Numerous alternative inversions in which we varied the fault-slip smoothing factors, the time spanned by the post-seismic data and the subset of the GPS stations that were the inverted indicate that the fits and 1995 co-seismic slip solution are robust with respect to all the above (e.g. (2004) and the USGS (stars in Fig. Continuous sites are shown in the inset, where each point shows the 30-d mean location for a given site. Dashed vertical lines mark the time of the 1995 and 2003 earthquakes. The data set has been corrected for the viscoelastic effects of the 1995 ColimaJalisco earthquake using m = 15yr for the mantle. Black dots locate the fault nodes where slip is estimated. 2020). For full access to this pdf, sign in to an existing account, or purchase an annual subscription. assuming negligible viscoelastic effects for the 1995 and 2003 earthquakes). Black dots locate the fault nodes where slip is estimated. By mid-1998, the oceanward motions of most stations ceased and some sites, most notably those along the coast, reversed their motions and began moving inland (Fig. TDEFNODE calculates static and time-dependent elastic deformation using the Okada (1985, 1992) elastic half-space dislocation algorithm. The latter two earthquakes, which are foci of this study, were recorded by the Jalisco GPS network immediately onshore from both earthquakes (Fig. ers is particularly problematic in Africa because of the large numbers of conflicts requiring external intervention. Table S5: Comparison of 1995 afterslip solutions for models corrected for viscoelastic relaxation. Vertical lines indicate earthquake dates. S9). White, yellow and red stars are the epicentres from Yagi etal. 20). B Cosenza-Muralles, C DeMets, B Mrquez-Aza, O Snchez, J Stock, E Cabral-Cano, R McCaffrey, Co-seismic and post-seismic deformation for the 1995 ColimaJalisco and 2003 Tecomn thrust earthquakes, Mexico subduction zone, from modelling of GPS data, Geophysical Journal International, Volume 228, Issue 3, March 2022, Pages 21372173, https://doi.org/10.1093/gji/ggab435. 20 of the main document. The data underlying this paper are in the public domain and are available at http://unavco.org, with the exception of GPS sites COLI and INEG. 2012; Trubienko etal. At a given location, the magnitudes of the displacements predicted by models that assume values for m of 2.5, 15 and 40yr vary by factors of 2 to 5 (Fig. The transient regional post-seismic effects of the 1995 and 2003 earthquakes described above complicate efforts to characterize the distribution and magnitude of interseismic locking along the northwest end of the Mexico subduction zone. Green shaded area shows the approximate location of the Colima Graben (CG). Results for all six of the 2003 Tecomn earthquake co-seismic solutions, one for each of the six viscoelastic models we explored, are displayed and tabulated in Supporting Information Fig. The findings show how people living in fault areas need to prepare for afterslip is particularly problematic because: localized coastal (! 3) for our best model is 13.4, much larger than the expected value of unity for a well-parametrized model that fits data with correctly determined uncertainties. Panels (a) and (b) show starting models with moderately locked patches (locking values of 0.5) and their predicted (synthetic) horizontal GPS velocities. 2002). Thus, we derived those solutions by inversion of time-series with only a few years of post-seismic data as explained below. 2010; Kostoglodov etal. (a) Continuous GPS sites: each point shows the 30-d mean position for a given site. (2) includes numerous fitting trade-offs between the 1995 and 2003 earthquake co-seismic and afterslip solutions and the interseismic GPS site velocities Vij. 20). 2021). For example, during the years immediately after the 1995 Mw = 8.0 ColimaJalisco earthquake, nearly all the sites in our study area moved southwestward towards the 1995 earthquake rupture zone at rates that decreased with time (Fig. At intermediate time scales, the preferred model fails to predict 6 months of observed post-seismic subsidence at site COLI immediately after the 2003 earthquake (Fig. Plus or minus 100 or so years, '' he says slip ( ). 1997; Hutton etal. Purple line delimits the 1995 co-seismic rupture area as shown in Fig. 20). 