The Dunhuang block, situated at the junction of the Tethyan domain, Tarim craton, North China craton and the Central Asian Orogenic Belt (CAOB), consists mainly of Archean–early Mesoproterozoic basement rocks and Paleozoic metamorphic and magmatic rocks. In this study we performed systematic analyses on the Duobagou metaigneous rocks, including zircon U-Pb ages and Hf isotopes and whole-rock geochemistry, to discuss the evolutionary history of the Dunhuang block and the surrounding continental fragments during the Paleozoic.
The metaigneous rocks in the Duobagou area have been identified to be amphibole plagiogneisses, amphibolites, and biotite plagiogneisses; the first two are classified as metamafic rocks, and the latter are metafelsic rocks. The whole-rock major and trace element and Sm-Nd isotopic compositions suggest that the amphibole plagiogneisses were probably produced by the fractionation of the amphibolites, and their protoliths were formed from a mantle wedge, which was associated with the emplacement of subduction-related arc basaltic magmas. The biotite plagiogneisses have geochemical affinities to typical Phanerozoic arc volcanic rocks, and in combination with the Nd-Hf isotopic compositions, these indicate that they were derived from island-arc magmas with varying degrees of crustal contamination. The Duobagou amphibole plagiogneisses and amphibolites have metamorphic ages of 425 ± 31 Ma and 442 ± 2 Ma, respectively, that are in agreement with the crystallization age (436 ± 5 Ma) of the biotite plagiogneisses. The geochemical and geochronological features of these metafelsic rocks provide substantial evidence of a major subduction-related Paleozoic magmatic event in the Duobagou area of the Dunhuang block.
The available data illustrate that the accretionary orogenesis in the Dunhuang block may have continued to the Carboniferous, and those in the Beishan orogenic belt most likely lasted until the Permian. However, those recorded in the North Altun–Qilian orogenic belt probably ended in the Ordovician. The temporal-spatial distributions of the orogeny-related metamorphic-magmatic rocks in the Dunhuang block seemingly match well with those in the CAOB, supporting the interpretation that the Dunhuang block was most likely involved in the orogenic activities of the CAOB during the Paleozoic.
Episodic magmatic and metamorphic activities are the prominent features in China, as in many other continents (Stein and Hofmann, 1994; Condie, 1995, 1998; Rogers, 1996). The Dunhuang block, located between the Beishan orogen (the southern margin of the Central Asian Orogenic Belt, CAOB) to the north and the Altun–North Qilian orogenic belt to the south (Fig. 1A), was previously considered a Precambrian terrane, representing the eastern part of the Tarim craton (Gansu Bureau of Geology and Mineral Resources, 1989; Mei et al., 1998; Xu et al., 1999; Lu et al., 2008; Meng et al., 2011) or the western extension of the North China craton (Zhang et al., 2012, 2013). However, the most recent studies on the newly recognized Paleozoic magmatic and metamorphic rocks (Fig. 1B) arrived at totally different new conclusions (Zhang et al., 2009; Meng et al., 2011; Zong et al., 2012; He et al., 2014; Wang et al., 2016a, 2016b, 2017a, 2017b; Zhao et al., 2016, 2017). The detailed studies on the amphibolite to high-pressure granulite facies metamorphic rocks indicate that the Dunhuang block underwent Paleozoic subduction and/or collision (Zong et al., 2012; He et al., 2014; Wang et al., 2016a, 2017a, 2017b; Zhao et al., 2016), which is closely related to the orogenic events of the CAOB. This is well corroborated by the numerous orogeny-related magmatic rocks in the Dunhuang block (Wang et al., 2016b; Zhao et al., 2017). Meanwhile, there is another very different view on the affiliation of the Paleozoic orogenic activities. Geochronological studies on the tonalite-trondhjemite-granodiorite (TTG) like rocks from the Danghe reservoir and garnet-bearing amphibolites in the Duobagou area provide evidence that the Paleozoic magmatic-metamorphic events were caused by the Altun–North Qilian orogeny (Zhang et al., 2009; Meng et al., 2011). Zhao et al. (2007) referred to the Dunhuang region as the Dunhuang tectonic belt, and proposed that it was reactivated from a stable block to an orogen, representing a Paleozoic accretionary orogenic belt of the southernmost margin of the CAOB between the Tarim craton (TC) and North China craton (NCC) that tectonically extends northward to the Beishan orogen and westward to the eastern South Tianshan belt.
