Structural overprint of a late Paleozoic accretionary system in north-central Chile ( 34 °-35 ° S ) during post-accretional deformation

In the Coastal Cordillera of central Chile a coherently preserved architecture of a late Paleozoic accretionary prism is exposed at 36°-35°S in close spatial association with a neighbouring area at 34°-35°S, where it is strongly modifi ed by post-accretional processes. Synand post-accretional structures can be distinguished relatively easily in this region studying the deviations from the original architecture. South of 35°S a transitional contact between two major units is observed, which refl ects a continuous change of the mode of accretion in the accretionary wedge before ~305 Ma: the structurally overlying metagreywacke of the Eastern Series exhibits structures typical of frontal accretion, i.e., subvertical chevron folds of bedding planes with an axial-plane foliation S1. With increasing fi nite strain structurally downwards, open F2 folds develop associated with a S2-foliation which becomes gradually fl attened as it rotates into a subhorizontal orientation. S2 is the penetrative transposition foliation in the structurally underlying Western Series. It affects the continent-derived metagreywacke series as well as metabasite intercalations of oceanic origin and was formed during basal accretion. This principal evolution of the accretionary system places fi rm constraints on the original architecture also in regions where it was destructed after accretion. Accretion ceased at ~225 Ma, when a major tectonic change from a convergent to an extensional/strike-slip regime occurred. Although the development of the margin in central Chile is largely characterized by extension during Mesozoic and Cenozoic times, two pronounced episodes involving shortening of the forearc particularly affected the Western Series north of 35°S: 1. Expressions of strike-slip activity during Jurassic times involve local steepening of the originally fl at S2-foliation planes, local rotation of the stretching lineation L2 into the N-S direction, tight upright folding of the S2 foliation and refolding about steep axes with associated vertical cataclastic left-lateral shear zones. 2. The left-lateral reverse Pichilemu-Vichuquén fault at the boundary between both units is a prominent brittle structure that formed at ~100 Ma concomitant with basin closure and acceleration of exhumation rates in the forearc. Similar contractional structures occur along the coast further north, where both units partly disappeared by subduction erosion most likely during these deformation episodes. The transition between nearly unaffected accretion systems in the south and disrupted and partly subducted ones in the north occurs at 35°S.


Introduction
Accretionary systems at long-lived convergent continental margins such as those in the circum-Pacifi c region are time-restricted phenomena and are usually strongly disrupted by post-accretional deformation processes, when activity at these margins continued under changing plate geometries and kinematics.Hence it is a general problem to distinguish between syn-and post-accretional structures in order to provide better data for tectonic models describing the formation of accretionary systems and other plate confi gurations at active margins.Here we present an example from central Chile, where a nearly coherent accretionary wedge architecture occurs in close vicinity to one strongly modifi ed by post-accretional processes.
The formation of the metamorphic basement within the Coastal Cordillera of Chile between 26° and 55°S marks the last increment of lateral growth of the continental margin of South America during late Paleozoic times in the north and Mesozoic times at its southernmost end.In the area between Pichilemu and Constitución (34°-35°30'S; Fig. 1) González-Bonorino (1971) and Aguirre et al. (1972) fi rst proposed that this basement represents deep-seated parts of a palaeo-accretionary wedge within a paired metamorphic belt, which was active in late Paleozoic times (Hervé et al., 1984).Whereas the architecture of this accretionary system was strongly overprinted in northern Chile during Mesozoic times, it is largely well preserved in central Chile, where its characteristics have recently been studied in detail (Glodny et al., 2005(Glodny et al., , 2006;;Willner, 2005;Willner et al., 2005;Richter et al., 2007).Traditionally two units of contrasting structural evolution are recognized in central Chile (Hervé et al., 1984): The Western Series of mixed oceanic and continental rocks is regarded as a high pressure/low temperature unit, whereas the Eastern Series composed of metagreywackes only experienced a low pressure/high temperature metamorphic overprint related to intrusion of an arc batholith.The concept developed in this region was later adapted to the entire basement of the Chilean Coastal Cordillera between 26° and 55°S defi ning it as a collage of accretionary prisms of diverse origin and age (Hervé, 1988).Revisiting the original type FIG. 1. Geological map of the Coastal Cordillera at Pichilemu and Constitución (34°-35°40'S) with cross section AB.Mapping information in this paper was also compiled from Godoy (1970), González-Bonorino (1971), Moraga (1981), Gana and Hervé (1983), Hervé et al. (1984), Klähn (1999) andBravo Espinosa (2001).
