Low-temperature thermochronology constraints on the evolution of the Eastern Kunlun Range, northern Tibetan Plateau Open Access

The deformation history of the northern Tibetan Plateau as a result of the Cenozoic India-Asia collision is key to understand the growth of the plateau, and the overall structural pattern has important implications for deciphering deformation mechanisms within the orogen (Fig. 1A) (e.g., England and Houseman, 1986, 1989; Molnar et al., 1993; Clark and Royden, 2000; Yin and Harrison, 2000; Yin, 2010; Clark, 2012). There are two end-member models for deformation across and the development of the northern plateau: (1) deformation gradually propagating northward from the Himalayan collisional front to the northern plateau margin (e.g., Tapponnier et al., 2001; Clark, 2012; Wang et al., 2014; Yu et al., 2019a, 2019b; Zheng et al., 2017; Wang et al., 2020); (2) early deformation across most of the Cenozoic orogen shortly after initial India-Asia collision, exploiting preexisting weaknesses such as older suture zones and subsequent out-of-sequence propagation of deformation from the Eastern Kunlun Range in the south to the northernmost plateau (e.g., Yin and Harrison, 2000; Yin et al., 2007a; Chen et al., 2019; Li et al., 2019, 2020; Wu et al., 2019a; Zuza et al., 2019; Bian et al., 2020; Chen et al., 2020; Yu et al., 2020). These two models predict different timings of continental deformation, location and kinematics of crustal structures, and cooling and exhumation histories in the northern Tibetan Plateau. Whichever model more adequately quantifies deformation across the plateau can improve our knowledge of orogenic mass balance and broad orogen kinematics (e.g., Yakovlev and Clark, 2014; Ingalls et al., 2016). For example, assumption of a constant bulk strain rate across the orogen allows first-order predictions of crustal shortening across the entire orogen (Fig. 1B) (Zuza et al., 2020).

The spatio-temporal evolution of the Eastern Kunlun Range plays an important role in the above issues, and the following describes three broad cooling and uplift events derived from published thermochronology results (Fig. 1C). ⁴⁰Ar/³⁹Ar thermochronology revealed a range-wide Mesozoic cooling event that was locally overprinted by a Cenozoic cooling event at ca. 30–20 Ma (e.g., Mock et al., 1999; Y. Liu et al., 2005; Wang et al., 2005). Apatite fission-track (AFT) studies suggested that the Eastern Kunlun region experienced rapid and widespread cooling at ca. 20–10 Ma, possibly related to Cenozoic range uplift in response to the Indo-Asian collision (Jolivet et al., 2001; Wang et al., 2004; Y. Liu et al., 2005; Yuan et al., 2006; McRivette et al., 2019; Wu et al., 2019a; Tian et al., 2020). Additional published low-temperature thermochronology results suggested early Cenozoic uplift of the Eastern Kunlun Range (e.g., Clark et al., 2010; Wang et al., 2016, 2017a; Liu et al., 2017a).

The Cenozoic Kunlun transpressional system in the Eastern Kunlun Range is one of the largest intracontinental structures of the northern Tibetan Plateau (e.g., Wang et al., 2011a), and it is associated with reactivation of a preexisting tectonic weakness zone (e.g., Wu et al., 2019a) (Fig. 1C). The transpressional system consists of the main left-slip Kunlun fault, the triangular-shaped, west‐northwest–trending Qimen Tagh, and northwest-trending Bayan Har thrust belts (e.g., Yin et al., 2007b, 2008), the northwest-striking Qinghai Lake right‐slip fault system (e.g., Wang and Burchfiel, 2004; Duvall and Clark, 2010; Yuan et al., 2011, 2013), and the Hoh Xil horsetail transtensional system (e.g., Kidd and Molnar, 1988; Peltzer et al., 1999; Jolivet et al., 2003; Taylor and Yin, 2009; Yin, 2000) (Fig. 1C). The deformation timing of the Kunlun transpressional system is evidenced by growth strata revealed by seismic profiles and surface geologic mapping, and analyses of scattered low-temperature thermochronology samples (Fig. 1C). Miocene‐to‐present strike‐slip activity on the Kunlun fault and the associated strain pattern can be explained by clockwise rotation of the left-slip fault and its wall rock as a bookshelf‐fault system, which has been proposed for the northern plateau as a result of distributed north‐south right‐lateral shear (e.g., Zuza et al., 2017; Wu et al., 2019a, 2020b).

