Jurassic back-arc and Cretaceous hot-spot series In the Armenian ophiolites — Implications for the obduction process

Abstract

The identification of a large OIB-type volcanic sequence on top of an obducted nappe in the Lesser Caucaus of Armenia helps us explain the obduction processes in the Caucasus region that are related to dramatic change in the global tectonics of the Tethyan region in the late Lower Cretaceous. The ophiolitic nappe preserves three distinct magmatic series, obducted in a single tectonic slice over the South Armenian Block during the Coniacian–Santonian (88–83 Ma), the same time as the Oman ophiolite. Similar geological, petrological, geochemical and age features for various Armenian ophiolitic massifs (Sevan, Stepanavan, and Vedi) argue for the presence of a single large obducted ophiolite unit. The ophiolite, shows evidence for a slow-spreading oceanic environment in Lower to Middle Jurassic. Serpentinites, gabbros and plagiogranites were exhumed by normal faults, and covered by radiolarites. Few pillow-lava flows have infilled the rift grabens. The ophiolite lavas have hybrid geochemical composition intermediate between Arc and MORB signatures: (La/Yb)_N = 0.6–0.9; (Nb/Th)_N = 0.17–0.57; (¹⁴³Nd/¹⁴⁴Nd)_i = 0.51273–0.51291; (⁸⁷Sr/⁸⁶Sr)_i = 0.70370–0.70565; (²⁰⁷Pb/²⁰⁴Pb)_i = 15.4587–15.5411; (²⁰⁸Pb/²⁰⁴Pb)_i = 37.4053–38.2336; (²⁰⁶Pb/²⁰⁴Pb)_i = 17.9195–18.4594. These compositions suggest they were probably formed in a back-arc basin by melting of a shallow asthenosphere source contaminated by a deeper mantle source modified by subducted slab-derived products. ⁸⁷Sr/⁸⁶Sr ratios and petrological evidence show that these lavas have been intensely altered by mid-oceanic hydrothermalism as well as by serpentinites, which are interpreted as exhumed mantle peridotites. The gabbros have almost the same geochemical composition as related pillow-lavas: (La/Yb)_N = 0.2–2.3; (Nb/Th)_N = 0.1–2.8; (¹⁴³Nd/¹⁴⁴Nd)_i = 0.51264–0.51276; (⁸⁷Sr/⁸⁶Sr)_i = 0.70386–0.70557; (²⁰⁷Pb/²⁰⁴Pb)_i = 15.4888–15.5391; (²⁰⁸Pb/²⁰⁴Pb)_i = 37.2729–37.8713; (²⁰⁶Pb/²⁰⁴Pb)_i = 17.6296–17.9683. Plagiogranites show major and trace element features similar to other Neo-Tethyan plagiogranites (La/Yb)_N = 1.10–7.92; (Nb/Th)_N = 0.10–0.94; but display a less radiogenic Nd isotopic composition than basalts [(¹⁴³Nd/¹⁴⁴Nd)_i = 0.51263] and more radiogenic (⁸⁷Sr/⁸⁶Sr)_i ratios. This oceanic crust sequence is covered by variable thicknesses of unaltered pillowed OIB alkaline lavas emplaced in marine conditions. ⁴⁰Ar/³⁹Ar dating of a single-grain amphibole phenocryst provides a Lower Cretaceous age of 117.3 ± 0.9 Ma, which confirms a distinct formation age of the OIB lavas. The geochemical composition of these alkaline lavas is similar to plateau-lavas [(La/Yb)_N = 6–14; (Nb/Th)_N = 0.23–0.76; (¹⁴³Nd/¹⁴⁴Nd)_i = 0.51262–0.51271; (⁸⁷Sr/⁸⁶Sr)_i = 0.70338–0.70551; (²⁰⁷Pb/²⁰⁴Pb)_i = 15.5439–15.6158; (²⁰⁸Pb/²⁰⁴Pb)_i = 38.3724–39.3623; (²⁰⁶Pb/²⁰⁴Pb)_i = 18.4024–19.6744]. They have significantly more radiogenic lead isotopic compositions than ophiolitic rocks, and fit the geochemical compositions of hot-spot derived lavas mixed with various proportions of oceanic mantle. In addition, this oceanic + plateau sequence is covered by Upper Cretaceous calc-alkaline lavas: (La/Yb)_N = 2.07–2.31; (Nb/Th)_N = 0.08–0.15; (¹⁴⁴Nd/¹⁴³Nd)_i = 0.51271–0.51282; (⁸⁷Sr/⁸⁶Sr)_i = 0.70452–0.70478), which were likely formed in a supra-subduction zone environment. During the late Lower to early Upper Cretaceous period, hot-spot related magmatism related to plateau events may have led to significant crustal thickening in various zones of the Middle-eastern Neotethys. These processes have likely hindered subduction of some of the hot and thickened oceanic crust segments, and allowed them to be obducted over small continental blocks such as the South Armenian Block.

