Abstract

Recent uplift in Crete, along the Hellenic trench, has exposed the internal portions of Alpine nappes that were transported south along north-dipping thrusts from the middle Cretaceous to the Oligocene (100–20 Ma). Flysch deposited along the leading, southern edge of the Alpine foreland basin dates the timing of nappe transport and contains twinned calcite in synorogenic limestones and calcite veins that record the stress-strain fields associated with nappe transport and stacking for a 17 Ma period between 35 and 18 Ma. The flysch ages young away from the craton thereby documenting thin-skin thrusting and nappe motion, for the first time, toward the trench.

Fossiliferous marine limestones and calcareous shales were deposited in marginal basins on the leading edges of four thrust-nappes in Crete: the higher and older flysch deposits (Asteroussia, Pindos) are underlain by younger flysch deposits of the Tripolitza and Plattenkalk nappes. The ages of these flysch deposits are well constrained and the flysch sequences are now found deformed between nappes without visible signs of Alpine metamorphism, except the lowest Plattenkalk flysch. Field observations in the flysch deposits include variable bedding, vein, fold axis, and kinematic orientations. Twinned calcite in the flysch limestones (16 samples, n = 425) and veins (23 samples, n = 612) generally preserve subhorizontal, in-transport shortening and vertical extension; where there is a strain overprint (high negative expected values [NEV]), a vertical shortening strain with transport-parallel extension (~N-S) is preserved. Strain magnitudes are greater in the vein sample suite, and differential stress magnitudes responsible for twinning are −230 bars for the entire sample suite. The tectonic evolution of the region involved thin-skinned shortening with south-vergent nappe formation, from north to south, followed by thrust motion on the Cretan detachment. Today, metamorphosed rocks (i.e., Phyllite-Quartzite unit [PQU]) overlie the basal Plattenkalk nappe placing earlier exhumed rocks of the Asteroussia, Tripolitza, and Pindos nappes on top of high-pressure (HP)-metamorphic rocks of the PQU without the deposition of any flysch.

INTRODUCTION

Synorogenic sedimentation in the Alps is well documented, where the terms molasse, mélange, and flysch were first introduced (Studer, 1827a, 1827b; Vassojevich, 1948; Hesse, 1964). Synorogenic sedimentation in thrust belts was first identified as a vital process in the understanding of the timing of thrust sheet motions by Rubey and Hubbert (1959) and Armstrong and Oriel (1965) in the Idaho-Wyoming belt. Further refinements on the ages of synorogenic sediments in this belt, by Dorr et al. (1977) and Wiltschko and Dorr (1983), in the context of a migrating foreland basin (Jordan, 1981), established the “younging toward the craton” progression of thrust sheet motions for this and other thrust belts (Dahlstrom, 1969; Chapple, 1978; Davis et al., 1983; Dahlen et al., 1984).

In Crete, the peak metamorphic mineral ages lie between 20 and 30 Ma (Seidel et al., 1982; Jolivet et al., 1996; Zulauf et al., 2007), were part of the Alpine orogen, and are on the leading edge of the evolving convergent margin between Africa and Eurasia as defined by the north-dipping African slab to the south of the Mediterranean Accretionary Complex and various andesitic volcanoes (e.g., Santorini) to the north (Fig. 1). The current subduction-related tectonism has produced significant topographic relief in Crete since the Pliocene (Meulenkamp et al., 1988, 1994). Active normal faults (see Fassoulas, 2001, his fig. 9) expose the internal structures of the older Alpine belt, namely the four stacked, south-vergent nappes that are separated by dated flysch deposits (Seidel, 1968; Fytrolakis, 1972; Bonneau, 1973; Richter et al., 1978; van Hinsbergen and Meulenkamp, 2006). The flysch deposits define a regional “younging away from the craton (i.e., southerly) thrust-nappe succession,” and the flysch deposits contain sparry, fossiliferous limestones and synorogenic calcite veins with mechanically twinned calcite. Both the regional thrust-nappe progression kinematics and calcite twinning strains in the flysch rocks constrain the dynamics of thrust-nappe relations within the Alpine orogen in Crete, including the debated origin of the “Cretan detachment” as a normal fault (Lister et al., 1984; Fassoulas et al., 1994; Kilias et al., 1994; Jolivet et al., 1996; van Hinsbergen et al., 2005), thrust fault (Bonneau, 1984; Xypolias and Doutsos, 2000; Doutsos et al., 2000; Zulauf et al., 2002; Kokkalas and Doutsos, 2004; Chatzaras et al., 2006; Xypolias and Kokkalas, 2006; Klein et al., 2008a, 2008b; Xypolias et al., 2008), or both (Kilias et al., 1994; Fassoulas et al., 1994; Thomson et al., 1999). Our goal in studying the Crete flysch deposits was to use the synorogenic mechanical twinning in flysch limestone and vein calcite to better understand the stress-strain fields associated with development of the thrust-nappe stacks and compare this with calcite strain results within the adjacent nappes (Zulauf et al., 2002; Klein et al., 2008a, 2008b) to reconstruct the path and geodynamics of the orogen.

