1 3

Int J Earth Sci (Geol Rundsch)DOI 10.1007/s00531-014-1111-9

ORIGINAL PAPER

The Rhodope Zone as a primary sediment source of the southern Thrace basin (NE Greece and NW Turkey): evidence from detrital heavy minerals and implications for central‑eastern Mediterranean palaeogeography

L. Caracciolo · S. Critelli · W. Cavazza · G. Meinhold · H. von Eynatten · P. Manetti

Received: 29 April 2014 / Accepted: 15 November 2014 © Springer-Verlag Berlin Heidelberg 2014

a common origin of tectonic units exposed in NW Tur-key (Biga Peninsula) with those of NE Greece and SE Bulgaria (Rhodope region). The entire region displays (1) common extensional signatures, consisting of com-parable granitoid intrusion ages, and a NE-SW sense of shear (2) matching zircon age populations between the metapelitic and metamafic rocks of the Circum-Rhodope Belt (NE Greece) and those of the Çamlica–Kemer com-plex and Çetmi mélange exposed in NW Turkey. Detri-tal heavy mineral abundances from Eocene–Oligocene sandstones of the southern Thrace basin demonstrate the influence of two main sediment sources mostly of ultramafic/ophiolitic and low- to medium-grade meta-morphic lithologies, plus a third, volcanic source limited to the late Eocene–Oligocene. Detrital Cr-spinel chem-istry is used to understand the origin of the ultramafic material and to discriminate the numerous ultramafic sources exposed in the region. Compositional and strati-graphic data indicate a major influence of the metapelitic source in the eastern part (Gallipoli Peninsula) during the initial stages of sedimentation with increasing contribu-tions from metamafic sources through time. On the other hand, the western and more external part of the south-ern Thrace margin (Gökçeada, Samothraki and Limnos) displays compositional signatures according to a mixed provenance from the metapelitic and metamafic sources of the Circum-Rhodope Belt (Çamlıca–Kemer complex and Çetmi mélange). Tectonic restoration and composi-tional signatures provide constraints on the Palaeogene palaeogeography of this sector of the central-eastern Mediterranean region.

Keywords Sediment provenance · Thrace basin · Circum-Rhodope Belt · Biga Peninsula · Detrital heavy minerals · Cr-spinel geochemistry

Abstract Detrital heavy mineral analysis coupled with a regional geological review provide key elements to re-evaluate the distribution of the Rhodope metamorphic zone (SE Europe) in the region and its role in determin-ing the evolution of the Thrace basin. We focus on the Eocene–Oligocene sedimentary successions exposed in the southern Thrace basin margin to determine the dis-persal pathways of eroded crustal elements, of both oceanic and continental origins, as well as their differ-ent contributions through time. Lithological aspects and tectonic data coupled with geochemistry and geochronol-ogy of metamorphic terranes exposed in the area point to

Electronic supplementary material The online version of this article (doi:10.1007/s00531-014-1111-9) contains supplementary material, which is available to authorized users.

L. Caracciolo (*) · S. Critelli Dipartimento di Biologia, Ecologia e Scienze della Terra (DIBEST), Università della Calabria, via P. Bucci, cubo 4b, Rende (CS), Italye-mail: luca.caracciolo@unical.it

L. Caracciolo Chemostrat Ltd., Sandtrak Unit 1, Ravenscroft Court, Buttington Enterprise Park, Welshpool, UK

W. Cavazza Dipartimento di Scienze Biologiche, Geologiche e Ambientali, Università di Bologna, Bologna, Italy

G. Meinhold · H. von Eynatten Abteilung Sedimentologie/Umweltgeologie, Geowissenschaftliches Zentrum der Georg-August-Universität Göttingen, Göttingen, Germany

P. Manetti Dipartimento di Scienze della Terra, Università degli Studi di Firenze, Florence, Italy

Int J Earth Sci (Geol Rundsch)

1 3

Introduction

The Thrace basin is a complex system of depocentres located between the Rhodope–Strandja massifs to the north and west and the Biga Peninsula to the south. The south-eastern margin of the basin is now covered by the Marmara Sea and deformed by the North Anatolian Fault Zone (NAFZ). The Thrace basin is the largest and thick-est Tertiary sedimentary basin of the eastern Balkan region and constitutes an important hydrocarbon province (Turgut et al. 1991; Turgut and Eseller 2000; Siyako and Huvaz 2007). Most of the Thrace basin-fill deposits range from the Early Eocene (Ypresian) to the Late Oligocene; maxi-mum total thickness, including the Neogene-Quaternary succession, reaches 9,000 m (Yıldız et al. 1997; Turgut and Eseller 2000; Siyako and Huvaz 2007). The older part of the basin fills crops out extensively along the basin mar-gins, but it is covered by Plio-Quaternary deposits in the basin centre. Volumetrically, most of the Eocene–Oligocene sedimentary succession is made of basin-plain turbidites (Aksoy 1987; Turgut et al. 1991). Sedimentation along the basin margins was characterised by carbonate deposits with nummulitids during the Eocene and by deltaic bodies pro-grading towards the basin centre in the Oligocene (Sümen-gen et al. 1987; Sümengen and Terlemez 1991). The west-ern margin of the basin was characterised already in the Eocene by a series of coarse-grained fan-deltas prograding eastward and feeding the basin-plain turbidites. The south-ern basin margin features abundant ophiolitic and carbon-ate olistostromes emplaced during the Eocene–Oligocene.

The Thrace basin was long interpreted as a forearc basin which developed in a context of northward subduction (Görür and Okay 1996). This interpretation was challenged by more recent data emphasising the lack of both a coeval magmatic arc and a subduction complex associated with the basin. All these elements—along with the correspondence between subsidence pulses in the basin and lithospheric stretching in the metamorphic core complexes of southern Bulgaria and the northern Aegean region—indicate instead that the Thrace basin was likely the result of (1) post-orogenic collapse after the continental collision related to the closure of the Vardar Ocean and/or (2) upper-plate extension related to slab retreat of the Pindos remnant ocean (see Cavazza et al. 2014, for a review). An increasing number of studies, combining kine-matic framework, regional geochronological and stratigraphic data, refine the definition of the Thrace depocentres as a hanging-wall supradetachment basin evolving into a strike-slip basin from the late Miocene to the present (e.g. Dinter et al. 1995; Beccaletto et al. 2007; Kilias et al. 2013).

Largely on the basis of compositional data, different authors determined the main provenance signatures and sediment dispersal pathways for the western (NE Greece), north-western (SE Bulgaria) and southern (NW Turkey)

sub-depocentres of the Thrace basin (e.g. Meinhold and BouDagher-Fadel 2010; Caracciolo et al. 2011, 2012, 2013; d’Atri et al. 2012; Maravelis et al. 2007; Maravelis and Zeli-lidis 2010, 2013; Cavazza et al. 2014). Three main sources are generally identified as the main contributors to the infill of the Thrace basin: (1) the Rhodope–Strandja massifs (mostly for W and NW Thrace and the main depocentre), (2) the Circum-Rhodope Belt (CRB; W and (?)SW Thrace) and (3) the Biga (?intra-Pontide) and Izmir–Ankara suture zones (S Thrace). Furthermore, an extensive volcanic source was also active, delivering huge quantities of volcaniclastic debris throughout the Thrace basin between the late Eocene and the Miocene. Detrital ophiolitic material has been docu-mented almost exclusively along the southern margin of the basin (Limnos/Lemnos, Samothraki, Gökçeada and Gal-lipoli Peninsula; Fig. 1). Its origin has been recently debated (cf. Maravelis and Zelilidis 2013; Caracciolo et al. 2012). There are different possible sources for these types of grains: (1) the meta-ophiolitic complex of the Circum-Rho-dope Belt, (2) the ophiolitic material exposed in the Biga Peninsula of NW Turkey, (3) the peridotite bodies of the Izmir–Ankara–Erzincan (IAESZ) or intra-Pontides suture zones and (4) the Lesvos “ophiolite”. Mainly based on pal-aeocurrents and compositional data, Maravelis and Zelilidis (2010, 2013) suggested that the ophiolite source for the turbidite succession exposed in the island of Limnos was a southern outer arc ridge identified in the island of Les-vos. From petrographic evidences, Caracciolo et al. (2011, 2012) proposed that Limnos and the western Thrace basin (Evros sub-basin, NE Greece) were supplied by a common source, the CRB, delivering low-grade metamorphic mate-rial to the western sector and meta-ophiolitic detritus to the southern sector (Limnos). d’Atri et al. (2012) stated that the Izmir–Ankara and/or Biga (?intra-Pontide) suture zones were the likely source for detrital Cr-spinel and ophiolite debris of the sediments exposed in the Gallipoli Peninsula of NW Turkey. The problem is tackled in the present study through a thorough characterisation of the geology of NW Turkey coupled with information obtained from the analy-sis of detrital heavy minerals. The chemistry of Cr-spinels of the Thrace basin constrains both the source for this type of sediments and Palaeogene palaeogeographic/geodynamic reconstructions of the SE European margin. Addressing the questions to be answered correctly is fundamental and only possible by considering the Thrace basin and the Biga Peninsula according to geological boundaries rather than to political borders.