In this contribution, we integrated systematic geomorphological analysis, field observations, and AFT thermochronology along three traverses across the Eastern Kunlun thrust belts to document the deformation-related exhumation of the Eastern Kunlun Range (Figs. 2–4). We focused on AFT analysis together with thermal history modeling of 22 Paleozoic granitoid samples collected in fault-bounded ranges to elucidate the cooling history of this region since the Cretaceous. The data shed light on the basin-range evolution across the northern plateau, which experienced multiple phases of growth.

The Eastern Kunlun Range is bounded by the Cenozoic Qaidam Basin to the north and the Hoh Xil Basin to the south (Fig. 1C). The range consists of the Cenozoic left-slip Kunlun fault and Neoproterozoic–early Mesozoic Kunlun arcs (e.g., Yin and Harrison, 2000; Cowgill et al., 2003; Gehrels et al., 2003; Ding et al., 2013; Wu et al., 2016, 2019a, 2019b; Dong et al., 2018; McRivette et al., 2019). The Kunlun arcs, which intruded Precambrian gneiss and Neoproterozoic to early Mesozoic low-grade metasedimentary rocks, are associated with the evolution and closure of the Kunlun ocean(s) (e.g., Pan et al., 2004; Wu et al., 2016, 2017, 2019a). Proterozoic metamorphic basement is exposed along the eastern segment of the range, where it is thrust over the early Paleozoic arc sequences (e.g., Qinghai BMGR, 1991; Pan et al., 2004; Wu et al., 2019a, 2020a). Unconformably overlying this unit is an upper Paleozoic passive continental margin sedimentary sequence (e.g., Qinghai BMGR, 1991; Yin and Harrison, 2000; Pan et al., 2004; Wu et al., 2019a). Triassic turbidite is widespread along the southern flank of the Kunlun fault, which is covered by late Cenozoic reddish sediments as an angular unconformity (e.g., Qinghai BMGR, 1991; Yin and Harrison, 2000; Pan et al., 2004). The Kunlun fault is mostly exposed along the Triassic suture related to the closure of the Neo-Kunlun ocean (i.e., part of the Paleo-Tethyan orogenic system), which is interpreted as the Cenozoic reactivation of the paleo-tectonic weakness zone (e.g., Yin, 2010; Zuza et al., 2017; Wu et al., 2019a).

Based on the structural styles and their geometric relationship to the Kunlun fault, the wall rocks of the Kunlun fault system may be divided into four portions (e.g., Zuza and Yin, 2016, 2017; Wu et al., 2019a): (1) the triangular-shaped, west-northwest–trending Qimen Tagh thrust belt to the northwest (e.g., Yin et al., 2007b, 2008); (2) the Qinghai Lake right-slip system to the northeast (e.g., Wang and Burchfiel, 2004; cf. Duvall and Clark, 2010; Zuza and Yin, 2016, 2018, 2019; Wu et al., 2019a, 2020b), which includes the right-slip Riyueshan fault and the Wenquan fault; (3) the northwest-trending Bayan Har thrust belt to the southeast (e.g., Yin et al., 2007b, 2008; Duvall et al., 2013); and (4) the Hoh Xil horsetail system to the southwest (e.g., Kidd and Molnar, 1988; Peltzer et al., 1999; Yin, 2000; Jolivet et al., 2003; Taylor and Yin, 2009) (Fig. 1C). The Qimen Tagh and Bayan Har thrust belts are asymmetric with respect to the left-slip Kunlun fault, and their orientations suggest that they have accommodated NNE-SSW–oriented shortening. The Qinghai Lake right-slip system was suggested as a rotation or bookshelf structure within two rotating systems in northern Tibetan Plateau, i.e., the Kunlun and Haiyuan transpressional systems (e.g., Duvall et al., 2013; Zuza et al., 2017, 2019; Wu et al., 2020b). The Hoh Xil horsetail system consists of a series of left-slip faults that branch off from the Kunlun fault and general NNE-striking normal faults (e.g., Jolivet et al., 2003).