area in north-central Chile (Fig. 1) it was shown that accretion activity was restricted to about 100 Myrs duration (319 Ma to 224 Ma) being exhumed at a rate of 0.2-0.6 mm/yr with erosion as the major agent of exhumation during continuous cyclic mass fl ow (Willner et al., 2005).Both series have Carboniferous depositional ages (Willner et al., 2008).The Western Series was overprinted at high pressure greenschist facies apart from local blueschist facies conditions (350-420°C, 7-11 kbar), whereas the Eastern Series was contemporaneously overprinted at 3 kbar with progressively rising temperature of 400°-720°C towards a batholith in the east that intruded in the east representing the Late Palaeozoic magmatic arc (Willner, 2005).It was shown by Richter (1977) and Richter et al. (2007) that the two major units represent two different principal modes of accretion: initial frontal accretion in the Eastern Series that gradually changes to basal accretion (underplating to the base of the wedge) in the Western Series.Particularly, the traverse east of Constitución (35°30'S) was shown to represent a complete and coherent architecture of the accretionary system (Willner, 2005;Richter et al., 2007).
Particularly in northern Chile, episodic destruction of the continental margin by shortening-related subduction erosion during Mesozoic times is supposed to have affected the metamorphic basement (e.g., Scheuber et al., 1994;Scheuber and González, 1999).Kay et al. (2005) proposed that subduction erosion affected the margin at 34°S during late Tertiary times.Glodny et al. (2006) argued that the entire margin of north-central Chile north of the Lanalhue fault at 38°S was affected by subduction erosion already around 270 Ma, i.e., during the time of accretion.In this paper, we challenge these contrasting views and reassess the nature of the link between processes leading to destruction and shortening of the outermost margin and structures actually observed in the fi eld.We consider subduction erosion as tectonic removal of continental forearc crust from the toe of the accretionary wedge.
The aim of this paper is to distinguish original accretionary structures and their orientations, from later structures that reorientated these original structures.The later structures are related to a stepwise destruction of the accretionary system and were caused by episodic events during continuing activity along the convergent margin and possibly by subduction erosion.We will show that the type region where the characteristics of the metamorphic basement of the Coastal Cordillera were originally described, also represents a transition between the largely coherent architecture exposed in south-central Chile and strongly disrupted and overprinted occurrences in north-central and northern Chile.

The Eastern Series
The Eastern Series represents the structurally upper unit of the accretionary system in central Chile, which is entirely composed of metagreywacke and metapelite of turbiditic origin with well preserved sedimentary structures such as graded and convolute bedding, load casts and crossbedding.The characteristic structures are best observed in the eastern part of the entire outcrop area, but become continuously obliterated by increasing deformation intensity toward the west: an orthostratigraphic coherence at outcrop scale and upright to slightly westvergent chevron folds of bedding planes on a meter to decameter scale with a pronounced axial-plane cleavage S 1 (Fig. 2A) F 1 fold axes strike NE-SW in the southern part of the study area and NNW-SSE in the northern part.This is consistent with a prominent change of strike of bedding/S 1 -planes as well as of the transition zone to the Western Series (Fig. 1).
Along two prominent traverses from E to W, i.e., the Río Maule transect (Godoy, 1970;Richter et al., 2007) and along the abandoned railway line east of Pichilemu, it can be observed that a second foliation S 2 continuously evolves from a widely spaced foliation parallel to subhorizontal west-vergent open F 2 -folds (Figs.2A, B) with a penetrative transposition foliation S 2 that becomes increasingly fl attened towards the west, where it becomes the dominant regional foliation in the Western Series.A SW-NEtrending stretching lineation L 2 is expressed on early formed quartz veins, which were transposed parallel to the subhorizontal S 2 -planes.The D 2 fabrics indicate substantial vertical fl attening and ductile thinning.The shortening axis rotated from a subhorizontal to a subvertical position from the eastern area where D 1 structures are dominant towards deeper structural levels in the west where D 2 becomes pervasive.
At shallow crustal levels the entire Eastern Series was intruded by the batholith of the Late Paleozoic magmatic arc and overprinted by a pervasive high temperature/low pressure metamorphism at about 300 Ma (Willner, 2005;Willner et al., 2005).The batholith had its main intrusive pulse at that time, but its activity continued until ~250 Ma (Willner et al., 2005).Statically grown porphyroblasts (biotite, andalusite, staurolite) in the metapsammopelitic rocks of the Eastern Series may locally be slightly rotated and stretched (Fig 2C).Stretching as evidenced by quartz-fi lled tension gashes in the porphyroblasts is generally in the direction of the regional stretching lineation L 2 , which can be observed with similar orientation in the Eastern and Western Series.This means that during and after HT-metamorphism of the Eastern Series the structural development was continuously infl uenced by the same deformation fi eld as during deformation of the Western Series.This observation is in line with previous Ar/Ar dating (Willner et al., 2005) indicating that HT-metamorphism in the Eastern Series was contemporaneous with accretion and high-pressure metamorphism in the Western Series.

