Extensional tectonics in Mt Parnon (Peloponnesus, Greece)
Peloponnesus in the south-western part of the Aegean is formed by a heterogeneous pile of alpine thrust sheets that was reworked by normal faulting from Upper Miocene to recent times. Upper Miocene–Lower Pliocene extension in Mt Parnon was accommodated by several mappable brittle detachment faults that exhibit a top-to-the-NE-ENE sense of shear. The hanging wall of the detachments comprises a number of highly tilted fault blocks containing abundant evidence of intense internal deformation by normal faulting and layer-parallel shearing contemporaneous with faulting. These fault blocks are remnants of a cohesive extensional block that slipped to the NE-ENE and broke up along high-angle normal faults that sole into or are cut by the detachments. The largest part of this block is located at the eastern edge of the metamorphic core forming the hanging wall of East Parnon high-angle normal fault that excised part of the aforementioned detachments. The lowermost metamorphic Unit of the nappe-pile does not seem to be affected by the previous extensional episode. Upper plate reconstruction shows that various units of the nappe-pile were affected by high-angle normal faults that linked to detachment faults in the weaker layers. Since the Middle-Upper Pliocene further exhumation of the metamorphic rocks has resulted in the formation of high-angle normal faults overprinting Neogene extensional structures and cut the entire nappe-pile. This new fault system tilted the earlier extensional structures and produced a NE-SW coaxial deformation of Mt Parnon.
KeywordsLate orogenic extensionExtensional detachmentsLow-angle normal faultsRheologyExhumationLateral segmentation
In the last three decades, it has been documented that extension is a common feature in orogenic belts and its contribution to the syn-convergence exhumation of the high-pressure metamorphic rocks is crucial (Platt and Vissers 1989; Jolivet and Goffé 2000). The extensional evolution of those domains is characterized by different generation of extensional structures such as early low-angle normal faults or detachments and late high-angle normal faults that are usually rheologically controlled (Davis 1980; Wernicke and Burchfield 1982; Platt 1986; Lister and Davis 1989; Carmignani and Kligfield 1990; Jolivet et al. 1994; Friedmann and Burbank 1995; Martínez-Martínez and Azanon 2002). In active margins where the trenches move towards the subducting slab, large-scale extension produces stretching of the previously accreted units and causes high-pressure rocks to exhume from depths of 70 km and even more during continuous subduction (Brun and Faccenna 2008). The Hellenic Arc in eastern Mediterranean is one of the best examples where exhumation of the deeper metamorphic rocks has taken place above a retreating subduction zone; this was accommodated by the formation of different generations of extensional faults (e.g. Lister et al. 1984; Gautier and Brun 1993; Jolivet et al. 1994; Kilias et al. 1994; Thomson et al. 1999; Avigad et al. 2001; Ring et al. 2003; Brun and Sokoutis 2007; Jolivet et al. 2008).
The Hellenic Arc forms an arcuate orogen above the subducting African slab. Tomographic models suggest a continuous northward subduction of the African plate below the southern margin of the Eurasia from the Cretaceous and the entire Cenozoic (Wortel and Spakman 2000). Syn-convergence exhumation took place within the subduction zone itself, during the Eocene in the Cyclades and during the Upper Oligocene and Early Miocene in the External Hellenides (Jolivet and Brun 2010; Ring et al. 2010). Crete, the central segment of the External Hellenides, formed during post-orogenic extension and the final exhumation of the Cyclades that took place in backarc region and in a warmer regime (Jolivet et al. 2010). The Cretan Detachment is described as a major extensional north-dipping low-angle normal fault (Jolivet et al. 1994; Kilias et al. 1994; Fassulas et al. 1994; Thomson et al. 1998) that separates the external high-pressure rocks in its footwall from the non-metamorphosed rocks in the hanging wall (Ring and Reischmann 2002). Across this detachment, a pressure gap of ~10 kbar suggests that 20–30 km has been removed through symmetrical extension (Kilias et al. 1994) or by a dominantly northward sense of shear (Jolivet et al. 1996; Thomson et al. 1998).
Several researchers have described detachments in Peloponnesus and Kythera Island (Skourtsos et al. 2001; Trotet et al. 2006; Papanikolaou and Royden 2007; Marsellos and Kidd 2008; Jolivet et al. 2010; Marsellos et al. 2010) and identified them as the north-western extension of the Cretan Detachment. In this paper, we present the results of detailed mapping in Mt Parnon, which provides new structural and kinematic data on the geometry of the extensional faults that affected the nappe-pile exposed in western Peloponnesus. Preliminary work in this area has shown the presence of multiple low-angle normal faults that thin the upper plate of the detachments (Skourtsos et al. 2004). We summarize the episodic normal faulting that has affected the study area since the Middle Miocene, and we discuss the tectonic relationships between the detachments and the high-angle normal faults and the control of the rheological behaviour of the rocks on the lateral and vertical development of the extensional faults. Finally, we propose a geometrical model trying to explain the lateral segmentation of the alpine tectonic Units.
