Geochemistry and petrology of the Early Miocene lamproites and related volcanic rocks in the Thrace Basin, NW Anatolia

Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Contents lists available at ScienceDirect

Journal of Volcanology and Geothermal Research journal homepage: www.elsevier.com/locate/jvolgeores

Geochemistry and petrology of the Early Miocene lamproites and related volcanic rocks in the Thrace Basin, NW Anatolia Yalçın E. Ersoy a,⁎, Martin R. Palmer b, İbrahim Uysal c, İbrahim Gündoğan a a b c

Dokuz Eylül Üniversitesi, Mühendislik Fakültesi, Jeoloji Mühendisliği Bölümü, TR-35160 İzmir, Turkey School of Ocean and Earth Science, NOC, University of Southampton, European Way, Southampton SO14 3ZH, UK Karadeniz Teknik Üniversitesi, Jeoloji Mühendisliği Bölümü, TR-61080 Trabzon, Turkey

a r t i c l e

i n f o

Article history: Received 3 February 2014 Accepted 18 June 2014 Available online 16 July 2014 Keywords: NW Anatolia Thrace basin Miocene volcanism Ultrapotassic volcanism Leucite lamproite

a b s t r a c t The extensional Thrace basin (NW Anatolia) contains an association of early Miocene diopside–leucite–phlogopite (Doğanca) and diopside–phlogopite (Korucuköy) lamproites with Oligocene medium-K calc-alkaline andesites (Keşan volcanics), early Miocene shoshonitic rocks (Altınyazı trachyte) and middle Miocene Na-alkaline basalts (Beğendik basalts). The Doğanca lamproite (K2O = 5.1–5.5 wt.%; K/Na =2.78–2.89; MgO =11.4–11.8 wt.%) consists of olivine (Fo71–86), diopside (Al2O3 = 1.0–5.0, Na2O =0.2–0.6), phlogopite (TiO2 = 1.1–9.4, Al2O3 = 11.1–13.9), spinel (Mg# = 22.9–32.6; Cr# = 64–83.4), leucite, apatite, zircon, Fe–Ti-oxides and magnetite in a poikilitic sanidine matrix. The potassic volcanic units (lamproites and trachytes) in the region have similarly high Sr and low Nd isotopic compositions (87Sr/86Sr(i) = 0.70835–0.70873 and 143Nd/144Nd(i) = 0.51227–0.51232). The major and trace element compositions and Sr–Nd–Pb isotopic ratios of the shoshonitic, ultrapotassic and lamproitic units closely resemble those of other Mediterranean ultrapotassic lamproites (i.e., orogenic lamproites) from Italia, Serbia, Macedonia and western Anatolia. The Beğendik basalts show intraplate geochemical signatures with an Na-alkaline composition, an absence of Nb negative anomalies on primitive mantle-normalized multielement diagrams, as well as low Sr (~0.70416) and high Nd (0.51293) isotopic ratios; and include olivine (Fo72–84), diopside, spinel, Fe–Ti-oxides and magnetite. The Oligocene Keşan volcanics were emplaced in the earlier stages of extension in Thrace, and represent the typical volcanic products of post-collisional volcanism. The continental crust-like trace element abundances and isotopic compositions of the most primitive early Miocene ultrapotassic rocks (Mg# up to 74) indicate that their mantle sources were intensely contaminated by the continental material. By considering the geodynamic evolution of the region, including oceanic subduction, crustal accretion, crustal subduction and post-collisional extension, it is suggested that the mantle sources of the potassic volcanic units were most likely metasomatized by direct subduction of continental blocks during accretion and assemblage of various Alpine tectono-stratigraphic units. Overall, the magma production occurred in an extensional tectonic setting that controlled the core-complex formation and related basin development, with the middle Miocene Beğendik basalts being derived from asthenospheric sources during the late stages of extension. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Lamproites are mantle-derived volcanic rocks with ultrapotassic geochemical affinities (MgO and K2O N 3 wt.%; K2O/Na2O N2; K2O + Na2O/Al2O3 N 1; Foley et al., 1987), and are of particular importance to petrologists because they reflect the distinct geochemical features of the highly enriched mantle sources from which they originated (e.g., Nelson et al., 1986; Bergman, 1987; Foley et al., 1987; Prelević et al., 2008, 2013). These rocks are unusual in that they show both mantle- (high-MgO, Ni contents) and crust-like (high incompatible trace element and radiogenic Sr isotope compositions) geochemical ⁎ Corresponding author. Tel.: +90 232 301 73 46. E-mail address: [email protected] (Y.E. Ersoy).

http://dx.doi.org/10.1016/j.jvolgeores.2014.06.016 0377-0273/© 2014 Elsevier B.V. All rights reserved.

features. Those derived in post-orogenic extensional settings (orogenic lamproites) have most commonly been described from along the Alpine–Himalayan system in Spain (Benito et al., 1999; Duggen et al., 2005), Italy (Peccerillo, 1998; Avanzinelli et al., 2008), Serbia & Macedonia (Altherr et al., 2004; Prelević et al., 2005; Yanev et al., 2008), Turkey (Çoban and Flower, 2006; Ersoy and Helvacı, 2007; Akal, 2008; Prelević et al., 2012), and China (Ding et al., 2003; Gao et al., 2007; Zhao et al., 2009). Orogenic lamproites are generally associated with other calc-alkaline and potassic (or even ultrapotassic) rock types and are frequently interpreted as having originated from orogenic mantle sources that have been intensely metasomatized (enriched) during subduction and collision events throughout the closure of the Tethyan oceanic branches. The mechanism of the enrichment process is, however, still debated. Some authors favor long-term isolation of

144

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

the subducted sediments and/or accretionary complexes in mantle domains (e.g., Prelević et al., 2005; Peccerillo and Martinotti, 2006; Gao et al., 2007; Tommasini et al., 2011), while others suggest direct subduction of continental crust into the mantle during crustal accretion (e.g., Schreyer et al., 1987; Arnaud et al., 1992; Parkinson and Kohn, 2002; Zhao et al., 2009; Çoban et al., 2012; Ersoy et al., 2012; Prelević et al., 2013). Eastern Rhodope and northwestern Anatolia are part of the Alpine– Himalayan orogenic system. In these regions, post-orogenic extension led to the formation of high-K, calc-alkaline to shoshonitic and locallydeveloped ultrapotassic volcanic extrusives during late Paleogene to Miocene times (Aldanmaz et al., 2000; Marchev et al., 2004; Altunkaynak and Genç, 2008; Dhont et al., 2008; Kirchenbaur et al., 2012a). In both areas, a number of continental blocks were accreted along suture zones, via oceanic subduction. Continental collision and continental subduction then led to high- to ultrahigh-pressure metamorphism, followed by extensional processes that are typically associated with magma generation (Kostopoulos et al., 2000; Mposkos and Kostopoulos, 2001; Jahn-Awe et al., 2012; Kirchenbaur et al., 2012b; Ersoy and Palmer, 2013, and references therein). In the eastern Rhodopes, the magmatic activity started during the latest Eocene and continued up to the early Miocene (Christofides et al., 2004; Marchev et al., 2004; Kirchenbaur et al., 2012a), and in the northern part of W Anatolia, the magmatic activity commenced during the early Eocene (52–47 Ma) and continued up to the middle Miocene (Aldanmaz et al., 2000; Altunkaynak and Genç, 2008; Kürkçüoğlu et al., 2008; Gülmez et al., 2012). Although this pattern of magmatism has been linked to a general southward migration of Aegean magmatic activity in parallel to retreat of the Hellenic subduction zone (e.g., Fytikas et al., 1984; Ring et al., 2010), Ersoy and Palmer (2013) pointed out that this hypothesis conflicts with observations that Oligocene to Miocene magmatic activity also took place to the north of the early Eocene magmatic rocks. The Thrace basin is located between eastern Rhodope and western Anatolia, and is one of the largest extensional basins in the region. It began to develop during the Eocene, marking the beginning of extensional tectonics in the N Aegean. The sedimentary infill of the basin contains Oligocene–Miocene volcanic rocks, lying to the north of the Eocene magmatic belt in NW Anatolia, and thus precluding simple southward migration of magmatism with time. Oligocene medium-K calc-alkaline rocks, high-MgO, shoshonitic to ultrapotassic rocks, including lamproites, and Na-alkaline basaltic rocks occur in the basin fill. This association, may therefore link the eastern Rhodope Oligocene–Miocene magmatic belt to that of Western Anatolia. Here, we present the first detailed description of Miocene leucite lamproitic rocks in northwest Turkey, including whole-rock geochemical data (major and trace elements, and Sr–Nd–Pb isotopes) from the associated rocks from the Keşan region in the Thrace basin. 2. Geological setting and sample locations The northern part of the Aegean region (to the north of the Vardar and İzmir–Ankara suture zones), including the eastern Balkans, NW Anatolia and the Thrace Basin, lies in the Eastern Mediterranean orogenic domain, which was shaped by closure of the Tethyan oceans (e.g., Şengör and Yılmaz, 1981; Okay and Tüysüz, 1999; Stampfli, 2000). Pre-Tertiary basement units in this region comprise the Strandja Massif to the north, the Sakarya Zone of the Pontides to the south and the Rhodope Massif to the west, which constitute the dissected Rhodope–Pontide fragment (Fig. 1; e.g., Ricou et al., 1998; Okay et al., 2001). The Strandja Massif is composed of a Paleozoic basement and Triassic metasedimentary cover, and is bound by the Thrace fault zone in the south (e.g., Natalin et al., 2012; and references therein). The Sakarya Zone is composed of Paleozoic granitoids and Triassic subduction–accretion complexes, which are unconformably overlain by Jurassic and younger sediments (Okay et al., 1996). The Rhodope

Massif consists of pre-Alpine and Alpine nappes of continental and oceanic units, which were assembled during subduction and closure of the Vardar Ocean (e.g., Bonev and Stampfli, 2011). The NE part of the Rhodope Massif is separated from the late Cretaceous Srednogorie magmatic arc by the Maritza fault zone (Boccaletti et al., 1978; Georgiev et al., 2012). The NE part of the Rhodope and southern part of the Strandja massifs are covered by Tertiary sedimentary units of the Thrace Basin (Elmas, 2011; Kilias et al., 2013; Fig. 1). The Rhodope Massif was exhumed as a number of extensional metamorphic core complexes that comprise: (1) the Southern Rhodope Core Complex, (2) the central Rhodope Core Complex (Arda dome), and (3) the Kesebir and Biala Reka domes (e.g., Dinter and Royden, 1998; Bonev and Beccaletto, 2007; Brun and Sokoutis, 2007; Jahn-Awe et al., 2012). In the eastern Rhodopes, calc-alkaline to shoshonitic rocks (basalt to rhyolite) are located mainly in the Borovitsa and Momchilgrad-Arda volcanic areas, with radiometric ages ranging from late Eocene to Oligocene (~ 39–25 Ma; Dhont et al., 2008 and references therein). To the southeast of the Momchilgrad-Arda volcanic area, the Krumovgrad alkaline basaltic rocks (with intraplate geochemical features) were emplaced at 28–26 Ma (Marchev et al., 2004). The calc-alkaline felsic volcanism continues further south (the Evros volcanic area), with ages of ~ 33–19 Ma (Christofides et al., 2004). To the southeast of the Evros volcanic area, lie the andesitic to rhyolitic Hisarlıdağ volcanics (35.0 ± 0.9 Ma; Ercan et al., 1998). Overall, these ages indicate that volcanism in the region continued until the early Miocene in the south of eastern Rhodopes and Thrace. Large exposures of early Miocene volcanic rocks are also located on the islands of Samothrace (Vlahou et al., 2006) and Limnos (Pe-Piper et al., 2009), and the Biga Peninsula (Aldanmaz et al., 2000). Importantly, basaltic to andesitic volcanic rocks were also emplaced further south in the Biga Peninsula and south of the Marmara Sea during the Eocene (~52–37 Ma; Ercan et al., 1998; Altunkaynak and Genç, 2008; Kürkçüoğlu et al., 2008; Gülmez et al., 2012), arguing against simple southward migration of magmatism throughout the Aegean region. Finally, during the late Miocene, small Na-alkaline basaltic extrusives with OIB-type geochemical affinity were emplaced in the Biga Peninsula (Ezine alkaline basalts) and in the eastern part of the Thrace basin (Thrace alkaline basalts, Ercan et al., 1998; Aldanmaz et al., 2006). The study area (Fig. 2) is located in the S-SW part of the Thrace Basin. Overall, the basin is a triangular extensional basin, including Eocene to Pliocene sedimentary units which are up to 9 km thick (e.g., Elmas, 2011; Kilias et al., 2013 and references therein). Kilias et al. (2013) proposed that the basin developed as a supradetachment basin, related to exhumation of the metamorphic rocks of the Rhodope Massif during middle–late Eocene to Oligocene times. According to Elmas (2011), the stratigraphy of the study area begins with turbidites of the middle–late Eocene Keşan Formation, and continues upwards with fine-grained fluvial sediments of the late Eocene to early Oligocene Mezardere, followed by the late Oligocene to early Miocene Danişment formations. These units are then unconformably overlain by middle–late Miocene sedimentary rocks of alluvial, fluvial and lacustrine origin. Additionally, several volcanic units with small outcrops and distinct compositions were emplaced during the Oligocene–Miocene. To the south of Keşan, andesitic–dacitic volcanic rocks occur as dykes, lava flows and domes (of which one sample (K-20) has been analyzed in this study). These volcanic rocks yield a 26.2 ± 0.5 Ma K–Ar age (Ercan et al., 1998), and are thus coeval with the Hisarlıdağ volcanics that lie in the southwest of the Thrace Basin. A small volcanic unit lies to the north of the study area, close to Altınyazı village (Fig. 2), which crosscuts the sedimentary rocks of the Danişment Formation and yields 21.27 ± 0.24 and 18.47 ± 0.20 Ma K–Ar radiometric ages (Ercan et al., 1998). To the east of the study area, south of Doğanca village, a small volume lava dome, which contains large mica phenocrysts, cuts the lignite-bearing sedimentary units of the late Oligocene–early Miocene Danişment formation. On the basis of its petrographic and geochemical features, this unit is named here as the Doğanca lamproite.

