ArticlePDF Available

Abstract and Figures

S U M M A R Y The subduction plate interface along the Nicoya Peninsula, Costa Rica, generates damaging large (M w > 7.5) earthquakes. We present hypocenters and 3-D seismic velocity models (V P and V P /V S) calculated using simultaneous inversion of P-and S-wave arrival time data recorded from small magnitude, local earthquakes to elucidate seismogenic zone structure. In this region, interseismic cycle microseismicity does not uniquely define the potential rupture extent of large earthquakes. Plate interface microseismicity extends from 12 to 26 and from 17 to 28 km below sea level beneath the southern and northern Nicoya Peninsula, respectively. Microseismicity offset across the plate suture of East Pacific Rise-derived and Cocos-Nazca Spreading Center-derived oceanic lithosphere is ∼5 km, revising earlier estimates suggesting ∼10 km of offset. Interplate seismicity begins downdip of increased locking along the plate interface imaged using GPS and a region of low V P along the plate interface. The downdip edge of plate interface microseismicity occurs updip of the oceanic slab and continental Moho intersection, possibly due to the onset of ductile behaviour. Slow forearc mantle wedge P-wave velocities suggest 20–30 per cent serpentinization across the Nicoya Peninsula region while calculated V P /V S values suggest 0–10 per cent serpentinization. Interpretation of V P /V S resolution at depth is complicated however due to ray path distribution. We posit that the forearc mantle wedge is regionally serpentinized but may still be able to sustain rupture during the largest seismogenic zone earthquakes.
Content may be subject to copyright.
Geophys. J. Int. (2006) 164, 109–124 doi: 10.1111/j.1365-246X.2005.02809.x
GJI Seismology
Seismogenic zone structure beneath the Nicoya Peninsula, Costa
Rica, from three-dimensional local earthquake P- and S-wave
tomography
Heather R. DeShon,
1
Susan Y. Schwartz, Andrew V. Newman,
2
Victor Gonz´alez,
3
Marino Protti,
3
LeRoy M. Dorman,
4
Timothy H. Dixon,
5
Daniel E. Sampson
1
and Ernst R. Flueh
6
1
University of California-Santa Cruz, Earth Sciences Department & IGPP, Santa Cruz, California, USA
2
Georgia Institute of Technology, Atlanta, Georgia, USA
3
Observatorio Vulcanol´ogico y Sismol´ogico de Costa Rica-UNA, Heredia, Costa Rica
4
University of California-San Diego, Scripps Institution of Oceanography, La Jolla, California, USA
5
University of Miami, Rosenstiel School of Marine and Atmospheric Science, Miami, Florida, USA
6
IFM-GEOMAR and SFB574, Kiel, Germany
Accepted 2005 September 21. Received 2005 September 6; in original form 2004 October 14
SUMMARY
The subduction plate interface along the Nicoya Peninsula, Costa Rica, generates damaging
large (M
w
> 7.5) earthquakes. We present hypocenters and 3-D seismic velocity models
(V
P
and V
P
/V
S
) calculated using simultaneous inversion of P- and S-wave arrival time data
recorded from small magnitude, local earthquakes to elucidate seismogenic zone structure. In
this region, interseismic cycle microseismicity does not uniquely define the potential rupture
extent of large earthquakes. Plate interface microseismicity extends from 12 to 26 and from
17 to 28 km below sea level beneath the southern and northern Nicoya Peninsula, respectively.
Microseismicity offset across the plate suture of East Pacific Rise-derived and Cocos-Nazca
Spreading Center-derived oceanic lithosphere is 5 km, revising earlier estimates suggesting
10 km of offset. Interplate seismicity begins downdip of increased locking along the plate
interface imaged using GPS and a region of low V
P
along the plate interface. The downdip
edge of plate interface microseismicity occurs updip of the oceanic slab and continental Moho
intersection, possibly due to the onset of ductile behaviour. Slow forearc mantle wedge P-
wave velocities suggest 20–30 per cent serpentinization across the Nicoya Peninsula region
while calculated V
P
/V
S
values suggest 0–10 per cent serpentinization. Interpretation of V
P
/V
S
resolution at depth is complicated however due to ray path distribution. We posit that the forearc
mantle wedge is regionally serpentinized but may still be able to sustain rupture during the
largest seismogenic zone earthquakes.
Key words: Costa Rica, earthquake location, microseismicity, Middle America subduction
zone, seismic velocities, tomography.
1INTRODUCTION
Underthrusting earthquakes occurring along subduction megath-
rusts account for greater than 80 per cent of the seismic moment re-
leased worldwide (Pacheco & Sykes 1992). Great (M
w
> 8.0), large
(M
w
> 7), and tsunamigenic earthquakes at convergent margins
cause much damage and loss of life along heavily populated coastal
zones. Understanding how and where seismogeniczone earthquakes
Now at: University of Wisconsin-Madison, Department of Geology and
Geophysics, 1215 W. Dayton St., Madison, WI 53706, USA. E-mail:
hdeshon@geology.wisc.edu
occur is a major focus of the international scientific community. Re-
cently, studies along the Middle America subduction zone offshore
Costa Rica–Nicaragua have led to new insights into the complex in-
teractions of thermal, mechanical, hydrological and compositional
processes generating seismogenic zone seismicity.
Images of regional seismicity within the Middle America sub-
duction zone in northern Costa Rica have been limited in the
past by the spatial coverage of available permanent networks of
single-component short-period seismometers. Earthquakes reported
through country-wide catalogues by Observatorio Vulcanol´ogico
y Sismol´ogico-Universidad Nacional de Costa Rica (OVSICORI-
UNA) and theRed Sismol´ogico Nacional (RSN-ICE) haverelatively
C
2005 The Authors 109
Journal compilation
C
2005 RAS
110 H. R. DeShon et al.
Figure 1. Overview map of the Nicoya CRSEIZE experiment. Cocos Plate
oceanic crust formed at the East Pacific Rise (EPR) and at the Cocos-Nazca
Spreading Center (CNS-1 and CSN-2) (Barckhausen et al. 1998, 2001)
subducts along the Middle America Trench (MAT) offshore northern Costa
Rica. The EPR–CNS plate suture occurs along a triple junction trace (bold
long dashed line); other tectonic boundaries include a propagator trace (short
dashed line) and a ridge jump (dotted line). Initial database locations using
1-D IASP91 velocity model are scaled by local magnitude (red circles).
The Nicoya Experiment recorded interseismic seismogenic zone seismicity,
crustal seismicity and the main shock and aftershock sequence of the 2000
Nicoya earthquake along the outer rise (yellow star with Harvard Centroid
Moment Tensor solution). The M
w
7.7 1950 and M
w
6.9 1978 earthquakes
(blue stars) and the 1950 aftershock region (blue line) provide one seis-
mogenic zone estimate (Avants et al. 2001). NP: Nicoya Peninsula; NG:
Nicoya Gulf; and OP: Osa Peninsula. Bathymetry is from von Huene et al.
(2000).
large depth errors (>5 km), especially for those events occurring
north and west of the primary networks (Quintero & Kissling 2001).
Increased station coverage using temporary three-component seis-
mometers on land and offshore the Nicoya Peninsula significantly
improved the resolution of microseismicity along the shallow plate
interface (Newman et al. 2002; DeShon & Schwartz 2004). The
Nicoya Peninsula passive array component of the Costa Rica Seis-
mogenic Zone Experiment (CRSEIZE) consisted of 20 short-period
and broad-band seismometers that recorded across the peninsula
from 1999 December to 2001 June and 14 broad-band ocean bot-
tom seismometers (OBS) that recorded from 1999 December to
2000 June (Fig. 1). OBS were improved versions of those describe
in Sauter et al. (1990) and Jacobson et al. (1991). The array recorded
over 10 000 earthquakes,and analyst-reviewed P- and S-wave arrival
times exist for2000 local events.This extensivedata set andstation
distribution allows for improved accuracy of earthquake locations
with significantly reduced depth errors.
In this study P-wave arrival time and SP traveltime delays
recorded by the CRSEIZE Nicoya Peninsula seismic array are in-
verted to solve for best fitting hypocenters, station corrections and
compressional and shear wave velocities. The resulting 3-D seis-
mic velocity information is combined with the improved precision
relocations of subduction-related seismicity to constrain material
properties beneath the Nicoya Peninsula, Costa Rica. Study goals
include:
(1) characterizing the geometric range of seismic activity within
the subduction zone,
(2) correlating variations in seismicity to velocity structure near
the plate interface and
(3) comparing the up- and downdip limits of seismogenic zone
seismicity to available thermal models and physical properties of
the d`ecollement.
2GEOLOGIC AND TECTONIC
SETTING
Oceanic Cocos Plate subducts beneath continental Caribbean Plate
offshore northern Costa Rica at 83–85 mm yr
1
,20
counter-
clockwise to orthogonal subduction at the Middle America Trench
(MAT) (Fig. 1) (DeMets 2001). Coastal subsidence and uplift along
the central coast range and the Nicoya Gulf are consistent with
trench-normal locking along the plate interface (e.g. Marshall &
Anderson 1995). Lundgren et al. (1999), Iinuma et al. (2004), and
Norabuena et al. (2004) have inverted GPS site velocities to ob-
tain estimates of interseismic locking along the plate interface be-
neath the peninsula. Norabuena et al. (2004), using data collected as
part of CRSEIZE, found a component of along-strike crustal sliver
transport and a variably coupled trench-normal plate interface, with
strongest locking occurring at 14 ± 2 and 38 ± 2kmdepth be-
low sea level (BSL), updip and downdip of interseismic interplate
seismicity. The resolution of the GPS inversion, however, precluded
examination of along-strike variability in locking beyond the Nicoya
Peninsula.
Seismic potential for the Nicoya Peninsula region has been esti-
mated based on the rupture limits of the 1950 M
w
7.7 underthrust-
ing earthquake and consideration of regional seismicity and plate
coupling (Fig. 1) (G¨uendel 1986; Protti et al. 2001). The most re-
cent damaging earthquake occurred on 1978 August 23 (Fig. 1).
This M
w
6.9 earthquake, originally located offshore the peninsula
at 25.7 km depth by G¨uendel (1986), had a focal mechanism consis-
tent with thrusting along the plate interface, but comparison of the
original hypocenter to the plate interface defined by refraction data
(Christeson et al. 1999; Sallar`es et al. 1999) places the earthquake
within the oceanic crust. Avants et al. (2001) relocated both the
1950 and 1978 earthquakes reported by G¨uendel (1986) relative
to the 1990 M
w
7.0 Nicoya Gulf earthquake (Protti et al. 1995b).
These relocations place the older earthquakes slightly inboard and
south of the original locations, more consistent with rupture along
the plate interface (Fig. 1) (Avants et al. 2001). We assume consis-
tent mislocation of these main shocks due to station geometry and
propagation path errors and accordingly adjust the 1950 aftershock
area, which was originally defined by G¨uendel (1986) using arrivals
from a local temporary network. This provides a revised estimate
of the potential seismogenic zone rupture along the Nicoya seismic
gap (Fig. 1).
Though the bathymetry of the incoming Cocos Plate near the
Nicoya Peninsula appears uniformly smooth seaward of the trench,
it exhibits variability in lithospheric history and shallow thermal
structure (e.g. Shipley et al. 1992; Kimura et al. 1997; Spinelli &
Underwood 2004). Oceanic lithosphere subducting beneath the
peninsula derives from both the Cocos-Nazca Spreading Center
(CNS) and East Pacific Rise (EPR) (Fig. 1) (Barckhausen et al.
2001). Lithosphere formed at the CNS between 21.5 and 23 Ma
subducts beneath the southern Nicoya Peninsula while lithosphere
formed at the EPR subducts in the north. At the MAT, there is an
1Maage difference between the two lithospheres that corresponds
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
Seismogenic zone structure beneath the Nicoya Peninsula 111
to a small (200 m offset) fracture zone (Walther & Flueh 2002).
Due to hydrothermal circulation, heat flow through EPR crust is
anomalously low (20–40 mW m
2
) compared to conductive cool-
ing models (100 mW m
2
) (Stein & Stein 1992; Langseth & Silver
1996; Harris & Wang 2002; Fisher et al. 2003). However, CNS
crustal heat flow is largely consistent with conductive cooling mod-
els (105–115 mW m
2
). The separation between hydrothermally
cooled and normal oceanic crust follows a meandering pattern on
the EPR crust and correlates to the plate suture at the MAT (Fisher
et al. 2003; Hutnak 2006).
Current estimates of crustal structure and plate interface geom-
etry shallower then 50 km depth come from wide-angle refraction
studies conducted on and offshore the Nicoya Peninsula. Walther
& Flueh (2002) imaged the Cocos Plate across the EPR–CNS plate
suture in a trench-parallel direction. Their results indicated normal
oceanic crustal structure and a low-velocity body near the plate su-
ture, which was interpreted as an igneous intrusion at the base of the
EPR-derived crust. Refraction line 300 imaged the trench-normal
CNS side of the suture (Fig. 2) (Ye et al. 1996); along this transect
the top of the Cocos Plate extends at a constant dip angle of 7
to
14 km depth, 90 km from the MAT. Across the EPR side of sub-
duction, refraction line 101 extended from 20 km seaward of the
MAT well landward of the volcanic arc (Fig. 2) (Christeson et al.
