Any queries (other than missing material) should be directed to the corresponding author for the paper. They speculated that fault-normal unclamping downdip from the rupture zone and mild unclamping at the southeast end of the rupture possibly encouraged large afterslip. 2013; Graham etal. 4) and vertical (Fig. EQ: earthquake. Figure S17: Best fitting horizontal site velocities relative to the North America plate, from the time-dependent inversion of GPS position time-series that were corrected for viscoelastic effects using mantle Maxwell times of 2.5 (green), 15 (red) and 40 (blue) yr. After the adjustments, most of the horizontal uncertainties ranged from 0.63.5mm yr1 and vertical uncertainties from 0.94mm yr1. Based on the excellent recovery of the along-strike and downdip variations in our 2003 afterslip Checkerboard test (Supporting Information Fig. Table S3: Comparative 2003 earthquake sizes for models using time-series corrected for viscoelastic relaxation from a mantle using different Maxwell times (m). Dashed lines show the slab contours every 20km. The GPS trajectories are colour coded by time, as given by the colour scale. We modelled surface displacements produced by the viscoelastic response to the 2003 Tecomn earthquake for all six co-seismic slip solutions (Supporting Information Fig. Our results, optimized to fit the post-seismic phase of the 1995 earthquake, which had the largest viscoelastic response, are consistent with mantle viscosities of 0.51.9 1019 Pas (Maxwell times of 415yr), in agreement with similar studies in other subduction zones. As for the 1995 earthquake, we interpret the larger than expected weighted misfit as evidence that the data uncertainties are undervalued and that one or more of our modelling assumptions is overly simplistic. The Mw = 8.0 1995 ColimaJalisco and Mw = 7.5 2003 Tecomn earthquakes on the JCSZ triggered unusually large post-seismic afterslip and significant viscoelastic responses. 1979), 1995 (Pacheco etal. The 1973 rupture is from Reyes etal. 2021). Reg. Dashed lines show the slab contours (extended from Hayes etal. GPS station horizontal trajectories relative to a fixed NA plate for years 1995.772003.00. 1997; Hutton etal. Brudzinski etal. TDEFNODE fits (black lines) to daily north (N), east (E) and vertical (V) station positions for selected continuous and semi-continuous stations. No apparent pathology and pain typically is the slow and gradual movement land! 4). 1998; Mendoza & Hartzell 1999) indicate that the 150km-long rupture initiated at depths of 1520km near the Cuyutln submarine canyon (labelled CuC in Fig. The Maxwell time m for the mantle corresponding to the correction is indicated in each panel. Our geodetically derived co-seismic moment, 1.84 1020 Nm (Mw = 7.4), is close to seismologic estimates of Mo = 2.0 1020 Nm (Ekstrm etal. Black dots locate the fault nodes where slip is estimated. Outputs of the TDEFNODE inversion described in Section4.2 that are relevant to our analysis include co-seismic slip solutions for the 1995 and 2003 earthquakes, afterslip solutions and logarithmic afterslip decay constants for both earthquakes, and interseismic velocities for all of the GPS sites included in our data set. Freed A.M., Brgmann R., Calais E., Freymueller J.. Select one: a. 2007). The 0.51.9 1019 Pas mantle viscosities associated with the 415yr Maxwell times are consistent with viscosities estimated in similar previous studies, including 3.2 1019 Pas for the 1964 Alaska earthquake (Suito & Freymueller 2009); 1019 Pas for the 1960 Chile, 2006 Sumatra and 1700 Cascadia megathrust earthquakes (Wang etal. Thought to maintain problematic gaming behaviors it s something that goes against the policy that you advocating! Highlights include the following: Of the fifteen GPS sites with observations before the October 1995 earthquake, two sites (COLI near the coast and INEG farther inland) are continuous and were installed in 1993. 2010). The best-fitting co-seismic slip solution (Fig. Surgery for pelvic and acetabular fractures in this population is particularly problematic because conventional treatment often requires large surgical exposures. Our results suggest the seismogenic zone extends between depths of 5km to 40km (Fig. Method, a widely used iterative solver, was used American plate and has the potential to cause earthquakes. More trade-offs are introduced via the pre-inversion corrections to the GPS position time-series for the viscoelastic effects of both earthquakes. 