The Western Series
The Western Series, which structurally underlies the Eastern Series north of 35°S, is dominated by two contrasting lithotypes: predominant continental metapsammopelitic phyllites and intercalated lenses of oceanic origin (metabasite with occasionally associated metachert, Fe-or Mn-rich quartzite and serpentinite).In spite of their different rheological properties both major rock types show the same deformational overprint and the same sequence of fabric formation, although with differing intensity.Primary structures are mostly erased in the metapsammopelitic rocks, but are locally present in the intercalated metabasites as pillows and hyaloclastitic fabrics (Fig. 2D, E).Further relics are a transposed S 1 -foliation, and abundant quartz veins of a fi rst generation formed during prograde dissolution.Both are folded by rootless intrafolial isoclinal F 2 folds during formation of the predominant transposition foliation S 2 (Figs.3A, B).Early formed planar structures as bedding, S 1 -relics and prograde quartz veins are transposed into parallelism with S 2 .Pressure solution and intracrystalline plasticity are regarded as the principal deformation mechanisms (Richter et al., 2007).A stretching lineation L 2 marked by elongated quartz grains is associated with the S 2 -foliation that generally plunges perpendicular to the strike of the foliation planes in the E-W direction (Fig. 4) except for a zone north of Lago Vichuquén, where it prominantly deviates to a N-S-direction.The S 2 -transposition foliation becomes continuously more closely spaced in the transition zone between the Western and Eastern Series, where F 2 -folds continuously become rootless or destroyed.Tightening of F 2 is associated with rotation of F 2 axial planes and S 2 into a shallowly east-dipping position.F 2 -fold axes are consistently subparallel to the stretching lineation L 2 (Fig. 4).There is no evidence that this parallelism was caused by rotation during noncoaxial deformation.Where the S 2 -transposition foliation is locally rotated away from its original subhorizontal position within the metapsammopelites, a further crenulation foliation S 3 is formed that may locally become the penetrative foliation (Fig. 3D).The corresponding stretching lineation L 3 is subparallel or at a small angle to L 2 (Fig. 4).The penetrative transposition foliations S 2 and S 3 are subhorizontal to moderately inclined south of 35°S, but generally steeply dipping north of 35°S (Fig. 4).
During both deformation episodes omnipresent mineral phases like phengite, chlorite, epidote and Na-Ca-amphibole depict a preferred orientation that indicate peak metamorphic PT conditions at 350-420°C, 7-11 kbar (Willner, 2005).However, subsequently all fabrics recrystallized statically with retrograde growth at the rims of these minerals as evidenced, for instance by mimetic growth of oriented minerals in F 2 -crenulation fold hinges (Willner, 2005).This proves that the penetrative fabrics were formed at maximum depth and during the earliest stage of exhumation.Post-D 2 quartz veins of a second generation crosscut the S 2 -foliation and were deformed during the D 3 episode.A third generation of quartz veins developed as tension gashes (Fig. 3E) at the end of crystal growth and recrystallisation under static conditions during exhumation of the Western Series rocks.The tension gashes formed perpendicular to the stretching lineation and are the locally last deformation structure formed under the deformation fi eld related to the accretion process.

Processes at the end of accretion
Three concomitant features mark the end of accretion and a major tectonic change.Upper Triassic sediments were unconformably deposited onto the Eastern Series and the late Paleozoic batholith (Thiele and Morel, 1981).According to Charrier (1979) tectonic development during the late Triassic and Jurassic was controlled by a simple morphology of locally subsiding basins and broad rising elevations that formed a NNW-SSEtrending en echelon pattern oblique to the present margin.The oldest zircon FT ages in the basement at 220 Ma (Willner et al., 2005) coincide with the time of deposition of the oldest overlying clastic sediments (Corvalán, 1976).However, cooling under 280°C is regionally remarkably uniform at ~200 Ma coinciding with a pronounced drop of the exhumation rate from 0.2-0.6 mm/yr to 0.04-0.06mm/yr as evidenced by fi ssion track data of zircon and apatite between 232 and 98 Ma (Willner et al., 2005).Similar tectonic conditions with a long-term crustal stability also prevailed in southcentral Chile (Glodny et al., 2006) At 224 Ma the granite of Constitución intruded the Western Series (Willner et al., 2005).This is an indication for the end of accretion, because it requires decoupling of the subducting and overriding plate.Similar granites have been described by 3. Overprinting of the original system 3.1.Variation of fi nite strain data: a tool for small scale resolution To estimate fi nite strain variation in rocks of the Western and Eastern Series, R f /Φ analyses in metagreywackes and X-ray texture goniometry (XTG) on phyllosilicate-rich rocks were carried out.For comparison of the results of both methods the conventional octahedral shear strain γ oct was determined.Both methods and their application are described in detail in Richter et al. (2007).Along the Río Maule traverse in the southern part of the study area, these authors documented a continuous structural transition between the Eastern and Wes-tern Series caused by continuous variation of the strain parameters.This fi nding was interpreted to refl ect a gradual change of the mode of accretion.For this study we apply this approach also to the northern part of the study area (Fig. 5A) to get further hints for its partly discontinuous structure.The combination of these data with those in Richter et al. (2007) is given in fi gure 5B.In fi gures 5A and 5B the conventional octahedral shear strain, γ oct., has been plotted and contoured.