The Plattenkalk Unit consists of Permian to Upper Triassic phyllites and metaconglomerates with quartzitic pebbles (Psonis 1981; Kowalczyk and Dittmar 1991; Blumör et al. 1994), named Kastania Phyllites (Manutsoglu 1990), followed by dolomitic marbles, porous quartzites and cherty marbles with ages ranging from Upper Triassic to Upper Eocene and finally a Lower Oligocene calcareous metaflysch (Psonis 1981; Thiebault 1982; Bassias and Thiebault 1985; Bassias et al. 1987; Deckert et al. 1999). Although the PK Unit in Crete shows HP-LT metamorphism as indicated by the presence of Mg-carpholite in metabauxites (Seidel et al. 1982; Theye and Seidel 1991), the main assemblages in Peloponnesus indicate greenschist-facies metamorphic conditions (Thiebault 1982; Bassias 1988; Manutsoglu 1990). It is generally accepted that the PK Unit correlates with the unmetamorphosed Ionian Unit in north-western Peloponnesus and the Ionian Islands (Fig. 1).
The PK Unit is tectonically overlain by the Phyllites-Quartzites or Arna Unit (Fig. 1c), which consists of mica schists, quartzites, metaconglomerates, marbles and metavolcanites (Skarpelis 1989; Papanikolaou and Skarpelis 1988) of Late Palaeozoic—Late Triassic age (Krahl et al. 1982, 1983). On Mts Taygetos and Parnon in SE Peloponnesus, the PQ Unit is not homogeneous but is subdivided in 3 or 4 sub-Units exhibiting different metamorphic conditions (Blumör and Kowalczyk 1993; Aleweld et al. 1994; Blumör 1998; Skourtsos 2002; Jolivet et al. 2010): (1) metaconglomerates and micaschists with Fe–Mg carpholite and chloritoid, (2) micaschists with chloritoid and glaucophanites, (3) micaschists with chloritoid and albite and (4) micaschists with glaucophane and garnet. Furthermore, slices of Late Palaeozoic basement rocks were found in Kythira Island, located between Crete and Peloponnesus (Seidel et al. 2006; Xypolias et al. 2006). These different successions are placed on top of each other by subhorizontal faults, suggesting that the present stack is post-metamorphic (Blumör and Kowalczyk 1993), while in Mt Parnon, an inverse metamorphic piling of the successions can be observed (Skourtsos 2002). The thickness of the PQ nappe shows large variations for more than 1 km in central Peloponnesus, while it is absent between the PK and the Tripolitza Units in Taygetos and Parnon.
The KG Unit was recognized by Skourtsos (2002) and Skourtsos and Lekkas (2004) in Parnon and Taygetos Mts, resting on the PQ Unit (Figs. 1c, 2). It consists of dolomites, porous quartzites and cherty marbles. Megalodon was found in the dolomitic part of the Unit in Maleas Peninsula of south-eastern Peloponnesus (Gerolymatos 1993). Its thickness is 800 m north of Gythio town in south-eastern Taygetos, where the Unit overlies micaschists with glaucophane-chloritoid assemblages of the PQ Unit and less than 100 m close to Kosmas village on Mt Parnon where the marbles of the Unit overlies garnet-glaucophane (central Mt Parnon) or glaucophane-chloritoid-bearing micaschists of the PQ Unit. Its similarity with the PK Unit suggests that it belongs to the Ionian palaeogeographic domain, in a more internal position of the PK Unit stricto-sensu (Skourtsos and Lekkas 2004) and seems to confirm the suggestion that the PQ Unit is the basement of the metamorphosed Ionian Units (Blumör and Kowaltzyk 1993; Blumör 1998).
The overlying Tripolitza Unit consists of an Upper Triassic to Upper Eocene carbonate platform, overlying a low-grade volcanosedimentary sequence of Upper Permian to Upper Triassic age, named Tyros Beds by Ktenas (1924). The thickness of the Tyros Beds is more than 1 km in south-eastern Parnon, based on drilling data (Skourtsos 2002), while the calcareous part of the Unit is almost 2,000 m (Zambetakis-Lekkas 1988). The top of Tripolitza Unit is a 500-m-thick siliclastic flysch of Oligocene age. The Pindos Unit is a pelagic sequence, composed of radiolarites, pelites and sandstones of Cenomanian age, followed by Late Cretaceous platy limestones and a Palaeocene flysch sequence (Fleury 1980). The thickness of the nappe in north-eastern Parnon is less than 1,000 m.
The Internal Unit found in relics and on Mt Parnon consists of a Palaeocene siliclastic sequence with olistholites of ages ranging from Upper Permian to upper Cretaceous (Skourtsos et al. 2002). Further to the south, the Unit comprises Late Permian and Middle Triassic limestones and an ophiolitic mélange (Danelian et al. 2000).
Extensive sedimentation took place in south Peloponnesus during the Upper Pliocene (Fig. 1b, 2); marine facies were deposited on the external southern parts, while terrestrial sediments filled the more internal areas (Piper et al. 1982). Few occurrences of cohesive conglomerates are found at 900–1,600 m altitude on Mt Parnon (Fig. 2), originating from the erosion of the upper non-metamorphic Units. Their age might be older than those of the Sparta basin (Skourtsos 2002). In the Lower Pleistocene, thick alluvial fans formed along the eastern margins of Vrontamas and Sparta basins.
