Episodic magmatism of the Gongga batholith (eastern Tibet) revealed by detrital zircon U-Pb geochronology: Insights into phased Xianshuihe fault activity and plateau growth Open Access

Understanding the onset and periods of magmatism in a specific location is crucial for understanding crustal extension, tectonic history, and geodynamic processes (Wang et al., 2010; Swanson-Hysell et al., 2014; Li et al., 2017; Wu et al., 2023). Geologic investigation by collecting bedrock samples for geo-chronological study is a standard method for determining the ages of intrusions (Li and Zhang, 2013; Searle et al., 2016; Roberts and Searle, 2019; Chen et al., 2020; Wu et al., 2021). Geological surveys in sparsely populated regions are significantly limited by various natural environmental factors, which hinders a comprehensive understanding of the geological history of these areas. Detrital zircons obtained from these regions offer valuable insights, as their U-Pb ages indicate the timing of magmatic intrusion and provide a natural sampling of the drainage basins. Theoretically, the timing and episodes of magmatism in the drainage area of modern river sands can be comprehensively revealed by detrital zircon U-Pb dating, which is routinely used to perform provenance analysis of sedimentary rocks (Liu et al., 2012; Gehrels and Pecha, 2014; Sun et al., 2020; Ma et al., 2023), but only a few case studies have explored this in the batholith.

The Gongga batholith, one of the largest intrusions in the eastern margin of the Tibetan Plateau (Zhang et al., 2017), is distributed along the curved section of the Xianshuihe fault (Fig. 1). The Xianshuihe fault is one of the most active fault systems in southwestern China and exhibits a significant left-lateral slip rate (Bai et al., 2018; Chevalier et al., 2018), yet its evolutionary history is still a subject of debate (Li and Zhang, 2013; Zhang et al., 2017; Lai and Zhao, 2018; Chen et al., 2020; Wu et al., 2021). Previous studies suggested that activity of the Xianshuihe fault and Cenozoic magmatism of the Gongga batholith were linked and reinforced one another (Chen et al., 2020). Additionally, it was proposed that the Gongga batholith’s magmatic history could shed light on the mechanisms and growth processes that underlie the eastern Tibetan Plateau’s boundary (Hu et al., 2022; Jiang et al., 2022). The magmatic evolution of the Gongga batholith has significant effects on both the extension of the plateau and fault activity (Searle et al., 2016; Roberts and Searle, 2019). Because of the steep terrain and the substantial portion of the batholith covered by glaciers or vegetation, the complete spatio-temporal pattern of magmatic activity in the Gongga batholith has long been ambiguous (Figs. 1B and 2B; Li and Zhang, 2013; Searle et al., 2016).

According to reported zircon U-Pb and mica Ar-Ar dating results from bedrock samples, the Gongga granitoid rocks were intruded during the Late Triassic and the Pliocene (Li and Zhang, 2013; Roberts and Searle, 2019; Chen et al., 2020; Wu et al., 2021; Hu et al., 2022). Among them, the genesis of Cenozoic granites is generally considered to be related to the collision of India and the Eurasian continent. Roberts and Searle (2019) proposed that the source of the Neogene (ca. 25–5 Ma) crustal melts in the Gongga area probably was the Neoproterozoic crust equivalent to the adjacent Kangding Complex. More recently, Hu et al. (2022) divided the Cenozoic magmatism into three stages: (1) Lower and middle crustal metamorphic rocks may have partially melted during the Eocene–Oligocene (ca. 50–25 Ma) due to the thickening of the crust and the upwelling of mantle magma; (2) the magma involved more supracrustal meta-sedimentary material with minimal involvement of the mantle during the early to mid-Miocene (ca. 20–10 Ma), and this magmatic activity may have facilitated the crust’s softening and the strike-slip movement of the Xianshuihe fault; and (3) the asthenosphere upwelled during the late Miocene (ca. 10–4 Ma), triggering the melting of metamorphic rocks on the Tibetan Plateau’s eastern margin. However, with scattered bedrock samples, it is difficult to accurately depict all of the stages and characteristics of magmatic intrusion. On the eastern side of the Gongga massif, the entire Gongga batholith is drained by multiple tributaries of the Yala River (Figs. 2 and 3). Detrital zircons from the modern river sands of these tributaries offer an excellent opportunity for identifying magmatism in the drainage basins. Due to the close spatial relationship between the Gongga batholith and the Xianshuihe fault (Fig. 2), the batholith’s magmatic activity has been closely correlated with the periodic activity of this fault. The activity of large-scale strike-slip faults in the Tibetan Plateau, such as the Xianshuihe, Anninghe, and Ailao Shan–Red River faults, is closely related to the regional geodynamic mechanisms (Leloup et al., 1995; Tapponnier et al., 2001). Detrital zircon geochronology makes the Gongga batholith an ideal location for defining the magmatic episodes that occurred during the evolution of the fault system and plateau growth. We discuss the provenance of detrital zircons, summarize the history of magmatism, and explore the relationship among the periods of magmatism in the Gongga batholith and the activity of the Xianshuihe fault.