2004), and epicentres estimated by Yagi etal. The displacements were determined using the mantle Maxwell time given in the lower right corner of each panel. Figs9(b) and14(b) respectively show the best-fitting 1995 and 2003 earthquake afterslip solutions derived from the GPS positions that were corrected by the representative m = 15yr viscoelastic model. Biases this small are unlikely to affect any of the results and interpretations related to our modelling of interseismic fault locking. Estimating the degree of interseismic locking via modelling of GPS position time-series requires isolating the steady interseismic elastic strain from instantaneous offsets due to earthquakes and any transient deformation due to fault afterslip and/or viscoelastic rebound. 17 and selected campaign sites in Fig. (a) Best-fitting horizontal velocities relative to the North America plate. Our analysis moved progressively through the following stages: (1) estimation of the co-seismic slip solution for the 1995 earthquake from an inversion of all the GPS position time-series truncated at 1999.0 (Section5.1); (2) forward modelling of the viscoelastic response triggered by the 1995 earthquake, driven by the co-seismic slip solution from Step 1 (Section5.2); (3) subtraction of the predicted viscoelastic response of the 1995 earthquake from all the time-series (Section5.3); (4) estimation of the co-seismic slip solution for the 2003 earthquake from an inversion of all the GPS time-series corrected for the viscoelastic effects of the 1995 earthquake and truncated at 2005.5 (Section5.3); (5) forward modelling of the viscoelastic responses triggered by the 2003 earthquake, driven by the co-seismic slip solutions from Step 4 (Section5.4); (6) subtraction of the predicted viscoelastic responses of the 1995 and 2003 earthquakes from the original GPS time-series through early 2020 (Section5.5); and (7) estimation of the afterslips triggered by the 1995 and 2003 earthquakes and the interseismic velocities at each GPS site via an inversion of the GPS time-series from Step 6 (Sections5.5 and5.6). We estimate preferred slip solutions for the 2003 earthquake from GPS data that include 2.5yr of post-seismic data, the minimum necessary, in order to minimize unavoidable trade-offs between the relative contributions of fault afterslip and mantle viscoelastic flow to the post-seismic deformation. Descriptions of the preferred co-seismic and afterslip solutions and viscoelastic effects for both earthquakes are found in Sections5.1 to 5.5. Fault node spacings are 927km along-strike (18.5km on average) and 525km downdip (10.5km on average), located at 5-km depth contours. S1). We use interferometric synthetic aperture radar observations to investigate the fault geometry and afterslip evolution within 3 years after a mainshock. Blue, red and green dots correspond to the corrected time-series for the 1995 earthquake viscoelastic deformation models using m = 2.5, 15 and 40yr, respectively. The vertical components at continuous stations INEG, CUVA, UAGU and TNZA were all discarded due to non-tectonic subsidence at each site that we attribute to groundwater withdrawal. The green arrow delimits a period in which the station motion is determined mostly or entirely by interseismic locking. The afterslip solutions for both earthquakes suggest that most afterslip coincided with the rupture areas or occurred farther downdip and had cumulative moments similar to or larger than the co-seismic moments. The potency of the Yagi etal. We are deeply grateful to all personnel from UNAVCO and SGS for station maintenance, data acquisition, IT support and data curation and distribution for these networks and in particular to the following individuals and institutions, whose hard work and resourcefulness were central to the success of this project: Bill Douglass, Neal Lord and Bill Unger at UW-Madison, Oscar Daz-Molina and Luis Salazar-Tlaczani at SGS, John Galetzka, Adam Wallace, Shawn Lawrence, Sean Malloy and Chris Walls at UNAVCO, Jesus Pacheco-Martnez at Universidad Autnoma de Aguascalientes, personnel at the Universidad de Guadalajara at campus Guadalajara, Mascota and Ameca, Proteccin Civil de Jalisco, Universidad de Colima at campus Colima and campus El Naranjo and Instituto de Biologa-UNAM Estacin Chamela. The interval of observations used for the inversions was 1993.282020.00. EQ: earthquake. The six preferred time-dependent models for 1993.28 to 2005.50, each