Geochemistry and detrital zircon geochronology of the Bagong Formation clastic rocks in the Queerchaka area, central part of the Qiangtang Basin: evidence for the Late Triassic chemical weathering and tectonic evolution

Late Triassic was not only a key episode for the evolution of the Qiangtang Basin but also a critical interval marking the final closure of the Paleo-Tethys Ocean on the Qinghai–Tibet Plateau. However, the nature of the Late Triassic Qiangtang Basin and its response to the evolution of the Paleo-Tethys Ocean remain unclear. Here, elemental geochemistry and detrital zircon U–Pb dating data of clastic rocks from the Upper Triassic Bagong Formation in the Queerchaka area (central part of the Qiangtang Basin) were integrated to infer the chemical weathering history, provenance, and tectonic evolution during the Late Triassic. The Bagong Formation may be assigned to the Carnian–Norian of the Late Triassic. Chemical Index of Alteration (CIA) values of the studied samples, which were unaffected by sedimentary recycling and K-metasomatism, reveal a moderate chemical weathering (CIA = 68.3–75.8). Multiple geochemical proxies and zircon U–Pb age distributions suggest a felsic provenance for the Bagong Formation in the Queerchaka area. This provenance was primarily derived from the complex rocks of the Central Uplift Belt, with a minor contribution from the Hoh Xil–Jinshajiang complex rocks. The Late Triassic Northern Qiangtang sub-basin may be a back-arc compound foreland basin that linked to a double subduction and collisional orogenic process involving the northward subduction of the Longmuco–Shuanghu Paleo-Tethys branch Ocean and the southward subduction of the Hoh Xil–Jinshajiang Paleo-Tethys Ocean. These findings provide a foundation for deciphering the nature and dynamic processes of the Late Triassic Qiangtang Basin.

The chemical (including major and trace elements) and mineralogical compositions of clastic rocks are largely governed by source-rock properties, weathering intensity, transport processes, deposition conditions, and diagenesis (McLennan et al., 1993). With improvements in analytical techniques, zircon U–Pb geochronology and geochemistry of clastic rocks are now commonly employed for tracing sedimentary provenance and distinguishing tectonic settings during deposition (Xu et al., 2007; Ma et al., 2017; Wang et al., 2023; Shi et al., 2026), which offering greater reliability than conventional methods based solely on detrital grain composition of sandstones. The elemental geochemistry of clastic rocks can be used to infer chemical weathering intensity, provenance, and tectonic setting (Verma and Armstrong Altrin, 2013; Moradi et al., 2016; Wang et al., 2018), whereas detrital zircon U–Pb ages not only constraint the maximum deposition age but also illuminate source-to-sink system during the evolution of orogenic belts and associated sedimentary basins (Xu et al., 2007; Cawood et al., 2012; Ding et al., 2013; Fan et al., 2015; Sharman and Malkowski, 2020). Sedimentary basins developed under different tectonic settings exhibited markedly distinct signatures in provenance. Therefore, provenance tracing of clastic sediments can effectively reconstruct the nature and evolutionary history of these basins (Xu et al., 2007; Cawood et al., 2012; Wang et al., 2023; Li et al., 2025).

The Qiangtang Basin, situated along the central–northern margin of the Tibetan Plateau, belongs to the eastern Tethyan tectonic domain that contains the world’s most prolific oil and gas resources (Tan et al., 2002; Wang et al., 2004; 2020; Wu et al., 2014). It is a marine sedimentary basin with the lowest degree of onshore exploration, the greatest oil and gas resource potential, and the highest likelihood of achieving a breakthrough in exploration in China (Tan et al., 2002; Wang et al., 2004; Wu et al., 2014). In recent years, with the ongoing advancement of hydrocarbon exploration in the Qiangtang Basin, outcrops investigations and the deployment of geological survey wells have revealed widely distributed argillaceous source rocks and favourable oil-gas shows in the Upper Triassic strata (Wang et al., 2022a). These findings have provided important geological data for the assessment of hydrocarbon resources in this basin. Moreover, the Late Triassic was a vital period in the evolution of the Qiangtang Basin, during which the basin experienced multiple tectonic events, including the extension of the Bangong–Nujiang Meso-Tethys Ocean in its south and the demise of both the Longmuco–Shuanghu Paleo-Tethys Ocean in its center and the Hoh Xil–Jinshajiang Paleo-Tethys Ocean in its north (Gehrels et al., 2011; Yan et al., 2016; Wang et al., 2022b). As a result of these complex evolutionary histories, the nature of the Qiangtang Basin in the Late Triassic remains contentious, and a variety of different views have been developed, including a back-arc basin (Li et al., 2001), a fore-arc basin (Yin and Harrison, 2000), a foreland basin (Wang et al., 2001; 2004), a peripheral foreland basin (Li et al., 2003), a quasicraton basin (Gao et al., 2006), and an epicontinental sea basin (Chen and Wang, 2009).

The Late Triassic strata are extensively developed in the Qiangtang Basin (Zhan and Tan, 2020; Liu et al., 2022a; Zhan et al., 2025). Recently, Zhan et al. (2025) systematically summarized the lithological associations, sedimentary sequences, depositional environment, contact relation, fossil, and chronostratigraphy of the Upper Triassic based on the previous research and latest drilling data. The Jiapila, Bolila, and Bagong formations represent the lower part of the Upper Triassic strata in the Southern and Northern Qiangtang Depressions and the Xiaoqiebao, Jiaomuchaka, and Zana formations represent the Upper Triassic lithostratigraphic framework in the Central Uplift Belt, which may be attributed to the Carnian–Norian age (Zhan et al., 2025). In addition, the Nadi Kangri Formation represents the Late Triassic Norian-Rhaetian volcanic-sedimentary sequence in the Qiangtang Basin, which overlies the lower part of the Upper Triassic strata either conformably or with angular unconformity (Zhan et al., 2025). These Upper Triassic successions serve as the best carrier for deciphering the Late Triassic history of continental convergence–breakup in the eastern Tethyan domain, as well as for constraining the nature of the Qiangtang Basin itself. Prior researches on the provenance and tectonic affinity of the Upper Triassic clastic rocks have primarily focused on the Tumengela in the southern part of the Northern Qiangtang Depression (Feng et al., 2010; Wang et al., 2017) and Zangxiahe Formations in its northern part (Song et al., 2013). Such studies have predominantly been based on the whole-rock geochemical compositions of clastic rocks and detrital grain of sandstone (Feng et al., 2010; Song et al., 2013; Wang et al., 2017), whereas detrital zircon U–Pb geochronological data remain comparatively limited. Given that sedimentary strata faithfully document basin-filling processes, determining the provenance and tectonic setting of the Upper Triassic Bagong Formation in the central part of the Qiangtang Basin is essential for reconstructing the evolutionary history of the Longmuco–Shuanghu Paleo-Tethys Ocean and for deciphering the nature of the Qiangtang Basin during the Late Triassic.