The Gongga batholith is situated at the junction of the Songpan-Ganzi, Chuandian, and Yangtze blocks (Fig. 2A), which exhibit complex tectonic evolution. The Songpan-Ganzi block split apart from the Yangtze and South China plates during the Early–Middle Devonian (Yang et al., 1995; Zhang et al., 2006; Roger et al., 2010), and the three blocks finally collided and merged during the Late Permian to Triassic, forming a series of complex fold zones along the plate boundaries (Fig. 1B; Roger et al., 2010; Li et al., 2015d; Hu et al., 2017; Xu et al., 2021b).

The Xianshuihe fault is a large active fault system that extends for over 1400 km from the Tibetan Plateau interior (Ganzi) to the Xiaojiang River near the Ailao Shan–Red River fault zone (Zhang et al., 2003). Several major earthquakes occurred along this fault, including the 2010 M_s 7.1 Yushu, 2014 M_s 6.3 Kangding, and 2022 M_s 6.8 Luding earthquakes (Tobita et al., 2011; Jiang et al., 2015; Li et al., 2022; An et al., 2023). The Xianshuihe fault plays a key role in controlling the plateau’s uplift and crustal shortening, and the lateral extrusion of its southeastern margin (Li and Zhang, 2013; Zhang et al., 2003; Chen et al., 2020; Lei et al., 2022). Studies conducted on this fault have revealed a complex and long-lasting deformational history (Tapponnier et al., 2001; Chen et al., 2020; Lei et al., 2022). Previous research on the formation and evolution of the Xianshuihe fault mainly focused on the Gongga batholith (Roger et al., 1995; Li and Zhang, 2013; Searle et al., 2016). Although a number of structural and geochronological analyses have been conducted on the Gongga batholith (Chen et al., 2020), the onset of the fault system’s activity is controversial and has been constrained to ca. 32–27 Ma, prior to ca. 12 Ma, ca. 9 Ma, or ca. 5 Ma (Roger et al., 1995; Li and Zhang, 2013; Searle et al., 2016; Zhang et al., 2017; Yan et al., 2018).

The Gongga batholith has a width of ~15 km and stretches for ~120 km along the main active trace of the Xianshuihe fault zone in a NNW–SSE direction (Figs. 2 and 3). This batholith is traversed by several branches of the Xianshuihe fault zone (Fig. 2; Zhang et al., 2004; Wang et al., 2008; Li and Zhang, 2013). Gongga Mountain (with a peak of ~7756 m) is the highest mountain on the eastern margin of the Tibetan Plateau (Zhang et al., 2017). During the Late Triassic, the Paleo-Tethyan oceanic plate subducted beneath the Tibetan Plateau to form the Gongga batholith (Zhang et al., 2017). During the Cenozoic, the batholith was intruded by granites produced by several magmatic episodes, which resulted in a complex and varied suite of rocks (Fig. 2C; Searle et al., 2016; Roberts and Searle, 2019; Chen et al., 2020) that includes biotite granite, granodiorite, and monzogranite, with local occurrences of porphyry (Searle et al., 2016; Zhang et al., 2017; Roberts and Searle, 2019). The Triassic and Jurassic granites are mainly distributed in the southern and western parts of the batholith, whereas the Cenozoic granites predominately occur in the northern part of the massif (Figs. 2 and 3; Li and Zhang, 2013; Searle et al., 2016; Roberts and Searle, 2019; Chen et al., 2020; Wu et al., 2021).

