The Longshou Shan of western China is the northern backstop of the Cenozoic Himalayan-Tibetan orogen and occupies a key linkage between the Tarim continent and North China craton which separate the pre-Cenozoic Tethyan orogenic system and Central Asian orogenic system. Therefore, the Paleoproterozoic–Paleozoic evolution of this region is critical to understanding the extent of overprinting Cenozoic deformation, construction of the Eurasian continent, and relationships between the pre-Cenozoic Tethyan orogenic system and Central Asian orogenic system. Here we present detailed field observations and results of geochronological and major and trace element and Sr-Nd isotope geochemical analyses of samples from the Longshou Shan to decipher its complex Paleoproterozoic–Paleozoic tectonic history. Our results show that the Paleoproterozoic basement rocks of the Longshou Shan were part of the North China craton and involved in Paleoproterozoic northern North China orogeny. A ca. 965 Ma granitoid in the Longshou Shan provides key evidence for a spatial linkage between northern Tibetan continents, the North Tarim continent, and the North China craton in the early Neoproterozoic. The presence of Early Ordovician granitoids and arc volcanic rocks in the Longshou Shan suggest that bivergent subduction of Qilian oceanic lithosphere occurred during the early Paleozoic. Crustal shortening and thickening during Ordovician–Carboniferous orogenesis are evidenced by the presence of several unconformities in the Longshou Shan. Late Carboniferous arc granites exposed in the study area are likely associated with the southward subduction of the Paleo-Asian Ocean to the north and with Permian siliciclastic strata sourced from a proximal arc-subduction system, based on detrital zircon ages. Although the tectonic history of the Longshou Shan can be traced back to Neoproterozoic time, most of the recorded deformation and uplift of the region occurred during the early Paleozoic Qilian orogeny and late Paleozoic Central Asian orogeny. Furthermore, we interpret that the several orogenic events recorded in the Longshou Shan (i.e., northern North China, Qilian, and Central Asian orogenies) are spatially and temporally correlative along strike with those recorded in the Tarim and North China cratons.

The Cenozoic Tibetan Plateau of western China was constructed during the most recent phase of development of the Tethyan orogenic system (e.g., Şengör and Natal'in, 1996; Yin and Harrison, 2000; Wu et al., 2020; Wu et al., 2016), which is bounded to the north by the Tarim continent and North China craton (Fig. 1A). North of the Tarim continent and North China craton is the Central Asian orogenic system, which evolved over ~1 b.y. during the evolution of the Paleo-Asian Ocean (e.g., Xiao et al., 2004; Windley et al., 2007; Wan et al., 2018). Despite being integral to the construction of the Eurasian continent, the Proterozoic–Phanerozoic tectonic histories of the Tethyan orogenic system and Central Asian orogenic system have yet to be clearly linked in space and time (Fig. 1A) (e.g., Şengör and Natal'in, 1996; Yin and Nie, 1996; Kusky et al., 2007; Zhao et al., 2018; Zuza and Yin, 2017; Li et al., 2018; Zhang and Gong, 2018; Zuza et al., 2018; Wu et al., 2016, 2021a). Critical unknowns regarding their assembly and genetic relationships include the timings and settings of petrogenesis and metamorphism of Precambrian basement rocks, the style and timing of Phanerozoic deformation, and their pre-Cenozoic configuration (Fig. 1A).

Critical unknowns regarding the Proterozoic–Phanerozoic evolution of the Tethyan orogenic system and Central Asian orogenic system are particularly evident in the Longshou Shan area, located just north of the Qilian Shan and Tibetan Plateau (Fig. 1B). Previous work has proposed that the Longshou Shan represents either the western extension of the North China craton since the Paleoproterozoic or the eastern extension of the North Tarim continent (Fig. 1B) (e.g., Zhao et al., 2005; Zhai and Santosh, 2011; Zhang et al., 2012; Gong et al., 2016; Zhang and Gong, 2018; Wu et al., 2021a). Alternatively, other workers have suggested that the Longshou Shan represents a block of basement rocks that was separated from the Yangtze block of southern China or other Gondwanan subcontinents (Fig. 1B) (e.g., Li et al., 2004; McKenzie et al., 2011; Dan et al., 2014; Song et al., 2017). Neoproterozoic (ultra)mafic intrusions (ca. 832–827 Ma) located within the Longshou Shan have been interpreted to be related to continental rifting (e.g., Li et al., 2004, 2005; Tung et al., 2013; Tang et al., 2014; Zhang et al., 2010) (Fig. 2). However, the Neoproterozoic magmatic record is poorly resolved, which has limited our understanding of the tectonic setting during that time. In addition, the precise timings, subduction polarities, and tectonic regimes that operated during the Paleozoic in the Longshou Shan remain inadequately constrained. These unknowns are centered on the petrogenesis of ca. 444–414 Ma arc granitoids and ca. 424–421 Ma (ultra)mafic rocks exposed in the Longshou Shan (Fig. 2), which were possibly generated during the development of either the Central Asian orogenic system to the north or the Qilian orogen to the south located between the North China craton and Qaidam block (e.g., Zhang and Gong, 2018; Song et al., 2017; Liu et al., 2020a, 2020b; Zhang et al., 2017; Wang et al., 2020; Wu et al., 2021a; Zeng et al., 2016, 2021; Duan et al., 2015). Further complicating these uncertainties is evidence for Carboniferous passive-margin sedimentation and magmatic quiescence in the Qaidam block to the south, which contrasts with evidence for Late Carboniferous arc magmatism in the Central Asian orogenic system including a diverse record of granitic intrusions (e.g., Cowgill et al., 2003; Wu et al., 2016; Zhang et al., 2007, 2009).