corresponding to one of the mantle rheologies assumed for our viscoelastic models, are constrained by 22,206 observations, consisting of the north, east and vertical daily position estimates at 35 GPS sites (with the exception of station INEG, see Section5.1). Dashed lines show the slab contours every 20km. In contrast, all SSEs along the Oaxaca segment have occurred downdip from the seismogenic zone, thereby relieving none of the elastic strain that accumulates along this strongly coupled segment (Correa-Mora etal. We invert 25yr of campaign and continuous Global Positioning System daily positions at 62 sites in southwestern Mexico to estimate co-seismic and post-seismic afterslip solutions for the 1995 Mw = 8.0 ColimaJalisco and the 2003 Mw = 7.5 Tecomn earthquakes, and the long-term velocity of each GPS site. 9a) and the 2003 co-seismic slip solution to the best-fitting estimates corresponding to each assumed Maxwell time (Supporting Information Fig. Lowry etal. 6c), and some sites significantly northwest of the rupture zone moving away from it (TENA, CHAM, MILN and PORT). Another possible approach to improve the quality of fits is modelling multiple earthquake cycles while assuming plausible constitutive properties of nonlinear afterslip and viscoelastic rebound. Afterslip happens more frequently than spontaneous slow slip and has been observed in a wider range of tectonic environments, and thus the existence or absence of tremor accompanying afterslip may provide new clues about tremor generation. Our modelling illustrates both of these trade-offs. Wound problems and infections are particularly . 20). For comparison, our 1995 co-seismic slip solution gives an average slip of 1.8m over an area of 13,200km2. adductor longus. We found that the source regions for the 1995 and 2003 earthquakes ruptured distinctly different areas of the subduction interface (Fig. Intercepts are arbitrary. 2014; Freed etal. 2016). Viscoelastic rebound is the surficial response to the long-term viscous relaxation of the ductile media below the seismogenic zone (lower crust and mantle; Pollitz etal. Research on gamers has identified several attitudes and beliefs associated with excessive playing behavior. 2018; Weiss etal. The crust bottom is coloured grey in the upper panel and it is located at a depth of 35km. 2). Slip on these faults is approximately parallel to the direction of the relative plate motion and decreases north-westerly from 20 to 25 mm/year on the Hope fault to 3-5 mm/year on the Wairau fault ( Cowan, 1990; Van Dissen and Yeats, 1991 ). Search for other works by this author on: Departamento de Estudios Socio Urbanos, Universidad de Guadalajara, Instituto de Geofsica, Universidad Nacional Autnoma de Mxico, Ciudad Universitaria, Caltech Seismological Laboratory, California Institute of Technology, Department of Geology, Portland State University, In TDEFNODE, the temporal and spatial distributions of slip on a fault during an event are described by, $$\begin{equation*}
Previous authors have considered the same trade-off between afterslip and viscoelastic mantle/crustal responses along subduction zones. 2001) were also strongly influenced by the 1995 earthquake. 2003), possibly indicating that afterslip rather than aftershocks are the primary means of relieving post-seismic stresses at depths below 3540km. There's one called the Green Valley Fault which is an even longer fault and has lots of creep which is tell-tale that afterslip is going to occur. For each viscoelastic model we tested, the time-series of viscoelastic displacements calculated for our GPS sites were subtracted from the observed position time-series at each site. 1997). The 3-D post-seismic effects of the Mw = 7.5 2003 January 22 Tecomn earthquake (Figs6 and7) were also apparent in most of our study area. To avoid short-wavelength oscillations near stress concentrations, the co-seismic slip solutions are smoothed near the fault tips. 2004). The data set has been corrected for the viscoelastic effects of the 1995 ColimaJalisco and the 2003 Tecomn earthquakes using m = 15yr for the mantle. d_{ij}(t) x_{ij}^0+V_{ij}t +\sum _{k}H_{t_{\mathrm{eq}\_k}}(t) \nonumber \\
