Whole-Rock and Apatite Geochemistry of Late Triassic Plutonic Rocks in the Eastern Songpan-Ganzi Orogenic Belt: Petrogenesis and Implications for Tectonic Evolution

To constrain the late Triassic tectonic evolution of the Songpan-Ganzi orogenic belt, we present new whole-rock and in situ apatite geochemistry for plutonic rocks in its eastern margin. The Taiyanghe pluton can be classified into two rock types: dioritic and granitic rocks. The former exhibits low SiO2 and MgO contents but high Al2O3, Th, LREE contents, and Th/Yb and Th/Nb ratios, as well as low Ba/La and Ba/Th ratios and enriched Sr-Nd isotopic compositions, which, together with apatite geochemistry and Nd isotopes, indicate that they were derived from low degrees of partial melting of lithospheric mantle metasomatized by sediment-derived melts. The latter is characterized by high Sr and low Y and Yb, with elevated Sr/Y and (La/Yb)N ratios, implying an adakitic affinity. Notably, their similar Sr-Nd isotopic compositions indicate an origin from partial melts of a newly underplated lower crust. The Maoergai granitic rocks, characterized by high Sr and low Y and Yb contents with high Sr/Y and (La/Yb)N ratios, are indicative of adakitic rocks. In combination with the enriched whole-rock Sr-Nd isotopes and the apatite Nd isotopic data, we suggest that they were generated by the partial melting of the ancient thickened mafic lower crust. The Markam and Yanggonghai felsic granitoid rocks are peraluminous and similar to typical S-type granitoids, indicating an origin from remelting of the Triassic metasedimentary rocks. Based on the temporal-spatial relationship of the late Triassic plutonic rocks in the orogenic belt, we suggest that these rocks were formed in association with the roll-back and subsequent break-off of a subducted slab of the Paleo-Tethys Ocean. During the subduction, the formation of the Maoergai adakitic rocks was triggered by slab roll-back, whereas the magmatic “flare up” (ca. 216–200 Ma) was likely caused by slab break-off. This indicates that the final closure of the Paleo-Tethys Ocean happened in the end of the Triassic or Early Jurassic.Orogenic belts are important sites where voluminous magmatic rocks with diverse lithologic and geochemical characteristics are produced [1-3]. However, the diversity of magmatic rocks developed in orogenic belts has been a topic of debate concerning their sources, magmatic processes, and geodynamic settings involved in petrogenesis [1-7]. Particular attention has been usually focused on the geodynamic framework, which can be generally regarded to be grouped into two types according to the temporal relationship to the tectonic evolutionary process of orogenic belts: subduction and postcollision [4, 8]. Magmatic rocks with diverse geochemical characteristics in orogenic belts offer a critical window to understand the tectonic evolution of these stages [1, 3, 4, 7]. Therefore, the tectonic evolution of orogenic belts could be accurately reconstructed by investigating the temporal-spatial variability and geochemical signatures of these diverse magmatic rocks.The Songpan-Ganzi orogenic belt is widely considered to be mainly formed as a consequence of the closure of the Paleo-Tethys Ocean and subsequent continental collision between the Yangtze, Qiangtang, Kunlun, and North China terranes at the Triassic [9-12]. It mainly consists of a thick Triassic metasedimentary succession of 5–15 km [9-15], which was intruded by abundant Late Triassic plutonic rocks of ca. 231–200 Ma [12, 16, 17]. These intrusions reveal diverse petrologic and geochemical compositions that can be mainly characterized as I-, A-, and S-type granitoids [12, 17-20]. These rocks have been extensively studied as they have recorded deep geodynamic processes and provide a unique opportunity to understand the geodynamic evolution of the Songpan-Ganzi orogenic belt [12, 16-18, 20-22]. However, these Late Triassic plutonic rocks with different petrologic and geochemical characteristics gave rise to contrasting geodynamic models involved in their formation. Correspondingly, no consensus has been reached mainly concerning: (1) what specific geodynamic setting was actually responsible for the diverse magmatism in the Songpan-Ganzi orogenic belt at the Late Triassic, that is, continent (or arc)-continent collision [18, 22] or subduction of the Paleo-Tethys Ocean [12, 16] and (2) when did the Paleo-Tethys Ocean, located in the southern section of the Songpan-Ganzi orogenic belt, finally close, in the Late Triassic [19, 22] or the Early Jurassic [23, 24]. Differentiating these contrasting geodynamic models and further improving our understanding of the time closure of the Paleo-Tethys Ocean require a comprehensive investigation regarding the temporal-spatial variability and geochemical signatures of the Late Triassic plutonic rocks. However, most of the previous works mainly focused on the whole-rock geochemistry and rarely concentrated on mineralogical geochemistry to discuss the petrogenesis and tectonic evolution. Therefore, comprehensive research on whole-rock and mineralogical geochemistry could provide not only new constraints on petrogenetic mechanisms of these plutonic rocks but also give a better understanding on the tectonic evolution of the Songpan-Ganzi orogenic belt during the Late Triassic.In this contribution, we conducted a combined study of whole-rock and apatite geochemical analyses on the variety of plutonic rocks in the eastern Songpan-Ganzi orogenic belt. The integration of these data with previous research studies can provide insights into the petrogenesis of these plutonic rocks and the geodynamic history of the Songpan-Ganzi orogenic belt in particular concerning the time closure of the Paleo-Tethys Ocean during the Late Triassic.The Songpan-Ganzi orogenic belt is a vast triangular-shaped domain (one-fifth of the Tibetan Plateau, >2.2 × 105 km2) in the northern Tibetan Plateau (Figure 1(a)) [9-12]. It formed as a consequence of continent-continent collisions between the North China, Yangtze, and Qiangtang terranes, following the subduction of the Paleo-Tethyan Ocean plate (Figure 1(a)) [9-12]. It is located between the Qiangtang, Qiadam-Kunlun, and Yangtze terranes and is separated from them by the Jinshajiang suture, Anyimaqen suture, and Longmen Shan thrust belt, respectively (Figures 1(a) and 1(b)) [9, 11].The Songpan-Ganzi orogenic belt consists of two main compositions: the Triassic stratigraphy and Late Triassic-Early Jurassic magmatic rocks (Figure 1(b)) [10-12]. The Triassic stratigraphy is almost exclusively composed of a thick (5–15 km) marine flysch deposit, known as the Xikang Group (Figure 1(b)) [9-11], which experienced simultaneous folding and high-temperature/medium-low pressure metamorphism [9-12, 25-27]. Intensive Late Triassic-Early Jurassic plutonic magmatism is dominated by granitoid rocks with subordinate mafic-intermediate lithologies, which intruded into the Triassic marine metasedimentary unit. In addition to mafic-intermediate lithologies with calc-alkaline feature [12, 16, 28], the granitoid rock series exhibit large petrological and geochemical variations, with three main groups identified to date: (1) I-type series characterized by aluminous to weakly peraluminous feature (e.g., the Yanggonghai and Maoergai plutons; Figure 1(c)) [12, 16-18, 20-22, 28-30]; (2) S-type series of highly evolved peraluminous granitoids (e.g., the Markam pluton) associated with Li-metal mineralization (Figure 1(c)) [17, 31-35]; and (3) A-type series with alkaline feature (e.g., Nyanbaoyeche pluton; Figure 1(c)) [12, 17, 19].In this study, the samples were collected from the Taiyanghe, Yanggonghai, Maoergai, and Markam plutons in the eastern Songpan-Ganzi orogenic belt (Figure 1(c)). The Taiyanghe pluton, which has an area of approximately 75 km2, is petrologically diverse and dominated by low SiO2 dioritic rocks and high SiO2 granitic rocks (Figures 2(a)–2(c)). Minerals in the low SiO2 dioritic samples are dominated by hornblende and plagioclase, with or without K-feldspar, quartz, and biotite, with accessory apatite, titanite, and zircon minerals (Figures 2(a) and 2(b)). Minerals in the high SiO2 granitic rocks include K-feldspar, plagioclase, biotite, and quartz, with minor accessory apatite, titanite, and zircon (Figure 2(c)). The Yanggonghai and Maoergai plutons have a respective area of ~1050 km2 and ~ 60 km2 and are composed of undeformed and medium-grained granitoid rocks, ranging from