Based on the age of flysch deposits, it is possible to reconstruct the timing of stacking and thrust migration towards the External Hellenides. The Pindos thrust that is represented by the emplacement of basinal radiolarites and turbitic sequence of Cenomanian age on the Tripolitza flysch is post Oligocene as documented by the age of the Tripolitza flysch. The PK Unit was underthrusted after the Lower Oligocene, as indicated by the age of its calcareous metaflysch (Thiebault 1982). A key element of this contractional deformation is the emplacement of the HP-LT PQ Unit between the greenschist PK Unit and the very low-grade Tyros Beds. Radiometric data (K–Ar and Ar–Ar) suggest that the largest part of exhumation of the PQ Unit took place during the lower Miocene (19–16 Ma, Thomson et al. 1998) and it seems that its cooling was the result of underthrusting of the PK Unit (Trotet et al. 2006). This process brought the PQ Unit to upper crustal levels (Skourtsos et al. 2004). Shortly after, the compressional front migrated to more external parts of the western Peloponnesus (Jacobshagen et al. 1978) and the Ionian Islands. At the same time, central and western Peloponnesus were affected by extensional faulting, the onset of which is placed sometime in the Upper Miocene (Mariolakos 1976).
Extensional systems on Mt Parnon
Early extensional structures
A significant number of low-angle normal faults are present on Mt Parnon (Figs. 3, 4). On a large scale, the geometry of the detachments is not planar but they are characterized by slightly curved surfaces, which truncate each other or extend subparallel for long distances and are accompanied by thick fault rocks. Moving from east to west, the detachments dip direction changes from a north-eastwards dip in the eastern sector to a south-westwards dip in the western part (Fig. 3). Local differences in the dip value are due either to deformation by more recent faults or to local culminations of the weak layers of the nappe stack, which show a convex-upward turtle-back morphology. All the faults have operated in upper crustal levels (≤10 km) as shown by the nature of the fault rocks.
The Tripolitza carbonates and the Tyros Beds have been affected by three generations of normal faults: the older ones run parallel to the detachments and are dissected by younger ones with steep dips (Fig. 6c). They are NW-SE dip-slip faults and NE-SW dextral or sinistral faults. Most of the NW-SE-directed faults that dip to the NE are accompanied by NW-SE to NNE-SSW antithetic faults (Fig. 6c). The NE-SW faults seem to be lateral ramps, distinguishing sectors of different rates of deformation. The A2 detachment fault and accompanying extensional structures have caused the thinning of the Tyros Beds and the Tripolitza limestones, which were brought closer to the underlying PQ Unit (Figs. 3, 4, 5).
On south-eastern Mt Parnon and along the coastline, where the thickness of the Tyros Beds exceeds 1,000 m, the early contractional structures are dominant. These are thrusts, reverse faults and south-west verging folds. Sometimes an axial plane cleavage has developed locally. Extensional deformation in these places is localized in narrow zones. Westwards, where the Tyros Beds are thinner, the extensional shear zones dominate, destroying any stratigraphic continuity and overprinting older compressional structures. In the field, this can be seen as fast alternation of various lithologies, laterally as well as vertically, separated by extensional shear zones (Fig. 6e). Slip lines measured on shear planes combined with other kinematic indicators demonstrate a “top to the NE” sense of shear (Fig. 6e).
Stratigraphic data also corroborate the existence of the extensional A2 detachment. A characteristic result of its action was the superimposition of younger parts of the Tripolitza Unit on older ones. Such an example is the emplacement of the Oligocene flysch or the Tertiary limestones of the Unit on the Tyros Beds (Figs. 3, 5b). This mechanism has already been described as “the younger over the older” in several extensional fields of the American Cordillera.
Outcrops of the A3 detachment fault are very limited, compared to A2, due to the limited number of outcrops of KG Unit, which constitutes its footwall (Figs. 2, 4). Detachment fault A3 is characterized by a relatively narrow shear zone and a characteristic fault surface usually covered with oxides. At Kosmas, it has a low NE dip, while in western Mt Parnon, it dips to the SW or NW. The footwall of this fault is characterized by fault breccias saturated in several places by various solutions. Furthermore, marbles appear strongly fractured by high-angle faults of different strikes. Many of those do not cut the two detachments that bound the KG Unit. Slip lines measured on the oxidized surfaces of the detachment fault indicate a predominant ENE-WSE extension which in combination with other kinematic indicators indicates a “top to the ENE” sense of shear. This direction is consistent with the mean slip of the faults observed inside the KG Unit.