In this study, we present systematic geochemical (major and trace elements) and detrital zircon U–Pb geochronological data for the Upper Triassic Bagong Formation clastic rocks in the Queerchaka area. The primary aims were to (1) constrain the depositional age of the Bagong Formation in the Queerchaka area; (2) reconstruct the chemical weathering history during deposition; (3) determine the provenance and tectonic settings; and (4) elucidate the tectonic evolution during the Late Triassic. These findings could serve as a foundation for constraining the nature of the Late Triassic Qiangtang Basin and for understanding its response to the closure of the Longmuco–Shuanghu and Hoh Xil–Jinshajiang Paleo-Tethys oceans.

In the Queerchaka area of the central Qiangtang Basin, the exposed thickness of the Bagong Formation exceeds 1,000 m, although neither its top nor its bottom is exposed (Figure 1C). A total of 24 clastic rock samples were collected sequentially from bottom to top along the section, including 1 sandstone sample and 23 mudstone samples. The detailed sampling positions are shown in Figure 1C. Of these, all 23 mudstones were subjected to major- and trace-element analysis, and one sandstone sample was used for detrital zircon geochronology.

Major element concentrations were determined using the melting method on a ZSX primus III XRF spectrometer at the laboratory of the Chengdu Center China Geological Survey (Geosciences Innovation Center of Southwest China). The instrument was operated at 70 mA and 40 kV, and the detailed analytical procedures followed GB/T 14506.28-2010. Data quality was monitored through duplicate analyses (one duplicate per eight samples) and the national standard GBW07107, with analytical precision better than 5%. Loss on ignition (LOI) was determined by placing 1 g of dried powder in a corundum crucible and heating it in a muffle furnace at 1,000 °C for 90 min. After cooling, the weight loss after heating was then recorded. Trace element contents were measured by acid digestion on an Agilent 7850 ICP–MS at the laboratory of the Chengdu Center China Geological Survey (Geosciences Innovation Center of Southwest China), following GB/T 14506.30-2010, with analytical precision also better than 5%.

For detrital zircons U–Pb dating, zircon grains were first separated from the sandstone sample using magnetic and heavy-liquid separation techniques. Approximately 200 grains were randomly selected, then mounted in epoxy resin and polished to expose grain interiors for subsequent analysis. Cathodoluminescence (CL) images were acquired using a TESCAN field emission scanning electron microscope to reveal the internal textures of individual zircon grains prior to U–Pb isotope dating. U–Pb zircon dating was performed on a laser ablation inductively coupled plasma-mass spectrometer (LA–ICP–MS) instrument at Wuhan SampleSolution Analytical Technology Co., Ltd (Wuhan, China), whose GeolasPro ablation system is equipped with a MicroLas optical assembly and a COMPexPro 102 ArF 193-nm excimer laser. The ICP–MS instrument is Agilent 7,900. The laser was operated at a repetition rate of 5 Hz with a spot size of 32 µm. Zircon standard 91500 was used as an external reference material for calibration of isotopic fractionation (Hu et al., 2021). Data processing was carried out using ICPMS DataCal software, and weighted mean age calculation as well as concordia diagrams were generated using Isoplot/Ex version 3. In this study, the available zircon U–Pb isotopic ages have a concordance of >90% and an analytical error (1σ) of <3.0%.

Elemental geochemistry of the Bagong Formation clastic rocks

Major oxides

Major element concentrations of the Bagong Formation mudstone samples from the Queerchaka section are presented in Supplementary Table S1. Among these samples, SiO 2 and Al 2 O 3 are the most abundant components, ranging from 49.3% to 61.2% (avg. 55.5%) and from 14.4% to 20.5% (avg. 18.3%), respectively (Supplementary Table S1), followed by TFe 2 O 3, whose contents varying from 5.72% to 7.85% (avg. 6.91%). The alkali metal oxide K 2 O is markedly more enriched than Na 2 O, with concentrations of 2.34%–4.71% (avg. 3.42%) and 0.86%–1.99% (avg. 1.28%), respectively (Supplementary Table S1). The K 2 O/Na 2 O ratios of all studied samples range from 1.79 to 3.57 (avg. 2.74), significantly exceeding 1, which is probably attributable to the weathering of feldspar into clay minerals. For most of the studied samples, the abundance of the alkaline earth metal oxide MgO is slightly higher than that of CaO, with abundances of 2.14%–4.25% (avg. 2.82%) and 0.45%–6.10% (avg. 1.69%), respectively (Supplementary Table S1). The concentrations of TiO 2, MnO, and P 2 O 5 are all below 1.0% (Supplementary Table S1). When normalized to the UCC (upper continental crust), the major element compositions of the studied samples show that SiO 2 and Na 2 O are depleted in all samples, whereas Al 2 O 3, TFe 2 O 3, K 2 O, MgO, TiO 2, MnO, and P 2 O 5 are slightly enriched, and CaO shows a notable depletion in most samples (Figure 2A).

A strong positive relationship is observed between K 2 O and Al 2 O 3 contents (R 2 = 0.647; Figure 3A), indicating that K 2 O is mainly derived from clay minerals rather than potassium feldspars. In contrast, no significant relationship is observed between Al 2 O 3 +K 2 O and SiO 2 contents (R 2 = 0.237; Figure 3B), suggesting that most of the Si in the samples originates from quartz rather than clay minerals. The mudstone samples show a pronounced positive correlation between CaO and MgO (R 2 = 0.926; Figure 3C), and the MgO + CaO contents are also strongly positively correlated with LOI (R 2 = 0.684; Figure 3D), suggesting that the majority of CaO and MgO are hosted in carbonate minerals. Furthermore, TFe 2 O 3 exhibits a notable positive relationship with MgO (R 2 = 0.483; Figure 3E), implying that part of the iron may be present in ankerite. Finally, a marked positive correlation is observed between TiO 2 and Al 2 O 3 +K 2 O (R 2 = 0.456; Figure 3F), indicating that Ti is mainly associated with clay minerals, possibly in the form of heavy minerals.

