Neotethyan intraoceanic microplate rotation and variations in spreading axis orientation: Palaeomagnetic evidence from the Hatay ophiolite (southern Turkey)

Earth and Planetary Science Letters 280 (2009) 105–117

Contents lists available at ScienceDirect

Earth and Planetary Science Letters j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e p s l

Neotethyan intraoceanic microplate rotation and variations in spreading axis orientation: Palaeomagnetic evidence from the Hatay ophiolite (southern Turkey) Jennifer Inwood a,⁎, Antony Morris a, Mark W. Anderson a, Alastair H.F. Robertson b a b

School of Earth, Ocean and Environmental Sciences, University of Plymouth, Drake Circus, Plymouth PL4 8AA, UK School of Geosciences, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, UK

a r t i c l e

i n f o

Article history: Received 22 September 2008 Accepted 14 January 2009 Available online 23 February 2009 Editor: L. Stixrude Keywords: ophiolite palaeomagnetism rotation tectonic Neotethys Turkey

a b s t r a c t Insights into tectonic processes operating in ancient ocean basins are provided by analyses of fragments of oceanic lithosphere preserved as ophiolites during collisional orogenesis. Here we present a palaeomagnetic analysis of the Upper Cretaceous Hatay (Kizil Dağ) ophiolite of Turkey that provides evidence for intraoceanic microplate rotation and variations in ridge axis orientation in a Neotethyan ocean basin. Magnetizations at 46 sites are shown to be predeformational in origin and rotated from the relevant reference direction. A net tectonic rotation approach to the analysis of the data provides information on permissible net rotation poles and angles and allows uncertainties in input vectors to be considered. Results demonstrate that all levels of the ophiolite have been rotated anticlockwise by angles in excess of 90° around steeply inclined axes. The Hatay ophiolite formed in the same supra-subduction zone spreading system as the Troodos ophiolite (Cyprus), which is known to have rotated 90° anticlockwise in an intraoceanic setting in the Late Cretaceous to Early Eocene. By considering our results in the context of the known timing of the Troodos rotation, we infer that 50–60° of rotation of the Hatay ophiolite took place as part of an areally extensive “Troodos microplate”. This phase of rotation was triggered by initial impingement of the Arabian continental margin with the Neotethyan subduction trench, consistent with models for modern day oceanic microplate rotation in complex convergent plate boundaries. The Hatay ophiolite then became detached from the actively rotating microplate and was emplaced onto the Arabian margin in the Maastrichtian, undergoing a further 30–40° of anticlockwise rotation during thrusting. Back-stripping of rotations allows correction of the Hatay sheeted dykes to their initial orientations. The restored dyke trend of 020° differs from that inferred previously for the Troodos sheeted dyke complex, demonstrating a primary variation in orientation of Neotethyan spreading axes. Such variability is commonly observed in modern spreading systems in marginal basins; these may act as analogues for the supra-subduction zone spreading inferred for many ophiolites. © 2009 Elsevier B.V. All rights reserved.

1. Introduction It is now accepted that ophiolites represent fragments of oceanic lithosphere preserved during collisional orogenesis. Ophiolites provide fundamental insights into oceanic tectonic processes associated with their formation at spreading centres and subsequent deformation during plate convergence, and are also important for regional palaeogeographic reconstructions. Numerous ophiolitic units are exposed throughout the eastern Mediterranean region and are interpreted to have mainly formed by supra-subduction zone spreading within Neotethyan oceanic basins during Late Cretaceous time. The most extensively studied unit is the Troodos ophiolite of

⁎ Corresponding author. Present address: Borehole Research Group, Department of Geology, University of Leicester, University Road, Leicester LE1 7RH, UK. Tel.: +44 116 2523327; fax: +44 116 2523918. E-mail addresses: [email protected] (J. Inwood), [email protected] (A. Morris), [email protected] (M.W. Anderson), [email protected] (A.H.F. Robertson). 0012-821X/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2009.01.021

Cyprus, which has been uplifted without complex internal tectonic disruption, leaving its spreading fabric largely intact. Palaeomagnetic research on the Troodos ophiolite has shown that tectonic rotations are a fundamental crustal response to oceanic extensional and transform faulting (Allerton and Vine, 1987; Bonhommet et al., 1988; Morris et al., 1990; Allerton and Vine, 1991; Hurst et al., 1992; Morris et al., 1998). The ophiolite and its sedimentary cover also preserve a unique palaeomagnetic record of oceanic microplate rotation that may be linked to plate-scale geodynamic interactions (Clube et al., 1985; Clube and Robertson, 1986; Abrahamsen and Schonharting, 1987; Robertson 1990). More recently, palaeomagnetic results have been reported from the Upper Cretaceous Baër-Bassit ophiolite of Syria (Morris et al., 2002; Morris and Anderson, 2002). This formed in the same southern Neotethyan oceanic basin as the Troodos ophiolite but was subsequently emplaced onto the Arabian continental margin during the Maastrichtian, undergoing extensive tectonic dismemberment (Al-Riyami et al., 2000, 2002). Extreme and locally variable anticlockwise rotations are observed in the Baër-Bassit units (Morris et al., 2002) that may be related in part to neotectonic

106

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

activity, but which more likely reflect Late Cretaceous emplacementrelated tectonic processes and/or intraoceanic rotation as documented for the Troodos ophiolite. The large Upper Cretaceous Hatay (or Kizil Dag) ophiolite of southern Turkey is crucial to understanding the pattern and significance of tectonic rotations in oceanic crust of the southern Neotethyan basin. This aerially extensive ophiolite is exposed 250 km to the NE of Troodos and 40 km to the north of Baër-Bassit (Fig. 1a). It was emplaced onto the Arabian continental margin together with the Baër-Bassit ophiolite as part of the same, large unit (Robertson, 2002). Internal tectonic disruption of the Hatay ophiolite is very much less than that of the Baër-Bassit unit, and it retains a clearly defined Penrose-type pseudostratigraphy like the Troodos ophiolite. Here we present the first palaeomagnetic data from this major terrane. Pre-deformational magnetic remanences are analysed using a net tectonic rotation approach that allows quantification of rotation axes and angles and their associated uncertainties. The results provide evidence for oceanic microplate- and emplacement-related rotation events and for primary variations in the orientation of spreading axes within the southern Neotethyan ocean basin, allowing comparison to processes operating in modern marginal basin systems. 2. Geological setting The Hatay ophiolite of Turkey and the related Baër-Bassit ophiolite of Syria comprise the westernmost part of the ‘Ophiolitic Crescent’ of the northern margin of Arabia (Ricou, 1971) (Fig. 1a). The Hatay massif covers 950 km2 (25× 45 km) and is composed of an ophiolite sequence up to 7 km thick. It is split into a large southwestern massif and a smaller northeastern massif by a high-angle fault, known as the Tahtaköprü Fault (Fig. 1b). The succession in the main massif (Delaloye and Wagner, 1984) begins with serpentinized tectonised harzburgite with local intercalations of dunite, wehrlite, lherzolite and feldspathic peridotites. The ultramafic rocks are separated from the overlying layered sequence by a 50–100 m thick shear zone (Dilek and Thy, 1998). The layered ultramafic and gabbroic rocks display cumulate textures and locally show a weaklyto moderately-developed foliation parallel to layering defined by the alignment of pyroxene and plagioclase crystals (Dilek and Thy, 1998). Isotropic gabbros become dominant higher in the plutonic section, and are intruded by small bodies of plagiogranites, leucocratic gabbro and dolerite. Dolerite dykes become abundant towards the top of the gabbros, passing upwards into a sheeted dyke complex. Locally, the gabbro-dyke contact is a low-angle shear zone marked by hydrothermal alteration. The best exposures of sheeted dykes occur along the coast in a 4.5 km long continuous section (Delaloye et al., 1980). Dykes there are generally subvertical and E–W striking, although some differences in orientation are observed along the section. Extrusive igneous rocks are not preserved in the main massif. The smaller ophiolitic massif to the NE of the Tahtaköprü Fault (Fig.1b) lacks the coherent internal structure and pseudostratigraphy observed in the main massif (Dilek and Thy, 1998). Here a highly attenuated upper crustal sequence is in tectonic contact with underlying upper mantle or lower crustal rocks. These relationships are best exposed in two localities near the villages of Kömürçukuru and Tahtaköprü. Extrusive sequences around Kömürçukuru are nearly 600 m thick and comprise massive and pillow lava flows with intercalated metalliferous sedimentary rocks (Erendil, 1984; Robertson, 1986) and are in faulted contact with isotropic gabbros beneath (Dilek and Thy, 1998). At Tahtaköprü, the approximately 400 m thick extrusive sequence overlies serpentinized peridotites along a gently SE-dipping normal fault (Dilek and Thy, 1998). The Hatay ophiolite forms a 7 km thick thrust sheet emplaced over limestones of the Arabian platform, from which it is separated by only a

