G33A-01
Earthquake cycle implications of the Jalisco/Colima GPS project, 1995-2007
This talk will summarize results from modeling of campaign and continuous GPS measurements in a 30+ station network that extends 350 km along the Pacific coast and 300 km inland in the states of Jalisco, Colima, and Michoacan. Measurements since March of 1995 have recorded in detail the Oct 9, 1995 M=8.0 Colima/Jalisco and Jan 22, 2003 M=7.4 Tecoman earthquakes, constituting the only earthquakes with magnitudes above 6.5 that have ruptured the northwestern end of the Middle America trench since 1932. Modeling of coseismic offsets for both earthquakes indicate that coseismic slip extended no farther downdip than the Pacific coastline, defining a shallower seismogenic zone than is typical for subduction zone. Near-term postseismic measurements demonstrate that both earthquakes were followed within hours to days by aseismic propagation of slip along the subduction interface to deeper areas beneath the continent, decaying logarithmically with time after the earthquakes. These are diagnostic of fault-centered afterslip, representing a transient frictional response to the earthquake. Measurements months to years after each earthquake clearly show protracted, decaying transient deformation, consistent with viscoelastic flow of the upper mantle and possibly lower crust to relax the seismically-elevated stresses at depth. The afterslip and viscoelastic transients are further superimposed on linear landward motions of the GPS sites, which are an elastic response to relocking of shallow seismogenic areas of the subduction interface. The site motions are poorly fit by models that exclude steady interseismic strain accumulation, afterslip, or viscoelastic transients, strongly suggesting that all three contribute to surface deformation. Interestingly, no episodic transient slip has been recorded in Jalisco or Colima since continuous GPS recording was initiated by INEGI in early 1993. Unlike the rest of western and southern Mexico, where ETS is pervasive and presumably caused by the existence of a wide region of relatively flat Cocos plate subduction, the absence of ETS in areas characterized by steeper Rivera plate subduction may indicate that the area of transitional slip associated with ETS is narrow or possibly non-existent along the northwesternmost 300 km of the trench. A new 10-station continuous network operated jointly by the Universidad de Guadalajara and the University of Wisconsin will improve the detection threshold for such transients if they occur.
G33A-02
Forearc Sliver Translation, a Lack of Arc-Normal Strain Accumulation, and Interplate Thrust Earthquakes: GPS Geodesy in Western Nicaragua
We have been investigating the kinematics of the Nicaraguan forearc using campaign GPS measurements of our geodetic network made over the last seven years (Turner et al., 2007). We currently have interseismic velocities for 18 campaign sites and have installed 10 additional sites in the backarc to investigate the nature of the transition from forearc sliver motion to stable Caribbean Plate motion. Our work focusing on the later issue is presented elsewhere at this meeting (Styron et al., 2007). Corrections for modeled coseismic offsets from the Jan. 13, 2001 Mw7.7 earthquake off the coast of El Salvador have been applied to our campaign site velocities. Some of our time-series are also strongly affected by coseismic and postseismic effects of the Oct. 9, 2004 Mw6.9 earthquake off of the coast of Nicaragua. The geodetic effects of this event are being removed from the affected time-series for interseismic velocity analysis. We have also derived interseismic velocities for five continuous GPS sites in the region. Our GPS results confirm previous predictions of northwest transport of a forearc sliver with an average Northwest velocity of ~15 mm yr-1, but show little evidence for an arc- normal component of strain accumulation associated with locking on the subduction interface. However, the amount of seismicity along this section of the Middle America Trench, including several recent large events such as the 1992 Mw7.6 and 2004 Mw6.9 earthquakes, indicates some amount of locking is present. Several possibilities may account for the apparent contradiction between the GPS results and observed seismicity. The locked zone may be too shallow and too far offshore for the arc-normal component to show up in our network, or the arc-normal signal may be masked by post-seismic effects from the 1992 offshore earthquake. If coupling between the downgoing slab and the overriding plate is weak or limited to a small seismogenic zone, then arc-parallel motion of the forearc sliver may be driven by a more strongly coupled region to the south in Costa Rica. We are developing regional FEM and half-space dislocation models constrained by the GPS-derived kinematics to distinguish between these possibilities and explore the possible driving mechanisms for sliver motion.