Phyllosilicate-rich rocks and metagreywackes show no differences in strain symmetry.The rather smooth westward increase in strain magnitude suggests a continuous transition from the Eastern into the Western Series for the entire southern part of the area south of Lago Vichuquén.The magnitude of strain progressively and gradually rises from the eastern part of the Eastern Series (γ oct.as low as 0.56) towards the Western Series at the coast (γ oct.values up to 1.61).Contours of γ oct.values are subparallel to the boundary between both units and the mean strike of the predominant foliation planes.A considerable change from a general NE-SW to a NNW-SSE direction again becomes conspicuous.A slight break in the strain distribution pattern by the Pichilemu-Vichuquén Fault in the northern part of the study area becomes apparent that can now be shown for the entire length of the fault between Pichilemu and Lago Vichuquén.However, west of the fault the number of data do not allow a good resolution.
The extensional direction derived from the strain analyses fi ts well in all cases with the stretching lineation measured in the fi eld.Orientation of the stretching lineation L 2 does not change along traverses from the Eastern into the Western Series.The sense of local shear parallel to the stretching lineation within the S 2 -foliation (derived from S-C relationships, σclasts and shear bands) is consist ently top-to WSW in the Eastern Series in the entire study area.In contrast in the Western Series the shear sense is alternating top-to west and top-to east with the exception of the anomalous area north of Lago Vichuquén with N-S transport directions (domain B in Fig. 4).

Cataclastic processes during destruction of the accretionary wedge
The most striking characteristic of the study area is a prominent contrast in the structural setting of the Western Series north and south of 35°S which refl ects different intensity of the deformational overprint after the end of accretion.Whereas in the entire Eastern and Western Series south of 35°S the accretionary structures largely occur in their original position, most effects of post-accretionary deformation are restricted to the Western Series north of 35°S.Deformational processes which caused these late deformational structures severely displaced the original orientation of the accretionary structures.Cataclastic deformation structures, observed only in the Western Series of the northwestern part of the study area, involve 1. megascale tight subvertical folds of the predominant foliation; 2. local subvertical kink bands in more pelitic schists; 3. local cataclastic strike-slip shear zones, and 4. megascale open folds with steep axes.For better resolution we distinguish two domains within the Western Series north of 35°S (Fig. 4): Domain A is mainly characterised by a steep dip of the principal foliation that originally formed horizontally as it is preserved in the southern study area.Domain B is also characterised by this subvertical predominant foliation, but additionally by a N-S-orientation of the principal stretching lineation, which deviates markedly from its normal E-W trend (see above).
South of 35°S the dominant transposition foliation in the Western Series is subhorizontal (Richter et al., 2007).This can also be observed in the entire western part of the Eastern Series where S 2 becomes prominent (Fig. 4, areas 1, 11) as well as at Punta Lobos in the Western Series (Fig. 4, area 3).Local undulation of dip in opposite directions can be explained by open folding.Such a megascopic open fold is prominently exposed at Punta Lobos (Fig. 6).Widespread late post-accretionary upright open folding of the similarly predominantly subhorizontal foliation within the Western Series was widely recognized in south-central Chile at 40°S by Martin et al. (1999).These authors relate this phenomenon to an episode of post-Permian transpression.Yet in domains A and B the general dip of the penetrative foliation planes is mostly subvertical (Fig. 4; areas 2, 4, 5, 7, 8, 9).This may be explained by tightening of these late large scale folds with subhorizontal fold axes.Prominent exposed megascopic examples where such megafolds may be reconstructed by regional mapping of metabasite intercalations are tight megafolds southwest of Pichilemu and at Punta Sirena (Figs. 6, 7).The distribution of foliation poles in stereoplots (Fig. 4, areas 2, 5 and 9) also refl ects such tight F 4 -megafolds.
Furthermore, kink bands are randomly observed in more pelitic lithologies, but only within domains A and B. They are invariably steep (65°-80° dip; Fig. 4) and mostly conjugate.The predominant foliation within these up to 5 cm thick kink bands is always rotated antithetically with respect to the dip of the kink bands.This means that they formed during episodes of subhorizontal shortening.However, the distribution of their strike (Fig. 4) indicates that they were partly rotated into an E-W-strike by folding around steep axes by a further deformation event.