REGIONAL GEOLOGY

The tectonic evolution of the Alpine belt in Crete was created by north-dipping, Miocene-Cretaceous subduction between Apulia and a Gondwana-derived microcontinent called Mesogea with Europe (Creutzburg and Seidel, 1975; Le Pichon and Angelier, 1979; Seidel et al., 1982; Bonneau, 1984; Hall et al., 1984). South-vergent nappes, which moved along north-dipping thrusts partially along underlying Triassic evaporites, represent the thin-skinned portion of the deformation; the final tectonic movements placed a nappe succession of metamorphic rocks (i.e., the Phyllite-Quartzite unit [PQU]) in fault contact (i.e., the Cretan detachment) between the underlying, locally metamorphosed carbonates of the Plattenkalk nappe (local aragonite pseudomorphs, lawsonite, Mg-carpholite + pyrophyllite + diaspore; HP-LT) and three unmetamorphosed upper nappes (Asteroussia, Pindos, and Tripolitza). Deformation and metamorphism of the lower nappes is subduction-related, as is indicated by blueschist-facies minerals such as glaucophane, carpholite, and chloritoid (400 °C, ~12 km depth), all of which define a fold-axis parallel N-S trending mineral stretching lineation in Crete. This lineation, which is approximately perpendicular to the recent strike of the subduction zone between Africa and Europe, has been interpreted in terms of subduction-related stretching (slab pull; e.g., Zulauf et al., 2002) and detachment faulting (e.g., Jolivet et al., 1996). Although blueschist-facies minerals are lacking within the upper nappes, anchizonal values of illite crystallinity and additional thermal indicators indicate elevated temperatures in the case of the Tripolitza unit. For the given tectonic setting, considerable overburden is required to explain these data (Feldhoff et al., 1991, 1993). Taking into account the current thickness of the Cretan nappe pile (~1–5 km), and the upper nappes in particular, the thermal structure (i.e., peak temperatures) indicates that the existence of significant lateral displacements of Miocene or younger ages may not be limited to the so-called Cretan detachment fault.

Complicating the interpretation of the structural geology is the E-W antiformal geometry of the tectonostratigraphic boundary between the upper and lower nappes, which often obscures exposure of the grandness and south-vergent nature of the nappes. This structure is related to the youngest event of folding that caused open, approximately E-W trending folds from map- to sample-scale (e.g., Meulenkamp et al., 1988, 1994; Dornsiepen and Manutsoglu, 1994; Zulauf et al., 2002). Furthermore, the whole nappe pile is affected by several episodes of normal and strike-slip faulting (e.g., Angelier et al., 1982; ten Veen and Kleinspehn, 2003; Fassoulas, 1998), all of which obscure the original (i.e., Miocene) structure of the nappe pile (e.g., Klein et al., 2008a, 2008b).

The so-called Cretan detachment represents a zone of cataclastic deformation that has variable thicknesses (270 m–2 km). The zone occurs between the Tripolitza unit and the PQU or the Plattenkalk unit (PKU) (i.e., the lower nappes) and mainly reworks the PQU. Within the PQU there are complex kinematic indicators (Xypolias and Kokkalas, 2006), strain variations (Doutsos et al., 2000; Xypolias and Doutsos, 2000; Xypolias et al., 2007, 2008), and younger, normal fault offsets (small top-to-the-north and south offsets in the PQU), and the PQU mimics the regional antiformal structure (see later). Exhumation rates based on U-He zircon data suggest cooling below 100 °C at 9–12 Ma for the lower nappes and 15–17 Ma for the upper nappes (J.M. Rahl, 2008, personal commun.) indicating that the nappe pile as exposed today must have been reworked after exhumation-related cooling of the HP rocks during upper Miocene–Pliocene times. Global Positioning System (GPS) arrays and fault-striae analysis in Crete indicate uplift began in Tortonian time (ca. 11 Ma) and intensified during the past 5 Ma (Pliocene–Recent), especially in the western part of the island (Pirazzoli et al., 1982; Meulenkamp et al., 1988, 1994), as the Hellenic subduction system evolves between Africa and Europe. The Tortonian compromises the time of the largest basin expansion and led to deposition of the Prina Complex in Crete (e.g., Fortuin and Peters, 1984).

Nappe Geology

The Asteroussia nappe consists of Jurassic and Cretaceous ophiolites (Figs. 2,202 and 3A, Table 1) and various metasedimentary and crystalline rocks, all of which represent a tectonic mélange thought to represent an ancient continent-ocean boundary that has been obducted to the south (Seidel, 1978; Seidel et al., 1981; Bonneau, 1984; Hall et al., 1984). Jurassic-aged flysch is overthrust by Asteroussia rocks in a few places in Greece (Vardar zone) as the highest structural unit; in Crete the flysch is Maestrichtian–upper Eocene in age. The Pindos nappe underlies the Asteroussia nappe and is composed of marine Upper Triassic–Paleocene limestones that include chert layers and nodules (Figs. 2, 3B, and 3C). Folded Pindos limestones bury flysch of Late Cretaceous age, and a younger flysch of Paleocene to Eocene age. The first flysch is related to closure of the Pindos Ocean, whereas the second flysch deposit resulted from Alpine thrust-nappe motion (Bonneau and Fleury, 1971; Table 1). Pindos structures overlie the Tripolitza sediments, which are a thick, monotonous sequence of gray, micritic limestones and dolomites of Upper Triassic to Eocene age (Creutzburg and Seidel, 1975; Bonneau and Karakitsios, 1979; Zambetakis-Lekkas et al., 1998; Fig. 2, Table 1); in western Crete the siliclastic Ravdoucha beds of Triassic age are interpreted as the sedimentary substratum of the carbonate Tripolitza unit (Sannemann and Seidel, 1976). Flysch of Eocene–Oligocene age is associated with the Tripolitza nappes (Bonneau, 1973; Seidel and Wachendorf, 1986).