Although there are existing published geochemical and geochronological data sets for the Longshou Shan area, most of these data lack a clear structural and tectonostratigraphic context for establishing a meaningful tectonic synthesis. In this contribution, we address the above issues by combining field-based lithotectonic observations with detrital zircon U-Pb geochronology and whole-rock major and trace element and isotope geochemistry across the Longshou Shan (Fig. 2). Our results provide new constraints on the number and setting of tectonomagmatic events recorded in the Longshou Shan (Fig. 3) and improve our understanding of the broader Proterozoic–Paleozoic tectonic evolution of central Asia, which is critical to evaluating the evolution of the Eurasian continent and deciphering the kinematics and extent of Cenozoic strain.

The northwest-trending, ~30-km-wide and ~300-km-long Longshou Shan is located along the southwestern margin of the North China craton and bounded to the south by the northwest- to west-striking, right-slip Longshoushan fault which was active during the early Paleozoic Qilian orogeny (e.g., Shi et al., 1995; Wu et al., 2021a) (Figs. 1B and 2). The lithostratigraphy of the Longshou Shan consists of crystalline basement rocks of the Paleoproterozoic Longshoushan Group overlain by Mesoproterozoic–Paleozoic sedimentary rocks (e.g., Gong et al., 2016; Zhang and Gong, 2018; Wu et al., 2021a). The Paleoproterozoic Longshoushan Group mainly consists of banded and lens-shaped migmatite intercalated with plagioclase amphibolite, marble, schist, quartzite, and orthogneiss (Fig. 3A). These rocks experienced amphibolite-facies regional metamorphism, based on a youngest detrital zircon age peak (Fig. 3B) (e.g., Gong et al., 2016; Wu et al., 2021a), and were subsequently folded and intruded by ca. 2.05 Ga leucogranites and Neoproterozoic ultramafic and mafic rocks (e.g., Gong et al., 2016; Zhang and Gong, 2018; Liu et al., 2020b; Wu et al., 2021a). The Mesoproterozoic–Paleozoic sedimentary sequence overlying the Paleoproterozoic Longshoushan Group includes the Mesoproterozoic Dunzigou Group at its base, which forms an unconformity atop the basement rocks and mainly consists of weakly metamorphosed siliciclastic and carbonate rocks (e.g., Gong et al., 2016; Wu et al., 2021a) (Fig. 3A). The Dunzigou Group is overlain by the Neoproterozoic Hanmushan Group (e.g., Gong et al., 2011), which consists of rift-related siliciclastic rocks, limestone, mafic rocks, and glacial deposits (Fig. 4A). Above the Hanmushan Group are Cambrian strata, which are largely exposed north to the city of Shandan and consist of quartz sandstone, slate, and volcanic rocks intruded by late Paleozoic granitoids (e.g., Xue et al., 2017; Liu et al., 2020b) (Fig. 3A).

Upper Silurian–Devonian strata unconformably overlie lower Paleozoic strata and include conglomerate and sandstone intercalated with amygdaloid basalt and andesitic tuff (e.g., Zhang et al., 2012; Zhang et al., 2017). Lower Carboniferous strata mainly include littoral-facies siliciclastic and carbonate rocks, whereas the upper Carboniferous–lower Permian strata consist of marine- to continental-facies rocks intercalated with coal layers and interbedded with dacite tuff (e.g., Pan et al., 2004; Xue et al., 2017; this study). Early Permian mafic dike swarms are widespread in the Longshou Shan (i.e., Dan et al., 2014; Zhang et al., 2017). During the Permian, the Longshou Shan experienced fluvial-lacustrine sedimentation, evidenced by Permian strata, which were subsequently placed atop Paleoproterozoic metamorphic rocks along a low-angle normal fault (Fig. 4B). Unconformably overlying upper Permian strata are Jurassic and Cretaceous terrestrial rocks. Cenozoic fluvial deposits overlie Cretaceous strata along a regional unconformity (Pan et al., 2004) (Fig. 3A).

