In order to reveal the tectonic evolution of the South Chinese Tianshan orogenic belt, we conducted structural, geochemical, and geochronological studies and identified granitic and volcanic rocks along the northern margin of the Tarim block. Zircon laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) U-Pb dating of two samples from granitic plutons yielded crystallization ages from 404.8 ± 2.0 Ma to 388.1 ± 2.2 Ma, and two samples of dacite and rhyolite yielded ages of volcanism at ca. 404 Ma, highlighting a middle Paleozoic igneous event. Geochemical data suggest that these Devonian igneous rocks are metaluminous, high-K calc-alkaline felsic volcanic and plutonic rocks. All the samples display relatively enriched light rare earth element (LREE) contents, weak or no negative Eu anomalies, and relative depletion in Nb, Ta, Ti, and P, but they show enrichment in K, Rb, and Ba. In situ zircon Hf isotopic results show a positive range of εHf(t). These geochemical and isotopic features suggest that this magmatic event probably occurred in an Andean-type arc. Combined with zircon U-Pb and Hf isotopic studies on the basement rocks, the results indicate multiphase Precambrian (ca. 2600–2300 Ma, 1900–1800 Ma, 1140–830 Ma) continental growth and minor involvement of basement in Devonian arc magmatism. Finally, we integrated the structural, geochemical, and geochronological data into a geodynamic model of the South Chinese Tianshan that emphasizes south-directed subduction of the South Tianshan Ocean along the northern margin of the Tarim block during the early to middle Paleozoic.
As the southernmost part of the Central Asia orogenic belt or Altaid orogenic collage, the Paleozoic Tianshan orogenic belt is a typical example of an accretionary and collisional orogen that played a major role in the assembly of Eurasia (e.g., Burtman, 1975; Coleman, 1989; Windley et al., 1990; Allen et al., 1993; Şengör et al., 1993; Brookfield, 2000; Şengör and Natal’in, 1996; Jahn, 2004; Xiao et al., 2004; Kröner et al., 2007; Windley et al., 2007; Q.C. Wang et al., 2010; Fig. 1). Its tectonic evolution has recently been a subject of debate, receiving much attention from the international geological community (Q.C. Wang et al., 2010; Wang et al., 2011). The tectonic assembly of the Central Asia orogenic belt is suggested to have occurred during Paleozoic time by the closure of the paleo-Asian or paleo-Tianshan Oceans and resulted from long-lasting, multiphase subduction-accretion of microcontinents, island arcs, fragments of oceanic islands, ophiolites, and accretionary complexes between the Siberia craton to the north and the Tarim block to the south (Jahn, 2004; Xiao et al., 2004; Kröner et al., 2007; Windley et al., 2007). Large volumes of igneous rocks in the Phanerozoic imply a significant addition of juvenile continental crust from a depleted-mantle source (e.g., Jahn, 2004; Xiao et al., 2012). Compared to other segments of the Paleozoic Tianshan orogenic belt, the Chinese Tianshan is a key area that includes typical lithological, magmatic, structural, and metamorphic elements of the Central Asia orogenic belt (B. Wang et al., 2010, and references therein). Because a detailed geological survey is lacking, many issues remain controversial in the Chinese Tianshan, such as the location of the suture zones, the timing of the subduction, the distribution of episodic magmatic arcs, and the direction of subduction (Q.C. Wang et al., 2010). Even the existence of a South Tianshan Ocean is debated (Gao et al., 2011). If it did exist, was it a backarc basin (Wang et al., 2008; Lin et al., 2009), coexistent with the Central Tianshan Ocean (Ge et al., 2012), or a long-lived Paleozoic ocean (Xiao et al., 2012)? How and when was the South Tianshan Ocean closed? In the northern part of the Tarim block, was there a large-scale magmatic arc that may have been related to closure of the South Chinese Tianshan Ocean? Several lines of evidence suggest the existence of an Andean-type continental arc (Ge et al., 2012). In this contribution, we present new field observations, geometric analysis, zircon U-Pb dating, Lu-Hf isotopes, and whole-rock geochemical data from intrusive rocks in the northern margin of the Tarim block to the northwest of Korla in order to reappraise the tectonic evolution of the South Chinese Tianshan.
The Tianshan extends east-west for over 3000 km from NW China to Kazakhstan and Kyrgyzstan across Central Asia, and it exhibits some of the highest subaerial topography on Earth because of its late Cenozoic reactivation by intracontinental deformation (Hendrix et al., 1994; Avouac et al., 1993; Sobel and Dumitru, 1997; Alexeiev et al., 2009; Q.C. Wang et al., 2010; Biske and Seltmann, 2010). Located between the Tarim Basin to the south and the Junggar Basin to the north (Fig. 1), the Chinese Tianshan is traditionally subdivided into three parts from north to south: the Kazakh–Yili–North Tianshan block, the Central Tianshan microcontinent, and the Tarim block, which were amalgamated during the late Paleozoic (Wang et al., 2009, 2011; Lin et al., 2009; Xiao et al., 2004, 2012; Fig. 1).
The Tarim block, situated to the south of the Tianshan orogenic belt (Fig. 1), is the largest Precambrian craton in the Central Asia orogenic belt. The South Chinese Tianshan forms the highest part of the Tianshan Mountains due to Cenozoic uplift, but this impedes detailed geological survey (Liu et al., 1996; Z.X. Zhu et al., 2008). This Paleozoic orogenic belt is named the Hark belt (Wang et al., 1994) or the South Tianshan (Kokshaal-Kumishi) accretionary complex or mélange zone (Xiao et al., 2008, 2012). In the South Chinese Tianshan, several ophiolitic mélange zones occur discontinuously over more than 1000 km from west of Aksu to south of Kumux (localities “W,” “H,” “K,” “S,” “Y,” and “T” on Fig. 1, respectively). Previous work indicated that these ophiolitic mélanges are tectonically enclosed by sheared and weakly metamorphosed pre–Middle Devonian rocks and are unconformably overlain by nonmetamorphic and undeformed late Lower Carboniferous (Serpukhovian) to Permian strata (Wang et al., 2011, and references therein). Based on different models of tectonic evolution, various interpretations of these ophiolitic mélanges have been proposed (Fig. 1; Xiao et al., 2012, and references therein).