G33A-03
Definition and Kinematics of the Nicaraguan Forearc and Stable Caribbean From GPS Geodesy
The Central American republic of Nicaragua's western margin is characterized by oblique subduction of the Cocos Plate, producing a volcanic arc in the overlying Caribbean Plate (DeMets, 2001). The arc sits in a structural depression with active faulting and related seismicity. Additionally, shear caused by oblique subduction requires accommodation in the forearc via elastic deformation or translation of a rigid forearc sliver. Recently reported GPS geodetic observations indicate nearly rigid forearc translation of ~15 mm yr -1 northwestward along the coast (relative to the stable Caribbean geodetic reference frame), with a small trenchward velocity component seen in the coastal sites (Turner et al., 2007). Here we report additional campaign GPS observations whose primary goal is to determine the nature of the boundary between the actively deforming forearc and the stable Caribbean plate to the east. To achieve this and to constrain rates of crustal deformation along this western boundary of the forearc, we have recently installed and occupied an arc- normal transect of 8 geodetic sites from the coast, through the Nicaraguan Depression, and into the stable Caribbean Interior Highlands (completed in early 2007). Along the transect, the new and existing sites are spaced 10- 15 km apart; some of these sites were initially installed in 2000 as part of our regional network. This transect is augmented by a less dense but more broad network of sites throughout western Nicaragua to observe changes in deformational style along strike.
G33A-04
Aseismic Slip Observed on the Faults in Mexicali Valley, Baja California, Mexico.
The Mexicali Valley, which is part of the Salton Trough, is located within an active tectonic region, in the boundary between the Pacific and North American plates; a region featuring a wide zone of transform faults associated with San Andreas fault system, and a zone of distributed deformation in the pull-apart center of Cerro Prieto. Since 1996, geotechnical instruments have operated in the Valley, for continuous recording of deformation phenomena. To date, the network includes three crackmeters, eight tiltmeters, and seven piezometers installed in the shallow aquifer; all instruments have sampling intervals in the 1 to 20 minutes range. The mainly vertical displacement at Saltillo fault (known before as southernmost part of Imperial fault) has been measured on a continuous basis since February 1996 by a crackmeter installed in Ejido Saltillo (Glowacka 1996; Nava and Glowacka, 1999). In 1998 a tiltmeter was installed on the fault and a second crackmeter , in a horizontal direction, about 60° from the fault strike, was installed about 1 km south of ES. Another crackmeter in the vertical plane crossing Morelia fault and a 3D Witness on the Cerro Prieto fault were installed in 2004. In 2003 seven digital water level meters were installed at depths up to 500 meters in the local piezometric wells, with the purpose of recording aquifer level changes. The observed deformation rates measured on the Saltillo fault are 6 cm/yr and 2 cm/yr for the vertical and horizontal components, respectively and occurs in steps (creep events), separated by months of quiescence, and large events account for 70 percent of the vertical displacement. Aseismic creep events have amplitudes of 1-3 cm and durations of 1-3 days. Much less precise measurements from the Cerro Prieto fault show vertical deformation on the fault with velocity around 3.1 cm/yr , and the rate does not depend on the local seismicity. Horizontal rates change slightly with the presence of seismicity when very small right lateral deformation (few mm/yr) can be observed. From the delay measured on different instruments installed on the Saltillo fault one can conclude that creep events have an apparent migration velocity of the order of 4 cm/s. Comparison between occurrence of creep events and the water level record done for the piezometer installed nearby, shows that in four cases a water level anomaly is associated with a creep event. We discuss some possible relation between creep phenomenon and water level changes, and tectonics, geology and anthropogenic activity in the Mexicali Valley. This research is sponsored partly by the CONACYT project 45997-F.
G33A-05
Slow slip events in the Mexican subduction zone (Guerrero-Oaxaca) observed by the analysis of continuous GPS
Aseismic slow slip events (SSE) have been observed in the last decade with GPS, in particular in the Mexican subduction zone, Guerrero-Oaxaca. Thanks to the effort of the UNAM (Universidad Nacional Autonoma de Mexico), 21 permanent GPS stations are currently operating within or close to this zone. We present the results of an analysis of the available continuous GPS data that we have conducted using the Gamit/Globk softwares developped by MIT (Massachusetts Institute of Technology) and SIO (Scripps Institution of Oceanography). The analysis includes up to date atmospheric loading models and mapping functions, what helps us to gain more precision in the positioning results. We have analyzed the complete time series running from 1997 (first GPS station installed in Cayaco) up to present. One of our major objectives is to study the time evolution of the slow slip events and to determine the amplitude and the direction of the slips. A first major result from our analysis is that we do not find in the time series any clear evidence of a number of the smallest transient slip events (i.e., total displacements of 1-2 cm at the coast) that had been reported before in Guerrero-Oaxaca and we suggest that some of them may actually be artefacts.
