1
Propagation of the 2001-2002 aseismic slow slip event and an interplate
coupling in the Oaxaca subduction zone, Mexico
Franco Sánchez, S.I.1, V. Kostoglodov1, K.M. Larson2, V.C. Manea1, M. Manea1, and J.A.
Santiago1
1
Institudo de Geofísica, Universidad Nacional Autónoma de México, Ciudad Universitaria, Del.
Coyoacan, 04510 Mexico D.F., Mexico
2
Department of Aerospace Engineering Science, University of Colorado, Boulder, CO 80309-
0429, USA
Abstract
A GPS (Global Positioning System) network has been established in the Mexican state of
Oaxaca with the aim to understand the subduction-related seismotectonic regime in the area.
Permanent GPS stations have been operating since 2000. Data from campaign GPS sites have
been acquired every year for the period between 2001 and 2004. The observed average
displacements corresponding to the epochs of 2001-2002 and 2002-2003 differ significantly
from the crustal displacements for the apparent interslip steady state phase, the epoch of 20032004. Continuous GPS stations in Guerrero (the neighboring state to the NW of Oaxaca)
recorded a large slow slip event (SSE, or silent earthquake) in 2001-2002 and a smaller SSE in
2002-2003. At least two continuous stations of the Oaxaca GPS network registered a propagation
of 2001-2002 slow slip event from Guerrero to Oaxaca. The anomalous displacements observed
on the campaign GPS sites in Oaxaca may be attributed to the effect of the 2001-2002 slow slip
event. The average 2001-2002 and 2002-2003 epoch displacements are the sum of the interslip
2
steady state and the slow aseismic slip displacements. The latter can be modeled by a
displacement pulse propagating along the Guerrero-Oaxaca coast. The model helps to identify
the undisturbed interslip epochs at each GPS site which can be used to estimate the interplate
coupling in the Oaxaca subduction zone.
Introduction
Subduction of the Cocos oceanic plate beneath the North America and Caribbean plates
in the Middle America Trench determines the seismotectonic regime along almost the entire
Pacific coast of Mexico (Figure 1). In order to understand the deformation processes in the
Mexican subduction zone, a GPS network has been established in the states of Guerrero and
Oaxaca. At the present time the network consists of 13 permanent and 42 campaign stations
(http://tlacaelel.igeofcu.unam.mx/ ~vladimir/gpsred/gpsred.html).
The Guerrero GPS network has already recorded four transient slow slip events. The first
1995-1996 slow slip event (SSE) was evidently associated with the 1995 Mw=7.3 Copala
earthquake (Larson et al., 2004). The second SSE (Lowry et al., 2001) occurred in 1998 in the so
called Guerrero seismic gap. This event was unambiguously recorded by only two permanent
stations: CAYA and POSW (http://h4.colorado.edu/gps_guerrero.html). The largest aseismic
slow slip event started in October 2001 in the same zone as the 1998 event. It lasted for about six
months, and covered an area of ~550 x 250 km2 (Kostoglodov et al., 2003). The equivalent
magnitude of this slow slip event was estimated as Mw=7.5 with a maximum modeled
displacement on the plate interface of ~22.5 cm (Iglesias et al., 2004). A small slow slip event
was recorded in Guerrero at the end of 2002.
3
All permanent GPS stations in Guerrero, one station in Morelos, one station on the
Popocatepetl volcano and two stations located in the neighboring state of Oaxaca recorded the
2001-2002 transient slip event. CAYA, located at the central part of the NW Guerrero seismic
gap (Figure 1), recorded the maximum aseismic crustal displacement (~5.2 cm south-southeast
lateral and ~4.4 cm uplift components) produced by this event. The time series registered by
other GPS stations show a systematic delay of the initial phase of this silent earthquake (Figure
2).
The campaign GPS data obtained in Oaxaca for the epochs of 2001-2002-2003 show
abnormal displacements, different from the expected displacements for a period of steady state
interslip strain accumulation (long-term deformation in the period between two subsequent SSE).
This observation is interpreted as a disturbance effect produced by the slow slip propagation
along the Oaxaca subduction zone. A simple model of the transient slip propagation along the
Guerrero-Oaxaca subduction zone allows us to trace a step-like slow displacement signal at each
campaign GPS site and to determine the intact epochs appropriate for the interslip steady state
velocities estimates.
Apparently the interslip period started in Oaxaca in 2003, after the 2001-2002 SSE has
run out. So the 2003-2004 GPS campaign data can be considered as belonging to the current
steady state epoch. The campaign sites velocities for steady state epochs then can be modeled to
assess the interseismic coupling on the subduction interface in Oaxaca.
Seismotectonic setting
The convergence of the Cocos and the North American plates along the Middle America
trench (MAT) is responsible for the crustal deformation and large subduction thrust earthquakes
4
in the Mexican state of Oaxaca. The relative convergence rate in the MAT increases from 5.9
cm/yr at the Guerrero-Oaxaca boundary to 6.4 cm/yr at the southeast of Oaxaca (DeMets et al.,
1994). The age of the Cocos plate in the Oaxaca segment of the MAT is less than 16 Ma
(Kostoglodov and Bandy, 1995). The plate interface is apparently subhorizontal in Guerrero
(Kostoglodov et al., 1996) and the NW Oaxaca (Torres Zamudio, 2002), gradually changing to
steep subduction geometry at the SE of Oaxaca (Pardo and Suarez, 1995).