Therefore, there are still two important unsolved questions. (1) Which orogenic belt could be responsible for the Paleozoic magmatism and metamorphism recorded in the Dunhuang block? (2) What is the extended range of the Paleozoic tectonothermal activities in the Dunhuang block? In contrast to earlier studies, which discussed the magmatism and metamorphism separately, we present a combined discussion on the Paleozoic tectonothermal activities of the Dunhuang block. In this study we performed systematic analyses on the metamafic and metafelsic rocks collected from the Duobagou area of the Dunhuang block. We present new zircon U-Pb ages and Hf isotopes, as well as whole-rock geochemical compositions, in order to (1) determine their formation and metamorphic ages, (2) constrain their sources and petrogenesis, (3) discuss the geodynamic processes involved, and (4) provide a further contribution to the present understanding of the evolution of the Dunhuang block.
The Dunhuang block is bounded by the Beishan orogen, Altyn Tagh fault, and Qiemo-Xingxingxia fault to the north, south, and northwest, respectively. Tectonically, it is situated at the junction of the Tethyan domain, TC, NCC, and the CAOB (Fig. 1A). The CAOB (also known as the Altaids), bounded on the north by the Siberia craton and on the south by the TC and NCC (inset, Fig. 1A), is one of the world’s largest accretionary orogens, characterized by multiple subduction-accretion and collision events (Şengör et al., 1993; Şengör and Natal’in, 1996; Windley et al., 2007; Xiao et al., 2009, 2015; Rojas-Agramonte et al., 2011; Wakita et al., 2013). On the southern margin of the CAOB, Paleozoic eclogite (Liu et al., 2011; Qu et al., 2011) and syncollisional granite are well recognized (Zhao et al., 2007; Zhang and Guo, 2008). The Altun–North Qilian orogenic belt, between the Alxa-Dunhuang and Altun-Qilian blocks (Zhang et al., 2015), also provides records of a well-developed Paleozoic high-pressure–low-temperature (HP-LT) metamorphic belt, ophiolitic mélange, and arc magmatic rocks that are related to subduction-accretion processes (Yang et al., 2002, 2008; Song et al., 2004; Qi et al., 2005; Wu et al., 2005, 2007; Zhang and Meng, 2005; Zhang et al., 2007, 2015).
In the Dunhuang block, the Archean–early Mesoproterozoic basement rocks were referred to as the Dunhuang Complex, consisting mainly of TTG and/or TTG-like gneisses and a series of medium- to high-grade metamorphosed supracrustal rocks (Mei et al., 1997; Lu et al., 2008). The outcrops of these rock units have an east-northeast–west-southwest trend. The TTG and/or TTG-like gneisses crystallized ca. 2.7–2.5 Ga and were overprinted by Paleoproterozoic (Zhang et al., 2013) and Paleozoic high-grade metamorphic events (Zong et al., 2013). The supracrustal rocks are dominated by greenschist to amphibolite facies metasedimentary rocks and amphibolite to granulite facies metavolcanic rocks. The metasedimentary rocks, including garnet-kyanite-mica schist, biotite-plagioclase gneiss, quartzite, and graphite-bearing marble (Mei et al., 1997; Yu et al., 1998; Lu et al., 2008), were deposited not earlier than 1.93 Ga and metamorphosed ca. 1.83 Ga (Wang et al., 2013). However, a study on the metapelites from the Hongliuxia area reported a new metamorphic age of 414 ± 3 Ma (Wang et al., 2017a). The Paleoproterozoic garnet-bearing amphibolites and high-pressure mafic granulites in the Hongliuxia area have been interpreted to be associated with the assembly of the Columbia supercontinent (Zhang et al., 2013; Wang et al., 2014). The amphibolites, mafic granulites, and eclogites, all of which have Paleozoic metamorphic ages (440–390 Ma), are interpreted to be related to orogenic processes (Zong et al., 2012; He et al., 2014; Peng et al., 2014; Zhao et al., 2016; Wang et al., 2017b).
Most of the previous studies were conducted in the eastern part of the Dunhuang block. In this contribution we focused on the metamafic and metafelsic rocks from the Duobagou area in the central part of the Dunhuang block.
Electron Microprobe Mineralogical Analysis
Compositional analyses of representative minerals were performed with a JEOL JXA-8100 electron microprobe at the Institute of Geology and Geophysics, Chinese Academy of Sciences (IGGCAS, Beijing, China). Operating conditions were set at an accelerating voltage of 15 kV, beam current of 10 nA, and electron beam diameter of 3 µm. Counting times were 20 s on peaks and 10 s on each background. Natural and synthetic phases were used as standards. Minerals were routinely analyzed for Na, Mg, Al, Si, K, Ca, Ti, Mn, Fe, Zn/Ba, Cr, and Ni. The data were processed with an online ZAF-type correction. The representative compositions of amphiboles are listed in Table 1.