Our study focused on three traverses located within the western, central, and eastern domains of the Eastern Kunlun Range, respectively (Figs. 2–4). The northwest-trending Qimen Tagh thrust belt is located at the western domain of the Eastern Kunlun Range, and the structural pattern of this thrust belt was described in Yin et al. (2007b, 2008). Nearly all major faults within the western traverse are south-directed (Fig. 2). Some of the thrusts are linked with east-striking left-slip faults that are sub-parallel to the Kunlun fault or terminate at northeast-striking left-slip faults associated with the Altyn Tagh fault system (e.g., Wu et al., 2020a, 2020b). Structures here include the Yiematan and Caishiling thrusts located south of Yousha Shan anticlinorium (Yin et al., 2007b), the Qimen Tagh imbricate thrusts located south of Gaskule syncline, and the western segments of the Adatan and Ayakum thrusts (e.g., Yin et al., 2007b; Wu et al., 2020b). The Proterozoic–early Paleozoic gneiss and/or schist and Paleozoic granite units were thrust over the Carboniferous and Paleogene sedimentary rocks by the south-directed Qimen Tagh imbricate thrusts (e.g., Yin et al., 2007b) (Figs. 5A and 5B). It is possible that these thrusts were inactive in the Quaternary based on a lack of obvious geomorphic signature (e.g., Wu et al., 2020b). The Adatan thrust projects under a fault-bend fold comprising Pliocene growth strata from the seismic profile of Yin et al. (2007b), which indicates a possible Pliocene activation age of this thrust. The Paleozoic granite was thrust over Triassic sedimentary rocks by the northern segment of the Ayakum thrust (e.g., Robinson et al., 2003; Dupont-Nivet et al., 2004; Yin et al., 2007b).

Structures of the central traverse here include the southeastern segment of the Ayakum thrust, the south-directed Narim thrust, the north-directed Golumb thrust, the south-directed Yeniugou imbricate thrusts, the Yeniugou left-slip fault, and the left-slip Kunlun fault (Fig. 3). The southeastern Ayakum thrust places Ordovician metamorphic sandstones and early Paleozoic granitoids over Carboniferous strata. The fault trace is covered by Quaternary river deposits in the east, obscuring its eastern extent. The south-directed Narim thrust places Paleozoic granitoid plutons over Neogene–Quaternary sedimentary rocks in the west, and the eastern extent of the fault cuts the Paleozoic–early Mesozoic granitoid plutons (e.g., Yin et al., 2007b) (Fig. 5C). The north-directed Golumb thrust juxtaposes lower Paleozoic metamorphic rocks over Carboniferous strata (e.g., Coward et al., 1988) (Fig. 5D). The Yeniugou imbricate thrusts place Proterozoic gneiss and Paleozoic granitoid plutons over the folded Triassic flysch complex, which is cut by the left-slip Kunlun fault in the south (e.g., Yin et al., 2007b; Wu et al., 2019a), whereas the Yeniugou left-slip fault is developed along the Kunlun river, which strikes parallel to the Kunlun fault as described in Wu et al. (2019a). At the outcrop scale, field evidence shows that the Paleozoic granitoid pluton is thrust over the Proterozoic schist (Fig. 5E).