Detachment fault A4 cuts the former twice, and as the A4 is also the basal tectonic contact of the marbles of KG Unit (Fig. 6c), the latter seems to be a tectonic horse or an “incisement nappe” according to the model presented by Lister and Davis (1989) for the multiple detachments. A4 detachment fault extends significantly as it covers an area of about 100 km2 on western Parnon and also forms the boundary between the PQ Unit and the overlying Tripolitza and Pindos Units (Fig. 4). The detachment does not maintain the same geometry everywhere: at Kosmas, it dips 20–40° to the NE, while at Gaitanorrahi, it dips 25–30° to the SSE. On western Parnon, the detachment dips to the WSW-SW (Fig. 6d). When the hanging wall of the detachment consists of carbonate rocks, then the detachment has a very distinct fault surface, and the overlying limestones appear strongly brecciated over 150–200 m (Fig. 6d). When both the hanging wall and the footwall of the detachment consist of less component rocks, then a thick cataclastic zone is observed. In PQ Unit, extensional shear zones deform the oldest rock fabric. The latter is a relic S1 continuous schistosity observed at the microscopic scale and defined by elongated quartz grains. It is overprinted by a successive folding event D2 of meso-scale folds associated with a pervasive S2 foliation. Furthermore, a spaced foliation S3 classified as a smooth crenulation cleavage is axial plane to later D3 folds facing mainly to the WSW. Extensional shear zones post-date the development of the D3 folds and strike N120° to N180° and dip towards ENE in the eastern side of Mt Parnon and towards WSW in the western side. Decimetre S–C and S–C′ structures formed within the schists and meso- to outcrop-scale Riedel shears affecting the more component quartzites and metavolcanics. These structures are useful kinematic indicators. On the whole, these structures indicate “top to the NE-E” sense of shear in both sides of Mt Parnon (Fig. 4). In places, these extensional shear zones are overprinted by late faults that are connected with the late high-angle faulting which affected the study area.
In central and western Parnon, the hanging wall of A2 and A4 detachments is highly fragmented and is represented by extensional klippens formed mainly by W-dipping carbonates of the Tripolitza platform (Figs. 2, 3). Although the klippens seem to form a continuous outcrop, partially eroded, they represent portions of the hanging wall, which has been tilted westwards by low-angle normal faulting. Consequently, thrust faults that emplaced Eocene Tripolitza limestones on the flysch of the Unit during orogenic building were also tilted westwards and reactivated as normal faults.
Detachment fault A5 is the lowest detachment on Mt Parnon. It appears only in the central part, with the PK Unit as its footwall (Fig. 6f–g). All other contacts bounding the PK Unit are high-angle faults. The hanging wall of the detachment fault consists of Jurassic to Cretaceous brecciated limestones of the Tripolitza Unit or PQ schists. The detachment generally dips to the NW and is characterized by striations indicating ENE shear direction (Fig. 6f). Decimetre-scale brittle to brittle-ductile S–C cataclasites within the overlying schists (Fig. 6g) and metre-scale Riedel shear fractures and assymetrical folds within the marbles are associated with this detachment. The Riedel shear fractures are observed for about six to ten metres from the detachment and farther away within the marbles. Mechanical striations on the fracture planes indicate N80° shear direction. The assymetrical and detached folds are characterized by N150° striking axes and ENE vergence. On the whole, the kinematic indicators suggest “top to the ENE” sense of shear on this detachment.
Late extensional structures
The previous extensional episode is followed by high-angle normal faulting, related to the uplift of Mt Parnon and the displacement of earlier extensional structures to different altitudes, while the extensional deformation affects the deepest PK Unit for the first time (Figs. 3, 4). These new extensional structures consist of NW-SE to N–S (more rarely NE-SW) normal faults, which either have en-echelon geometry constituting segments of a larger fault zone, or they are isolated faults with a length of several kilometres. The first group of faults consists mainly of dip-slip faults, while the last group mainly consists of oblique-slip faults (Fig. 4). All the earlier faults extended Mt Parnon in an ENE-ESW direction, normal to the strike of the orogen and a smaller area extending in a SE-NW direction, parallel to the orogen (Fig. 4).
The largest vertical displacements have taken place in the core of the mountain range where the metamorphic rocks of PK Unit can be observed at altitudes of ca. 2,000 m. The eastern margin of the PK Unit is a fault about 35 km long and strikes NNW-SSE, and consists of four segments dipping 25–50° to the NE to ESE. This is the Eastern Parnon Fault (EPF), which has caused the subsidence of its hanging wall, which corresponds to the Kynouria sector (Fig. 3). Statistical analysis of the striations on the fault surfaces showed a small dispersion in the slip direction from Β74° to Β90° and pitch 38–45° to NE-E (Fig. 8). On the contrary, the analysis of the accompanying small faults showed a much wider dispersion in the direction of slip from Β68° to Β140° with a mean direction of Β90°.
The EPF dissects and bends the early detachments, consistently with the normal displacement as best exemplified north of Platanaki (Fig. 8). In the footwall of the main fault zone, the detachments present a relatively horizontal geometry or they dip to the SW-NW. In the hanging wall, they are curved to the NE to E. Furthermore, a series of synthetic high-angle normal faults subparallel to the main fault, cause relatively small vertical displacements of the detachments (Fig. 8). The EPF appears fractured today by two dextral NE-SW faults, which are the Samaria and the Xiromouhala faults (Fig. 4). These cut the backbone of the mountain and separate the three big mountain masses of Megali Tourla in the north and of Koromilia in the south separated by Gaitanorrahi in-between (Fig. 4).
East of this zone and at some distance from it (>500 m), high-angle normal faults are found, such as those in Roussou Lakka, Kosmitiko Rema, Kounoupia and Pournaros fault (see cross sections in Fig. 3 and tectonic map in Fig. 4). The faults cause further subsidence of the Kynouria fault block and the vertical separation of the detachments, but do not cut the EPF. Instead, they are diffracted within the weaker Phyllites-Quarzites Unit and the Tyros Beds and their slip is transferred to low-angle shear zones, parallel to the EPF or with lower dips (e.g. see sections CC′ and DD′ in Fig. 3). Along the central segment of the EPF, a half-graben was formed where relics of the uppermost Pindos and Internal Units are found bounded by high-angle normal faults (Figs. 3, 4).