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

145

Fig. 1. Geological map of eastern Rhodope, Thrace and NW Anatolia. Compiled from geological maps of Greece (IGME, 1983) and Turkey (MTA, 2002) with additions from Christofides et al. (2004), Schenker et al. (2012), Georgiev et al. (2008, 2012), Jahn-Awe et al. (2012), Elmas (2012) and Kilias et al. (2013). CRD: Central Rhodope Dome.

The age of this lamproite is likely early Miocene on the basis of its stratigraphic relationships. NW of Keşan, around Korucuköy, evolved volcanic rocks (the Korucuköy lamproite) were emplaced during the late Oligocene or early Miocene. Small volume basaltic extrusives also occur to the west of Beğendik (the Beğendik basalts), where they cut the Mezardere and Danişment formations and yield a 15.0 ± 0.3 Ma K–Ar age (Ercan et al., 1998). 3. Petrography and mineral chemistry The Keşan volcanics are composed of andesitic rocks with a porphyritic texture, including plagioclase, hornblende, minor clinopyroxene, apatite and opaque phases in a matrix of plagioclase microlites. The Korucuköy lamproite includes olivine (partly altered to iddingisite and resorbed by carbonate), clinopyroxene and phlogopite (Fig. 3a). The Altınyazı trachyte is composed of small crystals of clinopyroxene, phlogopite, apatite and magnetite embedded in a poikilitic sanidine matrix (Fig. 3b). Small, and volumetrically minor (b 0.1%), plagioclase crystals are also present, together with minor amounts of fan-shaped secondary zeolites within the feldspar matrix. One sample (K-3) from the Altınyazı trachyte differs from the others in having a pilotaxitic matrix containing altered olivine and clinopyroxene phenocrysts. The phlogopites within both the Korucuköy lamproite and the Altınyazı trachyte are largely transformed into Fe–Ti oxide aggregates under subsolidus conditions (Fig. 3a and b). In contrast, the Doğanca lamproite contains large (up to 13 mm) poikilitic phlogopites with small leucite inclusions (Fig. 3c), euhedral diopside and olivine phenocrysts with spinel inclusions, together with apatite, magnetite, Fe–Ti oxides and zircon in a poikilitic sanidine matrix. The Beğendik basalts contain euhedral olivine and clinopyroxene phenocrysts with lesser amounts of orthopyroxene in a matrix of plagioclase microlites. Outer zones of the clinopyroxenes are surrounded by brown zones (Fig. 3d). The clinopyroxenes also

include earlier phases of orthopyroxene (enstatite) and clinopyroxene (pigeonite). 3.1. Analytical methods Mineral analyses were undertaken by electron probe micro analyzer (EPMA) at the Ludwig-Maximillian University, Munich, Germany, using an accelerating voltage of 15 kV and a beam current of 20 nA. The beam diameter of 1 μm applied for the analyses of olivine, pyroxene, chromite, magnetite and ilmenite, whereas it was defocused to 10 μm during the feldspar and phlogopite measurements. The measurements of Si, Ti, Al, Cr, Fe, Mn, Ni, Mg, Ca, Na, K was carried out using the Kα lines. Natural wollastonite, ilmenite, albite, chromite, K-feldspar and metallic Ni were used for the calibration. The counting times for peak and background were 20 and 10 s, respectively. 3.2. Olivine and pyroxenes Forsterite (Fo) contents of olivines from the Doğanca lamproite range from 70.6 to 85.6% (Appendix 1a). The olivines generally show normal zoning, with decreasing Fo and NiO and increasing CaO contents from core to rim (Fig. 4a and b). Similarly, the Fo contents of olivines from the Beğendik basalts vary between 72.4 and 84.1%, but for a given Fo composition, the CaO contents of olivines in the Beğendik basalts are slightly higher than those of the Doğanca lamproite. Clinopyroxenes of the Altınyazı trachyte, Doğanca lamproite and Beğendik basalts are generally diopside-rich and show similar compositions (En 40–49 Fs 09–13 Wo 44–47 and Mg# = 75.7–84.3, En 38–49 Fs 05–12 Wo 39–54 and Mg# = 75.6–89.5, En 30–46 Fs 08–18 Wo 44–51 and Mg# = 61.5–84.4, respectively; See Appendix 1b; and Fig. 4c and d). Wo contents of clinopyroxenes from the Doğanca lamproite are slightly higher than those of the Altınyazı trachyte. The

146

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Fig. 2. Geological map of the study area (modified from Elmas and Şengül, 2013). Also shown are sample locations and radiometric age data of the volcanic units. The samples with labels “T” in black, and radiometric ages are from Ercan et al. (1998). The samples from this study are labeled as “K” in red.

brown outer rims of the clinopyroxenes in the Beğendik basalts are characterized by high amounts of Al 2O 3 and TiO 2 (up to 8.7 wt.% and 5.6 wt.%, respectively; Fig. 4e and f). The Altınyazı trachyte and Beğendik basalts also contain a small amount of orthopyroxene with ~ En61Fs35Wo4 and ~ En54Fs45Wo1, respectively (Appendix 1b). Earlier phases of pyroxene that are surrounded by clinopyroxene phenocrysts in the Beğendik basalts have pigeonite compositions, with En51–55Fs36–37Wo07–11 and Mg# of 58–60 (Fig. 4e and f; Appendix 1b).

3.3. Phlogopite and feldspars The mica minerals from the Altınyazı trachyte and Doğanca lamproite are classified as phlogopite on the basis of their high Mg# and low Al contents (Fig. 5a). The Mg# of phlogopites from the Doğanca lamproite reach up to 83.7 (Appendix 1d), and their Al2O3 and TiO2 contents are 11.1–13.9 wt.% and 1.1–9.4 wt.%, respectively. The compositions of the sanidine matrix of the Altınyazı trachyte and Doğanca lamproite are Ab36–43An04–06Or51–59 and Ab20–53An01–08 Or41–77, respectively. Minor plagioclase in the Altınyazı trachyte is

labradorite in composition, with Ab32–41An57–67Or02–03 (Appendix 1d; Fig. 5b). 3.4. Titano-magnetite, ilmenite and chromite Titano-magnetite, with TiO2 contents ranging between 9.76 and 21.26 wt.%, is abundant in the Altınyazı trachyte, Doğanca lamproite and Beğendik basalts, and several ilmenite grains have also been analyzed in the Doğanca lamproite. The Mg# and Cr# of chromite inclusions embedded in olivine phenocrysts in the Doğanca lamproite are 22.9–32.6 and 64.0–84.6, respectively (Appendix 1c). 4. Whole rock major/trace elements, Sr, Nd and Pb isotopes 4.1. Analytical methods Nineteen samples were collected from the southern Thrace basin (Fig. 2). The samples were prepared by removing the altered surfaces and powdering in a tungsten carbide shatter box at Dokuz Eylül University, Turkey. Importantly, Rice et al. (2009) showed that tungsten

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

147

Fig. 3. Microscope views of the samples. (a) Sample K-15 from Korucuköy lamproite including olivine (Ol), clinopyroxene (Cpx) and small phlogopites (Phl) that are altered to Fe–Ti-oxides. (b) Sample K-7 from Altınyazı trachyte showing small Cpx and largely altered Phl microphenocryst (which are altered to Fe–Ti-oxides) in sanidine (Sa) matrix. (c) Sample K-17 showing large phlogopite phenocrysts with small diopside and apatite in sanidine matrix. Note that phlogopite poikilitically encloses leucite microcrysts. (d) Cpx and Ol phenocrysts in the Beğendik basalts. Note that the cpx phenocrysts are surrounded by brownish rims.

carbide milling (even for 25 min) does not result in significant contamination for the suite of elements and isotopes determined here. Element abundances were determined by inductively coupled plasma-atomic emission spectrometry (ICP-AES) (major elements) and inductively coupled-plasma mass spectroscopy (ICP-MS) (trace elements) in ACME Laboratory, Canada, following lithium–borate fusion and dilute nitric acid digestion of a 0.1 g sample. Loss on ignition (LOI) was determined as the weight difference after ignition at 1000 °C. The results of major and trace element analyses of the Miocene volcanic units in the study area are given in Table 1. Analyses of Sr and Nd isotope ratios in 8 samples and Pb isotope ratios in 6 samples were carried out at the University of Southampton (UK). The samples were dissolved in HF-HNO3 for 24 h on a hot plate at 130 °C. They were evaporated until dry, before adding a further 0.5 ml of concentrated HCl and 0.5 ml of concentrated HNO3 to the samples and evaporating until dry after each addition. For Pb analysis, 1.5 ml of hydrobromic (HBr) was added to the residue, the Teflon pot lid was replaced and the vessels placed on a hotplate for 1 h. The contents were then centrifuged for 5 min to produce a supernatent suitable for column chemistry. Isolation of Pb from the matrix was performed using AG1-X8 200–400 mesh anion exchange resin. The procedural blanks measured with the samples contained b50 pg of Pb. Pb isotope analyses were conducted on a VG Sector 54 thermal ionization mass spectrometer and MC-ICPMS (Neptune) at NOC. Both TIMS and MC-ICPMS techniques utilized the double spike technique to correct instrumental bias using a method outlined by Ishizuka et al. (2007). Pb standard NBS 981 gave results of, 16.9404 ± 32 (2SD) for 206Pb/204Pb, 15.4982 ± 30 for 207Pb/204Pb and 36.7225 ± 85 for 208Pb/204Pb for TIMS and 16.9403 ± 27 for 206Pb/204Pb, 15.4973 ± 21 for 207Pb/204Pb and 36.7169 ± 66 for 208Pb/204Pb for MC-ICP-MS. For Sr analysis, the Pb residue was evaporated and dissolved in 3 M HNO3. The Sr was isolated using Sr resin (Eichrom Industries, Illinois, USA). For Nd isotopic analysis, the Rare Earth Elements (REE) were initially separated by

cation exchange, before isolating Nd on Ln resin (Eichrom Industries, Illinois, USA) columns. Sr and Nd isotope ratios were measured on a nine-collector VG Sector 54 mass spectrometer, as the average of 150 ratios. Reported values are the average of 150 ratios obtained by measuring ion intensities in multidynamic collection mode normalized to 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219. Measured values for NBS SRM-987 and JNdi-1 were 87Sr/86Sr = 0.710297 ± 19 (2 SD, n = 58) and 143Nd/144Nd = 0.512096 ± 7 (2 SD, n = 64) during the measurement period. The Sr and Nd isotopic data presented here have been normalized to NBS SRM-987 (0.710248) and JNdi (0.512110). 5. Results 5.1. Major and trace elements Detailed geochemical classification of the rocks is shown in Fig. 6. Oligocene Na-alkaline rocks, shoshonitic–ultrapotassic (SHO–UK) rocks with nepheline- or olivine–hypersthene-normative mineralogy, and quartz-normative calc-alkaline rocks from the Rhodope region (Marchev et al., 2004; Kirchenbaur et al., 2012a) are shown for comparison. The geochemical data of Ercan et al. (1998) obtained from the volcanic units studied here are also included in the classification diagrams (with smaller symbols). The Oligocene Keşan volcanics are classified as andesite, latite and trachyte on a Total Alkalis vs. Silica (TAS) classification diagram (Fig. 6a). According to the K2O vs SiO2 plot (Fig. 6b), the Keşan volcanics have mainly medium-K calc-alkaline character, similar to calc-alkaline volcanic rocks from Rhodopes. The Korucuköy lamproite plots in the latite TAS field and show a shoshonitic affinity in the K2O vs. SiO2 diagram. On the basis of the criteria proposed by Foley et al. (1987), the Korucuköy lamproite samples have a weak ultrapotassic affinity (Fig. 6c). The Korucuköy lamproite has distinctly higher K contents than the Keşan volcanics and the other Oligocene calc-alkaline rocks.

