Figure 1 - uploaded by Gerardo Leon Soto
Content may be subject to copyright.
Map showing the location of the Isthmus of Tehuantepec and various tectonic elements. Inset at bottom left shows stations of the Veracruz- Oaxaca (VEOX) array.
Source publication
Shear-wave splitting measurements were made using S waves from local
earthquakes recorded by stations of the Veracruz-Oaxaca array, which was
deployed across the Isthmus of Tehuantepec. In this segment of the
Middle America Trench, the oceanic Cocos Plate subducts under the
continental North American Plate. Intraplate earthquakes within the
Cocos s...
Contexts in source publication
Context 1
... the Gulf of California ( Obrebski et al. 2006;van Benthem et al. 2008;Long 2010), and on subduction of the Rivera and Cocos plates at the Middle America Trench (MAT) (van Benthem 2005;Stubailo & Davis 2007, 2012aSoto et al. 2009;Ponce-Cortés 2012;Rojo-Garibaldi 2012;van Benthem et al. 2013). These results are summarized for central Mexico in Fig. S1. The only work involving local S-wave splitting was conducted by Soto et al. (2009) for the Rivera and westernmost Cocos Plate, but given the limited depth extent of the slab seismicity they could only resolve the continental crust. In the present study, we report shear-wave splitting measurements from slab events in the Isth- mus of ...
Context 2
... Isthmus of Tehuantepec is the narrowest continental re- gion in southern Mexico and joins the Gulf of Mexico with the Pacific Ocean (Fig. 1). The Cocos Plate subducts beneath the North American continent and shows a change in geometry across the Isthmus of Tehuantepec. Farther west, under Guerrero state, the Cocos Plate subducts subhorizontally (Pardo & Suárez 1995;PérezCampos et al. 2008;Husker & Davis 2009;Kim et al. 2010). By the time it reaches the Isthmus of ...
Context 3
... dips at an angle of ∼25 • (Pardo & Suárez 1995; Rodríguez-Pérez 2007;Kim et al. 2011;Melgar & Pérez-Campos 2011). Farther east, under Chiapas state, the Cocos slab dips at an angle of 45 • ( Bravo et al. 2004;Rodríguez-Pérez 2007). Located offshore, the Tehuantepec Ridge (TR) subducts beneath the North American Plate at the Isthmus of Tehuantepec (Fig. 1). The age of the oceanic lithosphere presents a sudden change across the TR estimated at 7 Myr ( Manea et al. 2005) where the ridge intersects the MAT. The convergence velocity between the Cocos and North American plates at the intersection of the TR with the MAT is 7.2 cm yr −1 (DeMets et al. 2010), whereas the trench is retreating at ...
Context 4
... subducting plates reaching a depth of ∼110 km and the location of volcanic arcs has long been recognized (e.g. Tatsumi & Eggins 1995;Syracuse & Abers 2006). The main, modern volcanic feature in central and southern Mexico is the Trans- Mexican Volcanic Belt (TMVB), but it is located northwest of the area encompassed by the present study (e.g. fig. 1 in Manea & Manea 2006). It is related to the subhorizontal subduction of the Cocos Plate taking place to the west. Volcanic activity is found again in the Los Tuxtlas Volcanic Field (LTVF) near the coast of the Gulf of Mexico. The LTVF is located southeast from the TMVB and around the northernmost end of the array used in the present ...
Context 5
... data set used in this study was recorded by the Veracruz-Oaxaca (VEOX) experiment which was a dense linear array of 46 broad- band seismometers deployed along a N-S profile across the Isthmus of Tehuantepec (Melgar & Pérez-Campos 2011); see Fig. 1. The array operated from 2007 June to 2009 ...