G33A-06
Seismicity variations associated with aseismic transients, Guerrero, Mexico, 1995- 2006
Aseismic transients with no obvious relation to large earthquakes are a newly recognized mode of deformation along subduction zones, transform plate boundaries and mega-landslide faults. Stressing in the up-dip seismogenic zone is increased due to down-dip slow slips, and it can be made more prone to failure in a large thrust earthquake. Thus, it is important for seismic hazard assessment, to search and identify patterns of spatiotemporal seismicity variation associated with transients and, in some cases, tremors. The Guerrero, Mexico, region is chosen for this study, because of the long-term (over 10 years, continuous and campaign) geodetic observation and abundance of seismicity in the shallow subduction zone. We search the GCMT (Global Centroid Moment Tensor) and NEIC (National Earthquake Information Center) catalogs for seismic events between 1995 and 2006 within the area of latitude 16 to 20N and longitude 98 to 102W, which covers the region affected by major transients since 1996. Earthquake depth is limited to be above 100 km, to roughly include events related to the shallow subduction process. A completeness magnitude of Mc ~ 4.5 is determined, for the NEIC events, based on the maximum likelihood method. Three large transients in 1998, 2001-2002 and recently 2006 [K. Larson, this meeting] are all temporally correlated with higher seismic rates in the studied area. Particularly, we found that transients are either preceded by normal-faulting earthquakes, presumably in the subducting slab given their large epicentral distance from the trench, or followed by shallow thrust-faulting earthquakes close the trench, although the GCMT and NEIC epicenters may vary by tens of km. In some cases, such as the transient from October 2001 to April 2002, both clusters are found bordering the transients, suggesting that the transient may act as a connection in stress transfer [much like the deep slip mechanism for transfer by Dmowska et al. (1988)]. The beginning of the 2006 transient coincided with two normal-faulting earthquakes in February and March, 2006. Thrust earthquakes are found to follow the 1996 and 1998 transients. Although so far no thrust earthquakes have been reported in GCMT after the 2006 transient, the NEIC catalog does show a cluster of seismicity with epicenters close to the trench, thus up-dip in the seismogenic zone. We also perform numerical simulations in the framework of rate and state-dependent friction to model subduction earthquake sequences and aseismic transients. When the normal-faulting earthquake is treated as a stress perturbation to the subduction interface, sequential aseismic transients can be triggered, and the timing of the next large thrust earthquake is affected by three factors, namely, when, where and how large is the perturbation.
G33A-07
Nonvolcanic tremors in the Mexican subduction zone
Nonvolcanic low frequency tremors (NVT) have been discovered and studied recently in Japan and Cascadia subduction zones and deep beneath the San Andreas Fault. The tremors activity is increasing during so-called silent earthquakes (SQ) in Japan and Cascadia. NVT clusters also migrate following the propagation of the SQ. The origin of the NVT is still unclear. The studies of NVT and SQ in different subduction zones are required to understand the cause for these phenomena. We discovered a number of NVT from daily spectrograms of continuous broad band records at seismic stations of Servicio Seismológico Nacional (SSN) an MASE project. The analyzed data cover a period of 2001-2004 (SSN) when in 2002 a large SQ has occurred in the Guerrero- Oaxaca region, and a steady-state interseismic epoch of 2005 and a new large SQ in 2006 (MASE). NVT occurred in the central part of the Mexican subduction zone (Guerrero) at approximately 200 km from the coast. We can not accurately localize the tremors because of sparse station coverage in 2001-2004. The MASE data of 2005-2006 show that NVT records in Mexico are very similar to those obtained in Cascadia subduction zone. The tremors duration is of 10-60 min, and they appear to travel at S-wave velocities. More than 100 strong NVT were recorded by most of the MASE stations with the epicenters clustered in the narrow band of ~40x150 km to the south of Iguala city and parallel to the coast line. NVT depths are poorly constrained but seem to be less than 40 km deep. We noticed a some increase of NVT activity during the 2001-2002 and 2006 SQs compared with an NVT activity for the "SQ quiet" period of 2003-2004 nevertheless. A lack of NVT for the period of 2-3 months after the SQ is apparent in 2002 and 2006.