1999; Sallar`es et al. 1999, 2001). Here the top of the subducted
slab extends from 5 km depth at the MAT to 15–16 km depth at the
coastline dipping shallowly at 6
, transitioning to 13
dip at 30
km from the trench (Christeson et al. 1999); by 40 km depth the
plate dips 35
(Sallar`es et al. 1999, 2001). These studies indi-
cated that the Nicoya ophiolite complex comprising the peninsula
extends offshore to form the margin wedge. P-wave velocities for
-87 -86 -85
9
10
11
EPR
CNS
A
Line 101
Line 300
Figure 2. Nicoya Experiment 3-D inversion data set. 611 earthquakes
within the network coverage were recorded by >10 P-wave arrivals (dark
grey circles) and are subdivided from the minimum 1-D inversion data set for
use in 3-D simultaneous inversion. Solid black lines mark seismic refraction
and reflection information discussed in the text (Line 101:Christeson et al.
1999 and Sallar`es et al. 1999, 2001; Line 300:Ye et al. 1996; A: Walther &
Flueh 2002). Triangles: seismic stations; White circle: 1-D model reference
station GUAI; diamond: GNS broad-band station JTS. Dashed lines: tectonic
boundaries defined in Fig. 1.
this sequence range from 4.5–5.0 km s
1
depending on depth, with
very low velocities (3.0 km s
1
) and high lateral velocity gradi-
ents within a small frontal prism composed of slope apron sediments
and margin wedge material (Christeson et al. 1999). The Caribbean
crust in northern Costa Rica is 40 km thick, nearly double the
average Caribbean Plateau crustal thickness, and exhibits smoothly
increasing velocities and densities with depth (Sallar`es et al. 2001).
Sallar`es et al. (2001) proposed a formation sequence for the Costa
Rica isthmus that includes magma generation, intrusion and under-
plating that thickened the Caribbean Plate and hydrated the forearc
mantle.
3METHODS
Local earthquake tomography (LET) techniques provide improved
earthquake locations and seismic velocity information that in turn
lends insight into the seismic and tectonic framework of a region.
Calculation of traveltime to solve for earthquake location depends
on earthquake phase picks, initial hypocentral location and veloc-
ity structure (e.g. Aki & Lee 1976; Crosson 1976; Ellsworth 1977;
Kissling et al. 1984; Thurber 1981, 1992). In this study we use the
LET algorithm SIMULPS13Q, an iterative damped least squares
solution for local earthquake data that utilizes approximate ray
tracing for computation of theoretical traveltimes (Thurber 1983;
Eberhart-Phillips 1990; Evans et al. 1994). Using this technique,
traveltime derivatives are extracted incrementally along small seg-
ments of the ray path, and velocities and hypocenters are iteratively
updated to minimize arrival time misfit (e.g. Thurber 1983, 1993;
Eberhart-Phillips 1986, 1990; Kissling et al. 1994). The complexity
and resolution of the resulting velocity model depends on station
spacing, earthquake distribution, subsurface structure and inversion
parametrization, including choice of damping parameters and grid
node spacing (Kissling et al. 2001).
We solve for the compressional wave velocity, V
P,
and the ratio
of the compressional and shear wave velocities, V
P
/V
S
, using the
compressional wave arrival time, P, and the time difference between
the shear and compressional waves, SP, respectively. For real data
sets, there is generally a significant decrease in quality and quantity
of S-wave arrival identification, thus using SP traveltimes to solve
for V
P
/V
S
accounts for the inherent coupling between V
P
and V
S
across the same structure and minimizes induced artefacts due to
differing resolution if V
P
and V
S
were calculated separately (e.g.
Eberhart-Phillips 1990; Eberhart-Phillips & Reyners 1997; Reyners
et al. 1999; Thurber & Eberhart-Phillips 1999; Wagner et al. 2005).
Because compressional and shear waves are equally affected by
density, changes in V
P
/V
S
can be directly related to Poisson’s ratio,
making it a useful tool for assessing petrologic properties.
Solutions of the linear approximation of the coupled
hypocenter-velocity problem, such as the parametrization used by
SIMULPS13Q, depend on choice of the initial reference model
(Kissling et al. 1994). Poorly defined 1-D a priori models can
introduce artefacts into the 3-D inversion that are not easily de-
tected using standard resolution or covariance estimates (Kissling
et al. 1994). Use of the ‘minimum 1-D velocity model’, or the opti-
mal least squares solution of the coupled hypocenter-velocity model
problem in 1-D for a given data distribution, decreases the possibil-
ity of artefacts induced by the initial reference model (e.g. Kissling
et al. 1994). We use a minimum 1-D velocity model calculated using
CRSEIZE Nicoyaarrival time datawithin the 1-D inversion program
VELEST (Kissling et al. 1995) as the initial starting model for 3-D
LET (Table 1) (DeShon & Schwartz 2004). The Nicoya minimum
1-D model P-wave velocities agree well with the velocities modelled
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
112 H. R. DeShon et al.
using wide-angle refraction data at the samedistance from the trench
as the 1-D reference station GUAI (Sallar`es et al. 1999, 2001). The
initial V
P
/V
S
inversion begins at 1.78, which represents the average
of the minimum 1-D V
P
/V
S
velocity model and is consistent with
reported V
P
/V
S
studies in Costa Rica (Protti et al. 1995b; Quintero
&G¨uendel 2000; DeShon et al. 2003). Station corrections from the
1-D inversion are not included as a starting parameter in the 3-D in-
version because the 1-D inversion values reflect model uncertainties
in calculating a flat-layered velocity model.
We relocate 611 earthquakes that occurred beneath the Nicoya
CRSEIZE stations using the minimum 1-D velocity model prior
to the full 3-D inversion (Fig. 2). These events have 10 P-wave
arrivals and a greatest azimuthal P-wave separation of 180
.
Recordings at Global Seismic Network (GSN) station JTS, a three-
component broad-band station east of the Nicoya Gulf (Fig. 2), are
included to expand the coverage of the array. Arrival quality is as-
signed to P- and S-wave data during initial processing (described
further in DeShon 2004). Individual arrival quality factors range
from 0–4, with 0 indicating highest quality picks (i.e. high signal-
to-noise ratio, impulsive arrivals) and 4 being low-quality picks not
appropriate for use in the inversion. Average P-wave picking er-
ror is estimated to be 0.10 s and S-wave picking error is 0.15–
0.20 s.
In order to reduce artificial velocity smearing, earthquakes should
be as evenly distributed as possible throughout the model such that
rays crossfrom multipledirections and angles(e.g. Eberhart-Phillips
1990; Evans et al. 1994). The majority of local earthquakes used
for inversion occur along the dipping plate interface and within
the continental crust landward of the OBS and, therefore, data are
not evenly distributed throughout the model space. Grid spacing
is defined to minimize variability in the number of ray paths that
sample each cuboid, or node-bounded parallelpiped. A coarse grid
spacing of20 × 20 km
2
places approximatelyone station per cuboid,
and a finer grid spacing of 10 × 10 km
2
is applied beneath the
peninsula where station spacing warrants (Fig. 3). Only grid nodes
-87 -86 -85
9
10
11
1
2
3
4
5
Figure 3. Coarse and fine grid spacing (crosses) used for progressive
3-D inversions. A 20-km-grid spacing was used to calculate the initial
3-D velocity structure and 10 km grid spacing was applied in regions of
higher resolution. Solid black lines denote profiles through the model shown
in later figures numbered sequentially from south to north. Triangles: seismic
stations.
Table 1. Initial 1-D V
P
and V
P
/V
S
velocity models used for 3-Dinversion
(modified from DeShon & Schwartz 2004).
Depth (km) V
P
(km s
1
) V
S
(km s
1
) V
P
/V
S
10.0 3.00 1.69 1.78
0.0 5.35 3.01 1.78
5.0 6.12 3.44 1.78
9.0 6.12 3.44 1.78
13.0 6.28 3.53 1.78
16.0 6.46 3.63 1.78
20.0 6.72 3.78 1.78
25.0 7.01 3.94 1.78
30.0 7.39 4.15 1.78
35.0 7.55 4.24 1.78
40.0 8.14 4.57 1.78
50.0 8.14 4.57 1.78
65.0 8.26 4.64 1.78
90.0 8.26 4.64 1.78
700.0 10.3 5.79 1.78
whose surrounding cuboids are sampled by 8rays are included
in the velocity inversion, and nodes outside of the station coverage
remain fixed. Depth grid spacing is similar to the minimum 1-D
model with minor revisions to regularize ray coverage (Table 1).
Progressiveinversion schemeshavebeen shown toyield smoother
velocity changes and fewer artefacts then non-progressive schemes
(e.g. Kissling et al. 1994; Eberhart-Phillips & Reyners 1999;
Reyners et al. 1999; Husen et al. 2000, 2003a). Hence, we pro-
gressively iterate from 1-D, to a coarse 3-D, then a fine 3-D model
such that results from the previous step are the initial conditions
for the following step. First, we solve for a coarse grid V
P
, then a
coarse grid V
P
and V
P
/V
S
, and finally a fine grid V
P
and V
P
/V
S
.
Extensive testing of alternative progression pathways indicated that
the chosen path yielded a smoothly varying velocity model with
fewer artefacts due to horizontal and vertically varying resolution,
low variance, and geologically reasonable V
P
and V
P
/V
S
(DeShon
2004).
Velocity and station parameter damping affect both the resulting
velocity model and resolution estimates (Kissling et al. 2001), and
properly choosing damping parameters ensures smoothly varying
velocity models while minimizing data variance (Eberhart-Phillips
1986). We analyse trade-off curves for individual iterations with
and without station corrections over a range of damping parame-
ters (1–800) to find that value which minimizes data variance while
maintaining low model variance(Eberhart-Phillips 1986). Choice of
damping parameter also depends on a priori data quality (Kissling
et al. 2001). Nicoya data picking errors are relatively low, so we
choose the lowest damping value that minimizes data and model
variance. For V
P,
we choose a preferred damping of 50 for the
coarse grid and 100 for the fine grid (Fig. 4a). Because the coarse
grid 3-D V
P
is the initial starting model for the V
P
plus V
P
/V
S
in-
version, the choice of damping for V
P
/V
S
proceeds with V
P
held
slightly overdamped (100). A preferred V
P
/V
S
damping of 50 is
chosen (Fig. 4b). Station corrections account for traveltime residu-
als not incorporated into the 3-D velocity model and for near surface
velocity characteristics not resolved due to data and network cov-
erage (Eberhart-Phillips & Reyners 1999). Station corrections are
damped in order to allow data residuals to be incorporated into the
velocity model rather than the correction terms during the first in-
version iteration. Calculation of station damping yields a preferred
value of 10.
Earthquake locations are held fixed in the first iteration to guar-
antee data residuals are incorporated into the velocity model rather
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
Seismogenic zone structure beneath the Nicoya Peninsula 113
0
0.01
0.02
0.03
0.04
0 2e-3 4e-3 6e-3
Data Variance (s )
2
Model Variance (km s
2
)
(a) P-wave
50
100
m s
2
)
0
0.02
0.04
0.06
0.08
0.10
0 2e-4 4e-4 6e-4 8e-4
Data Variance (s )
2
Model V
ariance (k
(b) S-P
50
Figure 4. Trade-off curves of data versus model variance for coarse (black)
and fine (grey) grid spacing. Curves are shown for single iteration inversions
for both P(a) and SP(b) arrival data with (solid) and without (dotted) station
corrections. For the trade-off curves including station corrections, damping
of 50 for P and SP data with a station damping of 10 best minimize data and
model variance. Fine grid and combined V
P
and V
P
/V
S
inversions to explore
damping begin from the appropriate preceding 3-D solution to maintain con-
sistency with the progression used in this study. Testing damping parameters
range from 5 to 800.
than systematic movement of earthquake locations or origin times
(i.e. Husen et al. 2000, 2003a). When locations are not held fixed on
the first iteration using the Nicoya data set, data misfit is expressed
as large velocity perturbations in the shallowest layers and yields
geologically unreasonable models (DeShon 2004).
4RESOLUTION ESTIMATES
Model resolution is assessed using multiple methods (e.g. Eberhart-
Phillips & Reyners 1999; Husen et al. 2000). To test amplitude and
geometry recovery of the data set, we construct a synthetic arrival
time data set using the hypocenters, arrival weighting and ray cov-
erage of the data input into the 3-D inversion. We design a synthetic
velocity model that reflects the broad scale geometry of the 3-D
solution but assign opposite sign to expected anomalies to mathe-
matically decouple the synthetic model from the real solution. This
is called the characteristic velocity model (Haslinger et al. 1999).
Computation of theoretical traveltimes through the characteristic
model is preformed using the ART approach in SIMULPS13Q. We
also examine resolution using derivative weighted sums (DWS: rela-
tive raydensity in the vicinity ofa model parameter) and thediagonal
element of the full resolution matrix (RDE). The spread function,
which is based on the relative values of the diagonal versus the rows
of the full resolution matrix (Toomey & Foulger 1989; Michelini &
McEvilly 1991), is calculated to better constrain velocity smearing
at individual nodes.
The characteristic V
P
model has a dipping 10 per cent anomaly
that generally parallels the Cocos Plate, but with a steeper dip, and
various crustal anomalies above the expected continental Moho.
The mantle wedge region is assigned to be 10 per cent faster than
the minimum 1-D model and is slightly offset landward in order
to assess horizontal smearing between the slab and wedge region.
All V
P
anomalies are assigned to be ±10 per cent of the initial
value at a grid point; V
P
/V
S
anomalies are either ±7.5 per cent or
±15 per cent of 1.78 (Fig. 5). Random noise is added to the synthetic
traveltime data set prior to inversion such that arrivals with quality
0 are assigned noise between 0.02 and 0.02 s. This is linearly
scaled so that arrivals with quality 3 are assigned noise between
0.18 and 0.18 s.