Context 6
... studies have been conducted to determine the anisotropy in other segments of the MAT using local S waves from hypocentres within the slab. In Mexico, in a region located west of the present study (Fig. S1), the MApping the Rivera Subduction zone (MARS) array was deployed over the subducted Rivera and westernmost Cocos plates (Soto et al. 2009). Only events in the depth range from 60 to 106 km were used because seismicity does not extend any deeper. Soto et al. (2009) obtained a mean delay time of ∼0.2 s and inferred that the anisotropy ...
Similar publications
We used high-resolution seismic tomography to probe into the arc and back-arc magmatism and volcanism of the Japan subduction zone. Prominent zones with low-velocity, high Poisson's ratio, high-attenuation and strong seismic anisotropy are revealed in the crust and uppermost mantle beneath the active volcanoes, and they exist in the central portion...
The Sunda Plate is a minor tectonic plate bounded by tectonically active convergent boundaries, below which are subducting: the Philippine Sea Plate to the east and the Indo‐Australian Plate to the south and west. It is thus an ideal tectonic setting for investigating the interaction between subduction and asthenospheric flow. To better understand...
Seismic anisotropy can provide valuable constraints for the study of subduction zone dynamics. This study presents a high‐resolution 3‐D azimuthally anisotropic shear wave velocity model down to 230 km beneath Alaska via ambient noise tomography and wave gradiometry method. The model reveals layered anisotropy patterns related to subduction tectoni...
Subduction zones are among the most dynamic tectonic environments on Earth. Deformation mechanisms of various scales produce networks of oriented structures and faulting systems that result in a highly anisotropic medium for seismic wave propagation. In this study, we combine shear wave splitting inferred from receiver functions and the results fro...
Competing geodynamic scenarios proposed for northern Apennines (Italy) make very different predictions for the orientation of strain in the upper mantle. Constraints on the pattern are offered by observations of seismic anisotropy. Previous study of the anisotropy beneath the northern Apennines used birefringence of core-refracted shear waves (SKS...
Citations
... A thin crust along the western part of the north arm of Sulawesi with about 20-25 km thickness has been proposed based on the receiver function analysis (Fauzi et al., 2021;Linang, 2023). Previous investigations in subduction zone settings support our inference that the dominant contributor of seismic anisotropy in the upper mantle is mantle wedge flow (e.g., Cao et al., 2021;León Soto & Valenzuela, 2013;Long & Wirth, 2013;Wang & He, 2020), though the upper plate crust and subducting slab may also contribute. However, it is likely to insignificant in comparison, as discussed above, unless the upper plate is of cratonic origin (Yuan et al., 2011), which is not the case here. ...
... Variation from trench-normal to slab-strike-normal anisotropy is also observed at other subduction settings along the Pacific plate boundary (e.g., León Soto & Valenzuela, 2013;Long & Wirth, 2013;Wang & He, 2020). It demonstrates that the most dominant flow type within a subduction zone is generally the corner flow, which Figure 7. Schematic cartoon illustrating the tectonic setting of multiple subduction zones (based on a model shown in Hall & Spakman, 2015) and interpreted mantle flows (red arrows) and lithospheric-asthenospheric deformation (blue arrow) in the study region. ...
... Given that the global average splitting time of continental crust is on the order of 0.1 s (Silver, 1996), we suggest that mantle wedge flow is the main source of local shear-wave splitting at intermediate depth, with potentially a certain degree of anisotropy also coming from the subducting slab and overriding lithosphere. Overall, we argue for the critical role of mantle wedge flow in explaining local S-wave splitting, as has been well established by other subduction zone studies (e.g., Cao et al., 2021;León Soto & Valenzuela, 2013;Long & Wirth, 2013;Wang & He, 2020). ...