Trace elements and rare earth elements (REE)

The trace element and REE data of the Bagong Formation mudstones from the Queerchaka section are reported in Supplementary Table S2. Among the large-ion lithophile elements (LILE), Rb, Cs, Sr, Ba, and Pb concentrations range from 74.6 to 168 ppm (avg. 112 ppm), 6.95–20.3 ppm (avg. 10.8 ppm), 78.7–222 ppm (avg. 100 ppm), 344–695 ppm (avg. 456 ppm), and 11.3–76.5 ppm (avg. 27.8 ppm, excluding sample QECK-10), respectively (Supplementary Table S2). Relative to the UCC, Cs shows slight enrichment, whereas Sr and Ba are depleted in all samples; Rb and Pb are enriched in most samples (Figure 2B). Among them, a significant positive correlation exists between Cs and Al 2 O 3 +K 2 O contents (R 2 = 0.4448; Figure 3G), indicating that Cs is mainly hosted in clay minerals. For the high-field-strength elements (HFSE), Ta, Nb, Hf, Zr, Th, and U concentrations range from 1.08 to 3.59 ppm (avg. 1.70 ppm), 12.8–17.9 ppm (avg. 15.1 ppm), 3.76–5.28 ppm (avg. 4.57 ppm), 46.3–171 ppm (avg. 142 ppm), 9.53–14.2 ppm (avg. 11.5 ppm), and 1.85–2.75 ppm (avg. 2.35 ppm), respectively (Supplementary Table S2). The UCC-normalized spider diagrams reveal that Zr, Hf, and U are slightly depleted, Nb and Ta are slightly enriched, whereas Th exhibits both depletion and enrichment (Figure 2B). The obvious positive relationship between Zr and Hf contents (R 2 = 0.988, excluding sample QECK-23; Figure 3H) reflects their similar geochemical behavior. In addition, both Nb and U display marked positive correlations with Al 2 O 3 +K 2 O (R 2 = 0.4225 and 0.5463; Figures 3I,J), indicating that these elements are primarily associated with clay minerals. Among the transition elements, V, Sc, Cr, Co., and Ni have concentrations of 107–174 ppm (avg. 147 ppm), 14.4–18.8 ppm (avg. 16.8 ppm), 73.3–106 ppm (avg. 95.4 ppm), 17.9–43.3 ppm (avg. 26.1 ppm), and 34.0–53.5 ppm (avg. 45.3 ppm), respectively (Supplementary Table S2). Compared with the UCC, Sc, V, and Co. are slightly enriched, and Cr is enriched in most samples, whereas Ni exhibits both depletion and enrichment (Figure 2B). Furthermore, Sc, V, Cr, and Ni all show significant positive correlations with Al 2 O 3 +K 2 O (R 2 = 0.7176, 0.8916, 0.6959, and 0.6038; Figures 3K–N), implying that these elements are also hosted in aluminosilicate phases such as clay minerals.

The total REE (ΣREE) abundances of the Bagong Formation mudstones from the Queerchaka section vary widely, ranging from 144 to 203 ppm (avg. 174 ppm), with most samples showing values slightly lower than that of the Post-Archean Australian Shale (PAAS; 185 ppm; Taylor and McLennan, 1985) (Supplementary Table S2). The mudstones exhibit significantly higher LREE contents (129–184, avg. 157) than HREE contents (14.1–19.0, avg. 16.9), yielding LREE/HREE ratios of 8.57–10.4 (Supplementary Table S2) and suggesting a marked fractionation between LREE and HREE. Chondrite-normalized REE patterns for the studied mudstones, together with those of PAAS and UCC, are shown in Figure 4. The REE distribution curves of the Bagong Formation mudstones are similar to that of PAAS, characterized by LREE enrichment and relatively flat HREE patterns. The Chondrite-normalized (La/Yb) N ratio values range from 7.98 to 9.96 (avg. 8.93; Supplementary Table S2), further confirming the pronounced fractionation between LREE and HREE. In addition, all mudstone samples display weak negative Ce anomalies and pronounced negative Eu anomalies, with δEu ranges from 0.63 to 0.68 (avg. 0.64) and δCe from 0.90 to 1.04 (avg. 0.97) (Supplementary Table S2). No correlation is observed between δEu and δCe (Figure 3O), suggesting that the δEu values are only weakly influenced by diagenesis (Shields and Stille, 2001).

Detrital zircon U–Pb age of the Bagong Formation sandstone

The lithology is gray lithic quartz sandstone, with quartz content of ca. 60%, lithic fragment content of ca. 25%, and feldspar content of ca. 20% (Figures 5A–1). The matrix content is relatively low and mainly consisting of micritic calcite. Quartz grains are poorly sorted and rounded and minor volcanic lithic fragments are observed (Figures 5A–1). No obvious alteration is observed under the microscope (Figures 5A–1). A total of 71 zircon grains of the Bagong Formation sandstone sample (T2) in the Queerchaka area were randomly analyzed, of which 67 zircon grains yielded available ages (concordance of >90% and 1σ analytical error <3.0%). The detrital zircon U–Pb age data are listed in Supplementary Table S3. CL images (Figures 5A–2) reveal that most zircon grains are subhedral to subrounded in shape, with grain sizes of approximately 60–150 μm. The majority of zircon grains exhibit clear oscillatory zoning, whereas a few grains show inherited cores and narrow bright rims (Figures 5A–2). The Th and U contents and Th/U ratios of the 67 dated grains are 121–1,696 ppm, 143–4295 ppm, and 0.13–2.16 (avg. 0.68), respectively (Supplementary Table S3). Most detrital zircons have Th/U ratios higher than 0.3 (60/67; Figure 5B; Supplementary Table S3). Together with the well-developed oscillatory zoning, this indicates that the majority of the detrital zircons are of igneous origin, with a minor contribution from metamorphic or hydrothermal sources (Wu and Zheng, 2004). The U–Pb ages of the 67 zircon grains span from the Mesoarchean to the Late Triassic, ranging between 3,185 and 231 Ma (Supplementary Table S3; Figure 5C). Among them, the ages are mainly concentrated in the Paleoproterozoic (2487–1746 Ma, n = 27) and Neoproterozoic (921–587 Ma, n = 16) (Supplementary Table S3). Other age populations are less abundant, including the Triassic (248–231 Ma, n = 6), Permian (277–258 Ma, n = 3), Devonian (419–410 Ma, n = 3), Silurian (440–431 Ma, n = 4), Ordovician (484–446 Ma, n = 3), Cambrian (515–490 Ma, n = 2), as well as the Neoarchean (2558–2554 Ma, n = 2) and Mesoarchean (3,185 Ma, n = 1) (Supplementary Table S3). The five youngest Late Triassic zircon grains yield ages ranging from 236 Ma to 231 Ma, giving a weighted average age of 233.9 ± 2.4 Ma (MSWD = 0.42) (Figure 5D).