Separating the upper Asteroussia, Pindos, and Tripolitza nappes from the lowest Platten-kalk nappe (see below) is the PQU—a metamorphic succession of contrasting rock sequences which comprise the Cretan detachment. Most authors treat this unit as a uniform slice, but significant differences exist in mineral provenance of contributing basement slices (e.g., Romano et al., 2006; Xypolias et al., 2008), illite crystallinity (e.g., Zulauf et al., 2002), and strain (e.g., Xypolias and Kokkalas, 2006; Barthelemes et al., 2004). Wachendorf et al. (1974) considered the PQU of eastern Crete to form a mélange, but recognizable lateral continuations of distinct lithological units contradict this interpretation (Haude, 1989). Several criteria, such as plant fossils and conglomerates, indicate that particular sections belonging to the PQU were deposited within a continental-to-shallow marine environment (Hall et al., 1984). On the other hand conodonts point to pelagic facies (e.g., Krahl et al., 1986) and a tentative palaeo-bathymetric sketch is given in Dornsiepen et al. (2001). The PQU probably was deposited on now-missing subducted continental crust also referred to as Mesogean crust (Hall et al., 1984; Thomson et al., 1999; Meier et al., 2004). Conodont faunae of western Crete indicate deposition of the PQU to extend from Pennsylvanian (Upper Carboniferous) to Upper Triassic (Krahl et al., 1983). In eastern Crete it extends from Permian to Upper Triassic (Kopp and Richter, 1983; Krahl et al., 1986; Kozur and Krahl, 1987; Haude, 1989). Several crystalline slices are sandwiched between slaty lithologies, metacarbonates, metavolcanics, and metaevaporites (anhydrite, gypsum) (Franz, 1991; Barthelmes et al., 2004; Romano et al., 2006). The absence of flysch (Table 1, Fig. 2) along this boundary suggested to early workers that the contact PQU/Tripolitza nappe was (originally) a stratigraphic and not a structural feature (Bonneau, 1984; Hall et al., 1984; Thomson et al., 1999). All the nappes are transported or offset by the Cretan fault, including flysch as young as 5 Ma in the northern Pelopponese (Laj et al., 1982), or younger (Underhill, 1989), so it is a young feature.

The structurally lowest unit (“the lower nappe”), the Plattenkalk nappe (Figs. 2 and 3D), consists of Mesozoic to Eocene metacarbonates (Bonneau, 1973) that are covered by limy metaflysch of upper Eocene–Oligocene age (Fytrolakis, 1972; Bizon and Thiebault, 1974; Krahl et al., 1983, 1986). The Plattenkalk limestones contain a number of primary sedimentary structures that reveal the grandeur of the nappes by delineating overturned limbs; metaevaporites are also associated with the Plattenkalk revealing the nature of the Triassic materials along the regional detachment. In western Crete the Upper Triassic to Lower Jurassic (Krahl et al., 1983) carbonate Tripali unit is sandwiched between PKU and the PQU, but flysch deposits are either missing or not preserved. The paleogeographic and tectonostratigraphic significance of the Tripali unit is still a matter of debate (Creutzburg and Seidel, 1975; Krahl et al., 1983; Karakitsios, 1987; Pomoni-Papaioannou and Karakitsios, 2002).

Nappe-Flysch Relationships

Asteroussia (Pelagonian), Tripolitza, and Pindos flysches are a product of a predominately siliciclastic deep-water turbidite system. They show the usual sedimentary features, e.g., well-developed Bouma cycles in different thicknesses and contain the typical neritic facies trace fossil assemblages. At their base, calcareous and fine-grained sediments are more abundant while olistostrome beds occur in the higher portions. The siliciclastic provenance for all these flysches is more or less enigmatic; the thicknesses of the flysch successions can exceed 1000 m, but the nappe deformation makes true thickness calculations difficult. The flysch of the PKU (Kalavros Beds in Crete, Vatia Beds in the Peloponnese) differ from the flysch deposits mentioned above. In the immediate hanging wall of Plattenkalk marbles in Crete, the lithology of the Plattenkalk flysch is dominated by marls and limestones. From the Peloponnese region, limestone turbidites and debrisites are known. Despite the penetrative deformation observed in flysch deposits, original sedimentary features are preserved in some parts clearly showing limestone turbiditic sequences although some sections lack Bouma A-sections. The preserved flysch thickness in Crete is small (<400 m; Fig. 2), such that the calciturbiditic dominance is often limited to the lower part of the flysch sequence. In the Peloponnese, some exposed sections show an increasing influx of siliciclastic material up-section pointing to a shift from shallow- to deep-water conditions. Again, poor regional structural control makes it difficult to create a reliable palinspastic reconstruction so as to delineate facies changes in the flysch.

Thin, discontinuous outcrops (10–100 m) of fossiliferous marine limestone and calcareous shales with local conglomeratic layers and olistostrome zones are common; some flysch deposits are spliced into the complexity of a faulted nappe or appear to be a regional deposit between two different nappes. Large olistostrome clasts are derived from hanging-wall rocks of the adjacent nappe. Flysch outcrops appear either monotonous and upright or intensely folded, and are always crosscut by a variety of calcite veins. Fold limb dips (5° to 70°) and fold plunges (0° to 60°) are variable; some folds are open, some are of a chevron style, and some are sheath folds. Stylolites are present in some of the limestone layers in many orientations. Only the Plattenkalk flysch contains evidence of metamorphism with minor lawsonite and carpholite. Striated surfaces are also present in all outcrops and on most calcite vein surfaces. Due to the different rheological characteristics, preserved sedimentary contacts between limestone and flysch do not exist nor are nappe-flysch fault contacts observed.

METHODS

Calcite twins mechanically at low differential stresses (~10 MPa; see Lacombe and Laurent, 1996; Ferrill, 1998), and twinning is largely independent of temperature and normal stress magnitudes in the uppermost crust. Twinning is possible along three glide planes and calcite strain-hardens once twinned; further twinning is possible in a crystal along either of the remaining two e{0112} planes at higher stress levels, provided that stress is oriented >45° from the initial stress orientation (Teufel, 1980). The application of twinned calcite to structural and tectonic problems has been primarily restricted to studies of limestones (e.g., Groshong, 1975; Engelder, 1979; Spang and Groshong, 1981; Wiltschko et al., 1985; Craddock et al., 1993), calcite veins (e.g., Kilsdonk and Wiltschko, 1988), or, more rarely, marbles (e.g., Craddock et al., 1991). Amygdule and vein calcite in basalts also yield interpretable results (DSDP Hole 433C, Craddock and Pearson, 1994; Keweenaw rift, Craddock et al., 1997; Iceland, Craddock et al., 2004). Rowe and Rutter (1990) and Burkhard (1993) have reviewed the variety of methods applied to utilizing twinned calcite in a host of geologic environments.