The characteristic thrust earthquakes with an average magnitude of M w ~7.5 (Singh et al.,
1983) are fairly frequent in Oaxaca (Singh et al., 1981). Figure 1 shows estimated rupture areas
of recent large thrust earthquakes occurred in this zone since 1960. Previous work concluded that
the plate interface is strongly coupled and appears to extend downdip for about 80 km (e.g.,
Singh et al., 2000). An important factor which may influence the seismic activity, crustal
deformation and interplate coupling is the subduction of the ocean floor bathymetric
irregularities (Kostoglodov and Ponce, 1994; Pardo and Suarez, 1995).
The subduction zone coupling should depend on the depth and configuration of the plate
interface, and vary as well with the timesince the last large subduction earthquake.
Data acquisition and processing
The GPS network in Oaxaca includes 3 permanent stations and 16 campaign sites. The
campaign stations are located mainly along the coast (Figure 1). In the Guerrero state the GPS
network, at the time of this study, encompassed 4 permanent stations (CAYA, ZIHP, ACAP and
IGUA). For this work we use continuous records from the YAIG and POSW continuous GPS
stations to the north of the Guerrero state (Figure 1). The POSW station located on the
5
Popocatepetl volcano, about of 360 km from the Pacific coast, has also recorded the slow
aseismic slip event occurred in Guerrero in 1998.
All campaign surveys in Oaxaca have been carried out during the dry season at the same
time period, in March-April of 2001-2004 to avoid the effect of variable weather conditions. The
measurements at most of the sites were carried out continuously for 72 hours. At a few sites in
2001, occupations were 8 hours per day during three consecutive days. The sampling rate is 30
seconds..
All GPS data presented in this paper were analyzed using the GIPSY software [ (Lichten
and Border,1987) which uses both GPS observables: carrier phase and pseudoranges.
Data processing was performed using a Precise Point Positioning (PPP) strategy, which analyzes
one station at time (Zumberge et al., 1977). Precise clocks and orbits are provided by JPL. The
GPS station coordinates are first estimated in a non fiducial reference frame (Heflin et al., 1992)
and then are transformed into ITRF2000 (International Reference Frame 2000, Boucher et al.,
2003) using Helmert transformation parameters supplied by JPL.
The velocities at the campaign sites were estimated for each epoch using the weighted
linear best fit to the daily positions of the two successive occupations. These velocities were then
converted form the ITRF2000 into the North America (NA) reference frame using the net NA
velocity in the ITRF2000 realization of Altamimi et al. (2002). Campaign data are not sufficient
to determine seasonal effects and antenna blunders. These effects should be considered in the
uncertainty estimates. Mao et al., (1999) showed that the contribution of white, flicker, and
random walk noise components into the uncertainties of the permanent GPS records is
significant. To scale the formal standard errors of the Oaxaca campaign data we accepted a
6
reasonable average scaling factor of 2 for both longitude and latitude position components
following Marquez-Azua and DeMets (2003).
Results
At the end of 2001, the permanent GPS stations CAYA and ACAP recorded the initial
phase of large SSE in Guerrero (Kostoglodov et al., 2003). The station CAYA, located in the
middle of the Guerrero seismic gap, was apparently the closest station to the “epicenter” of this
silent earthquake. The CAYA record reveals the highest uplift and horizontal displacement (~4.2
and ~5.2 cm, respectively, Yoshioka et al., 2004) produced by this SSE. Other permanent GPS
stations detected a change from the steady state interslip motion to the slow slip rebound motion
some time (up to ~3 months) later than the CAYA station (Figure 2). This observation suggests a
relatively slow lateral propagation of this SSE.
The results of the 2001-2002 and 2002-2003 epochs (Figure 3 A,B) compared with those
for the 2003-2004 epoch (Figure 3C) clearly show (Figure 3 D,E) the effect of the slow slip
propagation on the displacement vectors. For the consistency of results we presented the
displacements on the permanent GPS stations as those would be the campaign sites occupied at
the time of the Oaxaca surveys. Comparing these GPS epochs and the direction of the plate
convergence vectors one can envisage that the slow slip has perturbed the displacement vectors
in Oaxaca only down to the longitudes ~98-97°W during the 2001-2002 epoch, while during the
2002-2003 epoch the main influence of the propagating slow aseismic slip was in the area to the
east of 97°W. The 2003-2004 epoch apparently represents the steady state interslip deformations
corresponding to the recovered partial coupling along the plate interface. There is probably some
7
influence of the 2002-2003 minor magnitude SEE on the velocity estimates in Guerrero (Figure
3).