The eastern traverse is located at the junction between the eastern Qaidam basin and Eastern Kunlun Range across the west-striking thrust belt, which contains the Kunlun left-slip fault, the Wenquan right-slip fault, and the northern Qaidam thrust. Structures here mainly include the Dulan detachment fault, the Dulan thrust, the Elashan fault, the Hacipu fault, the Chahanwusu fault, the Gouli fault, the south Qaidam fault, and the Boluoer fault (Fig. 4). Recent reviews of this structural pattern are provided in Yin et al. (2007b), Chen et al. (2015), and Wu et al. (2019b]. A regional unconformity contact between the Cretaceous strata and Cenozoic sedimentary rocks is exposed in the Eastern Kunlun Range (Wu et al., 2019b), the Hoh Xil basin to the south (Fig. 5F), and the Qaidam basin to the north (Fig. 5G).

Fission-track thermochronology is based on crystal-lattice damage manifested as linear tracks resulting from the constant-rate spontaneous fission of trace levels of ²³⁸U in zircon and apatite grains. Fission tracks in apatite are incompletely annealed over the temperature range of 60–120 °C, which is termed the partial annealing zone (PAZ) (e.g., Gleadow, 1981; Gleadow et al., 2002; Ketcham et al., 2007). Cooling of a sample through the partial annealing zone with time is reflected by the distribution of lengths for the partially annealed tracks. We conducted apatite fission-track (AFT) analyses from 22 granite samples to determine the low-temperature thermal history of the Eastern Kunlun Range.

Fission-track ages were measured using the external detector method (Gleadow, 1981) and calculated using the zeta calibration method (Hurford and Green, 1983). Ages were calculated using the Zeta calibration method (Hurford and Green, 1983; Hurford, 1990) with a Zeta value of 322.1 ± 3.6 (1 sigma). Apatite and zircon grains were separated from ~5 kg materials for each sample using standard mineral separation techniques at the Institute of the Hebei Regional Geology and Mineral Survey in Langfang, China. Polished grain mounts were prepared and etched to reveal spontaneous fission tracks. Apatite grain mounts were etched in 6.6% HNO₃ at 25 °C for 30 seconds, and all samples were irradiated at the China Institute of Atomic Energy reactor facility, Beijing. Low-U muscovite external detectors covering apatite grain mounts were etched in 40% hydrofluoric acid at 25 °C for 20 min to reveal induced fission tracks. In order to increase the number of observable horizontal confined tracks, the samples were exposed to ²⁵²Cf (Donelick and Miller, 1991). Horizontal confined fission-track lengths (e.g., Gleadow et al., 1986; Laslett et al., 1987) were measured only in prismatic apatite crystals because of the anisotropy of annealing of fission tracks in apatite (Green et al., 1986). We assume that the 22 granite samples that were analyzed started at temperatures equal to, or hotter than, the AFT PAZ (60–120 °C). Counts of 35–45 per granite sample were performed on average (Table 1).

Because fission-track systematics in apatite are characterized by a PAZ approximately between 60 and 120 °C, AFT ages and measured fission-track length distributions can be inverted to produce suites of compatible thermal histories. We performed inverse modeling of the AFT data using the HeFTy v1.9.1 software of Ketcham (2005) and the kinetic annealing model for apatite of Ketcham et al. (2007); this model considers the Dpar values and the angle with c-axis parameters. A goodness-of-fit (GOF) value was used to estimate how well the modeled data fit measured values (Ketcham, 2005). The paths were accepted (green envelopes) when the GOF > 0.05 and rated as good (pink envelopes) for GOF > 0.5 (Ketcham, 2005). Inverse thermal history modeling was run for 100,000 paths for each sample, which in all cases resulted in at least 1000 acceptable paths. For all samples, initial time-temperature conditions were set at above the AFT PAZ at a time range of 200–0 Ma. This initial condition is justified because all of these studied rock types are thought to have originated from below the uppermost crust. The burial depth of the Cretaceous sediments (~1.4 km) would have been above the AFT PAZ. Therefore, the thermal modeling started with the AFT samples above the AFT PAZ in the Cretaceous, and sought to see what temperature-time paths were permissible for the studied samples, including the possibility of reheating (i.e., burial).