Both the tight upright folding of the transposition planes and the kink bands require local subhorizontal shortening incompatible with the deformation fi eld during basal accretion, where the last deformation structures after exhumation to upper crustal levels are represented by tension gashes.Unfortunately good outcrops only exist along the coast and it was so far impossible to fi nd similar semi-brittle shear zones inland for reconstruction of a major strike-slip system.However, prominent folds with steep axes that are presumably related to strike slip zones can be observed or reconstructed indirectly.A large-scale fold with a steep axis can be observed along the road east of Cahuil.In this zone mapping of the greenschist (Fig. 6, 7) reveals refolding of the subvertically dipping predominant foliation.Also the megafold at Punta Sirena (Fig. 6) has an unusual E-W strike presumably caused by rotation during refolding around steep axes.Furthermore, the stretching lineation L 2 is rotated from a regional NE-SW direction, as observed in the Eastern Series, into directions around E-W.The entire eastern fl ank of the Cahuil megafold was later overthrust by the Eastern Series along the Pichilemu-Vichuquén Fault (Figs. 4, 6, 7).
In the same way the narrow zone of domain B with the unusual N-S strike of the stretching lineation L 2 (Fig. 4, 5, 6) can be explained by localized rotation of the subvertical predominant foliation planes into parallelism with a major strike-slip zone.Further east these structures are obliterated by late brittle thrusting of the Eastern Series over the Western Series.
The described cataclastic structures may be best explained as being related to a megascale left-lateral strike-slip system of presumable N-S-strike (Fig. 6C) during a strike-slip episode with a subhorizontal maximum shortening axis oriented NW-SE.Early formed tight upright folds, kink bands and also minor shear zones were rotated into E-W and N-S directions, respectively.Both fl anks of the megashear zone are not exposed, but might be submerged offshore or buried beneath the Pichilemu-Vichuquén Fault (see below).

The Pichilemu-Vichuquén Fault: brittle destruction processes
The Pichilemu-Vichuquén Fault (PVF) in the northern part of the study area (Figs. 1, 4-8) was fi rst shown by Hervé et al. (1984) as a prominent brittle structure which cuts isogrades produced during the high temperature metamorphic overprint of the Eastern Series (Fig. 1) and which separates biotite-bearing schists of the Eastern Series from biotite-free phyllites of the Western Series.Earlier, Ernst (1975) referred to the PVF as the 'Coast Range Suture' prolonging this tectonic contact between both series along their entire boundary in central Chile and thus implying plate tectonic dimensions.However, the distribution of strain data (Fig. 5) shows that the fault only continues southward to Lago Vichuquén as it was already mapped by Moraga (1981).Hitting the contact with the late Paleozoic batholith, the PVF splits into three branches at its southern end: The eastern branch cuts into Triassic and Jurassic cover sediments showing a slight apparent left-lateral separation (Figs 1, 8).The intermediate branch cuts off the lowermost Triassic cover sediments and places Jurassic sedimentary rocks in direct contact with the metamorphic basement.The western branch, at the boundary between both series, cuts the strike of the predominant foliation, the megascale deformation pattern in the Western Series, produced during Jurassic times (Fig. 6), as well as the strain contour lines (Fig. 5).Left-lateral strike slip also becomes apparent by the displacement of the sillimanite-in isograde (Fig. 1) and possibly the strain contours (Fig. 5), although resolution of the latter is low.On the other hand, the PVF appears not to be fully continuous along its length, but presumably splits into two separate faults both starting within the Los Ciruelos granite pluton north of the village of Cahuil.Although boundaries of the pluton are diffi cult to map due to poor outcrop conditions, no major displacement by the fault becomes apparent.Cataclastic fabrics in the prolongation of the fault within the granite body suggests that it was affected by faulting (see below).The prominent difference of the strike of the PVF rather suggests two separate branches starting at the pluton.Parallel to the northern branch a repetition of the fault is observed further W, which cuts the eastern boundary of the Pichilemu granite pluton near the town of Pichilemu (Figs. 1, 6-8).
To obtain further information about the nature of the fault, we carried out a fault-slip analysis at six localities along the PVF (Fig. 8).The cataclastic rocks in the vicinity of the main trace of the Pichilemu-Vichuquén fault have a rubbly to fragmental appearance and show numerous mesoscopic brittle faults.The fault planes are characterized by anastomosing clayey gouge layers with thin zones of cataclasite, breccia and hematite-clay-coated fractured rock.Bleaching and alteration of intact rock occurs in the vicinity of faults.Weakly oriented phacoidshaped tectonic slivers of country rock within the fault zone are in the centimetre to decimetre scale.The fault-surfaces contain Riedel shears which caused lunate and crescentric structures at their intersections with the fault plane.In sections parallel to the striation and perpendicular to the fault plane the Riedel surfaces are characterized by fi ne seams of greyish-brown material.In addition to the Riedel planes, linear fi bre growth features formed in dilatant areas of sliding fault blocks and extension-rack arrays occur.