The synthetic arrival time data is used within the progressive
inversion scheme to estimate smearing, expected velocity recovery
and variable resolution. Due to the a priori data error, we cannot
resolve velocity perturbations of less than ±1–2 per cent at a given
depth. Fig. 5 consists of depth slices through the characteristic V
P
and V
P
/V
S
models and the recovered velocity models. Note that
V
P
and V
P
/V
S
resolution, and specifically the location of velocity
artefacts, varies significantly due to differences in data quality, data
quantity and ray coverage. Fig. 6 consists of a cross-section along
Profile 4 (Fig. 3) through a well-resolved region of the velocity
model where the OBS extend seaward of the MAT. Due to the lack
of OBS along the southern Nicoya Peninsula (Fig. 3, Profiles 1
and 2), V
P
and V
P
/V
S
resolution varies significantly along MAT
strike. There is little to no resolution of low-velocity layers above
5kmdepth for both V
P
and V
P
/V
S
(Fig. 5). Anomaly recovery
offshore is also consistently poorer than beneath the land due to
data distribution. The majority of earthquakes occur from 12 to
35 km depth, and at these depths V
P
anomaly recovery averages
80–100 per cent of the prescribed anomaly beneath the peninsula
and OBS array. In comparison, V
P
/V
S
anomaly recovery averages
>80 per cent only along the northern and central regions of the
peninsula. Below 40 km depth, recovery decreases to 50–80 per cent
for both V
P
and V
P
/V
S
, though the general pattern of the prescribed
anomalies remains (Fig. 5). The dipping nature of the earthquake
data set and a decrease in S waves recorded at the OBS introduce
two significant artefacts attributed to ray path geometry:
(1) a high V
P
artefact beneath the most landward OBS (Fig. 5,
depth 0 and 9 km) and
(2) a high V
P
/V
S
artefact parallel to the subducting slab at 30–
65 km depth (Fig. 6).
The dip, relative amplitudes and shape of features introduced in
the characteristic model are well recovered, thoughsome smearing is
evident in both the V
P
and V
P
/V
S
models (Fig. 6). As noted in plane
view, V
P
/V
S
resolution is significantly poorer than V
P
resolution.
Anomaly recovery decreases 170–220 km from the trench axis, and
hence the high V
P
,lowV
P
/V
S
velocity anomaly placed at 40–65 km
depth to represent the expected location of the forearc mantle wedge
could beresolved onlyat the tipof thecontinental Moho/oceanic slab
intersection. Amplitude recovery of the seaward tip of this anomaly
only reaches 65 per cent of the initial 10 per cent and 15 per
cent prescribed V
P
and V
P
/V
S
, respectively (Fig. 6). At these depths,
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
114 H. R. DeShon et al.
Figure 5. Depth planes through the characteristic velocity model (columns 1 and 3) and restoring resolution V
P
(column 2) and V
P
/V
S
(column 4) results
using synthetic arrivals times with noise. Colours represent velocity anomaly from the initial 1-D velocity model. Stations are shown as open triangles. Features
discussed in text marked with dashed lines.
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
Seismogenic zone structure beneath the Nicoya Peninsula 115
Figure 6. Representative cross-section along Profile 4 (shown in Fig. 3) through the characteristic V
P
and V
P
/V
S
velocity models (column 1), and the recovered
V
P
and V
P
/V
S
models (column 2). Colours represent velocity anomaly from the initial 1-D velocity model; contours represent absolute V
P
or V
P
/V
S
.Tand C
represent the MAT and coastline locations, respectively. Features discussed in text marked with dashed lines.
this anomaly recovery corresponds to a V
P
amplitude difference of
0.5 km s
1
and a V
P
/V
S
difference of 0.2 compared to the abso-
lute V
P
and V
P
/V
S
in the characteristic velocity model. The synthetic
V
P
/V
S
model clearly fails to recover the abrupt change in V
P
/V
S
along the dipping slab feature at 30 km depth (Fig. 6), limiting
interpretation of the oceanic crust/mantle V
P
/V
S
.
Calculation of the full resolution matrix for the last iteration of
each progressively finer inversion yields important information on
spatial resolution. DWS and RDE largely track over the model space
while spread function varies over the Nicoya Peninsula study area.
Based on the results from the characteristic V
P
and V
P
/V
S
models
(Figs 5 and 6), we use regions with spread function 2.5, RDE
>0.04, and DWS 50 to define the resolution boundaries of the
velocity model for interpretation purposes (Figs 7 and 8). Due to the
hypocenter and station geometry, velocities beneath the peninsula
and above the subducting slab are much better resolved than at the
edges of the model.
5RESULTS
The final fine grid results utilize 7598 P- and 5761 S-wave arrivals
(Fig. 3),and results havea combined rms of 0.14 s and data variances
of 0.010 s
2
for P and 0.035 s
2
for SP. The initial rms of the data
set prior to 1-D inversion calculated on a subset of 475 events was
0.43 s, and data variances were 0.29 s
2
for P and 0.32 s for SP
(DeShon 2004). Prior to 3-D inversion P and SP data variances
were 0.04 s
2
and 0.11 s
2
, respectively, with a total rms of 0.16 s.
Note the largest drop in variance occurs during 1-D inversion for
hypocenter parameters, velocity model and station corrections.
The mean absolute value changes in event locations following
3-D inversion are 0.54, 0.67 and 1.64 km (north, east, depth) rela-
tive to the 1-D inversion locations. Changes relative to initial cat-
alogue locations calculated using the IASP91 velocity model are
0.35 s origin time and 0.84, 1.01 and 2.86 km (north, east, depth).
The algorithm provides formal 1σ hypocenter uncertainties, which
do not account for systematic biases and may significantly under-
estimate true errors (e.g. Husen et al. 2003b). The 2σ hypocen-
ter uncertainty estimates are 0.10 s origin time and 0.58, 0.60 and
1.04 km (north, east, depth). The 2σ uncertainty spans the mean
location shift moving from the 1-D to 3-D inversion solution fairly
well.
The correlation between earthquake activity and seismic velocity
structure varies with depth. Fig. 7 shows depth slices through the
final 3-D model in absolute V
P
and V
P
/V
S
and in relative V
P
and
V
P
/V
S
anomalies compared to the 1-D initial model. Fig. 8 shows
V
P
and V
P
/V
S
cross-sections in absolute and relative anomaly per-
pendicular to the MAT (shown in Fig. 3) and includes earthquakes
located within 7.75 km of the cross-section plane. Regions with
DWS <50 and spread function >2.5 are masked in grey to indi-
cate areas of poorer resolution. For clarity, results are discussed by
feature.
5.1 Station corrections and surface geology
Structure above 5 km depth is not adequately sampled by this data
set. The peninsula largely appears faster than the 1-D model pre-
diction, possibly due to the high-velocity ophiolites composing the
Nicoya Complex underlying the peninsula. The small and spatially
dispersed P-wave station corrections for land seismometers indicate
little variancein near surface velocities (Table 2).S-wavecorrections
for the land stations reflect the presence of surface sediments. Cor-
rections are skewed towards high, positive values suggesting V
P
/V
S
is too high beneath the stations, an artefact noted with synthetic
data. High, positive P-wave corrections at OBS located along the
forearc are likely due to model sensitivity in this region (Figs 5 and
6). V
P
/V
S
station corrections for the OBS are relatively low, which
reflect the small number of SP traveltimes picked (generally <20
S-wave arrivals/station) on the noisier OBS.
5.2 Plate interface seismicity and velocities
Previous P-wave refraction studies along Line 101 (Fig. 2) pre-
sented plate interface interpretations for the central Nicoya Penin-
sula (Christeson et al. 1999; Sallar`es et al. 1999, 2001). Along
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
116 H. R. DeShon et al.
Table 2. Station locations with V
P
and V
S
station corrections for the Nicoya
Peninsula experiment.
Station Lat (
) Lon (
) P Correction (s)
S Correction (s)
BANE 9.93 84.96 0.05 0.33
BONG 9.75 85.21 0.13 0.45
COND 10.58 85.67 0.03 0.20
CORO 9.97 85.17 0.08 0.31
COYO 10.36 85.65 0.02 0.36
GRAN 10.34 85.85 0.10 0.19
GUAI 10.27 85.51 0.02 0.38
GUIO 9.92 85.66 0.18 0.47
HATA 10.43 85.28 0.15 0.25
HOJA 10.05 85.42 0.07 0.32
INDI 9.86 85.50 0.11 0.42
JTS 10.29 84.95 0.22 0.15
JUDI 10.17 85.54 0.01 0.28
MARB 10.06 85.75 0.09 0.37
NB01 9.50 86.20 0.00 0.00
NB03 9.70 86.16 0.14 1.07
NB04 9.70 86.30 0.00 0.00
NB05 9.88 86.20 0.01 0.07
NB06 9.78 86.03 0.05 0.14
NB07 9.68 85.94 0.01 0.07
NB08 9.72 85.76 0.39 0.51
NB09 9.84 85.90 0.40 0.00
NB10 9.95 86.00 0.35 0.21
NB11 10.05 85.90 0.45 0.15
NB12 9.92 85.80 0.46 0.00
NB13 9.83 85.68 0.38 0.02
NB14 9.75 85.55 0.27 0.00
PA PA 10.59 85.68 0.01 0.17
PARG 10.20 85.82 0.09 0.36
PNCB 9.59 85.09 0.01 0.42
PNUE 9.84 85.34 0.16 0.41
SAJE 10.61 85.45 0.11 0.28
SARO 10.84 85.61 0.15 0.11
TFER 10.21 85.27 0.06 0.34
VAIN 9.78 85.01 0.07 0.33
VIMA 10.14 85.63 0.08 0.19
Profile 4 (Fig. 3), which broadly parallels Line 101, the plate in-
terface is set to the refraction interpretations. The Nicoya Peninsula
minimum 1-D model used as the initial condition for velocity inver-
sion closely mimics expected velocities for the top of the oceanic
plate. Thus we expect the 0 per cent V
P
anomaly contour to represent
the top of the subducting slab (Figs 7 and 8). Comparison between
the refraction data and the 5 per cent to 0 per cent V
P
anomaly
contour is very good along Profile 4, and hence the 5 per cent to
0 per cent anomaly contour is used to guide the choice of the top of
the subducting plate interface to about40 km depth on other profiles.
Resolution beneath the OBS array limits identification of the plate
interface at <9kmdepth, but intuitively the plate interface should
move progressively deeper inland starting from the seafloor depth
at the trench (4 km; Fig. 7a). Below 40 km depth the plate in-
terface interpretation is based on the location of the interface from
refraction data and on the extension of plate curvature. We use a
continuous curvature interpolation technique to contour the surface
of the plate interface between cross-sections (Smith & Wessel 1990)
and project them onto the velocity images to facilitate interpretation
(Figs 7 and 8).
We distinguish between interplate and intraplate seismicity across
the Nicoya Peninsula by interpreting those earthquakes occurring
directly on or above the plate interface (5to+8 km) between
0–40 km depths as interplate seismogenic zone events. This method
helps to account for errors in plate interface interpretation. Calcu-
lation of focal mechanisms is ongoing, but the majority of events
interpreted as interplate in this study are consistent with under-
thrusting along the plate interface (Hansen et al., ‘Relocations and
focal mechanisms determined from waveform cross-correlation of
seismic data from the Nicoya Peninsula, Costa Rica’, submitted to
Bull. Seis. Soc. Am.). We define the up- and downdip limits of
interplate microseismicity using the range of interplate seismicity
(Fig. 9). The Nicoya earthquakes have low errors (1kmat2σ ),
and the Gaussian distribution approach outlined by Pacheco et al.
(1993) commonly utilized to define the updip and downdip limits of
seismicity (e.g. Husen et al. 1999; DeShon et al. 2003) may under-
estimate the extent of the plate interface hosting small magnitude
earthquakes. Consideration of data range places the updip limit for
CNS and EPR sides of the plate at 12 and 17 km depth, respectively,
and the downdip limits at 26 and 28 km depth, respectively. A robust
offset in depth at onset of microseismicity (5 km) remains at the
transition between CNS and EPR crust (Fig. 9). This depth offset
was first noted by Newman et al. (2002), who estimated an 10 km
offset, with EPR microseismicity beginning at 20 km depth BSL
and CNS microseismicity at 10 km depth BSL. The mean absolute
location differences between this study and Newman et al. (2002)
are 0.76, 0.99 and 2.28 km (north, east, depth). Along the subducted
plate suture, microseismicity extends nearly 5 km closer to the MAT
in map view, locally extending the updip extent of microseismicity
to 10 km depth (Fig. 9).
We examine five cross-sections along the peninsula to investi-
gate small-scale variability in earthquake locations and velocities
and to identify inter- and intraplate events. One source of error in
this approach results from the arbitrary cut-off used to distinguish
between shallow interplate and intraslab seismicity beneath 40 km
depth. Intraslab earthquakes may be misidentified as interplate, and
vice versa.
An 11 km southwestward step in the plate interface is apparent
near the projected downdip EPR/CNS plate suture, indicating that
the subducting CNS-origin crust dips more shallowly than does the
EPR-generated crust in the north (Figs 7a and b). Steepening of the
plate interface from south to north has been observed in previous
refraction and Wadati-Benioff zone studies along Costa Rica (e.g.