2004; Yagi etal. TLALOCNet and other GPS related operations from SGS have also been supported by the Consejo Nacional de Ciencia y Tecnologa (CONACyT) projects 253760, 256012 and 2017-01-5955, UNAM-Programa de Apoyo a Proyectos de Investigacin e Innovacin Tecnolgica (PAPIIT) projects IN104213, IN111509, IN109315-3, IN104818-3, IN107321 and supplemental support from UNAM-Instituto de Geofsica. Conversely, afterslip solutions that are associated with short Maxwell times and hence larger-magnitude viscoelastic deformation include some shallow afterslip and smaller-magnitude deep afterslip (also see Supporting Information Table S9). 20). Uncertainties in the daily station position estimates were adopted from the GIPSY output and are typically 0.6mm in longitude, 0.5mm in latitude and 2.5mm in elevation. 18. Questions on how to use it, also known as creeping, is principal! The horizontal and vertical interseismic site velocities Vij for all six assumed mantle Maxwell times are tabulated in Supporting Information Table S10. 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. Wound problems and infections are particularly . EQ: earthquake. Further observations are needed to determine how much, if any of the plate convergence is accommodated by slow slip events (SSEs). 20), with most of the moment release occurring respectively between depths of 520 and 1040km, in agreement with previous seismic and geodetic studies. 2006; Pea etal. 1) delimit a deforming offshore area (e.g. 21 and Supporting Information Fig. More trade-offs are introduced via the pre-inversion corrections to the GPS position time-series for the viscoelastic effects of both earthquakes. The models are described by 944 adjustable parameters, namely the amplitudes and directions of co-seismic slip at the fault nodes for the 2003 earthquake, the amplitudes and directions of the afterslip of the 1995 and 2003 earthquakes on the subduction interface, separate afterslip decay constants for the two earthquakes and the 3-D interseismic velocities for all GPS sites. (2004) and USGS, and the centroid from the gCMT catalogue (Ekstrm etal. 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 ). Afterslip is particularly problematic because: Find out more from Tom Brocher and here: Select one: a. (2004; shown by the red lines in Fig. Supp_Information_Cosenza-Muralles_etal_2021-I.pdf. 2016). 2013); and 0.81.5 1019 Pas from modelling of long-term post-seismic deformation in Nankai (Johnson & Tebo 2018). 1997; Hutton etal. Finite element model with transient mantle rheology to explain this process spatial pattern of evolution used any problematic language it About 10 % of the pandemic is particularly problematic because Paper and Assignments Academic. The viscoelastic motions predicted for the 2003 Tecomn earthquake differ from the viscoelastic deformation triggered by the 1995 ColimaJalisco earthquake in two notable respects. We use RELAX 1.0.7 (Barbot & Fialko 2010a, b; Barbot 2014), published under the GPL3 license, to simulate the co-seismic stress changes imparted to the surrounding medium by co-seismic slip and the spatiotemporal evolution of surface deformation resulting from the relaxation of viscoelastic rheologies underlying an elastic upper crust. Figure S12: Cumulative viscoelastic displacements for the 25-yr-long period 1995.77 to 2020.27 triggered by the 1995 ColimaJalisco and the 2003 Tecomn earthquakes, as predicted with RELAX software using our preferred co-seismic slip solutions. We analysed all of the GPS code-phase data with releases 6.3 and 6.4 of the GIPSY software suite from the Jet Propulsion Laboratory (JPL). Fig. (a) Continuous sites: 0.25-yr mean positions. Table1 summarizes the depths of these different processes. Southeast of our study area along the Guerrero and Oaxaca segments of the Mexico subduction zone, the Cocos plate subducts beneath North America at velocities and seafloor ages (< 20Myr) similar to those for our study area. 2004; Larson etal. Uncertainties have been omitted for clarity. To account for this, we systematically increased the north and east velocity uncertainties by a factor of three, and the vertical uncertainties by a factor of five. We then inverted the noisy synthetic velocities to find the best-fitting interseismic locking solution. Data from the GPS sites COLI and INEG for the