The main Xianshuihe fault extends along the eastern margin of the Gongga massif and is accompanied by a series of en echelon sinistral faults intersecting the batholith (Fig. 2) that form a classical left-lateral, strike-slip duplex system. Several significant fault splays aligned in a NW–SE direction intersect the granitic batholith (Searle et al., 2016). On the northern side of the Gongga batholith, where Triassic sandstone is exposed, folding and brittle deformation are evident near the main Xianshuihe fault. The folds primarily show upright structures with horizontal stretching lineation, which indicates shear deformation as their origin (Lai and Zhao, 2018; Chen et al., 2020). The intensity of deformation gradually diminishes from the main fault toward the periphery. Field observations indicate that the Xianshuihe fault displays a ductile fabric that, to the north of Kangding, is overprinted by the formation of brittle gouge zones (Figs. 2 and 3; Li et al., 2015a; Searle et al., 2016). The fault diverges from the Gongga batholith south of Kangding, extending along metamorphosed sedimentary rocks and the Proterozoic Kangding Complex (Searle et al., 2016; Chen et al., 2020).

The lithology of Late Triassic granites mainly comprises granodiorite and biotite plagiogranite (Searle et al., 2016; Chen et al., 2020). The Middle Jurassic granites mainly consist of porphyritic, coarse-grained monzogranites (Li et al., 2015a; Searle et al., 2016). These Middle Jurassic granitic plutons are not widely exposed and align with the NW–SE direction of the Xianshuihe secondary fault (Figs. 2 and 3; Chen et al., 2020). Previous field observations concluded that the emplacement of the Late Triassic and Middle Jurassic granites was not directly related to the formation and evolution of the Xianshuihe fault (Searle et al., 2016; Chen et al., 2020). The Cenozoic migmatitic zone, which is located in the eastern segment of the Gongga batholith (Figs. 2 and 3), originated from two stages of migmatization events at ca. 47–27 Ma and 27–20 Ma, based on zircon U-Pb and monazite Th-U-Pb geochronology (Liu et al., 2006; Li et al., 2016; Searle et al., 2016; Chen et al., 2020). The migmatitic zone is mainly composed of banded gneisses, migmatitic gneisses, and migmatites (Chen et al., 2020). The distribution of the migmatitic zone along the Xianshuihe fault (Figs. 2 and 3) indicates a spatially close connection with the fault. The first stage of migmatization exhibits strong flattening of microstructures, which is reflected in intensive deformation during the Eocene–Oligocene (Li and Zhang, 2013). Field investigation and microstructural study show a syn-shearing anatexis of the second-stage migmatization, with intensively developed horizontal lineations, which indicates a sinistral strike-slip regime of the Xianshuihe fault (Li et al., 2016). Typical eyeball-like structures of the palaeosome and leucosome also indicate that the lineations may be the result of ductile deformation (Chen et al., 2020). The granites of the Miocene are dominated by medium- and fine-grained monzogranites. Zhang et al. (2004) also reported the ⁴⁰Ar/³⁹Ar ages of mica and K-feldspar to be ca. 10 Ma and ca. 12 Ma, which reveals that a thermal event occurred during the sinistral shear of the Xianshuihe fault. Roger et al. (1995) suggested that the granites of this period were synkinematically emplaced with the Xianshuihe fault, as the minimum age of the deformation is within error of the emplacement age of the granite (11.6 ± 0.4 Ma and 12.8 ± 1.4 Ma). The shallowing of the magmatic source of the granites emplaced along the Xianshuihe fault during the Miocene, as suggested by the gradual decrease in Sr/Y ratios, has been linked to the change in the fault’s kinematics from compression to strike-slip (Lai and Zhao, 2018).

To attain a more comprehensive, time-integrated record of magmatic activity in the Gongga batholith, eight samples (GG21-1 to GG22-8) were collected from modern glacial and river sediments within the Gongga batholith (Fig. 4). Except for sample GG22-8, from the trunk Yala River, all samples are from small catchments that locally drain the glacier-capped massifs.