The North Sulawesi subduction zone is characterized by southward subduction of the Celebes Sea slab to a depth of ∼250 km, mainly overlying the Sangihe slab that subducts west from the Molucca Sea and penetrates the mantle transition zone. The palaeo‐subducted Sula slab dips northward and partially underlies both the Sangihe and Celebes Sea slabs. Adjacent subduction zones with horizontal overlapping subducting slabs in the upper mantle have unclear dynamic interactions. An extensive strike‐slip fault forms the western boundary of the active North Sulawesi subduction zone, providing an ideal setting to study mantle flow between overlapping slabs. We use local S‐wave and teleseismic S and SK(K)S waveform splitting analysis to measure seismic anisotropy in the northern Sulawesi region. Our observations reveal typical mantle wedge corner flow within the Sangihe subduction system. In the Gulf of Tomini, the observed trench‐oblique fast‐axis orientations above the Celebes Sea slab are likely a consequence of the interaction between two subducting slabs. The southernmost measurement with an E–W‐trending fast direction in the mantle wedge might be related to the subduction of the Sula slab. Furthermore, fault‐parallel fast‐axis orientations of anisotropy near the southern segment of the Palu‐Koro fault are attributed to large‐scale shearing across this lithospheric‐scale strike‐slip fault system. Overall, our observations suggest that the strain caused by lithospheric and asthenospheric deformation is mainly confined within the microplate, displaying a restricted flow pattern and localized effects due to the size of the plate boundaries, such as the Palu‐Koro fault.
... A thin crust along the western part of the north arm of Sulawesi with about 20-25 km thickness has been proposed based on the receiver function analysis (Fauzi et al., 2021;Linang, 2023). Previous investigations in subduction zone settings support our inference that the dominant contributor of seismic anisotropy in the upper mantle is mantle wedge flow (e.g., Cao et al., 2021;León Soto & Valenzuela, 2013;Long & Wirth, 2013;Wang & He, 2020), though the upper plate crust and subducting slab may also contribute. However, it is likely to insignificant in comparison, as discussed above, unless the upper plate is of cratonic origin (Yuan et al., 2011), which is not the case here. ...
... Variation from trench-normal to slab-strike-normal anisotropy is also observed at other subduction settings along the Pacific plate boundary (e.g., León Soto & Valenzuela, 2013;Long & Wirth, 2013;Wang & He, 2020). It demonstrates that the most dominant flow type within a subduction zone is generally the corner flow, which Figure 7. Schematic cartoon illustrating the tectonic setting of multiple subduction zones (based on a model shown in Hall & Spakman, 2015) and interpreted mantle flows (red arrows) and lithospheric-asthenospheric deformation (blue arrow) in the study region. ...
... Given that the global average splitting time of continental crust is on the order of 0.1 s (Silver, 1996), we suggest that mantle wedge flow is the main source of local shear-wave splitting at intermediate depth, with potentially a certain degree of anisotropy also coming from the subducting slab and overriding lithosphere. Overall, we argue for the critical role of mantle wedge flow in explaining local S-wave splitting, as has been well established by other subduction zone studies (e.g., Cao et al., 2021;León Soto & Valenzuela, 2013;Long & Wirth, 2013;Wang & He, 2020). ...
... In the overriding plate the main source of . Shear wave splitting measurements averaged over a grid from local intraslab earthquakes deeper than 50 km red, are shown as thick short lines as reported in León Soto & Valenzuela (2013). Source side splitting results (Lynner & Long 2014a) are plotted at their surface projection of the source and presented with thick, green bars labelled by 1, 2 and 3. Anisotropic parameters for measurement labelled as 1 are (φ = 64.9 ...
... In order to constrain the vertical distribution of the observed anisotropy, we compared our results with the anisotropic parameters measured in the mantle wedge from local intraslab events deeper than 50 km (León Soto & Valenzuela 2013). Even though the frequency content of local earthquakes and teleseismic phases is not the same (local earthquakes in León Soto & Valenzuela (2013) has dominant frequency ∼1.0 Hz), this comparison gives some insight into the anisotropy distribution at depth (Long & Silver 2008). ...