Two key issues related to the tectonics of the South Chinese Tianshan are still under debate. The first is the existence of the South Tianshan Ocean. Most researchers consider that the early Paleozoic passive continental margin of the northern Tarim block was bounded to the south by the Nalati fault, and the south Tianshan ophiolites were emplaced as nappes by south-directed thrusting of the high-pressure–ultrahigh-pressure (HP-UHP) belt of the SW Chinese Tianshan (Fig. 1; Windley et al., 1990; Allen et al., 1993; Hao and Liu, 1993; Wang et al., 1994; Zhou et al., 2001; Gao et al., 1998, 2011; Han et al., 2011). Based on stratigraphic and paleogeographic studies, Liu et al. (1996) argued for the existence of the South Tianshan Ocean. On the basis of tectonics and structural analysis, Ma et al. (1993), Charvet et al. (2007), Wang et al. (2008, 2011), and Lin et al. (2009) interpreted the South Tianshan Ocean as the backarc basin within the early Paleozoic south-facing subduction system.
If we consider the existence of the South Tianshan Ocean, it brings up a second issue, which concerns the existence of a magmatic arc during closure of the South Tianshan Ocean, for which there is no evidence in the western part of the Chinese Tianshan. Considering north-directed subduction of the South Tianshan oceanic plate coeval with a long-term accretionary process recognized in the South Tianshan (Kokshaal-Kumishi) area, the northern margin of the Tarim block has been interpreted as a long-lived passive margin during the entire Paleozoic (e.g., Carroll et al., 1995; Gao et al., 1998; Xiao et al., 2004, 2008, 2012). Based on structural analysis, Lin et al. (2009) considered that the closure of the South Tianshan backarc basin was accommodated by south-directed subduction of the Central Tianshan Ocean beneath the Tarim block, as suggested by the northward thrusting of the South Tianshan ophiolitic mélange over the Central Tianshan microcontinent (Fig. 2). Ge et al. (2012) reported a plutonic dike-type intrusion of Late Silurian age to the north of Korla, and therefore speculated that this was related to an Andean-type continental arc.
LITHOLOGY AND GEOMETRY OF THE KORLA-LIUSHUGOU SEGMENTS
To the north and northwest of Korla, Archean to Paleoproterozoic amphibolite and orthogneiss are thought to comprise a tonalite-trondhjemite-granodiorite (TTG) suite with U-Pb zircon ages of ca. 2.5–2.6 Ga (Fig. 3; Long et al., 2010, 2011; Shu et al., 2011). The deformed and weakly metamorphosed Neoproterozoic marine sequences, such as marble and limestone, are exposed to the east of Korla (Gao et al., 1993; Lu et al., 2008). Upper Cambrian to Middle Devonian strata are composed of nonmetamorphic, gently folded limestones, marls, and phosphatic rocks, and continental sandstones, which are considered as passive-margin sediments of the northern Tarim block. Upper Paleozoic to Mesozoic continental sedimentary rocks overlie the pre–Middle Devonian rocks unconformably (e.g., XJBGMR, 1993). In the Liushugou cross section, three lithotectonic units are described as follows (Fig. 3).
Lower Devonian Volcanic Rocks and Related Sediments
At Liushugou, a sequence of volcanic rocks thicker than 1.5 km is pervasively folded and fractured (Fig. 4A). These volcanic rocks are composed mostly of crimson to dark-green dacite, quartz porphyry, rhyolitic or dacitic porphyry, andesitic porphyry, andesitic lava, and rhyolitic tuff (Figs. 4B and 4C; XJBGMR-Halamaodun, 1972). On the geological map (XJBGMR, 1993), these igneous rocks are considered to be Middle Devonian, but our dating indicates an Early Devonian age (Fig. 3).
Further south, sedimentary rocks are bounded by a ductile strike-slip fault with a vertical foliation and subhorizontal lineation trending WNW-ESE (95°–120°). On the outcrop scale, the Lower Devonian volcanic rocks have been intensely sheared and mylonitized with a dextral sense of shear (Figs. 4D and 4E). The sedimentary rocks consist of laminated siliceous siltstone, mudstone, and limestone (Figs. 5A and 5B). Locally, mass-flow deposits or pebbly mudstones have been interpreted as turbidites (Wang et al., 2011). However, several thick-bedded lentoid limestone olistoliths with sizes from centimeters to kilometers occur in these rocks (Figs. 5C and 5D). These strata were previously considered to be Middle Devonian based on fossils of coral, such as Neospongophyllum sp. and Disphyllum cf. longiseptatum. However, based on recent paleontological work on blocks of radiolarian chert with Archocyrtium cf. procerum (Cheng), Archocyrtium venustum (Cheng), and Archocyrtium cf. ludicrum Deflandre, this turbiditic sedimentation was probably continuous until the Mississippian (Tournaisian or Visean; Gao et al., 1998; Liu, 2001; Zhu, 2007; Wang et al., 2011). In most cases, the matrix is strongly sheared or even mylonitized, so that a scaly fabric is developed in weakly metamorphosed pelite.
Although the Permian strike-slip faults have largely changed contact relationships, to the south of the Liushugou area, conglomerate, coarse sandstone, and bioclastic limestone overlie the volcanic rocks, indicating an Early Devonian unconformity (Fig. 5E). According to our field investigation, several fossils typical of Pennsylvanian age have been observed in the bioclastic limestone, such as Cancrinella sp. and Dictyoclostus sp., confirming the Pennsylvanian age of deposition. It is worth noting that these strata were first mapped as a Pennsylvanian unit (XJBGMR-Halamaodun, 1972); however, in adjacent areas, the age of equivalent strata has been confirmed as Mississippian by recent geological mapping (Li and Xu, 2007). This nonmetamorphosed unit was deformed by north-verging open folds during the late Paleozoic (Fig. 5F).