Numerous Proterozoic and Paleozoic granitoid plutons are exposed throughout the Longshou Shan (Fig. 3C; Table S11). Paleoproterozoic granitoids and granitic gneisses range in age from ca. 2.49 to 1.76 Ga and form the main component of the Longshou Shan basement metamorphic sequence (e.g., Gong et al., 2016; Wu et al., 2021a). The Neoproterozoic ultramafic Jinchuan intrusion (ca. 827 Ma) has been interpreted to be related to continental rifting (e.g., Li et al., 2005; Tung et al., 2013). Paleozoic plutons have ages between ca. 448 and 329 Ma and are mostly attributed to arc and/or syn- to post-orogenic magmatism (e.g., Zhang et al., 2017; Xue et al., 2017; Wang et al., 2020; Tang, 2015). In this study, we document additional magmatic pulses that occurred in the Longshou Shan based on field observations and geochronologic analyses of previously undated Paleoproterozoic, early Neoproterozoic, Early Ordovician, and late Carboniferous granitoids.

This work builds on detailed field mapping and documentation provided in Wu et al. (2021a), and here we present focused observations of the lithostratigraphy, contact relationships, and results of analyzed igneous samples. The present-day physiography and spatial distribution of rock exposures throughout the range are largely a product of Cenozoic southwest-directed thrust faulting (Fig. 2). The Longshoushan fault separates rocks of the Longshou Shan in the north from Mesozoic–Cenozoic strata of the low-lying Hexi Corridor (Fig. 1B). The Longshoushan fault strikes 290°–300° and dips ~70°NE and is traced from Hongyashan (around 38°23.5'N, 102°49.8'E) in the east to Shandan and Gaotai in the west and is inferred to continue to westward (Fig. 2). The subsurface structure of the Longshoushan fault is thought to be complex, with multiple north-directed splay thrusts. The root zone of one such thrust at depth may represent the uppermost section of basement of the Qilian orogen (Shi et al., 1995). At one location, an undeformed Paleoproterozoic pink-colored granite is thrust atop quartzoschist of the Paleoproterozoic Longshoushan Group (Fig. 4C). There, the lack of any significant topographic expression suggests that the thrust was active prior to the Cenozoic. This thrusting may have occurred in the early Paleozoic during the Qilian orogeny; alternatively, this thrusting may have occurred in the early Neoproterozoic, coeval with an enigmatic ca. 0.9–1.0 Ga magmatic event evidenced by an early Neoproterozoic granitic intrusion within schist of the Longshoushan Group (Fig. 4D). Scattered exposures of Early Ordovician volcanic rocks and pink-colored granitic intrusions occur within Mesoproterozoic strata with unclear contact relationships (Fig. 4E). A low-angle normal fault between the Permian and Proterozoic strata is traced 50 km to the west of the Longshou Shan, where the fault juxtaposes Cretaceous strata in its hanging wall over Proterozoic metamorphic rocks in its footwall (Fig. 4B) and has a displacement of ~100 m. These field relationships suggest that this normal fault may have been active during the Cretaceous and related to middle to Late Cretaceous regional extension (e.g., Wu et al., 2021b; Wang et al., 2022). A late Carboniferous granitoid pluton is thrust over the Cretaceous strata, and the field observations show that pink K-feldspar granitoid intrudes gray-colored granitoid (Fig. 4F).

