Neogene structure of the Andean Precordillera, Argentina: insights from analogue models

Analogue models combining different sets of preexisting structural weaknesses were developed to understand their evolution during regional ~ENE shortening. Strain analysis of simulations was performed with the GEODEF 1.1 software, a tool that allows to quantify deformation in plan view on the basis of displacement fields. Results showed up that regional NNE heterogeneities are reactivated as dextral reverse-slip structures, though NNE neoformed thrusts are also present. Likewise, dominant sinistral strike-slip motions have been obtained for reactivated second-order WNW structures whereas sinistral reverse-slip has been recorded for NW ones. Comparison of these results with structural, kinematic and paleomagnetic data supports partitioned dextral transpression for the northern Andean Precordillera since the Miocene. Moreover, models not only confirm sinistral strike-slip motions for WNW structures of the Precordillera but also suggest that they would represent preexisting crustal fabrics that were reactivated during the Andean orogeny. These cross-strike structures have played a significant role in the construction and evolution of the fold and thrust belt as they segmentate the activity of orogen-parallel structures.


Introduction
The Precordillera represents an almost northtrending morphostructural unit that was developed within the Andean foreland due to flat-slab subduction in the Pampean segment (Fig. 1; Ramos, 1999a;Ramos et al., 2002).This process is related to the subduction of the Juan Fernández Ridge between 28º and 33ºS since Miocene times and gives rise to the migration of deformation and arc magmatism into the foreland (Pilger, 1984;Kay et al., 1987;Allmendinger et al., 1990;Jordan et al., 1993;Ramos, 1999a).
The Andean Precordillera can be divided into two segments: northern and southern.The northern Precordillera comprises a NNE-trending fold and thrust belt (Allmendinger et al., 1990;Cristallini and Ramos, 2000) while the southern Precordillera evolved as a consequence of the reactivation of NNW to NNE Paleozoic-Triassic structures (Cortés et al., 2005(Cortés et al., , 2006;;Giambiagi et al., 2010).However, the tectonic evolution of the northern Precordillera is still controversial as different models have been proposed.The fold and thrust belt was defined by Allmendinger et al. (1990) and Cristallini and Ramos (2000), whereas Ré et al. (2001), Siame et al. (2005) and Álvarez Marrón et al. (2006) considered dextral transpression for the Andean deformation of the Precordillera.
The aim of this paper is to understand the structural evolution of the northern Precordillera on the basis of analogue models considering the role and interaction of major NNE structures and subordinated WNW cross-strike structures.These models have been interpreted with the software GEODEF 1.1 (Yagupsky, 2010) which let to quantify both incremental and finite strain.Consequently, the Andean deformation of the Precordillera is constrained by a comparative analysis between deformation patterns in models and geological data.

Regional framework
The northern Precordillera is located in the Andean foreland between the Frontal Cordillera to the west and the Sierras Pampeanas to the east (Fig. 1).It represents a fold and thrust belt that can be divided into three segments (Western, Central and Eastern) with distinctive structural and geological features (Ramos, 1999b and references therein).Western and Central Precordillera represent an east-verging thin-skinned fold and thrust belt, whereas the Eastern Precordillera shows west-verging thick-skinned deformation which is related to the Sierras Pampeanas structure (Fig. 2).Consequently, a thick-skinned triangle zone is developed between both belts (Zapata and Allmendinger, 1996).
Many authors have proposed different hypothesis concerning the Cenozoic tectonic evolution of the Precordillera.The first proposals considered the Precordillera as a fold and thrust belt that developed mostly during the Miocene due to nearly E-W shortening (Allmendinger et al., 1990;Cristallini and Ramos, 2000).Later, Siame et al. (2005) suggested that the slightly oblique convergence between Nazca and South American plates gives rise to dextral transpression related to a compressive regime.According to these authors, Plio-Quaternary deformations are partitioned between thrusting in the Eastern Precordillera and the Western Sierras Pampeanas and dextral strike-slip motions in the El Tigre fault (Fig. 2, Bastías et al., 1984;Siame et al., 1997;Cortés et al., 1999). However, Álvarez Marrón et al. (2006) proposed that deformation in the Precordillera is the result of dextral transpression that was developed since Neogene times and produced orogen longitudinal extension and orogen perpendicular compression.Likewise, Ré et al. (2001) and Japas et al. (2002a) have previously suggested a transpressive system between 22º and 33ºS based on tectonic fabric analysis.