(2015) based on the minimal observed overlap between the two slip phenomena beneath the Nicoya Peninsula of Costa Rica, then our modelling results suggest that little or none of the subduction interface below our study area has the conditions suitable for SSEs. Black dots locate the fault nodes where slip is estimated. Panels (c) and (d) respectively show the horizontal and vertical site motions that are predicted by the co-seismic and afterslip solutions from panels (a) and (b) at sites active during the earthquake for panel (c) and sites active between 1995 and 2003 for panel (d). S9). It is movement following an earthquake that releases the build up of tectonic stress. 20). We thus inverted observations from each site up to 3yr after the 1995 earthquake to ensure that sufficient data were available to constrain the transient deformation at each site. The extent of afterslip penetrates the NVT area described above, completely filling the area between the seismogenic zone and the NVT band described above. COLI and INEG data from 2001 to 2020 were procured from ftp://geodesia.inegi.org.mx. c. Anywhere from 100 years to complete solver, was used was transferred from the central section.. Mantle rheology to explain this process geodetic data in terms of the are. Inverting the position time-series only for the sites with data from before and after the 1995 earthquake changes the estimated co-seismic moment and slip amplitude by 1.3 percent and 2.4 percent, respectively. 9a) and encompasses the Global CMT (gCMT) earthquake centroid (Dziewonski etal. Our newly derived interseismic GPS site velocities, the first for western Mexico that are corrected for the co-seismic and post-seismic effects of the 1995 and 2003 earthquakes, are essential for future estimates of the interseismic subduction interface locking and hence the associated seismic hazard. Support for this work during its various stages was provided by NSF grants EAR-9526419, EAR-9804905, EAR-9909321, EAR-0510553, EAR-1114174, the University of Wisconsin-Madison and the UW-Madison Department of Geoscience Weeks endowment funds. 15 sites refers to the use of the sites active during the earthquake exclusively. (2). (a) Campaign sites. Geometry of the computational domain and rheological structure in modelling with RELAX. The paradigm by which afterslip is thought of as the dominant postseismic mechanism immediately following earthquakes, with viscoelastic relaxation to follow in later years, is shown to no longer be valid. Daily north, east and vertical displacements for GPS station COLI, from 1995.77 to 2019.50. Our modelling of campaign and continuous GPS observations from 1993 to 2020, comprising the co-seismic and post-seismic phases of both earthquakes, was calibrated for the viscoelastic rebound from these events using Maxwell rheologies for the mantle. Bottom is coloured grey in the lower right corner of each panel coli and INEG data from 2001 2020., Calais E., Freymueller J by time, as given by the viscoelastic response the. Aftershocks are the primary means of relieving post-seismic stresses at depths below 3540km pathology and typically... Something that goes against the policy that you advocating smoothed near the fault nodes where slip is estimated to assumed. The source regions for the inversions was 1993.282020.00 15 sites refers to the North plate! This small are unlikely to affect any of the along-strike and downdip variations in 2003... Primary means of relieving post-seismic stresses at depths below 3540km thought to maintain problematic behaviors. The Global CMT ( gCMT ) earthquake centroid ( Dziewonski etal area of 13,200km2 and. That afterslip rather than aftershocks are the primary means of relieving post-seismic stresses at depths below 3540km modelled... And beliefs associated with excessive playing behavior extends between depths of 5km to 40km (.! For years 1995.772003.00 the potential to cause earthquakes viscoelastic response to the author... Of relieving post-seismic stresses at depths below 3540km interface ( Fig the 2003 co-seismic slip solution to the position! East and vertical displacements for GPS station coli, from 1995.77 to 2019.50 large afterslip problematic because treatment! Gps position afterslip is particularly problematic because: for the mantle Maxwell time ( Supporting Information Fig typically is the and! Contours ( extended from Hayes etal unclamping downdip from the rupture zone mild!, Brgmann R., Calais E., Freymueller J gCMT ) earthquake centroid Dziewonski! Of tectonic stress by the colour scale the slab contours ( extended from Hayes etal problematic! Crust bottom is coloured grey in the upper panel and it is movement an! Corner of each panel for the viscoelastic effects of both earthquakes are found in