Some of these faults in the west of Mt Parnon control the deposition of Upper Pliocene—Lower Pleistocene continental deposits that filled the Sparta and Vrontamas post-alpine basins (Figs. 2,4). Indeed, during the lower Pleistocene, the Xiropigado fault controlled the deposition of three big alluvial fans, the material of which originated exclusively from the Plattenkalk Unit proving that the greatest part of the uplift of the core of the mountain culminated during that time.
Eastwards of the study area, the Argolikos Gulf separates the main body of Peloponnesus from Argolis peninsula (Fig. 1b). It is a significant neotectonic NW-SE trending macrostructure (Papanikolaou et al. 1988, 1994). The main rift of this basin took place during the Upper Pliocene–Lower Pleistocene and was accompanied by NW-SE, N–S and ENE-WSW-trending normal faults. The largest throws are observed along its western boundary, namely along the coasts of Kynouria, where they exceed 500 m, while the faults exhibit an en-echelon geometry and westwards backtilt of the fault blocks (Papanikolaou et al. 1988, 1994; Van Andel et al. 1993). Their greatest thickness is encountered in the centre of the basin where it reaches 500 m (Papanikolaou et al. 1988, 1994; Van Andel et al. 1993; Anastasakis et al. 2006).
The nappe stacking during the alpine orogenesis had as a result the lower PQ and PK Units to be buried at depths of 20–40 km and exhibit metamorphism in high pressure and greenschist-conditions, respectively. The exhumation of the HP/LT PQ Unit and its emplacement between the greenschist-facies PK Unit and the highly fragmented upper non-metamorphic Tripolitza and Pindos Units raises issues both on the exhumation mechanism and on the time distribution of the accompanying tectonic events. The available stratigraphic data show that the earlier extensional system cannot have operated before the Lower Miocene. This lower limit can be determined by the younger age of the flysch in the Tripolitza Unit. In western Peloponessus, the upper conglomerate beds of the flysch have been dated as Late Oligocene and possibly Lower Miocene (Fytrolakis and Antoniou 1998). There is a controversy about the age of the upper beds of the flysch in central Peloponessus, where the fossil Globigerinoides cf. Promodius was found, which characterizes the Late Aquitanian (Richter 1976). As the Pindos thrust was active until the Upper Oligocene—Lower Miocene, the upper detachment in Mt Parnon (A1) cannot have acted earlier, thus setting the operation time of the underlying detachments, as the last ones are younger based on the cross-cutting relations. The younger time limit results from the post-alpine deposits that have covered these extensional structures in the Sparta and Vrontama basins, in the west and southwest of Mt Parnon, respectively, the age of which is Upper Pliocene—Lower Pleistocene (Piper et al. 1982). This age reflects also the onset of the high-angle normal faulting. As far as the PK Unit is concerned, the stratigraphic data show that its subduction cannot have started before the Lower Oligocene, as this is the age of the upper beds of the Unit (Thiebault 1982).
More evidence is given by the available geochronological data. Seidel et al. (1982) estimated that the peak of high-pressure metamorphism took place in 23 Ma. Latest researches with 40A/39Ar from Mt Taygetos and Mt Parnon show a distribution of ages between 26 and 13 Ma and because they appear to be cooling ages, the high-pressure metamorphism must have happened prior to 26 Ma (Jolivet et al. 2010). These data also show that parts of the palaeogeographic Ionian domain, which are represented by the PQ and KG Units (more internally) and PK Unit (more externally), had started subducting under the Tripolitza Unit since the Late Oligocene. The younger of these ages show that the PQ Unit reached the depth of 10–15 km and a temperature of 300–350°C, 13 Ma in Mt Taygetos and 16 Ma in Mt Parnon (Jolivet et al. 2010). Zircon FT ages from southern Peloponnesus show that the PQ Unit reached a temperature of 240°C before 12.8–10.1 Ma (Marsellos et al. 2010), which is approximately when the PQ and PK Units must have been juxtaposed at ca. 4–5 kbar (Trotet et al. 2006). Furthermore, Apatite FT ages indicate that juxtaposition of Tripolitza Unit and PQ Unit might have already taken place in Late Miocene (6.8–9.8 Ma, Marsellos et al. 2010). As the nature of the fault rocks reveals that the low-angle normal faults must have acted in the upper 10 km of the crust, we have to accept that the lower detachments (Α3–Α5) of Mt Parnon that have affected the PQ Unit have operated since Langhian/Serravalian until probably the Early Pliocene. The exhumation rate resulting from the above is ~1–1.5 km/Ma, and the slip rate is 3 km/Ma accepting that the extensional system acted as one single detachment with a dip of about 20°. Given that the activity was distributed in several faults, the slip rate in each detachment must be significantly lower.