Depositional age of the Bagong Formation

The Bagong Formation predominantly occurs in the eastern part of the Northern Qiangtang Depression, while it exhibits conformable contacts with the underlying Bolila Formation and the overlying Eerlongba Formation (Zeng et al., 2021). However, the depositional age of the Bagong Formation remains highly controversial. Previous studies have identified several fossil assemblages from the lower part of the Bagong Formation in the Quemoco–Maiduo area, including ammonite fossil assemblages (Paratibetites cf. Wheeleri and Nodoti-betites cf. Nodosus) (Niu et al., 2003), bivalve fossil assemblages (Halobia convexa, Halobia yunnanensis, Halobia superbescens, and Halobia diserseinsecta) (Tang et al., 2007), and sporopollen fossil assemblages (Annulispora sp., Chasmatosporites sp., Klausipollenites sp., Ovalipollis ovalis, Ovalipollisbreviformis, Micrhystridium sp., Limatulasporites sp., Taeniaesporites sp., and Pinuspollenites) (Yu et al., 2019). Additionally, the upper part of the Bagong Formation has yielded bivalve fossils including Amonotis togtonheensis h, A. rothpletzi yushuensis, Cardium (Tulongocardium) xizhangensis, Cardium (Tulongocardium) martini, Trigonia (Kumatrigonia) huhxilensis, Bakevellia pannonica and Costatoria goldfussi (Tang et al., 2007). Among them, bivalve fossils such as H. convexa and H. yunnanensis are considered Carnian species, whereas A. rothpletzi yushuensis, C. goldfussi and B. pannonica are Carnian–Norian bivalve species. Bivalve fossils, including Amonotis togtonheensis - Cardium (Tulongocardium) xizhangensis, Cardium (Tulongocardium) martini - Trigonia (Kumatrigonia) huhxilensis, and H. superbescens - H. diserseinsecta, belong to the Early–Middle Norian assemblages. Moreover, previous researchers identified an approximately 6‰ negative excursion in both δ 13 C org and δ 13 C carb in the lower part of the Bagong Formation in the Quemoco area (Fu et al., 2020). This excursion is comparable to the carbon isotope anomalies observed in typical sections recording the global Carnian Pluvial Episode, suggesting that it might be linked to the global Carnian Pluvial Episode recorded in the Qiangtang Basin (Fu et al., 2020).

The maximum depositional age (MDA) determined from detrital zircon U–Pb ages has become a complementary method for constraining the age of sedimentary strata (Dickinson and Gehrels, 2009; Sharman and Malkowski, 2020). Multiple methods are available for calculating the MDA, such as Youngest Single Grain age (YSG), Youngest Graphical Peak age (YPP), Youngest Gaussian Fit age (YGF), and comparing the results from these different methods allows for a more robust constrains and helps ensure reliable interpretations (Sundell et al., 2024). The youngest zircon populations from the upper part of the Bagong Formation in the Well QZ-16 and from its top in the Quemoco area are 229–223 Ma (n = 19) and 227–217 Ma (n = 9), diving weighted average ages of 227.2 ± 1.2 Ma (n = 19; Zeng, 2021) and 222.8 ± 1.9 Ma (n = 9; Li and Wang, 2018), respectively. These results constrain the maximum depositional age of the upper Bagong Formation to the Carnian–Norian of the Late Triassic (Li and Wang, 2018; Zeng, 2021). In addition, the youngest zircon age group from the lower part of the Bagong Formation in the Queerchaka area ranges from 236 to 231 Ma (n = 5), with MDAs of 234.0–231.2 Ma determined by nine different calculation methods using DZmda software (Table 1; Sundell et al., 2024). This age constrains the maximum depositional age of the lower Bagong Formation to the Carnian (Late Triassic). We integrate the biostratigraphic constrains, maximum depositional ages, and carbon isotope excursions of the Bagong Formation in the region and infer that the Bagong Formation was most likely deposited during the Carnian–Norian interval.

Intensity of chemical weathering

Chemical composition and chemical weathering

The essence of chemical weathering lies in the transformation of feldspars in the source rock into clay minerals. This process is typically accompanied by the dissolution and mobilization of mobile elements such as K +, Ca 2+, and Na +, whereas relatively immobile elements including Ti 4+ and Al 3+ remain in the residue, resulting in systematic changes in bulk chemical composition (Nesbitt and Young, 1982). Based on this principle, several indices have been developed to quantify the degree of chemical weathering, including the Chemical Index of Alteration (CIA = [Al 2 O 3 /(Al 2 O 3 + CaO * + K 2 O + Na 2 O)] × 100) (Nesbitt and Young, 1982), Plagioclase Index of Alteration (PIA = [(Al 2 O 3 − K 2 O)/(Al 2 O 3 + CaO * + Na 2 O − K 2 O)] × 100) (Fedo et al., 1995), and Chemical Index of Weathering (CIW = [Al 2 O 3 /(Al 2 O 3 + CaO * + Na 2 O)] × 100) (Cullers, 2000). All oxides are expressed in molar proportions, and CaO* refers exclusively to the CaO fraction derived from silicate minerals only (Nesbitt and Young, 1982). To date, there is no direct method for distinguishing this silicate-bound CaO from that hosted in non-silicate minerals (i.e., carbonates and phosphates). Instead, the indirect correction method proposed by McLennan et al. (1993) is commonly adopted, where CaO * = min (CaO − 10/3P 2 O 5, Na 2 O) in molar units. In practice, however, these indices are subject to certain limitations. For example, the PIA is only applicable when the source rock contains plagioclase but no K-feldspar (Xu and Shao, 2018). The CIW fails to exclude Al contributed by K-feldspar, whereas the CIA is subject to relatively fewer interfering factors and has consequently become the most widely used index (Xu and Shao, 2018). However, CIA values can be affected by sedimentary differentiation, sedimentary recycling, post-depositional weathering during transport, and potassium metasomatism during diagenesis. Only after these interfering factors have been evaluated and excluded can the CIA values accurately assess the intensity of chemical weathering (Fedo et al., 1995; Roy and Roser, 2013; Wang et al., 2023; Xu and Shao, 2018). Moreover, owing to their relatively high compositional homogeneity and low permeability, argillaceous rocks preserve provenance information more faithfully and are thus preferred over other clastic rocks for assessing chemical weathering intensity (Xu and Shao, 2018).