Zircon U-Pb Dating and Hf Isotopic Analyses
Zircons were separated using heavy liquid and magnetic techniques and then hand-picked under a binocular microscope. To observe the internal structure and select a potential target site for U-Pb and Hf isotopic analyses, high-resolution cathodoluminescence (CL) imaging was conducted using a JXA8100 electron microprobe at IGGCAS.
Zircon U-Pb geochronological analyses were performed by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS) at IGGCAS, using an Agilent 7500a quadruple (Q)-ICP-MS, coupled with a 193 nm excimer ArF LA system. The analyses were conducted with a beam diameter of 44 µm, a repetition rate of 8 Hz, and a laser power of 100 mJ/pulse (Yang et al., 2006). Zircon 91500 was used as the standard, and silicate glass samples NIST 610 and GJ-1 were used to optimize the machine. U, Th, and Pb concentrations were calibrated using 29Si as an internal standard and NIST 610 as a reference material. The age calculation and plotting of concordia diagrams were performed using Isoplot version 3.0 (Ludwig, 2003).
In situ Hf isotopic analyses were conducted by multicollector (MC) ICP-MS at the IGGCAS, using the same Geolas-193 LA system. Lu-Hf isotopic compositions were analyzed near the same spots where U-Pb analyses were carried out. A beam diameter of 44 µm was used for all the samples. The repetition rate was 6 Hz and the laser energy was 15 mJ/cm2. Detailed analytical procedure and isobaric interference correction were reported in Wu et al. (2006) and Xie et al. (2008). During the analyses, the 176Hf/177Hf and 176Lu/177Hf ratios of the standard zircon (91500) were 0.282294 ± 15 (2σn, n = 20) and 0.00031 (Yang et al., 2006).
Whole-Rock Geochemical Analyses
Samples were grounded in an agate mill to ∼200 mesh. Major oxides were determined by wavelength-dispersive X-ray fluorescence spectrometry on fused glass at IGGCAS. The analytical precision was estimated to be better than 1% for elements with concentrations >0.5 wt%. Trace element concentrations were analyzed by ICP-MS II at IGGCAS. We used pure elemental standards for external calibration and granite as a reference material. The accuracy of the analyses was better than 2.5%.
For the Sm-Nd isotopic analysis, whole-rock powders were dissolved using a mixture of HF + HNO3 + HClO4 in Savillex Teflon screw-top capsule after being spiked with mixed 149Sm-150Nd tracers. Sm and Nd were separated using the classical two-step ion exchange chromatographic method and measured using a Triton Plus MC thermal ionization MS at IGGCAS. The whole procedure blank was <100 pg for Sm and Nd (Li et al., 2015, 2016). Nd isotopic ratios were corrected for mass fractionation by normalizing to 146Nd/144Nd = 0.7219. The JNdi-1 international standard sample was employed to evaluate instrument stability during the period of data collection. The measured value for JNdi-1 standard was 143Nd/144Nd = 0.512106 ± 0.000010 (n = 9, 2 standard deviations). U.S. Geological Survey reference material BCR-2 was measured to monitor the accuracy of the analytical procedures, and yielded 143Nd/144Nd = 0.512625 ± 0.000008. The εNd (t) (t is time) and TDM2 (depleted mantle model age) values were calculated based on the LA-ICP-MS U-Pb zircon ages. The values of 143Nd/144NdCHUR, 0 = 0.512638 (CHUR—chondritic unfractionated reservoir), 147Sm/144NdCHUR, 0 = 0.1967, 143Nd/144NdDM, 0 = 0.513153, 147Sm/144NdDM, 0 = 0.2137, and k = 6.54 × 10−12 were used in the calculation (Jacobsen and Wasserburg, 1984).
Petrography and Mineralogy
Fresh samples were collected from stream-cut exposures. The sample locations are presented in Figure 1B. Lithologically and structurally, the metamafic rocks can be identified to be amphibole plagiogneiss and amphibolite.