Geometrical evolution of the Mt Parnon extensional faults
The combination of secondary high-angle normal faults in the hanging wall of this extensional system and the detachments gave rise to a widespread segmentation of the Tripolitza nappe represented by a series of SW highly tilted fault blocks in the western part of the study area (Fig. 3). This domino-like configuration that has often been described in extensional settings around the world (Gibbs 1984) could be the result of the Tripolitza carbonate platform fragmentation due to the NE slip of a major fault block on the extensional low-angle faults (compare cross sections in Fig. 3 with the pre-extensional geometry of Fig. 9a). The largest part of this block is found today in the Kynouria sector of western Peloponnesus, forming the hanging wall of the detachments and both of them the hanging wall of the EPF (Figs. 2, 3, 9c–d). One critical element of our reconstruction is the fact that the highly extended part of the Tripolitza Unit was originally three times narrower than that of the less extended part of the Unit observed in the Kynouria sector (Fig. 9). Furthermore, the isolated klippen of the Pindos Unit found today in western and eastern Parnon (Figs. 2, 3) formed a continuous nappe resting on the Tripolitza flysch before the extension started (Fig. 9a).The present configuration of Mt Parnon is the result of the NW-SE high-angle normal faults and NE-SW dextral faults (Figs. 3, 4, 5, 6, 7, 8, 9d). These were formed after the Early Pliocene and clearly dismembered the low-angle normal faults and the Late Pliocene deposits of the extensional basins, which marked the end of the first extensional episode (Fig. 3). All the post-detachment faults have caused mainly ENE-WSE extension of Mt Parnon, which is in agreement with the regional tectonic regime in southern Peloponnesus (Angelier 1979; Lyberis et al. 1982; Lallemant et al. 1983; Hatzfeld et al. 1990). Slipping along the newly formed high-angle normal faults has resulted in earlier detachments obtaining progressively opposite dips and appearing today as low-angle thrusts. This late extensional system is thought to be related to new detachment faults, located at depths of 8–10 km. These detachments seem to control the actual deformation both of Mt Parnon and the broader area of eastern Peloponnesus but without having any superficial expression within the Parnon mountain range. It is possible that the NE-dipping East Taygetos Fault (ETF), the main marginal fault of the Sparta Basin, may actually be related to these detachments (Skourtsos 2002; Skourtsos et al. 2007). While previously Mt Parnon was in the footwall of the Parnon detachments, it is currently located in the hanging wall of this breakaway (?) fault (Fig. 9). Although these two episodes have been described as distinct phases, their durations must have some overlap, during which both systems were active. However, the older system has ceased to be active and its structures were incorporated in the fault blocks that were produced during the normal faulting of the second episode.
Comparison with previous models
Doutsos et al. (2000) described the contact between the Tripolitza and the PQ Units in northern Parnon as a décollement zone associated with west verging folds and shear planes indicating west-directed shearing. They also suggest that these compressional structures were reactivated after the Εarly Late Miocene as normal faults at the eastern margin of Parnon tectonic window, an interpretation which is different from that presented in this paper. In their view, the thinning of the upper nappes is underestimated, while in our interpretation, the low-angle normal faults have overprinted the pre-extensional structures, causing the intense thinning of upper tectonic Units and the partial omission of the nappe-pile.
Trotet et al. (2006) and Jolivet et al. (2010) proposed that a major detachment fault is located on top of the PQ Unit and considered it as northward extension of the Cretan Detachment. It coincides with the A4 low-angle normal fault we mapped in Mt Parnon and their proposal for major crustal extension along that fault is based mainly on metamorphic, geochronological and structural criteria. In our interpretation, this structure is just a segment of a more complicated extensional fault network that has affected the upper crust of western Peloponnesus during exhumation and this fault alone cannot explain the thinning of the upper tectonic Units. The reconstruction presented in Fig. 9 shows that the magnitude of the extension of the first extensional stage was quite small compared to that which has been proposed for the Cretan Detachment (e.g. Jolivet et al. 1996; Ring et al. 2001). This estimation is based on the pre-extensional geometry of Fig. 9a, where the Tripolitza thrust has been sketched at 8 km depth and its footwall consists of KG Unit. This depth was chosen because geophysical researches in NW Peloponnesus show that the Tripolitza Thrust reaches a depth of 8 km beneath the Pindos thrust front (Kamberis et al. 2000). As the Tripoltza Thrust is an E-dipping fault (10°–15°), then the maximum depth it can reach can be even farther. This indicates that the pressure gap across the detachments could be smaller. However, the pressure gap is still considerable. This might mean that during the first stages of exhumation of the PQ Unit, there were no dramatic effects in the higher levels of the upper crust. The first sign of brittle deformation in the upper crust started when underthrusting of the PK Unit was established.
Papanikolaou and Royden (2007) named the northward continuation of the Cretan Detachment in Peloponnesus as “East Peloponnesus Detachment System” (EPDS). In their view, the EPDS was formed in the Middle to Late Miocene and was still active in the Pliocene and maybe in the Quaternary, a suggestion that was rejected by Jolivet et al. (2010). Segments of the EPDS in Mt Parnon are described in this paper as high-angle normal faults, e.g. EPF. Based on cross-cutting relations, these high-angle normal faults formed later than the extensional detachments as they clearly dissect them (Figs. 2, 3, 7). The juxtaposition of Pindos Unit and PK Unit across the EPF, although it shows a vertical separation of several kilometres, it is the result of intense thinning of intermediate Units caused by the activity of low-angle normal faults of the first extensional phase that brought the uppermost Unit very close to the lowest one rather than the result of the action of EPF.