... In order to constrain the vertical distribution of the observed anisotropy, we compared our results with the anisotropic parameters measured in the mantle wedge from local intraslab events deeper than 50 km (León Soto & Valenzuela 2013). Even though the frequency content of local earthquakes and teleseismic phases is not the same (local earthquakes in León Soto & Valenzuela (2013) has dominant frequency ∼1.0 Hz), this comparison gives some insight into the anisotropy distribution at depth (Long & Silver 2008). Fig. 6 compares SKS splitting measurements (thin lines) from the present study with the local S splitting measurements (thick, red bars) from León Soto & Valenzuela (2013). ...
Shear wave splitting measurements in the Isthmus of Tehuantepec (IT), southern Mexico, inferred from teleseismic core phases are presented. Measurements were made along a south-to-north profile across the IT. The results show a predominantly trench-normal pattern of fast polarization orientations with averaged delay times up to 2.2 s. Fast orientations near the trench suggest a corner flow in the mantle wedge and an entrained flow in the subslab region. Away the trench, fast orientations are parallel to the Absolute plate Motion, suggesting that the anisotropy in that region is driven by a simple asthenospheric flow. A comparison with splitting measurements made in the Mexican subduction zone shows a 17º clockwise rotation of the fast orientations of between east and west Mexico. This is consistent with the observed change in orientation of 19º clockwise in the Middle America Trench (MAT). This suggests that the rotation of the fast orientations is controlled by the change of orientation in the MAT.
... a source of anisotropy (Holtzman & Kendall, 2010). Although there are exceptions (Hammond et al., 2010;Schlaphorst et al., 2017), most local splitting studies around the globe have suggested that the mantle wedge is the main anisotropic structure in the subduction system (Abt et al., 2009;León Soto & Valenzuela, 2013;Long & van der Hilst, 2006;Nakajima & Hasegawa, 2004). In the Alaska subduction zone specifically, anisotropy has been suggested to be present in the mantle wedge and in the subducting Pacific lithosphere and subslab asthenosphere (Christensen & Abers, 2010;Hanna & Long, 2012;Karlowska et al., 2021 (Argus et al., 2011). ...
Shear‐wave splitting observations can provide insight into mantle flow, due to the link between the deformation of mantle rocks and their direction‐dependent seismic wave velocities. We identify anisotropy in the Cook Inlet segment of the Alaska subduction zone by analyzing splitting parameters of S waves from local intraslab earthquakes between 50 and 200 km depths, recorded from 2015–2017 and emphasizing stations from the Southern Alaska Lithosphere and Mantle Observation Network (SALMON) experiment. We classify 678 high‐quality local shear‐wave splitting observations into four regions, from northwest to southeast: (L1b) splitting measurements parallel to Pacific plate motion, (L1a) arc‐perpendicular splitting pattern, (L2) sharp transition to arc‐parallel splitting, and (L3) splitting parallel to Pacific plate motion. Forward modeling of splitting from various mantle fabrics shows that no one simple model fully explains the observed splitting patterns. An A‐type olivine fabric with fast direction dipping 45° to the northwest (300°)—aligned with the dipping slab—predicts fast directions that fit L1a observations well, but not L2. The inability of the forward model fabrics to fit all the observed splitting patterns suggests that the anisotropy variations are not due to variable ray angles, but require distinct differences in the anisotropy regime below the arc, forearc, and subducting plate.
... Additionally, Lynner and Long (2014a) carried out source-side splitting measurements of subslab anisotropy using teleseismic S waves after accounting for anisotropy beneath the stations. León Soto and Valenzuela (2013) used S phases from local, intraslab earthquakes deeper than 50 km recorded by the VEOX experiment in order to measure anisotropy in the mantle wedge. Recent work has quantified anisotropy using SKS data from new SSN stations (Ponce-Cortés, 2012;van Benthem et al., 2013). ...
... The technique used to quantify splitting, however, is more general and can be applied to shear waves containing both SV and SH energy from local events (e. g. León Soto et al., 2009;León Soto and Valenzuela, 2013). ...