granodiorite to granite in composition. Minerals in these two plutons are dominated by plagioclase, K-feldspar, quartz, and biotite, with minor accessory apatite, monazite, allanite, and zircon (Figures 2(d)–2(f)). The Markam pluton has an area of ~525 km2 and is composed of undeformed and medium-grained two-mica granitoids, with minerals dominated by quartz, plagioclase, K-feldspar, biotite, and muscovite (Figures 2(g) and 2(h)).In total, thirty-five granitoid rocks were powdered to ~200 mesh in an agate mortar and a pestle. For major element analysis, powered samples were used to create fused glass beads, which were analyzed by X-ray fluorescence (XRF) using a Rigaku RIX 2000 XRF spectrometer at the Nanjing FocuMS Technology Co. Ltd, China. The analytical procedures are described in detail by Jiang et al. [36]. Calibration lines used for quantification were taken by bivariate regression of data recovered from thirty-six reference materials, encompassing a wide range of silicate compositions [37]. The analytical precision was between 1% and 5%. Trace element concentrations, including rare earth elements (REEs), were analyzed using an Agilent 7700× inductively coupled plasma mass spectrometer (ICP–MS; Hachioji, Tokyo, Japan) at the Nanjing FocuMS Technology Co. Ltd, China. The analytical procedure is described in detail by Jiang et al. [36]. Multiple analytical results of international standards (BHOV-2, AGV-2, BCR-2, and GSP-2) were consistent with the recommended values, and the analytical precision was <5%. In addition, we selected twenty-four and twenty-seven samples for whole-rock Sr and Nd isotopic composition analysis, respectively, using a Nu Instruments Nu Plasma II MC-ICP-MS (Micromass Isoprobe multi-collector mass spectrometer) (Wrexham, UK) and a Teledyne CETAC Technologies Aridus II desolvating nebulizer system (Omaha, Nebraska, USA; MC-ICPMS) at the Nanjing FocuMS Technology Co. Ltd, China. The analytic procedures are described in detail by Sun et al. [38]. The measured raw Sr and Nd isotope data were fractionation normalized to 86Sr/88Sr = 0.1194 and 146Nd/144Nd = 0.7219, respectively. International isotopic standards (NIST987 for Sr and JNdi-1 for Nd) are analyzed after each three studied samples to correct instrumental drift. Geochemical reference materials (BCR-2, AGV-2, BHVO-2, and STM-2) were analyzed for quality control.Apatite grains from intrusive rocks of the Taiyanghe, Yanggonghai, Maoergai, and Markam plutons were mounted in epoxy resin and then polished. Backscattered electron images were obtained using a SUPRA 55 SAPPHIRE spectrometer coupled to secondary electron and energy-dispersive X-ray spectrometry detectors at the Nanjing Hongchuang Geological Exploration Technology Service Co., Ltd. The images were used to characterize internal structures and select optimal sites for in situ elemental and Nd isotopic analyses. LA–ICP–MS (laser ablation-inductively coupled plasma-mass spectrometry) and LA–MC–ICP–MS (laser ablation multi-collector inductively coupled plasma mass spectrometry) analyses were undertaken sequentially at the same or nearby analysis sections.In situ trace elements were measured by LA–ICP–MS using a resolution M-50 LA system fitted with an Agilent 7900a ICP-MS at the Wuhan Sample Solution Analytical Technology Co., Ltd., China. The analytic procedures are described in detail by Liu et al. [39] and Zong et al. [40]. Laser sampling was conducted using a GeoLasPro laser ablation system incorporating a COMPexPro 102 ArF excimer laser (wavelength of 193 nm and maximum energy of 200 mJ) and a MicroLas optical system. An Agilent 7700e ICP-MS instrument was used to obtain the ion-signal intensities. Helium (He) was used as a carrier gas, and Argon (Ar) was applied as the make-up gas; He and Ar were mixed via a T-connector before entering the ICP system. A “wire” signal smoothing device was incorporated into this laser ablation system [41]. The spot size and frequency of the laser were 44 µm and 5 Hz, respectively. Trace element compositions were calibrated against various reference materials (BHVO-2G, BCR-2G, and BIR-1G) without using an internal standard [39]. Each analysis incorporated a background acquisition of approximately 20–30 seconds followed by 50 seconds of data acquisition from the sample. The Excel-based software ICPMSDataCal was used to perform off-line selection and integration of the background, analyze signals, perform time-drift correction, and provide quantitative calibration for trace element analyses [39].In situ Nd isotopic compositions were measured by LA–MC–ICP–MS using a Neptune Plus MC–ICP–MS (Thermo Scientific) coupled with a RESOlution M-50 193 nm LA system (Resonetics) and a GeoLas HD excimer ArF LA system (Coherent, Göttingen, Germany) at the Wuhan Sample Solution Analytical Technology Co., Ltd., China. In the LA system, He was applied as the carrier gas within the ablation cell and was subsequently merged with Ar gas. A small amount of nitrogen gas was added to Ar to increase the sensitivity of Nd isotopes [42]. A spot diameter of 90 µm was used for apatite analysis. A new signal-smoothing device was applied downstream of the sample cell to efficiently remove short-term variation in the signal and eliminate mercury from the background and sample aerosol particles [41]. The Neptune Plus was coupled to nine Faraday cups equipped with 1011 Ω resistors. Isotopic 142Nd, 143Nd, 144Nd, 145Nd, 146Nd, 147Sm, 148Nd, and 149Sm compositions were collected in Faraday cups using the static mode. The mass discrimination factor for 143Nd/144Nd was determined using 146Nd/144Nd (0.7219) with the exponential law. The 149Sm signal was adopted to correct for 144Sm interference on 144Nd by applying the 144Sm/149Sm ratio (0.2301). The mass fractionation of 147Sm/149Sm was corrected via the 147Sm/149Sm normalization by using the 147Sm/149Sm ratio (1.08680) and the exponential law. All data reduction for MC–ICP–MS Nd isotope ratios was performed using the “Iso-Compass” software [43]. Two natural apatite megacrysts (Durango and MAD) were measured as unknown samples to check the accuracy of the calibration method. The eight analyses of Durango apatite and thirty analyses of MAD apatite yielded average 143Nd/144Nd ratios of 0.512466 ± 0.000018 (2SD) and 0.511347 ± 0.000020 (2SD), which are consistent with the previously reported values (0.512490 ± 0.000018 [2SD] and 0.511322 ± 0.000053 [2SD]), respectively [44].The major and trace element contents of the representative samples from the Taiyanghe, Yanggonghai, Maoergai, and Markam plutons are given in online supplementary Table S1.The rocks from the Taiyanghe pluton can be divided into gabbroic-monzonitic and felsic granitic rocks in the SiO2 versus K2O + Na2O diagram, which will be referred to as diorite (Hb-gabbro to monzonite) and granite for simplicity in the following discussion (Figure 3(a)). The dioritic rocks exhibit varying SiO2 contents of 51.38–57.92 wt.% and Na2O + K2O contents of 4.13–8.84 wt.%, and samples with variable K2O (1.31–5.35 wt.%) contents plot into the intermediate-high-K calc-alkaline rock fields on the SiO2 versus K2O diagram (Figure 3(b)). Their A/CNK (molar Al2O3/[CaO+Na2O+K2O]) values range from 0.94 to 1.04, indicating metaluminous to weakly peraluminous composition signature (Figure 3(c)). Additionally, these studied dioritic rocks have variable Al2O3, Fe2O3T, P2O5, TiO2, Na2O, and CaO contents of 18.20–20.55 wt.%, 7.26–14.08 wt.%, 0.21–0.48 wt.%, 0.58–1.31 wt.%, 2.82–4.06 wt.%, and 3.74–7.51 wt.%, respectively, with low MgO (0.87–1.64 wt.%), Cr (4.6–27.2 ppm), and Ni (3.1–15.4 ppm) contents (Figure 4). In the Harker diagram, they display correlations between major trace elements and SiO2, and all exhibit weakly negative relationships between SiO2 and Fe2O3T, MgO, P2O5, TiO2, and CaO and weakly positive correlations between SiO2 and Al2O3 and Na2O (Figure 4). Furthermore, these dioritic rocks have low Mg# (18–23), which is not related to SiO2 (Figure 4(h)).The dioritic rocks have elevated total REE contents, ranging from 350.4 to 3299.9 ppm, with an average of 1072.5 ppm. Chondrite-normalized REE distribution patterns revealed significant enrichment in LREEs (light REEs) and depletion in HREEs (heavy REEs; [La/Yb]N = 11.96–68.42), with moderate to weak Eu (Eu/Eu* = 0.31–0.80) negative anomalies (Figure 5(a)). Primitive mantle-normalized trace element distribution pattern diagram indicates that the dioritic samples are characterized by large-ion lithophile element (LILEs, e.g., Rb, U, K, and Pb) enrichment and high-field earth element (HFSEs, such as Nb, Ta, and Ti) depletion (Figure 5(b)). In addition, they are also rich in Th and are characterized by elevated Th/Yb and Th/Nb and low Ba/La and Ba/Th ratios (Figures 6(c) and 6(d); online supplementary Table S1).Compared with the dioritic rocks, the granitic rocks exhibit elevated silica and total alkalis, with SiO2 varying from 70 to 74.15 wt.