Influence of sedimentary differentiation (grain size), sedimentary recycling, and K-metasomatism on CIA values

During transport and deposition, weathering products of source rocks are subject to sedimentary differentiation, which tends to reduce the clay mineral content in sandstones and thus yields CIA values lower than the actual values (Garzanti et al., 2009; Roy and Roser, 2013). The Al 2 O 3 typically represents the fine-grained fraction because it is extremely immobile and mainly resides in clay minerals and feldspar weathering residues, while SiO 2 generally represents the coarse-grained fraction under strong hydrodynamic conditions. Accordingly, the SiO 2 /Al 2 O 3 ratio can serve as a proxy of grain size (Ohta, 2004). In the present study, the SiO 2 /Al 2 O 3 ratios of the Bagong Formation mudstone samples range between 2.63 and 3.55 (avg. 3.06). Moreover, no correlation is observed between SiO 2 /Al 2 O 3 ratios and CIA values for the studied samples (R 2 = 0.005, n = 23; Figure 6A), suggesting that the CIA values were essentially unaffected by sedimentary differentiation and grain size variation.

Sediments that have undergone sedimentary recycling exhibit altered chemical composition and increased compositional maturity due to secondary weathering, which can result in elevated CIA values that no longer faithfully reflect the weathering intensity of the source area (Cox et al., 1995). The index of compositional variation (ICV) is the most widely used proxy for identifying sedimentary recycling (Cullers, 2000; Xu and Shao, 2018). The ICV proxy is calculated as (K 2 O + CaO* + Na 2 O + TFe 2 O 3 + MgO + MnO + TiO 2)/Al 2 O 3, with all oxide concentrations are expressed in molar units. As recycling progresses, clay mineral content increases while non-clay minerals decrease, and because Al 2 O 3 is significantly enriched in clay minerals relative to non-clay minerals (Cox et al., 1995; Cullers, 2000). Sediments with low compositional maturity typically have ICV >1, implying a first-cycle origin in tectonically active settings, whereas those with high maturity yield ICV <1, indicating recycling or intense chemical weathering of first-cycle material (Kamp and Leake, 1985). Accordingly, it has been suggested that samples with ICV >1 should be preferentially used when assessing weathering intensity, so as to minimize the influence of sedimentary recycling on CIA values (Xu and Shao, 2018). The ICV values of the Bagong Formation mudstones vary between 0.83 and 1.46 (avg. 1.04), with only a few samples exhibiting ICV values slightly below 1.0 (Supplementary Table S1; Figure 6B). This range points to a compositionally immature provenance and a first-cycle origin in an active tectonic regime for the Bagong Formation mudstones. In addition, sedimentary recycling can also be assessed using trace elements associated with heavy minerals (e.g., element Zr in zircon), which tends to be enriched during recycling (McLennan et al., 1993). Specifically, the Zr/Sc and Th/Sc ratios of clastic rocks, together with their binary diagram, provide a useful method for assessing compositional variations of source area and sedimentary recycling (McLennan et al., 1993; Wang et al., 2017; 2023). The Zr/Sc and Th/Sc ratios of the Bagong Formation mudstone are 2.95–11.9 (avg. 8.47; Supplementary Table S2) and 0.54–0.99 (avg. 0.69; Supplementary Table S2), respectively. As shown in the Zr/Sc–Th/Sc cross-plot (Figure 6C), all data points plot along the trend line of compositional variation rather than the trend line of sedimentary recycling, indicating that the sediments experienced no significant recycling during deposition. Furthermore, the CIA values show no correlation with either ICV values (R 2 = 0.046; Figure 6B) or Zr/Sc ratios (R 2 = 0.111; Figures 6C–1), confirming that the CIA values were essentially unaffected by sedimentary recycling. Therefore, the CIA values of the Bagong Formation mudstones from the Queerchaka section can reliably be used to constrain the intensity of chemical weathering.

Although potassium content in source rocks theoretically decrease with increasing weathering intensity, they may be elevated by K-metasomatism during diagenesis, thus making the CIA values lower than their true levels (Nesbitt and Young, 1982; Xu and Shao, 2018). The Al 2 O 3 (A)–CaO * +Na 2 O (CN)–K 2 O (K) ternary plot serves not only to evaluate whether sediments have been affected by K-metasomatism but also to correct for such alteration (K-correction) (Nesbitt and Young, 1982; Fedo et al., 1995). In the A–CN–K ternary diagram, the solid lines nearly parallel to the A–CN side represent the ideal weathering trend lines of various source rocks compositions. Deviations of data points from these ideal trends are generally attributed to K-metasomatism, with greater deviations indicating stronger K-enrichment (Yan et al., 2010; Wang et al., 2023; Xu and Shao, 2018). For the Bagong Formation mudstones from the Queerchaka section, all CIA values fall along the ideal weathering trend (Figure 6D), indicating that these samples were not significantly affected by K-metasomatism. Therefore, their CIA ratios can reliably reflect the degree of weathering in the source area.