thin (10s of metres) melange with no metamorphic sole present (Robertson, 1986, 2002). The contact is only exposed in two small areas making emplacement relations difficult to determine (Aslaner, 1973). However, kinematic indicators suggest emplacement towards the SE in the well-developed metamorphic sole of the smaller Baër-Bassit ophiolite to the south, which is interpreted as the thinned and structurally dismembered leading edge of the emplaced ophiolite sheet (Al-Riyami et al., 2002). The combined Hatay/Baër-Bassit thrust sheet was emplaced in the middle Maastrichtian, with the timing of this event precisely bracketed by the ages of the youngest carbonates in the underlying authochthon (early Maastrichtian) and the oldest postemplacement sedimentary cover sequences (late Maastrichtian). The post-emplacement sedimentary cover of the Hatay ophiolite (Fig. 1b) reaches a total thickness of around 3 km (Piskin et al., 1986). A Maastrichtian basal conglomerate horizon (Erendil, 1984; Piskin et al., 1986) overlies the ophiolite (Tinkler et al., 1981), and is succeeded concordantly by a series of claystones, sandstones, limestones and marls of Maastrichtian to Late Eocene age (Piskin et al., 1986). The Miocene sequence unconformably overlies either the older sedimentary sequences or the ophiolite (Boulton and Robertson, 2007) and is succeeded by Pliocene sandstones, marly limestones and claystones and then by Quaternary conglomerates, travertines, alluvium and beach sand (Boulton et al., 2006). 3. Sampling and methods We have sampled the Hatay layered sequence (ultramafic and gabbroic cumulates), the sheeted dyke complex and the extrusive sequence at 43 sites along key, representative sections (Fig. 1b) for palaeomagnetic analyses in order to quantify any tectonic rotations that have affected the ophiolite. Three additional sites were sampled within gabbros that are exposed as host-rock screens between dykes. An average of eight samples per site were drilled in situ using standard palaeomagnetic procedures. At one pillow lava site (ML01) samples were collected as small hand-samples from cooling-related joint blocks by attaching a flat disc which provided a suitable surface for orientation. Sampling was restricted to exposures that showed either consistent palaeovertical indicators (exposures of multiple, subparallel sheeted dykes) or palaeohorizontal indicators (laterally continuous, planar layering in gabbroic rocks; coherent sequences of pillowed and sheet lava flows). The orientations of these indicators were measured in the field to an accuracy of ±5°. A standard palaeomagnetic analysis based on simple tilt corrections would assume that these measured surfaces represent initially vertical/ horizontal planes. However, the net tectonic rotation approach adopted here allows uncertainties in initial orientations (e.g. potential palaeoslopes in lava sequences) to be incorporated into the analyses. Natural remanences of the ophiolitic samples were measured in the University of Plymouth palaeomagnetic laboratory using a Molspin fluxgate spinner magnetometer (noise level = 0.05 × 10− 3 A/m). Samples were subjected to stepwise alternating field (AF) and thermal demagnetization. Characteristic remanent magnetizations (ChRMs) were found using orthogonal vector plots and principal component analysis (Kirschvink, 1980) and site mean remanence directions were computed using Fisherian statistics (Fisher, 1953). A range of supporting rock magnetic analyses were performed on selected samples, both at the University of Plymouth palaeomagnetic laboratory, and more extensively at the Institute for Rock Magnetism, Minneapolis. These included hysteresis measurements (to determine magnetic domain states), low temperature susceptibility experiments (to identify magnetic inversions related to ferromagnetic mineralogy), and the determination of Curie temperatures.

Fig. 1. Location maps. (a) The eastern Mediterranean ‘peri-Arabian ophiolite crescent’ of Ricou (1971) extending eastwards from the Troodos ophiolite of Cyprus to the Semail ophiolite of Oman; (b) Simplified geological map of the Hatay ophiolite, showing the location of palaeomagnetic sampling sites. Sites were located in six main localities, with the first letter of each site designating the sampling area: K = Karaçay valley; J = Kisecik valley; C = Coastal section; I = Isikli; M = Kömürçukuru region; T = Tahtaköprü region.

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

107

108

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

4. Results and analysis 4.1. Magnetic mineralogy and palaeomagnetic results from the ophiolite The magnetic mineralogy experiments demonstrate the presence of fine magnetic grain sizes with pseudo-single domain properties capable of retaining a stable magnetization throughout the protracted

history of the ophiolite (Fig. 2). Magnetic mineralogy, however, displays variation dependent on crustal level. Clear Verwey transitions and high Curie temperatures (TCs) in the deeper crustal levels (ultramafic cumulates and cumulate gabbros) indicate that Ti-poor titanomagnetite/magnetite is the major carrier of the magnetic signal, whilst the absence of the Verwey transition and lower TCs of the higher crustal levels (extrusive rocks) indicate the presence of a more

Fig. 2. Magnetic characteristics of rocks from the Hatay ophiolite. (a) A Day plot (Day et al., 1977) of hysteresis parameters for samples from the cumulate rocks, sheeted dyke complex and the extrusive sequence of the Hatay ophiolite. The majority of samples contain pseudo-single domain grains capable of retaining a stable magnetization over geological timescales; (b) Typical examples of orthogonal demagnetization diagrams, showing well-defined stable end-point remanence directions in all lithologies isolated by both alternating field and thermal treatment. Solid circles = horizontal plane; open symbols = vertical N–S plane Note that the vertical projections are on the horizontal axes.

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

Ti-rich titanomagnetite. These mineralogies are compatible with acquisition of stable remanences during or soon after genesis of the ophiolitic crust at an oceanic spreading centre. Stable components of magnetization were isolated at all sites, following removal of minor secondary components during initial demagnetization. Typical examples of demagnetization behaviour are shown in Fig. 2. Most samples are dominated by univectorial, single component decay to the origin. Both AF and thermal demagnetization experiments yielded identical remanence directions (Fig. 2), although thermal demagnetization data are occasionally noisier than the AF data. Stable components of magnetization were identified from individual samples and subsequently combined to give a mean ChRM for each site. In situ magnetic remanences from all sites are shown in the stereographic equal area projections of Fig. 3 and are listed in Table 1. Magnetizations are predominantly of normal polarity, consistent with remanence acquisition during the Cretaceous long normal polarity interval (chron C34N; Cande and Kent, 1992). Directions are unrelated to the present-day geocentric axial dipolar field in the Hatay region and are generally directed towards the west or southwest (Fig. 3), indicating substantial anticlockwise tectonic rotation since magnetization acquisition. Magnetization directions of sites within host gabbro screens in the sheeted dyke complex are indistinguishable from those of the associated dykes. Reversed polarity remanences were observed at only one isolated locality in the northwestern corner of the ophiolite (Isikli; sites ID01–04; Fig. 1b), where a moderately dipping series of sheeted dykes is in faulted contact with the underlying plutonic section. These localised reversed directions suggest that magnetization acquisition in this single section occurred later than in the main ophiolite (and potentially may postdate crustal accretion).