The amphibole plagiogneiss, enclosed in metasedimentary rocks, occurs as layers or veins. However, it has no clear boundaries with amphibolite when they co-occurred in the field (Fig. 2A). It is gray in color and has well-developed schistosity defined by columnar amphiboles (Fig. 2B). With a dominant lithology of medium- to coarse-grained metamafic rocks, this rock has typical mineral assemblage of 30%–40% amphibole and 35%–45% plagioclase, with minor biotite and microcline, and accessory magnetite, zircon, and apatite. Amphibole grains are green to brown and mostly show ragged grain boundaries (Fig. 2C). The amphibole crystals are classified as magnesiohornblende and display no observable compositional zoning (Fig. 3A). The discrepant mineral compositions of amphiboles indicate variable levels of metamorphism from greenschist to epidote amphibolite to amphibolite facies (Fig. 3B). The biotites occur typically as small crystals (<0.2 mm), constituting between 3% and 5% by volume of the whole rock.
The amphibolite occurs as lenses in mica schist and/or gneiss or marble in the field (Figs. 2D, 2E). It is dark in color and consists almost exclusively of amphiboles. The amphiboles are commonly present as well-developed crystals and laths (Fig. 2F). The euhedral-subhedral crystals and straight boundaries of most amphibole grains are suggestive of minor alteration during retrograde metamorphism. All the analyzed amphiboles belong to the hornblende group, the majority being magnesiohornblende, with a few near the field of tschermakite (Fig. 3A). No systematic zoning was observed in the amphiboles, but their compositions indicate an epidote amphibolite to amphibolite facies metamorphism (Fig. 3B).
The biotite plagiogneiss is usually distributed in the metasedimentary rocks as veins in the field. It is gray to dark in color with remarkable gneissosity (Figs. 2G, 2H). The typical minerals include lath-like plagioclase (40%–50%), sugar-like quartz (25%–35%), flake-like biotite (5%–10%), minor microcline, and accessory minerals (Fig. 2I).
Zircon U-Pb Ages and Hf Isotopes
Amphibole Plagiogneiss (15DB-12)
Zircon grains from this sample are prismatic to subrounded in shape and variable in size, with crystal lengths of 50–100 µm. Most grains are characterized by typical oscillatory zoning, and a few have planar zoning or no zoning (Fig. 4). We selected 29 unbroken and crack-free zircon grains for U-Pb analyses; they mostly exhibit high Th/U ratios (>0.1), indicating a magmatic origin. The analytical results define a normal discordant line with an intercept age of 2485 ± 140 Ma (mean square of weighted deviates, MSWD = 6.0; Fig. 5A), representing the formation age of the protolith. Analyses of zircons with lower Th/U ratios (<0.1) define a weighted mean age of 425 ± 31 Ma (N = 6, MSWD = 8.8; Fig. 5A), which more likely represents the metamorphic age of the amphibole plagiogneiss.
We analyzed 26 dated spots for Hf isotopic compositions. The zircon grains with metamorphic age of 425 Ma yielded a large range of initial Hf compositions (176Hf/177Hf(i) = 0.282084–0.282729) and highly variable εHf (t) values (Fig. 6A), with TDM2 ages ranging between 880 and 2350 Ma. The magmatic zircons have initial Hf compositions changing between 0.280634 and 0.282508. Their corresponding εHf (t) values and TDM2 model ages range from −11.9 to + 12.0 and 1067–3609 Ma, respectively.
Zircons from the amphibolite sample are rounded to subrounded and small. All zircon grains have homogeneous internal structures and display no zoning (Fig. 4). We analyzed 30 zircon grains for U-Pb ages. The Th/U ratios of these zircon vary from 0.002 to 0.705, dominantly between 0.002 and 0.013, suggesting a metamorphic origin. The age data provide a concordant age of 442 ± 2 Ma (MSWD = 0.12; Fig. 5B), which is taken to represent the metamorphic age of the amphibolite. Of the 30 grains analyzed for U-Pb ages, 28 were selected for Hf isotopic analyses. They have variable initial Hf ratios (176Hf/177Hf(i) = 0.282444–0.282878) and positive εHf (t) values between + 1.7 and + 12.9, except one zircon grain (εHf (t) = −2.2; Fig. 6B). The zircons with positive εHf (t) values have TDM2 model ages in the range of 600–1567 Ma.
Biotite Plagiogneiss (15DB-32)
Zircon grains from this sample are prismatic in shape and range from 120 to 180 µm in the long axis. CL images reveal clear core-rim textures. Most of the zircon cores are dark and irregular with clear oscillatory zoning; even though there are few cores, they still display weak oscillatory zoning. All the zircon rims are bright and have no zoning (Fig. 4). A total of 48 analyses were performed on the zircon cores; the majority of the data plot on or near the concordia line (Fig. 5C), and yield a weighted mean 206Pb/238U age of 436 ± 5 Ma (MSWD = 9.7; Fig. 5D). These zircon grains have variable Th/U ratios (0.02–1.76) but mostly >0.1, and therefore this age is interpreted to represent the crystallization age of the protolith. Four spots intentionally shot on the inherited cores are below the concordia with much older ages, interpreted to be xenolithic in origin. We selected 45 dated zircons for Hf isotopic analyses; they show a narrow range in initial Hf compositions (176Hf/177Hf(i) = 0.282048–0.282260) and have negative εHf (t) values from −16.4 to −8.7 (Fig. 6B), with TDM2 model ages from 1967 to 2323 Ma.