Paleozoic granitoids occupy one third of the research area (Fig. 3). On the basis of petrological and geochronological studies, three groups of plutonic rocks can be distinguished. North of Korla, the Hulashan pluton, the largest one in this area (Fig. 3), is composed of medium- to fine-grained gray-colored monzogranite, two-mica granite, and granodiorite. Plagioclase, K-feldspar, biotite, and muscovite are the dominant phases. On the basis of geological mapping, this pluton is considered to be “late Variscan” in age (Carboniferous to Permian; XJBGMR-Halamaodun, 1972). The gneissic foliation is well developed along the northern margin of the pluton. In the northeastern and southern parts of the Hulashan, medium-grained alkali feldspar granite and granodiorite are well exposed (Fig. 3). Previous ICP-MS zircon U/Pb dating indicates similar ages for the granodioritic intrusion north of Korla city, clustering at ca. 420 Ma (Ge et al., 2012). To the north of the research area, undeformed biotite granite and two mica granite yielded U/Pb zircon sensitive high-resolution ion microprobe (SHRIMP) ages around 300 Ma, and these have been interpreted as postorogenic granites (Fig. 3; Z.X. Zhu et al., 2008).
ANALYTICAL METHODS AND SAMPLE DESCRIPTIONS
Zircon LA-ICP-MS U-Pb Dating
Zircons were separated from samples using standard density and magnetic separation techniques. Zircon grains, together with standard 91500 and Temora zircons, were mounted in epoxy mounts that were then polished to section the crystals in half for analysis. All zircons were photographed in transmitted and reflected light, and cathodoluminescence (CL) images were obtained in order to reveal their internal structures. The mount was vacuum-coated with high-purity gold for LA-ICP-MS analyses.
Samples were analyzed by the LA-ICP-MS at China University of Geosciences, Beijing. The instrument couples an ICP-MS (Agilent 7500a) and a UP-193 Solid-State laser (193 nm, New Wave Research, Inc.) with the automatic positioning system. Detailed analytical procedures are to be found in Song et al. (2010). Laser spot size was set to 25 μm in diameter. Zircon standard 91500 was used as an external standard for correction of U-Pb isotope fractionation. Zircon standard Temora was used as a monitoring standard to correct the deviation of age measurement and calculation. The common Pb correction was made following the method of Andersen (2002).
Major and Trace Elements
Four fresh rock samples were selected, crushed, and powdered in an agate mill. Elemental analyses were conducted at the Institute of Geology and Geophysics, Chinese Academy of Sciences (IGGCAS, Beijing). Major-element analyses were performed by X-ray fluorescence (XRF), with analytical uncertainties ranging from 1 to 3 wt%. Trace-element contents were determined using an Agilent 7500a inductively coupled plasma–mass spectrometer (ICP-MS). The detailed analytical procedure is the same as used by Zhang et al. (2012). An internal standard was used for monitoring drift in mass response during mass spectrometric measurements.
Zircon Lu-Hf Isotopes
Zircon Lu-Hf isotopic analysis was carried out on a Neptune multicollector ICP-MS equipped with a Geolas-193 laser ablation system at the Institute of Geology and Geophysics, Chinese Academy of Sciences. Detailed analytical procedures are given in Wu et al. (2006). Previously analyzed zircon grains for U-Pb isotopes were chosen for Lu-Hf isotopic analyses. The beam diameter was either 44 or 60 μm, with a laser repetition rate of 10 Hz at 100 mJ. During the analytical period, the weighted mean 176Hf/177Hf ratios of the zircon standards GJ-1 (mean value = 0.2820146) and MUD (mean value = 0.2825096) were in good agreement with reported values (Woodhead and Hergt, 2005; Morel et al., 2008).
In order to compare tectonic features and reveal the evolution of South Chinese Tianshan, we selected representative granitoids and basement rocks from the Tarim block to carry out zircon U-Pb, Hf isotopic, and geochemical analyses.
XJ045-2 and XJ050 were collected from the Liushugou area north of the Hulashan (Fig. 3). XJ045-2 was a mylonitic volcanic rock with a vertical foliation and subhorizontal lineation (Fig. 4D). It was composed mainly of quartz, plagioclase, biotite, and secondary muscovite. In the photomicrographs, deformed porphyroclasts indicate a dextral shear sense (Fig. 4E). XJ050 is a dacite from the bank of Liushugou (Fig. 4A). In thin section, this sample consists of plagioclase phenocrysts, quartz, K-feldspar, epidote, amphibole, and tuffaceous matrix (Fig. 4C).
Two coarse-grained granitoids were selected for dating to determine the emplacement age of the pluton. XJ046-1 is an alkali feldspar granite, composed of quartz (40%), alkali feldspar (45%), plagioclase (5%), and dark minerals (10%). XJ048-3 is a granite, composed of quartz (50%), alkali feldspar (25%), plagioclase (15%), and dark minerals (10%).
In order to make comparisons with the Paleozoic magmatic rocks, four samples were selected from the Kuluketage, north Korla area, which was considered to be the basement of the Tarim block (Fig. 3; XJBGMR, 1993). All four samples were collected along the National Road G208. Two samples, XJ030-1 and XJ031-1, were obtained from the northern margin of the Kuluketage massif (Fig. 3). XJ030-1 is a well-foliated vein of felsic rock in amphibolites (Fig. 6A). In thin section, XJ030-1 consists of quartz (∼30%), plagioclase (∼45%), alkali plagioclase (∼20%), and ∼5% dark minerals (Fig. 6B). XJ031-1 is a foliated marble interlayered with paragneiss (Fig. 6C). Pyroxene can be observed in the thin section within calcite, which is the dominant phase (Fig. 6D). TS126A and TS126B were collected from a foliated migmatite on the southern margin of the Kuluketage massif, adjacent to the north of the Korla city. TS126A is the leucosome and TS126B is the melanosome part of the migmatite (Fig. 6E). TS126A consists of quartz (25%), alkali feldspar (45%), hornblende (25%) and muscovite and epidote (<5%), while TS126B consists of quartz (25%), alkali feldspar (5%), and hornblende (65%) (Fig. 6F).