% and total Na2O + K2O varying between 7.68 and 8.09 wt.% (Figure 3(a)). These granitic samples are characterized by high K2O contents (4.86–5.77 wt.%), displaying high-K characteristics (Figure 3(b)). Furthermore, their A/CNK values range between 1.01 and 1.08, indicating weakly peraluminous composition signature (Figure 3(c)). Additionally, these granitic samples have low Al2O3 (12.99–14.61 wt.%), Fe2O3T (2.17–3.76 wt.%), MgO (0.15–0.42 wt.%), P2O5 (0.02–0.07 wt.%), TiO2 (0.10–0.26 wt.%), Na2O (2.32–2.86 wt.%), CaO (1.53–2.25 wt.%), Mg# (12–19), Cr (0.86–5.45 ppm), and Ni (0.37–3.26 ppm), which display no correlations with SiO2 in the Harker diagram (Figure 4).The granitic samples exhibit LREE-enriched chondrite-normalized distribution patterns and show significant fractionation of REE ([La/Yb]N = 37.57–270.16; [Gd/Yb]N = 0.83–1.31), with slightly negative to positive Eu (Eu/Eu* = 0.83–1.31) anomalies (Figure 5(c)). Additionally, all samples demonstrate pronounced depletion in HFSEs (Nb-Ta-Ti) and substantial enrichment in LILEs (Rb, U, K, and Pb) on the primitive mantle-normalized multielement distribution diagram (Figure 5(d)). Notably, these granitic samples feature elevated Sr (338–396 ppm) but low Y (4.0–11.3 ppm) and Yb (0.2–1.0 ppm) contents, with raised Sr/Y (29.9–98.8) and (La/Yb)N (37.57–270.16) ratios, clearly falling within the adakite field (Figure 7).The Yanggonghai and Maoergai are adjacent magmatic plutons and characterized by high SiO2 and total alkalis contents ranging from 72.26 to 73.73 wt.% and 7.29 to 8.31 wt.%, respectively (Figure 3(a)). These samples exhibit relatively high K2O contents of 4.68–4.93 wt.% for the Yanggonghai and 3.81–4.39 wt.% for the Maoergai, plotting into high-K calc-alkaline field, respectively (Figure 3(b)). They have high A/CNK ratios of 1.05–1.14 (1.12–1.14 and 1.05–1.13), with weakly to strongly peraluminous features (Figure 3(c)). All of these granitic rocks exhibit very similar Al2O3 (14.32–14.45 and 14.68–15.15 wt%), Fe2O3T (1.06–1.32 and 0.77–1.41 wt.%), MgO (0.24 and 0.16–0.44 wt.%), P2O5 (0.17–0.24 and 0.07–0.09 wt.%), TiO2 (0.17–0.20 and 0.11–0.23 wt.%), Na2O (3.01–3.19 and 3.39–4.22 wt.%), CaO (1.31–1.34 and 1.09–2.38 wt.%), Mg# (26–31 and 29–38), Cr (3.5–5.2 and 1.2–10.2 ppm), and Ni (0.9–5.7 and 0.3–7.5 ppm) contents (Figure 4). These element contents do not show any significant correlations with SiO2 on the Harker diagram (Figure 4).The chondrite-normalized REE distribution patterns indicate that both felsic granites from the Yanggonghai and Maoergai plutons exhibit LREE enrichment ([La/Yb]N = 21.14–37.51, 33.81–49.26) and weakly Eu (Eu/Eu* = 0.64–0.84, 0.43–0.58) negative anomalies (Figures 5(e) and 5(g)). Additionally, all samples display a strong depletion of HFSE (Nb, Ta, and Ti) and significant enrichment of LILEs (e.g., Rb, K, and U), as shown in primitive mantle normalized multielement distribution diagram (Figures 5(f) and 5(h)). Notably, the Yanggonghai granites are characterized by high concentrations of HREEs and Y (e.g., Y = 8.2–19.4 ppm and Yb = 0.6–1.3 ppm), resulting in low Sr/Y (14.68–21.79) and (La/Yb)N (21.14–37.51) ratios, while the Maoergai granitic rocks have high Sr (355–538 ppm) and low Y (4.7–8.4 ppm) and Yb (0.3–0.5 ppm) contents, with high Sr/Y (53.96–75.34) and (La/Yb)N (33.81–49.26) ratios, which also plot into the adakite field (Figure 7).The granitoids of the Markam pluton exhibit high SiO2 contents, ranging from 66.56 to 74.60 wt.%, placing them in the granodiorite and granite fields on the SiO2 versus K2O + Na2O diagram (Figure 3(a)). These granitoid samples display varying K2O (2.08–5.97 wt.%), Na2O (2.75–4.81 wt.%), and total K2O + Na2O ranging from 6.38 to 8.97 wt.%, with most having Na2O < K2O (Na2O/K2O = 0.50–2.31; Figure 3(a)). Most of the samples exhibit high-K calc-alkaline to shoshonitic characteristics, except for a sample that displays intermediate-K feature (Figure 3(b)). Their A/CNK values range from 1.06 to 1.19, positioning them within weakly to strongly peraluminous fields (Figure 3(c)). Additionally, these granitoids share variable Al2O3 (13.69–17.69 wt.%), P2O5 (0.09–0.24 wt.%), and low Fe2O3T (1.11–3.16 wt.%), MgO (0.20–0.78 wt.%), TiO2 (0.16–0.44 wt.%), CaO (0.56–3.27 wt.%), Mg# (27–36), Cr (1.4–11.6 ppm), and Ni (0.4–8.0 ppm; Figure 4). Furthermore, there are weakly negative relationships between SiO2 and Al2O3, Fe2O3T, and CaO, while they display a weakly positive correlation between SiO2 and P2O5 (Figure 4). There is no relationship observed between SiO2 and MgO, TiO2, Na2O, and Mg# (Figure 4).These granitoids exhibit similar chondrite-normalized REE distribution patterns, characterized by LREE ([La/Yb]N = 5.83–63) enrichment and varying degrees of negative Eu (Eu/Eu* = 0.21–0.80) anomalies (Figure 5(i)). In the primitive mantle-normalized multielement distribution diagram, most granitoids display significant negative anomalies in Nb, Ta, Ti, and P, but strong enrichments in Rb, Th, Pb, and K relative to neighboring elements (Figure 5(j)).The Sr-Nd isotopic data for the plutonic rocks from the Taiyanghe, Yanggonghai, Maoergai, and Markam plutons are listed in online supplementary Table S1 and illustrated in Figure 8. Both the dioritic and granitic rocks in the Taiyanghe pluton exhibit nearly identical 87Sr/86Sr(t) (0.7084–0.7090 and 0.7088-0.7089, respectively) as well as εNd(t) (–6.4 to –5.4 and –6.1, respectively) ratios (Figure 8). The granitic rocks from the Yanggonghai and Maoergai plutons share similar ranges of 87Sr/86Sr(t) (0.7075–0.7089 and 0.7075–0.7083) values, but the Maoergai granitic rocks display slightly higher εNd(t) (–8.7 to –8.0) ratios compared with these Yanggonghai granites (–9.6 to –9.4; Figure 8). In contrast to the Yanggonghai and Maoergai granitic rocks, the Markam granitoids exhibit more variable ranges of both 87Sr/86Sr(t) (0.7093–0.7184) and εNd(t) (–9.9 to –7.2) values (Figure 8).In situ trace elemental and Nd isotopic data for apatites from these different types of plutonic rocks are given in online supplementary Tables S2 and S3.The apatite grains from two Taiyanghe dioritic samples (SG21-10-62a and 63b) exhibit elevated Sr (335–486 and 377–554 ppm), Sr/Y (0.20–1.11 and 0.33–4.54), and (La/Yb)N (0.84–18.46 and 1.17–315.23) ratios, which demonstrate a clear positive correlation with those observed in the host rocks (Figure 9). Additionally, they also display high Ce/Y, Th/U, as well as La/Sm values (Figures 10(a) and 10(b)). Apatite grains from a Taiyanghe granitic sample similarly show significantly high Sr (253–363 ppm) contents and Sr/Y (0.32–3.58) and (La/Yb)N (8.70–152.78) ratios, all of which are positively correlated with those found in the host rock (Figure 9). Furthermore, these grains also exhibit elevated Ce/Y, Th/U and La/Sm ratios akin to those observed in the dioritic rocks (Figures 10(a) and 10(b)).The apatite grains derived from the dioritic rocks possess Nd isotopic compositions characterized by high 143Nd/144Nd(t) values ranging from 0.511998 to 0.512079, corresponding to εNd(t) values ranging between –7.23 and –5.64 (Figure 11). Similarly, the apatite grains obtained from the granitic sample show comparable 143Nd/144Nd(t) values ranging from 0.512008 to 0.512048 and εNd(t) values ranging from –6.96 to –6.18 (Figure 11). These results are similar to those of host rocks (143Nd/144Nd[t] = 0.512041–0.512094 and 0.512053 and εNd[t] [–6.4 to –5.4 and –6.1]; Figure 11). In addition, the consistency in Nd isotope ratios of apatite from these two distinct intrusive rock types is evident (Figure 11).Apatite grains from the Yanggonghai granitic sample (SG21-05-25a) exhibit predominantly low Sr (167–178 ppm) contents and Sr/Y (0.07–0.09) and (La/Yb)N (0.90–1.07) ratios, while those in the two Maoergai granitic samples (SG21-07-38a and SG21-07-40a) display high Sr (289–396 ppm) contents and Sr/Y (0.13–0.30) and (La/Yb)N (1.16–2.63) ratios, with positive correlation to those observed in the host rocks (Figure 9). In addition, all apatite grains from both Yanggonghai and Maoergai granitic rocks show low Ce/Y, La/Sm, as