress provinces with time, which requires intermittent permutations between the intermediate and least principal stresses. The central part of the TMVB is characterized by an active extensional fault network. In 1912 some of the faults ruptured to the surface in an M ~7.0 earthquake causing moderate damage in Mexico City (Suter, 2014), and the paleoseismicity studies summarized by Suárez et al. (2019) document that the major segments of this fault system ruptured repeatedly during the Holocene. However, open questions still remain, especially whether the segmented southern border of the fault network, 80 km long, ruptured across segment boundaries in prehistoric extreme events and what ground shaking such a throughgoing rupture would cause in the nearby Greater Mexico City metropolitan area. As for the slip rates and mean recurrence intervals reviewed by Suárez et al. (2019) these paleoseismicity results should be taken with a grain of salt. Most of the studies do not document displacements across the master faults but across secondary ruptures on their hanging wall side, notwithstanding that most of the slip is likely to occur along the primary fault surface itself. Other studies focus on splays near the tips of major faults, where the throw is attenuated. In both situations, only a minimum of the overall displacement is intercepted by the excavated trenches, and the slip rates in the center part of the master faults are likely to be significantly higher than estimated in these studies. Furthermore, earthquake extreme events often cluster in time (Ortuño et al., 2019; Salditch et al., 2020; Scholz, 2019 and references therein), and the rupture recurrence times can have such a high variance that the average becomes meaningless. Suárez et al. (2019) failed to develop an intensity attenuation model for the TMVB but applied instead a model for Southern California (Los Angeles Basin and San Andreas fault region) by Bakun (2006). They wrongly claim to have inverted in Suárez and Caballero-Jiménez (2012) their macroseismic data of the 1912 Acambay and 1920 Jalapa earthquakes using published attenuation equations for the much different geodynamic frameworks of Southern California, Hispaniola, and Italy, which borders on disinformation. Moreover, Suárez et al. (2019) did not take into account published intensity attenuation curves for seven early instrumental earthquakes in the TMVB spanning a broad range of instrumental magnitudes between 4 and 7 (Suter et al., 1996, their Figure 12). These curves show considerable differences in the decay of intensity with epicentral distance, which depends not only on the magnitude but also on focal depth, geometric spreading, and anelastic attenuation. Any generalized intensity attenuation model for the continental earthquakes within the TMVB will for that reason have a large uncertainty. Furthermore, the inversion applied by Suárez et al. (2019) does not resolve for the epicenter as they assume but for the point of maximum intensity, which highly depends on site amplification. A case in point is the intensity distribution of Michoacán-Guerrero subduction interface earthquakes, which do not have their maximum intensity in their epicentral region but in the Mexico City basin. In conclusion, the commented article remains intransparent. It lacks a self-critical appraisal of the applied methodology and a critical analysis of the underlying paleoseismic and macroseismic observations. These basic fallacies severely undermine the well-meant intent by Suárez et al. (2019) to further our understanding of the ground-shaking hazard in this densely populated region. No new data were generated for this comment. All data necessary to understand, evaluate, replicate, and build upon the comment are contained in the cited references. The referenced issue of the newspaper El Siglo Diez y Nueve was culled from the digital newspaper collection of the National Library of Mexico (Hemeroteca Nacional Digital de México; http://www.hndm.unam.mx). The input data for the referenced intensity attenuation curves in Suter et al. (1996) are preserved online (https://www.researchgate.net/profile/Max_Suter).
Referência(s)