In Figure 7, zircon CL images from four magmatic rocks were used to reveal internal structures, and zircon U-Pb ages are plotted in Figure 8. Meanwhile, four samples were collected from the basement rocks of Tarim block, the internal structures of analyzed zircons are shown in Figure 9 and the analytical data are plotted in Figure 10. U-Pb, geochemical, and Hf data for zircons are presented in Tables 1, 2, and 3, respectively.
U-Pb Zircon Geochronology
Volcanic Rocks from the Liushugou Area
Zircons from XJ045-2 and XJ050 have similar morphologies and are prismatic, transparent, and colorless (Fig. 7). They are mostly euhedral, with lengths ranging from 70 to 200 μm, and length to width ratios from 1:1 to 2:1. According to their internal structures revealed by CL imaging, all the zircons show few inherited cores, with clear oscillatory zoning (Fig. 7).
XJ045-2. Guided by the CL images, 23 analyses were performed on oscillatory zoned zircons, and 21 concordant ages were obtained, with discordance mostly within ±3%; two ages were omitted because of discordance (Table 1). Among these 21 analyses, the U concentrations range from 135 to 533 ppm, Th from 126 to 712 ppm, and Th/U ratios from 0.79 to 2.21 (Table 1). The 21 concordant 206Pb/238U ages cluster at 405–402 Ma, with a weighted mean age of 403.5 ± 2.1 Ma (n = 21, mean square of weighted deviates [MSWD] = 0.022), representing the age of the rhyolite eruption (Fig. 8).
XJ050. In this sample, 25 analyses were obtained from oscillatory zoned parts of 25 zircon grains, and they have discordances mostly within ±2%. U concentrations range from 59 to 280 ppm, Th from 42 to 220 ppm, and Th/U ratios from 0.57 to 1.04 (Table 1). Twenty-four analyses have indistinguishable 206Pb/238U ratios within analytical error, and a weighted mean age of 403.9 ± 2.0 Ma is calculated (n = 24, MSWD = 0.076) as the best estimate of the eruption age of this volcanic rock (Fig. 8). Analysis of XJ050-8 could be interpreted as an inherited age of 740 Ma (Table 1).
Granitoids from the Liushugou Area
In the research area, two samples of granitoids were selected for geochronological and geochemical work (Fig. 3). Zircons from these two samples are mostly euhedral, with the lengths ranging from 50 to 150 μm, and length to width ratios from 1:1 to 2.5:1 (Fig. 7). Oscillatory zoning is also the dominant feature in the internal structure of most zircons, and inherited cores are rare (Fig. 7).
XJ046-1. Twenty-four analyses were performed on these oscillatory zoned zircons, and 21 spots (except XJ046-1 9, 14, 18) with low discordance were used to calculate a crystallization age (Table 1). U and Th concentrations have a large range, from 60 to 1130 ppm, and from 55 to 1161 ppm, respectively, with high Th/U ratios (0.66–1.67; Table 1). However, all give concordant 206Pb/238U ages, with a weighted mean age of 388.1 ± 2.2 Ma (n = 21, MSWD = 0.070). This age is interpreted as the crystallization age of the alkali feldspar granite (Fig. 8).
XJ048-3. Based on the CL images, 25 analyses were performed on magmatic zircons. Except for one discordant point, 24 analyses have concordant ages with discordance within ±2%. They show relatively low U contents, from 123 to 460 ppm, Th from 125 to 1381 ppm, and Th/U ratios between 0.70 and 3.00 (Table 1). The 206Pb/238U ages of 24 analyses give a concordant cluster with a weighted mean age of 404.8 ± 2.0 Ma (n = 24, MSWD = 0.036), representing the crystallization age of this granite (Fig. 8).
Basement Rocks of the Tarim Block
Zircons from the four samples (XJ030-1, XJ031-1, TS126A, and TS126B) show similar morphologies. They are prismatic, transparent, and colorless, with lengths ranging from 50 to 200 μm and width to length ratios mostly ranging from 1:1 to 1:3. According to their internal structures revealed by CL imaging (Fig. 9), these zircons can be subdivided into two types. Type I zircons show core-mantle-rim structures. The inherited cores are dark, prismatic, or irregular, with clear oscillatory zoning; the enclosed mantles, usually less than 30 μm, are very bright without oscillatory zoning; and the rims, usually less than 10 μm, are dark (Fig. 9). On the contrary, the structure of type II zircons is relatively bright cores with or without narrow dark rims (Fig. 9).
Guided by the CL images, 133 zircon grains without evident inclusions and fractures from four samples were analyzed using LA-ICP-MS. For sample XJ030-1, 35 analyses were performed on core and mantle (Fig. 9). The U concentrations range from 65 to 460 ppm, Th from 34 to 307 ppm, and Th/U ratios from 0.12 to 0.80 (Table 1). A subordinate 207Pb/206Pb age group at ca. 2300–2500 Ma, with an upper-intercept age at 2494 ± 86 Ma (Fig. 10A), from the dark core of zircons of type I is interpreted as the crystallization age of the felsic vein, coeval with the recently documented Neoarchean to Paleoproterozoic gneissic TTG suite (Long et al., 2010, 2011; Shu et al., 2011). A major age group with a weighted mean age of 1810 ± 21 Ma (n = 11, MSWD = 2.5) was obtained from the bright mantles (Fig. 10A), interpreted to represent the age of a late metamorphism associated with the formation of the foliation (Fig. 6A). For sample XJ031-1, 30 analyses were performed on both core and mantle of zircons from the marble (Fig. 9). The U and Th concentrations are relatively low, 5–998 ppm and 0–88 ppm, respectively. Th/U ratios vary from 0.03 to 0.11, with a majority around 0.10. Based on the CL images, metamorphic alternation is preferred as the cause of the zircon growth (Table 1). Unlike sample XJ030-1, old ages (>2.0 Ga) are not found. Zircon analyses from this sample define a well-constrained upper-intercept age at 1748 ± 52 Ma (Fig. 10B), consistent with the age of a Paleoproterozoic event (such as migmatization and regional deformation) in the northern Tarim block (Shu et al., 2011; Zhang et al., 2012). It is worth noting that three analyses of regular bright grains of type II zircons, which possibly correspond to the mantle part of the type I zircons, give 206Pb/238U ages between 800 and 950 Ma. Although these ages are consistent with the age of a Neoproterozoic event in the northern Tarim block (W.B. Zhu et al., 2008; Shu et al., 2011; Zhang et al., 2011, 2012), without any metamorphic record, the origin of these zircons is still unclear.