Zircon grains were separated from detrital samples after crushing, using conventional magnetic and heavy-liquid separation procedures. The experimentation and subsequent data processing were conducted at the National Institute of Natural Hazards, Ministry of Emergency Management of China, Beijing, China. The laser stripping spot diameter was 30 μm, the frequency was 8 Hz, and the energy density was 4 J/cm². The backgrounds of ²⁰⁴Pb and ²⁰²Hg were more than 100 cps due to the use of high-purity nitrogen and helium as carrier gases. During the zircon U-Pb dating process, data collection by inductively coupled plasma–mass spectrometry (ICP-MS) adopted a single-point peak-hopping mode, with the dwell times set at 20 ms (²⁰⁴Pb, ²⁰⁶Pb, ²⁰⁸Pb, ²³⁸U, and ²³²Th) and 50 ms (²⁰⁷Pb), respectively. The operating conditions for laser ablation (LA)-ICP-MS are shown in Table S1 in the Supplemental Material.¹ Cathodoluminescence (CL) imaging was performed by the MIRA3 field emission–scanning electron microscope with a scanning time of 2 min and a voltage of 7kV, and clear CL pictures of the annular bands were obtained. Zircon crystals with no surface fractures and distinct internal annular bands were then selected as potential U-Pb dating points using transmitted light, reflected light, and CL images of zircon (Fig. 5).

Dynamic elemental fractionation, static elemental fractionation, and instrument sensitivity drift were calibrated prior to age determination. The Plešovice zircon standard (Sláma et al., 2008) with a ²⁰⁶Pb/²³⁸U age of 337.13 ± 0.37 Ma (2 standard deviation [SD]) was used to correct for instrumental mass bias and isotopic fractionation of Pb and U. The Harvard zircon standard 91500, a widely used international reference zircon with a ²⁰⁶Pb/²³⁸U age of 1062.4 ± 0.8 Ma (2 SD; Wiedenbeck et al., 1995), was also used. We repeated measurements on the two reference zircons over a six-day period, which yielded mean ²⁰⁶Pb/²³⁸U ages of 342 ± 30 Ma (2 SD) and 1062 ± 69 Ma (2 SD) for Plešovice and 91500, respectively. The selection of zircon sample signals, U-Th-Pb isotopic ratios, and age calculations was conducted using iolite software (Hellstrom et al., 2008; Paton et al., 2011), and IsoplotR (Vermeesch, 2018) was utilized for probability density plots and peak ages.

A total of 1913 detrital zircon grains were analyzed by U-Pb dating. Most of these grains exhibit oscillatory zoning, low-luminescence, and high-Th/U ratios (>0.1), which are indicative of a magmatic origin (Corfu et al., 2003). Some zircons, however, display overgrown edges and low-Th/U ratios (<0.1), which suggests modification by later metamorphic processes (Corfu et al., 2003). We excluded these zircons from subsequent analyses. Furthermore, zircons exhibiting a discordance of >10% were not considered in the interpretation. After eliminating zircons of metamorphic origin and high-age discordance, 1429 detrital zircon grains were selected to reveal the magmatism within the Gongga batholith. Summarized data are presented in Table S2 (see ^{footnote 1}). Probability density plots of zircon U-Pb age distributions are shown in Figures 6 and 7.

Zircon age distributions of the three samples (GG21-1, GG21-2, and GG21-3) collected from Hailuogou and Yanzigou catchments, near the main peak of Gongga Mountain, exhibit similar age peaks at ca. 200 Ma. In sample GG21-2, a few zircon grains date back to the Neoproterozoic, Paleozoic, and Late Jurassic, while the majority of zircons in all three samples were dated to the Late Triassic to Early Jurassic.