109

only tilt test formulation (Enkin and Watson, 1996) is independent of the structural history and assumes that the angle between the inclination and the identified palaeohorizontal at a site remains constant during rigid body rotation. A statistically significant improvement of clustering of inclinations upon tilt correction from sites with different structural orientations implies that a pre-tilt magnetization has been identified (Enkin and Watson, 1996). Data from the 19 palaeohorizontal sites from the Hatay layered gabbro and extrusive sections yield the following statistics: In situ Tilt − corrected

Iˆ = 48:6- + 36:3- = −18:2Iˆ = 32:5- + 5:2- = −4:9-

k = 4:2 k = 28:4

where Iˆ and k are the maximum likelihood estimates of the true mean inclination in degrees and the Fisher precision parameter respectively. Stepwise untilting gives a maximum k value of 29.0 at 90% untilting (Fig. 4), suggesting that remanences were acquired before significant tectonic disruption of the ophiolitic crust. Where a sample collection consists of several groups of sites, each from a separate coherent block or section, the declination information within each block is usable and need not be discarded. In these circumstances the “block-rotation Fisher” analysis of Enkin and Watson (1996) is applicable. This maximises the use of the remanence data and yields improved estimates of mean inclination for subsequent use in a parametric resampling tilt test formulation (Enkin and Watson, 1996). This approach yields the following statistics for the six blocks containing palaeohorizontal sites: In situ

Iˆ = 42:6- + 9:8- = −9:5Iˆ = 32:4- + 4:6- = −4:5-

k = 7:9

4.2. Timing of magnetization acquisition

Tilt − corrected

This is determined by using field tests of palaeomagnetic stability, the most common of which is the palaeomagnetic tilt test (McElhinny, 1964; McFadden and Jones, 1981). However, in regions where differential vertical axis rotations may have occurred, use of the standard area-wide tilt test based on full remanence vectors (declination and inclination) is invalid. In addition, palaeomagnetic data from sheeted dykes may be affected by components of rotation around dyke-normal axes which are impossible to identify from structural observations alone, again making standard tilt tests unreliable. In order to take account of these complications we adopt an alternative approach that uses the distribution of inclinations from sites where a palaeohorizontal can be determined. This inclination-

Stepwise untilting gives a maximum k value of 33.1 at 90% untilting (Fig. 4), again supporting acquisition of remanence prior to deformation (Enkin and Watson, 1996). In both cases a parametric re-sampling implementation of the tilt test (Enkin and Watson, 1996), using 1000 re-sampling trials and incorporating a circular standard deviation of 5° on the poles to palaeohorizontal surfaces, indicates an optimum untilting with 95% confidence limits close to 100% of untilting. This constitutes a positive result indicating acquisition of remanence prior to deformation of the sampled sequences (Enkin and Watson, 1996), and suggests that the observed magnetizations were acquired during or shortly after oceanic crustal genesis.

k = 32:3

Fig. 3. Lower hemisphere stereographic projections of site mean remanence directions (in situ/geographic coordinates, without tilt correction). Ellipses = projection of α95 cones of confidence around site mean remanences.

110

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

Table 1 Palaeomagnetic data from the Hatay ophiolite. Site

Lithology

n

In situ Dec

Dyke sampling sites Coastal section Group 1 (southern section) CD01 CD02 CD03 CG01 Group 2 (central section) CD04 CD05 CD12 Group 3 (northern section) CD06 CD07 CD08 CD09 CD10 CD11 Karaçay Valley KD01 KD02 KD03 KD08 KD09 KD10 KD11 KD12 KG02 KG03 Isikli ID01 ID02 ID03 ID04

α95

k

Structure

Lat.

Long.

Inc

Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Gabbro screen between dykes

11 11 9 6

196.1 201.0 208.3 194.6

46.2 61.3 64.6 59.6

35.1 125.9 72.9 166.9

7.8 4.1 6.1 5.2

130/83 317/80 116/80 –

36° 36° 36° 36°

07.89′N 07.89′N 07.89′N 07.89′N

35° 54.77′E 35° 54.77′E 35° 54.77′E 35° 54.77′E

Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes

6 8 10

243.6 262.1 231.9

66.7 51.4 46.9

36.1 26.8 23.7

11.3 11.0 10.1

160/66 190/74 162/56

36° 08.39′N 36° 09.02′N 36° 08.43′N

35° 54.63′E 35° 54.16′E 35° 54.60′E

Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes

7 7 10 9 7 14

342.1 298.4 341.3 328.4 336.9 322.9

69.8 81.2 72.7 76.3 62.6 73.4

224.2 50.3 132.2 53.8 55.1 105.2

4.0 8.6 4.2 7.1 8.2 3.9

010/72 010/72 016/68 360/78 003/69 006/74

36° 36° 36° 36° 36° 36°

09.47′N 09.47′N 09.51′N 10.11′N 09.91′N 09.79′N

35° 53.83′E 35° 53.83′E 35° 53.79′E 35° 53.25′E 35° 53.42′E 35° 53.51′E

Dolerite dykes cutting layered gabbros Dolerite dykes cutting layered gabbros Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Gabbro screen between dykes Gabbro screen between dykes

9 8 6 7 10 10 7 5 7 6

211.1 192.8 237.8 243.9 269.6 228.6 237.5 201.6 238.0 272.5

64.4 63.0 41.7 51.1 58.8 32.4 45.5 30.1 38.3 58.5

57.0 173.1 42.8 139.7 109.6 130.8 64.4 44.9 43.1 129.9

7.4 4.6 10.4 5.1 4.6 4.2 7.6 11.5 9.3 5.9

153/63 163/77 195/72 350/88 001/82 174/72 173/79 168/70 – –

36° 36° 36° 36° 36° 36° 36° 36° 36° 36°

11.40′N 11.40′N 11.60′N 14.77′N 13.67′N 12.37′N 12.45′N 12.37′N 11.60′N 13.67′N

35° 59.35′E 35° 59.35′E 35° 58.84′E 35° 58.45′E 35° 57.65′E 35° 57.60′E 35° 57.51′E 35° 57.54′E 35° 58.84′E 35° 57.65′E

Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes Dolerite sheeted dykes

6 8 7 9

131.8 153.9 139.8 142.6

− 11.7 − 35.9 − 18.8 − 11.9

76.2 225.2 209.4 175.8

7.7 3.7 4.2 3.9

040/45 026/48 040/43 030/38

36° 36° 36° 36°

19.70′N 19.83′N 20.02′N 20.02′N

35° 48.91′E 35° 48.90′E 35° 49.19′E 35° 49.19′E

Layered gabbros Layered gabbros

7 9

191.2 287.5

80.5 67.7

207.4 601.5

4.2 2.1

290/44 304/40

36° 11.20′N 36° 10.11′N

35° 52.22′E 35° 53.25′E

Layered gabbros Layered gabbros Layered gabbros Layered gabbros

4 9 6 8

305.7 271.6 274.0 268.2

46.3 56.2 47.3 38.9

55.2 56.6 50.7 57.0

12.5 6.9 9.5 7.4

342/45 355/45 342/45 351/41

36° 36° 36° 36°

17.03′N 17.07′N 17.01′N 17.37′N

36° 36° 36° 36°

Layered gabbros Layered gabbros Layered gabbros Layered gabbros

7 10 10 10

185.7 198.8 199.3 217.9

65.5 66.3 67.0 79.1

82.1 146.7 260.1 320.9

6.7 4.0 3.0 2.7

308/72 300/58 297/58 280/47

36° 36° 36° 36°

11.40′N 11.40′N 11.40′N 11.44′N

35° 59.35′E 35° 59.35′E 35° 59.35′E 35° 59.11′E

Layered gabbbro sampling sites Coastal section CC01 CC02 Kisecik Valley JC01 JC02 JC03 JC04 Karaçay Valley KC01 KC02 KC03 KC04

02.96′E 03.20″E 03.23′E 03.38′E

Lava sampling sites Kömürçukuru region ML01 ML02 ML03 ML04 Tahtaköprü region TL01 TL02 TL03 TL04 TL05

Basaltic Basaltic Basaltic Basaltic

pillow lavas sheet flow pillow lavas pillow lavas

7 8 6 4

226.1 292.7 278.9 262.3

27.0 17.6 13.9 22.0

109.9 75.8 37.4 63.7

5.8 6.4 11.1 11.6

118/33 171/36 171/36 171/36

36° 36° 36° 36°

25.86′N 25.78′N 25.58′N 25.64′N

36° 36° 36° 36°

08.70′E 08.38′E 08.44′E 08.46′E

Basaltic Basaltic Basaltic Basaltic Basaltic

pillow lavas pillow lavas pillow lavas pillow lavas pillow lavas

6 6 6 6 5

278.4 281.1 268.3 280.9 289.6

15.0 21.7 5.1 23.0 21.5

12.2 40.9 26.7 33.7 117.1

20.0 10.6 13.2 11.7 7.1

151/62 151/62 151/62 151/62 151/62

36° 36° 36° 36° 36°

23.42′N 23.42′N 22.68′N 22.68′N 22.68′N

36° 36° 36° 36° 36°

12.03′E 12.03′E 10.61′E 10.60′E 10.60′E

n = number of specimens; Dec = declination; Inc = Inclination; k = Fisher precision parameter; α95 = semi-angle of 95% cone of confidence; Structure = dip direction and dip of inferred palaeovertical/horizontal.