The whole-rock major and trace element compositions for the metamafic and metafelsic samples in the Duobagou area are listed in Table 4. The whole-rock Sm-Nd isotope analytical results as well as the calculated initial Nd ratios are presented in Table 5. Because the Nd isotope system is generally unaffected by metamorphism (McLennan and Hemming, 1992; Zhao et al., 1992; McDaniel et al., 1994), the Sm-Nd model ages of the amphibole plagiogneisses and amphibolites were calculated using their metamorphic ages.
Amphibole Plagiogneisses (15DB-15, 15DB-16)
The amphibole plagiogneisses have low contents of SiO2 (50.0–54.0 wt%), relatively high contents of total Fe2O3 (10.0–10.6 wt%), Al2O3 (15.2–16.2 wt%), and CaO (8.6–10.6 wt%), and moderate concentrations of MgO (6.1–6.4 wt%) and Na2O (2.6–3.6 wt%). All samples plot in the field of volcanic rocks (Fig. 7A) and have a basaltic affinity (Fig. 7B). In the AFM (Al2O3-FeO-MgO) diagram (Fig. 7C), they are in the field of tholeiitic series, and their Mg numbers vary slightly from 54 to 57. With respect to trace element concentrations, these rocks have relatively low total rare earth element (REE) contents ranging from 79 to 87 ppm. In the chondrite-normalized REE diagram (Fig. 8A), they have moderate light (L) REE-enriched patterns (LaN/YbN = 5.62–6.44) and minor positive or negative Eu anomalies (δEu = 0.92–1.06). On the spider plots normalized to primitive mantle (Fig. 8B), the rocks are characterized by remarkable negative Nb-Ta, Zr-Hf, and Ti anomalies, and pronounced positive spikes in K, Sr, U, and Pb.
For the amphibole plagiogneisses, the measured 147Sm/144Nd values are 0.1342–0.1412, while 143Nd/144Nd values are 0.512422–0.512525, corresponding to εNd (425) values of −0.8 to + 0.5 (Fig. 9A) and two-stage Nd TDM2 ages of 1.25–1.12 Ga.
Amphibolites (15DB-22, 15DB-23, 15DB-24, 15DB-25)
The amphibolites show a narrow compositional range in SiO2 (44.9–46.3 wt%) and are characterized by high contents of total Fe2O3 (11.0–11.2 wt%), MgO (14.6–14.9 wt%), CaO (11.6–12.6 wt%), and Al2O3 (11.3–11.9 wt%), and low contents of K2O (0.67–0.76 wt%) and Na2O (1.18–1.30 wt%), resulting in moderate Na2O/K2O ratios (1.55–1.91). The significantly high abundances of MgO in the amphibolites place them in the high-Mg series, and also result in relatively high Mg# numbers (75.3–75.9). These samples plot in or near the volcanic rock field (Fig. 7A) and indicate a tholeiitic protolith (Fig. 7C). In the chondrite-normalized REE diagram (Fig. 8A), they display moderate fractionation between LREEs and heavy (H) REEs (LaN/YbN = 3.62–4.38), with minor positive or negative Eu anomalies (δEu = 0.92–1.05). When normalized to primitive mantle (Fig. 8B), these amphibolite samples show large ion lithophile element (LILE) enrichment relative to high field strength element (HFSE) enrichment, coupled with conspicuous Nb-Ta and Zr-Hf depletion, and positive U, K, and Pb anomalies. It is noteworthy that the trace element patterns of the amphibolites are almost the same as those of the amphibole plagiogneisses.
The amphibolites have Sm contents of 4.58–4.64 ppm, and Nd contents of 24.0–24.4 ppm. The rocks have 147Sm/144Nd values varying from 0.1153 to 0.1156 and 143Nd/144Nd values from 0.511636 to 0.511654. Therefore, the corresponding εNd (442) values range from −5.0 to −14.6 (Fig. 9A), with TDM2 model ages of 2.40–2.43 Ga.