148

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Fig. 4. (a and b) BSE view of sample K-17 from the Doğanca lamproite, showing a large euhedral olivine phenocryst and the results of microprobe analyses (Mg#, CaO and NiO) along the profile shown on (a). (c and d) pyroxene classification for the samples. (e) BSE view of Diopside (Di) phenocryst from the Beğendik basalts, in which earlier enstatite (En) pigeonite (Pgt) inclusions are enclosed. (f) Microprobe analyses along the profile shown on (e).

The samples from the Doğanca lamproite fall in the phonotephrite field in the TAS diagram (Fig. 6a) and show a clear ultrapotassic affinity (Fig. 6c), with K2O contents in the range of 5.1–5.5 wt.%. Except for sample K-3, the samples of the early Miocene Altınyazı trachyte are classified as shoshonites in the TAS diagram. Sample K-3 has lower SiO2 and K2O (45.6 and 1.5 wt.%, respectively) and higher CaO and LOI (10.4 and 5.5 wt.%, respectively) than the other samples of the Altınyazı trachyte. These features may be indicative of secondary alteration processes, but the immobile element contents (e.g., TiO2, Th, Zr, Hf, Ta and Nb) are also different from the Altınyazı trachyte. Hence, the distinct geochemical composition of sample K-3 may be the result of a magmatic process (e.g., mixing) rather than alteration. The middle Miocene Beğendik basalts include sodic volcanic rocks, which are classified as hawaiite (Fig. 6a) with low SiO2 (45.7–46.5 wt.%) and high MgO contents (8.9–9.3 wt.%). The Altınyazı trachyte and the ultrapotassic rocks of the Korucuköy lamproite and Doğanca lamproite are compared with other lamproites from the Mediterranean region on classification diagrams for ultrapotassic mantle-derived rocks (Fig. 7), and reveal that the Doğanca and Korucuköy lamproite samples show similar compositions to lamproites from Serbia–Macedonia and the Menderes Core Complex in western Anatolia. Chondrite-normalized REE and primitive mantle (PM)-normalized trace element patterns of the volcanic units are shown in Figs. 8 and 9.

The andesite sample K-20 from the Keşan volcanics clearly differs from the Korucuköy lamproite in having lower REE and other trace element abundances (Fig. 8a). Sample K-20 also shows a steep light-REE (LREE) to medium-REE (MREE) and a flat MREE to heavy-REE (HREE) pattern. This sample is also characterized by enrichments of large ion lithophile elements (LILE) over neighboring high field strength elements (HFSE) on the PM-normalized diagram, resulting in troughs in Nb and Ta abundances (Fig. 8b). The trace element abundances of sample K-20 show similarities to those of the least incompatible trace element enriched samples of the Oligocene calc-alkaline rocks from Rhodopes. Samples from the Korucuköy lamproite show LREE-enrichments and nearly flat HREE patterns (Fig. 8a), with weakly developed negative Eu anomalies (Eu/Eu* = 0.8–0.9) and negative HFSE anomalies. Trace element patterns of the Korucuköy lamproite differ from those of the Oligocene calc-alkaline rocks from the Rhodopes, and are more comparable with the Miocene shoshonitic–ultrapotassic volcanic rocks from the Menderes Core Complex (MCC) in western Anatolia (Fig. 8b). On a PM-normalized multi-trace element diagram, the Doğanca and Korucuköy lamproites are characterized by LILE-enrichments over HFSE (Fig. 8d), and more closely resembles the trace element patterns observed in western Anatolian shoshonitic and ultrapotassic rocks, rather than those from the Rhodopes. The Altınyazı trachyte has similar REE and trace element patterns to those of the Doğanca lamproite (Fig. 9a and b), although the latter has

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

149

One of the three samples from the alkaline basaltic Beğendik basalts (K-5) has similar Sr and Nd isotopic ratios to those of alkaline rocks from Rhodopes and Thrace, and is isotopically comparable with OIB-like intra-plate rocks of the Mediterranean, which are derived from the convective mantle beneath the region (Lustrino and Wilson, 2007). In contrast, the other two samples show slightly increased 87Sr/86Sr(i) ratios with constant SiO2 and MgO values. Sample K-5 is also characterized by a lower 206Pb/204Pb(i) ratio that is similar to those of the Altınyazı trachyte, but the other two samples show higher 206Pb/ 204 Pb(i) ratios (Fig. 10d). 6. Discussion In the following sections, the origin of the volcanic units is examined in the context of previous observations from Rhodopes and Western Anatolia in order to develop a model of the geodynamic evolution of the region, and to explore the implications for lamproite petrogenesis. The main geochemical features of the volcanic units studied here are summarized in Table 3. 6.1. Origin and evolution of the volcanic units

Fig. 5. Major element composition of (a) micas (b) feldspars from the Altınyazı trachyte and Doğanca lamproite. Symbols as in Fig. 4 (apfu: atoms per formula unit).

steeper HREE patterns and a positive K anomaly. The REE abundances of sample K-3 resemble the shoshonitic samples of the Altınyazı trachyte (with slightly increased La), but it is has lower Rb, K, Zr and Hf, and higher Ta, Nb and Ti abundances. Samples from the Beğendik basalts show LREE enriched REE patterns (Fig. 9c), with no depletions in the HFSE (Fig. 9d). Na-alkaline basaltic rocks from Rhodopes and Thrace are also shown in Fig. 9d and it is apparent that the trace element patterns of the Beğendik basalts are comparable with those of the Thrace alkaline rocks, especially with regard to their HFSE patterns.

5.2. Sr, Nd and Pb isotopes The Sr, Nd and Pb isotope results are summarized in Table 2. The data are shown in Fig. 10, together with those from neighboring localities. Although the medium-K andesite (sample K-20) of the Keşan volcanics has higher SiO2 contents than the calc-alkaline rocks from Rhodopes, it has lower 87Sr/86Sr(i) and 143Nd/144Nd(i) ratios and is comparable to the Oligocene volcanic rocks of Rhodopes and the Biga Peninsula in NW Anatolia (Fig. 10a). The Korucuköy, Doğanca and Altınyazı samples all have high Sr and low Nd isotopic compositions that are comparable with those of the MCC SHO–UK rocks. Importantly, the samples show uniform Sr isotope ratios against variable SiO2 contents (Fig. 10c). The Pb isotopic compositions of the two samples from the Altınyazı trachyte are similar to those of shoshonitic and ultrapotassic rocks from Rhodopes, Serbia and Macedonia (Fig. 10b and d), and they all have lower 206Pb/204Pb(i), but similar 208Pb/204Pb(i) (Fig. 10d) and 207Pb/204Pb(i) ratios (not shown) to the western Anatolian (MCC SHO–UK) rocks. The Korucuköy lamproite has slightly higher 206Pb/204Pb(i) ratios and is similar to Rhodope volcanic rocks.

6.1.1. Keşan volcanics The Keşan volcanics with medium-K andesitic characteristics have higher SiO2 (~ 59 wt.%) and lower MgO (~ 2.5 wt.%) and Mg# values (~ 50) than the other rock groups considered here, while their Sr and Nd isotopic compositions are comparable with Oligocene volcanic units from Rhodopes and NW Anatolia (Fig. 10). These data indicate that the Keşan volcanics, as with the other Oligocene calc-alkaline intermediate to felsic units in the region, have undergone extensive fractional crystallization and possible crustal contamination processes, as discussed in previous studies (Ercan et al., 1998; Aldanmaz et al., 2000; Altunkaynak and Genç, 2008). The enriched LILE and LREE, over HFSE and Ti abundances on a PM-normalized multi-trace element diagram (Fig. 8) and Sr–Nd isotopic compositions (Fig. 10) indicate that this volcanic unit was derived from a metasomatized mantle source. This interpretation is supported by the Th/Yb vs. Nb/Yb plot (Fig. 11a), which shows that mantle sources affected by subduction-related enrichment are shifted toward higher Th/Yb ratios, whereas volcanic rocks that are not linked to subduction processes plot along the mantle array, which is described by variable degrees of melting and intra-plate enrichment events (e.g., Pearce, 1983). The Keşan volcanics (and the other Oligocene volcanic rocks from Rhodope and NW Anatolia) plot well above the mantle array as a result of the effects of subductionrelated enrichment, as well as fractional crystallization and crustal contamination. Taking into account the age of the Keşan volcanics (26.2 Ma; Ercan et al., 1998), and the geodynamic evolution of the region (late Cretaceous–Paleocene collision between Anatolide–Tauride block and the Sakarya zone), it is most likely that the metasomatic event responsible for the genesis of these volcanics was related to the northward subduction of the northern branch of the Neo-Tethys. Following this event, the extensional regime that started in the Eocene and continued up to the late Oligocene in the Rhodope (e.g., Brun and Sokoutis, 2007; Jahn-Awe et al., 2012; Kilias et al., 2013) was then responsible for melting of the previously metasomatized mantle to produce extensive medium- to high-K calc-alkaline magmatism in the Rhodope to NW Anatolia region (Aldanmaz et al., 2000; Christofides et al., 2004; Marchev et al., 2004; Vlahou et al., 2006; Altunkaynak and Genç, 2008). 6.1.2. Korucuköy and Doğanca lamproites and Altınyazı trachyte The early Miocene Altınyazı trachyte, Korucuköy and Doğanca lamproites are variably enriched in K (and other fluid-mobile incompatible trace elements), and have more primitive geochemical features than the Keşan volcanics. Their SiO2 values are generally b55 wt.% and their MgO contents are N 3 wt.% (Tables 1 and 3). The Korucuköy and Doğanca lamproite also have ultrapotassic affinities with high MgO

150

Table 1 Whole-rock major and trace earth element analyses of the volcanic units. (*) Analyses carried out in University of Southampton. Keşan

Korucuköy

Sample

K-20

K-12

K-13

K-14

K-15

SiO2 Al2O3 Fe2O3(t) MgO CaO Na2O K2O TiO2 P2O5 MnO LOI Mg# Cs Rb Ba Sr Pb Th U Zr Hf Ta Y Nb Sc Cr Ni V La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

58.91 16.05 5.23 2.68 6.22 2.52 1.74 0.50 0.10 0.10 5.70 50.38 2.8 60.1 402.0 556.4

56.83 13.34 5.35 4.82 5.79 2.81 5.59 1.14 0.96 0.11 2.60 64.09 6.4 204.4 2182.0 979.2

57.17 14.24 5.41 2.99 5.24 2.93 6.07 1.24 1.14 0.06 2.80 52.27 6.9 224.8 2971.0 1106.0