The results of AFT analyses from our samples are shown in Table 1, including AFT age, track length, and Dpar information (Fig. S1¹). We used the central age of these samples. For the remaining AFT samples that passed the chi-square test, pooled AFT ages are reported (e.g., Sobel et al., 2006a, 2006b; Table 1) (Fig. S1). The observed grain-age distributions were decomposed into different grain-age components for these samples using the mixture model function of the Density Plotter program (Vermeesch, 2012) (Fig. 6). Almost all analyses were significantly younger than their respective crystallization (Table 1), which suggests that samples have experienced postcrystallization cooling histories through the apatite PAZ (e.g., Galbraith and Laslett, 1993; Gallagher et al., 1998; Yuan et al., 2006). Thermal history modeling of the AFT ages, track lengths, and Dpar data are used to evaluate the cooling processes, which were conducted using the HeFTy program to produce time-temperature pathway models in this study (Ketcham, 2005; Ketcham et al., 2007, 2009) (Fig. 7).

We focused on the AFT sampling across the northeast-striking Ayakum, Adatan, and Qimen Tagh thrusts located south of Gaskule Lake (Fig. 2). There, Paleozoic granite sills and dikes intrude Proterozoic and early Paleozoic metamorphic sedimentary and arc sequence rocks or are thrust over Cenozoic sedimentary rocks. Five samples (i.e., WC071615-1, WC071716-5, WC071716-3, WC071716-4, and WC071716-1) were collected from the Paleozoic granites in the hanging wall of observed thrusts. The AFT ages ranged from 68 ± 4 Ma (sample WC071716-1) to 40 ± 3 Ma (sample WC071716-4) with Dpar values of 1.83–2.11 (Table 1). The mean track lengths of our Paleozoic AFT samples range from 12.5 ± 1.8 mm (sample WC071716-1) to 11.7 ± 1.9 mm (sample WC071716–5) (Table 1).

Almost all samples failed the chi-square test (P (χ2) < 5%; Galbraith and Green, 1990) indicating significant dispersion in the individual grain ages with the exception of sample WC071716-3 (i.e., P (χ2) = 5.2%) (Fig. S1 [^{footnote 1}]). The modeled peak ages of these age components from five samples of the western domain of the Eastern Kunlun Range are grouped into four populations based on their peak age correlations, termed P1 (ca. 37.3 Ma), P2 (ca. 49 Ma), P3 (ca. 69 Ma), and P4 (ca. 76.9 Ma), respectively. This demonstrates that multi-phase cooling affected the western Eastern Kunlun Range since the latest Cretaceous to late Eocene (Fig. 6).

Two samples yielded good thermal modeling results (i.e., WC071615-1 and WC071716-5), showing a similar cooling history (Fig. 7). The models show that these samples were cooled from a temperature above the upper limit of the AFT PAZ during the early Cretaceous, followed by thermal stagnation in the PAZ through its lower limit since the late Cretaceous and finally rapid exhumation from ca. 8–5 Ma to present (Fig. 7).