U-Pb Zircon Geochronology

Cathodoluminescence (CL) imaging and U-Pb dating of individual zircon grains from six igneous and two metasedimentary samples from the Longshou Shan were performed at the Key Laboratory of Continental Collision and Plateau Uplift, Chinese Academy of Sciences, Beijing, China. Zircons were analyzed via laser ablation multi-collector inductively coupled plasma mass spectrometry (LA-MC-ICP-MS) using the analytical procedures described in Gehrels et al. (2003). The fractionation correction and results were calculated using the software GLITTER 4.0 (Griffin, 2008). A common-Pb correction was applied following the method described in Andersen (2002). Age calculations and concordia plots were made using Isoplot 4.7 (Ludwig, 2003). Key field relationships of analyzed samples are shown in Figure 4. The U-Pb ages presented are 206Pb/238U ages for grains younger than ca. 1000 Ma and 206Pb/207Pb ages for grains older than ca. 1000 Ma (Table S2, footnote 1). To interpret the crystallization ages for the igneous samples, we examined only analyses with discordance of <10%. We focused on the dominant components of younger analyses and interpreted the weighted mean age of the youngest cluster (n ≥ 3) of zircon ages as the crystallization age. Generally, distinctly older age populations from analyzed CL cores were attributed to zircon inheritance, and distinct young zircon ages from analyzed CL bright rims are less straightforward to interpret but may result from Pb loss due to a younger metamorphic event. These inferences are consistent with the zircon Th/U ratio data (Fig. 5; Table S2). Uncertainties on individual analyses are reported at the 1σ level and plotted at the 2σ level in Figure 6. All uncertainties, including analytical and systematic, of the weighted mean ages are reported at the 95% confidence level. The weighted mean ages and probability density distribution were calculated and plotted using Isoplot 3.0 of Ludwig (2003). For the detrital zircon age data, we only considered analyses that were <30% discordant or <5% reversely discordant (e.g., Gehrels et al., 2003). Normalized probability density plots were used to compare the different data sets using George E. Gehrel's 2010 software provided by the Arizona LaserChron Center (Tucson, Arizona, USA) (Fig. 3B).

Paleoproterozoic granitoid sample WC09-15-14(4a) (Fig. 4C) yields 34 concordant 207Pb/206Pb ages that range from ca. 1669 Ma to ca. 2497 Ma (Fig. 6A). The largest population of 10 concordant analyses is clustered at ca. 2300 Ma with a weighted mean age of 2383 ± 11 Ma. A second population of seven concordant zircon ages yields a weighted mean age of 2170 ± 13 Ma, which we interpret as the crystallization age of this granitoid (Fig. 6A). In addition, a younger population of three concordant ages is clustered at ca. 1900 Ma. We interpret that the ca. 1900 Ma zircons are metamorphic and the ca. 2300 Ma zircon cores are inherited based on evidence from Th/U ratios and CL images (Figs. 5 and 6A).

Thirty-five (35) zircons were analyzed from the early Neoproterozoic granitoid sample WC09-16-14(3) (Fig. 4D) yielding 15 concordant ages that range from ca. 918 Ma to ca. 1317 Ma (Fig. 6B). The largest population of eight concordant analyses in the zircon CL rims are clustered at ca. 960 Ma with a weighted mean age of 965 ± 39 Ma, which we interpret as the crystallization age of the granitoid (Fig. 6B). In addition, an older concordant age population from zircon grain cores is clustered at 1306 ± 210 Ma, which is interpreted to be from inherited xenocrystic zircons (Fig. 6B). The discordant ages in this sample generally possessed zircon Th/U ratio <0.1 (Figs. 5 and 6B), consistent with a metamorphic signature.

Early Ordovician granitoid sample WC09-17-14(2b) (Fig. 4E) yields 32 concordant ages that range from ca. 432 Ma to ca. 1797 Ma (207Pb/206Pb) (Fig. 6C). The largest and youngest population of eight concordant analyses is clustered at ca. 500 Ma with a weighted mean age of 470 ± 13 Ma, which we interpret as the crystallization age of the granitoid (Fig. 6C). In addition, two older concordant inherited zircon age populations yield weighted mean ages of 1277 ± 250 Ma and 1764 ± 74 Ma (Fig. 6C).

Early Ordovician volcanic sample WC09-17-14(2c) (Fig. 4E) yields 32 concordant ages that range from ca. 452 Ma to ca. 2305 Ma (Fig. 6D). The youngest population of four concordant analyses is clustered at ca. 470 Ma with a weighted mean age of 478 ± 31 Ma, which we interpret as the crystallization age of the sample (Fig. 6D). Two older concordant zircon age populations yield weighted mean ages of 849 ± 49 Ma and 1431 ± 80 Ma (Fig. 6D). The largest population of concordant analyses is clustered at ca. 1800 Ma with a weighted mean age of 1801 ± 35 Ma, representing ages of inherited zircon grains (Fig. 6D). The sample also yields minor populations of slightly discordant ages clustered at ca. 1931 and ca. 2305 Ma (Fig. 6D).

Late Carboniferous K-feldspar granite sample WC09-17-14(4a) (Fig. 4F) yields concordant ages that range from ca. 294 Ma to ca. 334 Ma (Fig. 6E). The weighted mean age of 34 concordant analyses is 315 ± 3 Ma, which we interpret as the crystallization age of the K-feldspar granite (Fig. 6E). Late Carboniferous granitoid sample WC09-17-14(4b) (Fig. 4F) yields 35 concordant ages that range from ca. 289 Ma to ca. 352 Ma (Fig. 6F). The weighted mean age of 34 concordant analyses is 309 ± 3 Ma, which we interpret as the crystallization age of the granitoid (Fig. 6F).