4.3. Selection of reference direction Tectonic interpretation of palaeomagnetic data is achieved by comparing observed magnetization vectors with a reference (expected) magnetization vector, commonly calculated from an appropriate apparent polar wander path (APWP). The Hatay, BaërBassit and Troodos ophiolites are all interpreted to have formed in a

southern Neotethyan ocean basin that formed by the rifting of microcontinental fragments from the northern Gondwanan (African) margin. African APWPs for the Late Cretaceous (Westphal et al., 1986; Besse and Courtillot, 1991, 2002) yield expected directions of magnetization for the Hatay region with northerly declinations (within the limits of uncertainty of the mean poles) and inclinations of approximately 30°. However, reference directions calculated in this

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

111

4.4. Determination of net tectonic rotations

Fig. 4. Variation in the Fisher precision parameter with progressive untilting of sampled sites and localities where the palaeohorizontal can be inferred, indicating a positive inclination-only tilt test (Enkin and Watson, 1996). Thin line = results using site-level data; Thick line = results using the block-rotation formulation of the Enkin and Watson (1996) method.

way do not take into account the shortening (palaeolatitudinal shift) associated with plate convergence in the period between crustal formation at a Neotethyan spreading axis and southwards emplacement of the Hatay ophiolite onto the continental margin. This suggests that an appropriate reference inclination should be slightly steeper than that defined from African APWPs. A more appropriate reference inclination is provided, therefore, by the extensive tilt corrected palaeomagnetic data from the coeval Troodos ophiolite. A recent synthesis of all available data (Morris et al., 2006) indicates a mean Troodos inclination of 38°. However, the mean Troodos declination (273°; Morris et al., 2006) reflects the well-known Late Cretaceous to Eocene palaeorotation of the “Troodos microplate”. In this study, therefore, we adopt an unrotated reference magnetization vector of Dec = 000° (based on African APWPs), Inc = 38° (based on Troodos data).

Standard palaeomagnetic corrections for the effect of tectonic tilting upon magnetization directions involve rotating inferred palaeohorizontal/vertical surfaces back to horizontal/vertical around strike-parallel axes. The total deformation at a site is, therefore, arbitrarily decomposed into components of tilting and vertical axis rotation. In complexly deformed terrains, where fold axes are seldom horizontal and where multiple phases of deformation may occur, this procedure can introduce serious declination errors (MacDonald, 1980; Kirker and McClelland, 1996). It is more appropriate, therefore, to describe the deformation at a site in terms of a single rotation about an inclined axis, which restores both the palaeohorizontal/vertical to its initial orientation and the site mean magnetization vector to the appropriate palaeomagnetic reference direction. This single rotation may then be decomposed into any number of component rotations on the basis of additional structural data (Kirker and McClelland, 1996). The net tectonic rotation algorithm employed here is that devised by Allerton and Vine (1987) for use within the sheeted dyke terrane of the Troodos ophiolite of Cyprus, and which was later modified by Morris et al. (1998) to yield estimates of uncertainties in rotation axis orientations and angles. This technique can be applied to both palaeovertical and palaeohorizontal cases, with the key assumption being that no internal deformation of a sampled unit has occurred. Under this circumstance, the angle ß between the magnetization vector and the normal to the dyke/flow is constant during deformation (Allerton and Vine, 1987). The analysis involves finding an initial normal to the palaeovertical/horizontal plane which conserves the angle ß and is as horizontal/vertical as possible. The mean site magnetization vector (SMV) is then restored to the reference magnetization vector (RMV) and the present palaeosurface normal to its initial orientation. The rotation axis which allows this restoration is located at the intersection of the great circle bisectrix of the SMV and RMV and that of the present and initial palaeosurface normals (Fig. 5a). The net tectonic rotation is described by the azimuth and plunge of the axis of rotation, and the angle of rotation; a positive angle represents an anticlockwise rotation (Allerton and Vine, 1987). Two solutions are generated in those cases where dykes can be restored to the vertical, and additional criteria (e.g. compatibility with observed structures) may be used to

Fig. 5. An example of the net tectonic rotation analysis using data from site CD10 as an illustration. (a) The Allerton and Vine (1987) algorithm. SMV = site magnetization vector (in situ remanence); RMV = reference magnetization vector; PDN = present dyke normal; IDN = initial calculated dyke normal; R = axis of net tectonic rotation; dashed line indicates circle of radius ß (= angle between SMV and PDN) centred on RMV; subscripts 1 and 2 refer to alternative solutions; (b) multiple application of this method to all combinations of five estimates of SMV, RMV and PDN, distributed around their respective α95 cones of confidence (Morris et al., 1998); (c) gives 125 estimates of the net tectonic rotation axis for each solution (only solution R1 is shown). These define an envelope that represents a first-order approximation of the 95% region of confidence around the true rotation axis. The inset histogram illustrates the associated distribution of net tectonic rotation angles.

112

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

Table 2 Net tectonic rotation parameters. Site Dyke sites Coastal section Group 1 (southern section) CD01 CD02 CD03 Group 2 (central section) CD04 CD05 CD12 Group 3 (northern section) CD06 CD07 CD08 CD09 CD10 CD11 Karaçay Valley KD01 KD02 KD03 KD08 KD09 KD10 KD11 KD12 Isikli: Sheeted dyke complex ID01 ID02 ID03 ID04 Site

Preferred solution Axis

Angle

Initial strike/dip

056/82 015/78 027/75

157 160 147

016/90 028/90 170/90

066/66 092/72 100/67

99 85 103

158/90 180/90 167/90

030/54 039/59 033/54 033/57 019/55 031/58

87 102 88 97 87 94

182/90 021/90 011/90 002/90 177/90 007/90

066/66 061/67 108/73 041/81 032/75 126/79 087/80 116/78

117 131 105 117 101 122 113 146

167/90 169/90 025/90 016/90 011/90 024/90 013/90 041/90

311/51 348/52 322/51 317/46

86 75 82 80

191/90 184/90 191/90 357/90

Preferred solution Axis

Layered gabbro sites Coastal section CC01 CC02 Kisecik Valley JC01 JC02 JC03 JC04 Karaçay Valley KC01 KC02 KC03 KC04 Extrusive sites Kömürçukuru region ML01 ML02 ML03 ML04 Tahtaköprü region TL01 TL02 TL03 TL04 TL05

Angle

Initial dip

060/54 068/62

97 69

9 10

129/61 113/45 122/57 133/59

42 60 64 71

31 6 18 14

079/44 076/54 075/55 056/59

94 100 100 101

9 5 6 1

275/73 295/63 285/64

156 92 104

6 5 16 19

300/55 287/54 287/54 309/55 312/53

128 128 128 130 122

12 12 12 7 14

Axis = azimuth and plunge of rotation pole; Angle = angle of rotation (positive angle = anticlockwise rotation); Initial strike/dip = restored dyke orientation; Initial dip = restored dip of inferred palaeohorizontal.

select a preferred solution. Importantly, this method can resolve rotations of dykes around margin-normal axes that are impossible to observe in the field and which may result in misinterpretation of data when standard tilt corrections are employed (Borradaile, 2001; Morris and Anderson, 2002). In the modification of this method (Morris et al., 1998), the Allerton and Vine (1987) algorithm is applied to all combinations of