well as Th/U ratios (Figures 10(a) and 10(b)).In terms of Nd isotopic compositions, apatite grains from the Yanggonghai granite have lower 143Nd/144Nd(t) values ranging between 0.511809 and 0.511914 with more negative εNd(t) values of –11.20 to –9.15 compared with those from the Maoergai granitic samples, which range between 0.511832 and 0.511947 with εNd(t) values of –10.32 to –8.07 (Figure 11). It is worth noting that the variation range of 143Nd/144Nd(t) and εNd(t) for apatite is almost identical to those of host rocks (143Nd/144Nd[t] = 0.511893 and 0.511919–0.511921, respectively; εNd[t] = –9.6 and −8.6, respectively), from which apatite crystallized (Figure 11).The apatite grains from four Markam granitoid samples (SG21-09-52a and SG21-10-53a, 56a, and 59a) exhibit consistently low Sr (70.6–173 ppm) contents and Sr/Y (0.041–0.147) and (La/Yb)N (0.66–2.14) ratios, which display a positive correlation with those observed in the host rocks (Figure 9). Furthermore, these grains also show low Ce/Y, Th/U, as well as La/Sm ratios (Figures 10(a) and 10(b)). Their isotopic compositions of 143Nd/144Nd(t) range from 0.511816 to 0.512057, while their εNd(t) values vary between –10.88 and –6.35; both of which are similar to those of the host rocks (143Nd/144Nd[t] = 0.511872–0.512011 and εNd[t] = –7.2 to –9.9; Figure 11).The crystallization ages of intrusive rocks from the Taiyanghe, Yanggonghai, Maoergai, and Markam plutons have been studied by zircon U–Pb dating in recent years [16, 18, 22, 35]. Previous zircon U–Pb dating yielded crystallization ages of ca. 202–239 Ma for the Taiyanghe dioritic rocks, suggesting a long-term magmatic activity [16, 22]. Although no previous study has determined the crystallization age of the Taiyanghe granites using zircon U–Pb dating, our unpublished data obtained from a granitic sample (SG21-10-65a) from the Taiyanghe pluton yielded a weighted mean 206Pb/238U age of 212.5 ± 0.6 Ma, which is considered as the crystallization age for these granites. Previous zircon U–Pb SHRIMP dating revealed that the Yanggonghai and Maoergai granitoids with adakitic affinity share magmatic emplacement ages of 221 ± 3.8 and 216 ± 5.7 Ma, respectively [18]. However, we conducted monazite U–Pb dating for a Yanggonghai granitic sample (SG21-05-25a) with a weighted mean 206Pb/238U age of 198.2 ± 1.1 Ma (our unpublished data), taken to be the emplacement age of this granite. The Markam granitoids were previously analyzed by zircon U–Pb isotopes, which mainly give Late Triassic ages ranging from ca. 231 to ca. 200 Ma [16, 35].The whole-rock geochemistry for the rocks from the Taiyanghe pluton reveals dioritic compositions, characterized by low SiO2 contents ranging from 51.38 to 57.92 wt.% (Figures 3(a) and 4). Three potential magmatic processes could account for the formation of these dioritic rocks: (1) melting of mafic lower crust [45, 46]; (2) fractional crystallization of mafic magma [47, 48]; and (3) melting of a metasomatized mantle [49, 50]. Intermediate-acid magmatic rocks can be formed by partial melting of mafic lower crust, exhibiting low MgO contents and Mg# values [51, 52]. The observed low MgO, Cr, and Ni contents and Mg# values in the Taiyanghe dioritic rocks suggest their possible derivation from the mafic lower crust source (Figures 4(c) and 4(h)). However, it is worth noting that plutonic rocks originated from the mafic lower crust in the eastern Songpan-Ganzi orogenic belt typically exhibit high SiO2 contents (>65 wt.%) [18, 21, 22]. The significantly lower SiO2 contents in the Taiyanghe dioritic rocks compared with those derived from the mafic lower crust clearly rule out this source as a plausible origin for these samples. Furthermore, most of the analyzed Taiyanghe dioritic samples display considerably higher Nb/Ta ratios ranging between 23 and 29 than the average value reported for typical mafic lower crust (Nb/Ta = ~8.3) [53], further precluding the possibility that they were formed by partial melting of the mafic lower crust.There are continuous magmatic series (49.83–57.92 wt.% SiO2) in the Taiyanghe pluton [16, 22], possibly indicating a progressive evolution through fractionation crystallization that gives rise to the observed magmatic succession within the Taiyanghe pluton. Deschamps et al. [16] also proposed that the observed continuous magmatic series in the Taiyanghe pluton can be reasonably explained by the progressive fractionation of both amphibole and plagioclase minerals. However, our study on the Taiyanghe dioritic rocks along with previous data on low MgO dioritic rocks consistently show low MgO contents [16, 22], which is contrary to those of high MgO (5.02–9.98 wt.%) dioritic samples from the Taiyanghe pluton (Figure 4(c)). This discrepancy suggests that these low MgO dioritic rocks were unlikely derived from continuous differentiation of mafic magma. Furthermore, fractional crystallization from mafic magma would not result in a positive correlation between La and La/Sm [54]. The combined data for the Taiyanghe dioritic rocks in this study and previously reported low MgO dioritic rocks clearly align along a trend indicative of partial melting rather than fractional crystallization in the La versus La/Sm diagram (Figure 6(a)), implying that their direct derivation from partial melting of the mantle source rather than fractional crystallization of mafic magma.The high 87Sr/86Sr(t) and negative εNd(t) ratios and the arc magma characteristics of the Taiyanghe dioritic rocks were probably caused by crustal assimilation during magma ascent and emplacement. However, considering their variable SiO2 contents and almost consistent Sr-Nd isotopic compositions, we suggest that crustal contamination played a limited role in the formation (Figure 8). The Taiyanghe dioritic rocks with the enriched Sr-Nd isotopic characteristics possibly originated from a metasomatized mantle source (Figure 8). The enrichment in LILEs and the depletion in Nb, Ta, and Ti indicate that these rocks likely originated from a mantle source metasomatized by slab-derived hydrous fluid or melt phases prior to magma production (Figure 5(b)). Additionally, the low MgO contents of these dioritic rocks suggest an enriched mantle as a possible magma source compared with typical Mg-rich andesitic rocks originated from a depleted mantle (Figure 4(c); generally >6 wt.%) [55]. The results of the melting experiment indicate that partial melting of fertile peridotite at low degrees is enriched in SiO2, Al2O3, and Na2O and depleted in MgO, Fe2O3T, and CaO compared with higher-degree melts [56, 57]. In addition to low MgO contents, the Taiyanghe dioritic rocks exhibit high Al2O3 and Na2O but low Fe2O3T and CaO contents (Figures 4(a), 4(b), 4(f), and 4(g)4), which are consistent with partial melts from fertile mantle peridotite at low degrees [56, 57]. The presence of an abundance of hornblende along with extensive apatite and magnetite suggests a water-rich and oxidizing environment in the parental mafic magma [58], further supporting the origin of these dioritic rocks from metasomatized lithospheric mantle (Figures 2(a) and 2(b)). The relatively elevated (La/Yb)N and Dy/Yb ratios in the Taiyanghe diorites suggest potential partial melting of metasomatized lithospheric mantle within the spinal-garnet stability field (Figure 6(b)). Furthermore, the high Al2O3 abundances in the rocks indicate a significant presence of Al2O3-bearing minerals (such as spinel, garnet, and pyroxene) in their mantle source (Figure 4(a)), while their relatively high HREE contents (Y = 26.1–190 ppm and Yb = 2.25–17.1 ppm) reflect a source that is relatively deficient in garnet (Figure 7). The primary mineral phases containing Al2O3 in this mantle source are primary pyroxene and spinel [59]. Therefore, it is likely that the parental magma for the Taiyanghe dioritic rocks likely originated from hornblende-bearing mantle peridotite beneath the Songpan-Ganzi orogenic belt within a transitional stability region between garnet and spinel [60]. This fertile mantle source may have been generated through multiple metasomatic events involving melt/fluid phase derived from subduction-related processes.Previous studies have suggested that distinct trace elemental composition features in metasomatized mantle can be caused by different types of metasomatic agents, such as hydrous fluids and sediment melts [61]. Specially, the fractionation of Ba/La and Ba/Th can be easily attributed to the mobility of Ba in hydrous fluid phases, while