WNW cross-strike structures
Two systems of conjugated megashear zones at the study latitudes have been proposed (Japas, 1998;Ré et al., 2000Ré et al., , 2001;;Japas et al., 2002a, b;Ré and Japas, 2004): left-lateral NNW and right-lateral NNE transpressional sets, and left-lateral WNW and right-lateral ENE transtensional ones.Left-lateral WNW structures have been confirmed within the Hualilán Belt (Fig. 2) by Oriolo (2012) and Oriolo et al. (2013).These authors have remarked the role of these WNW and subordinated ENE cross-strike structures as the main structural control of magmatism emplacement due to the into-the-foreland migration associated with the flat-slab process.Kinematics of cross-strike structures shows dominant strike-slip displacements as well as a minor component of extension that would favor the magmatic output and emplacement.Ré et al. (2001), Oriolo (2012) and Oriolo et al. (2013) have also proposed the existence of other cross-strike structural belts, that would be equivalent to the one located in the Hualilán region.One of these is placed between Gualcamayo and Jachal localities (~30º S), where Chernicoff and Nash (2002) have demonstrated the relationship between Cenozoic magmatism and associated ore deposits and NW-WNW cross-strike structures.
During the last decades, physical models applied to understand strike-slip deformation combined with other geological processes diversified.Richard and Cobbold (1989) and Pinnet and Cobbold (1992) analyzed strain partitioning mechanisms related to transpression due to oblique convergence.Likewise, Le Guerroué and Cobbold (2006) studied the influence of erosion and sedimentation on strike-slip faults, whereas Corti et al. (2005), Mathieu and van Wyk de Vries (2011) and Mathieu et al. (2011) investigated relationships between transtension/transpression, magmatism and volcanism.α=15° between the layer boundary and the plate A, in order to favor the localization of deformation in this area.Within the silicone layer, second-order oblique discontinuities filled with sand were set out in models II and III considering β angles between them and the plate A of 100º and 130º, respectively (Fig. 3).The model was then covered with ~1 cm thick sand layers.
The value of α=15° was selected in order to represent major inherited structures of the Precordillera which have been interpreted as terrane boundaries (i.e., Giménez et al., 2008) and are slightly oblique to both directions of convergence of the Nazca plate and regional shortening in the Precordillera (Brooks et al., 2003).Subordinated oblique sand ribbons are induced to analyse the influence of pre-existing heterogeneities in the basal level.Their influence during shortening of the overlying sand pile can be compared with cross-strike structures described by Oriolo (2012) and Oriolo et al. (2013).Therefore, they were not included in model I in order to determine if they could be developed when they do not represent inherited features.
However, analogue models combining different sets of structures with strike-slip displacements are still scarce.Within this framework, this contribution provides information about this interaction and suggests an alternative structural and kinematic model for the Andean foreland system.