56.14 13.25 5.48 6.04 5.92 2.65 5.58 1.14 0.94 0.12 2.10 68.59 5.7 205.5 2065.0 907.0

8.4 2.7 104.5 2.6 0.5 17.6 5.1 13.0

42.6 9.1 596.4 16.1 2.3 22.3 40.7 15.0 301.1 109.3 129.0 63.6 133.7 15.9 65.1 9.6 2.2 6.4 0.8 4.2 0.8 2.0 0.3 2.0 0.3

45.8 9.5 662.6 17.5 2.4 23.9 41.4 16.0 342.1 112.3 144.0 67.5 142.5 17.4 68.7 10.6 2.4 7.3 0.9 4.2 0.8 2.4 0.4 2.1 0.3

39.1 7.8 550.0 14.3 1.7 20.5 34.6 15.0 294.2 135.4 137.0 55.8 123.5 15.2 56.2 9.2 2.2 6.7 0.8 3.6 0.7 2.0 0.3 2.2 0.3

55.56 13.03 5.55 5.50 6.08 2.61 5.51 1.13 0.95 0.13 3.30 66.26 3.5 190.5 2082.0 887.8 37.1 39.4 8.1 550.4 15.2 1.8 20.3 36.4 15.0 287.4 153.9 135.0 56.5 123.1 15.2 62.2 9.3 2.2 6.4 0.8 3.6 0.7 1.7 0.3 1.6 0.2

149.0 17.5 33.0 3.7 13.1 2.9 0.8 2.8 0.5 3.0 0.6 1.6 0.3 1.6 0.3

Altınyazı K-15*

3.23 5.12

4.5 244.5 2406.0 1159.0 37.1 54.8 10.5 769.4 20.2 3.3 30.3 49.8 20.6 355.6 225.1 154.8 76.5 162.3 20.8 79.6 12.8 1.9 7.8 1.1 5.4 1.0 2.7 0.4 2.5 0.4

Doğanca

K-1

K-2

51.63 15.44 7.64 5.07 7.30 2.87 4.14 1.04 0.88 0.16 3.00 56.80 7.7 141.0 2987.0 1479.0

51.21 15.54 7.62 5.21 7.30 2.89 3.97 1.04 0.89 0.16 3.40 57.53 6.3 135.4 2689.0 1179.0 30.7 45.5 9.5 570.3 15.5 1.5 23.7 29.7 20.0 164.2 56.5 174.0 48.4 110.8 15.1 62.7 10.3 2.4 7.1 0.9 5.0 0.9 2.8 0.4 2.5 0.4

50.7 11.0 636.7 17.7 2.2 26.3 36.4 20.0 184.7 56.7 172.0 52.7 118.6 15.9 67.0 10.4 2.5 7.2 1.0 5.3 1.0 2.7 0.4 2.4 0.4

K-2*

2.78 3.93

6.8 146.4 2726.0 1209.0 30.7 48.4 9.4 528.1 14.0 2.8 28.1 36.5 21.1 190.7 81.1 153.0 50.6 121.7 16.2 66.6 10.9 1.6 6.8 0.9 5.0 0.9 2.6 0.4 2.4 0.4

K-4

K-7

51.25 15.50 7.90 5.29 7.50 2.75 3.98 1.05 0.91 0.16 3.00 57.02 6.6 137.2 2648.0 1197.0

51.64 15.58 7.69 5.17 7.39 2.88 3.93 1.04 0.89 0.16 2.80 57.12 7.8 143.1 2757.0 1288.0 29.9 49.4 10.6 640.9 16.0 1.8 26.5 35.0 21.0 171.1 52.1 164.0 53.5 123.1 16.1 65.3 10.7 2.4 7.3 1.0 5.3 0.9 2.8 0.4 2.5 0.4

48.5 10.1 594.1 16.0 1.5 26.3 31.2 21.0 164.2 53.4 175.0 51.0 118.8 15.7 64.9 11.2 2.2 7.5 0.9 4.6 0.9 2.6 0.4 2.3 0.4

K-7*

2.71 3.87

9.0 155.8 2617.0 1154.0 29.9 48.7 7.8 245.2 6.8 2.2 27.6 36.4 23.6 185.0 77.8 153.3 51.4 116.9 16.4 67.2 11.1 1.7 6.9 0.9 4.9 0.9 2.5 0.4 2.3 0.3

Beğendik

K-3

K-17a

K-17b

K-18

K-5

45.61 14.87 8.43 6.56 10.36 3.81 1.47 1.47 1.09 0.15 5.50 60.66 5.1 27.8 1538.0 1346.0

47.56 10.93 8.19 11.76 7.97 2.14 5.15 1.98 0.89 0.12 2.70 73.99 2.2 250.0 1023.0 835.4

48.01 11.10 8.17 11.40 7.94 2.13 5.51 1.97 0.92 0.13 2.20 73.44 2.0 253.6 1095.0 900.4

47.66 10.80 8.30 11.83 8.11 2.03 5.05 1.98 0.93 0.13 2.60 73.85 1.9 249.3 1111.0 860.7

36.0 8.3 443.3 11.0 4.3 27.7 78.6 22.0 266.9 116.6 201.0 70.3 135.6 16.1 60.8 10.0 2.5 7.4 1.1 5.5 1.0 3.0 0.4 2.5 0.4

22.2 2.6 574.8 16.1 2.1 21.1 37.3 23.0 554.2 321.0 208.0 49.1 118.6 16.6 68.1 11.3 2.7 7.5 0.9 4.7 0.7 1.9 0.3 1.5 0.2

24.2 2.7 606.5 16.5 2.2 21.9 39.7 24.0 547.4 310.0 220.0 53.0 124.2 16.8 69.1 11.3 2.8 8.2 0.9 4.5 0.7 2.1 0.3 1.6 0.2

23.2 2.7 616.7 15.4 2.0 22.1 40.1 24.0 567.9 320.0 218.0 52.8 124.2 17.6 75.2 11.4 2.9 7.9 0.9 4.7 0.8 1.9 0.3 1.7 0.2

45.83 13.12 10.89 9.19 9.17 4.08 2.00 2.47 0.63 0.16 1.90 62.57 1.1 39.7 598.0 835.9 5.3 8.0 2.5 233.7 4.9 4.6 21.5 70.4 20.0 260.0 145.0 228.0 35.4 66.5 7.7 32.1 6.9 2.2 6.2 0.9 4.6 0.9 2.1 0.3 1.6 0.2

K-5*

3.83 1.99

1.1 40.0 533.8 744.2 5.3 7.5 2.2 246.9 5.3 5.0 22.3 70.7 20.6 275.4 171.3 212.8 32.2 63.7 7.6 31.2 6.8 2.1 6.0 0.9 4.5 0.8 1.9 0.3 1.5 0.2

K-6

K-8

45.72 12.93 10.51 9.03 9.02 3.79 1.99 2.44 0.62 0.15 3.30 62.99 1.2 46.6 573.0 828.6

46.54 13.15 10.60 8.90 9.66 3.13 2.20 2.45 0.65 0.16 2.00 62.46 1.5 51.2 652.0 1320.0 6.1 8.2 3.0 234.8 5.3 3.6 22.2 64.9 20.0 246.3 143.4 219.0 35.2 66.7 8.0 32.8 6.9 2.2 6.1 0.9 4.4 0.8 1.9 0.3 1.7 0.2

6.8 2.3 210.7 4.6 3.7 20.1 65.6 20.0 239.5 137.5 218.0 31.9 62.1 7.3 33.4 6.6 2.1 5.7 0.8 4.4 0.8 2.0 0.3 1.5 0.2

K-8*

3.21 2.35

1.3 51.6 633.9 1218.0 6.1 9.2 2.5 258.5 5.4 8.3 23.2 70.6 22.0 273.4 160.6 215.8 34.1 68.0 8.1 33.1 7.0 2.1 6.1 0.9 4.5 0.8 2.0 0.3 1.6 0.2

K-9

K-10

K-11

46.35 13.00 10.91 9.18 9.43 3.68 2.12 2.54 0.63 0.16 1.50 62.51 1.0 47.7 576.0 891.9

46.44 12.80 10.89 9.24 9.28 3.92 1.89 2.47 0.62 0.16 1.80 62.70 1.1 38.3 548.0 862.5

7.0 2.4 215.7 4.7 3.6 21.2 63.8 20.0 253.2 147.0 219.0 33.3 62.8 7.4 29.1 6.7 2.0 6.0 0.9 4.6 0.8 2.0 0.3 1.4 0.2

6.6 2.2 202.3 4.7 3.5 20.2 59.3 20.0 253.2 151.2 216.0 31.1 59.3 7.0 27.6 6.2 2.1 5.6 0.8 4.4 0.7 1.9 0.3 1.5 0.2

46.00 12.83 10.80 9.26 9.29 3.69 1.90 2.51 0.62 0.16 2.40 62.95 1.2 38.5 575.0 953.0 5.0 6.9 2.3 214.2 4.7 3.7 19.4 64.7 20.0 246.3 148.4 228.0 32.0 62.6 7.3 31.2 6.4 2.1 5.8 0.9 4.6 0.8 2.1 0.3 1.4 0.2

K-11*

4.05 2.12

1.4 46.3 615.0 946.5 5.0 7.4 2.1 244.1 5.1 4.7 23.2 69.2 27.6 276.5 174.3 219.3 31.7 63.1 7.6 31.1 6.8 2.1 6.1 0.9 4.5 0.8 2.0 0.3 1.6 0.2

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Unit:

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

151

Fig. 6. (a) Total alkali–silica (TAS), (b) K2O vs. SiO2 (c) K2O/Na2O vs. MgO diagrams for the volcanic rocks. IUGS fields in (a) and (b) are after LeMaitre (2002). Data for the Rhodope Oligocene calc-alkaline, shoshonitic–ultrapotassic (SHO–UK) and Na-alkaline rocks are from Christofides et al. (2004), Marchev et al. (2004) and Kirchenbaur et al. (2012a). Data for the Lesbos lamproite are from Pe-Piper et al. (2014). The rock analyses of Ercan et al. (1998) (with smaller symbols) are also included with the samples from this study.

contents (3–6 wt.% and ~ 11.5 wt.%, respectively) and Mg# values (52–69 and 73–74, respectively). These features indicate that the Korucuköy and Doğanca lamproite represent the most primitive

potassic volcanic units in the region. In particular, the very high MgO contents and Mg# values of the Doğanca lamproite (together with Ni and Cr contents of these lamproites of 310–312 and 547–567 ppm,

Fig. 7. Classification diagrams for the Altınyazı trachyte and Korucuköy and Doğanca lamproite (adapted from Foley et al., 1987). Ultrapotassic lamproites from Italy (Conticelli et al., 2009), Spain (Benito et al., 1999; Duggen et al., 2005), Serbia–Macedonia (Altherr et al., 2004; Prelević et al., 2005, 2008; Yanev et al., 2008), Afyon and Menderes Core Complex (MCC) (Floyd et al., 1998; Francalanci et al., 2000; Innocenti et al., 2005; Çoban and Flower, 2006; Akal, 2008; Ersoy et al., 2008, 2010; Prelević et al., 2012) are also shown for comparison.