each of five orientations for the three vectors input into the analysis (i.e. RMV, SMV, normal to palaeosurface; Fig. 5b). These orientations are distributed around the α95 circles for each vector (an α95 of 5° is assigned to the structural data). This yields 125 combinations of input vectors, and an output consisting of a minimum of 125 and maximum of 250 estimates of the net tectonic rotation axis and angle at each site. The envelope on a stereonet which completely encloses the set of estimated rotation axes (Fig. 5c) provides a first-order approximation of the associated 95% confidence region. This envelope is the only practicable means of describing the confidence region, since the net tectonic rotation technique does not yield rotation axis estimates which are symmetric around the mean estimate. The associated rotation angles can be plotted as histograms (Fig. 5c). This method indicates the range of rotation axes and angles which are possible at a site given the uncertainties in orientation of the various input vectors. 4.5. Net tectonic rotation solutions The analyses demonstrate that, without exception, sites have experienced major rotation around moderately to steeply inclined axes. The net tectonic rotation parameters found by single application of the Allerton and Vine (1987) technique to our data are given in Table 2, whereas the stereonets of Figs. 6 and 7 show the site-level envelopes of potential rotation axes within each locality together with histograms of estimated rotation angles compiled at the locality level. In the case of the sheeted dyke sites, we have accepted the net tectonic rotation solution involving anticlockwise rotation since this is consistent with the sense of displacement of in situ directions away from the reference direction (Fig. 3) and generally yields smaller rotation angles. The most extensive data come from the sheeted dyke sections exposed along the coastal road and in the Karaçay river valley. With the exception of sites in the southern section of the coastal exposures, net tectonic rotation solutions from these localities are very similar, with subvertical to steeply plunging E to NE-directed rotation axes and rotation angles of 85–100° (Fig. 6). Nearly identical rotation angles are also observed in sheeted dykes exposed in the Isikli area, but here rotation poles are less steeply plunging and NW-directed (Fig. 6). The Isikli dykes are less steeply dipping than in the main sheeted dyke complex (Table 1) and have been tilted in the hanging wall of a normal fault zone (Dilek and Thy, 1998). Hence the more moderately inclined net tectonic rotation poles reflect a combination of this component of tilting and a substantial rotation around a steeply plunging axis. Analysis of data from the southernmost sampled section of sheeted dykes along the coast (sites CD01–03) again demonstrates steeply plunging (to sub-vertical) rotation axes, but here rotation angles are significantly larger than in the sections to the north with a modal value of 150–160°. The net tectonic rotation analysis also provides information on admissible initial dyke strikes (see Table 2), but discussion of this aspect of the analysis is deferred to Section 5 below. Sites in lower crustal layered sequences along the coastal, Karaçay and Kisecik sections have also been rotated around broadly Edirected, moderately plunging rotation axes (Fig. 7a). Rotation angles vary within and between sections from 65–75° to 95–105°. The Allerton and Vine (1987) technique does not require that measured layering represents a true palaeohorizontal surface, but instead places the initial pole to layering as close to vertical as possible (and hence layering as close to horizontal as possible) during calculation of a single net tectonic rotation solution. However, a geometric consequence of the analysis in these cases is that the azimuth of the initial pole to layering is forced to lie along the declination of the reference magnetization vector. An alternative approach is to assume that the net tectonic rotation parameters from the sheeted dyke sites in each locality also describe the deformation in the layered sequences, and use these solutions to restore the orientation of

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

113

Fig. 6. Results of net tectonic rotation analyses of palaeomagnetic data from sites in the sheeted dyke complex of the main massif of the Hatay ophiolite.

layering to its pre-deformational position. This is particularly applicable in the case of the Karaçay valley section where dyke swarms (sites KD01–02) cross-cut the layered sequence (sites KC01– 04). Back-stripping the mean net tectonic rotation calculated for sites KD01–02 from the SMVs and present day poles to layering at sites KC01–04 restores the SMVs to a direction indistinguishable from the reference magnetization vector (Fig. 8a) and the layering to an approximately shallow easterly dipping orientation (note: this analysis provides additional justification for preferring anticlockwise solutions for the dyke sites since the alternative clockwise solutions result in layering dipping at unrealistic initial angles in excess of 70°). In addition, the orientation of dyke screens cutting the layered sequence in the Karaçay valley is in close agreement with the orientation of dykes in the main Karaçay sheeted dyke section, suggesting that the lower crustal layered sequences and sheeted dyke sequences have experienced little or no relative rotation. Hence, although no dykes cross-cutting the layered sequence were observed in the coastal section, it is appropriate to use the net tectonic rotation solutions from the northern section of the coastal sheeted dyke complex to restore the orientation of the layered sequence (to the immediate north) to its initial, pre-deformational position. Again, this results in an inferred initial orientation that dips shallowly towards the east (Fig. 8a) and SMVs close to the reference direction. The significance of these restored orientations is discussed below. Finally, net tectonic rotation solutions from localities in the extrusive sequences of the ophiolite within the NE massif (to the NE

of the Tahtaköprü Fault; Fig. 1b) indicate large anticlockwise rotations around moderate to steeply plunging W to WNW-directed axes (Fig. 7b). The analysis yields broad distributions of rotation angles with modal values of approximately 90–120°, apart from at site ML01 which yields higher rotation angles of 150–160°. This latter site is isolated from the other sites at the Kömürçukuru locality, which consist of semi-continuous exposures of pillowed and sheet flows with consistent orientations, and the result from this site is considered to be anomalous. Rotation angles at remaining sites are significantly larger than those seen in the main southwestern massif of the ophiolite. The structural style of the NE ophiolitic massif is characterised by tectonic juxtaposition by low angle normal faulting of upper crustal extrusive rocks with underlying serpentinized mantle peridotites (Dilek and Thy, 1998). The pillow and sheet flows are locally steeply dipping, e.g. in the Tahtaköprü locality, suggesting a significant component of tilting during normal fault displacement. Hence the larger net rotation angles may reflect a combination of tilting around subhorizontal axes and the major component of rotation around more steeply inclined axes seen elsewhere. Higher rotation angles in the NE massif could also potentially result from a small component of relative rotation of the ophiolitic massifs across the Tahtaköprü Fault. This structure has been interpreted as an accommodation zone related to ridge segmentation by Dilek and Thy (1998), on the basis of the different internal structures of the massifs on either side of the inferred fault. However, no kinematic information is currently available for this structure, which is not clearly exposed in

114

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

Fig. 7. Results of net tectonic rotation analyses of palaeomagnetic data. (a) Results from sites in the lower crustal layered sequence (ultramafic and gabbroic cumulates) of the main massif of the Hatay ophiolite; (b) Results from sites in the extrusive sequences of the northeastern massif of the Hatay ophiolite (to the NE of the Tahtaköprü Fault).

the field. In either case it is clear that the NE massif has experienced a similar history of large, bulk anticlockwise rotation to that documented for the main SE massif. 5. Discussion 5.1. Potential settings for rotation The analysis above demonstrates that large anticlockwise rotations are ubiquitous throughout the Hatay ophiolite, both geographically and pseudostratigraphically. Such rotations are also evident if a simpler analysis based on standard tilt corrections is performed (although this approach does not facilitate assessment of the effects of uncertainties in magnetization vectors and structural orientations). Differences in rotation angle between localities are minor compared to the magnitude of calculated net rotations. Overall the data are consistent with bulk rotation of the entire ophiolite thrust sheet of the order of 90°, but this could be of composite origin reflecting deformation under one or more of the following rotation scenarios: (i) deformation during crustal accretion by sea-floor spreading, e.g. at a ridge-transform intersection, as inferred for parts of the Troodos ophiolite (see review by Morris et al., 2006); (ii) post-spreading oceanic rotation, e.g. as part of a rotating microplate;

(iii) thrust emplacement onto the Arabian continental margin during the Maastrichtian; (iv) post-emplacement neotectonic deformation associated with the development of the modern plate tectonic configuration of the eastern Mediterranean. Regarding scenario (i), transform-related tectonism has been well characterised within the coeval Troodos ophiolite, and is restricted to areas adjacent to the fossil Southern Troodos Transform Fault Zone (STTFZ; MacLeod et al., 1990). The key evidence for transform tectonismrelated rotations is the presence of cross-cutting igneous units with significantly different directions of magnetization, indicating synchronous magmatic activity and fault block rotation. These relationships are not observed in the Hatay ophiolite, nor are the very large and localised differences in magnetization directions seen in the STTFZ. Scenario (iv) is inconsistent with palaeomagnetic data from the Tertiary sedimentary cover sequences of the Hatay ophiolite (Kissel et al., 2003; Inwood, 2005) that indicate magnetization directions close to present day north (combined mean direction: Dec=347°, Inc=36°, α95 =16°). These data suggest only minor post emplacement anticlockwise rotation has occurred, compatible with a small anticlockwise rotation of the Arabian Plate (e.g. Kissel et al., 2003; Gürsoy et al., 2009). These considerations therefore lead to the conclusion that the large bulk rotation of the underlying ophiolite most likely reflects oceanic microplate rotation (scenario (ii)), emplacement-related rotations (scenario (iii)), or a combination of both.

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

115

Fig. 8. Restoration of cumulate layering and dyke strikes to their pre-deformational orientations. (a) Back-stripping the mean net tectonic rotations observed in sheeted dyke complex sites KD01–02 and CD06–11 from the SMVs and orientation of layering in the associated layered sequences. This yields initial orientations of SMVs (circles) close to the reference magnetization vector (star) and cumulate layering dipping initially to the east (great circles and associated poles (crosses)). Small circles (shaded: KC01–04; unshaded: CC01–02) around SMVs indicate the associated α95 envelopes; (b) The distribution of initial dyke strikes at 24 sheeted dyke sites (3000 solutions) resulting from the preferred anticlockwise solutions of the net tectonic analysis approach.