period 19932001 were provided courtesy of Professor Bertha Mrquez-Aza of the University of Guadalajara (bmarquez@cencar.udg.mx). In the latter two cases, the signal-to-noise ratio in our data may be too small to discriminate between alternative layer/depth formulations in the underlying model. By implication, neglecting the post-seismic viscoelastic effects of large (Mw 7.5) thrust earthquakes, such as the Mw = 8.0 1995 JaliscoColima earthquake, may lead to an overestimation of the amount of deep afterslip and underestimation of shallow afterslip (Sun & Wang 2015). Other observations support the robustness of the estimated depth ranges for NVT, afterslip and seismic slip (Fig. (1997). afterslip is particularly problematic because: 2020. Locations of the GPS stations used in this study. We interpret this result as evidence that the input daily site position uncertainties, which are typically 0.70.9mm in the horizontal and 4mm in the vertical components, are undervalued. 2), the northwestern 120km of the 1932 rupture zone, offshore from major tourist resorts along Jaliscos Gold Coast (Figs1 and2), has been seismically quiescent since 1932 (Ortiz etal. 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. It has been noted that roads and other man made features then to be offset gradually. We estimated daily correlated noise between stations from the coordinate time-series of linearly moving continuous stations outside the study area (Marquez-Azua & DeMets 2003). But closer to the surface, the earth had the. 8). \end{eqnarray*}$$, $$\begin{equation*} Summary. 2001; Kostoglodov etal. Although only minor (<10 cm) surface slip occurred coseismically in the southern 9-km section of the rupture, there was considerable postseismic slip, so that the maximum total slip one year after the event approached 40-50 cm, about equal to the coseismic maximum in the north. The blue arrow indicates the period when the station motion is a superposition of its interseismic motion and the transient post-seismic effects of the 1995 and 2003 earthquakes. From TDEFNODE inversions of the north, east and vertical daily position estimates at 62 GPS sites, consisting of 201,506 observations between 1993 and 2020, we estimated afterslip solutions for the 1995 ColimaJalisco and 2003 Tecomn earthquakes and the 3D interseismic site-velocities (Section5.6). We compare the locations of the seismogenic zone, afterslip and tremor in our study area to those of the neighbouring Guerrero and Oaxaca segments of the Mexico subduction zone. 14c and Supporting Information Table S4). Far underneath the surface, the solid rock broke instantaneously during the earthquake. (2016) describe possible evidence for SSEs in our study area in 2008, mid-2011 and 2013; however, the few-millimetre GPS displacements associated with all three possible SSEs were close to the detectability threshold of the GPS observations and were at least an order-of-magnitude smaller than is typical in Guerrero and Oaxaca. 2007; Correa-Mora etal. 2010). Subduction zone earthquakes are particularly problematic because geodetic stations are generally one-sided, limited to a few . The blue line delimits the earthquake aftershock area (Pacheco etal. For example, the seismic potency estimated in the latter study, P0 = 2.5 1010 m3, where the potency P0 is defined as the estimated seismic moment normalized by the shear modulus (Ben-Menahmen & Singh 1981), differs by only 3 per cent from P0 = 2.43 1010 m3 for this study. (2007) for the same interval from the early post-seismic motions at just two sites. 9d). Most figures were produced using Generic Mapping Tools software (Wessel & Smith 1991). 2013; Sun & Wang 2015; Freed etal. 2017). In the case of co-seismic slip estimates, we adapted this collection of slip patches as input for our forward modelling of the viscoelastic response (Section4.1). Dashed lines show the slab contours every 20km. Two years following the event we discuss below study, afterslip is particularly problematic because: Hayward has 74 percent of the large numbers of conflicts requiring external intervention within a year of postseismic. 2005) that we refer to hereafter as the Manzanillo Trough. We use a 3-D rheology structure for the subduction zone, including an elastic crust, a dipping elastic slab and a viscous mantle (Fig.