Geological background

The integration of surface and subsurface data shows that the sedimentary fill of the northern, western and central

Int J Earth Sci (Geol Rundsch)

1 3

portions of the Thrace basin disconformably overlies the basement complexes of the Rhodope–Strandja massifs. The contact between Rhodopian basement and the base of the Thrace basin sedimentary fill is traceable southward in out-crop to the northern shores of the Gulf of Saros (Okay et al.

2010). Due to the effects of extensive strike-slip tectonism, the southern boundary (and basement) of the Thrace basin is less well defined. Nonetheless, progressively thinner Eocene sedimentary rocks extend from the Gallipoli Pen-insula into the northern part of the Biga Peninsula (Siyako

Fig. 1 a Schematic geological map of the Aegean and surrounding regions showing the main geotectonic units. b Geological map show-ing Palaeozoic–Mesozoic basement rocks (undifferentiated) and the

Thrace basin sedimentary cover of the Biga Peninsula (NW Turkey) and Rhodopian region (NE Greece, SW Bulgaria) (after Cavazza et al. 2014)

Int J Earth Sci (Geol Rundsch)

1 3

et al. 1989) where—before pinching out—they overlie again metamorphic rocks attributed to the Rhodopian domain by Okay and Satir (2000a), Beccaletto and Jenny (2004), Beccaletto et al. (2005), and Bonev and Beccaletto (2007).

In the following paragraphs, a synthetic description of the main tectono-stratigraphic units exposed in the Biga Peninsula (i.e. Çamlıca-Kemer metamorphics, Çetmi mélange, Ezine Group) (Fig. 2) is provided:

Çamlıca-Kemer complex—The Cretaceous (65–70 Ma) Çamlıca metamorphic complex is mostly composed of various types of schist intercalated with eclogitic amphi-bolite bands, marble and metabasite and is in tectonic con-tact with the mid-Cretaceous Denzigören ophiolite (Fig. 2). The Kemer mica schists are lithologically and structur-ally comparable to the Çamlıca mica schists, cropping out south-westward (Beccaletto et al. 2007; Fig. 2). The differ-ences between the two portions of the complex stay in the

maximum metamorphic conditions reached prior to exhu-mation. The Kemer mica schists display medium-grade metamorphic conditions, whereas those exposed in Çamlıca are typical of high-pressure/low-temperature (HP/LT) met-amorphic signatures, indicating a deeper burial of the latter with respect to the former. The mica schists contain garnet, phengite, paragonite, albite, epidote, calcite, chlorite and titanite. Metabasites constitute a small part (<5 %) of the metamorphic sequence (Aygül et al. 2012). In the metaba-sites, the mineral assemblage is garnet, barroisite, epidote, albite, titanite, quartz, phengite and chlorite. The lack of any correspondent geological element of the Çamlıca meta-morphics in Anatolia suggests that its affinity must lie with the large area of crystalline rocks that constitute the Rho-dope Massif (Okay and Satir 2000a, b). A further evidence of their mutual origin is represented by matching zircon ages from metasediments of the CRB and Çamlıca–Kemer complex ranging between 330 and 270 Ma (cf. Meinhold

Fig. 2 Geological maps of the study areas. a Western and southern Thrace basin (modified after d’Atri et al. 2012). The North Anatolian Fault (NAF) defines the limit between samples from “South Ganos” and “North Ganos” (see text for description). b Gökçeada Island

(after Akartuna 1950). c Samothraki Island (after Heimann et al. 1972). d Limnos Island (after Innocenti et al. 2009). The colour ver-sion of this figure is available in the web version of this article

Int J Earth Sci (Geol Rundsch)

1 3

et al. 2008). In its southern part, the Çamlıca–Kemer com-plex is in tectonic contact with the Çetmi mélange.

Çetmi mélange—In its northern part, the mélange is in tectonic contact with the Çamlıca mica schists, whereas in the southern part, it is tectonically underlain by high-grade metamorphic rocks of the Kazdağ Massif (Fig. 2) through the late Oligocene low-angle detachment fault that formed the Kazdağ Core Complex (Duru et al. 2004; Beccaletto et al. 2005). According to data from hydrocarbon explora-tion, the Çetmi mélange may also form the basement of the southern Thrace basin (Görür and Okay 1996). Main lithol-ogies are metavolcanic (basaltic, andesitic and pyroclastics) rocks, Upper Triassic limestones, serpentinites and diverse types of sandstones. According to Beccaletto et al. (2005), the Çetmi mélange has little in common with the mélanges from the Izmir–Ankara and intra-Pontide sutures of north-western Turkey, thus avoiding any direct correlation.

The Ezine Group—It is an enigmatic geological ele-ment in the framework of the eastern Mediterranean geol-ogy. It is made of two unit, namely the Ezine Group s.s. and the Denizgören ophiolite (Fig. 2). The first consists largely of recrystallised Permian–Triassic carbonate rocks with a slight to pronounced (bottom-to-top of the succession) detri-tal nature, whereas ophiolites are mostly serpentinised har-zburgites overlying a metamorphic amphibolite sole (Bec-caletto and Jenny 2004). The origin of the Ezine Group and especially of the Denizgören ophiolite remains unclear. The Ezine Group can be correlated to contemporary sedimentary successions in Greece, deposited during the opening of the Maliac/Meliata Ocean (Beccaletto and Jenny 2004). On the other hand, the age of the Denizgören ophiolite (ca. 125 Ma, age of the correspondent amphibolitic sole, Beccaletto 2004) has no time equivalent in the whole peri-Aegean region (Beccaletto and Jenny 2004). Based on the present-day geo-graphic position, many authors proposed a direct correlation of the Denizgören ophiolite with those exposed in the island of Lesvos. However, the age gap between the two ophiolite bodies (more than 100 Ma) makes this hypothesis unlikely. Beccaletto and Jenny (2004) pointed out a Rhodopian affin-ity and consequent juxtaposition of the Denizgören ophiolite together with the Ezine Group onto the western end of the Sakarya zone in post-Barremian times.

The CRB in southern Thrace—Okay et al. (2010) report on the existence of fragments of the CRB in the Mecidiye area (northern shore of Saros Bay) where a recrystallised lime-stone, calc-schist and phyllite series are overlain by a middle Eocene sedimentary succession consisting of conglomerates and nummulitic limestone. The same pattern is documented in the western Thrace basin (Evros sub-basin, Caracciolo et al. 2011) and in the island of Samothraki (Meinhold and BouDagher-Fadel 2010). The southern extension of the metamorphic basement in the Saros Bay (Gallipoli Penin-sula) consists of slate and phyllite and has been correlated

by several authors with the CRB of Greece (Kopp 1969; Kauffmann et al. 1976; Bonev and Stampfli 2003; Okay et al. 2010). Detrital zircons from sediments of the CRB of Greece show a polymodal age distribution with a major population at 370–290 Ma and minor populations at 2,100–1,780, 655–545 and 520–450 Ma (Meinhold et al. 2010; Meinhold and Kostopolous 2013). Tunç et al. (2014), although in different proportions, have recently determined the same population density in sediments of the CRB of NW Turkey.

Available stratigraphic and sedimentological data for the southern Thrace basin

A summary of the stratigraphy and sedimentology of the southern Thrace basin fill can be found in Okay et al. (2010), d’Atri et al. (2012) and Cavazza et al. (2014). Innocenti et al. (2009) and Maravelis and Zelilidis (2010) described the stra-tigraphy of Limnos Island. The successions exposed in the study area range mostly from the middle Eocene (Lutetian) to the Oligocene (Fig. 3). Lower Eocene sedimentary rocks are exposed at the base of the Tayfur succession (Karaağac Formation). The middle–upper Eocene transition is charac-terised by continental siliciclastic sediments followed by widespread nummulitic limestones (Meinhold and BouD-agher-Fadel 2010; Caracciolo et al. 2011, 2013; d’Atri et al. 2012). On the island of Limnos, the sedimentary succession is not older than Priabonian (Innocenti et al. 2009). From the late Eocene (Priabonian), most of the successions record deep-marine sedimentation, including carbonate olistoliths and significant volcanoclastic debris. A general coarsening upward trend is documented for the entire southern Thrace basin through time (Maravelis and Zelilidis 2010; d’Atri et al. 2012; Cavazza et al. 2014). Available palaeocurrent measurements (Limnos: Maravelis and Zelilidis 2010; south-ern Thrace: d’Atri et al. 2012) indicate that only the most proximal facies show northward palaeocurrents whereas most measured palaeocurrent indicators show an eastward palaeoflow, most likely the result of gravity flow deflection. S140° to S120° palaeoflow deflections/reflections were doc-umented by Caracciolo et al. (2013) on the island of Lim-nos. All these elements argue for a complex basin palaeo-physiography between the Eocene and the Oligocene. The present-day configuration results from the interplay of the Aegean regional NE-SW extension and the dismembering of the southern Thrace basin caused by the NAFZ. Armijo et al. (1999) suggested a minimum of 85 km south-westward migration for the Biga Peninsula, so that the palaeo-southern Thrace margin would need to be placed at the same latitude of the Rhodope–Thrace area (Bonev and Beccaletto 2007). Furthermore, a 25° clockwise rotation has been documented for the island of Limnos (Kondopoulou 2000). All these ele-ments complicate the possible interpretation from any prov-enance study in this area.