Experimental configuration
Three sandbox analogue experiments (I, II, III) were developed in order to simulate major structural features and their role during the evolution of the northern Precordillera.The models were carried out in a deformation sandbox with dimensions of 70x52x4 cm (Fig. 3), wide enough to avoid boundary effects.Orthogonal compression was applied to a rigid plate A linked to an acetate plate B (Fig. 3).They both were moved towards the right of the model with a velocity of 1x10 -5 m s -1 as the shortening increased while an acetate plate C was fixed to the modeling table.A 0.5 cm thick silicone layer was arranged between plates B and C with an angle The analogue materials were well sorted finegrained dry quartz sand and silicone.The sand has a density ρ=1,400 kg m -3 , an internal friction angle φ=32.7° and negligible cohesion C 0 <100 Pa (Yagupsky et al., 2008) and, therefore, is a good analog for the brittle upper crust (Davy and Cobbold, 1991).The silicone is a viscous material with a viscosity of 2.5x10 3 Pa s (Stewart, 1996, Likerman et al., 2013) that can be considered a Maxwell solid (Casassa et al., 1986).The selected configuration of sand overlying a silicone layer were used by other authors (Casas et al., 2001;Soto et al., 2007;Yagupsky et al., 2008;Leever et al., 2011;Likerman et al., 2013), supporting the validity of both set-up and materials.
The comparison between models and natural examples is only possible if the experiments are properly scaled.In this case, a geometric scale factor λ=2x10 -6 was considered, which represents the ratio between the length of the model and its equivalent length in the nature.Gravity ratio was g´=1, whereas density ratio was ρ´=0.6, considering a mean density ρ=2,500 kg m -3 for sedimentary rocks in the Precordillera (Perucca and Ruiz, 2014).Therefore, the stress ratio between models and nature was σ´=ρ´g´λ=1.2x10 - .The shortening rate was V=1x10 -6 mm y -1 and the shortening rate ratio was V´=2x10 -7 , if an average shortening rate V=5 mm y -1 for the last 20 Ma is assumed (Siame et al., 2005).

PIV and GEODEF processing
Photographs in plan view were taken every minute with a camera suspended above the model.These images were then processed using a PIV software (Sveen, 2004), that allows to quantify high-resolution displacement fields between two successive images within the grain-size range ( White et al., 2001;Adam et al., 2005).The software uses an optical correlation to obtain the displacement vectors (Sveen, 2004).
PIV results were reprocessed with the software GEODEF 1.1 (Yaguspky, 2010).GEODEF 1.1 considers the directional derivatives for each incremental vector between two pictures, allowing to calculate the incremental strain matrix for each point of the model.Finite strain for a specific n stage can be estimated as the product of all incremental strain matrices from the first to the n stage (Yagupsky, 2010).This method quantifies the deformation of systems (ellipticities, rotations, strain ellipses) even with extremely low shortening values (~2 mm).

Model I
Model I shows the simplest configuration and, thus, it does not include subordinated oblique heterogeneities.Figure 4 shows results obtained for this configuration.
The resulting structures are mostly thrusts (Fig. 4a-c), which is indicated by the presence of long axes of ellipses parallel to them supporting E-ENE shortening directions (Fig. 4d-f).However, a localized zone with WNW-NW long axes is also present over the western boundary of the silicone plate and it cannot be explained considering only thrusting.Additionally, this zone can be detected only after a certain amount of shortening.
In a first stage, shortening is mainly accommodated in en-échèlon structures that resulted from the reactivation of the silicone plate boundaries (Fig. 4a, d, g).The western and eastern structures consist of transpressional structures with respectively W-and E-vergence.En-échèlon structures then coalesce, giving rise to a sigmoid pattern (Fig. 4b, e, h).Particularly, the eastern structure shows higher values of ellipticity than the western one.The appearance of neoformed structures contributes to localized shortening, reducing the deformation rate of the early reactivated structures.During the third step, some thrusts developed out from the previously deformed zone bounded by the two main anisotropies, showing variable vergence direction (Fig. 4c).
Although deformation is registered all over the deformed zone placed between the NNE heterogeneities, clockwise rotations seem to be restricted to them (Fig. 4j-l).Areas with higher values of rotations seem to correlate with those with higher values of ellipticity supporting higher values of non-coaxial strain.Low counterclockwise rotation values were only recognized at the final stage of modeling and appear to be linked to the small bends of the NNE-striking structures (Fig. 4l).These bends seem to be related to ENE strike-slip faults that can be interpreted as transfer zones as they segmentate the chain: the northern domain shows deformation concentrated in the fold and thrust belt and its hinterland, and the southern one shows deformation localized in the fold and thrust belt and its foreland.