Protti et al. 1994, 1995a; Ye et al. 1996). Below 20 km depth this
offset is not apparent (Fig. 7c).
Interplate seismicity generally forms a dipping planar feature, but
in the southern Nicoya Peninsula microseismicity occurs in a more
diffuse region around the plate interface (Fig. 8a, Profiles 1 and 2).
Microseismicity along the south of the peninsula begins between 7.5
and 11.5 km depth but well landward of the plate interface (Figs 7a
and 8a, Profiles 1 and 2). By 13–14.5 km depth in the southern and
by 14.5 and 17.5 km in the northern Nicoya Peninsula microseis-
micity occurs along the plate interface (Fig. 8). Activity along the
EPR/CNS suture, approximately located along Profile 3 (Fig. 3),
notably decreases by 17.5–22.5 km depth (Fig. 8a). A decrease in
the spatial density of microseismicity across the CNS/EPR bound-
ary, with fewer events occurring on the EPR side, does not appear
to be due to a bias in earthquake selection since station coverage is
denser along the northern Nicoya Peninsula.
Interplate activity is associated with a negative V
P
anomaly that
parallels the plate interface to a depth of 20 km, with absolute ve-
locities of 5.2–5.8 km s
1
(Fig. 8a). Along the EPR interface the
updip limit of interplate microseismicity abruptly occurs where
the negative V
P
anomalies terminate (Fig. 8a, Profiles 3–5). Along
the northern Nicoya Peninsula, a small, negative V
P
/V
S
anomaly
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
Seismogenic zone structure beneath the Nicoya Peninsula 117
Figure 7. Depth planes through absolute and relative V
P
(column 1 and 2) and absolute and relative V
P
/V
S
(column 3 and 4). Earthquakes (solid circles) are
projected to nearest plane, and the plate interface contours at each depth (red line; PI) are interpreted as described in the text.
coincides with reduced V
P
(Fig. 8a). These absolute velocities are
broadlyconsistent with predicted valuesfor margin wedgeof Nicoya
Complex composition (Sallar`es et al. 1999, 2001) and oceanic
upper crust velocities (i.e. Walther & Flueh 2002). Langseth &
Silver (1996) and Ranero et al. (2003) proposed that normal faulting
within the Cocos Plate along the outer rise introduces water into the
oceanic crust, which could account for the noted V
P
reduction. Al-
ternately, the EPR oceanic crust may be hydrated prior to reaching
the outer rise normal faults due to hydrothermal circulation (Hutnak
et al. 2006).
5.3 Continental Moho location and mantle
wedge velocities
Previous studies placed the continental Moho between 32 and 38 km
depth beneath the Nicoya Peninsula and suggested that the velocity
contrast across the crust/mantle boundary is very small (Matumoto
et al. 1977; Protti et al. 1996; Quintero & Kulhanek 1998; Sallar`es
et al. 2000, 2001; Husen et al. 2003a; DeShon & Schwartz 2004).
Inspection of profiles through the V
P
model indicate a turnover in
the 7.0–7.2 km s
1
V
P
contours between 30 and 40 km (Fig. 8) that
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
118 H. R. DeShon et al.
Figure 8. (a) Cross-sections through relative V
P
(column 1) and V
P
/V
S
(column 2). Earthquakes (open circles) and stations (inverted triangles) are projected
to the nearest cross-section. Plate interface (black line) and continental Moho (dash-dot black line) interpretations are described in the text. Contours represent
absolute V
P
and V
P
/V
S
.Regions of lower resolution are shaded grey. (b) Resolution of profiles through the V
P
and V
P
/V
S
models. Shown is derivative weighted
sum (DWS). The thick black lines mark regions where DWS < 50, RDE > 0.04, and spread function 2.5. Profile locations from Fig. 3.
we infer to represent the continental Moho. Such P-wave veloci-
ties are consistent with lower continental crust based on refraction
data (Sallar`es et al. 1999, 2001). However, we recognize that the
error associated with this interpretation cannot be constrained due
to the expectedly small velocity contrast. To the south, the continen-
tal Moho intersects the downgoing slab at 30 km depth (Profile
2, Fig. 8a). The mantle forearc wedge exhibits slow P-wave veloc-
ities, averaging 6.8–7.0 km s
1
(Fig. 8a), and the downdip limit of
seismogenic zone seismicity is 5kmupdip of the intersection of
the continental Moho and the subducting slab. Along the northern
Nicoya Peninsula, the continental Moho intersects the slab at 30 km
depth withmantle wedge V
P
at 7.0–7.2km s
1
(Profiles 3–5,Fig. 8a).
V
P
/V
S
in the mantle wedge along EPR profiles ranges between 1.73
and 1.78, and interplate seismicity occurs 10 km updip of the
Moho/slab intersection (Fig. 8). Overall, along the Nicoya Penin-
sula the continental Moho intersects the downgoing slab between
30 and 34 km depth.
5.4 Caribbean crust seismicity and velocity
Refraction studies constrained continental crust V
P
to be 4.5–
7.0 km s
1
, increasing with depth (Christeson et al. 1999; Sallar`es
et al. 1999, 2001). Sallar`es et al. (1999, 2001) dividedthe continental
crust into upper, mid and lower levels with corresponding P-wave
velocities of 5.3–5.7 km s
1
(0–10 km BSL), 6.2–6.5 km s
1
(5–
20 km BSL) and 6.9–7.3 km s
1
(20–40 km BSL). Fig. 10 super-
imposes the crustal structure from Sallar`es et al. (1999, 2001) on a
generalized interpretation of subduction zone structure along Pro-
file 4 (see Fig. 3 for profile location). The V
P
of the two models
of continental crustal structure generally agree, with the largest V
P
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
Seismogenic zone structure beneath the Nicoya Peninsula 119
Figure 8. (Continued.)
discrepancy occurring beneath the coastline between 0 and 15 km
depth.
Decreasing velocity model resolution and the lack of refraction
data beneath the southern Nicoya Peninsula complicates interpreta-
tion of crustal versus interplate seismicity. Along Profile 1, seismic-
ity forms a dipping plane generally paralleling the plate interface but
outside the bounds used in this study (Fig. 8a). These earthquakes
occur within the upper plate, but the events are likely related to local
subduction processes.
5.5 Oceanic plate velocities and intraslab earthquakes
Earthquakes occurring within the oceanic crust form a sparsely
populated dipping feature that extends from 18 to >80 km depth
(Fig. 9). We primarily interpret the oceanic Moho by maintaining
acrustal thickness of 7 km, consistent with refraction information
(Fig. 10) (Ye et al. 1996; Christeson et al. 1999; Sallar`es et al. 1999,
2001; Walther & Flueh 2002). In the northern Nicoya Peninsula, the
oceanic mantle appears relatively fast in V
P
, ranging between 8.1
and 8.4 km s
1
.Variations in V
P
within the oceanic crust and man-
tle beneath the southern Nicoya Peninsula are likely due to low ray
coverage (i.e. poor solution quality). Because of the limited timing
and distribution of the OBS and seismicity, an insufficient number
of long ray paths penetrated the mantle beneath the oceanic crust
for analysis.
6DISCUSSION
Possible mechanisms controlling the transition from aseismic to
seismic behaviour along the updip limit of seismogenic zones
include the mechanical backstop model (Byrne et al. 1988),
temperature-controlled mineral transition models (Vrolijk 1990;
Hyndman et al. 1997), and combinations of pressure- and
temperature-dependent material characteristics such as variable
pore fluid pressure and cementation (Moore & Saffer 2001; Saffer
& Marone 2003; Moore et al. 2006). Comparisons of thermal mod-
elling to seismogenic zone seismicity rupture limits across many
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
120 H. R. DeShon et al.
0
20
40
60
)mk( htpe
D
0
20
40
60
)mk( htpeD
40 60 80 100 120 140 160 180
Distance along profile (km)
0
20
40
60
0
20
40
60
40 60 80 100 120 140 160 180
Distance along profile (km)
(a) CNS (Profile 1 and south)
(b) CNS (Profile 2)
(c) EPR (Profile 3)
(d) EPR (Profiles 4, 5, and north)
Figure 9. Cross-sections moving south to north through hypocenters resulting from 3-D inversion. Earthquakes are shown with error bars and divided
into earthquakes identified as seismogenic zone interplate seismicity (black) and intraplate seismicity (grey). Stations are projected onto the nearest profiles
(triangles). The preferred plate interface interpretation from each cross-section is shown for reference (dark grey line). Note the increasing dip angle of
subduction from south to north (from a to d). Profile locations are shown in Fig. 3. All seismicity and stations occurring south of Profile 1 and north of Profile
5 are included in cross-sections a and d, respectively.
Figure 10. Generalized interpretation of the Nicoya Peninsula 3-D velocity models and associated hypocenters. Colour and values indicate V
P
; V
P
/V
S
shown
in italics. The observed velocities agree well with P-wave velocities and crustal structure (dashed black lines) from wide-angle refraction data (Sallar`es et al.
2001). The updip limit of interplate seismicity (red circles) may correspond to the mid crust/oceanic slab interface, identified by the transition from 5.8 to
6.4 km s
1
P-wave velocities. The continental Moho (Mc) appears between 32 and 40 km depth BSL, and the downdip limit of microseismicity clearly occurs
updip of the continental Moho/oceanic slab intersection. The forearc mantle wedge may be 15–30 per cent hydrated based on decreased P-wave velocities and
slightly elevated V
P
/V
S
. The oceanic Moho (Mo) is well resolved, and oceanic mantle P-wave velocities are normal with slightly elevated V
P
/V
S
.
subduction zones have shown that seismicity generally occurs be-
tween the 100–150
C and 350
C isotherms (Hyndman & Wang
1993; Hyndman et al. 1997; Oleskevich et al. 1999; Currie et al.
2002; Harris & Wang 2002). For low-temperature subduction zones
and oceanic subduction zones, Hyndman et al. (1997) suggested the
downdip transition occurs when the downgoing plate encounters the
upper forearc mantle before temperatures reach 350
C, the onset of
ductility in common subduction component materials. Dehydration
of subducted sediments and oceanic crust at these depths can in-
troduce water into the continental mantle. Serpentinite, a hydrated
peridotite-related mineral assemblage possiblycreated by this water,
has been shown to be conditionally stable in a frictional sense, ex-
hibiting seismic behaviour under increasing applied stress or higher
rupture velocities (e.g. Hyndman & Peacock 2003).
Previous studies utilizing the CRSEIZE earthquake data set sug-
gested that the updip limit of microseismicity across the Nicoya
Peninsula broadly corresponded to the 100–150
C isotherm esti-
mate along the plate interface (Harris & Wang 2002; Newman et al.
2002). Spinelli & Saffer (2004) performed a 3-D thermal model for
the region and concluded that the onset of seismicity likely repre-
sented the transition to stick-slip behaviour due to fluid pressure
dissipation and increased effective normal stress. Both the Harris &
Wang (2002) and Spinelli & Saffer (2004) thermal models incor-
porated cooling due to 0–2 km of hydrothermal circulation on the
EPR-derived crust and varied shear heating along the subduction
thrust to assess the downdip thermal affects of EPR and CNS litho-
sphere. Because geodetically derived interface locking models and
prior large earthquakes show that the shallow subduction zone updip
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
Seismogenic zone structure beneath the Nicoya Peninsula 121
Figure 11. Geodetically inferred locking and interplate seismicity (orange
circles) along the plate interface. Interplate seismicity occurs downdip of a
major zone of locking which corresponds to nearly 75 per cent of the plate
rate in this region. Microseismicity concentrates where the plate creeps.
Geodetic model from Norabuena et al. (2004). Temperature isotherms from
Spinelli & Saffer (2004) updip and Harris & Wang (2002) downdip; the grey
box indicates the potential variance in the modelled 300
C isotherm. Thick
grey line marked ‘Moho’ represents the intersection of the subducting plate
with the continental Moho estimated in this study.
of 20 km depth is susceptible to failure in large earthquakes, the
CRSEIZE microseismicity does not constrain the upward extent of
potential subduction megathrust earthquakes in this region (Fig. 11)
(Norabuena et al. 2004).
Schwartz & DeShon (2006) compared the earthquake relocations
presented in this study with geodetic locking models (Norabuena
et al. 2004) and thermal models (Harris & Wang 2002; Spinelli
& Saffer 2004). They concluded that plate interface microseis-
micity more closely correlates to the 250
C isotherm estimate, a
temperature at which most mineral transitions in oceanic sediments
have already taken place (Fig. 11), while the onset of shallow
plate locking corresponds well with the modelled 100
C isotherm
(Fig. 11). Based on this observation, Schwartz & DeShon (2006)
suggested that two transitions in mechanical behaviour occur along
the plate interface:
(1) the transition to stick-slip behaviour as marked by increased
locking across the plate interface and
(2) the transition to a geodetically creeping zone that is capa-
bleofproducing microseismicity. The transition to creep and mi-
croseismicity may result from increased pore fluid pressure across
the plate interface resulting from basalt dehydration reactions in a
hydrated oceanic plate and/or decreased permeability in the upper
plate that both begin at 250
C. Small decreases in effective normal
stress maylead to fault weakening andgeneration of microseismicity
downdip of the more strongly coupled, geodetically locked region.
This hypothesis explains why the onset of microseismicity reflects
the thermal state of the incoming plate but corresponds to higher
temperatures.
