5.1 Provenance of the Triassic zircons: location and composition
The detrital zircon U–Pb ages for the eight samples from the Middle–Upper Triassic Xiaoquangou Group in SJB are categorized into six populations displayed in Fig. 5. Previous studies on paleocurrents show that the potential source areas lie south of the Tianshan Mountains (Li et al. 2012; Yang et al. 2013; Tang et al. 2014; Zhu et al. 2017; Liu et al. 2013, 2017; Wang et al. 2018b, 2019b). Therefore, a comparison of the detrital zircon U–Pb age distribution to the reported crystallization ages of intermediate–acidic igneous rocks from the NTS, CTS, Bogda Mountains, and STS–Tarim, enables determination of the provenance of Cambrian to Permian detrital zircons (Fig. 5; Wang et al. 2018b, 2019b). However, constraining the provenance of Triassic zircons is difficult because reports suggest limited or no Triassic magmatism in the Junggar Basin and adjacent areas. Greene et al. (2005) reported a granitoid pluton zircon U–Pb age of 229 ± 11 Ma for the Barkol Tagh (Eastern Tianshan). Yin et al. (2018) reported a biotite 40Ar–39Ar age of 239.5 ± 1.3 Ma for the CTS, while zircon U–Pb age yielded an age of 281.9 ± 1.2 Ma for the same sample. The BGMRXUAR (1993) reported thin amygdaloidal andesites interlayered with the Middle Triassic Kelamayi Formation in the Jimusaer area of Bogda Mountains, without providing precise ages. That is, direct field evidence of Triassic magmatic activity in the area is inexistent.
Constraining the provenance of Triassic zircons was in addition to dating the main reason for collecting the samples from west to east in different sections of SJB (Fig. 2). The detrital zircon U–Pb dating results reveal dominance of Triassic zircons in the Haojiagou and Xidalongkou sections (northern Bogda Mountains), with subordinate populations in the Well C, and Dabancheng and Yuergou sections (southern Bogda Mountains), and no or minor amounts in other regions like the NTS and CTS areas (e.g. the Sikeshu, Jiangongmeikuang, and Aiweiergou sections; Fig. 5). The proportions of the Triassic zircons decrease steadily from the Well C, and Haojiagou and Xidalongkou sections to the Dabancheng and Yuergou sections, and further in the Aiweiergou and Jiangongmeikuang sections (Figs. 2 and 5).The Triassic detrital zircons occur mainly in the Middle–Upper Triassic Xiaoquangou Group of west Bogda Mountains, with absence in the NTS and CTS areas, suggesting the Triassic zircons probably come from the Bogda Mountains, restricted to the Urumqi–Jimusaer region, and probably in the Haojiagou–Well C–Xidalongkou areas.
Moreover, some other evidences support that the Triassic zircons are probably source from the Bogda Mountains. (1) Paleocurrent observations show complicated drainage system with both northward- and southward-directed paleocurrents (Fig. 8). The paleocurrents are principally perpendicular to the present day Bogda Mountains in the western Bogda Mountains, while in the eastern Bogda Mountains the paleocurrents mainly northward-directed (Fig. 8). This proved that the western Bogda Mountains probably have uplifted and began to provide sources. Instead, the eastern Bogda Mountains (e.g., Zaobishan region) remain accept as a sink region with the sources from NTS and CTS. (2) The shapes of lithic fragments and Triassic zircon grains are angular to sub-angular (Fig. 7; Wang et al., 2018b), maybe indicate a proximal source area. The sedimentary lithic fragments occur in the Middle–Upper Triassic samples and increase upward with decrease in feldspar grains (Wang et al., 2018b), suggest less unroofing of granitoids and more unroofing of clastic rocks. These phenomena indicate that the Middle–Late Triassic provenance are probably come from the nearby Bogda Mountains which provide syndepositional and recycled sediments.