To assess the propagation characteristics of the 2001-2002 SSE we need to estimate the
initial time, duration and attenuation of the event pulse at each permanent GPS station. These
parameters can be easily determined for semi-continuous time series (e.g., for the ACAP, CAYA
and PINO stations) using parameterized hyperbolic tangent fit (Lowry et al., 2001) or subtracting
the linear regression estimates for the steady state interslip intervals in the cases where the GPS
stations have noticeable time gaps.
The propagation parameters of the 2001-2002 silent earthquake can be assessed from the
initial time, duration and attenuation versus distance plots presented in Figures 4 and 5. The
propagation rate, VSSE, of the slow slip is notably anisotropic, with the VSSE ≈ 5 km/day in the
direction perpendicular to the coastline and trench, and VSSE = 2.3 ± 0.1 km/day southeast along
the coast. Northwest along the coast the propagation rate seems to be slower, VSSE ≈ 1.2 km/yr.
The physical mechanism of this slow slip propagation is not yet clear, so the further study of
VSSE anisotropy may give us valuable information.
The geometric southeast lateral (along the coastline) attenuation of the displacement can
be approximated by an exponential function of the distance (Figure 5). The amplitude of slow
slip displacement decreases for more than 5-6 times at the distance of ~400 km. The duration of
the aseismic slip reduces nearly linearly with the distance as ~510-4 yr/km, so that at the distance
of ~375 km between the CAYA and OAXA stations the duration decays from ~6 months to
almost 4 months.
The slow event was still ongoing when the March 2002 Oaxaca GPS campaign was
carried out. The permanent stations CAYA and ACAP had already recorded over a half of the
8
slow event period. At the permanent station PINO located about 240 km southeast from the
CAYA station, the event was just starting. Permanent stations located further (than PINO) away
from CAYA have not yet recorded the slow event at the time of the 2002 campaign. The stations
ZIHU and OAXA started to record the aseismic slip only about one month after the Oaxaca
campaign (Figure 2).
During the 2001-2002 epoch, the displacement vectors of the Oaxaca temporary stations
located close to the Guerrero state did not agree with the direction of expected interseismic
steady state strain accumulation, which is congruent with the Cocos-North America convergence
(Figure 3a). The campaign stations located at the central part of Oaxaca show a different
behavior: the displacement vectors and the convergence vector are almost parallel. The campaign
stations located in the southeastern zone of Oaxaca show a larger east-west displacement,
suggesting an influence of the oblique convergence in this area.
During the 2002-2003 epoch, the direction of the displacement vectors has changed in all
stations. The stations in and close to Guerrero started to recover the interslip steady state motion,
while the displacement at the rest of the Oaxaca GPS sites deviated to the SE from their
displacement direction recorded in the 2001-2002 campaign. This observation indicates that the
slow slip event started in 2001 in Guerrero has apparently propagated down to SE of Oaxaca in
the beginning of 2003. Even though the SE propagation rate of the slow event of ~2 km/day
(Figure 4) is in favor of this hypothesis, it is necessary to verify it by tracing the slow event
displacement pulse at each campaign GPS site.
Assuming that the velocities during the interslip steady state period of elastic strain
accumulation is overall comparable at every campaign site in Oaxaca to that observed at the
PINO permanent station, we can use the curve approximating the latitude time series of PINO as
9
a pattern signal. Extrapolation of this step-like signal into the position of each campaign GPS site
is done applying estimated time delay, attenuation and duration relations. The modeled slow
event pulse migration (Figure 6) is fairly consistent with the time positions of the campaign sites
in the northwestern Oaxaca (sites COPA-ROBL in Figure 3 and shaded charts in Figure 6). The
velocity vectors at PAST and ROBL sites have not changed their directions from 2001-2002 to
2002-2003 epochs. Even though the pulse model still fits the 2001-2001-2003 positions at these
sites, we can not exclude the possibility that the pulse fit at PAST and ROBL sites is an artifact
and the SSE just ceased in this area. This conclusion is supported by a preliminary report of the
velocity measurements at a local GPS network concentrated between PUAN and ROBL (Cabral
et al, 2003).
The modeled pulse propagation in the southeast of Oaxaca (Figures 3) does not fit at all
the GPS sites positions (sites PUAN-SACR in Figure 6). The slow event in this area is amplified
and somewhat delayed (see HUAT record in Figure 6) compared to the modeling. This
observation does not agree with a simple slow pulse propagation model and suggests an
occurrence of a separate SSE, which has probably been triggered by large 2001-2002 SSE
propagated to Oaxaca from Guerrero (Figure 7). The secondary SSE originated in the transition
zone from the subhorizontal subduction plate interface in Guerrero-Oaxaca to normal steep
subduction interface geometry in Chiapas, and possibly causally related with this transition zone.
The main 2001-2002 SSE caused south-southwestward displacements of the GPS sites whereas
the secondary SSE in Oaxaca produced a predominant east-west displacement.
Configuration of the plate interface
A configuration of the interplate contact is an important constraint for the coupling
modeling. Based on the rough estimates of the subducted plate geometry (Pardo and Suárez,
10
1994, Torres Zamudio, 2002) one can perceive two areas with a distinct geometry of the plate
interface: 1- Central Oaxaca, similar to Guerrero, and 2- Southeastern Oaxaca and the Isthmus of
Tehuantepec, which is a transition from a subhorizontal subduction to “common”, steep
subduction in Chiapas. The boundary between these two areas passes in SW-NE direction,
approximately through 96°W (flanked by Puerto Angel and Huatulco, PUAN and HUAT in
Figure 3).