Alternately, the onset of microseismicity between 12 and 17 km
depth may mark a small modification in frictional stability across
the plate interface due to a change in composition of the upper
plate. Continental crustal structure interpretations for the Nicoya
Peninsula region based on P-wave refraction data suggested Nicoya
Complex flood basalts transition to a basalt–gabbro composition at
these depths (Sallar`es et al. 2001). Nicoya Complex materials are
expected tohavea lower V
P
than mid-crust level basalt–gabbro rocks
along the plate interface (Sallar`es et al. 2001), similar to imaged
V
P
at the onset of interplate seismicity (5.8–6.4 km s
1
;Fig. 10).
This model does not explain the along-strike depth transition at the
CNS–EPR plate suture nor the temperature correlation. However,
steepening of the slab across the suture could lead to a vertical offset
of the mid-crust/oceanic slab intersection seen in microseismicity.
The degree of subduction erosion beneath the Nicoya forearc
could vary spatially along-strike and explain local variations in seis-
micity along the peninsula. Subduction erosion may be indirectly
related to temperature if hydrofracturing of the upper plate plays
akey role in the erosional process, as proposed by Ranero & von
Huene (2000) and von Huene et al. (2004). Additionally erosion
could introduce a large amount of upper plate material into the sub-
duction channel (von Huene et al. 2004). The amount and depth
range of sediment, fault gouge, and basalt dewatering and consol-
idation is affected by temperature. The above described processes
could decrease pore fluid pressure and increase effective normal
stress or change the mechanical strength of the plate contact zone
with depth, perhaps explaining the variable onset of locking and
seismicity along the plate interface.
The downdip limit of interplate microseismicity along the Nicoya
Peninsula occurs at 26–28 km depth BSL. The continental Moho
intersects the downgoing slab in this region between 30 and 35
km depth, 95 and 105 km from the MAT, downdip of seismogenic
zone microseismicity (Fig. 10). Thermal modelling studies allow-
ing small amounts of shear heating have placed the 300–350
C
isotherm, the expected temperature controlling the onset to duc-
tile behaviour, less than 100 km from the trench (Harris & Wang
2002; Peacock et al. 2005). One caveat, however, is that, like the
updip limit, inspection of aftershock regions of M
w
> 6.9 earth-
quakes indicate that the potential seismogenic zone extends beyond
the downdip extent of the interplate microseismicity we recorded
in the interseismic period (compare Figs 1 and 11). Error bars as-
sociated with the 300–350
C isotherm overlap with the continental
Moho/oceanic slab intersection, precluding a unique interpretation
of processes involved in controlling the downdip extent of micro-
seismicity (Fig. 11).
DeShon & Schwartz (2004) reported multiple lines of evidence,
including receiver function modelling at GSN station JTS, for the
presence of serpentinized mantle wedge beneath northern Costa
Rica. Previous 3-D tomography studies of Costa Rica suggested
that the mantle wedge south of the Nicoya Peninsula was anhydrous,
though central Costa Rica forearc mantle appeared hydrated (Husen
et al. 2003a). The Nicoya Peninsula study region discussed here oc-
curs predominantly northwest of and has greater depth resolution
than the most recently calculated country-wide tomography study
(Husen et al. 2003a); in particular, the forearc mantle wedge be-
neath the peninsula appears localized in a region of poor resolution
in the country-wide model. In the Husen et al. (2003a) model, V
P
ranged from 6.8 to 7.2 km s
1
in the forearc mantle with significant
along-strike variability. This study indicates that low V
P
values (7.0–
7.6 km s
1
)vary little along-strike, suggesting fairly uniform hydra-
tion of the forearc mantle wedge.
The observed P-wave velocity reduction reported in this study
corresponds to serpentinization of 20–25 per cent (Fig. 10) (i.e.
Christensen 1966; Carlson & Miller 2003). Resolution modelling
suggests anomaly recovery in the forearc mantle region of 65–
70 per cent over the initial model. This implies V
P
anomalies may be
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
122 H. R. DeShon et al.
underestimated and that the degree of serpentinization in the forearc
wedge may be higher. The theoretical increase in V
P
/V
S
,which
should push V
P
/V
S
over 1.80 throughout the wedge, is not observed,
and ratios spatially vary from 1.70 to 1.80, or from anhydrous to 10
per cent serpentinization (Carlson & Miller 2003). Choice of V
P
/V
S
damping parameter additionally complicates the interpretation; tests
using lower V
P
/V
S
damping lead to greater model variance and
locally increased V
P
/V
S
amplitudes to >2.0.
7CONCLUSIONS
The depth extent and width of the seismogenic zone illuminated
by interseismic microseismicity varies along-strike of the Middle
America subduction zone. Beneath the southern Nicoya Peninsula,
interplate microseismicity extends from 12 km BSL, 70–75 kmfrom
the trench, to 26 km BSL, 100 km from the trench. In the northern
Nicoya Peninsula, interplate seismicity extends from 17 to 28 km
BSL, 75 to 88 km from the trench. The downdip limit of interplate
microseismicity on the EPR crust occurs 12 km closer to the trench
than it does on the CNS crust due to the steepening of the subducting
plate. The updip extent is 5kmdeeper north of the CNS–EPR
plate suture. Along the suture both the updip and downdip extent of
interplate microseismicity steps trenchward.
Geodetic modelling and interplate microseismicity provide ap-
parently disparate spatial constraints on seismogenic zone updip
and downdip limits (Norabuena et al. 2004; Schwartz & DeShon
2006; this study). The updip limit of interplate microseismicity is
associated with the termination of anomalously low V
P
along the
plate interface and the 200–250
C isotherm estimates from thermal
modelling. This change in velocity may represent a transition from
ophiolitic Nicoya Complex to mid-crust compositions with increas-
ing depth along the subduction thrust, while the onset of seismicity
may be a mechanical effect of changing composition in the upper
plate. Alternately, low V
P
may indicate increased hydration in the
upper and/or lower plates, and microseismicity may result due to
changes in pore fluid pressure with increasing depth due to dehy-
dration reactions within the subducting slab (Schwartz & DeShon
2006). The downdip edge of plate interface microseismicity nei-
ther directly correlates to the intersection of the continental Moho
and oceanic crust nor to the modelled seaward extent of the brit-
tle/ductile transition at 300–350
C. Low P-wave velocities in the
mantle wedge indicate hydration and serpentinization of 15–25 per
cent beneath the Nicoya Peninsula, allowing for a conditionally sta-
ble mantle wedge (DeShon & Schwartz 2004). Though the mantle
wedge may rupture during large earthquakes, during the current in-
terseismic period, the frictional stability state apparently does not
allow microseismicity nucleation.
Multiple studies resulting from the CRSEIZE program, including
the local earthquake tomography study presented here, have shown
that interseismic microseismicity along the Nicoya Peninsula mar-
gin does not reflect the extent of the potential rupture area of sub-
duction megathrust earthquakes. Hence, microseismicity may be a
temporally variable feature within the earthquake cycle as loading
increases along the plate interface, and, even when present, inter-
plate microseismicity may not uniquely define the full seismogenic
thrust interface. Although temperature may serve as a good proxy
for defining the current seismogenic zone, deciding which isotherm
describes the seismogenic zone over an entire seismic cycle is dif-
ficult; it may depend on the local subduction zone setting, and of
course is affected by uncertainties in thermal models. As exempli-
fied by the Costa Rica margin, the extent of locking along the sub-
duction thrust, historic rupture dimension, temperature, hydration
state and location of microseismicity must be used in combination
to assess the potential for earthquake rupture along the subduction
megathrust.
ACKNOWLEDGMENTS
We thank Alan Sauter and Sharon Escher for the OBS fieldwork and
data processing components of this project, GEOMAR for the use of
R/V Sonne cruise SO-144 leg 3b for OBS deployment, and the R/V
Reville for OBS recovery. Some instruments used in the field pro-
gram were provided by the PASSCAL facility of the Incorporated
Research Institutions for Seismology (IRIS) through the PASSCAL
Instrument Center at New Mexico Tech. Land data collected during
this experiment are available through the IRIS Data Management
Center, and OBS data are available by request through LMD at SIO.
The facilities of the IRIS Consortium are supported by the National
Science Foundation under Cooperative Agreement EAR-0004370.
Figures made using GMT (Wessel & Smith 1998) and TOMO2GMT
available through Stephan Husen. Constructive reviews by G. Laske,
C. Thurber, S. Husen, and an anonymous reviewer significantly im-
proved the manuscript. This work was supported by NSF grants
OCE 9910609 and EAR0229876 to SYS, OCE9910350 to LMD
and OCE 9905469 to THD.
REFERENCES
Aki, K. & Lee, W.H.K., 1976. Determination of three-dimensional anoma-
lies under a seismic array using first P arrival times from local earth-
quakes, 1. A homogeneous initial model, J. geophys. Res., 81, 4381–
4399.
Avants, M., Schwartz, S.Y., Newman, A.V. & DeShon, H.R., 2001. Large
underthrusting earthquakes beneath the Nicoya Peninsula, EOS Trans.
Am. geophys. Un., 82, F1266.
Barckhausen, U.,Roeser, H.A.& von Huene, R., 1998. Magnetic signature of
upper plate structures and subducting seamounts at the convergent margin
off Costa Rica, J. geophys. Res., 103, 7079–7094.
Barckhausen, U., Ranero, C.R., von Huene, R., Cande, S.C. & Roeser, H.A.,
2001. Revised tectonic boundaries in the Cocos Plate off Costa Rica:
implications for the segmentation of the convergent margin and for plate
tectonic models, J. geophys. Res., 106, 19 207–19 220.
Byrne, D.E., Davis, D.M. & Sykes, L.R., 1988. Loci and maximum size of
thrust earthquakes and the mechanics of the shallow region of subduction
zones, Tectonics, 7, 833–857.
Carlson, R.L. & Miller, D.J., 2003. Mantle wedge water contents estimated
from seismic velocities in partially serpentinized peridotites, Geophys.
Res. Lett., 30, 1250, doi:10.1029/2002GL016600.
Christensen, N., 1966. Elasticity of ultrabasic rocks, J. geophys. Res., 71,
5921–5931.
Christeson, G.L., McIntosh, K.D., Shipley, T.H., Flueh, E.R. & Goedde, H.,
1999. Structure of the Costa Rica convergent margin, offshore Nicoya
Peninsula, J. geophys. Res., 104, 25443–25 468.
Crosson, R.S., 1976. Crustal structure modeling of earthquake data, 1,simul-
taneous least squares estimation of hypocenter and velocity parameters,
J. geophys. Res., 8, 3036–3046.
Currie, C.A., Hyndman, R.D., Wang, K. & Kostoglodov, V., 2002. Ther-
mal models of the Mexico subduction zone: implications for the
megathrust seismogenic zone, J. geophys. Res., 107, 2370, doi:10.1029/
2001JB000886.
DeMets, C., 2001. A new estimate for present-day Cocos-Caribbean plate
motion: implications for slip along the Central American volcanic arc,
Geophys. Res. Lett., 28, 4043–4046.
DeShon, H.R., 2004. Seismogenic zone structure along the Middle America
subduction zone, Costa Rica, PhD thesis, p. 356, University of California,
Santa Cruz.
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
Seismogenic zone structure beneath the Nicoya Peninsula 123
DeShon, H.R. & Schwartz, S.Y., 2004. Evidence for serpentinization of the
forearc mantle wedge along the Nicoya Peninsula, Costa Rica, Geophys.
Res. Lett., 31, L21611, doi:10.1029/2004GL021179.
DeShon, H.R., Schwartz, S.Y., Bilek, S.L., Dorman, L.M., Gonzalez, V.,
Protti, J.M., Flueh, E.R. & Dixon, T.H., 2003. Seismogenic zone structure
of the southern Middle America Trench, Costa Rica, J. geophys. Res., 108,
2491, doi:10.1029/2002JB002294.
Eberhart-Phillips, D., 1986. Three-dimensional velocity structure in North-
ern California Coast Ranges from inversion of local earthquake arrival
times, Bull. seism. Soc. Am., 76, 1025–1052.
Eberhart-Phillips, D., 1990. Three-dimensional P and S velocity structure in
the Coalinga region, California, J. geophys. Res., 95, 15 343–15 363.
Eberhart-Phillips, D. & Reyners, M., 1997. Continental subduction and
three-dimensional crustal structure: the northern South Island, New
Zealand, J. geophys. Res., 102, 11 843–11 861.
Eberhart-Phillips, D. & Reyners, M., 1999. Plate interface properties in
the Northeast Hikurangi subduction zone, New Zealand, from converted
seismic waves, Geophys. Res. Lett., 26, 2565–2568.
Ellsworth, W.L., 1977. Three-dimensional structure of the crust and mantle
beneath the island of Hawaii, PhD thesis, p. 327, Massachusetts Institute
of Technology, Cambridge.
Evans, J.R., Eberhart-Phillips, D. & Thurber, C.H., 1994. User’s manual
for SIMULPS12 for imaging Vp and Vp/Vs:aderivative of the Thurber
tomographic inversion SIMUL3 forlocal earthquakesand explosions. U.S.
Geol. Surv. Open-File Report, 94–431, U.S. Government Printing Office.
Fisher, A.T., Stein, C.A., Harris, R.N., Wang, K., Silver, E.A., Pfender,
M., Hutnak, M., Cherkaoui, A., Bodzin, R. & H. Villinger 2003.
Abrupt thermal transition reveals hydrothermal boundary and role of
seamounts within the Cocos Plate, Geophys. Res. Letts., 30, 1550,
doi:10.1029/2002GL016766.
uendel, F., 1986. Seismotectonics of Costa Rica: an analytical view of the
southern terminus of the Middle America Trench, PhD thesis, p. 157,
University of California, Santa Cruz.
