The Late Triassic Longmu Co–Shuanghu suture zone is a metamorphic belt in Central Qiangtang, SW China, that is interpreted to have formed at the boundary between a Gondwana-derived block and Laurasia during the subduction of Paleo-Tethyan oceanic crust. Recent research has suggested that a Late Triassic (ca. 225–205 Ma) magmatic “flare-up” event took place on both sides of the metamorphic belt within Central Qiangtang coeval with exhumation of the metamorphic rocks. The age-equivalent Gangmari metamorphic belt and Riwanchaka Yangtze-type deposits of the South Qiangtang terrane are located ∼70 km from the Longmu Co–Shuanghu suture zone and the North Qiangtang terrane. We propose that Central Qiangtang underwent postcollisional extension in the Late Triassic. By integrating the geological features described here, we suggest that these complex geological phenomena were triggered by slab breakoff of the northward-subducting lithosphere of the Paleo-Tethys Ocean. Central Qiangtang is therefore an ideal area in which to verify the process of slab breakoff by geological observations.
The Longmu Co–Shuanghu suture zone is an east-west–trending metamorphic belt in the central part of the Qiangtang terrane (Central Qiangtang), SW China (Fig. 1; Li et al., 1995, 2006; Zhang et al., 2006). The zone consists of blueschist, eclogite, ophiolitic mélange, and metasedimentary rocks, and it is interpreted as a tectonic mélange that developed at the suture between the South and North Qiangtang terranes during subduction of Paleo-Tethys oceanic crust (Li, 2008). Based on distinctive strata, fossil assemblages, and paleomagnetic data prior to the late middle Permian (Li and Zheng, 1993; XZBGM, 1993; Li, 2008), the Longmu Co–Shuanghu suture zone is inferred to represent the boundary between the Gondwana-derived South Qiangtang terrane and the Yangtze-derived North Qiangtang terrane. The South Qiangtang terrane is characterized by metamorphosed glaciomarine diamictites that represent a typical Gondwana facies cold-water sedimentary unit (Li, 2008; Fan et al., 2014). The North Qiangtang terrane is characterized by unmetamorphosed sedimentary units that contain abundant warm-water fossils and are similar to the Yangtze-type sedimentary units of southern China (Li, 2005, 2008).
U-Pb zircon geochronology from ophiolite in the Longmu Co–Shuanghu suture zone suggests that the Paleo-Tethys Ocean had opened by the Early Cambrian (ca. 500 Ma; Wu, 2013; Hu et al., 2014), whereas U-Pb zircon geochronology of the magmatic arc suggests that subduction of oceanic crust began during the Mississippian (Hu et al., 2013; Jiang et al., 2014, 2015). Rocks interpreted as a Late Permian ocean-island unit were recently discovered in the Tianquan area, indicating that the Paleo-Tethys Ocean was still open in the Early Triassic (ca. 254–246 Ma; Fan et al., 2016).
Recently, one granodiorite sample from the Riwanchaka area was selected for zircon U-Pb dating (Fig. 2). Zircon U-Pb dating was performed by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) at the Geological Laboratory Centre, China University of Geosciences, Beijing, using an Agilent 7500a ICP-MS (Agilent Technologies Inc., USA). Detailed operating conditions and data reduction were performed following Wu et al. (2015). The results of the zircon dating are summarized in Table 1. The zircons from the sample are mostly short euhedral prisms, 150–300 μm in length. Most of the zircons are transparent and colorless, and cathodoluminescence (CL) images reveal concentric oscillatory zoning (Fig. 3). In addition, all the zircons have relatively high Th/U ratios (>0.1), in the range 0.15–0.19. These compositional characteristics are typical of magmatic zircons (Hoskin and Schaltegger, 2003). Furthermore, the U-Pb isotope contents of the zircons indicate nearly concordant U-Pb ages for 24 grains on the concordia diagram, and the spot analyses yielded 206Pb/238U ages in the range 214–231 Ma, with a weighted mean age of 222 ± 1 Ma (mean square of weighted deviates [MSWD] = 0.09). In addition, unmetamorphosed magmatic rocks and high-pressure (HP) to ultrahigh-pressure (UHP) metamorphic rocks (ca. 225–205 Ma) have been reported recently in the Central Qiangtang (e.g., Kapp et al., 2003; Li et al., 2006). Both the magmatism and the exhumation of HP-UHP metamorphic rocks occurred in the Late Triassic (Table 2).