Across a traverse of the Ayakum, Narim, Golumd, and Yeniugou thrusts and the left-slip Kunlun fault, 11 Paleozoic granite samples (i.e., WC3004164, WC2904162, WC03051612, WC2904164, WC2904161, WC3004161, WC2904163, WC3004163, WC0718163, WC0718161, and WC0718166) were collected to document the cooling history (Fig. 3). The AFT ages spread across a wide range from 91 ± 6 Ma (sample WC300416-1) to 11 ± 1 Ma (sample WC071816-1) with Dpar values of 1.36–1.98 (Table 1). The mean track lengths of these 11 samples range from 12.0 ± 2.0 mm (sample WC300414-1a) to 13.6 ± 1.4 mm (sample WC071816-3) (Table 1). Results of most samples passed the chi-square test (P(χ2) > 5%; Galbraith and Green, 1990) with the exception of samples WC300416-1 and WC300416-3 (Fig. S1 [^{footnote 1}]), which indicates that most samples were affected by a single monotonic cooling event. The modeled peak ages of these age components from 11 samples of the central domain of the Eastern Kunlun Range are grouped into five populations based on their peak age correlations, termed P1 (ca. 13.77 Ma), P2 (ca. 41.9 Ma), P3 (ca. 57.7 Ma), P4 (ca. 80.9 Ma), and P5 (ca. 105.3 Ma), respectively. This demonstrates that multi-phase cooling affected the central Eastern Kunlun Range since the late Cretaceous to Miocene (Fig. 6).

Eight samples yield good thermal modeling results, which exhibit three different thermal histories (Fig. 7). Samples WC290416-2, WC290416-4, WC300416-1, WC071816-6, and WC300416-3 cooled from a temperature above the upper limit of the AFT PAZ during the late Cretaceous, followed by thermal stagnation in the PAZ through its lower limit, and finally a strong pulse of rapid cooling since ca. 20–16 Ma (Fig. 7). Samples WC030516-12 and WC290416-1 are characterized by initial late Cretaceous cooling, residence within the PAZ, and finally a strong pulse of rapid cooling through the lower limit of the PAZ since ca. 10–5 Ma (Fig. 7). Sample WC090416-3 shows initial late Cretaceous cooling and residence in the PAZ, a strong pulse of rapid cooling through the lower limit of the PAZ since ca. 20 Ma, and finally rapid exhumation from ca. 10–12 Ma to present (Fig. 7).

Six samples (i.e., WC071115-2, WC071115-3, WC071115-5, WC0805142, WC0713151b, and WC071015-1a) were collected from the Paleozoic–early Mesozoic granitoids in the hanging wall of the thrusts (Fig. 4), and the AFT ages are spread over a wide range from 66 ± 4 Ma (sample WC071115–5) to 16 ± 1 Ma (sample WC071015-1a) with Dpar values of 1.61–2.27 (Table 1). The mean track lengths of these six samples range from 12.1 ± 2.2 mm (sample WC071315-1b) to 12.9 ± 2.1 mm (sample WC071115-2) (Table 1). Results of samples WC071115-2, WC071115-5, and WC071315-1b failed the chi-square test, whereas samples WC071115-3, WC080514-2, and WC071015-1a passed the chi-square test (e.g., Galbraith and Green, 1990) (Fig. S1). The modeled peak ages of these age components from six samples of the eastern domain of the Eastern Kunlun Range are grouped into four populations based on their peak age correlations, termed P1 (ca. 19.9 Ma), P2 (ca. 34 Ma), P3 (ca. 50 Ma), and P4 (ca. 69.4 Ma), respectively. This demonstrates that multi-phase cooling affected the eastern Kunlun Range since the latest Cretaceous to Miocene (Fig. 6).

Six samples yield good thermal modeling results, which exhibit four types of thermal histories (Fig. 7). Samples WC071315-1b and WC080514-2 cooled from a temperature above the upper limit of the AFT PAZ during the late Cretaceous, followed by thermal stagnation in the PAZ, and finally a strong pulse of rapid cooling through the lower limit of the PAZ since ca. 20 Ma (Fig. 7). Samples WC071115-2 and WC071115-3 exhibit thermal histories that are characterized by cooling from a temperature above the upper limit of the AFT PAZ during the Cretaceous, residence in the PAZ, and latest rapid cooling since ca. 30 Ma (Fig. 7). Sample WC071115-5 reveals cooling in the late Cretaceous to the bottom of the AFT PAZ followed thermal stagnation in the PAZ and cooling from the PAZ to the surface since the early Cenozoic (Fig. 7). Sample WC071015-1a exhibits cooling from a temperature above the upper limit of the AFT PAZ during the Oligocene and residence in the PAZ until it was cooled rapidly through the lower limit of the PAZ since 20 Ma (Fig. 7).