... It should also be noted that the orientation of the teleseismic fast axes shows a progressive clockwise rotation from the area of flat subduction to the west, to this area of steeper subduction, and finally to the Yucatán peninsula farther east (Figure 5a). The study by León Soto and Valenzuela (2013) also relied on VEOX data in the Isthmus of Tehuantepec. They made shear wave splitting measurements using S waves from deep, local intraslab earthquakes to constrain the characteristics of flow in the mantle wedge. ...
A review is presented of the shear wave splitting studies of the upper mantle carried out in Mexico during the last decade. When a seismic wave enters an anisotropic medium it splits, which means that a fast and a slow wave are produced. Two parameters are used to quantify anisotropy. These are the fast polarization direction and the delay time between the fast and the slow wave. An example of the measurement technique is presented using an SKS phase because most observations are based on teleseismic data. Results of two studies using local S waves from intraslab earthquakes are also discussed. Key aspects of the interpretation of splitting measurements are explained. These include the depth localization of anisotropy, the relationship between olivine fabrics and mantle flow, the role of absolute plate motion, and the role of relative plate motions with a special focus on subduction zones. An important motivation for studying seismic anisotropy is that it makes it possible to constrain the characteristics of upper mantle flow and its relationship to tectonic processes. Mexico has many diverse tectonic environments, some of which are currently active, or were formerly active, and have left their imprint on seismic anisotropy. This has resulted in a wide variety of mechanisms for driving mantle flow. Broadly speaking, the discussion is organized into the following regions: Baja California peninsula, Western Mexican Basin and Range, northern and northeastern Mexico, the Middle America Trench, the Yucatán peninsula, and lowermost mantle anisotropy. Depending on the unique characteristics encountered within each region, the relationship between anisotropy and mantle flow is explored.
... In the Mexican subduction zone, several studies now exist of upper mantle shear wave anisotropy using records of teleseismic SKS phases (van Benthem, 2005;Stubailo and Davis, 2007Bernal-Díaz et al., 2008;León Soto et al., 2009;Rojo-Garibaldi, 2011;Rojo-Garibaldi et al., 2011;Ponce-Cortés, 2012;van Benthem et al., 2013;Bernal-López et al., 2014Bernal-López, 2015;Stubailo, 2015), using S waves recorded locally (León Soto et al., 2009;León Soto and Valenzuela, 2013), and also source-side subslab anisotropy using teleseismic S phases (Lynner and Long, 2014a). In the area of flat-slab, all fast polarization directions φ are oriented roughly in the direction perpendicular to the MAT with delay times δt of about 1 s (Rojo-Garibaldi, 2011;Rojo-Garibaldi et al., 2011;Ponce-Cortés, 2012;van Benthem et al., 2013;Bernal-López et al., 2014Bernal-López, 2015;Stubailo, 2015). ...
... León Soto and R.W. Valenzuela, manuscript in preparation, 2016). Measurements using S waves from local, intraplate Cocos earthquakes established the existence of two distinct anisotropy patterns in the mantle wedge, separated by the 100 km isodepth contour (León Soto and Valenzuela, 2013). It should be noted that in this part of Mexico no active volcanic arc exists associated to subduction of the Cocos slab, probably due to a lack of dehydration fluids. ...
... On the other hand, for slab depths between 60 and 85 km, the data show some trench-parallel fast axes (average δt = 0.28 s), but the pattern is complex given that other measurements with their fast axes oriented in different directions are interspersed throughout the same region, where B-type olivine is expected. This complexity likely arises from the fact that for shallow earthquakes, the path through the mantle wedge is shorter, and thus the anisotropy contributions from the continental crust and also from the slab itself become more significant (León Soto and Valenzuela, 2013). Teleseismic measurements farther southeast along the MAT, to the east of the subducted extension of the Tehuantepec Ridge, show small delay times and variable orientations of the fast polarization directions, indicative of little mantle anisotropy (Ponce-Cortés, 2012). ...