Recent research has identified another metamorphic belt, the Gangmari metamorphic belt, some 70 km from the Longmu Co–Shuanghu suture zone. This belt contains blueschist, eclogite, and ophiolitic mélange units (Li, 2008; Zhai et al., 2011a) that define a Mississippian magmatic arc within the Gangmari area of the South Qiangtang terrane (Jiang et al., 2014, 2015; Fig. 2). Peng et al. (2014a) studied detrital zircons and fossils within the Mississippian Riwanchaka Formation in the Riwanchaka area, located adjacent to the Gangmari metamorphic belt, and determined that this formation has a Yangtze affinity and is similar to units within the North Qiangtang terrane. However, other research suggests that a Paleo-Tethyan oceanic slab was subducted northward beneath the North Qiangtang terrane (Zhai et al., 2013a; Peng et al., 2014a; Jiang et al., 2015). The Upper Triassic granodiorite investigated in this study was emplaced in the Riwanchaka Yangtze-type deposits. Consequently, a mechanism is needed to explain the emplacement of the Gangmari metamorphic belt and Riwanchaka Yangtze-type deposits, both of which were originally formed within a subduction zone and the North Qiangtang terrane but are now located within the South Qiangtang terrane, and a geodynamic mechanism for Late Triassic emplacement of magmatic rocks in the Central Qiangtang.
In this paper, we synthesize the geological events that occurred in Central Qiangtang and propose a model for the Late Triassic evolution of the Paleo-Tethys Ocean.
LATE TRIASSIC TECTONIC SETTING
The Upper Triassic Wanghuling Formation is considered to represent the first sedimentary cover after the closure of the Paleo-Tethys Ocean (Li et al., 2007). Paleomagnetism studies indicate that this closure had amalgamated the South and North Qiangtang terranes by the Late Triassic (Song et al., 2012). The Middle Triassic Gaco basalt (233 Ma), which has an ocean-island basalt affinity, seems to have erupted within an accretionary wedge, rather than onto oceanic basement (Zhu et al., 2006). Coeval Tuohepingco diabase (234 Ma) has been interpreted as the product of continental collision (Zhang et al., 2011). Eclogites of Central Qiangtang yield Lu-Hf mineral isochron ages of 233–244 Ma (Pullen et al., 2008) and zircon U-Pb ages of 230–237 Ma (Zhai et al., 2011a), which are interpreted as the timing of peak eclogite-facies metamorphism. Therefore, we suggest that the peak collision between the South and North Qiangtang terranes occurred during the Middle Triassic and that Central Qiangtang was undergoing postcollisional extension in the Late Triassic.
We have collected the published geochemical data from the Central Qiangtang, and their age data and locations are listed in Table 2. We found that Upper Triassic magmatic rocks have a bimodal affinity that is typical of extension-related magmatism (Fig. 4). In addition, plotting these data in the discrimination diagrams of Whalen et al. (1987) and Eby (1992) suggests that the rhyolites and dacites of the Wanghuling and Guoganjianian areas are geochemically consistent with A2-type silicic rocks (Fig. 5), which generally form during the postcollisional stage of orogenesis (Zhai et al., 2013a; Liu et al., 2014). The basalts of the Nadigangri Formation (ca. 220 Ma) are geochemically consistent with within-plate basalt and probably formed in an extensional setting rather than in a continental-margin, arc-related setting (Fu et al., 2010a). In addition, numerical experiments indicate that HP-UHP metamorphic rocks are always exhumed in an extensional setting (Beaumont et al., 2009). Based on these studies, we consider that the Central Qiangtang was an extensional environment in the Late Triassic, which followed the Middle Triassic continent-continent collision of the South and North Qiangtang terranes.
GEODYNAMIC MECHANISM FOR LATE TRIASSIC EMPLACEMENT OF MAGMATIC ROCKS
Continental collision is the natural consequence of continental drift and plate tectonics, and it is usually accompanied by extensive syn- and postcollisional magmatism, triggered by slab breakoff (Davies and Blanckenburg, 1995; van de Zedde and Wortel, 2001; van Hunen and Allen, 2011). In continental collision settings, the closure of the ocean results in subduction of continental lithosphere. Oceanic lithosphere sinks because of its negative buoyancy, whereas the buoyant continental lithosphere usually remains at the surface. This kinematic contrast between oceanic and continental lithosphere ultimately leads to the detachment of the oceanic slab from the continental lithosphere, known as slab breakoff (Davies and Blanckenburg, 1995; van Hunen and Allen, 2011). The slab window that forms as a result of slab breakoff results in partial melting of different source regions, including upwelling asthenospheric mantle, the detached oceanic crust, enriched lithospheric mantle, and even overlying crust. This partial melting produces mafic magmatism that geochemically resembles within-plate basalt, ocean-island basalt, or intermediate-silicic magmatism that forms in an extensional setting (Wortel and Spakman, 1992; Davies and Blanckenburg, 1995; Hildebrand and Bowring, 1999; Ferrari, 2004; van Hunen and Allen, 2011).