Weakly metamorphosed, coarse-grained sandstone sample WC09-16-14(14) of the Neoproterozoic Hanmushan Group was collected from the lower part of a moraine conglomerate layer (Fig. 4A). A total of 38 detrital zircon grains were analyzed, yielding concordant ages that range from ca. 998 Ma to ca. 2661 Ma (Fig. 6G). This sample has three major zircon populations of ca. 1484–1608 Ma, ca. 1806–2116 Ma, and ca. 2489–2661 Ma (Fig. 6G). A total of 100 detrital zircon grains were analyzed from Permian sandstone sample WC09-17-14(3) (Fig. 4B). Five grains from this sample yield discordant ages (Fig. 6H). Ninety-one (91) concordant ages from this sample range from ca. 269 Ma to ca. 446 Ma with one major zircon population of ca. 269–432 Ma (Fig. 6H). The weighted mean of the three youngest concordant zircon grains was 270 ± 17 Ma (MSWD [mean square weighted deviation] = 0.035), and we interpret this age to represent the maximum depositional age of this sandstone sample.

Whole-Rock Major and Trace Element and Sr-Nd Isotope Geochemistry

Seven granitoid samples were analyzed for major and trace element (Table S3, footnote 1) and Sr-Nd isotopic concentrations (Table S4) to determine their composition and generation setting. Sample preparation and analyses were performed at the State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences in Wuhan. Analytical details are described in Liu et al. (2004).

All granitoid samples are granite and quartz monzonite in composition with high SiO2 (64.84–79.55 wt%) (Fig. 7A) and low MgO (0.13–1.15 wt%), falling within the shoshonitic to tholeiitic fields (Fig. 7B). The samples are metaluminous to peraluminous and have aluminum saturation indices [i.e., A/CNK = molecular Al2O3/(CaO + Na2O + K2O)] of 1.26–2.25 (Table S3, footnote 1). All samples are variably enriched in light rare earth elements. Early Neoproterozoic and early Paleozoic granitoid samples show weak positive Eu anomalies, whereas other samples display weak to moderate negative Eu anomalies (Fig. 7C). All samples are enriched in large ion lithophile elements (i.e., Rb and U) and relatively depleted in high field strength elements (i.e., Nb and Ta) and elements reflecting plagioclase fractionation (i.e., Sr) (Fig. 7D).

Paleoproterozoic granitoid samples are characterized by high initial 87Sr/86Sr of 0.7160–0.9013 and high εNd(t) of −0.29 to 17.26 (with T2DM model ages of 1.76–2.70 Ga, DM is depleted mantle), whereas the Ordovician granitoid samples have initial 87Sr/86Sr of 0.7075–0.7261 and lower εNd(t) of −18.67 to −4.57 (with T2DM model ages of 1.55–2.71 Ga) (Fig. 7E; Table S4, footnote 1). Early Carboniferous granite samples also have relatively high initial 87Sr/86Sr of 0.7077–0.7081 and low εNd(t) of −16.81 to −15.82 (with T2DM model ages of ca. 2.37–2.45 Ga) compared to Silurian granitoid samples with initial 87Sr/86Sr of 0.6394–0.7080 and εNd(t) of −9.79 to −3.73 (with T2DM model ages of 1.29–1.54 Ga) (Fig. 7E; Table S4).