Introduction

The role of Oceanic Plateaus in the obduction processes of oceanic crust has still not been clearly established. We understand that their larger crustal thickness and buoyancy as compared to ‘standard’ oceanic crust does not allow them to subduct, in particular when they reach subduction zones soon after their formation (e.g., Ben-Avraham et al., 1981, Cloos, 1993, Abbot and Mooney, 1995, Abbot et al., 1997, Kerr and Mahoney, 2007). However, the reasons for oceanic crust obduction onto continental margins are still debated: (i) is obduction driven by subduction of continental crust? Or (ii) does it result from the intrinsic nature of the oceanic crust? In the first case, ophiolites are obducted due to the mechanical coupling of continental crust with the dense subducting slab (e.g., O'Brien et al., 2001, Guillot et al., 2003). Continental subduction may be facilitated by the thinned margins of continental domains following earlier phases of divergence rifting that precede oceanic crust emplacement (Guillot and Allemand, 2002). In the second case, a lower density of oceanic lithosphere might result from intra-oceanic hot-spot and magmatic arc events, which will lead to crustal thickening and a decrease in lithosphere density (Cloos, 1993, Abbot and Mooney, 1995). The emplacement of oceanic plateaus has a great influence either on the slab dip, but also on the cessation of subduction and on the onset of obduction as is proposed for the Ontong–Java plateau (Petterson et al., 1997). The ability of an oceanic plateau to resist subduction and eventually be transported onto continental crust depends on both crustal thickness and plateau age (Kerr and Mahoney, 2007). The older a plateau, the cooler and thus the less buoyant it will be. Alternative hypotheses for obduction involve rapid inversion of tectonic plate motions and rapid continental convergence (e.g., Agard et al., 2007). Obduction is ascribed to the presence of young oceanic crust in the hanging-wall of the subduction zone, as a result of subduction initiation at the Mid-Oceanic Ridge (e.g., Boudier et al., 1988, Nicolas, 1989); or to scalping of oceanic lithosphere (e.g., Agard et al., 2007 and references therein).

The case of Armenian ophiolites (Lesser Caucasus) is peculiar as recent investigations (Galoyan et al., 2007, Galoyan et al., 2009, Rolland et al., in press) have shown the presence of slow-spreading ophiolites in several locations. Further, the ophiolites were tectonically transported above the South Armenian Bloc or SAB (Zakariadze et al., 1983). Although some blueschists are locally found, these affect oceanic crust-derived rocks which underwent intra-oceanic subduction and exhumation within accretionary prisms (Rolland et al., 2009). In contrast, the underthrusted Armenian continental crust appears not to have been metamorphosed by any subduction event. Therefore, the obduction of the Armenian ophiolites might be explained by the intrinsic nature of the oceanic crust. However, the slow-spreading nature of the ophiolites, and in particular the fact that exhumed mantle forms a large part of the reconstructed ophiolite is rather in agreement with a relatively dense oceanic lithosphere.

In this paper, we report new geochemical data, including major and trace elements and Nd, Sr, Pb isotopes, on magmatic series from several Armenian ophiolites (i.e. Stepanavan (NW Armenia), Sevan (N Armenia), Vedi (central Armenia); Fig. 2). We identify three superposed levels of lavas corresponding to three distinct environments: (1) back-arc, (2) ‘OIB’-like and (3) arc. Moreover, we suggest that these ophiolite windows correlate with each other and be part of a unique obducted nappe. Tectonic transport of this nappe onto the SAB can be dated to the Coniacian–Santonian (88–83 Ma; Sokolov, 1977, Sosson et al., in press). Finally, the influence of oceanic plateau event in oceanic lithosphere rheology and its role in the obduction process is discussed.