The sample from the outlet of the Nanmenguangou catchment (GG21-4) yielded 192 zircon U-Pb ages ranging from Neoproterozoic to Neogene. In the probability density plot, the major age group is Jurassic (39.6%), and two secondary age groups are Triassic (26.6%) and Cenozoic (23.0%), respectively. Some zircon grains with Neoproterozoic and early Paleozoic U-Pb ages show residual cores and typical external oscillatory zones in CL images, which is compatible with the properties of inherited zircon. Sample GG21-4 exhibits several age peaks, including a major age peak at ca. 230–200 Ma and minor age peaks at ca. 200–180 Ma, ca. 180–160 Ma, ca. 155–135 Ma, ca. 50–25 Ma, and ca. 25–15 Ma.

Sample GG22-5, which was collected from the outlet of the Wangjiagou catchment, yielded 213 zircon U-Pb ages spanning from the Neoproterozoic to Neogene. In the probability density plot, a major group is Late Triassic in age (84.5%), and the secondary groups have Cenozoic (6.6%) and Jurassic ages (5.2%), respectively. Sample GG22-5 has a major age peak at ca. 230–200 Ma, with no identifiable minor age peaks.

The sample from the outlet of the Zheduo River (GG22-6) yielded 265 zircon U-Pb ages ranging from the Archeozoic to Neogene. In the probability density plot, a major group is Cenozoic in age (76.2%), and there are no secondary groups. Sample GG22-6 has a major age peak at ca. 25–15 Ma, with no identifiable minor age peaks.

Sample GG22-7, which was collected from the outlet of the Dalongbugou catchment, yielded 178 zircon U-Pb ages spanning from the Neoproterozoic to Neogene. In the probability density plot, a major group is Cenozoic in age (39.9%), and the secondary groups have early Paleozoic (26.4%) and Jurassic ages (16.3%), respectively. Sample GG22-7 exhibits several age peaks, including a major peak at ca. 50–25 Ma and minor peaks at ca. 550–500 Ma, ca. 230–200 Ma, ca. 180–160 Ma, and ca. 25–5 Ma.

Sample GG22-8, which was collected from the outlet of the trunk Yala River (Shenshugou), yielded 211 zircon ages that range from Archeozoic to Neogene. In the probability density plot, a major group is Paleoproterozoic in age (24.2%), and the secondary groups have Neoproterozoic (19.0%), Triassic (18.5%), and Cenozoic ages (16.1%), respectively. Sample GG22-8 exhibits two major age peaks at ca. 230–200 Ma and ca. 20–5 Ma, with no other discernible peaks.

Detrital zircon U-Pb ages from the rivers draining the Gongga batholith display multiple distinct peaks. We exclusively discuss the post-Triassic magmatism in the Gongga batholith since the zircon U-Pb data we obtained mainly comprise Triassic–Cenozoic ages. The scattered pre-Triassic U-Pb ages likely represent the ages of inherited zircons.

We incorporated all detrital zircon U-Pb ages into one probability density plot (Fig. 7). The ages are grouped into ca. 230–200 Ma, ca. 200–180 Ma, ca. 180–160 Ma, ca. 50–25 Ma, and ca. 25–5 Ma. The probability density plot of the samples collected in the Hailuogou, Yanzigou, and Wangjiagou catchments and the trunk Yala River all show a notable peak between ca. 230 Ma and ca. 200 Ma. Another Early Jurassic age peak (ca. 200 Ma to ca. 180 Ma) is evident in the probability density plot of the sample from the Hailuogou catchment. The probability density plot of the sample collected in the Nanmenguangou catchment stands out for possessing all age peaks. The probability density plot of the Dalongbugou catchment’s sample shows a zircon U-Pb age peak between ca. 50 Ma and ca. 25 Ma. Within the ca. 25–5 Ma period, the probability density plots of the samples from the trunk Yala River and the outlet of the Zheduo River to Dalongbugou both display a noteworthy peak. Further, within the <50 m.y. range, the number of zircon grains whose ages lie between ca. 50 Ma and ca. 25 Ma is significantly less than those aged between ca. 25 Ma and ca. 5 Ma, and the height of the age peak is also markedly lower. Notably, the Nanmenguangou catchment, Dalongbugou catchment, and the trunk of the Yala River all have older age peaks (>250 Ma). We suggest that the old zircon ages in those samples indicate the presence of inherited zircon cores. On the eastern side of the trunk Yala River is the adjacent Kangding Neoproterozoic complex (Figs. 2C and 3), and the ancient zircons in this sample also may have originated from this Neoproterozoic complex.