Metasedimentary rock samples of the Paleoproterozoic Longshoushan Group yield three prominent zircon age populations with peaks at ca. 1.85 Ga, ca. 2.05 Ga, and ca. 2.15 Ga and a minor older zircon age population at ca. 2.56 Ga (Fig. 3B). The overlying Mesoproterozoic Dunzigou Group has two major zircon age populations with peaks at ca. 1.82 Ga and ca. 2.05 Ga, in addition to two minor age populations at ca. 2.31 Ga and ca. 2.60 Ga (Fig. 3B). The youngest zircon age population of ca. 1.82–1.85 Ga and corresponding Th/U ratios from the samples of the Longshoushan and Dunzigou Groups are interpreted to reflect a regional metamorphic event, which has been identified across the North China craton (e.g., Zhang and Gong, 2018; Gong et al., 2012, 2016; Wu et al., 2021a). The ca. 2.05–2.15 Ga and ca. 2.31 Ga age peaks from the metasedimentary rock samples reflect local detrital sources of the Paleoproterozoic magmatic arc of the Longshou Shan and North China craton (e.g., Zhao et al., 2005; Zhang and Gong, 2018; Gong et al., 2012, 2016; Song et al., 2017; Kusky et al., 2016; Wu et al., 2018, 2021a). The Archean zircon U-Pb age peak and Lu-Hf isotopes are interpreted to reflect reworking of older crust of the North China craton, which suggests an Archean basement source in the region (Duan et al., 2015; Gong et al., 2016; Qin, 2012; Xue et al., 2017; Zhang et al., 2017) (Fig. S1, footnote 1). The ca. 2.17 Ga arc granitoid sample WC09-15-14(4a) that is thrust over the Longshoushan Group also shows a major inherited zircon age cluster at ca. 2.38 Ga (Fig. 6A). Evidence for the development of the Paleoproterozoic magmatic arc and a regional metamorphic event and the detrital zircon age population of Paleo- to Mesoproterozoic metasedimentary samples of the Longshou Shan have been reported for the northern part of the North China craton (e.g., Kusky et al., 2016; Liu et al., 2020a; Wu et al., 2018, 2021a; Zhang and Gong, 2018) (Figs. 3B and 3C). Although Archean basement is not reported in the Longshou Shan, the Lu-Hf-Sr-Nd isotopes of amphibolite-facies Longshoushan Group show evidence of reworking of ca. 2.4–2.7 Ga basement of the North China craton (Fig. 7E; Fig. S1) (e.g., Gong et al., 2016; Xue et al., 2017; Liu, 2008; this study). These results imply that the Longshou Shan basement rocks were once part of the North China craton and involved in the northern North China orogeny (e.g., Kusky et al., 2016; Wu et al., 2018, 2021a). Furthermore, evidence for ca. 2.2–2.0 Ga magmatism is relatively rare on Earth and may be useful for global Paleoproterozoic reconstruction of continental connections.

The detrital zircon U-Pb ages of Longshou Shan strata from Song et al. (2017) and sample WC09-16-14(14) of this study from the Neoproterozoic Hanmushan Group show two prominent zircon age populations with peaks at ca. 1.15 Ga and ca. 1.40 Ga and three minor age populations with peaks at ca. 1.0 Ga, 1.95 Ga, and 2.5 Ga (Fig. 3B). Proterozoic rocks with similar U-Pb age spectra have been reported in northern Tibet and the northern part of the North China craton, specifically the Hujiertu Formation of the Bayan Obo Group (e.g., Wu et al., 2017, 2021a; Zuza et al., 2018; Zhou et al., 2018; Liu et al., 2020a). In addition, the ca. 965 Ma age of granitoid sample WC09-16-14(3) correlates with prominent ages in northern Tibet and the Tarim continent, which implies that a Neoproterozoic magmatic arc may have stretched from the Longshou Shan region through the Qilian Shan to the Tarim continent (e.g., Gehrels et al., 2003; Zuza and Yin, 2017; Zuza et al., 2018; Wu et al., 2017, 2021a) (Fig. 8A). The ca. 0.9–1.0 Ga magmatic arc in northern Tibet and the Longshou Shan region is likely associated with the northward subduction of Paleo-Qilian oceanic lithosphere between the Kunlun-Qaidam and North China cratons (e.g., Wu et al., 2016, 2017; Zuza et al., 2018) (Fig. 8A). This correlative Neoproterozoic arc system would have been constructed on the southern margin of the unified North China–North Tarim craton. We infer this subduction system also led to the final amalgamation between South Tarim and North Tarim continents (e.g., Guo et al., 2005; Xu et al., 2013; Zuza and Yin, 2017; Zhao et al., 2021), and this long-active margin is important for reconstructions of Neoproterozoic plate configurations (e.g., Zuza and Yin, 2017).

Closure of the Paleo-Qilian Ocean is marked by the presence of ca. 830–600 Ma bimodal volcanic rocks and ca. 825–775 Ma post-collisional, extension-related granitoids throughout northern Tibet, the Longshou Shan, and North China craton (e.g., Wu et al., 2016, 2017; Zuza et al., 2018) (Fig. 8B). In addition, oceanic island basalt–type mafic volcanic rocks are interbedded with Neoproterozoic limestone and/or marble, possibly indicating that a Neoproterozoic rift-to-drift basin developed in northern Tibet and the Longshou Shan region during the openings of the Paleo-Asian and Qilian Oceans from north to south (e.g., Wu et al., 2021a). After the Neoproterozoic collision, passive margins of the Qaidam–Qilian Shan were likely adjacent to the North China craton, based on the presence of Neoproterozoic fossils in northern Tibet that match those found in ca. 550 Ma rocks of the North China craton (Pang et al., 2021).