Harris, R.N. & Wang, K., 2002. Thermal models of the Middle America
trench at the Nicoya Peninsula, Costa Rica, Geophys. Res. Lett., 29, 2010,
doi:10.1029/2002GL015406.
Haslinger, F. et al., 1999. 3-D crustal structure from local earthquake tomog-
raphy around Gulf of Arta (Ionian region), NW Greece, Tectonophysics,
304, 210–218.
Husen, S., Kissling, E., Flueh, E. & Asch, G., 1999. Accurate hypocentre
determination in the seismogenic zone of the subducting Nazca Plate in
northern Chile using a combined on-/offshore network, Geophys. J. Int.,
138, 687–701.
Husen, S., Kissling, E., & Flueh, E.R., 2000. Local earthquake tomography
of shallow subduction in north Chile: a combined onshore and offshore
study, J. geophys. Res., 105, 28 183–28 198.
Husen, S., Quintero, R., Kissling, E. & Hacker, B., 2003a. Subduction-zone
structure and magmatic processes beneath Costa Rica constrained by local
earthquake tomography and petrological modeling, Geophys. J. Int., 155,
11–32.
Husen S., Kissling, E., Deichmann, N., Wiemer, S., Giardini, D. & Baer, M.,
2003b. Probabilistic earthquake location in complex three-dimensional
velocity models: application to Switzerland, J. geophys. Res., 108(B2),
2077, doi:10.1029/2002JB001778.
Hutnak, M. et al., 2006. The thermal station of 18–24 Ma upper lithosphere
subducting below the Nicoya Peninsula, northern Costa Rica, in The Seis-
mogenicZone of Subduction Thrust Faults, eds Dixon,T.H. & Moore, J.C.,
Columbia University Press, New York, in press.
Hyndman, R.D. & Peacock, S.M., 2003. Serpentinization of the forearc
mantle, Earth planet. Sci. Lett., 212, 417–432.
Hyndman, R.D. & Wang, K., 1993. Thermal constraints on the zone of major
thrust earthquake failure—the Cascadiasubduction zone, J. geophys. Res.,
98, 2039–2060.
Hyndman, R.D., Yamano, M., & Oleskevich, D.A., 1997. The seismogenic
zone of subduction thrust faults, The Island Arc, 6, 244–260.
Iinuma, T. et al., 2004. Inter-plate coupling in the Nicoya Peninsula, Costa
Rica, as deduced from a trans-peninsula GPS experiment, Eart. Plan. Sci.
Lett. 223, 203–212.
Jacobson, R.S., Dorman, L.M., Purdy, G.M., Schultz, A. & Solomon, S.C.,
1991. Ocean Bottom Seismometer facilities available, EOS Trans. Am.
Geophys. Un., 72, 405–515.
Kimura, G., Silver, E.A. & Blum, P., 1997. Proc. Ocean Drilling Program,
Initial Reports 170, p. 458, Ocean Drill. Program, College Station, TX.
Kissling, E., Ellsworth, W.L. & Cockerham, R., 1984. Three-dimensional
structure of the Long Valley Caldera, California, region by geotomogra-
phy, U.S. Geol. Surv. Open File Rep., 84–939, 188–220.
Kissling, E., Ellsworth, W.L., Eberhart-Phillips, D. & Kradolfer, U., 1994.
Initial reference models in local earthquake tomography, J. geophys. Res.,
99, 19 635–19 646.
Kissling, E., Kradolfer, U. & Maurer, H., 1995. VELEST user’s guide-short
introduction, Institute of Geophysics and Swiss Seismological Service, 25
pp, ETH, Zurich.
Kissling, E., Husen, S. & Haslinger, F., 2001. Model parameterization in
seismic tomography: a choice of consequences for the solution quality,
Phys. Eart. planet. Int., 123, 89–101.
Langseth, M.G. & Silver, E.A., 1996. The Nicoya convergent margin-a
region of exceptionally low heat flow, Geophys. Res. Lett., 23, 891–
894.
Lundgren, P., Protti, M., Donnellan, A., Heflin, M., Hernandez, E. &
Jefferson, D., 1999. Seismic cycle and plate margin deformation in Costa
Rica: GPS observations from 1994 to 1997, J. geophys. Res., 104, 28 915–
28 926.
Marshall, J.S. & Anderson, R.S., 1995. Quaternary uplift and seismic cycle
deformation, Peninsula de Nicoya, Costa Rica, Geol. Soc. Am. Bull., 107,
463–473.
Matumoto, T., Othake, M., Latham, G. & Uma˜na, J., 1977. Crustal structure
of southern Central America, Bull. seis. Soc. Am., 67, 121–134.
Michelini, A. & McEvilly, T.V., 1991. Seismological studies at Parkfield: I.
Simultaneous inversion for velocity structure and hypocenters using cubic
b-splines parameterization, Bull. seism. Soc. Am., 81, 524–552.
Moore, J.C. & Saffer, D., 2001. Updip limit of the seismogenic zone beneath
the accretionary prism of southwest Japan: an effect of diagenetic to low-
grade metamorphic processes and increasing effective stress, Geology,
29, 183–186.
Moore, J.C., Rowe, C. & Meneghini, F., 2006. How can accretionary prisms
elucidate seismogenesis in subduction zones?, in The Seismogenic Zone
of Subduction Thrust Faults, eds Dixon, T.H. & Moore, J.C., Columbia
University Press, New York, in press.
Newman, A.V., Schwartz, S.Y., Gonzales, V., DeShon, H.R., Protti, J.M.
& Dorman, L., 2002. Along strike variability in the seismogenic
zone below Nicoya Peninsula, Costa Rica, Geophys. Res. Lett., 29,
doi:10.1029/2002GL015409.
Norabuena, E. et al., 2004. Geodetic and seismic constraints on some seis-
mogenic zone processes in Costa Rica, J. geophys. Res., 109, B11403,
doi:10.1029/2003JB002931.
Oleskevich, D.A., Hyndman, R.D. &Wang, K.,1999. The updip and downdip
limits to great subduction earthquakes: thermal and structural models
of Cascadia, south Alaska, SW Japan and Chile, J. geophys. Res., 104,
14 965–14 991.
Pacheco, J.F. & Sykes, L.R., 1992. Seismic moment catalog of large, shallow
earthquakes, 1900–1989, Bull. seism. Soc. Am., 82, 1306–1349.
Pacheco, J.F., Sykes, L.R. & Scholz, C.H., 1993. Nature of seismic coupling
along simple plate boundaries of the subduction type, J. geophys. Res.,
98, 14 1333–14 159.
Peacock, S.M., van Keken, P.E., Holloway, S.D., Hacker, B.R., Abers, G.A.
& Furgason, R.L., 2005. Thermal structure of the Costa Rica-Nicaragua
subduction zone, Phys. Eart. plan. Sci. Int., 149, 187–200.
Protti, M., G¨uendel, F. & McNally, K., 1994. The geometry of the Wadati-
Benioffzone under southern CentralAmerica andits tectonicsignificance:
results from a high-resolution local seismographic network, Phys. Earth
planet. Int., 84, 271–287.
Protti, M., G¨uendel, F. & McNally, K., 1995a. Correlation between the age of
the subducting Cocos plate and the geometry of the Wadati-Benioff zone
under Nicaragua and Costa Rica, in Geologic and Tectonic Development
of the Caribbean Plate Boundary in Southern Central America, Special
Paper 295, ed. Mann, P., GSA, Boulder, CO, 309–326.
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
124 H. R. DeShon et al.
Protti, M. et al., 1995b. The March 25, 1990 (M
W
= 7.0, M
L
= 6.8), earth-
quake at the entrance of the Nicoya Gulf, Costa Rica: its prior activity,
foreshocks, aftershocks, & triggered seismicity, J. geophys. Res., 100,
20 345–20 358.
Protti, M., Schwartz, S.Y. & Zandt, G., 1996. Simultaneous inversion for
earthquake location and velocity structure beneath central Costa Rica,
Bull. seism. Soc. Am., 86, 19–31.
Protti, M., G¨uendel, F. & Malavassi, E., 2001. Evaluaci´on del Potencial
ısmico de la Pen´ınsula de Nicoya, Editorial Fundacion UNA, Heredia,
Costa Rica, p. 144.
Quintero, R. &G¨uendel, F., 2000. Stressfield in Costa Rica, CentralAmerica,
J. Seis., 4, 297–319.
Quintero, R. & Kissling, E., 2001. An improved P-wave velocity reference
model for Costa Rica, Geof´ıs. Int., 40, 3–19.
Quintero, R. & Kulhanek, O., 1998. Pn-wave observations in Costa Rica,
Geof´ıs. Int., 37.
Ranero, C.R. & von Huene, R., 2000. Subduction erosion along the Middle
America convergent margin, Nature, 404, 748–752.
Ranero, C.R., Phipps Morgan, J., McIntosh, K. & Reichert, C., 2003.
Bending-related faulting and mantle serpentinization at the Middle Amer-
ica trench, Nature, 425, 367–373.
Reyners, M., Eberhart-Phillips, D. & Stuart, G., 1999. A three-dimensional
image of shallow subduction: crustal structure of the Raukumara Penin-
sula, New Zealand, Geophys. J. Int., 137, 873–890.
Saffer, D.M. & Marone, C., 2003. Comparison of smectite- and illite-rich
gouge frictional properties: application to the updip limit of the seismo-
genic zone along subduction megathrusts, Earth planet. Sci. Lett., 215,
219–235.
Sallar`es, V., Da˜nobeitia, J.J., Flueh, E.R. & Leandro, G., 1999. Seismic ve-
locity structure across the middle American landbridge in northern Costa
Rica, J. Geodyn., 27, 327–344.
Sallar`es, V., Da˜nobeitia, J.J. & Flueh, E.R., 2000. Seismic tomography with
local earthquakes in Costa Rica, Tectonophysics, 296, 61–78.
Sallar`es, V., Da˜nobeitia, J.J. & Flueh, E.R., 2001. Lithospheric structure of
the Costa Rican Isthmus: effects of subduction zone magmatism on an
oceanic plateau, J. geophys. Res., 106, 621–643.
Sauter, A., Hallinan, J., Currier, R., Barash, T., Wooding, B., Schultz, A. &
Dorman, L.M., 1990. A new ocean bottom seismometer, Conference Pro-
ceedings on Marine Instrumentation, MPL-U-125/90,Marine Technology
Society, Columbia, MD, 99–104.
Schwartz, S.Y. & DeShon, H.R., 2006. Distinct Geodetic and Seismic Up-dip
Limits to the Northern Costa Rica Seismogenic Zone: evidence for Two
Mechanical Transitions, in The Seismogenic Zone of Subduction Thrust
Faults, eds Dixon, T.H. & Moore, J.C., Columbia University Press, New
York, in press.
Shipley, T.H., McIntosh, K.D., Silver, E.A. & Stoffa, P.L., 1992. Three-
dimensional seismic imaging of the Costa Rica accretionary prism: struc-
tural diversity in a small volume of the lower slope, J. geophys. Res., 97,
4439–4459.
Smith, W.H.F, & Wessel, P., 1990. Gridding with continuous curvature
splines in tension, Geophysics, 55, 293–305.
Spinelli, G.A. & Saffer, D., 2004. Along-strike variations in under-
thrust sediment dewatering on the Nicoya margin, Costa Rica re-
lated to the updip limit of seismicity, Geophys. Res. Lett., 31, 4613,
doi:10.1029/2003GL018863.
Spinelli, G.A. & Underwood, M.B., 2004. Characterization of sediment
entering the Costa Rica subduction zone: implications for fluid parti-
tioning along the Nicoya Peninsula plate interface, Island Arc, 3, 432–
451.
Stein, C.A. & Stein, S., 1992. A model for the global variation in oceanic
depth and heat flow with lithospheric age, Nature, 359, 123–129.
Thurber, C.H., 1981. Earth structure and earthquake locations in the Coyote
Lake area, central California, PhD thesis, p. 339, Massachusetts Institute
of Technology, Cambridge.
Thurber, C.H., 1983. Earthquake locations and three-dimensional crustal
structure in the Coyote Lake area, central California, J. geophys. Res., 88,
8226–8236.
Thurber, C.H., 1992. Hypocenter-velocity structure coupling in local earth-
quake tomography, Phys. Earth planet. Inter., 7, 55–62.
Thurber, C.H., 1993. Local earthquake tomography: velocities and V
P
/V
S
theory, in Seismic Tomography: Theory and Practice, pp. 563–583, eds
Iyer, H.M. & Hirahara, K., Chapman and Hall, London, United Kingdom.
Thurber, C. & Eberhart-Phillips, D., 1999. Local earthquake tomogra-
phy with flexible gridding, Computers and Geosciences, 25, 809–
818.
Toomey, D.R. & Foulger, G.R., 1989. Tomographic inversion of local earth-
quake data from the Hengill-Grensdalur central volcano complex, Iceland,
J. geophys. Res., 94, 17 497–17 510.
von Huene, R., Ranero, C.R., Weinrebe, W. & Hinz, K., 2000. Quaternary
convergent margin tectonics of Costa Rica: segmentation of the Cocos
Plate and Central American volcanism, Tectonics, 19, 314–334.
von Huene, R., Ranero, C.R. & Vannucchi, P., 2004. A generic model
of subduction erosion, Geology, 32, doi:10.1130/G20563.1, 913–
916.
Vrolijk, P., 1990. On the mechanical role of smectite in subduction zones,
Geology, 18, 703–707.