CIA proxy for chemical weathering

As demonstrated above, the CIA values of the Bagong Formation mudstones in this study were not significantly affected by K-metasomatism, sedimentary recycling, or sedimentary differentiation. Hence, the CIA data from these samples can be used with confidence as reliable proxies for evaluating chemical weathering intensity. According to the conventional classification scheme (Feng et al., 2003), CIA values < 50 indicate an unweathered source area; 50–70 reveal a weak weathering; 70–80 point to a moderate weathering; and >80 reflect a intense weathering. The CIA values of the Bagong Formation mudstones from the Queerchaka section range between 68.3 and 75.8, with most samples falling between 70 and 80 (Figure 6D; Supplementary Table S1), indicating that their source area experienced moderate chemical weathering.

Provenance of the Bagong Formation

Evidence from elemental geochemistry

Owing to the poor aqueous solubility of their hydroxides and oxides, Al, Ti, and Zr do not readily migrate with surface water during weathering and are therefore regarded as stable, immobile elements (Moradi et al., 2016). Consequently, their contents and ratios in clastic rocks generally resemble those of their source rocks, making them widely used tracers for provenance, like the TiO 2 /Zr and Al 2 O 3 /TiO 2 ratios (Hayashi et al., 1997; Moradi et al., 2016; Wang et al., 2018). In details, low Al 2 O 3 /TiO 2 ratios varying between 3 and 8 reveal a mafic provenance, intermediate Al 2 O 3 /TiO 2 ratios varying from 8 to 21 reflect an intermediate source, and high Al 2 O 3 /TiO 2 ratios fluctuating from 21 to 70 are typically linked to a felsic provenance (Hayashi et al., 1997). The Bagong Formation mudstones from the Queerchaka area exhibit Al 2 O 3 /TiO 2 ratios of 20.0–28.1 (avg. 24.0), with all but one sample (QECK-1) exceeding 21, revealing a felsic provenance during deposition (Figure 7A). Additionally, the TiO 2 /Zr ratio increases progressively from felsic to mafic provenance. Specifically, felsic provenance yields low TiO 2 /Zr ratios less than 55, intermediate provenance shows moderate ratios varying between 55 and 200, and mafic provenance displays ratios greater than 200 (Hayashi et al., 1997). The TiO 2 /Zr ratios of the studied samples (excluding QECK-23) range from 48.1 to 62.4 (avg. 53.2), with 9 samples above 55 and the remainder below this threshold, suggesting a dominantly felsic source with minor input from intermediate compositions (Figure 7B). In the A–CN–K ternary plot, the weathering trends of the studied samples fall between the granodiorite and granite field (Figure 6D), further supporting a predominantly felsic provenance.

Moreover, some specific trace elements (including La, Cr, Th, Co., Sc, and Hf) are geochemically stable in clastic rocks and serve as effective indicators for provenance determination (Armstrong-Altrin et al., 2004; Xu et al., 2007). Generally, La, Th, and Hf are enriched in felsic rocks and their weathering products, whereas Cr, Co., and Sc are more abundant in mafic rocks and their altered products (Armstrong-Altrin et al., 2004; Cheng et al., 2024). Among the various discriminant diagrams, the La/Th–Hf (Floyd and Leveridge, 1987), Sc/Th–Cr/Th (Condie and Wronkiewicz, 1990), and La/Sch–Co/T (Condie, 1993) diagrams are widely used to distinguish provenance. The Bagong Formation mudstones in this study exhibit relatively low and concentrated La/Th ratios (2.69–3.63, avg. 3.23; Supplementary Table S2) and high Hf contents (3.76–5.28, avg. 4.57; Supplementary Table S2). On the La/Th–Hf plot, most studied samples plot within the field of felsic source, with only two samples falling into the field of mixed felsic/basic source-rock (Figure 7C). Additionally, these mudstone samples are characterized by low La/Sc and Co./Th ratios, ranging from 1.33 to 3.78 (avg. 2.31; Supplementary Table S2) and 1.78 to 2.65 (avg. 2.21; Supplementary Table S2), respectively. On the La/Sc–Co/Th plot, all samples are distributed between the andesite and felsic volcanic rock fields, revealing a predominantly felsic provenance (Figure 7D). The studied samples also exhibit relatively high Cr/Th ratios (6.37–10.5, avg. 8.40) and relatively low Sc/Th ratios (1.01–1.86, avg. 1.48). On the Sc/Th–Cr/Th cross-plot, all data points fall between the felsic volcanic rock and andesite fields, close to the granodiorite field, further supporting a mainly felsic provenance (Figure 7E). Moreover, in the Zr/Sc–Th/Sc plot, the studied samples align with the compositional variation trend and plot between the granite and andesite fields (Figure 6C), also indicating a dominantly felsic provenance.

Additionally, REEs are also commonly used for provenance tracing, because felsic source rocks generally exhibit significant high LREE/HREE ratios and pronounced negative Eu anomalies, while mafic source rocks tend to show low LREE/HREE ratios and no or weak Eu anomalies (Moradi et al., 2016; Wang et al., 2017). As described in section 3, the δEu values of the Bagong Formation mudstones in this study are not significantly affected by diagenesis; rather, they reflect the signature of the source rocks. The chondrite-normalized REE patterns display pronounced negative Eu anomalies, with δEu ranging between 0.63 and 0.68 (avg. 0.64; Supplementary Table S2; Figure 4), indicative of a felsic provenance. Furthermore, on the La/Yb–ΣREE diagram, all studied samples fall within the field of granite, further supporting a felsic source (Figure 7F).

Evidence from zircon U–Pb dating

U–Pb age spectra of detrital zircons from terrigenous clastics offer critical constraints on sedimentary provenance (Gehrels et al., 2011; Ding et al., 2013; Ma et al., 2017). The zircon grains of the Bagong Formation sandstone from the Queerchaka area are predominantly subhedral to rounded, exhibit clear oscillatory zoning and have high Th/U ratios (Figures 5A–2), suggesting a magmatic origin and long-distance transport (Wu and Zheng, 2004). The Queerchaka area, located in the southern part of the Northern Qiangtang Depression, is bordered to the south by the Longmuco–Shuanghu orogenic belt (Central Uplift Belt) and to the north by the Hoh Xil–Jinshajiang orogenic belt, both of which formed during the Late Triassic (Figure 1A). During this interval, the Southern Qiangtang terrane was characterized by carbonate platform deposits of the Riganpeicuo Formation (Wang et al., 2012; Hou et al., 2014), while the Bangong–Nujiang Ocean in its south was in an extensional stage (Fan et al., 2017). This paleogeographic pattern would have hindered the supply of detritus from the southern terranes in the southern of the Central Uplift Belt (including South Qiangtang–Lhasa–Himalaya terranes) to the Northern Qiangtang Depression. Therefore, we compared the detrital zircon age spectra of the Bagong Formation with pre-Late Triassic U–Pb ages from surrounding rocks (clastic, igneous, and metamorphic) of potential source areas—namely, the Central Uplift (Longmuco–Shuanghu orogenic belt), the Northern Qiangtang terrane, and the Hoh Xil–Jinshajiang orogenic belt (Figure 8). This comparison aims to better constrain the sedimentary provenance and to reconstruct the Late Triassic source-to-sink system for the Queerchaka area.