Paleostress (paleopiezometry of Engelder, 1993) responsible for twinning can be calculated in terms of its compressional (or tensile) orientation (Turner, 1953) and magnitude (Jamison and Spang, 1976; Rowe and Rutter, 1990). Strain ellipsoid axis orientations are computed using the calcite strain gauge (Groshong, 1972, 1974) and are quite accurate for strains ranging from 1%–17% (Groshong et al., 1984a). Strain magnitudes vary greatly, however, depending on factors such as lithology, grain size, and porosity, and are a function of twin thickness. Three orthogonal sections were used for each of the fold samples, and 1 to 2 sections for the cleavages samples. Thin twins (~0.5 μm) are dominant in our sample suite and are characteristic of calcite deformed below 200 °C (Ferrill, 1991, 1998; Ferrill et al., 2004). The calcite strain gauge technique also computes positive and negative expected values (PEV and NEV, respectively) for all the twins in a given thin section. A NEV for a twinned grain indicates that this grain was unfavorably oriented relative to the stress field that caused the twinning in the majority of the grains in a given thin section. A high percentage of negative expected values (>40%) indicates that a second, noncoaxial twinning event occurred (Teufel, 1980). Two twinning strains (PEV and NEV groups, respectively) can be analyzed separately. Strain fabrics are interpreted by plotting the sample's bedding relative to the three strain axes. If the maximum shortening axis (ϵ1) is ±20° from bedding this is a layer-parallel shortening strain (LPS). Shortening in the plane of a vein (±20°) would be vein-normal shortening (VNS).

Flysch limestone-calcite vein pairs for each nappe from Crete were analyzed. Stable isotopes were analyzed with a Finnigan MAT252 with a Kiel automatic carbon dioxide extraction unit at the University of Minnesota. Samples were reacted with 105% phosphoric acid for 300 s at 70 °C and compared to standards NBS-19, NBS-18, and LSVEC from NIST (National Institute of Standards and Technology) for carbon and oxygen and normalized to VPDB (Vienna Pee Dee belemnite). The reproducibility of a single measurement is better than 0.06‰ (1σ).

RESULTS

Field Observations

We defined the relationship of a flysch deposit to a nappe by the outcrop appearance of the sediments, their age (where known), and the local fault relations of the adjacent nappe. For example, if a flysch deposit was overthrust by Tripolitza limestones or fault-bounded by Tripolitza limestone, we classified this as a flysch derived from and deformed by the Tripolitza nappe. The clasts within the Tripolitza flysch are mainly derived from the older, structurally higher Pindos nappe. Structural elements in the flysch deposits (Fig. 4) always appear contained to the flysch, meaning folds, veins, etc. cannot be traced into overlying or underlying nappe sediments (Fig. 3). Tripolitza and Pindos limestones are easy to distinguish as thickness, color, and structural style largely differ. Similarly, the Asteroussia nappe rocks are distinctive (ophiolite scraps, etc.) and are only exposed in a few places. The Plattenkalk flysch is poorly exposed but has a distinctive greenish-brown color, is more massive, has a variety of rounded, cobble-sized clasts, and appears metamorphosed, particularly in central Crete (see also Kilias et al., 1994; Fassoulas et al., 1994; Jolivet et al., 1996). This classification is the basis for our strain analysis organization (Table 1).

Structural observations include calcite vein, fold axis, and lineation data (Fig. 5) and are combined into a common thrust transport direction across ~105° of curvature along the belt (Fig. 1). Calcite veins are prevalent in the Asteroussia, Pindos, and Tripolitza flysch outcrops and rare in the Plattenkalk exposures, and are in a variety of orientations (Fig. 5A). Folds in the flysch have a variety of styles (Fig. 4) that appear dominated by the ratio of limestone to shale; limestone-free outcrops do not have folds. Folds appear to be of one generation (F1) and are not refolded, and we did not observe any folding-related cleavage formation. Regionally, two F1 fold orientations are prevalent: folds that trend parallel to the thrust-nappe transport direction (some are sheath folds) and folds parallel to the axes of the overlying nappes (Fig. 5B). We observe similar fault striation and mineral lineation orientations in the flysches. Where kinematic offsets are interpretable, a top-to-the-south (parallel to thrust transport) shear sense is prevalent, with a less prevalent E-W, strike-slip (normal to transport) fabric present (Fig. 5C).

Flysch Metamorphism and Fluids

The Plattenkalk flysch is as metamorphic as the Plattenkalk limestones with P > 8 kbar and T = 300–400 °C for western and central Greece (Theye and Seidel, 1991, 1993; Rahl et al., 2005). Deposits of the Plattenkalk fly-sch in eastern Crete are less metamorphic (T ~200–250 °C; Soujon and Jacobshagen, 1997; Rahl et al., 2005) and do not show new formation of metamorphic minerals. Metamorphic minerals occurring in Plattenkalk marbles and intercalated bauxites are lawsonite and magnesio-carpholite, respectively (Seidel et al., 1982). Illite crystallinity and vitrinite reflection data suggest T ~170–200 °C for the Tripolitza flysch, whereas the Pindos flysch experienced conditions corresponding to high diagenesis (Feldhoff et al., 1991, 1993; Klein et al., 2008a, 2008b). Based on apatite fission track analysis the Asteroussia flysch should be free of Alpine metamorphism (Thomson et al., 1998). Stable isotopes of flysch limestone-vein pairs from a few flysch outcrops (Table 2,03, see Table 4 for site locations) demonstrate that the calcite veins are derived from flysch limestones and are locally reprecipitated. In a forthcoming paper, Klein and coworkers sampled travertine veins around Crete that crosscut all the nappes, and these δ18O and δ13C values are considerably more negative, and presumably of meteoric-geothermal origin.