152

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Fig. 8. Chondrite and Primitive Mantle (PM)-normalized REE and multi-element diagrams for the Keşan volcanics and Korucuköy and Doğanca lamproites. Normalizing factors for the chondrite and PM are from Sun and McDonough (1989), Palme and O'Neill (2004) respectively. Data for the Rhodope Oligocene calc-alkaline (CA) and shoshonitic–ultrapotassic (SHO–UK) rocks are from Marchev et al. (2004) and Kirchenbaur et al. (2012a). Data for the Miocene shoshonitic and ultrapotassic rocks from the Menderes Core Complex (MCC SHO–UK) are from compilation of Ersoy and Palmer (2013).

respectively) indicate that the composition of this unit closely reflects the mantle source composition, while the other samples with lower Mg# indicate that their geochemical features were modified by secondary processes (e.g. AFC) before they reached the surface. One sample from the Altınyazı trachyte (K-3) differs from the other samples from this unit, in that it has a Na-alkaline affinity, lower SiO2 and higher MgO contents and Mg# value (Table 2). The geochemical features of this sample will be discussed separately below. The trace element patterns of the most primitive samples from both the Korucuköy and Doğanca lamproites show variable enrichments of LILE and LREE over HFSE and HREE (Fig. 11a), indicating that all these rocks were derived from mantle sources which have been variably affected by subduction-related enrichment events. In spite of their very primitive nature (Mg# up to 74), the shoshonitic to ultrapotassic volcanic units are characterized by very high 87Sr/86Sr(i) (~ 0.7086) and low 143Nd/144Nd(i) (~ 0.5123) ratios. These features are common in Mediterranean primitive shoshonitic to ultrapotassic rocks, such as lamproites from Spain, Italy and Serbia–Macedonia, and are interpreted to reflect the products of low degree melting of anomalously enriched lithospheric mantle domains (Prelević et al., 2005; Peccerillo and Martinotti, 2006; Conticelli et al., 2007, 2009; Pe-Piper et al., 2009; Tommasini et al., 2011; Ersoy et al., 2012). Although the mantle enrichment process for the ultrapotassic magmas is still debated (see below), it is widely accepted that their geochemical features reflect high contributions from mature continental crust (Zhao et al., 2009; Çoban et al., 2012; Ersoy et al., 2012; Prelević et al., 2013). In order to quantitatively evaluate the petrogenesis of these volcanic rocks we have, therefore, used an average of available chemical data from the felsic Rhodope crustal basement (RCB) (Bonev et al., 2010) to represent this end member. We further

assume that the RCB (average mineral assemblage of K-feldspar(40%) + plagioclase (30%) + biotite(10%) + garnet(10%)) underwent high-degree melting, and that a melt composition obtained by 40% modal batch melting was added to a depleted mantle with by simple mixing. Because the original mantle composition for the high-MgO ultrapotassic magmas requires a depleted mantle source (e.g., Foley, 1992; Prelević et al., 2008), we chose an initially highly depleted mantle composition, that is typically produced in supra-subduction zone environments (SSZ peridotites; Parkinson and Pearce, 1998). The 87Sr/86Sr composition of typical SSZ mantle is 0.704 (from Andean mantle xenoliths; Conceição et al., 2005). To calculate the composition of the metasomatized mantle source, SSZ is mixed with crustal melt (derived from 40% melting of the RCB), in proportions of 1%, 3%, 5% and 10%. The 87Sr/86Sr ratios of these calculated sources range between 0.70484 and 0.70878, and are similar to that of the Doğanca lamproite (0.70858), which most closely reflects the nature of the mantle source of the volcanic rocks in the region. Hence, the composition of these mixed source compositions was used as the starting mantle composition for the melting models here. Condamine and Médard (2014) experimentally determined the melting of phlogopite-bearing lherzolite and harzburgite and concluded that melting of such metasomatized mantle can produce ultra-potassic magmas in post-collisional settings. By using the melting reactions presented in this study (phlogopite harzburgite: ol0.62(−0.99) + opx0.25(1.24) + phl0.10(0.70) + spl0.01(0.05) and phlogopite lherzolite: ol0.56(−0.58) + opx0.21(0.56) + cpx0.11(0.47) + phl0.10(0.49) + spl0.02(0.05); where the mineral and (melt modes) are indicated by subscript numbers, respectively), we performed non-modal fractional melting modeling on the calculated starting composition (the model parameters are given in Table 4). The results are shown on a Tb/Yb vs

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

153

Fig. 9. Chondrite and Primitive Mantle (PM)-normalized REE and multi-element diagrams for the Altınyazı trachyte and Beğendik basalts. Normalizing factors for the chondrite and PM are from Sun and McDonough (1989) and Palme and O'Neill (2004), respectively. Data for the Rhodope Oligocene calc-alkaline Na-alkaline rocks are from Marchev et al. (2004) and Kirchenbaur et al. (2012a). Data for the Miocene shoshonitic and ultrapotassic rocks from the Menderes Core Complex (MCC SHO–UK) are from compilation of Ersoy and Palmer (2013). Thrace Na-alkaline rocks are from Aldanmaz et al. (2006).

La/Nb diagram (Fig. 11b). Clearly, the phlogopite lherzolite mineral assemblage cannot produce the Tb/Yb fractionation seen in the studied samples. While ~ 4% fractional partial melting of a source obtained from 3% mixing of crustal melt into SSZ mantle can reproduce the

trace element ratios of the Doğanca lamproite, reproduction of the Korucuköy lamproite and Altınyazı trachyte compositions requires higher degrees (~5–6%) of melting of a source contaminated by lower degrees of crustal melts. The melting model is also illustrated on a

Table 2 Whole-rock Sr, Nd and Pb isotope data of the volcanic units. Unit

Keşan

Doğanca

Korucuköy

Altınyazı

Altınyazı

Beğendik

Beğendik

Sample

K-20

K-17b

K-15

K-2

K-7

K-5

K-8

Beğendik K-11

Age SiO2 (wt %) Mg# U (ppm) Th (ppm) Pb (ppm) Rb (ppm) Sr (ppm) Sm (ppm) Nd (ppm) 87 Sr/86Sr(m) (2σ) 143 Nd/144Nd(m) (2σ) 147 Sm/144Nd 87 Sr/86Sr(i) 143 Nd/144Nd(i) σNd(DM1)(Ma) 206 Pb/204Pb(m) 207 Pb/204Pb(m) 208 Pb/204Pb(m) 206 Pb/204Pb(i) 207 Pb/204Pb(i) 208 Pb/204Pb(i)

26.2 58.91 50.38 – – – 60.1 556.4 2.9 13.1 0.706380(± 6) 0.512605(± 6) 0.1321 0.706264 0.512583 1010.2 – – – – – –

20.0 48.01 73.44 – – – 253.6 900.4 11.3 69.1 0.708578(± 6) 0.512307(± 6) 0.0996 0.708347 0.512294 1120.7 – – – – – –

20.0 55.56 66.26 8.1 39.4 37.1 190.5 887.8 9.3 62.2 0.708521(± 6) 0.512320(± 3) 0.0907 0.708521 0.512320 1025.0 18.928 15.641 38.824 18.928 15.641 38.824

19.0 51.21 57.53 9.5 45.5 30.7 135.4 1179.0 10.3 62.7 0.708818(± 8) 0.512297(± 20) 0.0995 0.708728 0.512285 1134.2 18.638 15.696 38.953 18.580 15.693 38.861

19.0 51.64 57.12 10.6 49.4 29.9 143.1 1288.0 10.7 65.3 0.708787(± 5) 0.512282(± 4) 0.0999 0.708700 0.512270 1157.2 18.678 15.694 38.969 18.611 15.691 38.866

15.0 45.83 62.57 2.5 8.0 5.3 39.7 835.9 6.9 32.1 0.703363(± 12) 0.512951(± 4) 0.1309 0.703363 0.512951 364.6 18.612 15.695 38.902 18.612 15.695 38.902

15.0 46.54 62.46 3.0 8.2 6.1 51.2 1320.0 6.9 32.8 0.704575(±6) 0.512894(±4) 0.1277 0.704575 0.512894 451.8 18.871 15.661 38.910 18.871 15.661 38.910

15.0 46.00 62.95 2.3 6.9 5.0 38.5 953.0 6.4 31.2 0.704535(±7) 0.512947(±3) 0.1236 0.704535 0.512947 342.0 18.944 15.643 38.856 18.944 15.643 38.856

154

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Fig. 10. (a) 87Sr/86Sr(i)–143Nd/144Nd(i), (b) SiO2–87Sr/86Sr(i), (c) 206Pb/204Pb(i)–143Nd/144Nd(i) and (d) 206Pb/204Pb(i)–208Pb/204Pb(i) diagrams of the volcanic rocks. Data for the Rhodope Oligocene calc-alkaline (CA), shoshonitic–ultrapotassic (SHO–UK) and Na-alkaline rocks are from Marchev et al. (2004) and Kirchenbaur et al. (2012a). Thrace Na-alkaline rocks are from Aldanmaz et al. (2006). Data for the Miocene shoshonitic and ultrapotassic rocks from the Menderes Core Complex (MCC SHO–UK), South Aegean Volcanic Arc (SAVA) and NW-Anatolian Eocene volcanic rocks are from compilation of Ersoy and Palmer (2013). Serbian & Macedonian shoshonitic–ultrapotassic rocks are from Altherr et al. (2004) and Prelević et al. (2005).

PM-normalized multi-element diagram, on which a low-degree (2%) partial melt of a 3% contaminated mantle source is shown (Fig. 11c). Most of the element abundances, especially those of the HFSE, MREE and HREE, are comparable with those of the Doğanca lamproite, but the LILE and LREE abundances show slight differences. Exact reproduction of the observed concentrations by modeling of these elements is precluded because their behavior is controlled by many factors that cannot be fully constrained (e.g., their concentrations are likely highly variable in both the original mantle and contaminant crustal material). Although the Sr and Nd isotope systematics of the Miocene volcanic rocks of the North Aegean are similar to those of Miocene volcanic rocks from W Anatolia (MCC SHO-UK), their Pb isotope compositions form a distinct array toward lower 206Pb/204Pb(i) ratios (Fig. 10d). Together with lamproites from Lesbos, N Aegean (Pe-Piper et al., 2014) and Serbia and Macedonia (Prelević et al., 2005), the Altınyazı trachyte also lie on this array of relatively constant 208Pb/204Pb ratios, but significantly lower 206Pb/204Pb values. Hence, the Pb isotope composition of all the volcanic rocks in this region appear to be dominated by

continental crustal signatures – be they acquired from subduction of sediments and/or crustal fragments, or through crustal assimilation. Thus, even for the volcanic rocks that have not experienced significant degrees of crustal assimilation, the range in Pb isotope composition of the volcanic rocks likely reflects temporal and spatial variations in the source of subducted continental material, rather than any variations in the mantle source. Sample K-3 from the Altınyazı trachyte is characterized by distinct petrographic features with a lower SiO2 content and K2O/Na2O ratio (Fig. 6), and the absence of Nb and Ta anomalies on the PMnormalized trace element diagrams (Fig. 9). Hence it has a higher Nb/ Y (and Nb/Yb) ratio than the other samples. Although this sample is characterized by a higher LOI value (5.5 wt.%; probably due to the presence of serpentinized olivines) than the other sample (2.8–3.4 wt.%), its distinct geochemical character may not simply be due to secondary alteration. We speculate that the geochemical composition of this sample may have resulted from magma mixing processes between the potassic Altınyazı trachyte and Na-alkaline melts (i.e., Beğendik basalts). This process is modeled on the Th/Yb vs Nb/Yb plot, and suggests that mixing

Table 3 Summary of the geochemical features of the volcanic units. Units (this study)

AGE

Rock affinity

IUGS name

SiO2

MgO

Mg#

87

Keşan volcanics Altınyazı trachyte Altınyazı trachyte (K-3) Korucuköy lamproite Doğanca lamproite Beğendik basalts

26.2 21.3–18.5

Medium-K CA Shoshonitic Na-alkaline Ultrapotassic Ultrapotassic Na-alkaline

Andesite Shoshonite Hawaiite Latite Lamproite Hawaiite

58.9 51.2–53.5 45.6 55.6–57.9 47.6–48.0 45.7–46.5

2.7 5.1–5.3 6.6 3.0–6.0 11.4–11.8 8.9–9.3

50.4 56.8–57.5 60.7 52.3–68.6 73.4–74.0 62.5–63.0

0.70626 0.70870–0.70873

Early Miocene Early Miocene 15

Sr/86Sr(i)

0.70852 0.70835 0.70336–0.70458

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

155

Fig. 11. (a) Th/Yb vs Nb/Yb (Pearce, 1983), (b) Tb/Yb vs La/Nb, (c) Primitive Mantle (PM)-normalized multi-element spider diagram and (d) Y/Zr vs Nb/Yb plots for the investigated Thrace volcanic units. Non-modal fractional melting model of metasomatized sources (obtained from simple mixing of supra subduction zone peridotites with 40% partial melt from the average Rhodope crustal basement) are also shown on (b). Melting models are performed for 1%, 3% and 10% mixed mantle sources (yellow asterisks) by using phlogopite harzburgitic mineral assemblage (Condamine and Médard, 2014). The composition of a low-degree (2%) partial melt from 3% contaminated mantle source is compared with the Doğanca lamproite on (c). Simple mixing trends between Altınyazı trachyte and Beğendik basalts are also shown on (a), (b) and (d). See Table 4 for melting parameters, and text for details. PETROMODELER (Ersoy, 2013) is used during geochemical modelings.

between ~30–40% K-alkaline melts (Altınyazı trachyte) with ~60–70% Na-alkaline melts (Beğendik basalts) could produce the geochemical composition of sample K-3 (Fig. 11a). This scenario is also supported by the Tb/Yb vs La/Nb (Fig. 11b) and the Y/Zr vs Nb/Yb systematics (Fig. 11d). 6.1.3. Beğendik basalts The Beğendik basalts, together with the other Na-alkaline basaltic rocks in the region, lie along the mantle array on Fig. 11a. In addition, these rocks do not show negative Nb and Ta anomalies (Fig. 9d), thus the mantle source region of these basalts did not contain crustal materials; i.e., they were derived from a mantle source that had not

experienced subduction-related metasomatism. It is also important to note that the samples of the Beğendik basalts have the lowest 87Sr/ 86 Sr (0.703363) and highest 143Nd/144Nd ratios (0.512951) of the samples considered in this study. All these geochemical features are, therefore, comparable with those of the Oligocene alkaline basalts from the Rhodope (Marchev et al., 2004), the late Miocene Ezine alkaline basalts and the late Miocene Thrace alkaline basalts (Aldanmaz et al., 2006), which are interpreted to be the products of partial melts from asthenospheric mantle sources in the region. Petrographic study of the Beğendik basalts reveals the presence of orthopyroxene (enstatite) remnants surrounded by pigeonitic rims, that all occur in clinopyroxene (diopsidic) phenocrysts (Fig. 4e and f).