Comparatively little is known of the interactions between mantle and crust that occurred in response to the collision of the South and North Qiangtang terranes, compared with the well-documented India-Asia and Lhasa-Qiangtang collision zones. An important feature of Central Qiangtang is the extensive exposure of coeval Upper Triassic intrusions and volcanic rocks on both sides of the metamorphic belt (Fig. 1). Upper Triassic magmatic rocks have been reported from north of the Longmu Co–Shuanghu suture zone, which is consistent with northward subduction (Zhang et al., 2011; Zhai et al., 2013a). Recently, however, increasing numbers of coeval magmatic rocks, including basalts, dacites, rhyolites, adakitic rocks, and I- and S-type granitoids, have been identified within the Central Qiangtang area (Kapp et al., 2003; Wang et al., 2007, 2008; Fu et al., 2008, 2010a, 2010b; Zhai, 2008; Zhai et al., 2009, 2013a; Hu et al., 2010a, 2010b, 2010c, 2014; Chen, 2015; Zhang et al., 2014b; Li et al., 2015a, 2015b). The compositional diversity of Upper Triassic magmatic rocks within the Central Qiangtang area is indicative of multiple deep magma sources. However, normal geothermal gradients are generally insufficient to allow partial melting of the crust (e.g., Sandiford et al., 1998; McLaren et al., 1999). Using our data with the calculation method of Watson and Harrison (1983) indicates that rhyolites of the Wanghuling Formation have high whole-rock zircon saturation temperatures (844 °C to 865 °C, barring two samples that yielded a temperature of 779 °C) and formed from magmas that were generated by mantle-derived heat. In addition, many mafic enclaves occur throughout the Upper Triassic magmatic rocks in this area, providing further evidence of the involvement of mantle-derived material in the genesis of the magmatic rocks (Chen, 2015; Zhang et al., 2014b).
Previous researchers have proposed a model of slab breakoff for the northward-subducting Paleo-Tethys oceanic lithosphere to explain the compositional diversity of the Late Triassic magmatic flare-up (Zhai et al., 2013a; Peng et al., 2014b). However, we consider that the petrological associations of mafic rocks characterized by within-plate basalt geochemical features, adakitic rocks, coeval bimodal magmatic rocks, especially our identification of A2-type silicic rocks, and the high heat flow caused by the upwelling of mantle material provide robust evidence that supports an interpretation of slab breakoff. Moreover, the bulk of the Late Triassic mafic igneous rocks plotted are calc-alkaline, high-K calc-alkaline, and shoshonitic, rather than true tholeiites (Fig. 4B). This result is consistent with Tertiary mafic dikes that were considered to be the product of slab breakoff in the Alps (von Blanckenburg and Davies, 1995).
We thus envisage the following processes to explain the magmatic diversity of the Late Triassic Central Qiangtang. At this time, the contrasting behavior of oceanic (negatively buoyant) and continental (positively buoyant) portions of the same lithosphere led to slab breakoff after collision (Davies and von Blanckenburg, 1995; van Hunen and Allen, 2011). Hot asthenospheric mantle welled up through the slab window, inducing partial melting of the oceanic crust at the slab window edges and thereby forming the adakitic rocks. The upwelling caused a thermal anomaly and subsequent partial melting of the overlying mantle wedge. The upwelling asthenosphere, along with rebound of the positively buoyant, subducted continental crust, caused extension of the overlying, thickened, continental lithosphere. The ensuing rapid eruption of basaltic magmas resulted in limited fractional crystallization of mafic minerals, generating within-plate basalt characteristics. However, the heat provided by the basalts induced partial melting of the lower or middle crust, resulting in the development of voluminous silicic melts, including the bimodal volcanic rocks and coeval granitic intrusions.