Asteroussia Flysch

We sampled a calcite vein in the flysch below ophiolitic rocks of the Asteroussia nappe near the town of Arvi in east central Crete (Figs. 1 and 3A). The flysch limestone was too micritic for calcite twin analysis. The horizontal vein contains twinned calcite that preserves a shortening direction (127°, 2°) approximately parallel to the regional thrust-nappe transport direction, and vertical extension. An E-W transport lineation was described from the uppermost unit by Fassoulas (1998) and related to the formation of the upper nappe pile. The shortening strain is −0.8%, in response to a differential stress of ~–244 bars. There is no secondary strain overprint (25% NEVs). As the shortening strain axis is within the plane of the vein, the fabric interpretation is vein-parallel shortening (Tables 3 and 4; Fig. 6).

Pindos Nappe and Flysch

Eight flysch limestones (n = 286 twins) and seven calcite veins (n = 242 twins) were collected from Pindos nappe sites (Fig. 1, Tables 3 and 4). In the limestone strain suite, seven of eight samples record layer-parallel shortening, generally in the direction of the thrust-nappe transport direction, i.e., ~top-to-the-S (Fig. 6). The other limestone strain result preserves layer-normal shortening, but also a strain overprint (31% NEVs) that is layer-parallel. The calcite veins record a mix of vein-parallel and vein-normal shortening strains that are generally normal to the thrust transport direction, but parallel to transport lineations within the uppermost unit (~E-W; Fig. 6); Pindos veins recorded the highest strains of the sample suite (Table 4) and only modest NEVs (no strain overprint). Differential stresses for the Pindos samples are ~–240 bars.

The beaches and cliffs around Agio Pavlos, along the south coast of Crete (Fig. 1), expose folded Pindos limestones (Figs. 3B and 3C) thrust over flat-lying Pindos flysch. The outcrop contains interlayered Cretaceous white limestones and red cherts. The folds trend N-S with shallow, variable plunges and their axial planes dip to the east at 40°, all curious for the south terminus of a nappe that was transported from north to south and now dips to the south. The twinning strains are from the same limestone layer and both preserve a low-NEV, LPS fabric normal to the fold axis, which is normal (i.e., E-W) to the regional thrust-nappe transport direction. The calcite veins, which are parallel to a weak axial planar cleavage, contain a sub-horizontal shortening strain that is vein and thrust transport-parallel (N-S). There is no strain overprint (low NEVs), the vein strain magnitudes are high (−2.5%), and the differential stress magnitudes are regionally consistent (−245 bars). The Pindos flysch below the Pindos folds (one limestone, two veins) contain twinning strain results that are complex. The limestone has a high strain and preserves a NE-SW LPS fabric. The veins record ~E-W subhorizontal shortening strains, one that is vein-parallel and one that is vein-normal. There is no strain overprint recorded here, and the differential stresses are consistent. E-W shortening is possibly related to W-vergent folding and/or the emplacement of the uppermost unit/Asteroussia nappe.

Tripolitza Flysch

Flysch samples were collected from nine Tripolitza sites (Fig. 1, Tables 3 and 4). Four limestones (n = 71) and ten calcite veins (n = 309) were analyzed resulting in 5 limestone and 16 vein strain analyses. The limestones preserve LPS fabrics that are generally in the plane of regional thrust transport (Fig. 6), with one high NEV result which is a layer-normal (vertical) strain overprint. The limestones have the highest strain value in the sample suite. Ten vein strain analyses produced 16 results, 10 PEV and 6 NEV strain overprints with a mix of vein-normal and vein-parallel fabric interpretations. The shortening axes for the PEV strains are random, whereas the NEV strain overprints record a subvertical strain. Differential stresses for the Tripolitza samples are ~–240 bars.

Plattenkalk Flysch

Flysch samples were collected from three sites in eastern Crete (Fig. 1) resulting in four strain analyses from three flysch limestones and one calcite vein (Tables 3 and 4; Fig. 6). The three limestones preserve LPS fabrics with the shortening axis generally NW-SE. The vein preserves a N-S shortening strain that is in the plane of the vein. There is no strain overprint in these rocks, despite being in the lowest nappe, and the strain and differential stresses magnitudes are comparable to the rest of the suite. Differential stresses for the Plattenkalk samples are ~–215 bars. As the Plattenkalk flysch was subducted, the observed strain may result either from subduction-related shortening or thrusting. The stretching lineations of the Plattenkalk fly-sch in eastern Crete trend NNW-SSE and were interpreted to have formed near the metamorphic peak (e.g., Zulauf et al., 2002).

Santorini PQU

We also visited Santorini, the active andesitic volcano ~100 km north of Crete, which is presumably magmatically related to the north-dipping African plate (Fig. 1). On the eastern side of the inner caldera near the new harbor (Fig. 7A), the eruptive strata overly a complexly deformed north-dipping suite of schists, amphibolites, marbles, and quartzites (Fig. 7B; Davies, 1999) and are overlain by Triassic limestones that are coated by younger, layered travertine, all reminiscent of the PQU relations in Crete and Kythira. We observed top-to-the-south striation kinematics within the PQU, and twinning strain results preserve a N-S, subhorizontal shortening fabric with no strain overprint (Fig. 7C). Strain magnitudes and differential stresses are comparable with the entire sample suite (Table 4).