Models II and III
These models consider subordinated oblique discontinuities within the silicone layer.They were set up with two different orientations: β=100º (model II) and β=130º (model III).
Results obtained for model II reveal some similarities with those obtained in model I, as significant E-ENE shortening and localized clockwise rotations are present (Fig. 5).Ellipticity is higher at the backthrust differing with results obtained for model I, where the reactivated eastern heterogeneity has accumulated higher values of ellipticity.Moreover, the higher strain recorded by the western main fault shows two minimums along strike where the two sets of oblique faults are present (Fig. 5l).These localized areas also show deformation migration into the foreland, indicating strong segmentation of the chain by these structures.
WNW faults show neither significant shortening nor extension revealed by cross-sections of the models and very low values of ellipticity but they exhibit localized counterclockwise rotations (Fig. 5n-p).This would reflect dominant strike-slip displacements for these structures.Therefore, it can be interpreted that strain induced by localized pre-existing discontinuities overprints to the previous strain state.Counterclockwise rotations become evident with shortening up to ~3 cm.This would give rise to local sinistral strike-slip displacements that are active in a later step of the model evolution.Scarce thrusts develop in the foreland between WNW faults (Fig. 5d) and each segment of the western NNE fault bounded by these oblique structures grows independently (Fig. 5l).Therefore, these cross-strike structures comprise important deformational features since migration of deformation into the foreland appears to be segmented by them.
Model III evidences shortening and counterclockwise rotations associated with the reactivated oblique structures (Fig. 6).The amount of shortening over them is much higher than that obtained in model II.
NW structures reactivate prior to the development of the main NNE front and shortening is transferred into the foreland by these structures (Fig. 6l).NNE faults are strongly segmented along-strike (Fig. 6l).Therefore, reactivation of NW structures is interpreted as more efficient in migrating shortening into the foreland.
An along-strike zone of ENE-NE long axes is present in both models II and III but it seems to be more localized in the western anisotropy than in model I.This zone requires a certain amount of shortening (2.5-3.0 cm) to be developed, as it was also observed for model I.