Table 4 Parameters for the melting models shown in Fig. 11.

RCB (1) Crustal melt (2) SSZ mantle (3) SSZ + 10% melt (4) SSZ + 5% melt (4) SSZ + 3% melt (4) SSZ + 1% melt (4)

Sr

Th

U

Zr

Hf

Y

Nb

La

Ce

Pr

Nd

Sm

Gd

Tb

Dy

Er

Yb

Lu

87

179.59 160.79 16.03 30.51 23.27 20.37 17.48

10.31 29.40 0.00 2.94 1.47 0.88 0.30

3.37 7.69 0.04 0.80 0.42 0.27 0.12

126.88 336.73 0.10 33.76 16.93 10.20 3.47

10.72 14.38 0.00 1.44 0.72 0.43 0.15

19.51 28.54 0.06 2.91 1.49 0.92 0.35

11.16 25.65 0.01 2.57 1.29 0.78 0.26

34.09 48.12 0.00 4.81 2.41 1.45 0.48

45.81 107.34 0.01 10.74 5.37 3.23 1.08

17.31 12.40 0.00 1.24 0.62 0.37 0.13

24.77 48.95 0.00 4.90 2.45 1.47 0.49

10.74 10.62 0.00 1.07 0.54 0.32 0.11

9.01 8.15 0.00 0.82 0.41 0.25 0.08

5.41 1.06 0.00 0.11 0.05 0.03 0.01

7.42 6.06 0.01 0.61 0.31 0.19 0.07

5.60 2.91 0.01 0.30 0.15 0.10 0.04

5.49 2.50 0.02 0.27 0.14 0.09 0.04

3.96 0.40 0.00 0.04 0.02 0.02 0.01

0.71577 0.71307 0.70400 0.70878 0.70713 0.70615 0.70483

(1) RCB: Average composition of the Rhodope Continental Basement (from Bonev et al., 2010). (2) Obtained by 40% modal melting of RCB. (3) SSZ (supra subduction zone) mantle (average SSZ peridotites from Parkinson and Pearce, 1998). (4) obtained by simple mixing between SSZ mantle and crustal melt with ratios of 1%, 3%, 5% and 10%.

Sr/86Sr

156

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Clinopyroxenes of the Beğendik basalts also show brownish zone along their outer zones. These observations suggest that the Beğendik basalts underwent some magma mixing processes; but further data is needed to confirm this hypothesis. 6.2. Implications for ultrapotassic volcanism The process by which mantle domains are enriched in continental convergence zones is still debated. Some authors argue that subduction of sediments and/or accretionary prisms into the mantle, associated with long-term isolation of metasomatic agents derived from the subducted materials, is required to supply the high amounts of 87Sr, 144 Nd and 207Pb isotopes (Foley, 1992; Tommasini et al., 2011). Tommasini et al. (2011) suggest that the positive correlation between Th/La and Sm/La ratios seen in Mediterranean orogenic lamproites, a geochemical feature absent in steady-state oceanic subduction zone igneous rocks, reflects metasomatic effects of lawsonite and zoisiteepidote veins in progressively metamorphosed accreting mélange rocks under high-pressure conditions in the subduction channel (i.e., metasomatic effects of fluids/melts extracted from the subducted high-pressure mélange rocks). Th/La vs. Sm/La ratios of the volcanic units are shown in Fig. 12. The Doğanca lamproite, together with the shoshonitic–ultrapotassic rocks from eastern Rhodope and the Menderes Core Complex in western Anatolia tend to show variations in Sm/La at relatively constant Th/La. A similar pattern is observed in calcalkaline samples from the eastern Rhodopes and the NW Anatolia Eocene volcanics. However, the samples of the Altınyazı trachyte and the Korucuköy lamproite show a weak positive correlation between Th/La and Sm/La ratios, while the Doğanca lamproites plot along the same horizontal trend as normal arc lavas. These observations suggest that metasomatic agents derived from progressive metamorphism of mélange units played a limited role in generation of the Altınıyazı trachyte and Korucuköy lamproite compositions. An alternative explanation for the origin of ultrapotassic rocks requires more recent subduction of continental crust during crustal collision to create the required mantle source (e.g., Schreyer et al., 1987; Mahéo et al., 2002; Parkinson and Kohn, 2002; Zhao et al., 2009; Çoban et al., 2012; Ersoy et al., 2012; Ersoy and Palmer, 2013; Prelević et al., 2013). Eastern Rhodopes is an important area in which to examine which mechanism is more appropriate in the genesis of the ultrapotassic/lamproitic rocks, because the region contains high(Jahn-Awe et al., 2012; Kirchenbaur et al., 2012b) to ultra-high pressure metamorphic rocks (Kostopoulos et al., 2000; Mposkos and

Fig. 12. Th/La vs Sm/La (Tommasini et al., 2011) plots for the investigated Thrace volcanic units.

Kostopoulos, 2001) formed during the late Cretaceous to Eocene (e.g., Ricou et al., 1998). Ultra-high pressure mineral assemblages have been reported in both continental and oceanic assemblages (e.g., metapelites and eclogites), indicating that these rocks were subducted to depths of up to 150 km (Kostopoulos et al., 2000; Mposkos and Kostopoulos, 2001). Jahn-Awe et al. (2012) and Kirchenbaur et al. (2012b) both report that slices of these metamorphic units in the Rhodope massif underwent high pressure metamorphism during the late Cretaceous (~ 126 Ma) and Eocene (45–43 Ma). Given the crustal accretion history of the region, that developed during late Mesozoic to Cenozoic (e.g., van Hinsbergen et al., 2005), there is therefore no need to invoke long-term storage of subducted material in the mantle for hundreds of millions of years. Rather, we favor metasomatism of the mantle beneath the region as a direct consequence of upper crustal subduction events during closure of the Tethyan oceanic branches in the late Mesozoic and mid Cenozoic. Fluids or melts released directly from the subducted crustal rocks, as well as from the subducted mélange units, were thus most likely responsible for the intense metasomatism of the overlying mantle domains. Partial melts from such an intensely and heterogeneously metasomatized mantle, during post Eocene extensional tectonics, would then have produced the shoshonitic to ultrapotassic rocks and the leucite lamproite in Doğanca district. Prelević and Foley (2007) reported very high-Mg olivine xenocrysts including Cr-rich spinel inclusions in Mediterranean lamproites, indicating that highly refractory mantle materials were also involved in their genesis. Prelević et al. (2012) interpreted these highly depleted xenocrystic olivines (and related spinels) to have originated from shallow subduction of refractory peridotites beneath western Anatolia during the Late Cretaceous. In the geodynamic scenario outlined above, accretion of both oceanic peridotites (as indicated by the xenocrystic olivines, with very high Fo contents which probably originated from highly refractory peridotites) and crustal units (as indicated by crustlike trace element and isotopic ratios) were part of a continuous process during the Alpine orogeny, whereby subduction of depleted mantle was followed by subduction of continental crust to supply all the required components for lamproite genesis. It is proposed that orogenic ultrapotassic rocks, such as lamproites, are the products of first-stage melting of intensely metasomatized lithospheric mantle containing secondary veins of phlogopite, clinopyroxene and amphibole (vein-plus-wall-rock mechanism, Foley, 1992). According to this model, the high-K calc-alkaline rocks (such as andesites) were then produced by melting of the residual parts of the mantle (e.g., Avanzinelli et al., 2009). Therefore, the first products derived from melting of the metasomatic veins have higher incompatible element concentrations (and more radiogenic Sr) than the calc-alkaline products derived from melting of the surrounding harzburgitic residue. In the study area considered here, however, (and for western Anatolia, in general) the ultrapotassic and lamproitic rocks were also accompanied by calc-alkaline magma production. Furthermore, some of the calc-alkaline series have similar Sr isotope ratios to those of the shoshonitic–ultrapotassic series. Thus, these lines of evidence seem to contradict the vein-plus-wall-rock melting model. An alternative model suggests that the high incompatible element and radiogenic Sr contents of the calc-alkaline series (basaltic andesite to rhyolite) observed in western Anatolia, resulted from mixing between mantlederived (shoshonitic–ultrapotassic) and lower crust-derived magmas, in which the lower crustal melting was caused by mafic alkaline magmatic underplating (Ersoy et al., 2012). Overall, we believe that this model provides a more coherent explanation for the temporal and geochemical evolution of the volcanic rocks within the Thrace Basin. Finally, the magmatic activity in the region is clearly related to the extensional tectonics, which have operated since at least late Eocene, and caused exhumation of the metamorphic massifs (Dinter and Royden, 1998; Bonev and Beccaletto, 2007; Brun and Sokoutis, 2007; Jahn-Awe et al., 2012).