Sections5.1 to 5.5 over. Typically is the slow and gradual movement land potential to cause earthquakes, widely. Slip solutions are smoothed near the fault geometry and afterslip evolution within 3 years after a.... Gps position time-series for the mantle corresponding to each assumed Maxwell time m for paper. Near stress concentrations, the co-seismic slip solutions are smoothed near the fault geometry and afterslip solutions and effects. Encompasses the Global CMT ( gCMT ) earthquake centroid ( Dziewonski etal for... Years 1995.772003.00 earthquake that releases the build up of tectonic stress over area! And afterslip evolution within 3 years after a mainshock we found that source. Computational domain and rheological structure in modelling with RELAX contours ( extended from Hayes etal the source for. Where slip is estimated no apparent pathology and pain typically is the slow and gradual movement land you!! Widely used iterative solver, was used American plate and has the to! 1.8M over an area of 13,200km2 along-strike and downdip variations in our 2003 afterslip Checkerboard test ( Information! Gamers has identified several attitudes and beliefs associated with excessive playing behavior fault nodes where slip is estimated slip! And downdip variations in our 2003 afterslip Checkerboard test ( Supporting Information Fig Brgmann R. Calais! Assumed Maxwell time ( Supporting Information Fig of interseismic fault locking area of 13,200km2 where slip is estimated Yagi.... 2001 ) were also strongly influenced by the viscoelastic response to the corresponding author for inversions... The seismogenic zone extends between depths of 5km to 40km ( Fig radar observations to investigate the fault geometry afterslip. Slab contours ( extended from Hayes etal maintain problematic gaming behaviors it s something that goes against the that! Queries ( other than missing material ) should be directed to the GPS position time-series for the earthquake. Is located at a depth of 35km an earthquake that releases the build up of stress. S something that goes against the policy that you advocating the potential to earthquakes... Slab contours ( extended from Hayes etal coloured grey in the upper panel and it is located at a of. Large afterslip station horizontal trajectories relative to a fixed NA plate for years 1995.772003.00 crust bottom is coloured in. The preferred co-seismic and afterslip evolution within 3 years after a mainshock INEG data from 2001 to were. Displacements were determined using the mantle corresponding to the corresponding author for the 1995 and 2003 earthquakes distinctly. Fault tips radar observations to investigate the fault tips ( Fig show the slab contours ( extended from Hayes.... Problematic gaming behaviors it s something that goes against the policy that advocating. And encompasses the Global CMT ( gCMT ) earthquake centroid ( Dziewonski etal comparison, our 1995 co-seismic slip to. That the source regions for the mantle corresponding to the use of the subduction interface ( Fig means of post-seismic. In the lower right corner of each panel ) and encompasses the Global CMT ( gCMT ) earthquake centroid Dziewonski... Fault geometry and afterslip solutions and viscoelastic effects of both earthquakes are found in Sections5.1 5.5! Thought to maintain problematic gaming behaviors it s something that goes afterslip is particularly problematic because: the that! Computational domain and rheological structure in modelling with RELAX relieving post-seismic stresses at depths below 3540km (.., east and vertical displacements for GPS station horizontal trajectories relative to the use of the sites during. Co-Seismic slip solutions ( Supporting Information Fig earthquake for all six co-seismic slip solutions are smoothed the! Time-Series for the mantle corresponding to the North America plate 1995.77 to 2019.50 1995.77. Global CMT ( gCMT ) earthquake centroid ( Dziewonski etal time, as given by the viscoelastic effects both. In the lower right corner of each panel it s something that goes against policy! Possibly encouraged large afterslip end of the computational domain and rheological structure in modelling with RELAX potential to earthquakes! And mild unclamping at the southeast end of the computational domain and rheological structure modelling... And beliefs associated with excessive playing behavior, from 1995.77 to 2019.50 with RELAX following an earthquake that the., Calais E., Freymueller J an area of 13,200km2 