Biotite Plagiogneisses (15DB-28, 15DB-29, 15DB-30)
The biotite plagiogneisses have SiO2 contents ranging from 68.4 to 69.2 wt% and Al2O3 from 13.7 to 13.9 wt%. They are characterized by low Na2O/K2O ratios (0.76–0.81) and MgO contents (1.58–1.84 wt%), with Mg# values ranging from 45.3 to 46.7. In the An-Ab-Or (anorthite-albite-orthoclase) diagram (Fig. 7D), they display a granodioritic trend. The rocks are characterized by high total REE contents ranging from 186 to 211 ppm and show enrichment in LREEs and depletion in HREEs with pronounced negative Eu anomalies (δEu = 0.55–0.59) (Fig. 8A). They have low contents of Cr (30–40 ppm), Co (8–9 ppm), and Ni (8–10 ppm), and high Th/La ratios (0.44–0.46). In the primitive mantle-normalized trace element diagram (Fig. 8B), they have enrichments in U, K, and Pb but depletion trends in Nb, Ta, and Ti. The most notable features that differentiate the metafelsic from the metamafic rocks are the absence of pronounced negative Zr-Hf anomalies.
The biotite plagiogneisses have Sm and Nd contents ranging from 6.57 to 7.26 ppm and 33.5 to 36.9 ppm, respectively. Their 147Sm/144Nd and 143Nd/144Nd values range from 0.1187 to 0.1191 and 0.511937 to 0.511943, respectively, with initial 143Nd/144Nd values of 0.511597–0.511602. The εNd (436) values range from −9.2 to −9.3 (Fig. 9A) and TDM2 model ages are from 1.96 to 1.95 Ga.
Effects of Alteration and Metamorphism
Although the least altered samples were selected for chemical analyses, for metamorphic rocks, mobile elements are prone to change in concentrations during postmagmatic alteration or metamorphism. Therefore, the first step of identifying their original geochemical characteristics is to evaluate the element mobility.
All samples have relatively low loss on ignition (LOI) values (<2%), suggesting that they have not been strongly altered, hydrated, or carbonated. The negligible to small Ce anomalies indicate limited REE mobility (Polat and Hofmann, 2003). Each rock type shows coherent variations in the chondrite-normalized REE and primitive mantle-normalized trace element patterns (Figs. 8A, 8B), suggesting that the REEs and HFSEs were largely preserved during subsequent alteration and metamorphism, and thus can be used to assess the characteristics of their protoliths (Taylor et al., 1986; Middelburg et al., 1988). Zr is considered as one of the least mobile elements, having undergone only dilutions and concentrations caused by the transport of other elements into and out of a system (Pearce et al., 1992; Rolland et al., 2009), and thus is preferred as a reference to monitor the mobility of other trace elements. In Figure 10 Zr is used as the horizontal axis to demonstrate the alteration of other elements; Nd, Nb, Y, and Th have well-defined liner relationships relative to Zr for all samples, suggesting that these elements as well as Zr have behaved as immobile elements. The Zr-Rb and Zr-Sr plots are also shown in Figure 10, and the large scatter suggests that Rb and Sr have more or less been altered; therefore, they were not used in the petrogenetic evaluation of the samples. The HFSE and other elements showing well-preserved igneous characteristics were used to identify and assess the behavior of the primary magmatic source. Some major element data have, however, been plotted for classification and for general characterization.
Origin of Metamafic and Metafelsic Rocks
On the basis of petrological and geochemical features, the amphibole plagiogneisses and amphibolites are classified as metamafic rocks. The mineral chemistry of amphiboles suggests that these rocks have undergone greenschist to amphibolite facies metamorphism. With respect to trace element concentrations, the amphibole plagiogneisses have relatively higher element abundances in comparison with the amphibolites, but the similar trends in REE patterns (Fig. 8A) and trace element spidergrams (Fig. 8B) suggest a common genetic link; that is, they were probably derived from the same magmatic source, or the amphibole plagiogneisses were produced by the fractionation of the amphibolites. The assumption of fractionation of the amphibolites resulting in the formation of the amphibole plagiogneisses is corroborated by a series of variation diagrams (Figs. 11A, 11B). The negligible changes in the Nb/La and Nb/Ce ratios also fully support the derivation of the amphibole plagiogneisses from amphibolites. The selective enrichment of LREEs and LILEs over HFSEs, leading to relatively high LaN/YbN and LILE/HFSE ratios, are the typical characteristics of slab dehydration and mantle wedge melting in island-arc settings (Zack and John, 2007). Furthermore, their significant negative Nb-Ta, Zr-Hf and Ti anomalies suggest a close resemblance to typical subduction-related intraoceanic arc basalts (Pearce and Peate, 1995; Pearce, 2008). The very low Ta/Nb ratios (<0.1) are also consistent with slab fluid–dominated sources (Elliott et al., 1997). In various tectonic discrimination diagrams (Figs. 12A, 12B), the amphibolites plot in the field of island-arc basalt, while the amphibole plagiogneisses are in the fields of island-arc lavas and within-plate basalts. Such dual tectonic affiliations suggest that the amphibole plagiogneisses formed in a source dominated by an island-arc setting but with an addition of within-plate–like components. The presence of amphibole and biotite in the amphibole plagiogneisses requires high water contents (>3 wt%), indicating relatively wet crystallizing magmas (Gaetani et al., 1993; Sisson and Grove, 1993; Moore and Carmichael, 1998; Barclay and Carmichael, 2004; Smith et al., 2009). However, both the lithospheric and asthenospheric mantle, as well as the continental crust, have very low water contents, while a mantle wedge metasomatized by subduction-related fluids (e.g., McInnes et al., 2001) can have a water content as high as 12 wt% (Gaetani and Grove, 1998).