Int J Earth Sci (Geol Rundsch)

1 3

Fingerprints of possible ophiolitic sources

CRB ophiolites

Dated as Early Triassic are metasedimentary rocks in struc-turally lower position in the eastern CRB section, i.e. mostly the Makri unit. Dated Early Triassic meta-ophiolites in the eastern CRB are not known. Overlying Evros ophiolites are unmetamorphosed to hydrothermally metamorphosed at the ocean floor. Certainly, the lower section of these ophi-olites is metamorphosed to greenschist facies conditions (e.g. Bonev et al. 2010). It is exposed on the Chalkidiki Peninsula to the west and tectonically overlies the east-ern Rhodopes, in NE Greece and SE Bulgaria to the east (Figs. 1, 2). The Samothraki mafic suite is an in situ mag-matic complex comprising gabbros, sparse dykes and basalt flows and pillows cut by late dolerite dykes (Koglin 2008; Koglin et al. 2009a). SHRIMP zircon geochronology of a gabbro yielded 159.9 ± 4.5 Ma (early Late Jurassic), thus precluding any correlation with the Lesvos ultramafic com-plex further south (253.1 ± 5.6 Ma; Latest Permian; Koglin et al. 2009b). Different authors (Biggazzi et al. 1989; Koglin 2008) obtained similar ages (~170–150 Ma) from gabbros

of the CRB exposed in the Evros region. Cr-spinel data for the Evros ophiolite suite are not available. Since differ-ent authors pointed out the mutual origin for the ophiolites exposed in the CRB (Bébien et al. 1987; Tsikouras and Hatzipanagiotou 1998; Magganas 2002; Bonev and Becca-letto 2007; Zachariadis 2007; Koglin et al. 2009a; Meinhold and Kostopolous 2013), we use as reference the detrital Cr-spinel data from the Chalkidiki Peninsula and from Samo-thraki included in Meinhold et al. (2009) and Meinhold and BouDagher-Fadel (2010), respectively. Cr-number (Cr#: 0.28–0.90) and Mg-number (Mg#: 0.12–0.77) contents are rather heterogeneous indicating a mixed (ultra)mafic source of highly depleted peridotites of mainly harzburgite and minor lherzolite composition (Fig. 4) of dominant supra-subduction zone (SSZ) origin (Fig. 5). High-grade meta-morphic-derived Cr-spinel grains also occur.

Lesvos ophiolites

The Lesvos ophiolite consists of thrust sheets of man-tle peridotite overriding a tectonic mélange of metasedi-ments, metabasalts and a few metagabbros (Koglin et al. 2009b). Based on mineral compositions of spinel lherzolite

Fig. 3 Stratigraphic sections of the southern and western Eocene–Oligocene Thrace basin fill with associated palaeoflow measurements (modi-fied after d’Atri et al. 2012; Cavazza et al. 2014)

Int J Earth Sci (Geol Rundsch)

1 3

and harzburgites, Koglin et al. (2009b) proposed that the mélange and ultramafic rocks are more typical of an incipi-ent continental rift setting rather than of a typical ophiolite suite. The Lesvos Cr-spinels show a peculiar composition,

exhibiting high Al and low Fe concentrations and moder-ate amounts of Cr and Mg. A typical signature for Lesvos Cr-spinel is given by low Cr# values (0.11–0.37). Mg# val-ues range from 0.57 to 0.77. Cr# and Mg# values (Fig. 4)

Fig. 4 Cr- and Mg-numbers (Pober and Faupl 1988) for detrital Cr-spinel from the Eocene–Oligocene sediments of the Thrace basin. Top-left diagram reports compositions of Cr-spinel from potential source rocks (see text for references)

Int J Earth Sci (Geol Rundsch)

1 3

Fig. 5 Al2O3 vs TiO2 diagram with Cr-spinel discrimination fields (Kamenetsky et al. 2001). Top-left diagram reports compositions of Cr-spinel from potential source rocks (see text for references). ARC island-arc magmas, OIB ocean-island basalts, MORB mid-ocean

ridge basalts, LIP large igneous province, SSZ supra-subduction zone. The colour version of this figure is available in the web version of this article

Int J Earth Sci (Geol Rundsch)

1 3

suggest an Alpine-type peridotite origin for this ultramafic suite. Low TiO2 and high Al2O3 concentrations indicate a MORB peridotite origin for minor depleted Cr-spinels in Lesvos (Fig. 5).

Sakarya ophiolites

The Late Cretaceous IAESZ ophiolites in north-central Turkey are remnants of the Neo-Tethys and tectonically override the Anatolide–Tauride platform to the south (Sarıfakıoğlu et al. 2009). Pillow lavas, radiolarites and pelagic limestones occur in an ophiolitic mélange in the suture zone. Sarıfakıoğlu et al. (2009) provide Cr-spinel data from peridotites and massive and nodular chromitites from the western portion of the IAESZ (i.e. Izmir-Ankara segment). Cr# and Mg# values are 0.72–0.83 and 0.42–0.68, respectively (Fig. 4), whereas TiO2 and Al2O3 are 0.05–0.22 and 8.62–13.97 wt%, respectively (Fig. 5). These values indicate a SSZ-boninite origin for the IAESZ ophiolites.

Geological affinities between the Rhodope Zone and the Biga Peninsula

Possible affinities between the Rhodope Zone and the Biga Peninsula can be found focusing on lithological aspects (as well as associated stacking patterns) and comparing availa-ble protolith zircon ages. Lithologically, the CRB in Greece and the Çamlıca–Kemer metamorphic complex of the Biga Peninsula present several analogies. The latter is composed mainly of mica schist, calc-schist and marble with minor metabasite and serpentinite. Lower-grade metapelitic phyl-lite and eclogite enclaves occur in the Çamlıca complex. A similar rock assemblage, but of slightly lower metamorphic grade, can be traced in the CRB of NE Greece (e.g. Ivanov and Kopp 1969; Papadopoulos et al. 1989; Magganas et al. 1991; Magganas 2002). Although in different proportions, probability density distributions of detrital zircon U–Pb ages of the CRB (Meinhold et al. 2010) and the Çamlıca-Kemer complex (Tunç et al. 2014) indicate mutual age pop-ulations at 580–560 and 370–290 Ma. Meinhold and Kost-opolous (2013) pointed out to a strong Rhodopean/Strandja affinity for the (meta-)sediments of the eastern CRB. Based on the above-mentioned evidences, we suggest here that the metasediments of the Çamlica–Kemer metamorphic com-plex correspond to the Triassic–Jurassic low- to medium-grade metamorphic rocks of the Makri unit in NE Greece but they experienced a higher metamorphic overprint.

There are some striking similarities between the Cetmi mélange and the mélange-like units of the eastern Rhodope Zone in NE Greece and SE Bulgaria, like the reworking of Triassic limestones, Middle Jurassic–Lower Cretaceous radiolarians, the absence of Jurassic–Cretaceous passive margin lithologies and a major Cenomanian transgression

(Beccaletto et al. 2005). This model agrees with the one recently proposed for the equivalent CRB in NE Greece by Meinhold and Kostopolous (2013).