DISCUSSION

Thrust belts are characterized by a prethrust transport, layer-parallel shortening (LPS) fabric that is preserved across a sedimentary accretionary wedge where the shortening axis is parallel to the younger thrust transport direction. This regional shortening (−3% to −5%) is preserved by finite strain markers (Engelder and Engelder, 1977) and mechanical twins in limestone calcite in both the autochthonous foreland and the adjacent thrust belt. Far-field orogenic stresses are preserved up to 2000 km into the foreland (Ziegler, 1995; Craddock and van der Pluijm, 1989, 1999). The prethrust LPS fabric can be used to interpret the style of the orogenic margin: parallelism of a prethrusting LPS fabric between the foreland and thrust belt indicates that thrust sheet motions did not involve rotation out of the plane of thrust transport during a collisional orogen (Kilsdonk and Wiltschko, 1988, Appalachian orogen; Craddock and McKiernan, 2007, Mazatzal orogen). The lack of LPS fabric parallelism between a thrust belt and foreland indicates that thrust motions were accommodated by thrust and/or fold rotation in the thrust belt in conjunction with oblique motion along the subduction margin (Craddock, 1992, Idaho-Wyoming belt; Craddock et al., 2007b, Cape belt). Thrust belt development involves accretion of the thrust piggybacks from the margin toward the craton, dated by terrestrial synorogenic sedimentation, occasionally modified by thick-skinned thrust uplifts when the subduction slab dip (Royden and Burchfiel, 1987; Beaumont et al., 1996; Bird, 1988) or location (i.e., “slab roll-back”; e.g., Royden, 1993; Pfiffner et al., 2002) changes. Strain studies in thrust belts (see Siddans, 1983), including the use of quartz deformation lamellae and twinned calcite (see literature review in Craddock and Relle, 2003), are rich with documentation of layer-parallel shortening strains in the plane of thrust shortening. There are a few, rare reports of primary or overprint strains that are subvertical and/or layer-normal shortening. Mosar (1989) and Groshong et al. (1984b) report vertical, layer-normal shortening strains in some flysch rocks in the Pre-Alp and Helvetic nappes, respectively. Craddock and McKiernan (2007) have reported a combination of transport-parallel, layer-parallel, and layer-normal shortening axes in quartzites of the Baraboo interval across the Mazatzal orogen (1.65 Ga) and interpreted these results to be related to nappe burial. Similarly, Craddock et al. (2007a) have reported a thrust transport-parallel, layer-parallel shortening strain in lamprophyre dike calcite ocelli and dike-margin calcite veins in the foreland of the Penokean orogen (1.85 Ga). This calcite also recorded a secondary vertical shortening strain (high NEVs), the result of nappe burial; limestones just to the north, in the distal Penokean foreland, were not buried and do not record a twinning strain overprint.

The Alpine orogen has a complex oroclinal shape that resulted from the collision of the Apulia microplate with Europe (Coward and Dietrich, 1989; Stampfli et al. 1998; Rosenbaum and Lister, 2004). Thrustnappe stacks evolved with northerly transport north of the Insubric suture, with south-transported nappes on the south (Steck and Hunziker, 1994; Pfiffner et al. 2002). Crete represents a southern continuation of the south-vergent nappes where the regional tectonism is constrained by the ages of the synorogenic flysch deposits that span ca. 80 Ma (Figs. 2, 7, 8, and 9,902; Table 1). The Cretan flysch deposits are the best preserved marine synorogenic deposits in a thrust belt. These flysch deposits preclude significant synorogenic topography because the evolution of the thrust-nappe stack was at or near sea level, and that the thrust younging direction was southerly and toward the ancient north-dipping trench (Fig. 9A). The current exposure and elevation in Crete are due to the Recent evolution of the Hellenic trench and the Mediterranean Ridge complex, the emergence of Crete as a horst structure, and the offset and uplift on the Cretan fault after deposition of the Plattenkalk flysch at ca. 20 Ma; estimates of the amount of crust missing at the Cretan fault range from > 20 (e.g., Ring et al., 2001) to ~8 km (Rahl et al., 2005). If the Cretan fault operated at a low angle, a large horizontal displacement would be necessary (e.g., Ring et al., 2001). The fault zone brought a sequence of partially contrasting (both by age and metamorphic grade) and strongly deformed rocks to the surface known as the PQU (Creutzburg and Seidel, 1975; Seidel et al., 1982; Kopp and Richter, 1983; Krahl et al., 1986; Bonneau, 1984; Haude, 1989; Zulauf et al., 2002; Barthelemes et al., 2004; Klein et al., 2008a, 2008b; Zulauf et al., 2007). The presence of high-grade metamorphic rocks in the PQU suggests a change from thin-skinned nappe formation along the Triassic evaporite décollement (see Underhill, 1989) including shallow subduction of the Tripolitza unit to a tectonic regime that favored fault offset and thrust uplift of deep crustal rocks (Figs. 8, 9A, and 9B). The metamorphosed evaporates are present below the Plattenkalk nappe, within the Plattenkalk as diapirs, and as small scraps within the overlying Tripolitza flysch; the upper nappes mobilized along the evaporites and were stacked by sliding along the flysch between the nappes (Figs. 9A and 9B). The basement rocks within the PQU are Gondwana-derived, with similarities to Anatolian basement rocks (Romano et al., 2006). Changes in slab dip along the Andean trench correspond to changes in the regional shortening style (thick versus thin; Allmendinger and Gubbels, 1993), a modern analog used by Bird (1988) to model the Sevier (thin)-Laramide (thick) orogens in western North America, a model that may also be applicable to Crete alpine geodynamics. To explain the lithologic chaos and HP minerals found in the PQU, we favor a model where the Cretan detachment stepped downward from the evaporitic-flysch slip zone to a deeper level (400 °C; Chatzaras et al., 2006) thereby facilitating the uplift of the HP assemblage. Poor field relations and a lack of seismic reflection profiling to the northwest of the Kythira-Crete arc prevent improvements on the generic aspects of our schematic cross sections: the nature of the Pelagonian crust to the northwest is unknown (Xypolias and Doutsos, 2000, their Fig. 11) with respect to the PQU and the Cretan detachment; there are extreme variations on the PQU thickness and composition along strike (Xypolias et al., 2008); the role of Triassic evaporates along strike within the nappe pile and detachment is variable (Underhill, 1989; Doutsos and Frydas, 1994); the dip of the detachment cutting Apulian crust is unconstrained, as is the offset distance, thereby making a palinspastic restoration impossible. The presence of PQU rocks ~100 km to the northwest in Santorini (Figs. 7, 9B) may be part of the palinspastic constraint. Chatzaras et al. (2006, their fig. 11; see also Doutsos et al. [2000, their fig. 8] and Xypolias and Kokkalas [2006, their fig. 1]) proposed a model where flexure of the uplifting PQU-detachment caused delamination of the overlying Apulian crust whereby the crust above the detachment and below the thin-skinned nappes was down-dropped; Klein et al. (2008a, 2008b) modified this process of slab delamination (or ductile extrusion) by arguing for tectono-metamorphic variations within the PQU “stratigraphy.”