To evaluate the kinematics of the PVF the orientation of primary and secondary fault and foliation planes, plunging directions of striations, and the sense of relative displacement on these planes were mapped in order to determine principal strain axes (Marrett and Allmendinger, 1990).The scale of observation was usually held small (i.e., outcrop scale).The direction and sense of shear on these surfaces can be deduced from the orientation of associated fi bres and Riedel shear planes.The displacement of the measured fault planes was generally in the centimetre to decimetre range.The main argument for relating these minor faults to the kinematic evolution of the Pichilemu-Vichuquén fault is their spatial relationship to the trace of the main fault plane.Away from the main trace of the Pichilemu-Vichuquén fault usually no minor fault planes were observed.Hence the increasing number of secondary faults in the vicinity of the Pichilemu-Vichuquén fault is thought to imply a genetic link between the minor faults and the PVF.
A simple method has been used to determine the principal strain axes from these crosscutting faults and associated slickensides (program 'Fault Kinematics' written by R. Allmendinger).This method graphically constructs the principal incremental shortening and extension axes for a given population of faults.Each pair of axes lies in the movement plane of the fault (i.e., a plane perpendicular to the fault plane that contains the unit vector parallel to the direction of accumulated slip, and the normal vector to the fault plane).Furthermore, each pair of axes makes angles of 45° with each of both vectors.For distinguishing between the shortening and extension axes, information on the relative sense of slip is needed.Since the method converts the measurements into a fault-plane solution, the kinematic axes of a fault only portray a different and visually more convenient presentation of the original data.
Bingham distribution statistics for axial data were used to optimize clusters of kinematic axes of a fault array (Mardia, 1972).Riedel structures, fi bres and extension cracks indicate consistent up-dip displacements on the minor fault planes.The consistently subhorizontal shortening (Z) axes and subvertical extension (X) axes indicate E-W to SE-NW reverse faulting.Hence the resulting deformation fi eld completely differs from the one during formation of the accretionary structures.At locations 1, 2, 3, and 5 the Z axes are subnormal to the trace of the fault, whereas their oblique orientations at locations 4 and 6 might be due to some local rotation during strike-slip movements.Although the main fault is nowhere clearly exposed we infer that the eastern block (Eastern Series, the batholith, and the oldest cover sediments) is thrustover the western block along a steeply E-dipping main fault plane (Fig. 7).Although the maximum vertical displacement can only be estimated not to have exceeded 3-4 km, the PVF is an expression of a regionally important contractional episode.

The accretionary system
The two main structural units of the metamorphic basement of the Chilean Coastal Cordillera that may be traced over entire central Chile (Hervé, 1988) are fundamental features of a deep-seated accretionary system.Following the flow-field concept of Freehan and Brandon (1999) and Ring et al. (1999) and analogue experiments of Glodny et al. (2005), Richter et al. (2007) interpreted the structural characteristics of the Eastern Series rocks as frontally accreted sediments.Frontal accretion represents one of the two principal end members of accretion modes which is characterized by subhorizontal shortening and largely subhorizontal initial particle fl ow.This original architecture is preserved north and south of 35°S.Changes of the strike of F 1 -fold axes and S 1 -foliation are interpreted as an original change of the strike of the accretionary system.
Within the Western Series, the origin of the subhorizontal S 2 -foliation during ductile thinning at depth is thought to result from underplating of material to the base of the wedge.The observed structures formed by basal accretion following the fl ow-fi eld concept of Freehan and Brandon (1999) and Ring et al. (1999) and analogue experiments of Glodny et al. (2005).Basal accretion is characterised by subvertical shortening and a mainly subvertical particle fl ow path.Because both lithotypes of the Western Series have a similar structural inventory, tectonic intermingling of oceanic and continental rocks must have occurred during an early stage of subduction prior to basal accretion.
According to Willner et al. (2005) maximum burial of the accreted rocks of the Western Series was reached at 292-305 Ma at the same time when the bulk of the arc batholith intruded into the Eastern Series at around 300 Ma causing high temperature/ low pressure metamorphism that statically over printed the high level accretionary prism architecture.Thus a truly contemporaneous paired metamorphic belt originated in Central Chile.The change of the mode of accretion must have occurred before that time at ~305 Ma.The beginning of accretion processes was slightly before 320 Ma as indicated by exotic local garnet-bearing schist in the Western Series at Punta Sirena exhibiting maximal PT-conditions of 10-15 kbar, >600°C, an anticlockwise PT-path and the oldest (retrograde) metamorphic ages (Ar/Ar; phengite) of 320 Ma (Willner, 2005;Willner et al., 2005).