Short review of previous provenance studies in Southern Thrace

Previous provenance studies for the islands of Limnos (Maravelis et al. 2007; Maravelis and Zelilidis 2010, 2013; Caracciolo et al. 2011, 2012), Gökçeada (d’Atri et al. 2012), Samothraki (Meinhold and BouDagher-Fadel 2010), the Gallipoli Peninsula (d’Atri et al. 2012), the Korudag-Ganos Mt. region (d’Atri et al. 2012) and westernmost Thrace (Caracciolo et al. 2011, 2012; d’Atri et al. 2012) report in one way or another the existence of two or three sources (according to the single authors) for the infill of the southern Thrace basin. Keeping apart the interpretation provided by single authors, it is worth to note that sandstones exposed along the entire southern margin contain the same types of grains, rock and lithic fragments, and heavy minerals (although in variable amounts). Petrographically, they mostly show a quart-zolithic composition, with the feldspar content very lim-ited and, in most cases, related to volcanic input. Based on textural evidences and proportions of volcanic grains types, Caracciolo et al. (2011, 2012) suggested a common intermediate volcanic source for the succession exposed on Limnos Island and in the Evros sub-basin (Alexan-droupolis area) providing mostly vitric and felsitic neo-volcanic grains and appreciable intermediate to basic vol-canic rock fragments. d’Atri et al. (2012) described the same types of volcanic textures, with the basic/interme-diate textures (lathwork and microlithic) and neovolcanic plagioclases dominating the felsic component. Granitic/gneissic lithotypes are the most representative coarse-grained felsic rock fragments. The most important detrital elements consist of fine-grained lithics, reflecting a domi-nant supply from low- to medium-grade metasedimentary (slate, phyllite, mica schist) and meta-ophiolitic (diabase, serpentinite, glaucophane schist, chert) suites of the CRB. Same types of lithic grains were recognised by d’Atri et al. (2012) in the successions of Gökçeada and in those exposed on the Gallipoli Peninsula. Variable amounts of carbonate and siliciclastic sedimentary lithics were docu-mented. The petrostratigraphic evolution of the southern and western Thrace basin can be summarised as follows (for details, see Caracciolo et al. 2011; d’Atri et al. 2012; Cavazza et al. 2014).

Early-Middle Eocene

Sandstones from the south-western margin display a heterogeneous composition varying from quartzolithic

Int J Earth Sci (Geol Rundsch)

1 3

(Qm53F19Lt28) to feldspathic litharenites (Qm21F30Lt49) petrofacies in the Evros sub-basin and in the Gallipoli area. Most lithic fragments are from epimetamorphics of the CRB (Caracciolo et al. 2011; d’Atri et al. 2012; Cavazza et al. 2014) and from the ophiolite terranes of the Biga Pen-insula (d’Atri et al. 2012; Cavazza et al. 2014). Increasing contributions through time from the Upper Tectonic Unit (UTU) of the Rhodope Massif are documented (Caracciolo et al. 2011).

Late Eocene–Oligocene

Priabonian–Rupelian sandstones reflect a massive delivery of high-grade metamorphic detritus from the Rhodope Massif in the western sector, determining a quartzofeldspathic composition (Qm46F46Lt8). A vol-caniclastic petrofacies (Qm8F37Lt55) developed synchro-nously (late Eocene–early Oligocene) and is interbedded with the quartzofeldspathic material. Sandstones from the southern margin are arkosic litharenites (Qm13F48Lt39) including dominant neovolcanic material and intrabasinal carbonate grains diluting the epimetamorphic and ophi-olitic signal.

Sampling and methods

Forty-seven sandstone samples were selected for opti-cal heavy mineral analysis, covering the entire southern Thrace margin. For comparison, a set of samples from the Evros sub-basin (Alexandroupolis) has been included. Heavy mineral data from the Turkish portion of the south-ern margin were taken from d’Atri (2010). Heavy min-eral separates were obtained from the 63–125 µm grain-size fraction through standard separation technique using sodium polytungstate (ρ = 2.89 g/cm3). Detrital Cr-spi-nels from 11 selected samples were embedded in epoxy resin and polished for microprobe analysis. In total, 325 grains were chosen from the bottom and top of four dif-ferent formations (four samples from Limnos Island, two from Sarköy, three from Tayfur, two from Gökçeada). Data from the CRB of the Chalkidiki Peninsula (4 sam-ples, 202 grains) are from Meinhold et al. (2009). Data from Samothraki (3 samples, 202 grains) are from Mein-hold and BouDagher-Fadel (2010). Chemical data were obtained using a JEOL electron microprobe (JXA-8900 RL) equipped with five wavelength dispersive spectrom-eters operating at the Geoscience Centre Göttingen. Oper-ating conditions were 20 kV accelerating voltage with a beam current of 20 nA and a beam diameter of 3 µm to determine Mg, Al, Cr, Fe, V, Si, Ti, Mn, Ni and Zn concentrations.

Analytical results

Heavy mineral suites

Heavy mineral analyses were performed aiming at identify-ing possible contrasts or variations in heavy mineral suites that can be accounted to sediment provenance. Results are reported in Table 1 (as supplementary material). Heavy mineral abundances consist of high proportions of zircon, tourmaline and rutile (ZTR average 48 %), with a domi-nance of zircon and subordinate amounts of tourmaline and rutile. When present, Cr-spinel and garnet average at ca. 20 and 18 %, respectively. Subordinate quantities were obtained for titanite, epidote, allanite, anatase, pyroxene, chloritoid, staurolite, kyanite, glaucophane, amphibole, apatite and monazite. In general, types and association of occurring detrital heavy minerals (zircon, garnet, rutile, allanite, titanite, epidote, staurolite) point to a provenance dominated by low- to medium-grade metamorphic terranes, including rocks that experienced blueschist-facies meta-morphism (presence of glaucophane). The massive occur-rence of Cr-spinel clearly points out for a key role played by an ultramafic, (meta)ophiolitic, source. Clinopyroxene and clinoamphibole are related to volcanic sources. Gar-net, rutile, zircon and Cr-spinel indices (GZi, RuZi, CZi and CGi; Fig. 6) have been calculated according to Mor-ton and Hallsworth (1994, 1999) to further investigate the information hidden in the heavy mineral assemblage. Sam-ples from north of the Ganos segment of the north Anato-lian Fault (Fig. 2) are normally characterised by similar quantities of rutile, zircon and garnet and low amounts of Cr-spinel (except in the Korudağ succession) indicating a common provenance from low- and medium-grade meta-morphic terrains (Fig. 6, GZi vs RuZi). Samples from south of the Ganos fault show a higher variability. In fact, Ypre-sian to Bartonian (lower–middle Eocene) samples from the Karaağac and Fıçıtepe formations display high contents in zircon with negligible amounts of garnet, rutile and Cr-spi-nel. Middle Eocene to Oligocene sediments from Sarköy, Tayfur, Gökçeada and Limnos record increasing values of RuZi, GZi, CZi and CGi indexes, indicating the progressive influence of a new source in the basin.

The best way to investigate the information hidden in the heavy mineral assemblage is to treat them as a composi-tion. In order to avoid the non-negativity and constant-sum constraints on compositional variables, we performed the centred log-ratio transformation as introduced by Aitchison (1982, 1986, 1997). We applied principal component anal-ysis (PCA) for a set of orthogonal linear combinations of the variables (or in geometric terms, a rotation of the axes of the sample space), such that the new variables (or the projections of the data set onto the new axes) are mutually

Int J Earth Sci (Geol Rundsch)

1 3

uncorrelated. These new variables are called principal com-ponents (PC), and by construction they are obtained in decreasing order of variance. This makes the first few PC to be representative of the structural variability of the data set, leaving for the remaining PC just the random noise. One of the more effective ways to graphically illustrate com-position and principal components is the biplot. A more detailed guide about how to interpret biplots can be found in Aitchison (1997) and Caracciolo et al. (2012). At first glance, there are no clear correlations (expressed by the

angle between the variables, with 0° meaning dependence, 90° independence, 180° inverse proportionally) between heavy mineral species in the biplot (Fig. 7). A more care-ful analysis allows recognising a group of variables repre-sented by allanite + other epidotes + titanite. This group is placed opposite to Cr-spinel, pointing to an inverse dependence for this group of variables. Zircon, tourmaline, rutile and garnet do not show appreciable correlating ele-ments and place in between. Further information can be obtained by plotting sub-compositions, obtained by plot-ting two “depending” variables and a third independent one in a ternary plot. Results are reported in Fig. 9. All of the four selected ternary plots indicate a clear compositional difference between successions exposed to the north of the Ganos fault with those to the south. This difference is determined by the Zr–Rt–Grt association for the N Ganos and by the Cr-Sp–Grt–Rt–Ep association for the south-ern Ganos depozones. Samples from north of the Ganos fault, in particular those from the Korudağ and Kesan stratigraphic sections, reflect a mixed provenance from ultramafic, low- and medium-grade metamorphic sources, whereas those from Alexandroupolis do not show compo-sitional signatures indicating inputs from ultramafic rocks. Samples from south of the Ganos fault are characterised by dominant ultramafic inputs with important contributions from medium-grade metamorphic rocks and subordinate low-grade metamorphic sources.