5.2. Oceanic microplate rotation The Troodos and Hatay ophiolites are interpreted as remnants of oceanic lithosphere formed in the same southern Neotethyan ocean basin (Robertson, 2002), and are petrogenetically-related (Lytwyn and Casey, 1993). Although independent rotation of these ophiolites cannot be excluded, the sense and similar magnitude of rotation in each suggests a linked rotation history. Palaeomagnetic studies of the Troodos ophiolite and its in situ sedimentary cover show that the 90° anticlockwise palaeorotation of the Troodos microplate occurred over an extended period from the Late Cretaceous to the Early Eocene (Clube et al., 1985; Clube and Robertson, 1986; Morris et al., 1990), with 50–60° of rotation complete by the Maastrichtian, i.e. the time of tectonic emplacement of the Neotethyan crust now preserved in the Hatay and Baër-Bassit ophiolites. Hence, a major component of the bulk rotation of the Hatay unit documented here may have taken place as part of a “Troodos microplate”. In this scenario, impingement of the passive margin of Arabia with an intraoceanic subduction zone in the Late Cretaceous initiated anticlockwise rotation of the overriding supra-subduction zone crust, following previous models suggested by Clube and Robertson (1986) and Robertson (1990). As Africa–Eurasia convergence continued, a fragment of the rotated oceanic crust above the subduction zone became detached and emplaced upon the Arabian continental margin in the Maastrichtian (to form the Hatay/ Baër-Bassit ophiolites), but rotation of crust to the west (now represented by the Troodos ophiolite) continued until the Early Eocene. Clube and Robertson (1986) inferred that the Troodos microplate was essentially confined to the present day area of Cyprus. The northern boundary of the rotated unit was believed to lie between the Troodos Massif and the Kyrenia Range (northern Cyprus; Robertson and Woodcock, 1986), but would restore further north when Eocene and Late Miocene regional shortening is accounted for. The eastern boundary was placed between the Troodos and Hatay ophiolites (Clube and Robertson, 1986) on the basis of the difference in present day orientation between dykes in the sheeted complexes of these ophiolites. However, we demonstrate that this difference is a primary one (see below) and not the result of large relative rotations between the terranes. Geological constraints therefore do not delimit the present day eastwards extent of the Troodos microplate. Further palaeomagnetic investigations are now required in the more easterly emplaced ophiolites in southern Turkey (e.g. the Göksun, Ispendere,

Kömürhan and Guleman ophiolites; Rizaoglu et al., 2006; Robertson et al., 2007) in order to identify the areal extent of units that have experienced substantial anticlockwise rotation. However, we note that large clockwise rotations have been documented in the Oman (Semail) ophiolite (e.g. Weiler, 2000) at the eastern end of the ‘Ophiolitic Crescent’ (Fig. 1a). This suggests that the plate-scale geometry of the Arabian continental ‘indentor’ and its interaction with Neotethyan subduction trenches prior to final collision provided the key control on the distribution of anticlockwise- and clockwiserotated Neotethyan oceanic crust. Modern oceanic microplate rotation is observed in two tectonic settings: (a) between overlapping spreading centres in fast spreading mid ocean ridge systems, where propagation of offset ridge tips results in edge-driven rotation of the intervening oceanic lithosphere (e.g. the Juan Fernandez, Easter and Galapagos microplates, East Pacific Rise; Searle et al., 1993; Klein et al., 2005); and (b) on a range of scales in the forearc regions of convergent plate margins (e.g. offshore Papua New Guinea; Wallace et al., 2004; Martinez and Taylor, 1996), where collision of continental crust with subduction trenches results in the fastest GPS-determined tectonic block rotations (Wallace et al., 2005). This latter setting provides the closest modern analogue for Neotethyan oceanic microplate rotation. For example, collision of continental crust of the New Guinea Highlands with the Finisterre arc is regarded as the trigger for rapid rotation of the over-riding South Bismarck Sea plate (Wallace et al., 2004). Analysis of GPS data from such regions led Wallace et al. (2005) to propose a general mechanism for rotation in which a change from collision of a buoyant indentor to normal subduction along the strike of a convergent margin exerts a torque on the upper plate, leading to microplate rotation. This is directly comparable to the inferred plate configuration of southern Neotethys in the Late Cretaceous, with a transition from collision of the Arabian continental margin with a subduction trench in the east to ocean–ocean subduction further to the west (Clube and Robertson, 1986; Robertson, 1990). 5.3. Rotation during ophiolite emplacement Tectonic rotations during thrust sheet emplacement have commonly been inferred from palaeomagnetic data in convergent zones (e.g. McClelland and McCaig, 1989; Allerton, 1998; Muttoni et al., 2000; Platt et al., 2003). However, extreme rotations in such systems are localised phenomena within thrust zones that have experienced

116

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117

lower rotations overall (e.g. Platt et al., 2003). More commonly, thrustrelated rotations are of the order of a few 10s of degrees. It is unlikely, therefore that all of the Hatay bulk rotation could be attributed to thrust sheet rotation during tectonic emplacement. However, such a mechanism may reasonably be invoked to account for a residual rotation of 30–40° if it is assumed that the Hatay crust formed part of the Troodos microplate and rotated by 50–60° as part of this unit prior to emplacement in the Maastrichtian, as outlined above. Emplacement-related processes may also potentially account for observed smaller variations in rotation between sampling localities. Small variations in rotation history may also result from post-emplacement faulting (see Inwood et al., in press). Our preferred model of sequential intraoceanic microplate- and emplacement-related tectonic rotations is also consistent with the highly-rotated palaeomagnetic data reported previously from the structurally dismembered and less extensive Baër-Bassit ophiolite exposed to the south of Hatay in northern Syria (Morris et al., 2002; Morris and Anderson, 2002). This unit is interpreted as the highly deformed leading edge of the emplaced ophiolite sheet (Al-Riyami et al., 2002). However, in contrast to the Hatay ophiolite, intraoceanic and emplacement-related rotations in the Baër-Bassit unit are extensively modified by variable postemplacement rotations (Morris and Anderson, 2002) relating to the development of a major neotectonic strike-slip fault system that represents the expression of the plate boundary zone between the African (Arabian) and Eurasian (Anatolian) plates (Al-Riyami et al., 2000, 2002). 5.4. Initial dyke orientations The net tectonic rotation algorithm employed in our analysis (Allerton and Vine, 1987) provides constraints on initial dyke orientations in the case of data from sheeted dyke sections. The preferred anticlockwise solutions (Table 2) yield a restored dyke trend of ~ 020° (Fig. 8b). This direction represents the best estimate of the orientation of the Neotethyan spreading axis responsible for the generation of the Hatay ophiolitic crust. Similar analyses in the sheeted dyke complex of the Troodos ophiolite (Allerton, 1989) indicate an original average restored trend of 325°, when remanence data are compared to the westerly directed Troodos magnetization vector as a reference direction. Correcting for the palaeorotation of the Troodos ophiolite these data indicate an average orientation of 053° for the Troodos spreading axis. Hence, our data identify for the first time substantial differences (c. 33°) in the primary orientation of spreading axes within the southern Neotethyan ocean, between the Troodos and Hatay ridge segments. Differences of this magnitude have been observed in the complex spreading geometries developed in present-day marginal basins (e.g. Lau Basin, Taylor et al., 1996; Philippine Sea, Sdrolias et al., 2004; Manus Basin, Martinez and Taylor, 1996), further supporting the analogy between these modern systems and ancient supra-subduction zone spreading systems like the southern Neotethys. Finally, the geometric relationships revealed by our analyses (Fig. 8) suggest that by the time of remanence acquisition at sites KC01–04 the cumulate layering in the Karaçay sequences had been tilted by an average of 25°, and were subsequently intruded by the dyke screens sampled at sites KD01–02. We note the coincidence between the restored N–NNE strikes of the dykes and cumulate layering (Fig. 8). This is compatible with some models for the development of cumulate sequences involving downwards ductile flow of hot crystal mushes, resulting in layering that dips towards the spreading centre (e.g. Quick and Denlinger, 1993). 6. Conclusions The Hatay ophiolite of southern Turkey retains stable magnetizations which are unequivocally shown to be of pre-deformational