Deformation patterns
Results obtained for model I (Fig. 4a-c) are similar to those obtained by Casas et al. (2001) for equivalent initial conditions.N-NNW long axes supporting E-ENE shortening directions together with clockwise rotations reflect dextral transpression for structures generated by reactivation of the boundaries of the silicone plate in all three models.Moreover, the slight obliquity between the NNE strained zones and the direction of shortening supports the dextral transpression.On the other hand, NNE neoformed structures show almost no rotations and are parallel to long axes of ellipses, so they can be interpreted as genuine thrusts.
The absence of significant dip-slip displacements and the presence of counterclockwise rotations on WNW reactivated structures can be interpreted as the result of dominant sinistral strike-slip displacements in model II (Fig. 5).Deformation associated with WNW structures (Fig. 5m-o) appears once the model arises a certain amount of shortening (~3 cm).This implies that these structures overprint to the previous NNE ones and control late development of some of them.Likewise, the presence of these oblique structures gives rise to along-strike segmentation of NNE structures (Fig. 5k-l).
Model III shows similarities to models I and II but a much more significant component of dip-slip displacement in the NW structures (Fig. 6) revealed by high values of ellipticity and observations in cross-sections.Together with counterclockwise rotations, these results point out to sinistral reverse-slip motions for the NW structures.
If all three models are considered, it can be observed that oblique WNW-NW structures are not developed (model I) if there is not an inherited fabric of similar orientation that can be reactivated (models II and III).These results are consistent with several proposals (Flinch and Casas, 1996;Pohn, 2001;Japas et al., 2010;Oriolo, 2012;Oriolo et al., 2013) where cross-strike structures are described as pre-existing crustal fabrics.Moreover, similar geometrical and kinematical relationships between dextral shear zones and secondary oblique sinistral ones were presented by theoretical models of Jiang (1994) and examples from the Zagros fold and thrust belt (Hessami et al., 2001), the Atacama fault system (Cembrano et al., 2005) and the Cap de Creus (Carreras et al., 2013).However, oblique ENE-trending structures do develop even though previous structures are not present (Fig. 4i).
The along-strike zone with WNW-NW long axes of ellipses is clearly developed in all models.It is present only over the western margin in models II and III whereas it is wider in model I.The presence of this zone seems to be independent from ellipticity as it is always located in the western structure, which has lower ellipticity than the eastern one in model I but higher in models II and III.Likewise, it requires a certain amount of shortening to be developed that fits with the step when deformation from western and eastern structures start to overlap.This 2D distribution of strain axes could be explained either by WNW extension or NNE shortening.A WNW direction of extension could result from local extension of the hanging-wall such as described by Bonini et al. (2000).This mechanism would give rise to strike-parallel extension that would explain this pattern.However, cross-sections do not reveal extension in all models.Then, a NNE shortening direction is likely to be a most suitable interpretation that can be explained considering interference patterns.As this zone is developed due to interference patterns generated by overlapping of deformation from western and eastern structures, it would probably represent local areas of strain interference patterns that are progressively modified by regional dextral transpression.
Results and interpretations presented in this paper point out to a two-stage evolution model associated with heterogeneous strain partitioning.Three domains 1, 2 and 3 are considered to explain this model (Fig. 7).In the first stage (shortening lower than 2.5-3.0 cm), domains 1 and 2 (western and eastern structures, respectively) would represent areas of transpression, whereas domain 3 (areas of neo formed thrusts) is related to compression due to E-ENE shortening.In the second stage (shortening higher than 2.5-3.0 cm), domain 1 not only shows dextral transpression but also local interference patterns.Likewise, domains 2 and 3 behave as in the first stage.WNW sinistral cross-strike structures activate during this stage suggesting that they could be related to the change in deformation patterns in domain 1.
The experiments presented herein agree with Jones et al. (1997) about a partitioned transpressional system with strike-slip components of deformation localized in narrow zones and broader areas of relative coaxial shortening.However, this proposal shows a much more complex pattern of strain partitioning than the classical model from Tikoff and Teyssier (1994).This would be explained due to the presence of pre-existing structures in the silicone layer, whose margins are also slightly oblique to the regional shortening direction.In agreement with Carreras et al. (2013), the existence of structural weaknesses gives rise to complex patterns of strain partitioning, which is favored by their presence (Dewey et al. 1998).In addition, the observed localization of strike-slip components in the reactivated structures matches ideas of Lister and Williams (1983), which proposed that non-coaxial flow is favored by suitably oriented inherited structures.