Below we describe our AFT analyses, their structural positions, previous published low‐temperature thermochronology data (Table S1 [^{footnote 1}]), and their correlation with existing knowledge in the literature to construct a deformation history that constrains the evolution of the Eastern Kunlun Range. AFT analyses and thermal modeling suggest that the Eastern Kunlun Range experienced a complex multi-phase cooling history through early Cretaceous to the present day (Fig. 8).

Almost all samples from the Eastern Kunlun Range with good modeling paths and modeled peak age components reveal an initial phase of early Cretaceous cooling and relative slow exhumation over the early Cenozoic, which indicates that these samples may have experienced long-term residence in the apatite PAZ through the late Cretaceous to Eocene (Fig. 7). Similar observations have been made from previous published thermochronology data sets (e.g., Tian et al., 2020) (Fig. 8) and stratigraphic records in northeastern Qaidam Basin (e.g., Ritts and Biffi, 2000; Wu et al., 2011). The local early Cretaceous cooling may be related to the far-field effects of the closure of the Neo-Kunlun Ocean along the Kunlun-Anyemaqen suture zone (e.g., Yin and Harrison, 2000; Pullen et al., 2008; Wu et al., 2016, 2019a) and the collision between the Lhasa and the Qiangtang blocks along the Bangong-Nujiang suture zone in the south during the Middle Jurassic to early Cretaceous (e.g., Kapp et al., 2007; Guo et al., 2019). A quiescence stage between the late Cenozoic and Eocene is supported by thermal history modeling from published data and this study (Fig. 9A).

Our AFT samples WC071015-1a, WC071115-2, and WC071115-3 from the eastern domain of the Eastern Kunlun Range show rapid Oligocene cooling, which we interpret resulted from fault-related uplift and exhumation (Fig. 7). We note that the structural positions of these three samples are located in the hanging wall of the southern Qaidam thrust (Fig. 4), which is consistent with the previous low-temperature thermochronology analysis (Mock et al., 1999; Clark et al., 2010) (Fig. 8). The early Cenozoic cooling may imply that contractional structures were locally developed across the forebulge in the Eastern Kunlun Range (Wu et al., 2019b), which is consistent with the short-lived development of a small foreland basin along the northern margin of the Eastern Kunlun Range. The early Cenozoic deformation in the northern plateau suggests plate-boundary stress transferred rapidly shortly after the India-Asia collision. Compressive stresses resulting from the ongoing India-Eurasia collision until ca. 50 Ma and continued to be transferred farther north across the Paleo-Qaidam basin to the northern margin of the plateau (e.g., Zuza et al., 2018; An et al., 2020; Li et al., 2020) (Fig. 9B). Most of the analyzed samples with good modeling paths in the eastern and central domains of the Eastern Kunlun Range show cooling out of the PAZ toward the surface in the early to middle Miocene (from ca. 20 Ma), which indicates that the Eastern Kunlun Range experienced uplift after ca. 20 Ma (Fig. 7). In addition, Miocene and younger sedimentary rocks in the Qaidam basin to the north became dominated by material sourced from the Eastern Kunlun Range (e.g., Bush et al., 2016; Cheng et al., 2016; McRivette et al., 2019; Wu et al., 2019b) (Fig. 7), which supports the continued and increasingly rapid uplift of the Eastern Kunlun Range into the Miocene (Fig. 8). The Miocene deformation and rapid uplift are associated with the two-bounding fold-and-thrust belts (i.e., Fenghuoshan-Nangqian and Qilian Shan–Nan Shan thrust belts) that may have generated a flexural bulge across the middle axis of the Paleo-Qaidam basin (Fig. 9B).