In addition, some problems worth discussing. (1) About the location of source area: We attribute the Triassic zircons to the Bogda Mountains which are likely interlayered in the Middle–Upper Triassic Xiaoquangou Group as suggested by BGMRXUAR (1993). Greene et al. (2005) reported the Triassic granitoid plutons in the Barkol Tagh region (Eastern Tianshan), further confirmed the Triassic magmatic activity exist in the Tianshan region. Theoretically, both locations could be the provenance of Triassic zircons, but petrological characters do not support Barkol Tagh region because it is too far to provide sources to the western Bogda Mountains (Fig. 3). Instead, the Balikun area maybe accepted deposits from the Barkol Tagh as shown by the paleogeographic map (Fig. 3). (2) About the composition of provenance: The trace element compositions of the detrital zircons are sensitive to source rock type (Hoskin and Ireland, 2000; Belousova et al. 2002; Corfu et al. 2003; Yang et al. 2012; Wang and Li, 2017). The variations of representative elements and ratios of REEs for the zircons indicate that the Triassic zircons originated from intermediate–acidic igneous rocks, with probable dominance of intermediate igneous rocks (Fig. 9). This inference is further supported by the REEs abundance and chondrite–normalized REE patterns (Fig. 6). The Triassic zircons show close REEs concentrations and chondrite–normalized REE patterns to the Cambrian–Permian zircons, suggesting they were derived from similar source rock types (Fig. 6; Hoskin and Ireland 2000; Belousova et al. 2002). Previous researches attest that the Cambrian–Pennsylvanian and Permian zircons were likely sourced from the intermediate–acidic igneous rocks of the NTS and CTS (Wang et al. 2018b, 2019a, 2019b). Collectively, the Triassic zircons are of magmatic origin and credibly originate from the intermediate–acidic igneous rocks. (3) About the explanation of previous work: Tang et al. (2014) published detrital zircon spectrum for sample TS12-99, which mark as Lower Triassic strata in the map but introduce as Middle–Upper Triassic Xiaoquangou Group in the text. Actually, it should be Lower Triassic strata after communicated with the author. Fang et al. (2015) think their published sample Hao-01 is collected from the Middle–Upper Triassic Xiaoquangou Group, but the latitude and longitude (N43°39.128′, E87°13.229′) of this sample shows that it is collected from the Lower Jurassic Badaowan Formation. The relative percentage of Triassic zircon ages shows dramatic decrease from Xiaoquangou Group (Sample 17HJG-01) to Badaowan Formation (Sample Hao-01), probably revealing sources from Bogda Mountains have been eroded. We propose this is a slight uplift accompany by rapid denudation process during the Middle–Late Triassic.
5.2 Source-to-sink system of SJB in the Middle–Late Triassic
The detrital zircon U–Pb age probability and the Multidimensional scaling (MDS) plots were used in conjunction to establish the Middle–Late Triassic source-to-sink system of SJB (Figs. 5 and 10; Vermeesch, 2018). In the MDS plot, the thirteen samples from the Middle–Upper Triassic Xiaoquangou Group in SJB show distinct differences that generate four groups (Fig. 10).
Samples 15C-33, 16DLK-49 and 17HJG-01 (group 1) from the Well C, Xidalongkou and Haojiagou sections differ significantly from others due to high contents of Triassic zircons (Figs. 5 and 10), indicating a predominant contribution from the Bogda Mountains. Moreover, the samples from the Well C, Xidalongkou and Haojiagou sections contain subordinate the Carboniferous and Cambrian–Devonian zircons revealing that the NTS and CTS also provided sediments to the northern Bogda Mountains (Fig. 5). That is, the provenance of the Well C, Xidalongkou and Haojiagou sections are primarily attributed to the Bogda Mountains, with secondary contribution from the NTS and CTS.
Samples XJ12-06, 16TDG-06, and 17DDLK-09 (group 2) from the Aiweiergou, Taodonggou and Dongdalong sections are assigned similar sediment sources owing to their close links in the MDS plot (Fig. 10). The Mississippian and the Cambrian–Devonian magmatic rocks from the CTS are likely major sources for group 2, considering the relation of their spectra (Fig. 5). Group 2 also displays minor Pennsylvanian zircon grains in the detrital zircon U–Pb age probability plot that are likely related to the NTS sources.
Samples 12SKS-07, 17JGMK-06, and 16DBC-04 (group 3) from the Sikeshu, Jiangongmeikuang, and Dabancheng sections generally display a limited age variation dominated by the Pennsylvanian (Fig. 5), and show proximity in the MDS plot (Fig. 10). The Pennsylvanian age distribution combined with the compositionally and texturally immature detrital grains, probably originate from the magmatic rocks of NTS with zircon U–Pb ages of 320–300 Ma (Fig. 5). Furthermore, a little Cambrian–Devonian and Mississippian age groups in the probability plot also suggest contribution to group 3 from the CTS. Particularly, the probability plot of detrital zircon ages for the Dabancheng section (sample 16DBC-04) shows an age group spanning 250–200 Ma, reflecting further contribution from the Bogda Mountains to the Dabancheng section.