Detrital Cr-spinel chemistry

The most effective way to obtain meaningful host-rock indi-cations from Cr-spinel chemistry is to measure elements that present a high variability, as Cr# (Cr/Cr + Al), Mg# (Mg/Mg + Fe2+), TiO2 and Fe3+# (Fe3+/Cr + Al + Fe3+) (e.g. Dick and Bullen 1984; Pober and Faupl 1988; Barnes and Roeder 2001; Kamenetsky et al. 2001). These vari-ables have proven useful to discriminate detrital Cr-spinel too (e.g. Pober and Faupl 1988; Lenaz et al. 2000; Lužar-Oberiter et al. 2009) and Cr# is generally used as an indica-tor of the degree of partial melting since it is strongly par-titioned into the solid phase. Mg# is generally controlled by the Cr and Al activity in spinels and generally decreases with increasing Cr# (Dick and Bullen 1984). Detrital Cr-spinels from the southern margin of the Thrace basin dis-play a wide range of Cr# and #Mg values. Most of the ana-lysed grains fall in the fields accounted by Pober and Faupl (1988) for lherzolites and harzburgites (Fig. 4). Although showing a wide distribution, Cr-spinel compositional trends are very similar in each of the selected stratigraphic successions with no appreciable differences according to stratigraphic age. Cr# and Mg# values are 0.16–0.93 and 0.1–0.8, respectively, suggesting a mixed ultramafic source from lherzolites to harzburgites. Furthermore, a few spinel

Fig. 6 Provenance sensitive heavy mineral indices determined on the 63–125 µm fraction of sandstone samples (calculated according to Morton and Hallsworth 1994, 1999). GZi = % garnet in total/gar-net plus zircon, RZi = % rutile in total/rutile plus zircon, CZi = % chrome spinel in total/chrome spinel plus zircon, CGi = % chrome spinel in total/garnet plus chrome spinel. Numbers beside symbols indicate the stratigraphic position within the corresponding succes-sion of analysed samples to facilitate the recognition of heavy min-eral trends. The colour version of this figure is available in the web version of this article

Int J Earth Sci (Geol Rundsch)

1 3

grains have very low values of Fe3+# (0.10–0.79) pointing to a metamorphic origin as being derived from high-grade metamorphic rocks (Barnes and Roeder 2001). Samples from the island of Limnos (Fig. 4) display the highest Cr# variability (0.16–0.85), with <5 % of them plotting below 0.35, the lower lherzolite field and Mg# ranging between 0.36 and 0.80. Most of the analysed grains show composi-tions of Cr-spinel from harzburgites and lherzolites. Apart from the low Cr# tail, the same pattern can be described for the successions of Samothraki, Gökçeada and Sarköy. Appreciable amounts of detrital Cr-spinels from Samoth-raki match the metamorphic spinel composition (Meinhold and BouDagher-Fadel 2010). The Tayfur succession resem-bles the same pattern described for the island of Limnos. According to Kamenetsky et al. (2001), residual spinels are mostly located in the middle of the field for the supra-subduction zone peridotites. In minor amounts, magmatic spinels are recognised between the MORB and arc basalt regions, whereas a negligible number of crystals fall within the MORB peridotite field (Fig. 5). The latter, having Cr# < 0.35, represent only 3.95 % of the analysed grains and display low TiO2 and high Al2O3, falling within the lherzolite field.

Discussion

Provenance deduced from heavy mineral analysis

In the studied samples, the following heavy minerals were recognised (in order of decreasing abundance): zir-con > Cr-spinel > garnet > rutile > epidote > titanite > tour-maline > allanite and common metamorphic-related min-erals as staurolite, kyanite and glaucophane. Amphibole,

monazite and pyroxenes rarely occur. The garnet/epidote/titanite/rutile/allanite association is well documented in the garnet-mica schists and metabasites exposed in the Kemer metamorphic complex (Aygül et al. 2012). Abundant Cr-spinel is attributed to the presence of numerous ophiolitic/ultramafic rocks surrounding the southern margin of the Thrace basin. Coupling the information from heavy miner-als (d’Atri 2010; d’Atri et al. 2012; this study) with those from sandstone framework petrology (Caracciolo et al. 2011; d’Atri et al. 2012; Cavazza et al. 2014), high con-centrations of zircon are attributed to widespread low-grade metamorphic rocks (phyllite and fine-grained schists) likely those from the Makri unit of the CRB and the Çamlıca complex. Compositional signatures point out for complex dispersal pathways delivering, in different proportions at different times, material from all of the accounted sources of sediment.

Early-middle Eocene (Ypresian-Lutetian)

Ypresian sediments from the Karağac Formation display very high zircon contents at the base, evolving upsec-tion to rutile + garnet (RuZi/GZi, Fig. 6) and Cr-spi-nel + Garnet (CZi/CGi, Figs. 6, 8) rich compositions. In terms of contributions, this suggests that the first input of sediments in southern Thrace has to be assigned to the low-grade metamorphic rocks of the CRB/Çamlica-Kemer complex, with increasing ophiolitic and medium-grade metamorphic material being delivered through time. The continental/deltaic nature of sediments and north-directed palaeoflows (Fig. 3) allow placing the eroded terranes immediately to the south of the Tayfur depo-centre (Fig. 9). Lutetian sediments south of the Ganos fault (Gökçeada, Fiçitepe) were mostly derived from the

Fig. 7 Biplot diagram showing the relationships among most representative detrital heavy minerals. Ep epidote, Grt garnet, All allanite, Tt titanite, Tur tourmaline, Rt rutile, Zr zircon, Cr-Sp chrome spinel, Pc principal component. The colour version of this figure is available in the web version of this article

Int J Earth Sci (Geol Rundsch)

1 3

erosion of ophiolitic terranes with increasing inputs from low- to medium-grade metamorphic sources through time. According to palaeoflows measurements (Fig. 3), ophiolitic terranes feeding the Lutetian basin need to be placed at the western margin of the basin (Fig. 9a). North of the Ganos fault (Alexandroupolis and Ganos), the sedi-mentation was mostly controlled by the erosion of the low-grade metamorphic rocks of the CRB in the western sector (Alexandroupolis), and by that of ultramafic rocks in the eastern part (Ganos).

Late Eocene–Oligocene (Priabonian-Chattian)

The relative abundance of heavy minerals from Priabonian and Oligocene sediments suggests a major role exerted by the metapelitic sources throughout the southern Thrace basin. Samples from Sarköy display a mixed composi-tion determined from the erosion of both metapelitic and ultramafic lithologies, likely the Kemer complex and the Çetmi mélange. However, the relative enrichment of gar-net and Cr-spinel with respect to rutile and zircon indi-cates an evolution from a mixed delivery of ultramafic and low-grade metamorphic rocks to the erosion of mixed

ultramafic and medium-grade metamorphic lithologies through time. Samples from Tayfur and Gökçeada docu-ment an increasing contribution from metapelitic rocks of the CRB/Kemer–Çamlica complex. Such higher concen-trations of zircon in samples from Tayfur point to a pos-sible sediment recycling beside a supply from low-grade metapelitic rocks supplied from the south (Figs. 3, 9). Apart from the higher amounts of zircon detected in the Tayfur succession, the same pattern can be described for the upper part of the succession exposed in Gökçeada. The increasing contribution from low-grade metapelitic rocks is in this case highlighted by the increase in the zircon–garnet association (Figs. 7, 9). Both heavy minerals evo-lutionary patterns and palaeoflows directions support the hypothesis of a similar evolution for these two depocen-tres. Priabonian sediments from the island of Limnos were mostly derived from a mix of dominant ultramafic source and zircon–garnet-rich metamorphic rocks. Increasing rutile and garnet contents in Oligocene sediments suggest a progressive influence of medium-to high-grade metamor-phic terranes placed to the SE and E (according to a mini-mum of 25° rotational restoration). Samples from the north of the Ganos fault display mixed compositional signatures

Fig. 8 Ternary plots representing sub-compositions deduced from the biplot figure. Ep epidote, Grt garnet, Rt rutile, Zr zircon, Cr-Sp chrome spinel. Grey areas represent zones of mixed provenance. Off-

set ternary plots represent most representative heavy mineral trends within sampled successions. The colour version of this figure is avail-able in the web version of this article

Int J Earth Sci (Geol Rundsch)

1 3

according to a provenance from low- to medium-grade metamorphic and ultramafic sources. Sediments from the Ganos Mountains and Korudağ reflect increasing contents in garnet and zircon and a corresponding decrease in Cr-spinel, pointing to a dominant role exerted by metamor-phic terranes between the late Eocene and Oligocene. Coeval samples from the Alexandroupolis basin display a garnet–rutile increase and a general zircon decrease according to a major contribution from Rhodopian terranes from the late Eocene.