The analyzed amphibolite samples have positive zircon εHf (t) values but negative whole-rock εNd (t) values. McDaniel et al. (1994) argued that simultaneous Sm/Nd and Nd isotope data on ∫Sm/Nd versus εNd (t) plot is required to more accurately interpret Nd isotope data. In Figure 9B, the analyzed amphibolites evolve along the mixing array, suggesting that the magma may have been derived from one or more additional sources. Therefore, the isotope data of the amphibolites indicate that their parental magma was sourced from a mixing source, involving the participation of both juvenile components and old materials. This is a reasonable explanation for the inconsistency between the zircon εHf (t) and whole-rock εNd (t) values. The highly variable εHf (t) values of the amphibole plagiogneisses indicate high Hf heterogeneity in the source. Such heterogeneity could have resulted from juxtaposed magmas with different Hf isotopic compositions in a single sample (Griffin et al., 2002; Kemp et al., 2007). Their negative Nb-Ta-Ti anomalies reflect crustal contamination of the source, while the negative εNd (t) and εHf (t) values suggest a significant contribution of preexisting crust. Therefore, the variable Nd-Hf isotopes of the amphibole plagiogneisses possibly indicate magma mixing between the depleted mantle and continental crust. This is consistent with the magma sources for the amphibolites, further supporting our interpretation that the amphibole plagiogneisses most likely formed from the fractionation of the amphibolites. However, more convincing evidence is needed to test this hypothesis, because the amphibolites lack crystallization ages. Based on the previous discussion, the Duobagou metamafic rocks most likely formed from a mantle wedge, which was associated with the emplacement of subduction-related arc basaltic magmas.
The biotite plagiogneisses are metavolcanic rocks and have a granodioritic affinity. When plotted on a primitive mantle-normalized multielement variation diagram, they show significant negative Sr and Ti anomalies. All the above-mentioned lines of evidence are very typical features of classical Phanerozoic arc volcanics (Peccerillo and Taylor, 1976; Martin, 1999). Moreover, the biotite plagiogneisses have calcic affinity (Fig. 13A) and plot well within the volcanic arc–postcollisional granites field in the Rb versus (Nb + Y) discrimination diagram (Fig. 13B), both of which are indicative of island-arc magmas with little crustal contamination. That notwithstanding, the negative whole-rock εNd (t) values and older TDM2 model ages (1.96–1.95 Ga), as well as the negative zircon εHf (t) values and much older TDM2 model ages (2.32–1.97 Ga) indicate a great amount of crustal involvement. These characteristics, together with the presence of older xenolithic zircons, have provided evidence for a preexisting older crust.
The amphibole plagiogneiss and amphibolite in the Duobagou area metamorphosed at 425 ± 31 Ma and 442 ± 2 Ma, respectively, and the biotite plagiogneiss from the same area crystallized at 436 ± 5 Ma, providing complete evidence for the Paleozoic tectonothermal activities in the study area. Geochemically, the biotite plagiogneisses have a pronounced resemblance to typical Phanerozoic arc calcic volcanics (Peccerillo and Taylor, 1976; Martin, 1999), and were most likely derived from island-arc magmas with significant crustal contamination. Considering all these characteristics, the Paleozoic (436 Ma) biotite plagiogneisses provide evidence for a major subduction-related Paleozoic magmatic event in the Duobagou area of the Dunhuang block.