A recent study of the Wadati-Benioff zone between Oaxaca and Chiapas (Bravo et al.,
2004) based on data from a local seismic network provides better insight on the plate interface
geometry. Seismicity cross sections D-E (Figures 7, 8, 11 in Bravo et al., 2004) can represent an
average configuration of the subducted slab in the second area (see Figure 10B). The estimates of
crustal thickness in the eastern Oaxaca and the Isthmus of Tehuantepec are ~35 km (Ligoría and
Ponce, 1993).
A combined 2D gravity and magnetic modeling as well as the intraslab seismicity are
used to constrain the slab geometry in the Central Oaxaca area (Figures 8). The gravity data are:
onshore from De la Fuente et al., 1994, offshore from GEODAS v4.0. The magnetic anomalies
are MAGSAT data reduced to the pole (RP) (Hinze et al., 1982). The configuration of the slab
accepted in the present modeling is similar to the gravity models of Guerrero (Kostoglodov et al.,
1996). For consistency, we used the same densities as in the paper of Kostoglodov et al. (1996):
upper mantle 3295 kg/m3, top of the oceanic plate (basalt and gabbro) 2700 and 2900 kg/m3,
oceanic lithosphere 3340 kg/m3, continental upper crust 2670 kg/m3, continental lower crust
3050 kg/m3. We also incorporated the transition from basalt to eclogite at ~70 km depth, which
is consistent with the recent thermal and metamorphic models for Guerrero (Manea et al., 2004).
For the eclogite we used a density of 3300 kg/m3.
11
The shape of the MAGSAT magnetic anomaly over subduction zones was attributed to
the magnetization contrast between the cold subducting oceanic crust (below Currie point of
~550C) and surrounding non-magnetic mantle (Clark et al., 1985, Vasicek et al., 1988). The
setting of the 550C Currie isotherm is consistent with recent thermal models of the Guerrero
subduction zone (Manea et al., 2004). The RP MAGSAT anomaly over the Oaxaca subduction
zone at the altitude of 350 km is induced by a constant 60,000 nT magnetic field (Hinze et al.,
1982). The following values for magnetic susceptibilities are used (see Figure 8): 0.058 SI for
the oceanic slab and 0.013 SI for the continental crust (Nagata, 1969, Thomas, 1984). To fit the
magnetic model with the observed anomaly we also included the effect of the Tuxtla alkaline
basalts, using a 3 km thick layer with a high magnetic susceptibility of 0.19 SI. This minor
contribution from the Tuxtla volcanic province was considered by Vasicek et al. (1988) in order
to fit the 3D magnetic model over Oaxaca with ~10 nT maximum of the MAGSAT scalar
anomaly.
The intraslab seismicity with magnitude Mw greater than 5 is shown in Figure 8 as
additional constraints on the model. The hypocenters distribution is rather consistent with the flat
subhorizontal section of the subducted slab, and also ascertains the dip of the slab in the
asthenosphere about of ~ 20.
Coupling models
Mechanical coupling between the subducting and overriding tectonic plates specifies the
proportion of the relative convergence motion converted into the deformation of the continental
plate edge. Elastic energy of this accumulated deformation could finally be rebounded as a large
subduction thrust earthquake and partially and gradually released by the SSEs. Consequently, the
12
coupling should vary depending on the stage of the seismic cycle and the influence of SSEs. It is
essential to distinguish the maximum value of coupling which is apparently achieved during
interslip steady state phase of plate interaction.
We apply a simple 2D elastic dislocation model of Savage (1983) to appraise the steady
state coupling in Oaxaca subduction zone. Since the subduction geometry significantly changes
from the Central to Southern Oaxaca and two distinct SSEs have occurred in these areas, it is
necessary to consider two different coupling models, one for each region.
For the Central and Northwestern Oaxaca, the 2003-2004 epoch data are apparently free
from the 2001-2002 SSE influence and thus can be used for the coupling modeling (figure 10A).
For other epochs only PAST and ROBL sites may probably represent undisturbed displacements
(see Figure 3 A, B). Almost all GPS sites in Oaxaca are located in the coastal zone (Figure 1),
fairly close to the trench. To obtain a reliable estimate of the coupled zone extension the model
should be restricted with the data from the distant inland sites. For that reason the estimate of
interslip velocity at OAXA permanent GPS station is crucial. OAXA positions records have a
noticeable scatter and time gaps (Figure 2). To obtain the better assessment of the interslip
velocity we apply a network solution (Larson et al., 2004) for OAXA and several GPS stations at
stable part of the North American plate (e.g., MDO1, PIE1). Figure 9 shows baseline OAXAMDO1 coordinates and the best linear fit estimates of OAXA velocities for the periods before
and after the 2001-2002 SSE. Because of the lack of longer period interslip data an average
velocity for both periods is then used for the modeling.