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

7. Conclusions The Eocene–Miocene Thrace extensional basin includes a number of small volume volcanic units which have variable geochemical features. These are: (1) the Oligocene medium- to high-potassic calc-alkaline andesites and dacites of Keşan volcanics, (2) the early Miocene shoshonitic–ultrapotassic rocks of Korucuköy and, Doğanca lamproites and Altınyazı trachyte, and (3) the middle Miocene Na-alkaline Beğendik basalts. The Keşan volcanics are typical products of postcollisional calc-alkaline volcanism, and were emplaced during extension of the Thrace basin. The Oligocene Keşan volcanics were emplaced in the early stages of extension in the Thrace basin and exhibit typical features of postcollisional medium- to high-potassic volcanic activity, and were derived from a previously metasomatized mantle source. Geochemical features of the most primitive samples (MgO N5 wt.%) from the Korucuköy and Doğanca lamproites (and also the Altınyazı trachyte) indicate they were derived from crustally metasomatized mantle sources. Taking into account the geodynamic evolution of the region, namely the presence of high- to ultra-high-pressure metamorphic continental/oceanic rocks developed during the late Cretaceous and Eocene, and the subsequent extensional tectonics and core complex formation, it is proposed that these rocks were derived from a mantle source that had been contaminated by fluids or melts derived from crustal rocks directly subducted into the mantle. These high-MgO potassic magmas were emplaced during enhanced stages of extension. During the middle Miocene, the Na-alkaline Beğendik basalts, with typical intraplate geochemical features (such as Na-alkaline chemical affinity, the absence of Nb negative anomalies on PM-normalized multi-element diagrams, as well as low Sr (0.70416) and high Nd (0.51293) isotopic ratios) were emplaced, indicating the presence of passively upwelling asthenospheric mantle beneath the region. Acknowledgments We thank Özgür Karaoğlu and M. Uğur Öven for their help during the field studies, Mustafa Çiçek for his help during the preparation of the polished samples for electron probe studies, Cüneyt Akal for comments on the petrographic studies and İ. Şentürk Ersoy and Pınar Bacıoğlu for their help during sample preparation. We also thank Dr. Melanie Kaliwoda for her help during electron microprobe studies at the Mineralogical State Collection, Munich. Agnes Michalik is thanked for her assistance with the geochemical analyses in Southampton. Michele Lustrino and an anonymous referee are thanked for their valuable comments and contributions to the manuscript. We also thank to Dr. Malcolm J. Rutherford for editorial handling. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jvolgeores.2014.06.016. References Akal, C., 2008. K-richterite–olivine–phlogopite–diopside–sanidine lamproites from the Afyon volcanic province, Turkey. Geol. Mag. 145, 570–585. Aldanmaz, E., Pearce, J.A., Thirlwall, M.F., Mitchell, J.G., 2000. Petrogenetic evolution of late Cenozoic, post-collision volcanism in western Anatolia Turkey. J. Volcanol. Geotherm. Res. 102, 67–95. Aldanmaz, E., Köprübaşı, N., Gürer, Ö.F., Kaymakçı, N., Gourgaud, A., 2006. Geochemical constraints on the Cenozoic, OIB-type alkaline volcanic rocks of NW Turkey: implications for mantle sources and melting processes. Lithos 86, 50–76. Altherr, R., Meyer, M.-P., Holl, A., Volker, F., Alibert, C., McCulloch, M.T., Majer, V., 2004. Geochemical and Sr–Nd–Pb isotopic characteristics of Late Cenozoic leucite lamproites from the East European Alpine belt (Macedonia and Yugoslavia). Contrib. Mineral. Petrol. 147, 58–73. Altunkaynak, Ş., Genç, Ş.C., 2008. Petrogenesis and time-progressive evolution of the Cenozoic continental volcanism in the Biga peninsula, NW Anatolia (Turkey). Lithos 102, 316–340.

157

Arnaud, N.O., Vidal, P., Tapponnier, P., Matte, P., Deng, W.M., 1992. The high K2O volcanism of northwestern Tibet: geochemistry and tectonic implications. Earth Planet. Sci. Lett. 111, 351–367. Avanzinelli, R., Elliot, T., Tommasini, S., Conticelli, S., 2008. Constraints on the genesis of potassium-rich Italian volcanic rocks from U/Th disequilibrium. J. Petrol. 49, 195–223. Avanzinelli, R., Lustrino, M., Mattei, M., Melluso, L., Conticelli, S., 2009. Potassic and ultrapotassic magmatism in the circum-Tyrrhenian region: significance of carbonated pelitic vs. pelitic sediment recycling at destructive plate margins. Lithos 113, 213–227. Benito, R., López-Ruiz, J., Cebriá, J.M., Hertogen, J., Doblas, M., Oyarzun, R., Demaiffe, D., 1999. Sr and O isotope constraints on source and crustal contamination in the high-K calc-alkaline and shoshonitic Neogene volcanic rocks of SE Spain. Lithos 46, 773–802. Bergman, S.C., 1987. Lamproites and other potassium-rich igneous rocks: a review of their occurrences, mineralogy and geochemistry. In: Fitton, J.G., Upton, B.G.J. (Eds.), Alkaline igneous rocks. Geological Society of London special publication, 30, pp. 103–119. Boccaletti, M., Manetti, P., Peccerillo, A., Stanisheva-Vassileva, G., 1978. Late Cretaceous high-potassium volcanism in eastern Srednogorie, Bulgaria. Bull. Geol. Soc. Am. 89, 439–447. Bonev, N., Beccaletto, L., 2007. From syn- to post-orogenic Tertiary extension in the north Aegean region: constraints on the kinematics in the eastern Rhodope–Thrace, Bulgaria–Greece and the Biga Peninsula, NW Turkey. Geol. Soc. Lond. Spec. Publ. 291, 113–142. Bonev, N., Stampfli, G., 2011. Alpine tectonic evolution of a Jurassic subduction–accretionary complex: deformation, kinematics and 40Ar/39Ar age constraints on the Mesozoic low-grade schists of the Circum-Rhodope Belt in the eastern Rhodope–Thrace region, Bulgaria–Greece. J. Geodyn. 52, 143–167. Bonev, N., Moritz, R., Márton, I., Chiaradia, M., Marchev, P., 2010. Geochemistry, tectonics, and crustal evolution of basement rocks in the Eastern Rhodope Massif, Bulgaria. Int. Geol. Rev. 52, 269–297. Brun, J.-P., Sokoutis, D., 2007. Kinematics of the Southern Rhodope Core Complex (North Greece). Int. J. Earth Sci. (Geol. Rundsch.) 96, 1079–1099. Christofides, G., Pécskay, Z., Eleftheriadis, G., Soldatos, T., Koroneos, A., 2004. The Tertiary volcanic rocks (Thrace, Northeastern Greece): petrology and K/Ar geochronology. Geol. Carpath. 55, 397–409. Çoban, H., Flower, M.F.J., 2006. Mineral phase compositions in silica-undersaturated ‘leucite’ lamproites from the Bucak area, Isparta, SW Turkey. Lithos 89, 275–299. Çoban, H., Karacık, Z., Ece, Ö.I., 2012. Source contamination and tectonomagmatic signals of overlapping Early to Middle Miocene orogenic magmas associated with shallow continental subduction and asthenospheric mantle flows in Western Anatolia: a record from Simav (Kütahya) region. Lithos 140–141, 119–141. Conceição, R.V., Mallmann, G., Koester, E., Schilling, M., Bertotto, G.W., Rodriguez-Vargas, A., 2005. Andean subduction-related mantle xenoliths: Isotopic evidence of Sr–Nd decoupling during metasomatism. Lithos 82, 273–287. Condamine, P., Médard, E., 2014. Experimental melting of phlogopite-bearing mantle at 1 GPa: implications for potassic magmatism. Earth Planet. Sci. Lett. 397, 80–92. Conticelli, S., Carlson, R.W., Widow, E., Serri, G., 2007. Chemical and isotopic composition (Os, Pb, Nd, and Sr) of Neogene to Quaternary calc-alkaline, shoshonitic, and ultrapotassic mafic rocks from the Italian peninsula: inferences on the nature of their mantle sources. In: Beccaluva, L., Bianchini, G., Wilson, M. (Eds.), Cenozoic volcanism in the Mediterranean area. Geological Society of America, 418, pp. 171–202. Conticelli, S., Guarnieri, L., Farinelli, A., Mattei, M., Avanzinelli, R., Bianchini, G., Boari, E., Tommasini, S., Tiepolo, M., Prelević, D., Venturelli, G., 2009. Trace elements and Sr– Nd–Pb isotopes of K-rich, shoshonitic, and calc-alkaline magmatism of the Western Mediterranean Region: genesis of ultrapotassic to calcalkaline magmatic associations in a post-collisional geodynamic setting. Lithos 107, 68–92. Dhont, D., Yanev, Y., Bardintzeff, J.-M., Chorowicz, J., 2008. Evolution and relationships between volcanism and tectonics in the central-eastern part of the Oligocene Borovitsa caldera (Eastern Rhodopes, Bulgaria). J. Volcanol. Geotherm. Res. 171, 269–286. Ding, L., Kapp, P., Zhong, D., Deng, W., 2003. Cenozoic volcanism in Tibet: evidence for a transition from oceanic to continental subduction. J. Petrol. 44, 1833–1865. Dinter, D.A., Royden, L., 1998. Late Cenozoic extension in northeastern Greece: Strymon Valley detachment system and Rhodope metamorphic core complex. Geology 21, 45–48. Duggen, S., Hoernle, K., van den Bogaard, P., Garbe-Schönberg, D., 2005. Post-collisional transition from subduction to intraplate-type magmatism in the westernmost Mediterranean: evidence for continental-edge delamination of subcontinental lithosphere. J. Petrol. 46, 1155–1201. Elmas, A., 2011. The Thrace Basin: stratigraphic and tectonic palaeogeographic evolution of the Palaeogene formations of northwest Turkey. Int. Geol. Rev. 54, 1419–1442. Elmas, A., 2012. Basement types of the Thrace Basin and a new approach to the preEocene tectonic evolution of the northeastern Aegean and northwestern Anatolia: a review of data and concepts. Int. J. Earth Sci. (Geol. Rundsch.) 101, 1895–1911. Elmas, A., Şengül, A., 2013. Miocene formations and NE-trending right-lateral strike-slip tectonism in Thrace, northwest Turkey: geodynamic implications. Int. Geol. Rev. 55, 705–729. Ercan, T., Türkecan, A., Guillou, H., Satır, M., Sevin, D., Şaroğlu, F., 1998. Marmara Denizi çevresindeki Tersiyer volkanizmasının özellikleri [Characteristic features of Tertiary volcanism around Marmara Sea]. Miner. Res. Explor. Bull. Turk. 120, 199–221. Ersoy, E.Y., 2013. PETROMODELER (Petrological Modeler): a Microsoft® Excel© spreadsheet program for modeling melting, mixing, crystallisation and assimilation processes in magmatic systems. Turk. J. Earth Sci. 22, 115–125. Ersoy, Y., Helvacı, C., 2007. Stratigraphy and geochemical features of the Early Miocene bimodal (ultrapotassic and calc-alkaline) volcanic activity within the NE-trending Selendi basin, western Anatolia, Turkey. Turk. J. Earth Sci. 16, 117–139.

158

Y.E. Ersoy et al. / Journal of Volcanology and Geothermal Research 283 (2014) 143–158