Th and LREE are generally considered less mobile than LILEs in these fluid phases [62-66]. Accordingly, these trace element variables can serve as solid indicators for potential contributions from hydrous fluids or sediment-derived melts to the mantle magma source area [63, 67]. Regarding the Taiyanghe dioritic rocks, their high Th and LREE contents along with elevated Th/Yb and Th/Nb ratios but low Ba/La and Ba/Th ratios clearly indicate that their magma source was influenced by sediment-derived melts (Figures 6(c) and 6(d)) [63, 64, 66]. In addition to the Late Paleozoic-Early Mesozoic Paleo-Tethys Ocean plate subduction process, it has been suggested that a Neoproterozoic oceanic plate subducted beneath the Yangtze Craton and approaching terranes (e.g., Yidun and Kunlun terranes) [28, 68-71]. These subduction processes may have contributed to the enrichment of the lithospheric mantle through metasomatism induced by slab-derived fluid or melt phases [28].Previous research has indicated that the elemental and isotopic composition characteristics of apatite can provide valuable insights into the origin and evolution of its host rock [72-75]. Consequently, apatite is frequently utilized to identify petrogenetic processes in igneous rocks [74, 75]. In this study, an in situ elemental analysis was conducted on apatite from the Taiyanghe dioritic rocks, revealing that most grains fall within the overlapping areas of crust-mantle-derived granitoids in the Ce versus Y and La/Sm versus Th/U diagrams (Figures 10(a) and 10(b)) [75]. The evidence suggests that the host magma of apatite grains was derived from a mantle source modified by crust-derived materials. Furthermore, apatite grains within the Taiyanghe dioritic rocks exhibit almost identical 143Nd/144Nd(t) (0.511998–512079) and εNd(t) (−7.23 to −5.64) ratios as their host rocks (143Nd/144Nd[t] = 0.512041–0.512074; εNd[t] = −6.4 to −5.8), indicating that they crystallized from their host magma (Figure 11). Based on the negative εNd(t) ratios combined with element ratios observed in these apatite grains, it is likely that the parental magma originated from an enriched mantle source, which aligns with whole-rock geochemical interpretations. Therefore, the integration of whole-rock geochemistry with apatite elemental and Nd isotopic composition characteristics suggests that the dioritic rocks from the Taiyanghe pluton were formed by low-degree partial melting of lithospheric mantle that has been metasomatized by sediment-derived melts.Previous research has suggested that the Taiyanghe granites are representative of S-type granitic rocks originated from a single Triassic metasedimentary source [16]. However, the Taiyanghe granites analyzed in this study exhibited relatively low A/CNK ratios, indicating that they may be I-type rather than S-type granites. Additionally, these granitic samples have very low P2O5 contents, which are inconsistent with typical S-type granitoids that possess high P2O5 abundances. It should be noted that the Taiyanghe granites share high Sr contents, Sr/Y and (La/Yb)N ratios and low Y and Yb contents (Figure 7), suggesting that they are adakitic rocks [76]. This conclusion is further supported by apatite geochemistry analysis. First, apatite grains from a Taiyanghe granitic sample display high Sr contents, Sr/Y, and (La/Yb)N ratios, indicative of an adakitic affinity for their host rock (Figure 9). Second, most of these apatite grains plot into the I-type granitoid rock field on a 10*Sr-LREE-10Y ternary discrimination diagram (Figure 10(c)) [77] and further fall within the adakite field on a 10/Eu/Eu*-10Sr-HREY ternary discrimination diagram (Figure 10(d)) [78]. Therefore, the whole-rock and apatite geochemical characteristics suggest that the Taiyanghe granites are I-type adakitic rocks instead of S-type granitoids.As previously mentioned, the distinct trace element characteristics of the Maoergai granites exhibit clearly adakitic affinities with high Sr contents and Sr/Y and (La/Yb)N ratios and low Y and Yb contents (Figure 7), which is consistent with previous research findings [18]. Additionally, apatite grains from the Maoergai granitic rocks also display high Sr contents and Sr/Y and (La/Yb)N ratios that are similar to those of the host rocks (Figure 9). Furthermore, on the 10Sr-LREE-10Y ternary discrimination diagram, apatite grains from the Maoergai granites plot into TTG (tonalite-trondhjemite-granodiorite) fields (Figure 10(c)) [77]. Although all apatites fall within the nonadakitic granitoid field, they cluster closely to the adakite field on a 10/Eu/Eu*-10Sr-HREY ternary discrimination diagram (Figure 10(d)) [78]. Thus, both whole-rock and apatite geochemical characteristics suggest that the Maoergai granites are adakitic rocks.Adakitic rocks can be produced through different magmatic processes, including (1) partial melting of subducted oceanic plate [79-81]; (2) fractional crystallization of basaltic arc magma [82, 83]; (3) partial melting of foundered or thickened mafic lower continental crust [84-86]; and (4) mixing of between crust-mantle-derived magmas [87].The Sr-Nd isotopic compositions of the adakitic granitic rocks from the Taiyanghe and Maoergai plutons differ significantly from those of the Jinshajiang and Anyimaqen Paleo-Tethyan Ocean ophiolites [88, 89], indicating that these adakitic rocks were unlikely formed by partial melting of the subducted oceanic plate (Figure 8). Additionally, the K2O (3.81–4.39 and 4.86–5.77 wt.%) contents in these adakitic rocks are much higher than those in sodic adakitic rocks (K2O = 1.97 ± 0.5 wt.%, n = 267) derived from partial melting of the subducted oceanic slab [80], further ruling out their origin from the subducted Paleo-Tethys Oceanic crust (Figure 3(b)).To generate a significant amount of Late Triassic adakitic rocks from the Taiyanghe and Maoergai plutons through basaltic magma differentiation, substantial coeval mafic rocks originated from the mantle would be necessary. For instance, the Camiguin and Central Mindanao Arc high-silica adakitic lavas were believed to have been produced by magma differentiation that was originated from a common mantle source [82]. The intensive Late Triassic magmatism is dominated by felsic granitoid rocks with subordinate mafic-intermediate lithologies, which contradicts a fractionation crystallization model. Although the Taiyanghe pluton contains dioritic and adakitic granitic rocks, suggesting that these high SiO2 granites possibly originated from the low SiO2 dioritic rocks by fractionation crystallization, the whole-rock compositional gap between the granitic and dioritic rocks clearly rules out continuous evolution through fractionation crystallization to produce the observed dioritic rocks (49.83–57.92 wt.% SiO2) and high SiO2 granitic rocks (70.00–75.07 wt.% SiO2) [16, 22].The formation of adakitic rocks via mixing of between crust-mantle-derived magmas is a possible process, as demonstrated by high-Mg andesitic rocks from Yanji area, NE China and Mount Shasta, US [87, 90]. However, such adakitic rocks derived from mixed origins typically exhibit high MgO (~8 wt.% for Mount Shasta; 3.64–4.36 wt.% for Yanji), Cr (~500 ppm for Mount Shasta; 128–161 ppm for Yanji), and Ni (~120 ppm for Mount Shasta; 86–117 ppm for Yanji) but low SiO2 (~58 wt.% for Mount Shasta; 60.92–62.20 wt.% for Yanji) contents. In contrast, the adakitic granites from the Taiyanghe and Maoergai plutons have low MgO (0.15–0.42 wt.% for Taiyanghe; 0.16–0.44 wt.% for Maoergai), Cr (0.86–5.45 ppm; 1.15–10.2 ppm), Ni (0.37–3.26 ppm; 0.34–7.46 ppm) but high SiO2 (70–74.15 wt.%; 72.26–73.73 wt.%; Figure 4(c); online supplementary Table S1). These geochemical characteristics of the adakitic granites from the Taiyanghe and Maoergai plutons thus preclude mixing of between crust-mantle-derived magmas for their formation.Partial melting of a foundered or thickened mafic lower crust is the most probable mechanism for generating adakitic rocks from the Taiyanghe and Maoergai plutons. These adakitic granites exhibit high SiO2 and low MgO, Mg#, Cr, and Ni contents (Figures 4(c), 4(h),), indicating their direct derivation from a thickened rather than a foundered mafic lower crustal source [52, 84-86]. Furthermore, these adakitic rocks with high K2O (3.81–4.39 and 4.86–5.77 wt.