origin by a positive inclination-only tilt test. Net tectonic rotation analysis of magnetization directions reveals ubiquitously large anticlockwise rotations at all sampling localities. The data are best explained by a model involving 50–60° rotation of the Hatay crust as part of an intraoceanic “Troodos” microplate within the southern Neotethyan basin, followed by a subsequent 30–40° anticlockwise rotation during thrust emplacement of the Hatay ophiolite onto the Arabian continental margin in the Maastrichtian. Minor variability in rotation solutions between localities may be due to a combination of emplacement- and post-emplacement-related faulting. The platescale geometry of the Arabian ‘indentor’ and its interaction with a segment of subduction trench, while normal subduction continued to the immediate west, is likely to have provided the key control on intraoceanic rotation of Neotethyan crust. A similar rotation mechanism has been proposed to explain rapid microplate rotation in present day complex convergent plate boundaries (Wallace et al., 2005). The geometric analysis of net tectonic rotations indicates an initial orientation of 020° for the Hatay sheeted dykes and hence the associated spreading axis. This trend differs from that inferred for the coeval Troodos spreading axis, revealing primary differences in ridge orientations within the southern Neotethys, comparable to those seen in some modern marginal basin spreading systems. Acknowledgements We acknowledge the support of Ulvican Ünlügenç (Çukurova University, Adana) throughout this project. We would also like to thank our field assistants, particularly Ali Kop. Inclination-only tilt tests were performed using software developed by Randy Enkin. A Visiting Fellowship (2004) from the Institute for Rock Magnetism in Minneapolis, University of Minnesota, USA, enabled the laboratory facilities at the Institute to be used to perform rock magnetic analyses. This research was financed by a University of Plymouth PhD studentship (Inwood) with additional financial support for fieldwork from the Geological Society of London (Fund for Fieldwork, 2002–3 and 2003–4) and the British Federation for Women Graduates (Johnstone and Florence Stoney Studentship, 2003–4). The authors would like to thank reviewers Rob Van der Voo and an anonymous reviewer for their constructive comments.

References Abrahamsen, N., Schonharting, G., 1987. Palaeomagnetic timing of the rotation and translation of Cyprus. Earth Planet. Sci. Lett. 81, 409–418. Allerton, S., 1989. Fault block rotations in ophiolites: results of palaeomagnetic studies in the Troodos complex, Cyprus. In: Kissel, C., Laj, C. (Eds.), Palaeomagnetic Rotations and Continental Deformation. NATO ASI Series C, vol. 254, pp. 393–410. Allerton, S., 1998. Geometry and kinematics of vertical-axis rotation in fold and thrust belts. Tectonophysics 299, 15–30. Allerton, S., Vine, F.J., 1987. Spreading structure of the Troodos ophiolite, Cyprus: some palaeomagnetic constraints. Geology 15, 593–597. Allerton, S., Vine, F.J., 1991. Spreading evolution of the Troodos ophiolite, Cyprus. Geology 19, 637–640. Al-Riyami, K., Robertson, A.H.F., Xenophontos, C., Danelian, T., Dixon, J.E., 2000. Mesozoic tectonic and sedimentary evolution of the Arabian continental margin in Baër-Bassit NW Syria. In: Malpas, J., Xenophontus, C., Panayides, A. (Eds.), Proceedings of the 3rd International Conference on the Geology of the Eastern Mediterranean, Nicosia, Cyprus, pp. 61–81. Al-Riyami, K., Robertson, A.H.F., Dixon, J., Xenophontos, C., 2002. Origin and emplacement of the Late Cretaceous Baër-Bassit ophiolite and its metamorphic sole in NW Syria. Lithos 651-2, 225–260. Aslaner, M., 1973. Iskenderun-Kirikhan bolgesindeki ofiyolitlerin jeoloji ve petrografisi. Bulletin M. T. A. Ankara. Besse, J., Courtillot, V., 1991. Revised and synthetic apparent polar wander paths of the African, Eurasian and Indian plates, and true polar wander since 200 Ma. J. Geophys. Res. 96 (B3), 4029–4050. Besse, J., Courtillot, V., 2002. Apparent and true polar wander and the geometry of the geomagnetic field over the last 200 Myr. J. Geophys. Res. 107 (B11). doi:10.1029/ 2000JB000050. Bonhommet, N., Roperch, P., Calza, F., 1988. Paleomagnetic arguments for block rotations along the Arakapas Fault (Cyprus). Geology 16, 422–425. Borradaile, G.J., 2001. Paleomagnetic vectors and tilted dikes. Tectonophysics 333, 417–426.

J. Inwood et al. / Earth and Planetary Science Letters 280 (2009) 105–117 Boulton, S.J., Robertson, A.H.F., 2007. The Miocene of the Hatay area, S Turkey: transition from the Arabian passive margin to an underfilled foreland basin related to closure of the Tethys Ocean. Sediment. Geol. 198, 93–124. Boulton, S.J., Robertson, A.H.F., Unlugenc, U.C., 2006. Tectonic and sedimentary evolution of the Cenozoic Hatay Graben, Southern Turkey: a two-phase model for graben formation. In: Robertson, A.H.F., Mountrakis, D. (Eds.), Tectonic Development of the Eastern Mediterranean Region. Special Publication of the Geological Society of London, vol. 260, pp. 613–634. Cande, S.C., Kent, D.V., 1992. A new geomagnetic polarity timescale for the Late Cretaceous and Cenozoic. J. Geophys. Res. 97, 13917–13951. Clube, T.M.M., Robertson, A.H.F., 1986. The palaeorotation of the Troodos microplate, Cyprus, in the Late Mesozoic–Early Cenozoic plate tectonic framework of the Eastern Mediterranean. Surv. Geophys. 8, 375–437. Clube, T.M.M., Creer, K.M., Robertson, A.H.F., 1985. The palaeorotation of the Troodos microplate. Nature 317, 522–525. Day, R., Fuller, M., Schmidt, V.A., 1977. Hysteresis properties of titanomagnetites: grain size and composition dependence. Phys. Earth Planet. Inter. 13, 260–267. Delaloye, M., Wagner, J.J., 1984. Ophiolites and volcanic activity near the western edge of the Arabian plate. In: Dixon, J.E., Robertson, A.H.F. (Eds.), The Geological Evolution of the Eastern Mediterranean. Special Publication of the Geological Society of London, vol. 17, pp. 225–234. Delaloye, M., Piskin, O., Selcuk, H., Vuagnat, M., Wagner, J.J., 1980. Geological section through the Hatay ophiolite along the Mediterranean coast, southern Turkey. Ofioliti 52/3, 205–216. Dilek, Y., Thy, P., 1998. Structure, petrology and seafloor spreading tectonics of the Kizildag Ophiolite, Turkey. Spec. Publ. Soc. Lond. 148, 43–70. Enkin, R.J., Watson, G.S., 1996. Statistical analysis of palaeomagnetic inclination data. Geophys. J. Int. 126, 495–504. Erendil, M., 1984. Petrology and structure of the upper crustal units of the Kizil Dag ophiolite. In: Tekeli, O., Concuoglu, C.M. (Eds.), Geology of the Taurus Belt, Proceedings, pp. 269–284. Fisher, R.A., 1953. Dispersion on a sphere. Proc. R. Soc. Lond., Ser. A 217, 295–305. Gürsoy, H., Tatar, O., Piper, J.D.A., Heimann, A., Koçbulut, F., Mesci, B.L., 2009. Palaeomagnetic study of Tertiary volcanic domains in Southern Turkey and Neogene anticlockwise rotation of the Arabian Plate. Tectonophysics 465, 114–127. doi:10.1016/j.tecto.2008.11.001. Hurst, S.D., Verosub, K.L., Moores, E.M., 1992. Paleomagnetic constraints on the formation of the Solea Graben, Troodos Ophiolite, Cyprus. Tectonophysics 208, 431–445. Inwood, J., 2005. The tectonic evolution of the Hatay (Kizil Dag) ophiolite, southern Turkey. PhD thesis. University of Plymouth. Inwood, J., Anderson, M.W., Morris, A., Robertson, A.H.F., in press. Successive structural events in the Hatay ophiolite of southeast Turkey: distinguishing oceanic, emplacement and post-emplacement phases of faulting. Tectonophysics, Special Issue. doi:10.1016/j.tecto.2008.10.037. Kirker, A., McClelland, E., 1996. Application of net tectonic rotation and inclination analysis to a high-resolution palaeomagnetic study in the Betic Cordillera. Spec. Publ. — Geol. Soc. Lond. 105, 19–32. Kirschvink, J.L., 1980. The least-squares line and plane and the analysis of paleomagnetic data. Geophys. J. R. Astron. Soc. 62, 699–718. Kissel, C., Laj, C., Poisson, A., Gorur, N., 2003. Paleomagnetic reconstruction of the Cenozoic evolution of the Eastern Mediterranean. Tectonophysics 362, 199–217. Klein, E.M., Smith, D.K., Williams, C.M., Schouten, H., 2005. Counter-rotating microplates at the Galapagos triple junction. Nature 433, 855–858. Lytwyn, J.N., Casey, J.F., 1993. The geochemistry and petrogenesis of volcanics and sheeted dikes in the Hatay (Kizildag) ophiolite, southern Turkey: possible formation with the Troodos ophiolite, Cyprus, along fore-arc spreading centres. Tectonophysics 223, 237–272. MacDonald, W.D., 1980. Net tectonic rotation, apparent rotation, and the structural tilt correction in palaeomagnetic studies. J. Geophys. Res. 85, 3659–3669. MacLeod, C.J., Allerton, S., Gass, I.G., Xenophontos, C., 1990. Structure of a fossil ridgetransform intersection in the Troodos ophiolite. Nature 348, 717–720. Martinez, F., Taylor, B., 1996. Backarc spreading, rifting and microplate rotation between transform faults in the Manus Basin. Mar. Geophys. Res. 18, 203–224. McClelland, E., McCaig, A.M., 1989. Palaeomagnetic estimates of rotations in compressional regimes and potential discrimination between thin-skinned and deep crustal deformation In: Kissel, C., Laj, C. (Eds.), Palaeomagnetic Rotations and Continental Deformation. NATO ASI Series C, vol. 254, pp. 365–379. McElhinny, M.W., 1964. Statistical significance of the fold test in palaeomagnetism. Geophys. J. R. Astron. Soc. 8, 338–340. McFadden, P.L., Jones, D.L., 1981. The fold test in palaeomagnetism. Geophys. J. R. Astron. Soc. 67, 53–58. Morris, A., Anderson, M.W., 2002. Palaeomagnetic results from the Baër-Bassit ophiolite of northern Syria and their implication for fold tests in sheeted dyke terrains. Phys. Chem. Earth Parts A B C 27 (25–31), 1215–1222.