Ersoy, E.Y., Palmer, M.R., 2013. Eocene–Quaternary magmatic activity in the Aegean: implications for mantle metasomatism and magma genesis in an evolving orogeny. Lithos 180–181, 5–24. Ersoy, Y., Helvacı, C., Sözbilir, H., Erkül, F., Bozkurt, E., 2008. A geochemical approach to Neogene–Quaternary volcanic activity of western Anatolia: an example of episodic bimodal volcanism within the Selendi Basin, Turkey. Chem. Geol. 255, 265–282. Ersoy, E.Y., Helvacı, C., Palmer, M.R., 2010. Mantle source characteristics and melting models for the early–middle Miocene mafic volcanism in Western Anatolia: implications for enrichment processes of mantle lithosphere and origin of K-rich volcanism in post-collisional settings. J. Volcanol. Geotherm. Res. 198, 112–128. Ersoy, E.Y., Helvacı, C., Uysal, İ., Karaoğlu, Ö., Palmer, M.R., Dindi, F., 2012. Petrogenesis of the Miocene volcanism along the İzmir–Balıkesir Transfer Zone in western Anatolia, Turkey: implications for origin and evolution of potassic volcanism in postcollisional areas. J. Volcanol. Geotherm. Res. 241–242, 21–38. Floyd, P.A., Helvacı, C., Mittwede, S.K., 1998. Geochemical discrimination of volcanic rocks associated with borate deposits: an exploration tool? J. Geochem. Explor. 60, 185–205. Foley, S.F., 1992. Vein-plus-wall-rock melting mechanisms in the lithosphere and the origin of potassic alkaline magmas. Lithos 28, 435–453. Foley, S.F., Venturelli, G., Green, D.H., Toscani, L., 1987. The ultrapotassic rocks: characteristics, classification, and constraints for petrogenetic models. Earth-Sci. Rev. 24, 81–134. Francalanci, L., Innocenti, F., Manetti, P., Savaşçın, M.Y., 2000. Neogene alkaline volcanism of the Afyon–Isparta area, Turkey: petrogenesis and geodynamic implications. Mineral. Petrol. 70, 285–312. Fytikas, M., Innocenti, F., Manetti, P., Mazzuoli, R., Peccerillo, A., Villari, L., 1984. Tertiary to Quaternary evolution of volcanism in the Aegean region. Geol. Soc. Lond. Spec. Publ. 17, 687–699. Gao, Y., Hou, Z., Kamber, B.S., Wei, R., Meng, X., Zhao, R., 2007. Lamproitic rocks from a continental collision zone: evidence for recycling of subducted Tethyan oceanic sediments in the mantle beneath southern Tibet. J. Petrol. 48, 729–752. Georgiev, V., Milovanov, P., Monchev, P., 2008. Strontium isotope studies of the late Alpine extensional magmatism in eastern rhodopes. Annual of the University of Mining and Geology “St. Ivan Rilski”, 51, p. 108. Georgiev, V., von Quadt, A., Heinrich, C.A., Peytcheva, I., Marchev, P., 2012. Time evolution of a rifted continental arc: integrated ID-TIMS and LA-ICPMS study of magmatic zircons from the Eastern Srednogorie, Bulgaria. Lithos 154, 53–67. Gülmez, F., Genç, Ş.C., Keskin, M., Tüysüz, O., 2012. A post-collision slab-breakoff model for the origin of the Middle Eocene magmatic rocks of the Armutlu–Almacık belt, NW Turkey and its regional implications. Geol. Soc. Lond. Spec. Publ. 372, 107–139. IGME, 1983. Geological map of Greece (1:500,000), Athens. Innocenti, F., Agostini, S., Di Vincenzo, G., Doglioni, C., Manetti, P., Savaşçın, M.Y., Tonarini, S., 2005. Neogene and Quaternary volcanism in Western Anatolia: magma sources and geodynamic evolution. Mar. Geol. 221, 97–421. Ishizuka, O., Taylor, R.N., Yuasa, M., Milton, J.A., Nesbitt, R.W., Uto, K., Sakamoto, I., 2007. Processes controlling along-arc isotopic variation of the southern Izu-Bonin arc. Geochem. Geophys. Geosyst. 8, Q06008. Jahn-Awe, S., Pleuger, J., Frei, D., Georgiev, N., Froitzheim, N., Nagel, T.J., 2012. Time constraints for low-angle shear zones in the Central Rhodopes (Bulgaria) and their significance for the exhumation of high-pressure rocks. Int. J. Earth Sci. (Geol. Rundsch.) 101, 1971–2004. Kilias, A., Falakis, G., Sfeikos, A., Papadimitriou, E., Vamvaka, A., Gkarlaouni, 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. Kirchenbaur, M., Münker, C., Schuth, S., Garbe-Schönberg, D., Marchev, P., 2012a. Tectonomagmatic constraints on the sources of eastern Mediterranean K-rich lavas. J. Petrol. 53, 27–65. Kirchenbaur, M., Pleuger, J., Jahn-Awe, S., Nagel, T.J., Froitzheim, N., Fonseca, R.O.C., Münker, C., 2012b. Timing of high-pressure metamorphic events in the Bulgarian Rhodopes from Lu–Hf garnet geochronology. Contrib. Mineral. Petrol. 163, 897–921. Kostopoulos, D.K., Ioannidis, N.M., Sklavounos, S.A., 2000. A new occurrence of ultrahighpressure metamorphism, Central Macedonia, Northern Greece: evidence from graphitized diamonds? Int. Geol. Rev. 42, 545–554. Kürkçüoğlu, B., Furman, T., Hanan, B., 2008. Geochemistry of postcollisional mafic lavas from the North Anatolian Fault zone, Northwestern Turkey. Lithos 101, 416–434. LeMaitre, R.W., 2002. Igneous Rocks: A Classification and Glossary of Terms: Recommendations of the International Union of Geological Sciences, Subcommission on the Systematics of Igneous Rocks. Cambridge University Press. Lustrino, M., Wilson, M., 2007. The circum-Mediterranean anorogenic Cenozoic igneous province. Earth-Sci. Rev. 81, 1–65. Mahéo, G., Guillot, S., Blichert-Toft, J., Rolland, Y., P cher, A., 2002. A slab breakoff model for the Neogene thermal evolution of South Karakorum and South Tibet. Earth Planet. Sci. Lett. 195, 45–58. Marchev, P., Raicheva, R., Downes, H., Veselli, O., Chiaradia, M., Moritz, R., 2004. Compositional diversity of Eocene–Oligocene basaltic magmatism in the Eastern Rhodopes, SE Bulgaria: implications for genesis and tectonic setting. Tectonophysics 393, 301–328. Mposkos, E.D., Kostopoulos, D.K., 2001. Diamond, former coesite and supersilicic garnet in metasedimentary rocks from the Greek Rhodope: a new ultrahigh-pressure metamorphic province established. Earth Planet. Sci. Lett. 192, 497–506. MTA, 2002. Geological Map of Turkey (1:500000), Mineral Research and Exploration (MTA) Bulletin, Turkey. Natalin, B.A., Sunal, G., Satır, M., Toraman, E., 2012. Tectonics of the Strandja Massif, NW Turkey: history of a long-lived arc at the northern margin of palaeo-tethys. Turk. J. Earth Sci. 21, 755–798. Nelson, D.R., McCulloch, M.T., Sun, S.-S., 1986. The origins of ultrapotassic rocks as inferred from Sr, Nd and Pb isotopes. Geochim. Cosmochim. Acta 50, 231–245.

Okay, A.I., Tüysüz, O., 1999. Tethyan sutures of Northern Turkey. In: Durand, B., Jolivet, L., Horthváth, F., Séranne, M. (Eds.), The Mediterranean Basin: Tertiary extension within the Alpine orogen. Geological Society of London Special Publication, 156, pp. 475–515. Okay, A.I., Satır, M., Maluski, H., Siyako, M., Monie, P., Metzger, R., Akyüz, S., 1996. Paleoand Neo-Tethyan events in northwestern Turkey: geologic and geochronologic constraints. In: Yin, A., Harrison, M. (Eds.), Tectonics of Asia. Cambridge Univ. Press, Cambridge, UK, pp. 420–441. Okay, A.I., Satır, M., Tüysüz, O., Akyüz, S., Chen, F., 2001. The tectonics of the Strandja Massif: late-Variscan and mid-Mesozoic deformation and metamorphism in the northern Aegean. Int. J. Earth Sci. (Geol. Rundsch.) 90, 217–233. Palme, H., O'Neill, H.St.C., 2004. Cosmochemical estimates of mantle composition. In: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry. 2. Elsevier, Amsterdam, pp. 1–38. Parkinson, C.D., Kohn, M.J., 2002. Continental subduction to depths of 200 km: implications for intra-continental ultrapotassic magmatism. In: Parkinson, C.D., Katayama, I., Liou, J. G., Maruyama, S. (Eds.), The Diamond-Bearing Kokchetav Massif. Kazakhstan. Universal Academy Press Inc, Tokyo, pp. 463–476. Parkinson, I.J., Pearce, J., 1998. Peridotites from the Izu–Bonin–Mariana Forearc (ODP Leg 125): evidence for mantle melting and melt–mantle interaction in a supra-subduction zone setting. J. Petrol. 39, 1577–1618. Pearce, J.A., 1983. Role of the sub-continental lithosphere in magma genesis at active continental margins. In: Hawkesworth, C.J., Norry, M.J. (Eds.), Continental Basalts and Mantle Xenoliths. Shiva Publishing, Nantwich, pp. 230–249. Peccerillo, A., 1998. Relationships between ultrapotassic and carbonate-rich volcanic rocks in central Italy: petrogenetic and geodynamic implications. Lithos 43, 267–279. Peccerillo, A., Martinotti, G., 2006. The Western Mediterranean lamproitic magmatism: origin and geodynamic significance. Terra Nova 18, 109–117. Pe-Piper, G., Piper, D.J.W., Koukouvelas, I., Dolansky, L.M., Kokkalas, S., 2009. Postorogenic shoshonitic rocks and their origin by melting underplated basalts: the Miocene of Limnos, Greece. Geol. Soc. Am. 121, 39–54. Pe-Piper, G., Zhang, Y., Piper, D.J.W., Prelević, D., 2014. Relationship of Mediterranean type lamproites to large shoshonite volcanoes, Miocene of Lesbos, NE Aegean Sea. Lithos 184–187, 281–299. Prelević, D., Foley, S.F., 2007. Accretion of arc-oceanic lithospheric mantle in the Mediterranean: evidence from extremely high-Mg olivines and Cr-rich spinel inclusions in lamproites. Earth Planet. Sci. Lett. 256, 120–135. Prelević, D., Foley, S.F., Romer, R.L., Cvetković, V., Downes, H., 2005. Tertiary ultrapotassic volcanism in Serbia: constraints on petrogenesis and mantle source characteristics. J. Petrol. 46, 1443–1487. Prelević, D., Foley, S.F., Romer, R.L., Conticelli, S., 2008. Mediterranean Tertiary lamproites: multicomponent melts in post-collisional geodynamics. Geochim. Cosmochim. Acta 72, 2125–2156. Prelević, D., Akal, C., Foley, F., Romer, R.L., Stracke, A., van den Bogaard, P., 2012. Ultrapotassic mafic rocks as geochemical proxies for post-collisional dynamics of orogenic lithospheric mantle: the case of southwestern Anatolia, Turkey. J. Petrol. 53, 1019–1055. Prelević, D., Jacob, D.E., Foley, S.F., 2013. Recycling plus: a new recipe for the formation of Alpine–Himalayan orogenic mantle lithosphere. Earth Planet. Sci. Lett. 362, 187–197. Rice, A., H.,N, Tait, J., Anderson, M.W., 2009. Use of tungsten carbide disc-mill in geochemistry: no evidence of contamination. Geophys. Res. Abstr. 11, EGU2009–EGU9350. Ricou, L.-E., Burg, J.-P., Godfriaux, I., Ivanov, Z., 1998. Rhodope and Vardar: the metamorphic and the olistostromic paired belts related to Cretaceous subduction under Europe. Geodin. Acta 11, 285–309. Ring, U., Glodny, J., Will, T., Thomson, S., 2010. The Hellenic subduction system: highpressure metamorphism, exhumation, normal faulting, and large-scale extension. Annu. Rev. Earth Planet. Sci. 38, 45–76. Schenker, F.L., Gerya, T., Burg, J.-P., 2012. Bimodal behavior of extended continental lithosphere: modeling insight and application to thermal history of migmatitic core complexes. Tectonophysics 579, 88–103. Schreyer, W., Massonne, H.J., Chopin, C., 1987. Continental crust subducted to depths near 100 km: implications for magma and fluid genesis in collision zones. In: Mysen, B.O. (Ed.), Magmatic processes: physicochemical principles. Geochem. Soc. Spec. Publ, 1, pp. 155–163. Şengör, A.M.C., Yılmaz, Y., 1981. Tethyan evolution of Turkey: a plate tectonic approach. Tectonophysics 75, 181–241. Stampfli, G.M., 2000. Tethyan oceans. Geol. Soc. Lond. Spec. Publ. 173, 1–23. Sun, S.-Su., McDonough, W.F., 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes. In: Saunders, A.D., Norry, M.J. (Eds.), Magmatism in the Ocean Basins. Geological Society of London, 42, pp. 313–345. Tommasini, S., Avanzinelli, R., Conticelli, S., 2011. The Th/La and Sm/La conundrum of the Tethyan realm lamproites. Earth Planet. Sci. Lett. 301, 469–478. van Hinsbergen, D.J.J., Hafkenscheid, E., Spakman, W., Meulenkamp, J.E., Wortel, R., 2005. Nappe stacking resulting from subduction of oceanic and continental lithosphere below Greece. Geology 33, 325–328. Vlahou, M., Christofides, G., Eleftheriadis, G., Pinarelli, L., Koroneus, A., 2006. Tertiary volcanic rocks from Samothraki island (north Aegean, Greece): Sr and Nd isotope constraints on their evolution. Geol. Soc. Am. Spec. Pap. 409, 283–304. Yanev, Y., Boev, B., Doglioni, C., Innocenti, F., Manetti, P., Pecskay, Z., Tonarini, S., D'Orazio, M., 2008. Late Miocene to Pleistocene potassic volcanism in the Republic of Macedonia. Mineral. Petrol. 94, 45–60. Zhao, Z., Mo, X., Dilek, Y., Niu, Y., DePaolo, D.J., Robinson, P., Zhu, D., Sun, C., Dong, G., Zhou, S., Lou, Z., Hou, Z., 2009. Geochemical and Sr–Nd–Pb–O isotopic compositions of the post-collisional ultrapotassic magmatism in SW Tibet: petrogenesis and implications for India intra-continental subduction beneath southern Tibet. Lithos 113, 190–212.