Models and the Neogene structure of the Precordillera
Dextral transpression obtained in models by the presence of N-NNE thrusts and clockwise rotations are consistent with proposals from Siame et al. (2005) and Álvarez Marrón et al. (2006) for the Precordillera.Particularly, strain partitioning into areas dominated by dextral oblique-slip or dip-slip displacements obtained in simulations presented herein supports the hypothesis of Siame et al. (2005).Within this framework, domain 1 would be equivalent to the dextral strike-slip El Tigre fault (Bastías et al., 1984;Siame et al., 1997;Cortés et al., 1999) located in the western margin of the Precordillera (Fig. 7a, b).Differences between transpression obtained in models and dominant strike-slip displacements recorded on the El Tigre fault could be explained due to kinematic changes in the later, as some authors have suggested (Fazzito, 2011;Fazzito et al., 2011). However, Álvarez Marrón et al. (2006) described a positive flower structure developed as the result of transpression in the Iglesia basin, located immediately to the north of the El Tigre fault, which would match better the kinematics observed in the models.Neoformed thrusts from domain 3 would be the analogous of the Precordillera fold and thrust belt (Allmendinger et al., 1990;Cristallini and Ramos, 2000) whereas domain 2 would be equivalent to the Tulum fault system (Fig. 7a, b), where both reverse and dextralstrike slip deformation have been recently registered (Perucca and Ruiz, 2014).Furthermore, these equivalences allow interpreting that the obliqueslip displacements may result from reactivation of preexisting basement structures that correspond to Paleozoic terrane boundaries (Giménez et al., 2008;Perucca and Ruiz, 2014).
Minor differences between models and data from the Precordillera could be explained considering the complexity of the strain partitioning processes (Jones and Tanner, 1995;Chemenda et al., 2000) that were not considered in the analogue models.In addition, the lack of a wide database of kinematic indicators in the Precordillera faults avoids carrying out more detailed analysis and comparisons.
Regional E-ENE shortening directions obtained by strain patterns from the models are similar to those obtained for the Precordillera by Siame et al. (2005) on the basis of kinematic data.Clockwise rotations observed in the models are related to dextral transpression and are comparable to paleomagnetic data from Japas and Ré (2012) and Vizán et al. (2013).Likewise, these clockwise rotations are equivalent to rotations of kinematic axes (Siame et al., 2005;Oriolo et al., 2013) within the Precordillera.These rotations around vertical axes could be also explained by changes in the direction of convergence between the Nazca and South America plates.However, Somoza and Ghidella (2005) have provided data that supports a rather constant direction of convergence during the last 26 My.
Comparison between analogue models and structural and geophysical data from the Hualilán area (Oriolo, 2012;Oriolo et al., 2013) suggests that model II is the best analogue of cross-strike structures in the Precordillera (Fig. 7a, b).Strain axes and rotations obtained in model II reflect dominant sinistral strike-slip motions and are also similar to those obtained through kinematic indicators in the Hualilán belt by Oriolo (2012) and Oriolo et al. (2013).Moreover, temporal relationships are equivalent (cross-strike structures postdating regional thrusting) in both model and field data.The Hualilán belt would be related to the Pie de Palo Norte lineament (Fig. 2; Oriolo et al., 2013), which also shows equivalent geometry and kinematics to those observed in the models (Zapata, 1998).The Pie de Palo Norte lineament was considered to be the boundary of a Mesozoic rift basin (Martínez and Colombi, 2011), indicating a pre-Neogene age for this structure.Similar relationships have been also reported in the northernmost Precordillera (~29º45'S) by Chernicoff and Nash (2002).These authors recognized sinistral NW structures that may represent reactivated preexistent features and crosscut and rotate structures of the fold and thrust belt.
Model III shows similar deformation patterns with those obtained by analogue modeling of rift inversion by Yagupsky et al. (2008) and Yaguspky (2010).These results are both in agreement with structural data from the southern Precordillera (Fig. 7a, c), where a Cenozoic sinistral transpression has been reported due to reactivation of Paleozoic-Triassic NNW-NNE structures (Cortés et al., 2005(Cortés et al., , 2006;;Giambiagi et al., 2010;Terrizzano et al., 2010).Particularly, the structures obtained in the models would be equivalent to the Barreal-Las Peñas belt (Fig. 7a; Terrizzano et al., 2010) and similar systems located therein.
The absence of WNW structures in model I supports the idea that these structures do not developed unless they represent pre-existing fabrics (models II and III).This is consistent with hypotheses that suggest NW-WNW cross-strike structures as pre-Neogene structures that were reactivated during the Miocene (Chernicoff and Nash, 2002;Oriolo, 2012;Oriolo et al., 2013).Different shortening rates between along-strike segments of the Precordillera could have favored the development of WNW structures as neoformed transference zones.However, all previous studies (Allmendinger et al., 1990;Cristallini and Ramos, 2000;Álvarez Marrón et al., 2006) point out to deformation and uplifting of the northern Precordillera as a single block and consequently the hypothesis of WNW neoformed structures is discarded.