117

Morris, A., Creer, K.M., Robertson, A.H.F., 1990. Palaeomagnetic evidence for clockwise rotations related to dextral shear along the Southern Troodos Transform Fault, Cyprus. Earth Planet. Sci. Lett. 99, 250–262. Morris, A., Anderson, M.W., Robertson, A.H.F., 1998. Multiple tectonic rotations and transform tectonism in an intra-oceanic suture zone, SW Cyprus. Tectonophysics 299, 229–253. Morris, A., Anderson, M.W., Robertson, A.H., Al-Riyami, K., 2002. Extreme tectonic rotations within an eastern Mediterranean ophiolite Baër-Bassit, Syria. Earth Planet. Sci. Lett. 202, 247–261. Morris, A., Anderson, M.W., Inwood, J., Robertson, A.H.F., 2006. Palaeomagnetic insights into the evolution of Neotethyan oceanic crust in the eastern Mediterranean. In: Robertson, A.H.F., Mountrakis, D. (Eds.), Tectonic Development of the Eastern Mediterranean Region. Geological Society of London Special Publication, vol. 260, pp. 351–372. Muttoni, G., Lanci, L., Argnani, A., Hirt, A.M., Cibin, U., Abrahamsen, N., Lowrie, W., 2000. Paleomagnetic evidence for a Neogene two-phase counterclockwise tectonic rotation in the Northern Apennines (Italy). Tectonophysics 326, 241–253. Piskin, O., Delaloye, M., Wagner, J.J., 1986. Guide to the Hatay geology. Ofioliti 11, 87–104. Platt, J.P., Allerton, S., Kirker, A., Mandeville, C., Mayfield, A., Platzman, E.S., Rimi, A., 2003. The ultimate arc: differential displacement, oroclinal bending, and vertical axis rotation in the external Betic-Rif arc. Tectonics 22 (3), 1017. doi:10.1029/ 2001TC001321. Quick, J.E., Denlinger, R.P., 1993. Ductile deformation and the origin of layered gabbro in ophiolites. J. Geophys. Res. 98 (B8), 14015–14027. Ricou, L.E., 1971. Le croissant ophiolitique péri-arabe, une ceinture de nappes mises en place au Cretacé superieur. Rev. Géogr. Phys. Géol. Dyn. 13, 327–349. Rizaoglu, T., Parlak, O., Hoeck, V., Fikret, I., 2006. Nature and significance of Late Cretaceous ophiolitic rocks and their relationship to the Baskil granitic intrusions of the Elazig region, SE Turkey. In: Robertson, A.H.F., Mountrakis, D. (Eds.), Tectonic Development of the Eastern Mediterranean Region. Geological Society of London Special Publication, vol. 260, pp. 327–350. Robertson, A.H.F., 1986. The Hatay ophiolite (southern Turkey) in its eastern Mediterranean tectonic context: a report on some aspects of the field excursion. Ofioliti 11, 105–119. Robertson, A.H.F., 1990. Tectonic evolution of Cyprus. In: Malpas, J., Moores, E.M., Panayiotou, A., Xenophontos, C. (Eds.), Ophiolites: Oceanic Crustal Analogues. Cyprus Geological Survey Department, pp. 235–252. Robertson, A.H.F., 2002. Overview of the genesis and emplacement of Mesozoic ophiolites in the Eastern Mediterranean Tethyan region. Lithos 65 (1–2), 1–67. Robertson, A.H.F., Woodcock, N.H., 1986. The role of the Kyrenia Range Lineament, Cyprus, in the geological evolution of the eastern Mediterranean area. Philos. Trans. R. Soc. Lond., A 317, 141–177. Robertson, A.H.F., Parlak, O., Rizaoglu, T., Ünlügenç, U.C., Inan, N., Taslí, K., Ustaömer, T., 2007. Tectonic Evolution of the South Tethyan ocean: evidence from the Eastern Taurus Mountains (Elazig region, SE Turkey). In: Ries, A.C., Butler, R.W.H., Graham, R.H. (Eds.), Deformation of the Continental Crust: The Legacy of Mike Coward. Geological Society of London Special Publication, vol. 272, pp. 231–270. Sdrolias, M., Roest, W.R., Müller, D.R., 2004. An expression of Philippine Sea plate rotation: the Parece Vela and Shikoku Basins. Tectonophysics 394 (1–2), 69–86. Searle, R.C., Bird, R.T., Rusby, R.I., Naar, D.F., 1993. The development of two oceanic microplates: Easter and Juan Fernandez microplates, East Pacific Rise. J. Geol. Soc. Lond. 150, 965–976. Taylor, B., Zellmer, K., Martinez, F., Goodliffe, A., 1996. Sea-floor spreading in the Lau back-arc basin. Earth Planet. Sci. Lett. 144 (1–2), 35–40. Tinkler, C., Wagner, J.J., Delaloye, M., Selcuk, H., 1981. Tectonic history of the Hatay ophiolites south Turkey and their relation with the Dead Sea rift. Tectonophysics 72, 23–41. Wallace, L.M., Stevens, C., Silver, E., McCaffrey, R., Loratung, W., Hasiata, S., Stanaway, R., Curley, R., Rosa, R., Taugaloidi, J., 2004. GPS and seismological constraints on active tectonics and arc–continent collision in Papua New Guinea: implications for mechanics of microplate rotations in a plate boundary zone. J. Geophys. Res. 109. doi:10.1029/2003JB002481. Wallace, L.M., McCaffrey, R., Beavan, J., Ellis, S., 2005. Rapid microplate rotations and backarc rifting at the transition between collision and subduction. Geology 33, 857–860. Weiler, P.D., 2000. Differential rotations in the Oman ophiolite: paleomagnetic evidence from the southern massifs. Mar. Geophys. Res. 21, 195–210. Westphal, M., Bazhenov, M.L., Lauer, J.P., Pechersky, D.H., Sibuet, J.C., 1986. Palaeomagnetic implications on the evolution of the Tethys Belt from the Atlantic Ocean to the Pamirs since the Triassic. Tectonophysics 123, 37–82.