U-Pb ages of detrital zircons provide information on their provenance and origin. We used ArcGIS to extract the watershed basins of the rivers studied and identify potential sources of detrital zircons to determine the provenance of the zircons within the Gongga batholith (Fig. 2B). Next, we reviewed published U-Pb zircon data from bedrock samples in the possible provenance (Table S3, see ^{footnote 1}; Fig. 7). The majority of the zircons come from granitic rocks, and 91.6% of the bedrock age data are Triassic–Cenozoic. Like our detrital zircon sample U-Pb age groups, these ages are categorized into ca. 230–200 Ma, ca. 200–180 Ma, ca. 180–160 Ma, ca. 65–45 Ma, ca. 45–25 Ma, ca. 25–5 Ma, and ca. 5 Ma–present (Fig. 7). However, there are no significant age peaks in the pre-Triassic bedrock sample data, so it cannot be compared with our detrital data. Consequently, it is most likely that the post-Triassic detrital zircons are from the metamorphic basement intruded by the Gongga batholith.

Several significant discrepancies were identified after a thorough comparison of the published bedrock U-Pb data and the U-Pb ages of detrital zircons collected from several catchments draining the Gongga batholith (Figs. 2C and 3). We visualized previously published data and our data using a scatter plot (Fig. 8). In the Hailuogou catchment (samples GG21-1 and GG21-2), although the bedrock sample zircon U-Pb ages are dominantly Late Triassic, which has led to the prior designation of Late Triassic magmatic intrusions (Li and Zhang, 2013; Searle et al., 2016; Roberts and Searle, 2019; Chen et al., 2020), the zircon U-Pb ages of these two detrital samples reveal a dominance of Early Jurassic zircons with proportions of 70.5% and 58.2%, respectively. Hence, we consider that these two periods of magmatic activity may correspond to different geological events and should be described separately. In the Nanmenguangou catchment (GG21-4), three detrital zircon U-Pb ages peaks have been identified in the Jurassic (with a range of ca. 200–130 Ma), and one has been identified in the Eocene–Oligocene (with a range of ca. 50–25 Ma), which were not reported from the bedrock U-Pb data (Fig. 8). This could be due to the fact that the Jurassic and Eocene–Oligocene intrusions have been obscured by glaciers or flora, which limited bedrock sampling (Fig. 2B). Detrital sample GG22-5, from the outlet of the Wangjiagou catchment, shows an age peak in the Late Triassic (with a range of ca. 220–200 Ma; Fig. 6), which was not reported from the bedrock U-Pb data (Fig. 8). The zircon U-Pb age data from the detrital samples in the trunk Yala River catchment (GG22-8) show a distinct Late Triassic peak at ca. 230–200 Ma that is not reflected in the bedrock data, which indicates that a large portion of the Late Triassic magmatic intrusions in the northern part of the Gongga batholith have not been discovered.

Based on the U-Pb age probability plots of the detrital zircons, we separate the Late Triassic–Early Jurassic magmatism into two periods (Figs. 6 and 7). The subduction of the Paleo-Tethys oceanic plate beneath the Tibetan Plateau (i.e., the Indosinian orogeny) induced the Late Triassic magmatism that formed the Gongga batholith (Zhang et al., 2017). Li et al. (2015a) discovered that the Gongga batholith contains many Jurassic leucogranites that exhibit localized reticulation. Searle et al. (2016) proposed that the leucogranites were formed in the late Indosinian by partial melting of continental crust and then metamorphosed by tectono-thermal events during the Middle Jurassic (Chen et al., 2020). Leucogranites are typically regarded as the result of late-stage processes in continental (subduction) collisions, occurring within localized zones of high-temperature, low-pressure metamorphism within subducted continental crust (Guo and Li, 2007, 2009). According to earlier research, Jurassic A-type granite is distributed extensively throughout the Gongga batholith (Wu et al., 2021), and it is thought that the formation of A-type granite is connected to post-collisional extensional settings (Shen et al., 2011). Thus, we propose that postcollisional extension subsequent to the closure of the Paleo-Tethys Ocean may have caused the magmatic activity during this time (Fig. 9). Furthermore, this conclusion is well supported by research conducted in other parts of the southeastern Tibetan Plateau, including the Eastern Songpan-Ganzi accretionary-orogenic wedge, Nianbaoyeche, and Wulaxi (Zhang et al., 2007; de Sigoyer et al., 2014; Zhou et al., 2014).