Along-strike segmentation and cross-strike structures
Segmentation models suggest that faults are divided into discrete units that behave distinctively during rupture cycles (Schwartz and Coppersmith, 1986).These units are separated by segment barriers than inhibit propagation of rupture towards other segments (Aki, 1979(Aki, , 1984)).Several authors have applied these segmentation models mostly on single fault seismic cycles and their rupture patterns (Schwartz and Coppersmith, 1984;Mueller and Talling, 1997;Pizzi and Galadini, 2009;Lin et al., 2011).
Models presented herein reveal that WNW cross-strike structures behave as persistent barriers of rupture along major NNE faults.They give rise to local bends as well as they segmentate the strain patterns (Fig. 5f-h).These results suggest that cross-strike structures not only behave as barriers of single faults ruptures as many authors have already remarked (Schwartz andCoppersmith, 1984, 1986;Pizzi and Galadini, 2009;Lin et al., 2011) but also develop a significant role in the evolution and construction of orogens.

Concluding remarks
Analogue models presented herein provide data about deformation patterns resulting from the interaction of different set of structures.Comparison between these results and structural and geophysical data from the Precordillera let to outline these main conclusions: -NNE heterogeneities are reactivated as dextral reverse structures that accommodate both E-ENE shortening and clockwise rotations.Neoformed NNE structures are also developed as genuine thrusts.These results support partitioned dextral transpression for the northern Andean Precordillera related to a compressive regime, confirming previous proposals from Siame et al. (2005).
However, dextral strike-slip displacements would not be restricted to the El Tigre fault but also to the eastern margin, as the models and recently published data suggests (Perucca and Ruiz, 2014).-WNW and NW weaknesses are reactivated as sinistral strike-slip (β=100º) or sinistral transpressional (β=130º) structures that do not develop unless they represent inherited structures.WNW structures are equivalent to those described in the northern Precordillera as they show similar orientation and kinematics and they postdate regional thrusting.NW anisotropies show similarities with Paleozoic-Triassic structures reactivated during the Andean orogeny in the southern Precordillera.-Cross-strike structures have a major role in the segmentation of the thrust belt.Therefore, they have to be carefully considered in further works related to seismic hazards within the Precordillera, as they are probably behaving as barriers of fault rupture.

FIG. 1 .
FIG. 1. Location of the Pampean flat-slab segment of the Central Andes.Contours of depth (km) of the oceanic slab (after Anderson et al., 2007) and main morphostructural units are shown.Area from figure 2 is indicated.

FIG. 3
FIG. 3. a. Schematic representation of the model setting.Shortening direction is shown (black arrows); b.Plan view.The angles between the plate A and the silicone layer (α) and the secondary oblique heterogeneities (β) are indicated.

FIG. 4 .
FIG. 4. Deformation stages of model I, considering main structures (a-c), long axes of finite strain ellipses (d-f), cumulative ellipticity (g-i) and cumulative rotations (j-l).Values of ellipticity and rotations (negative rotations, clockwise; positive rotations, counterclockwise) are shown in lateral bars.White lines indicate schematic position of anisotropies.Small arrows show areas of local counterclockwise rotations.

FIG. 5 .
FIG. 5. Deformation stages of model II, considering main structures (a-d), long axes of finite strain ellipses (e-h), ellipticity (i-l) and rotations (m-p).Dotted lines show position of the basal WNW anisotropies.Values of ellipticity and rotations (negative rotations, clockwise; positive rotations, counterclockwise) are shown in lateral bars.Small arrows show areas of local counterclockwise rotations.

FIG. 7 .
FIG. 7. Comparison of the structure of the Precordillera and analogue models.a. Structural map of the Precordillera (modified from Fig. 2); b.Model II (modified from Fig. 5d) showing the main structural features comparable to the northern Precordillera.Domains 1, 2 and 3 as well as sinistral WNW structures equivalent to the Hualilán belt are shown; c.Model III (modified from Fig. 6d) showing the main structural features comparable to the southern Precordillera.Sinistral reverse-slip structures equivalent to those present in the southern Precordillera are indicated.