Provenance deduced from Cr-spinel chemistry

Cr-spinel is considered to have a considerable chemical and mechanical stability (e.g. Morton and Hallsworth 1994, 1999). Consequently, we assume that possible modifications of its chemistry induced by chemical weathering are negli-gible. On the other hand, its survivability potential to low-T metamorphism is extremely high due to its pristine forming conditions. Thus, any possible temperature-induced further modifications (diagenesis, low-temperature metamorphism)

Fig. 9 Palaeogeographic reconstruction for the Eocene–Oligocene showing basement highs to the north and south supplying detritus to the southern Thrace basin. The colour version of this figure is available in the web version of this article

Int J Earth Sci (Geol Rundsch)

1 3

can be excluded, and the composition of detrital Cr-spinel can be taken as representative of its source. Careful analysis of the Cr# vs Mg# and Al2O3 vs TiO2 plots provides useful insights on the provenance of Cr-spinels. If we compare Cr-spinel compositions from possible source rocks, it is clear that there is a remarkable difference between Lesvos and the remaining available data (Figs. 4, 5). Peridotites from Sakaryan ophiolites display a different composition, with a group of Cr-spinels from chromitites plotting slightly outside of the harzburgites field. Conversely, Sakaryan harzburgites and the CRB peridotites group plot in the harzburgites field and show a wide distribution. The low Cr# values (coupled with the minor depletion) of Cr-spinels from Lesvos allow excluding the latter among the possible sources for the ophi-olitic material occurring in the southern Thrace sediments. A minor exception could be represented by Cr-spinels from the island of Limnos displaying Cr# < 0.35 in <3 % of the ana-lysed grains. Excluding Lesvos as the source candidate, it is highly probable that the southern portion of the Thrace basin received its (ultra)mafic detritus from Vardarian ophiolites with similarities to the Chalkidiki, Evros and Samothraki ophiolite complexes, as these complexes comprise various types of mafic and/or ultramafic rocks of mid-ocean ridge, island-arc and supra-subduction zone affinities (Tsikouras et al. 1990; Tsikouras 1992; Tsikouras and Hatzipanagiotou 1998; Magganas 2002; Pe-Piper and Piper 2002; Zachariadis 2007; Bonev and Stampfli 2009; Koglin 2008; Koglin et al. 2009a; Meinhold and BouDagher-Fadel 2010). Possible con-tributions from the Sakarya Zone cannot be excluded from detrital Cr-spinel chemistry. A review of available geody-namic and geological data can shed light on the problem.

Geodynamic affinities between the Rhodope Zone and the Biga Peninsula

There are several geodynamic features indicating a mutual tectono-sedimentary evolution of the Thrace basin margins, likely the Rhodope s.l. and metamorphic rocks exposed on the Biga Peninsula. All possible considerations necessarily need to take into account the position of the Vardar subduc-tion zone, locally located underneath the Rhodope during the Maastrichtian–Palaeocene and the southward shift of the subduction zone already during the Eocene (e.g. Bonev and Beccaletto 2007; Caracciolo et al. 2012; Jolivet et al. 2013). Key geodynamic signatures are as follows:

• Granitoid intrusions that followed the closure of the Vardar Ocean occurred synchronous in the eastern Rho-dopes and the Biga Peninsula at ca. 53 Ma (Bonev and Beccaletto 2007; Beccaletto et al. 2007). A second gen-eration of granitoid intrusions is recorded in both the eastern Rhodopes and Biga Peninsula between 32 and 21 Ma (Del Moro et al. 1988).

• NE–SW- and NNE–SSW-directed Early Eocene syn-orogenic extension is recorded in both the eastern Rho-dopes and Biga Peninsula (Krohe and Mposkos 2002; Bonev and Beccaletto 2007). The Miocene post-oro-genic extension is found in the Kazdağ Massif (Duru et al. 2004; Cavazza et al. 2009).

• Comparable crustal thickness (ca. 30 km) and uniform lithospheric vertical distribution between the Rhodope Zone and the Biga Peninsula (Papazachos and Scordi-lis 1998; Saunders et al. 1998; Bonev and Beccaletto 2007).

• According to palinspastic restoration from literature (Armijo et al. 1999), the Biga Peninsula needs to be shifted at least 85 km to the north-east. Between the Eocene and the Oligocene, the southern Thrace basin was structured as a graben (Karfakis and Doutsos 1995; Bonev and Beccaletto 2007) with its northern margin constituted by the Rhodope s.s. and the CRB (deliver-ing material into the northern Evros sub-basin) and the southern margin made up of CRB equivalent units (e.g. the Çamlica-Kemer complex) and Çetmi mélange. The eastward palaeoflows preserved in most of deep-water sediments suggest the presence of metapelitic palaeo-highs in the SW and central part of the basin (Fig. 9).

Conclusions

The review of geological data indicates that the pre-Eocene basement exposed on the Biga Peninsula needs to be con-sidered as part of the Rhodope metamorphic zone. North of the Kazdağ Massif, the Çetmi mélange and the Çamlıca-Kemer metamorphic complex exhibit the same rock asso-ciation documented by several authors in the Circum-Rho-dope Belt (CRB) of NE Greece and SE Bulgaria. The entire Rhodope/Biga Peninsula zone displays (1) similar granitoid intrusion ages, in particular those related to the closure of the Vardar Ocean (at ca. 53 Ma) and those at the Oli-gocene–Miocene transition (32–21 Ma), (2) similar direc-tion of kinematic indicators pointing to a common NE-SW synorogenic extension, (3) matching zircon populations between the CRB and the Çamlıca–Kemer metamorphic complex.

• The stratigraphic architecture of the southern Thrace basin presents several analogies along its margin, with most lower–middle Eocene sediments being continental/siliciclastic and passing up-section, during the Bartonian, into nummulitic limestone (Fig. 3). From the upper Eocene to Oligocene, the southern margin underwent complex geological conditions determined by the interplay of rapid subsidence and magmatism giving place to an extensive deep-water sedimentation.

Int J Earth Sci (Geol Rundsch)

1 3

• Heavy mineral assemblages indicate that the oldest exposed Ypresian sandstones from the Tayfur sub-depo-centre were mostly derived from the erosion of the Çam-lica complex. Lutetian and Priabonian sandstones record the progressive uplift and erosion of ultramafic terranes of the Biga Peninsula and increasing contributions from the metapelitic sources of the Kemer-Çamlica complex. Priabonian–Rupelian sandstones document the influ-ence of a volcanic source, delivering material throughout the whole southern Thrace basin. Oligocene sandstones reflect dominant contributions from garnet-rich mica schists and decreasing amounts of material from ultra-mafic sources. The successions north of the Ganos fault display a higher degree of mixing between metapelitic and ultramafic rocks of the CRB/Kemer–Çamlica and increasing influence from Rhodopian terranes.

• Detrital Cr-spinel compositions display very homog-enous distributions in all of the analysed successions reflecting a provenance from a mixed (ultra)mafic source of highly depleted peridotites, mostly harzbur-gites and subordinate lherzolites.

The possible inclusion of Lesvos among the possible sources of ultramafic material is rejected according to a clear difference in Cr# and Mg# numbers. Contributions from the Sakarya/intra-Pontide–IASZ are here disregarded for a number of reasons. Cr-spinel compositions from the Sakaryan terranes south of the Biga Peninsula cluster in narrow areas, especially those from chromitites. Looking at Al2O3 and TiO2, Sakaryan Cr-spinels cluster in the fields for island-arc (ARC) and low-Al2O3 supra-subduction zone (SSZ) peridotites that cannot be directly excluded as pos-sible sources. Being the Çamlica-Kemer complex and the Çetmi mélange already exposed at the time of deposition, and considered the depositional styles of analysed succes-sions, the presence of fluvial routing systems delivering material from the Sakaryan terranes (Karakaya Complex) seems unlikely. On the other hand, the distribution pattern and the depleted character of analysed Cr-spinel grains indicate supplies from (ultra)mafic detritus from Vardar-ian ophiolites, namely the Chalkidiki, Evros and Samoth-raki ophiolite complexes, as they comprise various types of mafic and/or ultramafic rocks of mid-ocean ridge, island-arc and supra-subduction zone affinities.

Acknowledgments Dr. F. Muto and Dr. R. Dominici from the Uni-versity of Calabria and Dr. N. Kolios are gratefully acknowledged for their help in the field. This research has been carried out within the MIUR-ex60% Projects (Relationships between Tectonic Accretion, Volcanism and Clastic Sedimentation within the Circum-Mediterra-nean Orogenic Belts, 2006–2013; Resp. S. Critelli), the 2009 MIUR-PRIN Project 2009PBA7FL_001 “The Thrace sedimentary basin (Eocene-Quaternary) in Greece and Bulgaria: stratigraphic deposi-tional architecture and sediment dispersal pathway within post-oro-genic basins.” (Resp. S. Critelli).