In addition to the Central Qiangtang arc, another two east-west–oriented belts of Upper Triassic magmatic rocks have been identified in northern Tibet. The Kunlun and Hohxil-Yidun arcs are thought to have resulted from the subduction of Paleo-Tethys Ocean crust along the Anyimaqen-Kunlun and Jinshajiang suture zones, respectively (e.g., Roger et al., 2003; Reid et al., 2007; Weislogel, 2008; Wang et al., 2011). However, to date, there is no evidence that another two Late Triassic magmatic arcs were produced by the subduction of the Longmu Co–Shuanghu suture zone. In addition, the significant distance between these areas suggests that these three magmatic arcs belong to three different suture zones. As such, we consider the three suture zones in northern Tibet to be branches of the Paleo-Tethys, all of which host magmatic arcs produced by the subduction and closure of the Paleo-Tethys Ocean in the Late Triassic.
EXHUMATION OF THE METAMORPHIC ROCKS
Recent studies have indicated that closure of the Paleo-Tethys Ocean by northward subduction (Li et al., 1995; Li, 2008) led to the formation of subduction mélange and HP-UHP metamorphic rocks, which were exhumed during the Late Triassic (Kapp et al., 2003; Zhai et al., 2011a). Details of the mechanism that exhumed the metamorphic rocks back to the surface in Central Qiangtang are still unclear, but in accordance with Beaumont et al. (2009), Burov et al. (2014), and others, we suggest that this may have occurred in the channel between the subducting oceanic plate and the overlying North Qiangtang continental crust. This view is consistent with the composition of the mélange, which contains rocks with affinities to both continental crust and oceanic plate. Numerical simulations have suggested that several mechanisms contribute to HP-UHP exhumation, including buoyant uplift, cavity-driven flow, and expulsion as a strong continent “plunger” is inserted into the subduction channel (Warren et al., 2008a).
The presence of eclogite and metapelite within the Longmu Co–Shuanghu suture zone (Kapp et al., 2003; Zhai et al., 2011a) suggests that continental crust was subducted to mantle depths during the collision between the South and North Qiangtang terranes. Recently, Zhang et al. (2014b) identified continental materials in HP-UHP metamorphic rocks in Central Qiangtang and suggested that continental subduction was the precondition for slab breakoff and subsequent exhumation. As outlined already, we believe that the slab breakoff model can reasonably explain the extensional tectonic setting and the magmatic flare-up event. Numerical modeling results indicate that slab breakoff is not a prerequisite for exhumation (Warren et al., 2008a, 2008b), and the retreat of a subducting slab, triggered by buoyancy forces after slab breakoff, cannot be rejected as a mechanism for driving the metamorphic rocks upward from mantle depths to the surface (Jolivet and Faccenna, 2000). On the other hand, 40Ar/39Ar age data from HP-UHP metamorphic rocks, and zircon U-Pb data from magmatic rocks indicate that the timing of both peak exhumation of the metamorphic rocks and peak magmatism in Central Qiangtang was coeval in the Late Triassic (ca. 225–205 Ma; Fig. 6; Table 2). This geological phenomenon also appeared in the Alps, and research suggests that the short time interval between uplift of HP rocks and the onset of magmatism was triggered by the slab breakoff in the Tertiary (von Blanckenburg and Davies, 1995).
Hence, we suggest that both the magmatic flare-up and the exhumation of metamorphic rocks in the Late Triassic were triggered by slab breakoff after continental collision and subduction.
DEVELOPMENT OF THE GANGMARI METAMORPHIC BELT AND RIWANCHAKA YANGTZE-TYPE DEPOSITS
We propose a slab breakoff model to explain the magmatic flare-up event and the exhumation of metamorphic rocks of the Longmu Co–Shuanghu suture zone in the Late Triassic. The slab breakoff model has been supported by numerical modeling (Duretz et al., 2011; van Hunen and Allen, 2011) and teleseismic tomography (Lei and Zhao, 2007); however, direct geological evidence on the surface remains insufficient.