TS126A and TS126B were collected from a well-foliated migmatite along the highway north of Korla city (Fig. 6E). In order to reveal the age of the migmatization, 34 analyses were performed on the dark core of the zircons from TS126A (Fig. 9), the leucosome of the migmatite. The U concentrations range from 252 to 1114 ppm, and Th ranges from 10 to 1066 ppm. Th/U ratios vary from 0.01 to 4.59, with a majority around 0.10, probably indicating a metamorphic origin (Table 1). A well-constrained upper-intercept age at 1775 ± 22 Ma is defined (Fig. 10C).
For comparison with TS126A, sample TS126B was collected from the melanosome of the migmatite (Fig. 6E). Thirty-four analyses were obtained from the dark cores of zircons from TS126B (Fig. 9). U contents vary from 722 to 2552 ppm, Th from 96 to 9340 ppm, and Th/U ratios from 0.06 to 4.28 (Table 1). Similar to the leucosome, an upper-intercept age at 1779 ± 42 Ma is calculated (Fig. 10D). It is therefore concluded that zircons from both the leucosome and melanosome of the migmatite crystallized around 1800 Ma, indicating the age of migmatization.
Whole-Rock Geochemical Characteristics
Four granitic and volcanic rocks from the study area were analyzed for whole-rock geochemistry. Results are listed in Table 2. The two granitic rocks show higher concentrations of SiO2 and lower MgO and total Fe as Fe2O3 than those of volcanic rock samples. All igneous rock samples are rich in alkalis (7–8.5 wt%), with relatively high contents of K2O compared to Na2O (Table 2). On the diagram of SiO2–Na2O + K2O, all samples plot in the granodiorite/dacite or granite/rhyolite fields (Fig. 11A). Based on the correlation between SiO2 and Na2O + K2O–CaO, all rocks belong to calc-alkalic or alkali-calcic series (Fig. 11B). On the SiO2-K2O diagram, the four samples all plot in the high-K calc-alkaline field (Fig. 11C). Furthermore, their A/NK (1.15–1.51) and A/CNK (0.99–1.05) ratios indicate that they are metaluminous I-type granitoids (Fig. 11D; Chappell and White, 2001). These results are comparable with previous work (Ge et al., 2012; Fig. 11).
The samples have total rare earth element (REE) contents of 101–169 ppm and LaN/YbN of 8.11–14.03, except sample XJ048-3 (70.2) (Table 2). On the chondrite-normalized diagram, all samples display moderate light (L) REE–enriched patterns with weak or no negative Eu anomaly (Eu/Eu* = 0.6–0.8), except the sample XJ048-3 (Eu/Eu* = 1.17) (Fig. 12A). In the primitive mantle–normalized spidergrams (Fig. 12B), both volcanic and magmatic rock samples exhibit trace-element patterns with variable depletion in Nb, Ta, Ti, and P, but enrichment in K, Rb, and Ba. These features, consistent with previous studies by Ge et al. (2012), are similar to arc-related volcanic and magmatic rocks (Fig. 12).
Zircon In Situ Hf Isotopes
In total, 160 dated zircon grains from seven samples were selected for in situ Hf isotope analysis. LA-MC-ICP-MS zircon Hf-isotope measurements were performed after the U-Pb isotope analyses. Overall, the measured εHf(t) values range from −11.0 to +13.2 (Table 3).
Volcanic Rocks from the Liushugou Area
Eighteen single zircon grains from XJ045-2 and 19 from XJ050 were analyzed for Hf isotopes. The results show variable Hf isotopic compositions, with 176Hf/177Hf ratios ranging from 0.282668 to 0.282763 and 0.282469 to 0.282633, respectively. The calculated εHf(t) values of XJ045-2 range from +4.9 to +8.0, and their Hf depleted mantle model age (TDMC) ages are between 889 Ma and 1088 Ma. In contrast, εHf(t) values of XJ050 are relatively low, from −2.2 to +2.9, and their TDMC ages are older, from 1211 to 1535 Ma.
Granitoids from the Liushugou Area
Nineteen analyses of 19 zircons from XJ046-1 show variable Hf isotopic compositions, with 176Hf/177Hf ratios ranging from 0.282663 to 0.282942. The calculated εHf(t) values range from +4.1 to +13.2, and their TDMC ages are between 543 Ma and 1124 Ma, with a major cluster ca. 1000 Ma. In sample XJ048-3, 20 analyses of 20 zircons show variable Hf isotopic compositions, with 176Hf/177Hf ratios ranging from 0.282678 to 0.282824. The calculated εHf(t) values (+5.2 to +9.2) are analogous to those of XJ046-1. Similar TDMC ages were obtained between 819 Ma and 1068 Ma (Fig. 13).
Basement Rocks of the Tarim Block
In sample XJ031-1, 14 analyses of 14 zircons show variable Hf isotopic compositions, with 176Hf/177Hf ratios ranging from 0.280895 to 0.281426. The calculated εHf(t) values range from -11.0 to +7.4, and their TDMC ages are between 1970 Ma and 3134 Ma.
From the leucosome and melanosome of the migmatite, samples TS126A and TS126B, respectively, 31 and 34 zircons were chosen for Hf analyses. In TS126A, Hf isotopic compositions showed 176Hf/177Hf from 0.281366 to 0.281608. The calculated εHf(t) values range from −10.3 to −2.6, and their TDMC ages are between 2714 Ma and 3085 Ma. In TS126B, 176Hf/177Hf ratios vary from 0.281491 to 0.281683, and the calculated εHf(t) values have a relatively narrow range from −5.9 to −2.7 (except TS126B 16). The TDMC ages vary from 2454 Ma to 2841 Ma, i.e., a little younger than that of TS126A.