The calcite twinning fabrics and kinematics in the nappes preserve a transport-parallel LPS and top-to-the-south structural history across the stack of nappes and around the salient (Zulauf et al., 2002; Chatzaras et al., 2006; Xypolias et al., 2008) indicating that rotations between nappes during thrusting were minimal. Unfortunately for this study we do not have autochthonous foreland limestones to use as a reference strain marker for the adjacent strain studies in the nappes (see Craddock and van der Pluijm, 1989, 1999) and flysch deposits; the results in Crete are consistent with thrusting from north to south. Calcite strain studies in the Alps and alpine foreland record a layer-parallel shortening strain where shortening axes are parallel in the nappes and foreland (Craddock and Burkhard, 2006; see also Lacombe et al., 1990). The deformation recorded in the flysch deposits is more complex. Folding (F1) orientations are either parallel or normal to the inferred thrust-nappe transport direction. Folding styles vary, including sheath folds, and none of the folds is refolded. Calcite veins are common and in many orientations, and are simple, single filling veins with no preferred calcite optic axis orientations. Veins are planar, seem to have been derived locally from flysch limestones (i.e., no metamorphic overprint in the O,C isotopes; Table 2), and are often ornamented with striations. The striation kinematics preserve top-to-the-south motions or E-W, strike-slip offsets. Twinning strains in the flysch limestones (n = 15) record a crude transport-parallel shortening orientation that is bedding parallel; 13 results are LPS results with two being LNS strain overprints (high NEVs).

As the veins show abundant type I twinning, the inferred temperatures during the deformation of these veins was well below 200 °C. In most cases veins crosscut the folded flysch and thus appear to be slightly younger than the deformation of the country rock. Veins in the flysch (n = 20) preserve a more random shortening axis orientation with more VNS than VPS fabric interpretations. Strain magnitudes in the veins are higher than the limestone suite, although the differential stress responsible for the twinning is everywhere −230 bars according to the twin density method of Rowe and Rutter (1990). Strain overprints (high NEVs) are only seen in the Pindos and Tripolitza nappes (two limestones and six veins) and not the lower Plattenkalk rocks. The strain overprint is a vertical (LNS) shortening strain with transport-parallel extension (horizontal, N-S), the result of nappe burial (see below). We can thus conclude that the limestone LPS fabric forms early, strain hardens (rare twinning strain overprint), and the calcite veins form during nappe stacking (folding, limestones go into solution) and twin during that complex process. Curiously, the nappe-related differential stresses (−230 bars) seem to have been consistent throughout the deformation, which was probably fairly shallow (and <200 °C).

The folded Pindos limestones at Agio Pavlos (Fig. 3) present an interesting local deformation history that is not consistent with the regional strain-kinematic story. The folded Cretaceous limestones are oriented N-S, parallel to the regional inferred thrust transport direction. The limestone twinning strains (two samples from the same layer) record layer-parallel shortening normal (E-W) to the fold axes, but also normal to the thrust transport direction. Calcite veins (N-S, 90°) record subhorizontal shortening parallel to the fold axes, the veins, and parallel to the inferred thrust transport direction. Neither strain data set has any appreciable overprint (low NEVs). The underlying and adjacent Pindos flysch produced one limestone and two veins for strain analysis. The limestone recorded a layer-parallel, NE-SW shortening while the crosscutting veins preserve an E-W subhorizontal shortening. There is no strain overprint in the flysch rocks. The existence of E-W shortening within the Pindos unit is of special importance, as was already pointed out by Hall et al. (1984). In this context, the early E-W shortening and the absence of metamorphism may indicate that the deformation of the Pindos unit (based on the Agio Pavlos outcrops) started in a different tectonic setting, implying local and temporal diversity between the Pindos and the underlying units. One explanation might be a change in convergence rate between Africa and Europe or Mesogea and Europe, respectively. The E-W lineations reported in the Asteroussia nappe (Fassoulas, 1998) and the E-W shortening preserved in the Pindos limestones at Agio Pavlos suggest an early, perhaps local, deviation in the deformational history that is not observed in the lower nappes. The prevalence of active normal faults in Crete could also be responsible for local rotations of older strain fabrics; the south side of the Agio Pavlos folds is offset by a normal fault (E-W, 70°N dip) with dip-slip and oblique-slip striations.