%) contents are consistent with K-rich adakites formed by partial melting of thickened lower crust (Figure 3(c)). Additionally, based on their Sr-Nd isotopic compositions, it is highly probable that the Taiyanghe and Maoergai adakitic rocks were formed by partial melting of ancient mafic lower crust.Apatite from purely crust-derived felsic granitoids with peraluminous signature typically exhibits lower Ce, higher Y contents, and lower La/Sm and Th/U ratios compared with those crystallized from mantle-derived I-type granitoid rocks with metaluminous signature [75]. Interestingly, all apatite grains from the Maoergai adakitic granitic rocks consistently display low Ce/Y, La/Sm, as well as Th/U ratios (Figures 10(a) and 10(b)), aligning well with the compositional characteristics of crust-derived felsic granitoids [75], further supporting the origin of the Maoergai adakitic granites from a continental crustal source. Most apatite grains in the Taiyanghe granitic sample exhibit similar Ce/Y, La/Sm, as well as Th/U values to those found in the Taiyanghe dioritic rocks, indicating a genetic relationship between their host rocks (Figures 10(a) and 10(b)). However, as mentioned earlier, it is unlikely that the high SiO2 granites originated from the low SiO2 dioritic rocks by fractionation crystallization. Thus, it is more plausible that the Taiyanghe adakitic granites were likely originated from newly underplated lower crust, possibly represented by the dioritic rocks. This interpretation is further supported by whole-rock Sr-Nd isotopes and apatite Nd isotopic compositions in the subsequent section.Therefore, the aforementioned compelling evidence suggests that the adakitic granites from the Taiyanghe and Maoergai plutons are indeed adakites originated from continental crust, most likely originated from a thickened lower crust.The whole-rock Sr-Nd isotopic compositions, combined with apatite Nd isotopic characteristics, indicate that the Taiyanghe and Maoergai adakitic rocks were possibly originated from ancient magma sources (Figures 8 and 11). Previous studies have suggested the presence of an unexposed crystalline basement in the eastern Songpan-Ganzi orogenic belt, which shares similarities with the Yangtze continental block [18, 91, 92]. The Maoergai adakitic rocks were previously interpreted to be formed by partial melting of this unexposed crystalline basement represented by the Kangding complex [18]. Although the Taiyanghe adakitic granites exhibit elevated 87Sr/86Sr(t) and negative εNd(t) values (Figure 8), plotting into the Kangding complex, it is unlikely that they were formed by partial melting of the unexposed crystalline basement of the Yangtze continental block. Interestingly, both the adakitic granites and dioritic rocks from the Taiyanghe pluton exhibit consistent whole-rock 87Sr/86Sr(t) and εNd(t) as well as apatite εNd(t) ratios, suggesting a possible genetic relationship between them (Figure 8). An alternative explanation is that the Taiyanghe adakitic rocks were likely derived from partial melting of newly underplated lower crust, possibly represented by the dioritic rocks. Deschamps et al. [16] conducted zircon U–Pb dating on a diorite (SiO2 = 53.70 wt.%) from the Taiyanghe pluton with the main population ages ranging from ca. 239 to ca. 202 Ma and suggested that the earliest magmas were produced at ca. 250 Ma on the crust-mantle boundary according to Nd isochrons. The ages are thought to correspond to the long-lasting magmatic activity of the Taiyanghe pluton [16]. Mantle-derived magma continuously underplated the Songpan-Ganzi orogenic belt lower crust, and partial melting of preunderplated lower crust likely produced the Taiyanghe high SiO2 granites, which inherited the isotopic characteristics of the low SiO2 dioritic rock.The depth of partial melting for the formation of the Taiyanghe and Maoergai adakitic granites can be approximately estimated. The Nb/Ta ratios are highly sensitive to pressure changes and commonly used to estimate the depth at which the lower crust melted [93]. Both Nb and Ta elements are mainly controlled by rutile, which appears at around 15 kb [93], a necessary residual mineral phase present in the eclogite residue [93]. Melts generated under high pressure (rutile-bearing) conditions would have more elevated Nb/Ta ratios than those produced under low pressure (rutile-free) conditions. The low HREE and Y contents of the Taiyanghe and Maoergai adakitic granites clearly demonstrate that these adakitic granites were formed in the garnet stability zone (>10 kb). On the other hand, the Taiyanghe granites have significantly higher Nb/Ta ratios than the Maoergai granitic rocks (Figure 12(b)), indicating that the depth of partial melting for the Taiyanghe granitic rocks (>15 kb or 50 km) is greater than that for the Maoergai granites (10, 15 kb or 30–50 km). Previous U–Pb SHRIMP dating showed that the Maoergai adakitic rocks give a crystallization age of ca. 216 Ma [18], while our unpublished zircon U–Pb dating showed that the Taiyanghe granites have a crystallization age of ca. 212.5 Ma, suggesting progressive thickening of continental crust. Particularly, Deschamps et al. [16] suggested the long-lasting mafic magmatic activity of the Taiyanghe pluton range from ca. 230 to ca. 202 Ma and further explained that the earliest mafic magmas were produced at ca. 250 Ma on the crust-mantle boundary according to the Nd isochrons. Thus, mantle-derived magma continuously underplated the Songpan-Ganzi orogenic belt lower crust, and the continental crust underwent progressive thickening.Collectively, the whole-rock geochemistry in conjunction with apatite elemental and Nd isotopic characteristics indicate that the Taiyanghe and Maoergai adakitic granites were formed by partial melting of thickened lower crust. The Taiyanghe adakitic rocks were likely products of partial melting of a recently underplated lower crust, while the Maoergai adakitic rocks originated from an ancient mafic lower crust. The continental crust of the Songpan-Ganzi orogenic belt underwent gradual thickening due to continuous underplating of mantle-derived magma.The Markam pluton is commonly considered as representative of S-type granitoids based on its petrological and geochemical characteristics [16]. It is petrologically supported by the presence of Al-rich minerals such as muscovite and minor tourmaline (Figures 2(g) and 2(h)). Geochemically, the Markam granitoids with high SiO2, Al2O3, and P2O5 and low Fe2O3T and TiO2 contents reflect felsic granitoid compositions (Figures 4(a), 4(b), 4(d), and 4(e)). Meanwhile, most of these granitoid samples with high K2O contents exhibit peraluminous feature with high A/CNK ratios, further indicating a S-type granitoid affinity (Figures 3(b) and 3(c)). Additionally, corundum contents in these granitoids are relatively high according to CIPW normative calculations (C = 1.21–2.82). The petrological and geochemical composition features of the Markam granitoids indicate that they belong to the high-K to shoshonitic, peraluminous S-type granitoid series.In addition to the previously recognized adakitic rocks [18], the Yanggonghai pluton also contains non-adakitic granitic rocks that exhibit high SiO2, Al2O3, P2O5, Y, and Yb contents as well as low Fe2O3, TiO2, Sr contents and Sr/Y and (La/Yb)N ratios, indicating normal felsic granitic compositions (Figures 4(a), 4(b)). Furthermore, these granitic samples belong to high-K calc-alkaline series with high K2O contents (Figure 3(b)) and exhibit high Al saturation index (A/CNK > 1.1; Figure 3(c)). These geochemical characteristics suggest that the Yanggonghai non-adakitic granitic rocks are strongly peraluminous S-type granites. This conclusion is further supported by the presence of muscovite in the Yanggonghai granitic rocks (Figure 2(d)). Additionally, these Yanggonghai granitic rocks also have high corundum contents according to CIPW normative calculations (C = 2.01–2.37; online supplementary Table S1). Therefore, the geochemical and mineral signatures of the Yanggonghai granitic rocks studied here indicate strongly peraluminous S-type granites.The S-type geochemical features of the Markam and Yanggonghai granitoids are further supported by apatite elemental compositions. Apatite grains from these felsic granitoid rocks exhibit low Sr, Sr/Y, and (La/Yb)N ratios, which are similar to those of the host rocks (Figure 9). Additionally, all the apatites fall within the non-adakitic granitoid field on a 10/Eu/Eu*-10Sr-HREY ternary discrimination diagram (Figure 10(d)) [78], and they further plot toward S-type peraluminous magma field on a 10Sr-LREE-10Y ternary discrimination diagram (Figure 10(c)) [77]. Thus, the