Samples 15HXG-95, 17JJZG-10, 17BYH-61, and HX-01 (group 4) from the Haxionggou, Jingjingzigou, Baiyanghe and Yuergou sections show a mixture age variation of Pennsylvanian (320–300 Ma) and Mississippian (360–320 Ma), and have proximity in the MDS plot (Fig. 10). The two major peaks with age group spanning 320–300 Ma and 360–320 Ma, revealing the major provenance of group 4 is from NTS and CTS. Besides, the presence of some Triassic zircons in sample HX-01 from the Yuergou section suggests possible provenance from the Bogda Mountains.
The systematic analysis of the detrital zircons preserved in the Xiaoquangou Group allowed us to discriminate their source areas and establish the source-to-sink system in the Middle–Late Triassic for SJB. The provenance analysis of thirteen samples partitions SJB and adjacent areas into four groups (Fig. 11), which is consistent with groups from the MDS plot (Fig. 10). The Bogda Mountains, CTS, and NTS appear to be the dominant sources for group 1, group 2, and group 3, respectively (Fig. 10). The mixture of NTS and CTS appear to be the dominant sources for group 4 (Fig. 10). Although the source areas of SJB include the Bogda Mountains, NTS, and CTS in the Middle–Late Triassic, the western (NTS areas) and eastern (Bogda Mountains areas) parts of SJB exhibit inconsistencies in provenance patterns. In the Bogda Mountains, the provenance is attributed to the Bogda Mountains itself, NTS and CTS (Figs. 10 and 11). Of these, the Bogda sources mainly supply near regions, such as Haojiagou section, Xidalongkou section and Well C. In contrast, the NTS piedmont (the Sikeshu and Jiangongmeikuang sections) received detritus mainly from the NTS, with minor input from the CTS and no from the Bogda Mountains. In conclusion, the NTS accounts primarily for provenance of the western part of SJB (NTS areas), while sediments of the eastern part of SJB (Bogda Mountains areas) are principally from the Bogda Mountains and/or CTS (Figs. 3, 5, 10, 11).
5.3 Implication for the initial uplift of the Bogda Mountains
The Middle–Upper Triassic Xiaoquangou Group in the Well C, Haojiagou and Xidalongkou sections yields a complex age population with many syndepositional Triassic zircons from the Bogda Mountains (Fig. 5), indicating the initiation of sediment supply from the Bogda Mountains in the same area. We attribute this abrupt change to the initial uplift of the Bogda Mountains (referred as uplift stage; Wang et al. 2019b). Moreover, the Xiaoquangou Group in the Haojiagou and Xidalongkou sections contain lithic fragments (Wang et al. 2018b, 2019b), with the sediments characterized by moderate chemical weathering and sedimentary recycling (Wang et al. 2019a, 2019b). This suggests that the Bogda Mountains provided syndepositional magmatic rocks and recycled siliciclastics to the basin, probably pointing to a “sink–source–sink” process. In the cumulative frequency curves of the lag time (lag time = crystallization age – deposition age; Fig. 12), the sediments of the uplift stage show a near “bimodal lag time” including low and high lag time groups compared to mainly low lag time groups for the syn-rift and post-rift sediments (Fig. 12). The high lag time sediments probably arise from sedimentary recycling, while the low lag time sediments are attributed to the syndepositional magmatic activity. In addition, although the proportions of the Cambrian–Devonian, Mississippian, Pennsylvanian, and Permian populations diminish, these do not vanish compared to the underlying units (Wang et al. 2018b, 2019b). This implies that both the NTS and CTS continue providing sediments for sedimentation in the Bogda Mountains. This is further evidenced by the detrital-zircon U–Pb age cumulative frequency curves and the probability plot, which exhibit zircon age groups like the syn-rift and post-rift sediments (Figs. 5 and 12). We attribute this to Late Triassic uplift of the Bogda Mountains, although the relatively smooth relief of its topography allowed scouring by rivers and consequently the supply of detritus from the NTS and CTS (Figs. 3 and 11).