Metasedimentary samples of the Cambrian strata in the Longshou Shan have zircon age peaks at ca. 0.59 Ga, 0.82 Ga, 1.00 Ga, and 2.50 Ga and a maximum depositional age of ca. 536 Ma (Liu et al., 2020b) (Fig. 3B). The three younger age populations are widely distributed in the Longshou Shan and northern margin of northern Tibet, whereas the older Archean age peak is widely distributed in the North China craton (e.g., Liu et al., 2020b; Liu et al., 2020a; Wu et al., 2021a). These ages combined with Cambrian passive-margin strata in the Longshou Shan (Fig. 3A) may suggest that the initiation of northward Qilian oceanic subduction did not begin by Cambrian time (Fig. 8B). The oldest ophiolites of Qilian oceanic lithosphere exposed in northern Tibet are ca. 550 Ma (e.g., Shi et al., 2004). Evidence for ca. 517–483 Ma arc magmatism and the exposure of a Cambrian–Ordovician arc sequence in northern Tibet reflect the timing of initiation of southward subduction of Qilian oceanic lithosphere (e.g., Xiao et al., 2009; Song et al., 2013; Zuza et al., 2018; Li et al., 2021), whereas the ca. 470 Ma granitoid sample WC09-17-14(2b) and ca. 478 Ma arc volcanic sample WC09-17-14(2c) from the Longshou Shan suggest northward subduction of Qilian oceanic lithosphere (Fig. 8B). Molasse strata of the Devonian Laojunshan Group unconformably overlie the Cambrian passive- margin strata and yield detrital zircon age populations that reflect a similar provenance to that of sedimentary rocks of northern Tibet (Zhang et al., 2017; Li et al., 2020; Wu et al., 2021a). The presence of Silurian–Devonian syn- to post-collisional granitoids (i.e., ca. 440–400 Ma) in the Longshou Shan reflects that the final closure of the bivergently subducting Qilian Ocean occurred at ca. 440 Ma (Wang et al., 2020; Zhang et al., 2017; Tang, 2015; Duan et al., 2015; Li et al., 2021) (Fig. 8B).

The early Paleozoic Qilian orogeny led to regional uplift and terrestrial clastic sedimentation in the Longshou Shan during the Silurian–Devonian (Zhang and Gong, 2018; Wu et al., 2021a) (Fig. 3A). By the early Carboniferous, high topography related to the Qilian orogen subsided, and carbonate rocks were deposited across the region (Fig. 3A). By the late Carboniferous, initial subduction of Paleo-Asian oceanic lithosphere led to uplift, erosion, and formation of an unconformity between the early Carboniferous and late Carboniferous–early Permian strata (Fig. 3A). Late Carboniferous (ca. 309–315 Ma) arc plutons observed in the Longshou Shan are interpreted to be related to this subduction event (Fig. 8C). Ca. 330 Ma adakitic rocks in the Longshou Shan may have been sourced from melted oceanic lithosphere or deep crustal rocks during the early phases of subduction (e.g., Xue et al., 2017). By Triassic time, the Longshou Shan region was above sea level, based on the widespread presence of terrestrial strata. Given the regional unconformities observed in this study and evidence for the closure of the Paleo-Asian Ocean in the Permian (e.g., Xiao et al., 2003; Eizenhöfer et al., 2014; Fu et al., 2018), we interpret that an orogen developed in the Longshou Shan region during the Carboniferous–Triassic, which explains the presence of Permian and younger terrestrial strata. To test this, La/Yb ratios (Table S3, footnote 1) of Longshou Shan granitoid samples were normalized to chondritic reservoir values of McDonough and Sun (1995), converted to paleo–crustal thickness using the method of Sundell et al. (2021), and plotted versus their crystallization ages (Fig. 7F). The results show an overall thickening trend during the Ordovician–Carboniferous, which is supported by the presence of Phanerozoic unconformities (Fig. 3A).

The detrital zircon ages of the Permian sandstone sample WC09-17-14(3) show a dominant unimodal Permian–Carboniferous age population with a ca. 330 Ma peak, which suggests that the strata were proximal to the source arc system (Cawood et al., 2012). In addition, early Permian (ca. 280 Ma) mafic dike swarms are widespread between the southern portion of the Central Asian orogenic system and north of the Qilian orogen (i.e., Dan et al., 2014; Zhang et al., 2017) and were likely emplaced in an extensional tectonic setting in the region. Permian backarc extension possibly perturbed and influenced by the Tarim large igneous provinces (Xu et al., 2013, 2021). Normal faults exposed in the Longshou Shan region are likely related to this extensional event (Fig. 4B). Subsequent Triassic terrestrial sedimentation may have been related to collision and closure of the Paleo-Asian Ocean along the Solonker suture (e.g., Fu et al., 2018) (Figs. 3A and 8C). Thus, the tectonic evolution of the Longshou Shan is mostly reflective of the generation and subduction of Qilian and Paleo-Asian oceanic lithosphere during the Paleozoic (Fig. 8C). Generally, the Longshou Shan is correlative to the North China craton, which contains stronger cratonic material that resisted deformation. In contrast, the Qilian Shan to the south is composed of mélange arc-suture zone material that was readily deformed.