The detrital zircon age distribution of sandstone sample from the Bagong Formation exhibits three major peaks (at ca. 1894, 744, and 431 Ma) and two subordinate peaks (at ca. 2458 and 243 Ma) (Figure 5D). Among them, 19 Paleoproterozoic (1981–1746 Ma) and 9 Neoarchean–Paleoproterozoic zircons (2558–2396 Ma) constituted one major peak and one subordinate peak, respectively (Figure 5D). Zircon ages in the 2558–1746 Ma range show diverse morphologies, from euhedral to subhedral columnar and rounded grains, with some showing clear oscillatory zoning and others lacking it. Their corresponding peak ages at ca. 1895 and 2458 Ma correlate well with zircon age peaks from the Southern Qiangtang terrane and the Hoh Xil–Jinshajiang complex rocks (Figure 8). In addition, 15 Neoproterozoic zircons (921–720 Ma) yield a major peak at ca. 744 Ma (Figure 5D), which broadly matches zircon age peaks from the Hoh Xil–Jinshajiang complex rocks, Central Uplift Belt complex rocks, and Southern Qiangtang terrane (Figure 8). 12 Early Paleozoic zircons (515–410 Ma) show a major peak at ca. 431 Ma (Figure 5D), coinciding with peak ages from the Hoh Xil–Jinshajiang complex rocks, the Central Uplift Belt complex rocks, and the Southern Qiangtang terrane (Figure 8). Three Permian zircons (277–258 Ma) display euhedral columnar morphology, high Th/U ratios, and clear oscillatory zoning, which are interpreted as the proximal transport of the Permian volcanic rocks from the Central Uplift Belt (Zhai et al., 2004; Liang et al., 2020). Five Late Triassic zircons (236–231 Ma) also exhibit significant oscillatory zoning and euhedral crystals (Figures 5A–2), suggesting that they were likely sourced from short-distance transport of the Late Triassic (235–221 Ma) igneous rocks in the Shanzixingshan area (Ma et al., 2016; Lu et al., 2019). In summary, we propose that the provenance of the Upper Triassic Bagong Formation in the Queerchaka area (southern part of the Northern Qiangtang Depression) was dominated by the Central Uplift Belt complex rocks, with a possible minor provenance contribution from the Hoh Xil–Jinshajiang belt (Figure 8).

Tectonic setting

The tectonic setting of source rocks is widely considered to be a primary control on the chemical component of clastic rocks (Bhatia, 2004), with sediments from different tectonic settings exhibiting distinct geochemical signatures (Bhatia, 2004). Recently, Verma and Armstrong-Altrin (2013) proposed two multidimensional discrimination diagrams based on statistical analysis of major oxide data from Cenozoic sediments. Unlike traditional tectonic setting discrimination diagrams for source areas (e.g., Th–Sc–Zr/10 and La–Th–Sc ternary plots; Bhatia and Crook, 1986), these new diagrams can effectively distinguish rift, collision, and island-arc settings during deposition and have therefore been widely adopted (Moradi et al., 2016; Wang et al., 2023). In their scheme, clastic rocks are subdivided into high-silica ((SiO 2) a dⱼ = 63–95%) and low-silica ((SiO 2) a dⱼ = 35–63%) types, where “adj' denotes oxide contents recalculated to 100% after exclusion of LOI (more details seen in Verma and Armstrong-Altrin, 2013). In this study, we applied these two multidimensional discrimination diagrams to determine the tectonic setting during deposition of the Bagong Formation mudstones from the Queerchaka area. The (SiO 2) adj concentrations of the 23 mudstone samples range from 57.3% to 65.3% (Supplementary Table S1), with 20 samples classified as low-silica and 3 as high-silica. On both the low-silica and high-silica discrimination plots, nearly all data points (except QECK-8) fall within the collision field (continent–continent collision) (Figures 9A,B), indicating that the Bagong Formation mudstones from the Queerchaka area were deposited in a collisional tectonic setting.

Furthermore, the cumulative distribution of the age differences between detrital zircon crystallization and deposition can provide independent constraints on tectonic settings (Cawood et al., 2012). This method is primarily based on the principle that detrital zircon age spectra effectively reflect the tectonic setting of sedimentary basins (Cawood et al., 2012). Convergent plate margins are typically characterized by a high proportion of zircon with ages close to the depositional age, collisional settings show a broader age range, mixing near-syndepositional zircons with those derived from ancient basement, and extensional settings are marked by zircons significantly older than the depositional age, indicating a lack of contemporaneous magmatism (Cawood et al., 2012). The detrital zircon age distribution of the Bagong Formation sandstone in the Queerchaka area exhibits multiple age peaks, which differ notably from the depositional age of the strata (Figure 5D). The youngest 5% of the detrital zircon crystallization ages (CAs) show a difference of less than 150 Ma from the depositional age (DA), indicating a convergent or collisional compressional setting in the source area (Figure 10A) (Cawood et al., 2012). Furthermore, the youngest 30% of the CAs exhibit a difference significantly greater than 100 Ma from the DA, further suggesting that the source area of the Upper Triassic Bagong Formation in the central part of the Qiangtang Basin was likely in a collisional compressional setting (Figure 10A) (Cawood et al., 2012). Meanwhile, the cumulative proportion curve of detrital zircon ages from the Queerchaka section falls almost entirely within the collisional-setting field (Figure 10B), which further corroborates this interpretation. This collisional setting is consistent with the inferences drawn from the geochemical discrimination diagrams.

Implication for the Late Triassic tectonic evolution

Integrating the analyses of depositional age, chemical weathering, provenance, and tectonic setting, a tectonic evolution model during the Late Triassic along the central part of the Qiangtang Basin is established in Figure 11.