Based on the detrital zircon U-Pb age probability plots (Figs. 6 and 7), we divide the Cenozoic epoch into two stages. The first stage of magmatism, which occurred during ca. 50–25 Ma, shows a relatively small amount of age data (22.9% of the Cenozoic sample) and a small age peak on the overall age probability plots (Fig. 7). The Xianshuihe fault zone underwent intense migmatization during ca. 47–27 Ma, which resulted in formation of the migmatitic zone that outcrops along the fault (Figs. 2C and 3; Li and Zhang, 2013). Recent comprehensive reviews of published bedrock U-Pb and ⁴⁰Ar-³⁹Ar data from the Gongga batholith indicate that the region underwent crustal shortening from ca. 47 Ma to ca. 27 Ma (Chen et al., 2020). The microstructures of the first-stage migmatization (ca. 47–27 Ma) also reflect the compressional deformation along the Xianshuihe fault (Li et al., 2016). This is consistent with the timing of early Cenozoic magmatic activity that we obtained from detrital U-Pb zircon age data (Fig. 7). Chen et al. (2020) posited that the migmatization indicates the timing of the initiation of the Xianshuihe fault. Therefore, based on our results, those of previous studies, and the record of magmatism, we propose that the ca. 50–25 Ma magmatic stage may indicate a particular stage of activity on the Xianshuihe fault (Fig. 7; Chen et al., 2020; Li et al., 2015a; Searle et al., 2016).

The second stage of magmatism, which took place between ca. 25 Ma and ca. 5 Ma, has a higher concentration of age data (74.4% of the Cenozoic sample) than the first stage. A higher peak was observed in the overall age probability plots during this stage, as well (Fig. 7). According to previous research, the Xianshuihe fault was quite active during this period, and its movement resulted in the occurrence of large-scale intrusion in the region’s crust (Chen et al., 2020). Based on thermochronological and paleomagnetic data, the intense counterclockwise rotation of the Sichuan basin initiated in the Miocene, and the intense activity of the Xianshuihe fault zone on its western side should have occurred no later than this period (Wang et al., 2014; Tong et al., 2019). Furthermore, magnetostratigraphic study in the Xiaolongtan basin located at the southern tip of the Xianshuihe-Xiaojiang fault indicates that the activation of the southern segment of the fault system initiated at ca. 12.7 Ma (Li et al., 2015c). Field investigation and microstructural studies show a syn-shearing anatexis of the second-stage migmatization (ca. 27–20 Ma; Li et al., 2016; Chen et al., 2020). Roger et al. (1995) also suggested that the granites of this period were syn-kinematically emplaced with the Xianshuihe fault. Neogene granite isotopic data from the Gongga batholith suggest that magmatism may have softened the crust and promoted Xianshuihe fault activity during ca. 20–10 Ma (Hu et al., 2022). There, we suggest that the ca. 25–5 Ma magmatic stage is linked to another stage of activity on the Xianshuihe fault, and the activity on the fault is significantly more intense than that of the previous stage, as indicated by the higher age peak in the overall age probability density plots (Fig. 7) and the shift from compression to strike-slip along the fault during this period (Lai and Zhao, 2018).

Our detrital zircon U-Pb results unveil two distinct stages of Cenozoic magmatism in the eastern Tibetan Plateau, which are likely associated with the two discernible phases of Xianshuihe fault activity. We propose that these two magmatic stages in the eastern Tibetan Plateau coincide temporally with two magmatic-tectonic phases (Tapponnier et al., 2001; Liu-Zeng et al., 2018; Cao et al., 2021, 2022; Spicer et al., 2021) and may represent two major plateau-growth events (Fig. 10).