Subducting plates around the globe display a large variability in terms of slab geometry, including regions where smooth and little variation in subduction parameters is observed. While the vast majority of subduction slabs plunge into the mantle at different, but positive dip angles, the end-member case of flat-slab subduction seems to strongly defy this rule and move horizontally several hundreds of kilometers before diving into the surrounding hotter mantle. By employing a comparative assessment for the Mexican, Peruvian and Chilean flat-slab subduction zones we find a series of parameters that apparently facilitate slab flattening. Among them, trench roll-back, as well as strong variations and discontinuities in the structure of oceanic and overriding plates seem to be the most important. However, we were not able to find the necessary and sufficient conditions that provide an explanation for the formation of flat slabs in all three subduction zones. In order to unravel the origin of flat-slab subduction, it is probably necessary a numerical approach that considers also the influence of surrounding plates, and their corresponding geometries, on 3D subduction dynamics.
... In the region southwest of the 100 km isodepth contour, where the slab is shallower, the anisotropy pattern is less clear given that fast axes are oriented in different directions. For these shallower earthquakes, the path through the mantle wedge is shorter, and thus the anisotropy contributions from the continental crust and also from the slab itself become more significant (LEÓ N SOTO and VALENZUELA 2013). It should be noted that there is no active andesitic volcanism in the Isthmus of Tehuantepec even as the Cocos slab reaches depths in excess of 110 km (RODRÍGUEZ-PÉ REZ 2007; CASTRO-ARTOLA 2010), a depth where this type of volcanism is common in subduction zones worldwide (TATSUMI and EGGINS 1995;SYRACUSE and ABERS 2006). ...
Shear wave splitting measurements were made using SKS and SKKS waves recorded by the Meso-American Subduction Experiment, which was deployed in southern Mexico starting at the coast of the Pacific Ocean and running north toward the Gulf of Mexico. In this segment of the Middle America Trench the oceanic Cocos plate subducts under the continental North American plate. The active volcanic arc is located at the southern end of the Trans-Mexican Volcanic Belt. Unlike most subduction zones, however, the volcanic arc is not subparallel to the trench. In the fore-arc, between the trench and the Trans-Mexican Volcanic Belt, the Cocos slab subducts subhorizontally. Beneath the volcanic belt, however, the slab dives steeply into the mantle. A marked difference in the orientation of the fast polarization directions is observed between the fore-arc and the back-arc. In the fore-arc the fast axes determined using SKS phases are oriented NE–SW, in the same direction as the relative motion between the Cocos and North American plates, and are approximately perpendicular to the trench. Physical conditions in the subslab mantle are consistent with the existence of A-type olivine and consequently entrained mantle flow is inferred. Strong coupling between the slab and the surrounding mantle is observed. In the back-arc SKS fast polarization directions are oriented N–S and are perpendicular to the strike of the slab. Given the high temperatures in the mantle wedge tip, the development of A-type, or similar, olivine fabric throughout the mantle wedge is expected. The orientation of the fast axes is consistent with corner flow in the mantle wedge.
... Based on numerical studies Schellart and Moresi (2013) suggest that a driving mechanism potentially inducing a compressional regime in the back arc is favored by poloidal asthenospheric flow in the mantle wedge. León-Soto and Valenzuela-Wong (2013) measure shear-wave splitting using S waves from local earthquakes recorded near the coast of the Gulf of Mexico in the Isthmus of Tehuantepec. A geometrical distribution of earthquakes and seismic stations allowed a detailed sampling of the mantle wedge. ...
... A geometrical distribution of earthquakes and seismic stations allowed a detailed sampling of the mantle wedge. The results of León-Soto and Valenzuela-Wong (2013) Table 1. Source Parameters of the Earthquakes in the Southern Gulf of Mexico. ...