The east-west–trending Gangmari metamorphic belt of the South Qiangtang terrane lies ∼70 km south of the Longmu Co–Shuanghu suture zone of Central Qiangtang. The 40Ar/39Ar age data indicate that the timing of exhumation of the Gangmari metamorphic belt is consistent with that of the northern metamorphic belt (Zhai et al., 2011a). Moreover, Riwanchaka Yangtze-type deposits located adjacent to the Gangmari metamorphic belt have a Yangtze affinity and are similar to strata of the North Qiangtang terrane (Li et al., 2005; Li, 2008). These features suggest that the Gangmari metamorphic belt and Riwanchaka Yangtze-type deposits were derived from the north and were emplaced in the South Qiangtang terrane, probably in the Late Triassic (Y. Liu, 2014, personal commun.). Considering the contemporaneous exhumation of both the Longmu Co–Shuanghu suture zone and Gangmari metamorphic belt metamorphic rocks, we suggest that the Gangmari metamorphic belt was emplaced during exhumation in the Late Triassic. An emplacement model suggests that the Gangmari metamorphic belt and Riwanchaka Yangtze-type deposits were emplaced during continental collision under an extrusion setting (Y. Liu, 2014, personal commun.), but the model does not account for the coeval magmatic flare-up event, and exhumation of metamorphic rocks, which indicate that the Central Qiangtang was undergoing postcollisional extension.
We envisage that the initial continental collision was followed by deep subduction of continental crust, which led to slab breakoff because of the contrasting behavior of oceanic and continental portions of the same lithosphere. Following slab breakoff, positive buoyancy forces drove the subducted continental crust back up the subduction zone, vertically displacing the tip of the slab by ∼60 km and exhuming the HP-UHP metamorphic rocks from depth to the surface (Duretz et al., 2012); however, this interpretation requires further verification from geological observations. Based on this discussion, we propose the following process to explain the emplacement of the Gangmari metamorphic belt, and Riwanchaka Yangtze-type deposits. The Gangmari and Riwanchaka areas were located far from the trench before continental subduction (Fig. 7A). During continental subduction, the Gangmari and Riwanchaka areas reached the trench, where a metamorphic belt was formed in the Gangmari area, while Yangtze-type deposits were formed in the Riwanchaka area during growth of a thrust stack (ca. 245–226 Ma; Fig. 7B). Deep subduction of continental crust was followed by slab breakoff (ca. 225–205 Ma; Fig. 7C), which triggered the retreat of the oceanic slab, carrying the Gangmari and Riwanchaka areas away from the trench again and separating the Gangmari metamorphic belt and Riwanchaka Yangtze-type deposits from the Longmu Co–Shuanghu suture zone and the North Qiangtang terrane (Fig. 7D).
Slab breakoff is the most effective way to trigger the formation of HP and UHP metamorphic rocks as well as the subsequent roll back and exhumation that would be needed to expose these rocks at the surface (von Blanckenburg and Davies, 1995). Consequently, slab breakoff models have been proposed for many suture zones around the world. However, not all suture zones are associated with exhumed HP-UHP metamorphic rocks. Our research in the Central Qiangtang area suggests that the Gangmari metamorphic belt and Riwanchaka Yangtze-type deposits were derived from the Longmu Co–Shuanghu suture zone and the North Qiangtang terrane, indicating in turn that eduction of the deeply subducted South Qiangtang continental crust is another piece of critical evidence needed to support exhumation. In addition, the distance between allochthonous units and the subduction zone is equal to the minimum distance of slab eduction. Therefore, the minimum exhumation rate can be estimated by combining this distance with the timing of peak exhumation. For example, the distance between the Gangmari metamorphic belt–Riwanchaka Yangtze-type deposits and the Longmu Co–Shuanghu suture zone is ∼70 km, and peak exhumation occurred over a period of ∼20 m.y. (ca. 225–205 Ma), so this yields a minimum exhumation rate of up to ∼3.5 mm/yr for HP-UHP metamorphic rocks in the Central Qiangtang area.
We propose that Late Triassic geological events in the Central Qiangtang, including a magmatic flare-up event, exhumation of HP-UHP metamorphic rocks, and emplacement of the Gangmari metamorphic belt and Riwanchaka Yangtze-type deposits away from the Longmu Co–Shuanghu suture zone and North Qiangtang terrane, were all part of a chain reaction of events triggered by slab breakoff of the northward-subducting Paleo-Tethys Ocean lithosphere slab.
We thank Franz Neubauer and anonymous reviewers for constructive reviews and Kurt Stuewe for careful editorial handling. This research was funded by the National Science Foundation of China (grants 41072166, 41272240, and 41273047), China Geological Survey projects (grants 1212011221093, 1212011121248, and 12120113036700), and Project 2015018 of the Supported Graduate Innovation Fund of Jilin University.