Paleoproterozoic deformation, amphibolite-facies metamorphism, and magmatism in the Longshou Shan suggest that the region is the western extension of the northern part of the North China craton. The three Neoproterozoic and Paleozoic orogenic events recorded in the Longshou Shan correlate in space and time with those that occurred in the Tarim craton to the west and the North China craton to the northeast. Specifically, Neoproterozoic deformation and magmatism link to the west with those that occurred in the North Tarim craton (e.g., Gehrels et al., 2003). Early Paleozoic deformation links with that recorded in the Tarim craton to the west and the Qinling orogen to the southeast (e.g., Yin and Nie, 1996). Lastly, the spatial extent of late Carboniferous–Triassic deformation associated with the subduction and closure of Paleo-Asian oceanic lithosphere can be extended along the margins of the Tarim craton to the west and the North China craton to the east (e.g., Xiao et al., 2003; Eizenhöfer et al., 2014; Zuza and Yin, 2017; Fu et al., 2018).

Geologic and geochronological evidence shows that the northern limit of the Cenozoic Himalayan-Tibetan orogen and Tibetan Plateau has persisted since initial India-Asia collision at ca. 58 Ma directly south of the Longshou Shan (Fig. 1A) (e.g., Yin et al., 2008; Clark et al., 2010; Zuza et al., 2020; Li et al., 2021). Our interpretation that the Tarim continent links through the Longshou Shan to the North China craton suggests that this continuous cratonic continental strip may have acted as an important backstop for Cenozoic orogeny. Although the Longshou Shan experienced deformation related to Neoproterozoic, early Paleozoic, Mesozoic, and Cenozoic events, as outlined above, the main collisional or suturing events appeared to have occurred south or north of the range (Fig. 8). Therefore, this work supports the interpretation that Cenozoic collision-related strain occurred most significantly in zones of preexisting collisional belts and suture zones, with shortening and crustal thickening focused in the mechanically weak, low-viscosity extent of the Tethyan orogenic system (e.g., An et al., 2020; Bian et al., 2020; Chen et al., 2020). The relatively weak Tethyan orogenic system may have been locally strengthened by Permian mantle plume activity (Xu et al., 2021; Liu et al., 2021). Interestingly, Cenozoic reactivation of the Central Asian orogenic system to the north, across the Tarim–North China cratonic strip, involves different styles of deformation, including more strike-slip and normal faulting (e.g., Webb and Johnson, 2006; Yin, 2010). Thus, understanding the pre-Cenozoic tectonic evolution of central Asia is critical to understanding initial conditions that influenced the kinematics and dynamics of Cenozoic collision-induced tectonism.

We appreciate Science Editor Andrea Hampel, Associate Editors Jeff Lee and Francesco Mazzarini, and three reviewers, Dr. Jianxin Zhang, Dr. Andrew K. Laskowski, and Dr. Anas Abbassi, for their critical, careful, and very constructive reviews that have helped improve the clarity and interpretations of the original draft. This research was supported by grants from the Basic Science Center for Tibetan Plateau Earth System (CTPES, grant 41988101-01), the Second Tibetan Plateau Scientific Expedition and Research Program (grant 2019QZKK0708), National Natural Science Foundation of China (grants 42072001 and 41702232), the National Key Research and Development Project of China (grant 2016YFC0600303), and the China Geological Survey (grant DD20160083).

1Supplemental Material. Table S1: Summary of geochronology results of magmatic intrusions in the Longshou Shan area, western China. Table S2: LA-MC-ICP-MS results for zircon U-Pb ages of igneous and sedimentary samples. Table S3: Major and trace elements for the igneous samples from the Longshou Shan, western China. Table S4: Summary of Sr-Nd isotopes of intrusions in the Longshou Shan, western China. Figure S1: Plots of zircon εHf(t) versus 206Pb/238U age for the Longshou Shan area, western China. Data are from Duan et al. (2015), Gong et al. (2016), Qin (2012), Xue et al. (2017), and Zhang et al. (2017). Please visit https://doi.org/10.1130/GEOS.S.19285499 to access the supplemental material, and contact editing@geosociety.org with any questions.
Science Editor: Andrea Hampel
Associate Editor: Francesco Mazzarini
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.