The Eocene to Oligocene magmatic episode (ca. 50–25 Ma) is generally coeval with the lithospheric mantle-derived alkaline magmatism and carbonatites in the eastern Tibetan Plateau (Chung et al., 1998; Xu et al., 2021a). During this period, the mantle imprinted geochemical characteristics of this episodic magma in the Gongga batholith that indicate active mantle-crust interaction (Hu et al., 2022). During the Eocene to Oligocene, the central and eastern–southeastern parts of the Tibetan Plateau are thought to have experienced substantial crustal thickening and surface uplift that are documented by lines of geological and paleontological evidence (Spicer et al., 2021; Ding et al., 2022). Crustal shortening and thickening are supported by studies of the late Eocene to Oligocene folding and thrusting in the eastern Tibetan Plateau (Li and Zhang, 2013; Cao et al., 2019). Stable isotopes (oxygen and carbon isotopes of carbonates), plant fossils, and (La/Yb) ratios in adakitic rocks (Hoke et al., 2014; Li et al., 2015b; Tang et al., 2017; Su et al., 2019; Xiong et al., 2020; He et al., 2022; Wang et al., 2022) all indicate that the eastern Tibetan Plateau approached high-elevation prior to the Miocene, which implies vigorous tectonic uplift during the early Cenozoic. This conclusion is also supported by the sedimentary records of the marginal seas of Asia, which indicate a sudden increase in sediment flux since the Oligocene (Clift, 2006). Therefore, we suggest that during this period, the crustal thickening, tectonic uplift, and primary plateau-growth pattern were dominated by compression/shortening in the northeastward direction, with no or less material extrusion (Cao et al., 2021, 2022).

The Miocene to Pliocene magmatic episode (ca. 25–5 Ma) is generally simultaneous with the records of deformation and exhumation along the thrust fault and strike-slip fault system in the eastern Tibetan Plateau (Lei et al., 2022). During the Miocene to Pliocene (ca. 25–5 Ma), due to the continuous indentation of the northeastern corner of the Indian plate and the sufficient crustal thickening caused by the plateau growth in the early Cenozoic, significant extrusion of the lithospheric material of the plateau began. The tectonic stress in the eastern Tibetan Plateau gradually underwent a clockwise rotation, which aligns with the structural kinematics of the major fault systems (Fig 1; England and Molnar, 1990; Gao et al., 2017). According to previous paleomagnetic studies, the clockwise rotation of the Indochina block occurred after the late Oligocene (25.7 ± 2.5 Ma; Tong et al., 2021), and the western part of the Chuandian block also underwent a clockwise rotation of ~20° after the middle Miocene (Otofuji et al., 1998; Tong et al., 2015; Wang et al., 2016, 2023; Gao et al., 2017). The principal compressive stress within the eastern Tibetan Plateau may have rotated to the northwest–southeast direction in the late Cenozoic, triggering intense left-lateral slip accompanied by the emplacement of Miocene granites along the Xianshuihe fault (Roger et al., 1995; Zhang et al., 2017). Therefore, we posit that the “lateral extrusion” model may provide a better explanation of plateau deformation and growth in the eastern Tibetan Plateau during the late Cenozoic.

We performed U-Pb geochronology on detrital zircons taken from the modern glacial and river sediments within the Gongga batholith to further constrain the chronology and periods of magmatism in the Gongga batholith located along the Xianshuihe fault at the eastern margin of the Tibetan Plateau. Our latest detrital zircon U-Pb data demonstrate that the Gongga batholith exhibits five main periods of magmatic activity at ca. 230–200 Ma, ca. 200–180 Ma, ca. 180–160 Ma, ca. 50–25 Ma, and ca. 25–5 Ma. The Late Triassic–Early Jurassic magmatism was related to postcollisional extension following the closure of the Paleo-Tethys Ocean. Moreover, the two periods of Cenozoic magmatic activity are coeval with the two main phases of plateau growth in the eastern Tibetan Plateau associated with the gradual intensification of the activity of the Xianshuihe fault.

Constructive reviews by associate editor Francesco Mazzarini, reviewer Wenyong Duan, and another anonymous reviewer are gratefully acknowledged. This study was supported by the National Key Research and Development Program of China (grant no. 2022YFF0800903) and the National Natural Science Foundation of China (grant no. 42073052).