An originally continuous transition between the structurally overlying frontally accreted rocks (similar to the Eastern Series) and the structurally underlying basally accreted rocks (similar to the Western Series) can be assumed as a fundamental feature of the metamorphic basement along the entire Chilean coastal Cordillera between 26° and 55°S irrespective of the depositional or metamorphic age of the accretionary systems.However, this transition was clearly demonstrated at two localities only: in the Constitución area (35°30'S) by Richter et al. (2007) and in the Chonos Archipelago (43°S) by Davidson et al. (1987).All other basement complexes are isolated and exhibit either structures typical of frontal accretion or basal accretion.For instance, structures similar to those of the Eastern Series were observed in Southern Chile e.g., in the Eastern Andean Metamorphic Complex or the Madre de Dios Accretionary Complex, structures equivalent to those of the Western Series in the Diego de Almagro Metamorphic Complex (Hervé et al., 2000).Hence it is of importance to unravel structures related to postaccretional processes that isolated these partial metamorphic complexes.Some of these structures are overprinting the accretionary complexes themselves during their exhumation.
Based on the consistent orientation of underplating related stretching lineations in the lowermost parts of the Eastern Series and within the Western series, as well as associated shear-sense indicators in the lower part of the accretionary wedge (Fig. 5B), we propose that there was a relatively high degree of coupling between the subducting and the overriding plate.Due to the relatively high degree of plate coupling the direction of subduction of the oceanic plate may be deduced from the mean orientation of the stretching lineations.Subduction was mainly from westerly directions with neglegible oblique components, from WNW south of 35°S to WSW north of 35°S.It is notable that the strike (1) of the F 1folds in the Eastern Series, (2) of the transition zone between the series, (3) of the penetrative foliation, and (4) of the magmatic arc is perpendicular to these directions.These relationships can be regarded as primary in the entire study area.

The Jurassic post-accretion development
The orientation of the subhorizontal maximum shortening axis of the cataclastic phenomena in the Western Series north of 35°S strongly deviates from that during accretion in the Western Series.Therefore the cataclastic phenomena are unrelated to the accretion process and are thought to relate to a deformation episode which severely destroyed the original structure of the accretionary system.However, they are confi ned to the narrow zones of domains A and B only.
Some considerations concerning the age of the cataclastic phenomena are possible: the temperature during formation of the cataclastic structures has to be below 280°C, which is the annealing temperature of fi ssion tracks in zircon.Cooling of the rocks of the study area below the zircon annealing temperature was dated at 206-218 Ma in the study area (Willner et al., 2005) giving an upper age limit for the cataclastic deformation.On the other hand, the local semi-brittle shear zones are cut by undeformed basalt and rhyolite dykes.A rhyolite dyke in the study area was dated at 138 Ma (Willner et al., 2005) consistent with Jurassic to Lower Cretaceous ages of the widespread bimodal magmatism in the forearc further north (Irvine et al., 1988).Hence, the cataclastic deformation episode is largely confi ned to Jurassic times.Similar semi-brittle strike-slip zones of considerable size cross-cut by Jurassic dykes are widely observed further north along the coast of north-central Chile (Rebolledo and Charrier, 1997;Irvine et al., 1988).Summarizing the cataclastic structures may be best explained as related to a left-lateral strike-slip system (Fig. 6C) formed during a Jurassic timerestricted episode of strike-slip movement.Most likely this represents the southernmost occurrence of similar phenomena widespread in northern Chile during Jurassic times.Left-lateral strike-slip during Jurassic times, causing stretching of the outermost continental margin, becomes prominent in northern Chile known as the Atacama fault system (e.g., Scheuber et al., 1994;Scheuber and González, 1999).On the other hand, the strike-slip system may have commenced during late Triassic times and thus could be related to the marked tectonic change throughout central Chile, which involved basin formation in the former accretionary wedge and granitoid intrusion into the Western Series.Yet granitoid intrusion into the Western Series involves plate decoupling.A marked change in the tectonic setting from a convergent to an extensional regime with a characteristic accompanying magma productivity (as observed in Chile) was recently studied by numerical modeling (Gorczyk et al., 2007).Their results show that the change from a high degree of plate coupling during accretion to plate decoupling during slab retreat was mainly controlled by the convergence rate.

Cretaceous shortening episode
The PVF represents a late brittle reverse fault system with a minor left lateral component.It must be younger than Jurassic because it cuts the Jurassic sediments.Willner et al. (2005) associated this young shortening event with an episodic acceleration of cooling and hence exhumation as deduced from the interpretation of apatite fi ssion track ages (80-113 Ma) and modeling of track lengths.Bravo Espinosa (2001) proposed the closure of Mesozoic intra-arc basins in the area by mid-Cretaceous times.Arancibia (2004) dated ductile reverse shearing along the prominent Silla del Gobernador shear zone further north (32°S) at 109±11 Ma (Ar/Ar; white mica grains).Due to widespread apatite fi ssion track ages in north-central Chile around ~100 Ma a mid-Cretaceous shortening event, responsible for the closure of Mesozoic basins and associated with an increased erosion rate, is now widely accepted (e.g., Parada et al., 2005).Thus, the mid-Cretaceous shortening episode affected a wider part of the forearc causing closure of the Mesozoic basins and acceleration of exhumation (Willner et al., 2005;Belmar et al., 2004).Accelerated exhumation must be due to enhanced erosion, which was probably caused by crustal thickening and the associated development of topography.