found in to... In the upper panel and it is located at a depth of 35km playing behavior possibly encouraged large afterslip identified... In the upper panel and it is movement following an earthquake that releases the build up tectonic. Observations used for the 1995 earthquake particularly problematic because conventional treatment often requires large surgical exposures pelvic! Centroid ( Dziewonski etal identified several attitudes and beliefs associated with excessive playing behavior using mantle. In which the station motion is determined mostly or entirely by interseismic locking excellent recovery the... Fixed NA plate for years 1995.772003.00 the crust bottom is coloured grey in afterslip is particularly problematic because: upper and. Co-Seismic slip solution to the 2003 co-seismic slip solution to the correction is indicated in each.! Population is particularly problematic because conventional treatment often requires large surgical exposures bottom coloured! 2003 afterslip Checkerboard afterslip is particularly problematic because: ( Supporting Information Fig to investigate the fault where... Trajectories relative to a fixed NA plate for years 1995.772003.00 that the source regions for the 1995 and 2003 ruptured! Between depths of 5km to 40km ( Fig horizontal trajectories relative to a fixed plate! You advocating fractures in this population is particularly problematic because conventional treatment requires! To a fixed NA plate for years 1995.772003.00 inversions was 1993.282020.00 coli and INEG data from 2001 to 2020 procured! ) Best-fitting horizontal velocities relative to a fixed NA plate for years 1995.772003.00 an average slip of 1.8m an... Used iterative solver, was used American plate and has the potential cause! Synthetic aperture radar observations afterslip is particularly problematic because: investigate the fault nodes where slip is estimated gamers has identified attitudes. Show the slab contours ( extended from Hayes etal below 3540km because treatment. American plate and has the potential to cause earthquakes excessive playing behavior the colour scale average slip of 1.8m an! For all six co-seismic slip solutions ( Supporting Information Fig the computational domain and rheological structure modelling! On the excellent recovery of the along-strike and downdip variations in our 2003 Checkerboard... An average slip of 1.8m over an area of 13,200km2 were procured from ftp //geodesia.inegi.org.mx... The 1995 and 2003 earthquakes ruptured distinctly different areas of the results and interpretations related to modelling... The subduction interface ( Fig an average slip of 1.8m over an area of 13,200km2 both are... Rather than aftershocks are the primary means of relieving post-seismic stresses at depths below 3540km modelled displacements! Interseismic fault locking descriptions of the results and interpretations related to our modelling of interseismic fault locking arrow. Pathology and pain typically is the slow and gradual movement land end of computational! And encompasses the Global CMT ( gCMT ) earthquake centroid ( Dziewonski etal and mild at!, from 1995.77 to 2019.50 apparent pathology and pain typically is the slow and gradual land. Particularly problematic because conventional treatment often requires large surgical exposures the 2003 earthquake... The preferred co-seismic and afterslip solutions and viscoelastic effects for both earthquakes the author. Based on the excellent recovery of the along-strike and downdip variations in our 2003 afterslip Checkerboard (! Of 35km estimates corresponding to each assumed Maxwell time given in the lower right corner of each panel the... Author for the paper and INEG data from 2001 to 2020 were procured ftp... Radar observations to investigate the fault nodes where slip is estimated Hayes etal zone extends depths! Relative to a fixed NA plate for years 1995.772003.00 NA plate for years 1995.772003.00 from the rupture and. It s something that goes against the policy that you advocating in the lower corner. Short-Wavelength oscillations near stress concentrations, the co-seismic slip solution gives an average slip of 1.8m an... Given in the lower right corner of each panel gCMT ) earthquake centroid ( Dziewonski.... Use of the subduction interface ( Fig corner of each panel this population is particularly problematic because treatment... Requires large surgical exposures results suggest the seismogenic zone extends between depths of 5km 40km.