Basement Rocks of the Kuluketage Massif, Northern Tarim Block, and the Precambrian Tectonic Evolution
As the largest Precambrian craton in the Central Asia orogenic belt, the Tarim block is a key to understanding the tectonic framework of the Central Asian orogenic belt (Kröner et al., 2007; Windley et al., 2007; Xiao et al., 2008, 2012). In the northern Tarim block, where the basement is well exposed in the Kuluketage massif (Fig. 1), four tectono-thermal events have been previously documented: (1) Neoarchean high-grade TTG gneiss formation at ca. 2600–2300 Ma (Long et al., 2010; Shu et al., 2011; Zhang et al., 2012, and references therein); (2) Paleoproterozoic magmatism and metamorphism around 1900–1800 Ma (Lu et al., 2008; Shu et al., 2011; Zhang et al., 2012); (3) early Neoproterozoic magmatism, which is considered to be related to the assembly of the Tarim block (1140–830 Ma; Shu et al., 2011; Ge et al., 2012); and (4) rift-related magmatism around 800–600 Ma (W.B. Zhu et al., 2008; Shu et al., 2011).
Among our analyses, a subordinate 207Pb/206Pb age group at ca. 2500–2300 Ma with an upper-intercept age at 2494 ± 86 Ma is identified from the dark core of zircons in XJ030-1 (Fig. 10A), indicating a ca. 2.5 Ga event. Like the Nd model ages (TDM) ranging from 3.2 to 2.2 Ga of Hu et al. (2000), the Hf crustal model ages (TDMC) of all our analyzed zircons show a similar range between 3.2 Ga and 2.6 Ga (Fig. 13B). These data indicate that the Archean basement was the dominant source during the ca. 1800 Ma tectono-thermal event, consistent with magmatic zircon structures (Fig. 9). A Neoarchean to Paleoproterozoic gneissic TTG suite has been discovered recently and has been interpreted as the product of a juvenile crustal growth event (Long et al., 2010, 2011). All the information has been revealed by radiometric dating of inherited cores of zircons, and it is still difficult to identify this event from structural and petrological studies.
Zircons from all of our samples from the basement record a Paleoproterozoic tectono-thermal event at ca. 1800 Ma (Fig. 10). Based on zircon morphology and Th/U ratios (the majority around 0.1), the zircon growth can be attributed to a metamorphic event. TS126A and TS126B show that the age around 1800 Ma is related to migmatization. In the Kuluketage massif, granulite-amphibolite–facies metamorphism and contemporaneous magmatism have been documented (Lu et al., 2008; Zhang et al., 2011, 2012). Some authors have interpreted this event as a response to addition of the Tarim block to the Columbia supercontinent (Rogers and Santosh, 2002; Zhao et al., 2002, 2004; Shu et al., 2011; Zhang et al., 2012). However, because of limited work, the cause of this tectono-thermal event remains unclear.
As mentioned already, three analyses of zircons from sample XJ031-1 reveal 206Pb/238U ages between 800 and 950 Ma (Fig. 10B). On the basis of zircon textures and Th/U ratios, these three grains suggest an origin through metamorphic growth from aqueous fluids (Long et al., 2010). In the Kuluketage massif, the early Neoproterozoic magmatism was significant (1140–860 Ma, with the peak around 920 Ma; Shu et al., 2011; Ge et al., 2012). Even if these ages correspond well to the age of the Aksu blueschists, because of a lack of detail geological constraints, it is still difficult to explain the tectonic evolution with the geochronological data only. The younger rift-related magmatism (830–800 Ma) is well documented in the same area (W.B. Zhu et al., 2008; Long et al., 2011; Zhang et al., 2011, 2012, and references therein). Our zircon Hf crustal model ages (TDMC) from the middle Paleozoic arc-related rocks range mainly from 0.5 to 1.6 Ga (a major peak at ca. 1.0 Ga and a subordinate one at ca. 1.4 Ga), also implying an important parental magma source at 1.1–0.8 Ga (Fig. 13B). It is worth mentioning that the Neoproterozoic blueschist of ca. 800 Ma age, recognized south of Aksu, is probably related to a collisional orogeny (Nakajima et al., 1990; Liou et al., 1996; Fig. 1). Thus, integrated with other studies, a large volume of crustal growth seems more likely during the late Mesoproterozoic to early Neoproterozoic (Shu et al., 2011; Ge et al., 2012).
In summary, the new ages of widespread basement rocks and episodic crustal growth revealed by zircon geochronology enable us to decipher the history of the crustal evolution of the Tarim block.
Petrogenesis of the Magmatic Arc along the Northern Margin of the Tarim Block
Our work on the ca. 400 Ma granitic and volcanic rocks on the northern margin of the Tarim block highlights a middle Paleozoic igneous event. The northern Tarim margin has been interpreted as a long-lived passive margin during the entire Paleozoic (e.g., Carroll et al., 1995; Gao et al., 1998; Xiao et al., 2004, 2008, 2012). Recently, more and more evidence of calc-alkaline magmatism suggests the existence of a volcanic arc, which argues for an active margin along the northern Tarim block, at least during the Silurian to Early Devonian (Zhu, 2007; Z.X. Zhu et al., 2008; Ge et al., 2012). Thus, the tectonic setting of the northern Tarim block margin becomes particularly important in understanding the tectonic evolution of the south part of the Central Asia orogenic belt. A middle Paleozoic magmatic event was identified previously as Carboniferous to Permian (XJBGMR-Halamaodun, 1972). According to our study in the Liushugou area of the northern Hulashan massif, two samples (XJ46-1 and XJ48-3) from granitic plutons (granite and alkali feldspar granite) yield crystallization ages from 404.8 ± 2.0 Ma to 388.1 ± 2.2 Ma, and two samples from the dacite (XJ050) and rhyolite (XJ045–2) yield ages of ca. 404 Ma. Although these ages are a little younger than those of Ge et al. (2012), these results highlight Early Devonian magmatic activity on the northern margin of the Tarim block.