Wagner, L.S., Beck, S. & Zandt, G., 2005. Upper mantle structure in the
south central Chilean subduction zone (30
to 36
S), J. geophys. Res.,
110, B01308, doi:10.1029/2004JB003238.
Walther, C.& Flueh,E., 2002. Remnantof the ancientFarallonPlate breakup:
alow-velocity body in the lower oceanic crust of Nicoya Peninsula,
Costa Rica—evidence from-angle seismics, Geophys. Res. Lett., 29, 1939,
doi:10.1029/2002GL015026.
Wessel, P. & Smith, W.H.F., 1998. New, improved version of the Generic
Mapping Tools Released, EOS Trans. Am. geophys. Un., 79, 579.
Ye , S., Bialas, J., Flueh, E.R., Stavenhagen, A. & von Huene, R., 1996.
Crustal structure of the Middle American Trench off Costa Rica from
wide-angle seismic data, Tectonics, 15, 1006–1021.
C
2005 The Authors, GJI, 164, 109–124
Journal compilation
C
2005 RAS
... Figura 8. Sismicidad interplacas interciclo (franja rosada) y ruptura de grandes terremotos en el segmento Pacífico Sureste y comparación con los segmentos Noroeste y Central, en el margen convergente de Costa Rica. La sismicidad interciclo fue determinada con base en registros de la RSN y otros estudios con redes temporales de tierra y fondo marino (Husen et al., 2002;DeShon et al., 2003;DeShon et al., 2006;Arroyo et al., 2014a). Se muestra también, para los sismos principales de algunas de las secuencias ocurridas en los últimos 40 años en la zona interplacas, el área de ruptura según: deslizamiento cosísmico para Sámara 2012 (Liu et al., 2015), imagen tomográfica de un monte submarino subducido para Cóbano 1990 (Husen et al., 2002), e inversión de ondas de cuerpo para Quepos 1999 y Golfito 1983 (Tajima y Kikuchi, 1995). ...
... La zona sismogénica interplacas en el margen Pacífico costarricense (Figura 8) ha sido estudiada especialmente en los segmentos Noroeste y Central, a través de tomografías con sismos locales (DeShon et al., 2006;Arroyo et al., 2009), el análisis y la relocalización de secuencias sísmicas importantes (e.g. Husen et al., 2002;DeShon et al., 2003;Chaves et al., 2017), modelado geodésico (Feng et al., 2012;Kobayashi et al., 2014;Protti et al., 2014) y otros fenómenos sísmicos (Outerbridge et al., 2010). ...
... Se ha señalado que la porción acoplada del límite de placas en las zonas de subducción ocurre generalmente en un área restringida, con límites superior e inferior (Byrne et al., 1988;Vrolijk, 1990;Oleskevich et al., 1999;Moore y Saffer, 2001;Moore et al., 2007;Ranero et al., 2008). En el margen costarricense, la sismicidad que sucede en la región interplacas en el periodo entre grandes terremotos ha sido captada y relocalizada con alta calidad gracias a la instalación de redes sísmicas temporales que han incluido estaciones de fondo marino en los segmentos Noroeste y Central (DeShon et al., 2006;Arroyo et al., 2014a). Además, esta sismicidad interplacas interciclo (en adelante denominada "sismicidad interciclo") también ha podido ser registrada rutinariamente con mejor calidad gracias al aumento de cobertura con estaciones de la RSN en todo el país. ...
Article
Full-text available
Between August and November 2018, a seismic sequence took place in the Dulce Gulf, in the vicinity of Golfito, a city in Southeastern Costa Rica. The main shock had a moment magnitude (Mw) of 6.1 and was widely felt in Costa Rica and Western Panama, with maximum Modified Mercalli intensities of VI. In this region, the oceanic Cocos Ridge, riding on top of the Cocos Plate, subducts beneath the Panama Microplate. Using the seismic records from the National Seismological Network of Costa Rica, in this work the seismicity is relocated using the double-difference technique, and an analysis of its temporal and geographic distribution together with the focal mechanism and intensities of the strongest events are presented. The results show that the sequence occurred at the interplate seismogenic zone, within the rupture area of the 1983 Golfito earthquake (7.4 Mw), between 12 and 27 km depth, in a cluster dipping 35º northeast underneath the Dulce Gulf. Based mainly on these results and on previous seismic sequences, it is here proposed that the seismogenic zone in Southeastern Costa Rica has an extension of ~160 x 45 km. Further, during the Golfito sequence the rupture of an inverse fault (5.9 Mw) also took place within the Cocos Plate beneath the Dulce Gulf, as well as seismicity along right-lateral strike-slip faults (4.6-5.6 Mw) in the Panama Microplate, 50 km away of the Dulce Gulf. The analysis of the interseismic interplate seismicity contributes to a better understating of the dynamics of the seismogenic zone. This is of particular relevance in Southeastern Costa Rica, where at least six damaging earthquakes of Mw > 7 have occurred since 1803, implying the imminent risk of an upcoming big earthquake in this area.
... In the Nicoya Peninsula, the seismicity that occurs in the interplate region during the seismic cycle, that is, in the period between major earthquakes, has been recorded and relocated with high quality (e.g. DeShon et al., 2006). There, the up-dip and down-dip limits of the interseismic seismicity range in a narrow area beneath the peninsula, between depths from ~15 to 25 km (DeShon et al., 2006). ...
... DeShon et al., 2006). There, the up-dip and down-dip limits of the interseismic seismicity range in a narrow area beneath the peninsula, between depths from ~15 to 25 km (DeShon et al., 2006). During the latest largest event in Nicoya in 2012 (Mw 7.6), the rupture propagated beyond this known area of interseismic seismicity, both in the up-dip and down-dip directions (e.g. ...
Article
Full-text available
Costa Rica is a seismically active region located in a subduction zone. Combining earthquake data from the National Seismological Network and several Central American catalogs from 1522 to 2020, we present a new estimation for the Gutenberg-Richter relationship. The seismic catalog compiled with ~122,000 earthquakes was studied by calculating the magnitude of completeness (Mc) and by applying a space-time window declustering method. The residual catalog and a previously proposed seismic zonation were used to determine the b-value, the maximum magnitude (Mmax), and the mean recurrence interval (MRI). Our results show a temporal and geographic variation in the Mc, decreasing from 7.0 since 1793 to 3.0 since 1995, and with the lowest estimate (2.5) in Central Costa Rica. The b-values of 0.85 for the entire catalog and ~ 0.83 for the interplate zones, are similar to other regions worldwide with young subducting slabs. There is a general trend of higher and more variable b-values among the upper-plate (average 0.90) and intraslab (1.14) zones as compared to the interplate regions (0.85). The upper-plate variability is explained in terms of the diversity in geological units and faulting style, whereas in the interplate and the subducting slab is connected to the stress level imposed by different seafloor morphologies and hydration states along the margin. Our data suggest a seismic potential of moment magnitude (Mw) 7.9–8.0 for the Nicoya and Southern Nicaragua interplate zones and for the Limon region. The Gutenberg-Richter distribution shows that a Mw 7.0 has a longer MRI for the intraslab zones (~72 years) than for the subduction interplate region (∼15 years) and plate and microplate boundaries (∼40–45 years). Interplate earthquakes have occurred recently (1991 Mw 7.7 and 2012 Mw 7.6) but no major intraslab since 1948 (Mw 7.0).
... de los límites de placas, de las fallas corticales (e.g. Denyer, Montero y Alvarado, 2003), de la zona sismogénica interplacas del margen Pacífico (e.g.DeShon et al., 2006;Arroyo, Husen y Flueh, 2014;) y del techo de la placa Coco bajo Costa Rica (e.g., ...
Article
Full-text available
RESUMEN: La Red Sismológica Nacional (RSN) localizó 3745 sismos locales durante el año 2021, lo cual es equivalente a la energía sísmica liberada por un evento de magnitud (Mw) 6,9. La sismicidad se localizó principalmente en siete zonas con 50-350 temblores. Hubo 209 sismos (~6% del total) sentidos por la población, incluyendo 31 con Mw entre 5,0 y 6,7. La mayoría de estos sismos fueron superficiales (~88% < 30 km), de magnitud baja (~52% Mw < 3,9) y fueron originados en fallas de las placas Caribe y Panamá (~69%). El sismo mayor del año, de Mw 6,7, se originó en la Zona de Fractura de Panamá, sin generar daños ni víctimas. Todos los sismos de Mw ≥ 5,5 tuvieron sus epicentros mar adentro y alejados de centros de población por lo que no se reportaron intensidades altas. Palabras clave: estadística sísmica; energía sísmica; intensidad sísmica; subducción; Zona de Fractura de Panamá. ABSTRACT: The National Seismological Network (RSN) located 3745 local earthquakes during 2021, which is equivalent to the seismic energy released by an event of magnitude (Mw) 6.9. The seismicity was mainly located in seven zones with 50-350 earthquakes. There were 209 felt events (~6% of the totality), including 31 with Mw between 5.0 and 6.7. Most of these earthquakes were shallow (~88% < 30 km), had a low magnitude (~52% Mw < 3.9), and were originated in faults within the Caribbean and Panama plates (~69%). The largest earthquake of the year of Mw 6.7 originated in the Panama Fracture Zone, without causing damage or casualties. Because all the earthquakes with Mw ≥ 5.5 were centered offshore and far from the population high intensities were not reported.
... Syracuse et al. (2008) iluminaron tomográficamente Nicaragua y la mitad noroeste de Costa Rica usando redes temporales. Otras tomografías locales de detalle se han obtenido para áreas menores, usando sismos registrados por redes combinadas de estaciones a lo largo de la costa Pacífica y de fondo marino, tales como DeShon et al. (2006) para la península de Nicoya, y Arroyo et al. (2009) y Dinc et al. (2010 para el Pacífico Central. ...
Article
Full-text available
RESUMEN: El Caribe Sur de Costa Rica fue en 1991 el escenario del último gran terremoto (Mw 7,7) entre la placa Caribe y la microplaca de Panamá. Pese a la alta sismicidad y avances en la cobertura de las redes sísmicas, esa zona carece de imágenes tomográficas detalladas. Usando los paquetes VELEST y SIMULPS, presentamos los modelos de velocidad unidimensional y tridimensional resultantes de la inversión de 1208 sismos registrados entre 1998 y 2020 por la Red Sismológica Nacional (RSN). Esta nueva tomografía muestra una franja de bajas velocidades que se inclina desde el Caribe hacia el suroeste hasta los 50 km de profundidad debajo de Talamanca, incluyendo el hipocentro del terremoto de Limón y sismicidad hasta los 30 km. Se interpreta esta configuración como la subducción de la placa Caribe debajo de la microplaca de Panamá y el terremoto de Limón como un evento de zona sismogénica interplacas. La geometría determinada proporciona un nuevo marco para interpretar la compleja tectónica del sureste de Costa Rica. ABSTRACT: In 1991, the Southern Caribbean of Costa Rica was the setting of the latest large earthquake (Mw 7.7) between the Caribbean plate and the Panama microplate. Despite the high seismicity and advances in seismic network coverage, this area lacks detailed tomographic images. Using the VELEST and SIMULPS packages, we derive the one-dimensional and three-dimensional velocity models resulting from the inversion of 1208 earthquakes recorded between 1998 and 2020 by the National Seismological Network (RSN). This new tomography shows a band of low velocities that inclines towards the southwest down to depths of 50 km below Talamanca, including the hypocenter of the Limon earthquake and seismicity down to 30 km. This configuration is interpreted as the subduction of the Caribbean plate under the Panama microplate and the Limon earthquake as an interplate seismogenic zone event. The determined geometry provides a new framework for interpreting the complex tectonics of southeastern Costa Rica.
... ,Kapinos et al. (2016) andCordell et al. (2019). On the other hand,Worzewski et al. (2011) with a study of MT in Central America where the Nazca, Cocos and Caribbean plates interact observe this important decrease in resistivity in the same zone whereDeShon et al. (2006) observe a decrease in the S wave velocities, both studies propose that this is a zone of serpentinization of the mantle. ...
Article
Full-text available
To analyze the process of subduction of the Nazca and South American plates in the area of the Southern Andes, and its relationship with the tectonic and volcanic regime of the place, magnetotelluric measurements were made through a transversal profile of the Chilean continental margin. The data-processing stage included the analysis of dimensional parameters, which as first results showed a three-dimensional environment for periods less than 1 s and two-dimensional for periods greater than 10 s. In addition, through the geomagnetic transfer function (tipper), the presence of structural electrical anisotropy was identified in the data. After the dimensional analysis, a deep electrical resistivity image was obtained by inverting a 2D and a 3D model. Surface conductive anomalies were obtained beneath the central depression related to the early dehydration of the slab and the serpentinization process of the mantle that coincides in location with a discontinuity in the electrical resistivity of a regional body that we identified as the Nazca plate. A shallow conductive body was located around the Calbuco volcano and was correlated with a magmatic chamber or reservoir which in turn appears to be connected to the Liquiñe Ofqui fault system and the Andean Transverse Fault system. In addition to the serpentinization process, when the oceanic crust reaches a depth of 80–100 km, the ascending fluids produced by the dehydration and phase changes of the minerals present in the oceanic plate produce basaltic melts in the wedge of the subcontinental mantle that give rise to an eclogitization process and this explains a large conductivity anomaly present beneath the main mountain range.