References

Aitchison J (1982) The statistical analysis of compositional data (with discussion). J R Stat Soc Ser B 44:139–177

Aitchison J (1986) The statistics analysis of compositional data. Chapman & Hall, London 416 pp

Aitchison J (1997) The one-hour course in compositional data anal-ysis or compositional data analysis is simple. In: Pawlowsky-Glahn V (ed) Proceedings of IAMG’97: the third annual confer-ence of the International Association for Mathematical Geology, International Center for Numerical Methods in Engineering (CIMNE), Barcelona (E), pp 3–35

Akartuna M (1950) Imroz Adasında Bazı Jeolojik Müs ahadeler. Tür-kiye Jeoloji Kurumu Bulteni 2(2):8–18 (in Turkish)

Aksoy Z (1987) Depositional environment of the sequences in the Barbaros–Kesan–Kadiköy–Gaziköy region (southern Thrace). In: Proceedings of the 7th petroleum congress of Turkey, Ankara, pp 292–311 (in Turkish)

Armijo R, Meyer B, Hubert A, Barka A (1999) Westward propagation of the north Anatolian fault into the northern Aegean: timing and kinematics. Geology 27:267–270

Aygül M, Gültekin T, Okay AI, Satir M, Meyer HP (2012) The Kemer Metamorphic Complex (NW Turkey): a subducted continental margin of the Sakarya zone. Turk J Earth Sci 21:19–35

Barnes SJ, Roeder PL (2001) The range of spinel composition in ter-restrial mafic and ultramafic rocks. J Petrol 42:2279–2302

Bébien J, Baroz F, Caperdi S, Venturelli G (1987) Magmatismes basiques assosiés à l’ouverture d’un basin marginal dans les Hellénides internes au Jurassique. Ofioliti 12:53–70

Beccaletto L, Jenny C (2004) Geology and correlation of the Ezine Zone: a Rhodope fragment in NW Turkey? Turk J Earth Sci 13:145–176

Beccaletto L, Bartolini A-C, Martini R, Hochuli PA, Kozur H (2005) Biostratigraphic data from the Çetmi melange, northwest Turkey: palaeogeographic and tectonic implications. Palaeogeogr Pale-oclimatol Palaeoecol 221:215–244

Beccaletto L, Bonev N, Bosch D, Bruguier O (2007) Record of a Paleogene syn-collisional extension in the north Aegean region: evidence from the Kemer micaschists (NW Turkey). Geol Mag 144:393–400

Biggazzi GA, Del Moro F, Innocenti F, Kyriakopoulos P, Manetti P, Papadopoulos P, Norilliri A, Magganas M (1989) The magmatic intrusive complex of Petrota, West Thrace: age and geodynamic significance. Geol Rhodopica 1:290–297

Bonev N, Beccaletto L (2007) From syn- to post-orogenic Tertiary extension in the north Aegean region: constraints on the kin-ematics in the eastern Rhodope–Thrace, Bulgaria–Greece and the Biga Peninsula, NW Turkey. In: Taymaz T, Yilmaz Y, Dilek Y (eds) The geodynamic of the Aegean and Anatolia. Geological Society, London, Special Publications, 291, pp 113–142

Bonev NG, Stampfli GM (2003) New structural and petrologic data on Mesozoic schists in the Rhodope (Bulgaria): geodynamic implications. Comptes Rendus Geosci 335:691–699

Bonev N, Stampfli G (2009) Gabbro, plagiogranite and associ-ated dykes in the supra- subduction zone Evros Ophiolites, NE Greece. Geol Mag 146:72–91

Bonev N, Spikings R, Mortiz R, Marchev P (2010) The effect of early Alpine thrusting in late-stage extensional tectonics: evidence from the Kulidzhik nappe and the Pelevun extensional allochthon in the Rhodope Massif, Bulgaria. Tectonophysics 488:256–281

Caracciolo L, Critelli S, Innocenti F, Kolios N, Manetti P (2011) Unraveling provenance from Eocene-Miocene sandstones of the Thrace Basin, NE Greece. Sedimentology 58:1988–2011

Caracciolo L, von Eynatten H, Tolosana-Delgado R, Critelli S, Man-etti P, Marchev P (2012) Petrological, geochemical, and statistical analysis of Eocene-Oligocene sandstones of the western Thrace Basin, Greece and Bulgaria. J Sediment Res 82:482–498

Int J Earth Sci (Geol Rundsch)

1 3

Caracciolo L, Critelli S, Innocenti F, Kolios N, Manetti P (2013) Reply to the discussion by Maravelis and Zelilidis on “Unrav-elling provenance from Eocene-Oligocene sandstones of the Thrace Basin, North-east Greece” by Caracciolo et al. (2011). Sedimentology 60:865–869

Cavazza W, Okay AI, Zattin M (2009) Rapid early-middle Miocene exhumation of the Kazdağ metamorphic core complex (Western Anatolia). Int J Earth Sci 98:1935–1947

Cavazza W, Caracciolo L, Critelli S, d’Atri A, Zuffa GG (2014) Pet-rostratigraphic evolution of the Thrace Basin (Bulgaria, Greece, Turkey) within the context of Eocene-Oligocene post-collisional evolution of the Vardar-Izmir-Ankara suture zone. Geodinamica Acta. doi:10.1080/09853111.2013.858943

d’Atri A (2010) Provenienza dei sedimenti arenitici nel Bacino di Tra-cia (Eo-Oligocene, Turchia nord-occidentale e Grecia nord-orien-tale). Ph.D. thesis, University of Bologna

d’Atri A, Zuffa GG, Cavazza W, Okay AI, Di Vincenzo G (2012) Detrital supply from subduction/accretion complexes to the Eocene-Oligocene post-collisional southern Thrace Basin (NW Turkey and NE Greece). Sediment Geol 243–244:117–129

Del Moro A, Innocenti F, Kyriakopoulos C, Manetti P, Papadopoulos P (1988) Tertiary granitoids from Thrace (northern Greece)–Sr isotopic and petrochemical data. Neues Jahrbuch für Mineralogie Abhandlungen 159:113–135

Dick HJB, Bullen T (1984) Chromian spinel as a petrogenetic indica-tor in abyssal and alpine-type peridotites and spatially associated lavas. Contrib Miner Petrol 86:54–76

Dinter DA, Macfarlane AM, Hames W, Isachsen C, Bowring S, Royden L (1995) U–Pb and 40Ar/39Ar geochronology of the Sym-volon granodiorite: implications for the thermal and structural evolution of the Rhodope metamorphic core complex, northeast-ern Greece. Tectonics 14:886–908

Duru M, Pehlivan S, Sentürk Y, Yavas F, Kar H (2004) New results on the lithostratigraphy of the Kazdağ Massif in Northwest Turkey. Turk J Earth Sci 13:177–186

Görür N, Okay AI (1996) A fore-arc origin for the Thrace Basin, NW Turkey. Geol Rundsch 85:662–668

Heimann KO, Lebkuchner H, Kretzler W (1972) Geological map of Greece, Samothraki sheet, scale 1:50000. Athens (Institute of Geology and Mineral Exploration)

Innocenti F, Manetti P, Mazzuoli R, Pertusati P, Fytikas M, Kolios N, Vougioukalakis GE, Androulakakis N, Critelli S, Caracciolo L (2009) Geological Map (Scale 1:50000) of Limnos Island (Greece): explanatory notes. Acta Vulcanol 20–21:87–97

Ivanov R, Kopp K-O (1969) Zur Tektonik des thrakischen Alttertiar-Beckens. Geotekton Forsch 31:117–132

Jolivet L, Faccenna C, Huet B, Labrousse L, Le Pourhiet L, Lacombe O, Lecomte E, Burov E, Denèle Y, Brun JP, Philippon M, Paul A, Salaün G, Karabulut H, Piromallo C, Monié P, Gueydan F, Okay AI, Oberhansli R, Pourteau A, Augier R, Gadenne L, Driussi O (2013) Aegean tectonics: strain localisation, slab tearing and trench retreat. Tectonophysics 597–598:1–33

Kamenetsky V, Crawford AJ, Meffre S (2001) Factors controlling chemistry of magmatic spinel: an empirical study of associated olivine, Cr-spinel and melt inclusions from primitive rocks. J Pet-rol 42:655–671

Karfakis J, Doutsos T (1995) Late orogenic evolution of the Circum-Rhodope Belt, Greece. Neues Jahrbuch für Geologie und Paläon-tologie, Monatsheft 5:305–319

Kauffmann G, Kockel F, Mollat H (1976) Notes on the stratigraphic and paleogeographic position of the Svoula Formation in the innermost zones of the Hellenides (Northern Greece). Bulletin de la Société Géologique de France 18:225–230

Kilias A, Falalakis G, Sfeikos A, Papadimitriou E, Vamvaka A, Gkar-laouni C (2013) The Thrace basin in the Rhodope province of NE

Greece—a tertiary supradetachment basin and its geodynamic implications. Tectonophysics 595–596:90–105

Koglin N (2008) Geochemistry, petrogenesis and tectonic setting of ophiolites and mafic-ultramafic complexes in the Northeastern Aegean Region: New trace element, isotopic and age constraints. Ph.D. thesis, University of Mainz, Germany, 136 pp