The plate interface geometry is accepted from the gravity modeling described in the
previous section. The model that plausibly fits the observed velocities (Figure 10A) includes a
shallow seismogenic zone that is frictionally coupled or locked and located between ~55 and 105
13
km from the trench. Farther and deeper, a partially coupled transition zone extends over
approximately 100 km. The rest of the plate interface slips freely.
For the Southeastern Oaxaca (southeast from PUAN), the 2001-2002 and 2003-2004
epochs are free of the SSE disturbance but may represent two different tectonic stages: pre- and
post-SSE. An average profile (Profile 2 in Figure 7) representing the interface geometry in this
area is adopted from Bravo et al., 2004 (cross section D in this paper). The coupled interface is
located between 40 km and 100 km from the trench along the profile strike. Figure 10B shows a
feasible model of coupling distribution along this profile. It is rather problematic to apply a
simple 2D model for the essentially 3D plate interface, and particularly in the case of oblique
subduction in the Southeastern Oaxaca. Since the direction of the profile, which is parallel to the
convergence vector, is also oblique to the coast and trench, we model observed velocities as a
function of the site’s distance from the trench along the profile.
The velocity vectors are not parallel to the profile direction for several GPS sites. The
model in Figure 10B intends to fit the velocities considering two sets of the data: total vector
amplitude and its projection on the profile strike. Thus the higher dispersion of these data means
the larger deviation of the velocity vector from the convergence direction, and consequently the
lower fitting weight of the site.
Discussion and conclusions
A detailed analysis of the largest ever registered slow slip event reveals the following
features of this SSE in the subduction zone of Mexico:
- Kostoglodov et al. (2003) noticed that the 2001-2002 SSE nucleated close to CAYA
GPS station, and the major displacement of ~ 0.2 m on the subduction interface was mostly of
14
north-south direction, not parallel to the direction of interslip steady state crustal motion and to
the convergence velocity vector. Our pulse propagation model confirms that the best linear fit to
the observed along-coast arrival times can be achieved only when the SSE would have its
“epicenter” close to (apparently northwest off) CAYA GPS station, in the middle of Guerrero
seismic gap.
- There is strong anisotropy in the velocity of SSE propagation. Along the Pacific coast of
Mexico the propagation velocity is VSSE = 2.3 ± 0.1 km/day, which is about a half of the velocity
component normal to the coastline, VSSE ≈ 5 km/day. The observations of future SSEs should
provide additional insights into the physics of VSSE anisotropy.
- The SSE displacement pulse is characterized by the geometric attenuation and inverse
relation between the pulse duration and the propagation distance.
- The simulation of the propagating SSE pulse at different campaign GPS sites shows that
the observed displacement is in agreement with the modeled one only in the Central Oaxaca
where the subduction interface is subhorizontal (according to our gravity and magnetic
anomalies model). The SSE obviously ceased somewhere at ~97°W on the Pacific coast of
Oaxaca where the PAST and ROBL sites do not show significant velocity changes between the
observational epochs. To the East of this point the modeled displacement pulse is too small to fit
the observed displacements at the campaign sites. This means that the model of single SSE pulse
propagation cannot explain the second SSE episode in the Southeastern Oaxaca subduction zone.
- The second SSE was apparently delayed with respect to the propagation of the main
2001-2002 event, which has probably been a triggering factor. The second SSE had a dominant
east-west displacement on the plate interface with a transitional geometry from subhorizontal to
15
normal dip subduction. The obliquity of the convergence in the southeastern Oaxaca may have a
role for this specific direction of slow rebound.
- The average coupling models are distinct for the Central and Southern Oaxaca. The
interslip coupling in the Central Oaxaca (Profile 1 in Figure 7) is consistent in general with the
coupling estimates for the steady state period in Guerrero (Kostoglodov et al., 2003). The
coupled interface extends from ~55 km to 210 km from the trench, along with the coupling
decreasing down dip from κ ≈ 1 to 0.7 (Figure 10A). For the Southern Oaxaca (Profile 2 in
Figure 7) the coupled zone, with κ ≈ 0.6-0.7, is less then 60 km, located from 40 km to ~100 km
from the trench (along the Profile 2 strike). This drastic difference in the average coupling
extension from the Central to Southern Oaxaca subduction zone is obviously related with a
change in the plate interface geometry.
- To examine the detailed variation of interplate coupling along the Oaxaca subduction
zone a dense GPS network of permanent stations is needed, with a distance between stations less
then 30 km, particularly in the coastal area.
Acknowledgements
This study was supported by G25842-T, 37293-T CONACyT and IN104599 PAPIIT grants.
The authors thank Patricia Julio Miranda, Wallis Hutton and Javier Alvarado for their help in the
survey campaigns. The POSW GPS station is maintained by the CARDI Lab at the Instituto de
Geofisica, UNAM, and the data are in the public access at the UNAVCO data base
(http://archive.unavco.org/query/pss). NSF EAR-0125618 provided funds to the University of
Colorado to partially support Guerrero measurements. Carlos Canet helped to improve the
manuscript.