The Devonian, Permian, and Middle Triassic radiolarian faunas discovered in the Hoh Xil area indicate that this area was dominated by marine environments prior to the Middle Triassic (Wang et al., 2005). Moreover, Early–Middle Triassic igneous rocks with island arc affinity have been successively identified in the eastern part of the Qiangtang Basin (Roger et al., 2003; Chen et al., 2016) and are interpreted as a magmatic arc formed by southward subduction of the Jinshajiang Paleo-Tethys Ocean beneath the Northern Qiangtang terrane (Wang et al., 2011; Ding et al., 2013; Zhang et al., 2014). Previous studies have also identified Middle Triassic syn-collisional granites in the Feimatan area, adjacent to the Jinshajiang suture zone (240 Ma; Li et al., 2003), suggesting that collision along the western margin of the Jinshajiang suture zone may have occurred prior to the Ladinian of the Middle Triassic. Meanwhile, continuous Triassic strata, including the Kanglu, Yingshuiquan, Kangnan, and Tumengela Formations, were deposited along the southern margin of the Northern Qiangtang Depression, exhibiting two fining-upward then coarsening-upward sedimentary cycles (Zhu et al., 2013; Wang et al., 2004; Sui et al., 2026). The provenance of the Middle–Lower Triassic sediments points to a continental island arc setting (Liang et al., 2020), which has been interpreted as a sedimentary response to the northward subduction of the Longmuco–Shuanghu Paleo-Tethys Ocean beneath the Northern Qiangtang terrane during the Early Triassic. Recent data show that basalts, rhyolites, and granites formed in an intraplate extensional tectonic setting during the late Middle Triassic to the early Late Triassic (242–235 Ma) are sporadically exposed from west to east in the central part of the Qiangtang Basin. These rocks are considered to be products of asthenosphere upwelling and partial melting associated with slab breakoff of the Longmuco–Shuanghu oceanic lithosphere following the collision between the Northern and Southern Qiangtang terranes (Zhang et al., 2011; Ma et al., 2016). Paleomagnetic data suggest that the collision between the Northern and Southern Qiangtang was completed by the early Late Triassic, resulting in a unified Qiangtang terrane (Song et al., 2012), during which the Longmuco–Shuanghu Ocean closed and a composite orogenic system developed, which further supported by the presence of Late Triassic syn-collisional S-type granites (Huang et al., 2007; Zheng et al., 2015).

Furthermore, the high to ultrahigh pressure metamorphic rocks along the Longmuco–Shuanghu suture zone yield metamorphic peak ages concentrated at ca. 240 Ma, which are thought to record the collisional orogenic event between the Northern and Southern Qiangtang terranes (Zhang et al., 2006; Zhai et al., 2011). Regionally distributed paleo-weathering crusts at the top of the coal-bearing strata of the Upper Triassic Tumengela Formation (a contemporaneous heterotopic facies of the Bagong Formation), together with the continental volcanic–sedimentary sequences of the Nadigangri Formation in the Northern Qiangtang Depression, reveal that the Triassic Qiangtang Basin gradually shrank and was uplifted to land (Fu et al., 2009; Wang et al., 2008). Combined magmatic, metamorphic, and sedimentary evidence suggests that the collision and amalgamation of the Kunlun, Southern Qiangtang, and Northern Qiangtang terranes were completed by the early Late Triassic, accompanied by the closure of both the Longmuco–Shuanghu and Hoh Xil–Jinshajiang oceans (Zhang et al., 2011; Peng et al., 2015; Wang et al., 2022b). In this study, we have demonstrated that the Bagong Formation clastic rocks in the Queerchaka area were deposited in a collisional tectonic setting (Figures 9, 10), consistent with the Late Triassic collisional compressional regime of the Northern Qiangtang Depression (Li et al., 2003; Song et al., 2012; Wang et al., 2017; 2018; Liu et al., 2022b; Shen et al., 2024), and their provenance was mainly derived from erosion of the Central Uplift Belt complex rocks (Figure 8). On basis of the previously published data and our new results, we conclude that the Northern Qiangtang Depression exhibited the characteristics of a back-arc compound foreland basin during the Late Triassic, driven by double subduction and collision orogeny involving the southward subduction of the Hoh Xil–Jinshajiang Ocean and northward subduction of the Longmuco–Shuanghu Ocean (Figure 11).

We report new elemental geochemical and zircon U–Pb geochronological data for the Upper Triassic Bagong Formation clastic rocks in the Queerchaka area (central part of the Qiangtang Basin), with the aim of constraining their chemical weathering intensity and tectonic evolution. The main conclusions are summarized as follows: The Bagong Formation is inferred to be of Carnian–Norian in age (Late Triassic) age, on the basis of the regional biostratigraphic constraints, maximum depositional ages, and carbon isotope excursions.

The CIA values of the Bagong Formation mudstone from the Queerchaka area were not significantly affected by sedimentary differentiation, K-metasomatism, or sedimentary recycling. The CIA values of the studied mudstone samples are 68.3–75.8, with most samples falling between 70 and 80, suggesting a moderate degree of chemical weathering in the source area.

Multiple geochemical proxies, including the Al 2 O 3 –TiO 2, TiO 2 –Zr, La/Sc–Co/Th, Sc/Th–Cr/Th, La/Th–Hf, La/Yb–∑REE, and A–CN–K plots, together with obvious negative Eu anomalies (0.63–0.68, avg. 0.64), consistently indicate that the provenance of the Upper Triassic Bagong Formation in the Queerchaka area was primarily felsic rocks. Detrital zircon U–Pb geochronology of the sandstone further reveals that the sediments supply for the Bagong Formation were mainly derived from the Central Uplift Belt, with a minor contribution from the Hoh Xil–Jinshajiang belt.

Both major-element multidimensional discrimination plots and detrital zircon age spectra suggest that the Bagong Formation in the Queerchaka area was deposited in a collisional tectonic setting. In combination with previously published data, we conclude that the Late Triassic Northern Qiangtang Depression exhibited the characteristics of a back-arc compound foreland basin, which was driven by a double subduction and collisional orogenic system involving northward subduction of the Longmuco–Shuanghu Paleo-Tethys Ocean and southward subduction of the Jinshajiang Paleo-Tethys Ocean.