Koglin N, Kostopoulos D, Reischmann T (2009a) Geochemistry, petrogenesis and tectonic setting of the Samothraki mafic suite, NE Greece: trace-element, isotopic and zircon age constraints. Tectonophysics 473:53–68

Koglin N, Kostopoulos D, Reischmann T (2009b) The Lesvos mafic ultramafic complex, Greece: ophiolite or incipient rift? Lithos 108:243–261

Kondopoulou D (2000) Palaeomagnetism in Greece: Cenozoic and Mesozoic components and their geodynamic implications. Tec-tonophysics 326:131–151

Kopp KO (1969) Geologie Thrakiens VI: Der Coban Dag westlich Alexandroupolis. Geotekton Forsch 31:97–116

Krohe A, Mposkos E (2002) Multiple generations of extensional detachments in the Rhodope Mountains (northern Greece): evi-dence of episodic exhumation of high-pressure rocks. In: Blun-dell DJ, Neubauer F, von Quadt A (eds) The timing and location of major ore deposits in an evolving orogen. Geological Society, London, Special Publications 204, pp 151–178

Lenaz D, Kamenetsky VS, Crawford AJ, Princivalle F (2000) Melt inclusions in detrital spinel from the SE Alps (Italy-Slovenia): a new approach to provenance studies of sedimentary basins. Con-trib Mineral Petrol 139:748–758

Lužar-Oberiter B, Mikes T, von Eynatten H, Babić L (2009) Ophi-olitic detritus in Cretaceous clastic formations of the Dinarides (NW Croatia): evidence from Cr-spinel chemistry. Int J Earth Sci 98:1097–1108

Magganas AC (2002) Constraints on the petrogenesis of Evros ophi-olite extrusives, NE Greece. Lithos 65:165–182

Magganas A, Sideris C, Kokkinakis A (1991) Marginal basin-volcanic arc origin of metabasic rocks from the Circum-Rhodope Belt, Thrace, Greece. Mineral Petrol 44:232–252

Maravelis A, Zelilidis A (2010) Petrography and geochemistry of the late Eocene-early Oligocene submarine fans and shelf deposits on Lemnos Island, NE Greece: implications for provenance and tec-tonic setting. Geol J 45:412–433

Maravelis A, Zelilidis A (2013) Discussion: unravelling provenance from Eocene-Oligocene sandstones of the Thrace Basin, North-east Greece” by Caracciolo et al. (2011), Sedimentology, 58, 1988–2011. Sedimentology 60:860–864

Maravelis A, Konstantopoulos P, Pantopoulos G, Zelilidis A (2007) North Aegean sedimentary basin evolution during the late Eocene to early Oligocene based on sedimentological studies on Lemnos Island (NE Greece). Geol Carpath 58:455–464

Meinhold G, BouDagher-Fadel MK (2010) Geochemistry and biostratigraphy of Eocene sediments from Samothraki Island, NE Greece. Neues Jb Geol Paläontol Abh 256:17–38

Meinhold G, Kostopolous D (2013) The Circum-Rhodope Belt, north-ern Greece: age, provenance, and tectonic setting. Tectonophysics 595–596:55–68

Meinhold G, Reichsmann T, Kostopoulos D, Lehnert O, Matukov D, Sergeev S (2008) Provenance of sediments during subduction of Palaeotethys: detrital zircon ages and olistolith analysis in Palae-ozoic sediments from Chios Island, Greece. Palaeogeogr Palaeo-climatol Palaeocol 263:71–91

Meinhold G, Kostopoulos D, Reischmann T, Frei D, BouDagher-Fadel MK (2009) Geochemistry, provenance and stratigraphic age of metasedimentary rocks from the eastern Vardar suture zone, northern Greece. Palaeogeogr Palaeoclimatol Palaeoecol 277:199–225

Int J Earth Sci (Geol Rundsch)

1 3

Meinhold G, Reischmann T, Kostopoulos D, Frei D, Larionov AN (2010) Mineral chemical and geochronological constraints on the age and provenance of the eastern Circum-Rhodope Belt low-grade metasedimentary rocks, NE Greece. Sed Geol 229:207–223

Morton AC, Hallsworth CR (1994) Identifying provenance specific features of detrital heavy mineral assemblages in sandstones. Sed Geol 124:3–29

Morton AC, Hallsworth CR (1999) Processes controlling the com-position of heavy mineral assemblages in sandstones. Sed Geol 124:3–29

Okay AI, Satir M (2000a) Upper Cretaceous eclogite-facies metamor-phic rocks from the Biga Peninsula, Northwest Turkey. Turk J Earth Sci 9:47–56

Okay AI, Satir M (2000b) Coeval plutonism and metamorphism in a latest Oligocene metamorphic core complex in northwest Turkey. Geol Mag 137:495–516

Okay AI, Özcan E, Cavazza W, Okay N, Less G (2010) Basement types, Lower Eocene series, Upper Eocene olistostromes and the initiation of the southern Thrace Basin (NW Turkey). Turk J Earth Sci 19:1–25

Papadopoulos P, Arvanitides ND, Zanas I (1989) Some preliminary geological aspects on the Makri unit (phyllite series); peri-Rho-dopian zone. Geol Rhodopica 1:34–42

Papazachos CB, Scordilis EM (1998) Crustal structure of the Rho-dope and surrounding area obtained by non-linear inversion of P and S travel times and its tectonic implication. Acta Vulcanol 10:339–345

Pe-Piper G, Piper DJW (2002) The igneous rocks of Greece. The anatomy of an orogen. Beiträge zur regionalen Geologie der Erde 30:1–573

Pober E, Faupl P (1988) The chemistry of detrital chromian spinels and its implications for the geodynamic evolution of the Eastern Alps. Geol Rundsch 77:641–670

Sarıfakıoğlu E, Hayrettin Ö, Çolakoğlu A, Sayak H (2009) Petrol-ogy, mineral chemistry, and tectonomagmatic evolution of Late Cretaceous suprasubduction-zone ophiolites in the Izmir-Ankara-Erzincan suture zone, Turkey. Int Geol Rev 52:187–222

Saunders P, Priestley K, Taymaz T (1998) Variations in the crustal structure beneath western Turkey. Geophys J Int 134:373–398

Siyako M, Huvaz O (2007) Eocene stratigraphic evolution of the Thrace Basin, Turkey. Sed Geol 198:75–91

Siyako M, Bürkan KA, Okay AI (1989) Tertiary geology and hydro-carbon potential of the Biga and Gelibolu peninsula. Turk Assoc Pet Geol Bull 1(3):183–199 (in Turkish with English summary)

Sümengen M, Terlemez I (1991) Stratigraphy of the Eocene depos-its of southwestern Thrace region. Gen Dir Miner Res Explor J 113:17–30

Sümengen M, Terlemez I, Sentürk K, Karaköse C (1987) Stratigra-phy, sedimentology and tectonics of the Gelibolu Peninsula and southwestern Thrace Basin. Internal Report of the Maden Tetkik ve Arama Enstitüsü 8128, Ankara (in Turkish)

Tsikouras B (1992) The ophiolites of the island of Samothrace. Ph.D. thesis, University of Patras, Greece, 435 pp (in Greek with Eng-lish summary)

Tsikouras B, Hatzipanagiotou K (1998) Petrographic evolution of an ophiolite fragment in an ensialic marginal basin, northern Aegean (Samothraki Island, Greece). Eur J Mineral 10:551–567

Tsikouras B, Pe-Piper G, Hatzipanagiotou K (1990) A new date for an ophiolite of the northeastern margin of the Vardar zone, Samo-thraki, Greece. Neues Jahrbuch für Mineralogie Monatshefte 11:512–527

Tunç IO, Yiğitbas E, Sengün F, Wazeck J, Hofmann M, Linnemann U (2014) U-Pb zircon geochronology of northern metamorphic massifs in the Biga peninsula (NW Anatolia-Turkey): new data and a new approach to understand the tectonostratigraphy of the region. Geodin Acta. doi:10.1080/09853111.2013.877242

Turgut S, Eseller G (2000) Sequence stratigraphy, tectonics and depo-sitional history in eastern Thrace Basin, NW Turkey. Marine Pet-rol Geol 17:61–100

Turgut S, Türkaslan M, Perinçek D (1991) Evolution of the Thrace sedimentary basin and its hydrocarbon prospectivity. In: Spencer AM (ed) Generation, accumulation, and production of Europe’s hydrocarbons: European Association of Petroleum Geoscientists. Special Publication, London, pp 415–437

Yıldız A, Toker V, Sengüler I (1997) The nannoplankton biostratig-raphy of the middle Eocene–Oligocene units in southern Thrace Basin and the surface water temperature Variations. Turk Assoc Petrol Geol Bull 9:31–44

Zachariadis P (2007) Ophiolites of the eastern Vardar Zone, N. Greece. Ph.D. thesis, University of Mainz, Germany, 131 pp