16
References
Altamimi, Z., P. Sillard, and C. Boucher, ITRF2000: A new release of the International
Terrestrial Reference Frame for earth science applications, J. Geophys. Res., 107 (B10),
2214, doi:10.1029/2001JB000561, 2002.
Blewitt G., An automatic editing algorithm for GPS data, Geophys. Res. Lett., 17, 199-202, 1990.
Blewitt G., Carrier phase ambiguity resolution for the Global Positioning System applied to
geodetic baselines up to 2000 km, J. Geophys. Res., 94, 10, 187-10, 203, 1989.
Boucher C., Z. Altamimi, and P. Sillard, The 2000 International Terrestrial Reference Frame
(ITRF2000), Technical Note 31, IERS, 2003.
Bravo, H., C. J. Rebollar, A. Uribe, and O. Jimenez, Geometry and state of stress of the WadatiBenioff zone in the Gulf of Tehuantepec, Mexico, J. Geophys. Res., 109, B04307,
doi:10.1029/2003JB002854, 2004.
Cabral, E., F. Correa Mora, B. Marquez Azúa, and C. DeMets, A GPS study of the subduction
interface beneath Oaxaca, Mexico, GEOS, Union Geofïsica Mexicana, 23, 182, 2003.
Clark, S.C., Frey, H. and Thomas, H.H., Satellite magnetic anomalies over subduction zones: the
Aleutian Arc anomaly. Geophysical Research Letters, 12: 41-44, 1985.
De la Fuente Duch, M.F., Mena, M., and Aiken, C.L.V., Cartas Gravimétricas de la Republica
Mexicana. I. Carta de anomalía Bouguer, Instituto de Geofísica, UNAM, 1994.
DeMets C., Gordon R., Argus D. and Stein S., Effect of recent revisions to the geomagnetic
time-scale on estimate of current plate motions, Geophys. Res. Lett., 21, 2191-2194, 1994.
GEODAS vs. 4. 0, Marine Trackline Geophysics, U.S. Department of Commerce, National
Oceanic and Atmospheric Administration.
17
Heflin M., Beriger W., Blewitt G., Freedman A.P., Hurst K. Lichten S., Lindqwister J., Vigue
Y., Webb F., Yunck T. P. , Zumberge J., Global geodesy using GPS without fiducial sites,
Geophys. Res. Lett., 19, 131-134, 1992.
Hinze, W.J., Von Frese, R.R.B., Longacre, M.B. and Braile L.W., Regional magnetic and gravity
anomalies of South America. Geophysical Research Letters, 9: 314-317, 1982.
Iglesias, A., S.K. Singh, A. R. Lowry, M. Santoyo, V. Kostoglodov, K. M. Larson, S.I. FrancoSánchez and T. Mikumo, The silent earthquake of 2002 in the Guerrero seismic gap,
Mexico (Mw=7.4): inversion of slip on the plate interface and some implications, Geofisica
Int., 43, 309-317, 2004.
Kelleher, J. and W. McCann, Bouyant zones, great earthquakes, and unstable boundaries of
subduction, J. Geophys. Res., 81, 4885-4896, 1976.
Kostoglodov, V., S. K. Singh, J. A. Santiago, S. I. Franco, K. M. Larson, A. R. Lowry and R.
Bilham, A large silent earthquake in the Guerrero seismic gap, Mexico, Geophys. Res.
Lett., 30, 1807, 2003.
Kostoglodov V., W. Bandy, J. Domínguez, and M. Mena, Gravity and seismicity over the
Guerrero seismic gap, Mexico, Geophys. Res. Lett., 23, 3385-3388, 1996.
Kostoglodov V. and W. Bandy, Seismotectonic constraints on the convergence rate between the
Rivera and North American plates, J. Geophys. Res., 100, 17,977-17,989, 1995.
Kostoglodov, V. and L. Ponce, Relationship between subduction and seismicity in the Mexican
part of the Middle America Trench, J. Geophys. Res., 99, 729-742, 1994.
Larson, K. J. Freymuller, S. Philipsen, Global plate velocities from the Global Positioning
System, J. Geophys. Res., 102, 9961-9981, 1997.
Larson, K., V. Kostoglodov, A. Lowry, W. Hutton, O. Sanchez, K. Hudnut, and G. Suarez,
18
Crustal Deformation Measurements in Guerrero, Mexico, J. Geophys. Res., 109, No. B4,
B04409 10.1029/2003JB002843, 2004.
Leick, A., GPS Satellite Surveying, 560 pp, Wiley-Inter-science, New York, 1995.
Lichten S., and J. Border, Strategies for high precision Global Positioning System orbit
determination, J. Geophysics Res., 92, 12,751-12,762, 1987.
Ligorría, J.P., and Ponce, L., Estructura cortical en el Istmo de Tehuantepec, México, usando
ondas convertidas, Geofísica Internacional, 32, 89–98, 1993.
Lowry, A.R., K.M. Larson, V. Kostoglodov, R. Bilham, Transient slip on the subduction
interface in Guerrero, southern Mexico, Geophys.Res.Lett., 28, 3753-3756, 2001.