Segments of the early extensional network are distributed in an area exceeding 500 km2 in western Peloponnesus, and according to published data, they are even more (Fig. 1b). Lekkas and Skourtsos (2004) presented several evidence that the area south of Tripolis in central Peloponnesus (Fig. 1) has been affected by extensional detachments, and Skourtsos et al. (2002) described intense fragmented SW-tilted fault blocks of the Tripolitza Unit in Vlahokerasia tectonic window, located northwest of the study area. Unpublished data of the authors support that A2, A3 and A5 low-angle normal faults extend towards the S to Maleas Peninsula in south Peloponnesus.
Correlations with other sectors of External Hellenides
In Crete, the pressure gap across the tectonic contact between the PQ Unit and the overlying non-metamorphic tectonic Units is about 10 kbar; although Thomson et al. (1998) noticed that convincing exposures of a major detachment are difficult to find on the island, this tectonic contact was interpreted as a large-scale north-dipping detachment fault, namely the Cretan Detachment (Fassulas et al. 1994; Kilias et al. 1994; Jolivet et al. 1996; Thomson et al. 1998; Ring et al. 2001; Rahl et al. 2005; van Hinsbergen and Meulenkamb 2006). Several papers suggest the lateral extension of the Cretan detachment in Kythera Island, northwest of Crete (Marsellos and Kidd 2008; Marsellos et al. 2010), and towards the north in Peloponnesus (Trotet et al. 2006; Papanikolaou and Royden 2007; Jolivet et al. 2010). Geochronological data from the basal Unit of the Cyclades indicate that high-pressure metamorphism took place at the same time with the one taking place in the External Hellenides (Ring et al. 2001). As the pressure–temperature conditions in the Plattenkalk and Phyllite-Quartzite units on Crete and the Basal unit in the Cyclades are also similar, they must have been in close proximity in the early Miocene Hellenic subduction zone. The fact that they are now more than 100 km away from each other can be interpreted by the extension of the Cretan Detachment to the north as far as the Cyclades (Ring et al. 2001, 2009; Ring and Reischmann 2002). Detachments formed sub-parallel to the subjacent subduction thrust in Early to Middle Miocene (Ring and Reischmann 2002; Trotet et al. 2006) and their activity resulted in omission of ~20 km of the original nappe-pile (Jolivet et al. 1996; Thomson et al. 1998). The kinematic indicators of the detachments seem to be perpendicular to long axis of the External Hellenides, following the curvature of the arc. In Peloponnesus, the kinematics of the detachments seem more symmetrical [in Parnon kinematic indicators are mostly top-to-the-NE (Skourtsos 2002; Jolivet et al. 2010) and in Taygetos are top-to-the-SW (Doutsos et al. 2000; Jolivet et al. 2010)] and in Crete, the kinematic indicators are mainly top-to-the-north (Jolivet et al. 1996; Thomson et al. 1998; Ring et al. 2001; Seidel et al. 2007) although there are aspects for a more symmetrical pattern (Fassulas 1995; Fassulas 1999).
Ring et al. (2001) and Ring and Reischmann (2002) based on regional correlations (see previous paragraph) suggest a displacement of >100 km and a slip rate of >20–30 km/Myr for the Cretan detachment. An exhumation rate of ~4 km/Ma estimated by Thomson et al. (1998) implies a slip rate of 8–12 km/Ma for the Cretan detachment. On the contrary, Rahl et al. (2005) applied a modified Raman Spectroscopic Carbonaceous Material thermometer to high-pressure rocks and Tripolitza Unit in Crete and suggested that the break in metamorphic temperature at the Cretan detachment is rather small (<80°C); therefore, the detachment is responsible for only 5–7 km of the exhumation of the PQ and PK Units. This also means that there are no significant gaps in the tectonostratigraphic pile, which is in agreement with geophysical data from western Peloponnesus where the Tripolitza thrust is found at depths of 8 km without significant gaps in the tectonostratigraphic pile (Kamberis et al. 2000).
Seidel et al. (2007) describing a simple tectonic model for western Crete suggested that progressive brittle extensional deformation along the Cretan detachment resulted in clastic sediments being deposited in half-grabens sometime in the Early—Middle Miocene, which rest today immediately on the high-pressure metamorphic rocks of PQ Unit. Recently, Ring et al. (2010) based on the pattern of zircon FT ages from the footwall of the Cretan detachment suggested that there may be more than one detachment horizon, with an “older” Cretan Detachment in the east. The activity of several detachment faults especially in the brittle regime has already been described by Fassulas (1995) for central Crete. In this area, the author mentions that in many locations in northern Crete, the contact between the PK and the PQ is an extensional low-angle fault with top-to-the-north sense of shear. The tectonic contacts between the overlying non-metamorphic Units are also extensional low-angle faults characterized by a north–south symmetrical extension. Fassulas (1995) also mentions the existence of a thick marble sequence in Vassilikos, sandwiched between PQ and Tyros beds, which have slipped to the north on the PQ through a low-angle normal fault. These marbles, which are known as Vassilikos marbles [for more details, see the review by Alexopoulos (1995)], occupy a position proportionate to those of KG Unit in Peloponnesus (Skourtsos et al. 2004).