In our study, two volcanic rocks (rhyolite and dacite) and two granites were analyzed for major and trace elements and Lu-Hf isotopes. These samples are all metaluminous, with A/CNK = 0.99–1.05, and high in silica (SiO2 > 66%), but low in MgO (0.21%–1.00%), with Al2O3 contents of 11.99%–14.68%. Negative Nb-Ta, P, and Ti anomalies and weak enrichment of K, Rb, and Ba relative to neighboring elements are shown in the incompatible trace-element spidergram. As in a previous study (Ge et al., 2012), these geochemical features suggest that these rocks probably were formed in an Andean-type continental arc (Figs. 11 and 12). A late early Paleozoic magmatic arc north of Korla city has been recently reported (Ge et al., 2012). Forty percent of the zircons that these authors selected yielded inherited ages, and εHf(t) values mostly ranged from −14.2 to 0, implying mixed sources from old crust and juvenile material. However, in our samples, only one inherited zircon (740 Ma) was found (Table 1; Fig. 7). Furthermore, Hf isotopes of the middle Paleozoic igneous rocks (volcanic rocks and granites) display relatively positive εHf(t) values, with only a small fraction of negative εHf(t) zircons. Thus, mantle-derived magma or remelting of juvenile crust mixed with subordinate supracrustal material is preferred here as the origin of these igneous rocks (Huang et al., 2012).
Zircon U-Pb geochronological, geochemical, and Lu-Hf isotopic data support the interpretation of a high-K calc-alkaline Andean-type continental arc along the northern margin of the Tarim block. Recently, north of the Tarim Basin, detrital zircon dating in Devonian sandstone from drilled samples identified a Silurian age population. Most zircon ages cluster around 460–414 Ma, especially around 436–423 Ma (Liu et al., 2012). Based on heavy mineral analysis, the Silurian clastic sediments, which correspond to a fluvial/delta system, were mainly provided by the partially uplifted areas north of the Tarim block (Liu et al., 2012; Lin et al., 2012).
This evidence indicates that the northern margin of the Tarim block was an active margin at least from 440 Ma to 380 Ma (Z.X. Zhu et al., 2008; Ge et al., 2012; Fig. 14), necessitating a southward subduction of the South Tianshan Ocean.
Tectonic Framework of South Chinese Tianshan and the Polarity of Subduction in the South Tianshan Ocean
The South Chinese Tianshan is separated from the Tarim block to the south by the Xingdi fault (Figs. 1 and 14). During the Neoproterozoic to early Paleozoic, the northern Tarim block is considered to have been a passive margin with continuous sedimentation of Sinian limestones, Cambrian–Ordovician limestones, marls, and phosphatic rocks, and Silurian clastic rocks. In the North Korla–Liushugou area, a sedimentary hiatus resulted from an early Paleozoic orogeny. Arc-related magmatic rocks intruded into the Precambrian basement with coeval volcanic rocks deposited on these basement rocks (Fig. 3). Similar magmatic rocks have been reported elsewhere, such as an arc-type granitoid north of Korla (Fig. 3; Ge et al., 2012), and an arc-related dacite to the north of the Xingdi fault, east of Kuluketage (Zhang and Sun, 2010). The mélange unit in the South Tianshan suture zone, and the turbiditic sequences situated north of the arc-related igneous sequences indicate that the northern margin of the Tarim block had changed to an active margin during the Late Silurian to Early Devonian (Figs. 3 and 14).
The ophiolite fragments of Wuxiapaerkan, Heiyingshan, Kulehu, Serikey, Yushugou, and Tonghuashan along the South Chinese Tianshan indicate the existence of the South Tianshan Ocean (Fig. 1; Shu et al., 2002; Charvet et al., 2007, 2011; Lin et al., 2009; Xiao et al., 2008, 2012; Wang et al., 2009, 2011, and references therein). This early Paleozoic ocean is generally considered as a backarc basin (Shu et al., 2002; Wang et al., 2008, 2011; Lin et al., 2009; Charvet et al., 2007, 2011) that was formed due to the convergence between the Central Tianshan Ocean and the Central Tianshan microcontinent or an independent ocean (Ge et al., 2012). No matter which model is adopted, the closure of the South Tianshan Ocean, from the view of our work, was accommodated by south-directed subduction beneath the Tarim block, contemporaneous with arc-related igneous rocks. A similar tectonic interpretation is indicated by olistoliths of Tarim-type Sinian and Cambrian carbonates within Silurian sedimentary rocks, and the northward thrusting of the South Tianshan ophiolitic mélange over the Central Tianshan microcontinent (Wang et al., 2011). Our work suggests that the northern margin of the Tarim block changed into an active margin with arc magmatism and volcanism during the Late Silurian to Early Devonian. Z.X. Zhu et al. (2008) reported a diorite north of Kuqa with a SHRIMP U-Pb age of 387 ± 8 Ma (Fig. 14). Farther east, a whole-rock 39Ar/40Ar age of 375 ± 6 Ma was obtained from a dacite on the north side of the Xingdi fault (Zhang and Sun, 2010; Fig. 14). These results suggest that arc magmatism in the northern Tarim block lasted until the Late Devonian. According to previous work (Wang et al., 2011), muscovite 40Ar/39Ar dating on mylonitic pelites from the mélange indicates that the backarc basin began to close in the Middle–Late Devonian and was consumed in the Mississippian (368–356 Ma). The sedimentary rocks of the north Korla–Liushugou area, where shallow-marine bioclastic limestones, coarse sandstones, and conglomerates cover the folded volcanic rocks unconformably, record the collision (Fig. 5E). A similar conclusion has also been supported by postcollisional magmatism around 300 Ma (Konopelko et al., 2007; Z.X. Zhu et al., 2008; Ren et al., 2011; Wang et al., 2011). The early Paleozoic South Chinese Tianshan orogenic belt was later strongly overprinted by Permian transcurrent tectonic and thereafter by Cenozoic intracontinental shortening (Lin et al., 2009).