geochemistry of apatite grains from the Yanggonghai and Markam felsic granitoids also indicates that the host rocks have a possible S-type affinity.Experimental petrological and geochemical studies suggested that the peraluminous S-type granitoids are commonly generated by partial melting of crustal materials involving metagraywacke and metapelite [94-97]. The S-type granitoid samples from the Markam and Yanggonghai plutons feature enrichment in Rb, Th, U, and Pb and Eu negative anomalies, indicating a crustal magma source (Figures 5(f) and 5(j)). Meanwhile, the Markam and Yanggonghai granitoids have high Al2O3/TiO2, Rb/Sr, and Rb/Ba but relatively low CaO/Na2O ratios (Figure 13), which clearly plot close to the sediment-derived end-member [96, 98]. Furthermore, these S-type granitoids with evidently high 87Sr/86Sr(t) and negative εNd(t) ratios almost plot into the Sr-Nd field of the Triassic metasedimentary rocks (Figure 8), clearly reflecting that metasediments were a source for the formation of these S-type granitoids. Thus, the Markam and Yanggonghai granitoids were likely formed by partial melting of crust materials, represented by the Triassic metasedimentary rocks.Apatite grains from the Yanggonghai and Markam granitoids clearly share low Ce and high Y contents and low La/Sm and Th/U ratios, plotting into purely crust-derived felsic granitoids with a peraluminous signature (Figures 10(a) and 10(b)) [75]. In addition, the samples with low Sr, Sr/Y, and (La/Yb)N ratios for both apatite grains and the whole rocks suggest that these granitoid rocks were possibly derived from a sedimentary rock source (Figures 7 and 9), with weathering processes causing the low Sr abundance in the source rocks of S-type granitoid rocks [72, 99]. These apatite grains from both the Yanggonghai and Markam S-type granitoids also have low 143Nd/144Nd(t) and negative εNd(t) ratios, indicating that their host rocks were possibly originated from ancient magma sources (Figure 11), consistent with their whole-rock isotopic interpretation. Therefore, the integrated element and Nd isotope compositions of apatite grains in these S-type granitoids reflect that their host rocks were possibly derived from partial melting of the ancient continental crust materials, possibly represented by the Triassic metasedimentary rocks.The initial opening of the Paleo-Tethys Ocean is widely accepted as the Devonian-Carboniferous, which has been attributed to the detachment of the east Asia continental complex (e.g., North China, Tarim, South China) from the eastern margin of the Gondwana supercontinent as extended collage of terranes or continental slivers [10, 100, 101]. The latest studies revealed that the closure of the Paleo-Tethys Ocean was likely diachronic [10]. The final closure of the eastern branch of the Paleo-Tethys Ocean has been commonly considered to have happened in the Middle-Late Triassic, leading to the final collision of the South-North China Cratons [102-104]. Although a twofold-subduction system of the western branch of the Paleo-Tethys Ocean plate beneath the Kunlun and Qiangtang terranes has been well constrained [12, 20, 28, 105], there is still no consistence regarding the disputed issue of the final closure time of the western branch of the Paleo-Tethys Ocean due to two contrasting point of views: (1) subduction of the Paleo-Tethys Ocean plate could have lasted until the Late Triassic-Early Jurassic according to the time span of arc magmatism happened to both the eastern Kunlun and Yidun arcs [[20, 105-116]] and (2) the Late Triassic abundant intrusive and volcanic activities in the eastern Kunlun and Yidun arcs were formed in a postcollision setting, indicating that the subduction of the Paleo-Tethys Ocean plate terminated in this period [117-120]. In this regard, a twofold-subduction system has been suggested to reasonably account for the final closure of the Paleo-Tethys Ocean between the Qiangtang and Kunlun terranes [12, 24, 28]. The double subduction system is also regarded to induce strong folding and thrusting of the Triassic metasedimentary cover in the Songpan-Ganzi orogenic belt [121].Although the double subduction system of the Paleo-Tethys Ocean plate is well rebuilt in the Triassic, the geodynamic background related to the widespread magmatic activity with ages in the range of ca. 231–200 Ma occurring in the Songpan-Ganzi orogenic belt is still controversial [12, 16-22, 122]. These magmatic rocks share a large category of petrologic and geochemical features including high-K calc-alkaline, with I-type signatures, with or without adakitic affinity, peraluminous S-type granitoids, as well as high-K alkaline A-type granitoids [12, 16-22, 122]. These Late Triassic intrusive and volcanic rocks have been previously attributed to be formed in a postcollisional regime [18, 19, 21, 22]. Particularly, a key factor led previous studies to demonstrate that the Late Triassic A-type granitoids and crust-derived adakitic rocks exposed in the Songpan-Ganzi orogenic belt were generated at a postcollisional setting in this region [18, 19, 21, 22, 29]. In a postcollisional regime, the adakitic and alkaline A-type of magmatism have been commonly considered to be derived from partial melting of the thickened or delaminated lower crust of the Songpan-Ganzi orogenic belt, with or without contributions from mantle materials [18, 19, 21, 22, 29].The formation of the Late Triassic magmatic rocks exposed in the Songpan-Ganzi orogenic belt formed at a postcollisional background is a plausible explanation; however, this interpretation was recently considered to be challenged [12, 16, 20, 28, 122]. This is because an increasing body of evidence suggests a new understanding of the subduction history of the Paleo-Tethys Ocean plate and demonstrates that it was not only subducted to the south under the Yidun and Yushu arcs but also had a history of north subduction beneath the Songpan-Ganzi orogenic belt since the Late Triassic [16, 20, 24, 28, 105, 122]. The Late Triassic plutonic rocks in the Songpan-Ganzi orogenic belt were thus considered to be unlikely formed in a postcollision environment in the recent years [16, 28, 122]. In contrast, they were recently regarded to be formed by the subduction of the Paleo-Tethys Ocean plate to the north beneath the Songpan-Ganzi orogenic belt according to different types of evidence [12, 16, 28, 122]. First of all, the Taiyanghe dioritic rocks in this study are characterized by the enrichment in LILEs (Rb, U, K, and Pb) and depletion in HFSEs (Nb, Ta, and Ti) and calc-alkaline major element compositions, indicative of typical of arc-related magmatism (Figure 5(b)). Second, some Late Triassic high-K calc-alkaline intrusive rocks with geochemical characteristics are similar to those of arc-related magmatic rocks [16, 28, 122]. These intrusive rocks with enrichment in LILEs (Rb, Ba, and Sr) and depletion in HFSEs (Nb, Ta, and Ti) were suggested to be formed by partial melting of the mantle wedge metasomatized by slab-derived fluid or melt phases [16, 20, 122]. Third, the range of isotope ages for the Songpan-Ganzi orogenic belt plutonic rocks is similar to arc-related magmatism with the youngest ages both above the southern subduction region in the Yushu and Yidun arcs (ca. 263–203 Ma) and above the northern subduction region in the eastern Kunlun terrane (ca. 225–200 Ma) [20, 106, 123-126]. Furthermore, previous studies suggested that the compiled zircon Hf and whole-rock Sr-Nd isotopic compositions of the Songpan-Ganzi orogenic belt intrusive rocks are similar to those of the Triassic Yidun and Kunlun plutonic rocks, which thus indicated that the Songpan-Ganzi orogenic belt magmatic activity was produced at the continental or island arc [20]. Finally, a great deal of field investigations combined with 40Ar/39Ar age for detrital muscovite and U-Pb age for detrital zircon from the Triassic sedimentary cover reflect that the continuous deposition from the middle to the Late Triassic had not been interrupted [127-129], indicating there is no arc-continent or continent-continent event occurring around the Songpan-Ganzi orogenic belt until the end of the Triassic [23].Therefore, based on the above important evidence, the Paleo-Tethys Ocean plate still existed and still subducted under the Yidun arc and Songpan-Ganzi orogenic belt during the Late Triassic. The final closure time of the Paleo-Tethys Ocean is preferred to be in the Early Jurassic or end of the Late Triassic, which is further