As a possible reason for major changes of the governing thrust-direction the slight acceleration of convergence between Africa and Europe at ca. 32 Ma (lower Oligocene) might be a possibility (Rosenbaum et al., 2004; J.M. Rahl, 2008, personal commun.). Although Crete appears to have been relatively stable during Alpine convergence (Kissel et al., 1984), Neogene block rotations (<5 Ma) are very significant within the external Hellenides and Crete in particular (Duermeijer et al., 1998, 2000). Such block rotations might also account for the observed misalignment between fold axes and the direction of shortening as well, but shortening would then have to postdate the rotation. Otherwise the calcite strain tensor would have been passively rotated as well. Relatively young, approximately arc-normal compression is indicated by brittle structures, such as reverse faults, folds, and thrusts (Fortuin and Peters, 1984; Zulauf et al., 2002; Chatzaras et al., 2006), NE-SW compression-compatible AMS fabrics of Pleistocene sediments (Duermeijer et al., 2000), and recent earthquake data for eastern Crete (Bohnhoff et al. 2005). Taking into account these data, parts of the calcite strain tensor solutions presented here could result from compressional episodes during the late Miocene to recent structural evolution of the Cretan nappe pile, although this interpretation is inconsistent with prevalence of LPS fabrics in thrust belts that strain harden and do not record a twinning strain overprint. In summary, there is evidence of E-W shortening preserved as a weak lineation in the upper nappes (Fassoulas, 1998) and by the E-W LPS twinning fabric preserved by the Pindos nappe Cretaceous limestones at Agio Pavlos (Fig. 3B, Table 4); both of these fabrics are local. The localized E-W fabrics seem anomalous compared to the remainder of the calcite strain data: the nappe host rocks, Triassic–Oligocene in age, preserve a LPS twinning fabric in the plane of thrust transport, similar to the twinning strains in nappe-hosted calcite veins. As flysch was deposited in the foreland basin to the south of the evolving nappes, the flysch limestones were shortened parallel to bedding, folded as the flysch was overthrust, then went into solution locally precipitating as randomly oriented calcite veins, preserving random twinning strains. This flysch deposition-deformation pattern spanned 80 Ma (deposition) with deformation telescoped into a 17 Ma window (Table 1, Figs. 2, 7); this progression is presented schematically (Fig. 9) as younger normal faults have dismembered the primary, peak Alpine structures (Fig. 9B) making a palinspastic restoration impossible (see Fassoulas, 2001; Fig. 9). Twinning strain studies of syn-faulting carbonates in the Cretan fault zone also preserve LPS N-S fabrics (Zulauf et al., 2002) and top-to-the-south kinematics that suggest the Cretan nappe pile formed by Neogene thrusting postdating the exhumation of HP rocks in Crete (Fig. 8; Klein et al., 2008a, 2008b) as a young, episodic crustal offset based on the strong differences in metamorphic degree within the PQU. If the Cretan detachment (i.e., PQU) was active as a postnappe normal fault we would expect to see consistent, regional top-to-the-north normal fault kinematic indicators within the detachment plane and a regional vertical shortening strain (calcite twin and other methods) in and below the PQU; this is not observed in Crete, Kythira, nor in Santorini (Figs. 7, 9).

CONCLUSIONS

Field and calcite twinning strain studies in Alpine orogen flysch deposits in Crete record a complex deformational path as these marine deposits were squeezed between successive thrust-cored nappes with local carbonate dissolution-reprecipitation and without metamorphism. Fold orientations and kinematic indicators and calcite twinning fabrics suggest deformation in the flysch rocks was both transport-parallel and transport-normal. Flysch limestones generally preserve a LPS fabric in the plane of thrust transport, whereas the locally derived, synorogenic calcite veins record a more random shortening pattern partly demonstrating the different rheological properties of the flysch sediments as compared to the encasing, thrust-cored nappe limestones. The flysch limestone calcite strain hardened, then locally dissolved to form synthrusting calcite veins, which record the chaos of flysch being deformed between nappes and transported southward. A twinning strain overprint is preserved in the Pindos flysch (one limestone) and Tripolitza flysch (one limestone; six veins) and records a vertical shortening strain overprint, which we interpret to be related to burial by successive stacking of nappes. Fly-sch ages range from middle Cretaceous to Oligocene (80 Ma) and bracket the age of motion of each thrust-nappe sequence which youngs to the south, away from Aegean (Mesogea) micro-plate toward the paleotrench, to a deformational period of 17 Ma between 35 and 18 Ma. This is the reverse of all previously documented thrust synorogenic relations, which record thrust accretion younging toward the craton; one shortcoming of understanding thrust belt evolution is the volcanic burial near the trench where younging away from the craton would be expected but is not observed near the volcanic arc. Perhaps the Cretan Alps evolved in a unique accretionary setting without arc volcanism whereby these structures can be observed. The Cretan detachment was initially a thrust, displacing the upper Asteroussia, Pindos, and Tripolitza nappes over the lower, youngest Plattenkalk nappe thereby uplifting and deforming the chaotic grouping of metamorphosed and sheared rocks forming the PQU, which is the décollement. The Cretan fault had a deep root and likely a large displacement, and its offset did not correspond with any flysch deposition. Presently, the model does not fully explain why the Pindos nappe appears to be nonmetamorphic, whereas the assumed peak pressures and temperatures of the PQU are 300–400 °C and >8 kbar, respectively, nor do we have a clear explanation of localized, E-W fabrics in the upper nappes including the N-S folds in the Pindos nappe at Agio Pavlos.

Craddock was funded by a DFG Mercator Professorship to work on this project while in Germany. Field and laboratory assistance from Hayley Campbell, Carolyn Kilberg, Michael Rieser, Julia Ries, Tanya Romer, and Sonya Vols is gratefully acknowledged. Similarly, this project benefited from the participation of Macalester students in a field course to Crete and Santorini in January 2004, which included field time with Charalampos Fassoulas in the Ida Mountains. This effort is a portion of Klein's doctoral research (Frankfurt). A detailed manuscript review by Paris Xypolias is greatly appreciated.