Geodynamic Evolution of the South Chinese Tianshan
Field work and laboratory studies in the north of Korla to Liushugou area allow us to characterize the geometry and tectonic framework of the South Chinese Tianshan. Our new geochronological, geochemical, and structural results are integrated in the tentative geodynamic model proposed by previous workers (Charvet et al., 2007; Wang et al., 2009, 2011; Lin et al., 2009; Fig. 15).
In the early Paleozoic, the Central Tianshan Ocean (“South Tianshan Ocean” or “South Tianshan Palaeo-Ocean” of previous works; cf. Windley et al., 1990; Gao et al., 1998; Laurent-Charvet et al., 2002; Charvet et al., 2007) separated the Yili–north Tianshan block from the Tarim block. The southward subduction of oceanic lithosphere began, perhaps as early as Late Ordovician, forming a magmatic arc installed upon the Central Tianshan continental crust that, at that time, formed the northernmost part of the Tarim continent (Fig. 15A; Hopson et al., 1989; XJBGMR, 1993; Laurent-Charvet et al., 2002; Ma et al., 2006; Xu et al., 2006; Yang et al., 2006). During the early part of the Silurian, a backarc basin opened and split the Central Tianshan microcontinent from the main part of the Tarim block. Relics of this oceanic basin are represented by the South Chinese Tianshan ophiolitic mélange (Fig. 15B; Xiao et al., 2008, 2012; Wang et al., 2011; Fig. 14). An oceanic setting has also been indicated by the normal mid-ocean-ridge basalt (N-MORB) affinity of gabbros and mafic lavas in the ophiolitic mélange (Ma et al., 1993; Long et al., 2006; Charvet et al., 2007; Wang et al., 2009, 2011). The age of the South Tianshan Ocean is constrained by gabbro zircon SHRIMP ages as 425 ± 8 Ma and 392 ± 5 Ma from the Kulehu and Heiyingshan ophiolites, respectively (Long et al., 2006; Wang et al., 2011). Farther east, granulite-facies meta-mafic rocks in the Yushugou ophiolite were dated at 392–378 Ma (Fig. 14; Jiang et al., 2001; Zhou et al., 2004). The Upper Devonian to Lower Carboniferous radiolarian-bearing siliceous muddy matrix of the South Chinese Tianshan ophiolitic mélange can be correlated with Middle Devonian–Lower Carboniferous radiolarian cherts (Gao et al., 1998; Liu, 2001; Zhu, 2007; Wang et al., 2011). From Late Silurian to the Mississippian, the closure of the South Chinese Tianshan Ocean was accommodated by south-directed subduction beneath the Tarim block, as suggested by calc-alkaline magmatic rocks along the northern margin of the Tarim block, especially in the Korla-Liushugou area (Fig. 15C). The northward thrusting of the South Chinese Tianshan ophiolitic mélange over the Central Tianshan microcontinent reveals the polarity of plate subduction (Wang et al., 2009, 2011; Lin et al., 2009). The regional unconformity between Upper Carboniferous rocks and pre–Lower Carboniferous tectono-stratigraphic units in South Chinese Tianshan indicates the age of South Tianshan Ocean closure (Chen et al., 1999; Zhou et al., 2001). Recently, mylonitic pelites from the mélange, which have top-to-the-north sense of shear, yielded muscovite 40Ar/39Ar ages of 368–356 Ma (Wang et al., 2011). This indicates that the backarc basin began to close in the Middle–Late Devonian and finally disappeared in the Mississippian. The peak age of metamorphism of (U)HP eclogite along the northern margin of the Central Tianshan microcontinent and the subsequent exhumation of these metamorphic rocks indicate the continuation of the southward subduction of the Central Tianshan Ocean beneath the Central Tianshan microcontinent in the late part of the Mississippian (around 320 Ma; Su et al., 2010; Li et al., 2011; Lin et al., 2009; Wang et al., 2009; de Jong et al., 2009; Fig. 15D). The vast amount of terrigenous deposits and postorogenic magmatism recorded the final closure of the Central Tianshan Ocean during the Pennsylvanian (Fig. 15E; Konopelko et al., 2007; Z.X. Zhu et al., 2008; Dong et al., 2011; Han et al., 2011).
In summary, the following conclusions can be drawn from this study.
(1) At the northern margin of the Tarim block, zircon LA-ICP-MS dating indicates that volcanic rocks and granitic rocks were emplaced in the middle Paleozoic (404.8 ± 2.0 Ma to 388.1 ± 2.2 Ma) in the Liushugou-Korla area. Geochemical data indicate that these Devonian magmatic rocks are metaluminous, high-K calc-alkaline felsic volcanic and plutonic rocks. Geochemical and isotopic features suggest that magmatism probably occurred in an Andean-type arc.
(2) New age data and Hf isotope signatures from zircons of the basement rocks of the Tarim block show that multiphase (ca. 2600–2300 Ma, 1900–1800 Ma, 1140–830 Ma) tectono-thermal events were recorded during formation of the basement. Subordinate crustal materials from the basement were involved in Devonian arc magmatism. This conclusion is comparable to results from the adjacent Kuluketage region of the northern Tarim block.
(3) The distribution of magmatic arcs, from north of Kuqa to north of Korla, suggests that the south-directed subduction lasted from Late Ordovician to Devonian time, and possibly to the Mississippian. The final closure of the South Chinese Tianshan Ocean occurred in the Pennsylvanian, accompanied by north-directed thrusting, terrigenous deposits, and postorogenic magmatism.
Field and laboratory work was supported by Chinese National 973 Project (no. 2009CB825008) and the Major State Project of Special Research No. 2011ZX05008-001. We thank X. Yan and S. Yang for help with zircon cathodoluminescence imaging, L. Su and Y. Li for help with zircon U-Pb analysis, and Z. Liu and Y.H. Yang for help with zircon Hf analysis. M. Faure, D. Cluzel, and J. Charvet are deeply acknowledged for their helpful discussions. We are indebted to Robert Hall and two anonymous reviewers for their constructive comments and suggestions, which led to significant improvement of the manuscript.