supported by an apparent change from marine to nonmarine deposition [130], by the existence of Early Jurassic marine sedimentation strata in the western Songpan-Ganzi orogenic belt [131], and by peak metamorphism and deformation with the Early Jurassic age occurring in both the Jinshajiang suture and the Songpan-Ganzi orogenic belt complex [30, 109, 132].Due to the numerous plutons with main ages ranging from ca. 231 to ca. 200 Ma, the Songpan-Ganzi orogenic belt has a long-term magmatic activity with ca. 30 Ma, which is similar to the Paleogene-Eocene plutonic magmatism in the Gangdese belt [8, 133, 134]. The Paleogene-Eocene intrusive rocks with a large variety of geochemical and petrologic features in the Gangdese belt involve a variety of sources including thickened lower crust, metasomatized mantle, and the ancient Indian continental crustal materials [133-135]. The formation of these magmatic rocks was commonly regarded to be inevitably related to the last stage of closure of the Neo-Tethys Ocean including slab roll-back and break-off processes [8]. The Late Triassic intrusive rocks with a large variety of petrologic and geochemical characteristics in the Songpan-Ganzi orogenic belt also include a variety of magmatic sources (sedimentary cover, lower crust, and metasomatized mantle) [12]. Importantly, these magmatic rocks were recently considered to be generated at the latest stage of closure of the Paleo-Tethys Ocean [12, 20, 28]. The presence of the representative plutonic rocks examined in this study was observed at this critical stage. Thus, we suggest that roll-back and then break-off of the subduction of the Paleo-Tethys Ocean plate could provide a reasonable mechanism for explaining the diverse petrological and geochemical characteristics of the widespread Late Triassic magmatic rocks in the Songpan-Ganzi orogenic belt [12, 20, 28].As mentioned above, the Songpan-Ganzi orogenic belt has a long-term magmatic activity (ca. 231–200 Ma) with ca. 30 Ma. The early stage of magmatism in the Songpan-Ganzi orogenic belt is dominated by adakitic rocks (including Maoergai adakitic rocks) and minor I- and S-type granitoid and mafic rocks (ca. 231–216 Ma) [16, 18, 20-22, 122]. These rocks were closely associated with roll-back of the Paleo-Tethys Ocean plate (Figure 14(a)). The involved slab roll-back induced rise of the hot asthenosphere, which provided enough convective heat to melt the Paleo-Tethys Ocean plate, thickened the mafic lower crust, and metasomatized the lithospheric mantle and sediment cover (Figure 14(a)), producing high-K calc-alkaline magma, with or without adakitic rocks and S-type granitoid magma [12, 17, 20, 28]. Slab roll-back of the Paleo-Tethys Ocean is also supported by recent provenance studies conducted on the Triassic marine depositions from the Songpan-Ganzi orogenic belt and adjacent regions [131, 136]. These studies suggest that depositions originally originated from proximal continents or terranes and sedimented into a remnant of the Paleo-Tethys Ocean prior to roll-back, and then, sediments from distant continent sources were deposited, possibly induced by roll-back of the Paleo-Tethys Ocean slab [131, 136].The roll-back subduction-related magmatic activity was followed by the younger peak magmatism (ca. 216–200 Ma) [12, 17, 20, 22, 28]. These rocks feature low SiO2 and high MgO and Mg#, indicating more mantle-derived composition contributions. According to the magmatic “flare up” (ca. 216–200 Ma) in the Songpan-Ganzi orogenic belt [12, 16, 20, 22, 24, 28], this peak magmatic activity was likely triggered by break-off of the Paleo-Tethys Ocean plate (Figure 14(b)). Slab detachment of the denser part of the Paleo-Tethys Ocean plate in the hot asthenospheric mantle form a gap window, and subsequent sinking of the plate causes hot asthenosphere upwelling giving rise to a quick increase in temperature in the overlying materials and results in melting of metasomatized lithospheric mantle, newly underplated lower crust, and sedimentary rocks. Noticeably, a dominant magmatic “flare-up” at ca. 216 Ma in the Yidun arc was regarded to be caused by slab break-off of the Paleo-Tethys Ocean [115, 125, 137], giving a critical evidence for syncollisional magmatism.The break-off of the Paleo-Tethys Ocean plate also caused the formation of the Nianbaoyeche A-type granites (ca. 211 Ma) in the Songpan-Ganzi orogenic belt [19]. A-type granitoids are generally regarded as a special category of granitoid rocks, which generally occur in postorogenic or intraplate geodynamic backgrounds [138-140]. The Late Triassic A-type granitoids exposed in the Songpan-Ganzi orogenic belt possibly indicate that the Paleo-Tethys Ocean has already closed during this period. However, recent studies have indicated that A-type granitoids could be formed in various dynamic backgrounds throughout geological time, encompassing both within-plate backgrounds and plate boundaries, rather than being limited to anorogenic or rifting tectonic backgrounds [140-142]. For example, the Carboniferous Baiyanghe A-type granite porphyry in the western Junggar was generated at a back-arc tectonic setting related to the subduction of the Irtysh-Zaysan oceanic lithosphere under the Zharma-Saur arc [143]. Thus, the Late Triassic A-type granitoids in the Songpan-Ganzi orogenic belt were likely produced at an extensional back-arc tectonic setting induced by the break-off of the northern subduction of the Paleo-Tethys oceanic lithosphere under the Songpan-Ganzi orogenic belt.To summarize, an integrated model for explaining the formation of the Late Triassic plutonic rocks with a large variety of petrologic and geochemical features exposed in the Songpan-Ganzi orogenic belt is probably that: (1) ca. 231–216 Ma magmatism was caused by the Paleo-Tethys Ocean slab roll-back (Figure 14(a)); (2) the Paleo-Tethys Ocean slab break-off at ca. 216–200 Ma triggered the widely partial melting of both mantle and crust materials (Figure 14(b)); and (3) the final closure of the Paleo-Tethys-Ocean took place at the end of the Late Triassic or the Early Jurassic (Figure 14(c)).The Taiyanghe dioritic rocks were probably generated by low degrees of partial melting of a lithospheric mantle metasomatized by sediment-derived melts.The Taiyanghe and Maoergai granites with adakitic geochemical characteristics were likely derived from partial melting of the thickened lower crust based on the whole-rock and apatite geochemistry. The Taiyanghe adakitic rocks originated from the partial melting of a newly underplated lower crust, possibly represented by the dioritic rocks, whereas the Maoergai adakitic rocks were derived from partial melting of the ancient lower crust.The Markam and Yanggonghai felsic granitoids likely originated from the Triassic metasedimentary source.The diversity of intrusive rocks exposed with the eastern Songpan-Ganzi orogenic belt was likely formed by the roll-back and then break-off of a subducted Paleo-Tethys plate. The final closure of the Paleo-Tethys-Ocean occurred at the end of the Late Triassic or the Early Jurassic.The Taiyanghe dioritic rocks were probably generated by low degrees of partial melting of a lithospheric mantle metasomatized by sediment-derived melts.The Taiyanghe and Maoergai granites with adakitic geochemical characteristics were likely derived from partial melting of the thickened lower crust based on the whole-rock and apatite geochemistry. The Taiyanghe adakitic rocks originated from the partial melting of a newly underplated lower crust, possibly represented by the dioritic rocks, whereas the Maoergai adakitic rocks were derived from partial melting of the ancient lower crust.The Markam and Yanggonghai felsic granitoids likely originated from the Triassic metasedimentary source.The diversity of intrusive rocks exposed with the eastern Songpan-Ganzi orogenic belt was likely formed by the roll-back and then break-off of a subducted Paleo-Tethys plate. The final closure of the Paleo-Tethys-Ocean occurred at the end of the Late Triassic or the Early Jurassic.The data used in this study are available in the supplementary files.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Financial support for this study was provided by the National Natural Science Foundation of China (Nos. 91955203 and 92162211), the Nanjing University Excellence Initiative and the Geological Survey of China (DD20230340). We thank the reviewers and editor for their useful comments and handling of the manuscript, respectively.