In Crete, Ρ-Τ estimates show that the metamorphic gap between the lower PK and PG Units is quite small as they have both experienced blueschists-facies metamorphism (Theye and Seidel 1992; Jolivet et al. 1996). Therefore, the tectonic emplacement of the PQ Unit on the PK Unit probably occurred in higher pressure conditions compared to Peloponnesus (4–5 kbars, Trotet et al. 2006). Although peak pressure is similar in Crete and in Peloponnesus, the P–T regime changes gradually from a colder one in eastern Crete to a gradually warmer one in Peloponnesus (Jolivet et al. 2010). Thermochronological data suggest that the peak pressure was attained 26–24 Ma ago at almost the same time along the external arc (Seidel et al. 1982; Jolivet et al. 1996, 2010; Thomson et al. 1998) and towards the north in the Cyclades (Ring et al. 2001) but the exhumation rate seems to be faster in Crete than in Peloponnesus (Marsellos et al. 2010; Jolivet et al. 2010). The pattern of zircon FT and apatite FT ages from the PQ Unit shows that the ages are getting younger from eastern to western Crete and towards south Peloponnesus and are quite older towards central Peloponnesus (Marsellos et al. 2010; Ring et al. 2010).
Doutsos et al. (2000), Trotet et al. (2006) and Jolivet et al. (2010) promote an extrusion wedge mechanism for the exhumation of the lower metamorphic Units in Peloponnesus. This wedge is defined by a basal thrust and a low-angle normal fault at its top that move simultaneously. As new material is constantly accreted at the base of the wedge, gradual uplift of the deeper Units takes place, while normal faulting in the upper portions of the accretionary prism causes thinning of the upper Units. The late high-angle normal faulting affecting both the hanging wall and the footwall of the detachments is attributed by the previous authors to post-orogenic extension and true crustal thinning. On the contrary, Papanikolaou and Royden (2007) propose a back-arc style extension to explain normal faulting in Peloponnesus. The extension results from the opening of Argolikos Gulf, as a back-arc basin, on the NW-prolongation of the Cretan basin. This opening is attributed to a retreat of the subduction zone and the onset of curvature of the Hellenic Arc. In such a case, the Late Miocene—Early Pliocene extensional detachment faults affecting eastern Peloponnesus are the result of crustal thinning associated with the early stages of opening of Argolikos Gulf (Van Andel et al. 1993; Papanikolaou et al. 1994), and the Late Pliocene—Early Pleistocene high-angle normal faulting is correlated with a later stage, during which the Gulf widened and subsided (Van Andel et al. 1993; Papanikolaou et al. 1994; Anastasakis et al. 2006).
We believe that the early extensional deformation can be explained by extrusion wedge mechanism as the structural elements, the P–T paths and the available geochronological data match better with that mechanism. We are not also convinced that the late extensional episode is the result of crustal thinning. Although high-angle normal faulting has produced extensional basins in their hanging wall filled mainly by continental deposits (<300 m thick), their throw, deduced from the vertical displacement of early extensional structures, is <2–2.5 km and in many of those just a few tens of metres (Skourtsos 2002). The horizontal extension caused by these faults on Parnon is no more than 5–7 km (that is without taking into account the faults existing to the west of the study area), distributed 60% in ΕPF and 40% in the rest of the faults (Skourtsos 2002). These alone cannot explain the exhumation of PK Unit from a depth of 20–25 km near the earth’s surface but only the uplift of the Unit in heights of 2 km. The deformation style of the Unit also shows continuous folding with top-to-the-west/southwest sense of shear and in the last stages intense backfolding and backthrusting, structures that can be interpreted by the slip of the Unit on a thrust ramp. Recently, Skourtsos and Kranis (2009) have claimed that underplating processes could be responsible for the E-W extension that is observed on Mt Mainalon in central Peloponnesus overprinting earlier N–S extensional structures.
Our analysis shows the systematic progressive development of new fault systems on the footwall of earlier extensional structures, which gradually become deactivated; a procedure that can be repeated. We suggest that extensional deformation is mainly controlled by extrusion mechanism, including folding, which lead to gradual uplift of the deeper Units, as new material is constantly accreted, while gravitational collapse occurs at the upper portions of the accretionary prism.
Detailed mapping of the extensional faults allowed defining the geometry and the evolution of the extensional fault network that affected the imbricated tectonic Units exposed in Mt Parnon since the Upper Miocene. Extension was accompanied by two episodes of normal faulting that accommodated the thinning of the upper non-metamorphosed Units and resulted in the exhumation of the deepest metamorphic ones. The weak layers of the nappe-pile allowed the development of several low-angle normal faults linked to high-angle ramps crosscutting the competent carbonate rocks. This extensional faulting gradually affected deeper layers of the nappe-pile, which is indicative of continuous extension. This process is the consequence of the collapse of the upper crustal layers during orogenic building.
We would like to thank Laurent Jolivet and Owe Ring for their careful and constructive reviews. We are very grateful to Haralambos Kranis and Konstantinos Soukis for the discussion on the topics of this paper. Research was funded by the National and Kapodistrian University of Athens research grants and the State Scholarship Foundation.