... ,Kapinos et al. (2016) andCordell et al. (2019). On the other hand,Worzewski et al. (2011) with a study of MT in Central America where the Nazca, Cocos and Caribbean plates interact observe this important decrease in resistivity in the same zone whereDeShon et al. (2006) observe a decrease in the S wave velocities, both studies propose that this is a zone of serpentinization of the mantle. ...
Preprint
Full-text available
In order to analyze the process of subduction of the Nazca and South American plates in the area of the Southern Andes, and its relationship with the tectonic and volcanic regime of the place, magnetotelluric measurements were made through a transversal profile of the Chilean continental margin. The data processing stage included the analysis of dimensional parameters, which as first results showed a three-dimensional environment for periods less than 1s and two-dimensional for periods greater than 10s. In addition, through the geomagnetic transfer function (tipper), the presence of structural electrical anisotropy was identified in the data. After the dimensional analysis, a deep electrical resistivity image was obtained by inverting a 2D and a 3D model. Surface conductive anomalies were obtained beneath the central depression related to the early dehydration of the slab and the serpentinization process of the mantle that coincides in location with a discontinuity in the electrical resistivity of a regional body that we identified as the Nazca plate. A shallow conductive body was located around the Calbuco volcano and was correlated with a magmatic chamber or reservoir which in turn appears to be connected to the Liquiñe Ofqui fault system and the Andean Transverse Fault system. In addition to the serpentinization process, when the oceanic crust reaches a depth of 80 - 100km, the ascending fluids produced by the dehydration and phase changes of the minerals present in the oceanic plate produce basaltic melts in the wedge of the subcontinental mantle that give rise to an eclogitization process and this explains a large conductivity anomaly present beneath the main mountain range.
... The seismological constraints on the crustal thickness throughout Costa Rica range ∼27-42 km in northern Costa Rica (Deshon et al., 2006;Linkimer et al., 2010;MacKenzie et al., 2008;Protti et al., 1996). In central to southern Costa Rica the depth to the upper plate Moho has been imaged by receiver function analysis (Dzierma et al., 2011) to be 30-40 km, whereas tomographic models place it at ∼30 km (Dinc et al., 2010), and active source imaging shows it to be 40 km beneath the volcanic front (Hayes et al., 2013). ...
Article
Full-text available
Plain Language Summary The presence of low‐velocity zones in the mantle wedge has long been attributed to the hydration of mantle peridotite, which in turn becomes serpentine minerals. However, the direct links between low‐velocity anomalies, the degree of mantle serpentinization, and water circulation are not firmly established due to the lack of direct experimental data from partially serpentinized peridotites. Because these variables have large uncertainties, it is often difficult to determine the water content of the mantle wedge using seismic observations. In this study, we measured the acoustic velocity through synthetic olivine–antigorite aggregates, which are the proxies for mantle wedge lithologies. We explore two important equations between the acoustic velocity and the degree of serpentinization as well as pressure. Combined with geophysical observations, we could map the degree of serpentinization and water contents of the subducting slabs around the Pacific Ocean. Our results provide a clue to study the water transport between the surface and the Earth's interior.
Article
Full-text available
Central America is one of the most active seismic zones in the World, due to the interaction of five tectonic plates (North America, Caribbean, Coco, Nazca and South America), and its internal deformation, which generates almost one destructive earthquakes (5.4. ≤. Mw. ≤. 8.1) every year. A new seismological zonation for Central America is proposed based on seismotectonic framework, a geological context (tectonic and geological maps), geophysical and geodetic evidence (gravimetric maps, magnetometric, GPS observations), and previous works. As a main source of data a depurated earthquake catalog was collected covering the period from 1522 to 2015. This catalog was homogenized to a moment magnitude scale (Mw). After a careful analysis of all the integrated geological and seismological information, the seismogenic zones were established into seismic areas defined by similar patterns of faulting, seismicity, and rupture mechanism. The tectonic environment has required considering seismic zones in two particular seismological regimes: a) crustal faulting (including local faults, major fracture zones of plate boundary limits, and thrust fault of deformed belts) and b) subduction, taking into account the change in the subduction angle along the trench, and the type and location of the rupture. The seismicity in the subduction zone is divided into interplate and intraplate inslab seismicity. The regional seismic zonation proposed for the whole of Central America, include local seismic zonations, avoiding discontinuities at the national boundaries, because of a consensus between the 7 countries, based on the cooperative work of specialists on Central American seismotectonics and related topics.
Preprint
Full-text available
In order to analyze the process of subduction of the Nazca and South American plates in the area of the Southern Andes, and its relationship with the tectonic and volcanic regime of the place, magnetotelluric measurements were made through a transversal profile of the Chilean continental margin. The data processing stage included the analysis of dimensional parameters, which as first results showed a three-dimensional environment for periods less than 1s and two-dimensional for periods greater than 10s. In addition, through the geomagnetic transfer function (tipper), the presence of structural electrical anisotropy was identified in the data. After the dimensional analysis, a deep electrical resistivity image was obtained by inverting a 2D and a 3D model. Surface conductive anomalies were obtained beneath the central depression related to the early dehydration of the slab and the serpentinization process of the mantle that coincides in location with a discontinuity in the electrical resistivity of a regional body that we identified as the Nazca plate. A shallow conductive body was located around the Calbuco volcano and was correlated with a magmatic chamber or reservoir which in turn appears to be connected to the Liquiñe Ofqui fault system and the Andean Transverse Fault system. In addition to the serpentinization process, when the oceanic crust reaches a depth of 80 - 100 km, the ascending fluids produced by the dehydration and phase changes of the minerals present in the oceanic plate produce basaltic melts in the wedge of the subcontinental mantle that give rise to an eclogitization process and this explains a large conductivity anomaly present beneath the main mountain range.
Article
Full-text available
We present a new approach to determine precise and reliable hypocenter locations in the tectonically complex region of Switzerland. A three-dimensional (3-D) P wave velocity model to be used for earthquake relocation is obtained by simultaneously inverting arrival times of local earthquakes for hypocenter locations and 3-D P wave velocity structure. A 3-D P wave velocity model derived from controlled source seismology (CSS) is used as an initial reference model. The final 3-D model thus combines all available information from both CSS and local earthquake data. The probabilistic, nonlinear formulation of the earthquake location problem includes a complete description of location uncertainties and can be used with any kind of velocity model. In particular, the combination of nonlinear, global search algorithms, such as the Oct-Tree Importance Sampling, with probabilistic earthquake location provides a fast and reliable tool for earthquake location. The comparison of hypocenter locations obtained routinely by the Swiss Seismological Service (SED) to those relocated in the new 3-D velocity model using a probabilistic approach reveals no systematic shifts but does exhibit large individual shifts in some epicenter locations and focal depths. We can attribute these large shifts in part to large uncertainties in the hypocenter location. Events with a low number of observations (<8) and no observation within the critical focal depth distance typically show large location uncertainties. Improved hypocentral locations, particularly for mine blasts and earthquakes whose routine hypocenter locations had been questionable, confirm that improved velocity model and probabilistic earthquake location yield more precise and reliable hypocenter locations and associated location uncertainties for Switzerland.
Article
Full-text available
A geotomographic inversion method was used to derive the three-dimensional velocity distribution of Long Valley caldera, California, from local earthquake data. 2450 local events with more than 8 recordings (observations from the 30 permanent and/or 83 temporary stations throughout the area under study) have been selected from the approximately 10,000 events recorded and located between May 1980 and June 1984 by the U.S. Geological Survey. The 31,730 observations were used to determine the velocity perturbations of 3328 blocks thus leading to an overdetermination factor of about 6.5. The main results of the geotomographic inversion suggest the presence of a zone of low p-velocity (decreased 3%to 5% (0.20 to 0.30 km/s) beneath the southern part of the resurgent dome and the south moat of the caldera between 3km and 7km depth. This anomaly seens to overlie a weaker and broader zone of low p-velocity (decrease of approximately 1% ) in the depth range of 5km to 14 km. The latter zone strikes south from beneath the Mono Craters through the Long Valley caldera and runs southeast beneath the south moat and into the Sierran block. While the seismic activity inside the caldera coincides with the area of low p-velocity the majority of the events outride the caldera lies in zones of higher p-velocity. Kissling, E., W. L. Ellsworth and R. S. Cockerham, Three-dimensional structure of the Long Valley Caldera, California region by geotomography, Proceedings of a Conference on Active Tectonic Magmatic Prozesses beneath Long Valley Caldera, Eastern California, U.S. Geol. Surv. Open File Rep. 84-939. 1984.
Article
Full-text available
High-resolution seismicity data from Costa Rica and southern Nicaragua have been used to image the geometry of the Wadati-Benioff zone under southern Central America. The data include 9514 events with computed horizontal and vertical errors smaller than 4 and 5 km respectively. We found that under the Nicaragua-Costa Rica border the Wadati-Benioff zone contorts (from steep to shallow dip angles, northwest to southeast) but does not show evidence of a brittle test, as postulated by others. Further to the southeast-northeast from Puerto Caldera-the Wadati-Benioff zone does, however, show a tear (the Quesada Sharp Contortion) at intermediate depths (h >70 km). Overall the dip angle of the Wadati-Benioff zone decreases from 84° under Nicaragua to 60° under central Costa Rica. This geometry of the Wadati-Benioff zone and other tectonic features related to the subduction of the Cocos plate beneath the Caribbean plate have been integrated into a model that correlates them with along-trench variations in age of the subducted Cocos plate. -from Authors
Article
Full-text available
Differing rates and styles of Quaternary deformation along the Costa Rican fore arc reflect segmentation of the trench corresponding with three contrasting domains of subducting seafloor offshore. Rapid upward flexure of the southern fore-arc segment results from the subduction of the buoyant Cocos Ridge, whereas moderate deformation along the central forearc segment reflects the subduction of buoyant seamounts. In contrast, the northern Costa Rican forearc deforms in response to the subduction of relatively dense sea floor devoid of major bathymetric anomalies. Quaternary geomorphic evidence and earthquake oral histories from the Península de Nicoya, within the northern Costa Rican fore arc, demonstrate arcward tilting of the fore-arc crust, with discrete uplift events occurring during large subduction earthquakes.
Article
Full-text available
Version 3.1 of the Generic Mapping Tools (GMT) has been released. More than 6000 scientists worldwide are currently using this free, public domain collection of UNIX tools that contains programs serving a variety of research functions. GMT allows users to manipulate (x,y) and (x,y,z) data, and generate PostScript illustrations, including simple x-y diagrams, contour maps, color images, and artificially illuminated, perspective, and/or shaded-relief plots using a variety of map projections (see Wessel and Smith [1991] and Wessel and Smith [1995], for details.). GMT has been installed under UNIX on most types of workstations and both IBM-compatible and Macintosh personal computers.
Article
Full-text available
New seismic and geodetic data from Costa Rica provide insight into seismogenic zone processes in Central America, where the Cocos and Caribbean plates converge. Seismic data are from combined land and ocean bottom deployments in the Nicoya peninsula in northern Costa Rica and near the Osa peninsula in southern Costa Rica. In Nicoya, inversion of GPS data suggests two locked patches centered at 14 ± 2 and 39 ± 6 km depth. Interplate microseismicity is concentrated in the more freely slipping intermediate zone, suggesting that small interseismic earthquakes may not accurately outline the updip limit of the seismogenic zone, the rupture zone for future large earthquakes, at least over the short (˜1 year) observation period. We also estimate northwest motion of a coastal "sliver block" at 8 ± 3 mm/yr, probably related to oblique convergence. In the Osa region to the south, convergence is orthogonal to the trench. Cocos-Caribbean relative motion is partitioned here, with ˜8 cm/yr on the Cocos-Panama block boundary (including a component of permanent shortening across the Fila Costeña fold and thrust belt) and ˜1 cm/yr on the Panama block-Caribbean boundary. The GPS data suggest that the Cocos plate-Panama block boundary is completely locked from ˜10-50 km depth. This large locked zone, as well as associated forearc and back-arc deformation, may be related to subduction of the shallow Cocos Ridge and/or younger lithosphere compared to Nicoya, with consequent higher coupling and compressive stress in the direction of plate convergence.
Article
A worldwide catalog of shallow (depth <70 km) and large (Ms ≥ 7) earthquakes recorded between 1900 and 1989 has been compiled. The catalog is shown to be complete and uniform at the 20-sec surface-wave magnitude Ms ≥ 7.0. The catalog is accompanied by a reference list for all the events with seismic moment determined at periods longer than 20 sec. Using these seismic moments for great and giant earthquakes and a moment-magnitude relationship for smaller events, a seismic moment catalog for large earthquakes from 1900 to 1989 is produced. The seismic moment released at subduction zones during this century constitutes 90% of all the moment released by large, shallow earthquakes on a global basis. The seismic moment released in the largest event that occurred during this century, the 1960 southern Chile earthquake, represents about 30 to 45% of the total moment released from 1900 through 1989. -from Authors
Article
Seismic data from explosions and local earthquakes are used to determine shallow crustal structure near Managua, Nicaragua, and to derive a complete crustal model for northern Costa Rica. A new method, referred to as the minimum apparent velocity (MAV) method, has been developed and applied in the analysis of local earthquake data. The results indicate that the MAV method of multilayer analysis, when combined with well-determined station corrections, can be quite useful in regions of high seismicity. Based upon the analysis presented here, the total thickness of the crust beneath the central volcanic province of northern Costa Rica is about 43 km. Four crustal layers are identified with compressional-wave velocities of 2.6, 5.1, 6.2, and 6.6 km/sec. The upper mantle velocity is 7.9 km/sec. The major features of this model closely resemble those of other seismically active, marginal zones of the Pacific.