Manea, V.C., Manea, M., Kostoglodov, V., Currie, C.A., and Sewell, G., Thermal structure,
coupling and metamorphism in the Mexican subduction zone beneath Guerrero. Geophys.
J. Int., 158, 775-784, 2004
Mao, A., C. G. A. Harrison, and T. H. Dixon, Noise in GPS coordinate time series, J. Geophys.
Res., 104, 2797– 2816, 1999.
Marquez-Azua Bertha and Charles DeMets, Crustal velocity field of Mexico from continuous
GPS measurements, 1993 to June 2001: Implications for the neotectonics of Mexico, J.
Geophys. Res., 108, No. B9, B092450 10.1029/2002JB002241, 2003.
McCaffrey R., Oblique plate convergence, slip vectors, and forearc deformation, J. Geophys.
Res., 97, 8905-8915, 1992.
Nagata, T., Reduction of geomagnetic data and interpretation of anomalies, in The Earth’s Crust
and Upper Mantle Geophysical Monograh, edited by P.J. Hart, American Geophysical
Union, 13: 391-398, 1969.
19
Nuñez-Cornú, F. and L. Ponce, Zonas sísmicas de Oaxaca, México: sismos máximos y tiempos
de recurrencia para el periodo 1542-1988, Geofísica Int. 28, 587-641, 1989.
Pardo M. and G. Suárez, Shape of the subducted Rivera and Cocos plates in southern Mexico:
Seismic and tectonic implications, J. Geophys. Res., 100, 12357-12373, 1995.
Singh, S.K., J. Havskov, and L. Astiz, Seismic gaps and recurrence periods of large large
earthquakes along the Mexican subduction zone, Bull. Seism. Soc. Am. 71 (3), 827-843,
1981.
Singh, S.K., M. Rodríguez, and L. Esteva, Statistics of small earthquakes and frequency of
occurrence of large earthquakes along the Mexican subduction zone, Bull. Seism. Soc. Am.
73 (6), 1779-1796, 1983.
Singh, S.K., M. Ordaz, L. Alcantara, N. Shapiro, V. Kostoglodov, J.F. Pacheco, S. Alcocer, C.
Gutierrez, R. Quaas, T. Mikumo, E. Ovando, J. Aguirre, D. Almora, J.G. Anderson, M.
Ayala, C. Javier, G. Castro, R. Duran, G. Espitia, J. Estrada, E. Guevara, J. Lermo, B.
Lopez, O. Lopez, M. Macias, E. Mena, M. Ortega, C. Perez, J. Perez, M. Romo, M.
Ramirez, C. Reyes, R. Ruiz, H. Sandoval, M. Torres, E. Vazquez, R. Vazquez, J.M.
Velasco, J. Ylizaturri, The Oaxaca earthquake of 30 September 1999 (M (sub w) = 7.5); a
normal-faulting event in the subducted Cocos plate, Seism. Res. Lett., 71, 67-79, 2000.
Thierry, G., Gipsy - OASIS II How it works..., manual for class of GIPSY, 109 pp, Jet Propulsion
Laboratory, California Institute of Technology, Pasadena, USA, 1996.
Thomas, H.H., Petrologic model of the northern Mississippi Embayment based on satellite
magnetic and ground based geophysical data. Earth Planet. Sci. Lett., 70: 115-120, 1984.
20
Torres Zamudio, A., El campo gravimétrico y la estructura de la zona de subducción en Oaxaca,
Mexico, B.S. –Thesis, Escuela Superior de Ingeniería y Arquitectura, Instituto Politécnico
Nacional, pp. 89, 2002.
Vasicek, J.M., Frey, H.V., and Thomas, H.H.. Satellite magnetic anomalies and the Middle
America Trench. Tectonophysics, 154, 19-24, 1988.
Zumberge
J.F., F.H. Webb,
M.B. Heflin,
D.C. Jefferson, M.M. Watkins, Precise point
positioning for the efficient and robust analysis of GPS data from large networks, J.
Gephys. Res., 102, 5005-5017, 1997.
Yoshioka, S.,
T. Mikumo, V. Kostoglodov, K. M. Larson, A. R. Lowry, and S. K. Singh,
Interplate coupling and a recent aseismic slow slip event in the Guerrero seismic gap of the
Mexican subduction zone, as deduced from GPS data inversion using a Bayesian
information criterion, Phys. Earth Planet. Inter., 146, 513-530, 2004.
21
Figure captions
Figure 1. The study area encompasses Guerrero and Oaxaca states of Mexico. Shaded ellipselike areas annotated with the years are rupture zones of the most recent large thrust earthquakes
(M ≥ 6.5) in the Mexican subduction zone. Triangles show locations of permanent GPS stations.
Small hexagons indicate campaign GPS sites. Arrows are the Cocos-North America convergence
vectors from NUVEL-1A model (DeMets et al., 1994). Double head arrow denotes the extension
of the Guerrero seismic gap.
Figure 2. Daily time series of relative latitude at several permanent GPS stations in Guerrero and
Oaxaca. Dashed vertical lines mark apparent initial and ending dates of the silent earthquake at
CAYA station. Arrows indicate the initial phases of the slow slip at different stations. Some GPS
records are not complete because of technical failures. Shaded vertical bands show the survey
periods at the campaign GPS sites.