The earthquakes of 23 May 2007 (Mw 5.6) and 29 October 2009 (Mw 5.7) that occurred in the southern continental margin of the Gulf of Mexico were studied by conducting a full wave inversion of teleseismic P waves. In this region, there is a band of shallow seismicity along the continental margin that extends from the Isthmus of Tehuantepec to the city of Veracruz, Mexico. The focal mechanism of the 2009 event shows reverse faulting at a high angle and is very similar to that of the 1959 Jáltipan earthquake (Mw 6.4) and to the 1973 Veracruz event (Mw 5.3). The focal depths between 22 and 27 km of these three earthquakes are unusually deep for continental events. The location of this band of seismic activity and the focal mechanisms suggest a process of crustal shortening of the southern margin of the Gulf of Mexico. This compressive regime appears to be induced by the subduction of the Cocos plate to the south, in a manner that is reminiscent of the crustal shortening process observed in the Andes. The 2007 earthquake took place north of this region, seaward of the city of Tuxpan. The focal mechanism shows strike slip faulting and the hypocentral depth obtained from the body wave inversion is approximately 7 km. The source mechanism of this earthquake and the depth in the upper crust where it took place suggest a different tectonic process than the one observed near the Isthmus of Tehuantepec.
Cocos intraslab earthquakes were used to make shear-wave splitting measurements to explore the factors that control seismic anisotropy and to study the mantle wedge flow patterns in southeastern Mexico, where the Cocos plate subducts beneath the North American plate. Cocos intraslab earthquakes reach depths of 250 km, making it possible to sample the mantle wedge. The Silver and Chan (1991) Silver and Chan (1991) covariance method was used to measure the splitting parameters: the fast polarization direction (ϕ) and the delay time (δt). The measurements can be divided into three regions: (1) northwest and (2) southeast of the Tehuantepec Ridge extension (i.e., the subducted Tehuantepec Ridge within the Cocos slab) and (3) the region above the subhorizontal Cocos slab. (1) In the first region, northeast of the 100 km isodepth contour of the Cocos slab, the fast axes are trench-perpendicular. This can be explained assuming the development of A-type olivine fabric and the existence of 2-D corner flow driven by the downdip motion of the Cocos slab. Southwest of the 100 km isodepth contour, measurements show trench-parallel fast polarization directions that are also consistent with corner flow, albeit in a serpentinized mantle wedge. Right above the Tehuantepec Ridge extension (northeast of the 100 km isodepth contour of the subducting slab), a change in the fast polarization directions from trench-normal to trench-parallel while going from northwest to southeast is observed and signals a change in the mantle flow pattern possibly through a vertical tear in the Cocos slab. 3-D toroidal flow drives subslab mantle material around this slab edge and into the mantle wedge. (2) In the second region, the measured fast polarization directions show a trench-parallel orientation that is interpreted to result from southeastward trench-parallel flow through a serpentinized mantle wedge tip and also through a mantle wedge core made up of A- or C-type olivine fabrics with their fast axes oriented parallel to the flow direction. Trench-perpendicular fast polarization directions are observed beneath the fore-arc region of the Central America Volcanic Arc, near Tacaná Volcano, and trench-parallel polarizations are observed beneath the arc. These orientations could be explained by assuming the presence of B-type olivine fabric in the mantle wedge tip, A- or C-type olivine fabric in the mantle wedge core, and trench-parallel flow, so that the orientations of the fast axes become perpendicular to the mantle flow direction beneath the fore-arc and parallel to it beneath the arc. Lastly, (3) in the third region, over the flat slab, the observed delay times (0.04–0.42 s) are consistent with crustal anisotropy magnitudes, and the fast polarization directions seem to be controlled by the orientations of fault systems and alignments in foliations. Therefore, crustal faults and folds seem to be the dominant factors controlling the observed anisotropy.