The possible role of subduction erosion
'Subduction erosion' as a mechanism of subduction of continental crust is frequently proposed in northern, but also in central Chile to explain missing structures beyond the present coastline (e.g., Glodny et al., 2006;Kay et al., 2005).Only at 35°30´S the entire width of the exposed late Paleozoic accretionary wedge onshore and its extension to about 50 km offshore allows to reconstruct a coherent model for the convergent margin during late Paleozoic times (Willner, 2005).This contrasts to the suggestion of Glodny et al. (2006) that north of the Lanalhue Fault Zone at 38°S (Fig. 1, inset map) the accretionary wedge was shortened by subduction erosion during late Paleozoic times, because the width of the accretionary wedge doubles south of the fault.Yet no structures can be related to this proposed omission of the continental margin and accretion was defi nitely continuing in the north during the proposed time of the subduction erosion during Permian times (Willner et al., 2005).Furthermore, magmatic activity in the late Paleozoic arc continued.We argue that the Lanalhue Fault Zone can be much better explained as a transform fault between two segments of contrasting slab geometry during formation of the Late Paleozoic accretionary systems.
However, north of 35°S the late Paleozoic forearc becomes narrower and nearly disappears between 34° and 30°S, where the late Paleozoic arc is directly exposed at the coast.Kay et al. (2005) pointed to this effect and proposed a latest Tertiary time for the observed omission of the forearc to explain eastward migration of the magmatic arc at that time, however, without presenting corresponding structures in the forearc.In contrast, we propose that this effect must rather be related to the two episodes of strike-slip faulting and shortening, because distribution of their structures coincides with the area of presumed cut-off of the fore-arc.The Jurassic episode of strike-slip faulting did apparently not involve any crustal thickening.On the one hand, considerable acceleration of exhumation was proposed by Willner et al. (2005) for the study area for the mid-Cretaceous event suggesting some crustal thickening.This shortening episode seems to have affected the entire forearc at least up to 30°S (Arancibia, 2004;Parada et al., 2005).On the other hand, no marked crustal thickening evidently occurred in the forearc since latest Tertiary times.The last marine sediments overlying the metamorphic basement in the study area at its northernmost end, the Miocene Navidad Formation, was merely uplifted by ~100-200 m at the coast (Encinas et al., 2006).Evidently deformation and shortening during the latest Andean orogeny was rather confi ned to the magmatic arc (Kay et al., 2005) and the eastern foreland leaving the forearc in the Coastal Cordillera as a rigid block or ram.Hence it is unlikely that subduction erosion in the fore-arc occurred as late as the latest Andean orogeny.

Summary
The basement area in the Chilean Coastal Cordillera between Pichilemu and Constitución represents an ideal type region, where the development of the late Paleozoic accretionary wedge can be studied.Structures related to a strike-slip and a later shortening event in the forearc during Mesozoic times are exposed in the northwestern part of the area providing clues to the type of destruction of the accretionary system.These destructive events made any reconstructions of the original late Paleozoic accretionary wedges so far diffi cult along the coast of north-central and northern Chile north of 35°S.
FIG. 2. A. subvertical F 1 chevron folds strongly overprinted by subhorizontal F 2 -folds at the coast of Tanumé (34°13'S) in the Eastern Series; B. close view of subhorizontal F 2 -folds near A; C. late-tectonic andalusite porphyroblasts stretched parallel to the regional stretching lineation with quartz-fi lled tension gashes (see arrows) at locality B; D. relic pillow lava fabric at Infi ernillo (34°24´S) in greenschists of the Western Series; E. relic hyaloclastite fabric near locality D.
FIG. 5. A. Map showing values of the conventional octahedral shear strain, γ oct., as obtained from R f /Φ and XTG measurements and contouring of their distribution for the northern part of the study area.Note the offset of contours by the Pichilemu-Vichuquén Fault; B. Map compilation including the data of Richter et al. (2007) showing: 1. the contouring of the conventional octahedral shear strain, γ oct., for the entire study area and 2. observed transport directions of local noncoaxial structures (arrows).
FIG. 6. A. Detail map of domain A showing mesoscopic structures using exposed greenschist as tracer horizons; B. Detail map of the area at Infi ernillo west of Pichilemu affected by left-lateral cataclastic strike-slip zones; cross sections AB and CD are shown in Fig. 7; C. Schematic sketch of presumed mega-scale left-lateral strike-slip system active during formation of the Jurassic cataclastic structures.Earlier formed tight megascopic folds were refolded.

FIG. 8 .
FIG. 8. Detail map of the Pichilemu-Vichuquén Fault with kinematic analyses of minor faults along the principal fault.Lower hemissphere, equal area stereograms show the fault surfaces and striae as well as the calculated principal strain axes.Note the splitting of the fault into three branches at its southern termination.