Zhao et al. (2017) conducted systematic studies on the Paleozoic intermediate-acid intrusive rocks from the Dunhuang block, and concluded that these rocks could be subdivided into three suites from north to south, based on their temporal-spatial distribution, petrogenesis, and tectonic setting (Zhu et al., 2014; Zhao et al., 2017). On the basis of the reported age data of the Paleozoic magmatic-metamorphic events, the Dunhuang block displays characteristic north-south zoning. The Duobagou and Mogutai areas, in the central part of the Dunhuang block, are tectonically located on the same zone. This is corroborated by the Duobagou metaigneous rocks and Mogutai high-pressure granulites, both of which document a subduction-related Paleozoic magmatic event (Zong et al., 2012; He et al., 2014). However, it does not represent the subduction boundary, because much younger arc-related intrusive rocks have been reported (Zhao et al., 2017).
The Altun–North Qilian orogenic belt, located in the southern Dunhuang block, documents numerous Paleozoic accretion- and/or collision-related tectonothermal events. The diagnostic features include a HP-LT metamorphic belt, ophiolitic mélange, and arc rocks associated with subduction collision (Zhang et al., 2015). In the North Altun, the ophiolite mélange are believed to have formed in various tectonic settings during the early Paleozoic (524–448 Ma) (Yang et al., 2002, 2008). The numerous subduction accretion–related magmatic rocks indicate that the commencement of the subduction accretion of the North Altun oceanic crust occurred ca. 481–467 Ma (Qi et al., 2005; Hao et al., 2006; Wu et al., 2007; Kang et al., 2011). The younger S-type granites (446–431 Ma) indicate that the closure of the paleo–North Altun Ocean was not later than 446 Ma (Chen et al., 2003; Wu et al., 2005, 2007). In the North Qilian orogenic belt, the temporal-spatial distributions of the ophiolitic mélange nearly recorded the entire evolution of the paleo–North Qilian Ocean; the beginning of the ocean was defined by the earliest ophiolites (550–529 Ma) (Shi et al., 2004; Hou et al., 2005; Song et al., 2013), and the subduction initiated in the early Paleozoic (Cambrian) (Meng et al., 2010; Xia et al., 2012). The backarc basin was formed in the Early Ordovician (Xia and Song, 2010; Song et al., 2013).
The Beishan orogenic collage, occupying the southernmost part of the CAOB, also developed Paleozoic ultrahigh-pressure metamorphic rocks, ophiolitic mélange, and collision-related magmatic rocks (Zuo et al., 1991; Ren et al., 2001; Guo et al., 2006; Yu et al., 2006; Zhao et al., 2007; Zhang and Guo, 2008; Mao et al., 2010; Ao et al., 2010, 2012; Liu et al., 2011; Qu et al., 2011). The Liuyuan eclogites and their surrounding rocks support a model that the eclogites started as oceanic crust in the paleo–Asian Ocean, and defined a process of subduction and exhumation from Ordovician to Silurian (Qu et al., 2011). The earliest subduction event in the Beishan was recorded by the Yueyashan-Xichangjing suprasubduction zone–type ophiolite (533 Ma), and it lasted from the early Cambrian to the Carboniferous–Permian (Xiao et al., 2004, 2008; Windley et al., 2007; Ao et al., 2012).
The accretionary orogenesis in the North Altun–Qilian orogenic belt most likely ended in the Ordovician, while it lasted until the Permian in the Beishan orogenic belt. The available data illustrated that the accretionary orogenesis in the Dunhuang block may have continued to the Carboniferous, supporting an argument that the Dunhuang block was involved in the orogenic activities of the CAOB.
1. Lithologically and structurally, the metamafic rocks in the Duobagou area of the Dunhuang block have been subdivided into amphibole plagiogneiss and amphibolite, and the fractionation of the amphibolite probably resulted in the formation of the amphibole plagiogneiss.
2. The metamafic rocks were most likely derived from a mantle wedge, which was associated with the emplacement of subduction-related arc basaltic magmas. These two kinds of rocks have consistent metamorphic ages of 442–425 Ma.
3. The metafelsic rocks have features indicative of island-arc magmas with crustal contamination, together with their crystallization age of 436 Ma, indicating a subduction-related Paleozoic magmatic event in the Duobagou area of the Dunhuang block.
This work was financially supported by the National Key Research and Development Program of China (2017YFC0601205), 973 Program (2014CB440801), National Natural Science Foundation Projects (41272107, 41230207, 41202150, 41390441, 41102132, and 41302167), National 305 Projects (2011BAB06B04–1) and the China Postdoctoral Foundation Fund Project (2015M580133). This paper is a contribution to International Geological Correlation Program 592.