Figure 3. The results of GPS campaign surveys in Oaxaca. A, B, C – Maps of the horizontal
displacements (solid arrows) during the 2001-2002-2003-2004 epochs, accordingly. Error
ellipses show two formal standard deviations. The displacement vectors in Guerrero (sites
ACAP, CAYA, IGUA, YAIG) have an evident influence of the slow slip event for the 20012002-2003 epochs. Open arrows are the Cocos-North America convergence predictions from the
NUVEL-1A model (DeMets et al., 1994). Locations of permanent and temporary GPS stations
are shown by triangles and circles respectively. D – Hodographs of the horizontal velocities at
22
each site for the 2001-2002-2003-2004 epochs. E – Comparison of the horizontal displacement
vectors for the 2001-2002-2003-2004 epochs.
Figure 4. Propagation of the SSE of 2001-2002 in Guerrero and Oaxaca. A. SSE arrival time as
a function of the distances along and normal to the coast line strike. The GPS stations locations
are shown by yellow circles. The anisotropy of the SSE propagation is obvious: the SSE
propagates faster (~5.0 km/day) in the direction perpendicular to the trench. B. The best linear
regression fit to the observed arrival times (the along-coast distances are calculated relative to the
position of CAYA station). Southeast along the coast velocity of SSE propagation is about ~ 2.3
km/day (thick red line). Including ZIHU arrival time (ZIHU is the only GPS station located
northwest from CAYA) slows the along-coast propagation velocity down to 2.0 km/day (thin
green line).
Figure 5. Propagation parameters of the SSE 2001-2002. CAYA station is assumed as an initial
(nucleation) point of the slow event. A. Attenuation of the displacement pulse (Geometrical
factor); B. Decrease of the displacement pulse duration as a function of distance.
Figure 6. Relative latitudinal position of the campaign sites and modeled displacements of the
2001-2002 slow slip event. A best fit displacement pulse for the PINO record is accepted as a
pattern signal for this slow event. The displacement pulses for other GPS sites are modeled
applying the estimates of propagation velocity, attenuation and duration (Figure 4, 5) to this
pattern signal. A dashed arrow displays a modeled arrival time of the silent earthquake at each
campaign site. Shaded charts correspond to the sites with a reasonably good fit of modeled signal
23
to the observed displacements. Starting from PUAN the modeled signal does not fit the observed
displacements at the campaign sites located eastward from this station. The modeled
displacement pulse at HUAT GPS station is faster than the apparent SSE phase in the observed
data.
Figure 7. Preferred model of the SSE propagation along the cost of Guerrero and Oaxaca. The
main slow slip event (with a predominant north-south component) starts at the end of 2001 at
CAYA, Guerrero, and propagates into the Western Oaxaca where it gradually ceases in the
middle of 2002. The next, smaller and relatively delayed slow slip occurs in the Southeastern
Oaxaca. It is probably triggered by the 2001-2002 SSE and has the westward slip direction.
Figure 8. Gravity and magnetic anomalies modeling used to constrain the interplate contact
geometry in the Central Oaxaca area. Position of the profile is shown in Figure 7 (Profile 1). For
other explanations see text: Configuration of the plate interface.
Figure 9. Time series of daily relative position of OAXA with respect to MDO1 (McDonald
Observatory, Texas) permanent GPS station. Lines show the best fit linear trends before and after
the SSE. The values of average interslip velocity components (corrected for a secular bias
produced by the rigid North America plate motion, see Larson et al., 2004) are specified in the
upper left side of each chat. Vertical dashed lines indicate the period of the SSE.
Figure 10. Elastic half space dislocation models which fit the observed velocities at the GPS
sites in Oaxaca for the undisturbed steady state interslip epochs. Dashed (red) and solid (blue)
24
curves are the modeled uplift and horizontal (negative is compressional) displacement rates,
correspondingly. Thick line segments on the plate interface represent a coupled zone annotated
with the degree of interplate frictional coupling. A. Central Oaxaca, Profile 1 in Figure 7. The
modeled data are mostly from the 2003-2004 epoch, open circles are for horizontal and
diamonds for vertical velocity components correspondingly. Oaxaca velocity estimate is average
for the pre-SSE and post-SSE epochs. B. Eastern Oaxaca, Profile 2 in Figure 7, includes the data
of 2001-2002 and 2003-2004 epochs. The squares are the average horizontal velocities for the
2001-2002 epoch and circles are the data for the 2003-2004 epoch. Plate interface geometry is
constrained by the gravity anomalies modeling for Profile 1, and seismicity data (Bravo et al.,
2004) for Profile 2. The observed campaign uplift rates have large errors and are not shown for
the most of the campaign surveys.
25
Figure 1
26
Figure 2
27
Figure 3 A
28
Figure 3 B
29
Figure 3 C
30
Figure 3 D,E
31
Figure 4 A, B
32
Figure 5 A, B
33
Figure 6
34
Figure 7
35
Figure 8
36
Figure 9
37
Figure 10 A,B