The Late Miocene to present (ca. 10 Ma) pulse of deformation is well documented by samples in the central and western domains of the Eastern Kunlun Range and is interpreted to reflect the reactivation of thrusts and initiation of left-slip Kunlun fault (Fu and Awata, 2007; Chen et al., 2011; Tian et al., 2020) (Figs. 7 and 9C). Although the extent of this deformation is not well preserved because of later erosion or overprinting, we further suggest that the left slip along the Kunlun fault is possibly associated with the development of the earlier phase thrusts within the Eastern Kunlun Range (e.g., Chen et al., 2011). Pliocene cooling is observed in modeling results of the samples from the western domain of the Eastern Kunlun Range and two samples (i.e., WC030516-12 and WC290416-1) in the central domain (Fig. 7); this cooling is interpreted as the initiation of later stage thrusts. These AFT samples are from the hanging wall of the north-dipping parallel Qimen Tagh imbricate thrusts, the Adatan thrust, the Ayakum thrust, the Golumd thrust, and the Kunlun fault, and record the thermal history related to slip along these thrusts (Figs. 2 and 3). Seismic profiles showing that the Adatan and Ayakum thrusts developed under Pliocene growth strata, combined with our new AFT modeling results and published data from the western end Jianxiashan thrust belt of Wu et al. (2020b) and eastern end thrust of Duvall et al. (2013), are consistent with the initiation of the foreland contraction structures of the Eastern Kunlun Range and basins along its two flanks. Furthermore, the uplift of the Eastern Kunlun Range extends laterally northwestward to the Qimen Tagh and southeastward to the Bayan Har (Figs. 1C and 10).

The evolution of the Himalayan-Tibetan orogen and the growth of the Tibetan Plateau depend on the structural framework of the thrust belts that bound the plateau, including the Himalaya to the south and Qilian Shan to the north (e.g., DeCelles et al., 2002; Yin, 2006; Webb et al., 2011, 2017; Haproff et al., 2019; Wu et al., 2019a, 2019b; Zuza et al., 2019; An et al., 2020; Bian et al., 2020). Early Cenozoic deformation shortly after the initial India-Asia collision has been observed across most of the northern Tibetan Plateau (Fig. 1B), which suggests plate-boundary stress transferred rapidly across the orogen (e.g., Yin et al., 2007b; Li et al., 2019, 2020; Zuza et al., 2019; Wu et al., 2020b; Bian et al., 2020; Yu et al., 2020) rather than the classic continuum deformation models that predict progressively northward-propagating deformation (e.g., England and Houseman, 1986; Zheng et al., 2010, 2017; Yu et al., 2019a, 2019b). Following Eocene–Oligocene Qilian Shan–northern Qaidam deformation along the northern boundary of the Tibetan Plateau, Oligocene deformation occurred along the southern Qaidam (Fig. 10), and the southern Qaidam thrust developed westward during the Oligocene (e.g., Chen et al., 2011; Duvall et al., 2013). Miocene deformation jumped back to the Eastern Kunlun Range during the partitioning of the Paleo-Qaidam basin into the Hoh Xil and Qaidam sub-basins (Fig. 10). The Miocene uplift of the Eastern Kunlun Range should be associated with the activation of the Eastern Kunlun thrusts and the later initiation of Kunlun left-slip fault in the Late Miocene (Fig. 10). The dip-slip shortening and strike-slip faulting at the western and eastern terminations of the left-slip Kunlun fault occurred since the latest Miocene and Pliocene (e.g., Duvall et al., 2013; Wu et al., 2020a, 2020b) (Fig. 10).

Field observations, AFT analyses, and thermal history modeling provide constraints on the complex Cretaceous to Cenozoic exhumation history of the Eastern Kunlun Range and the multi-phase intracontinental growth history of the northern Tibetan Plateau. Low-temperature thermochronology analyses reveal four stages of cooling history of the Eastern Kunlun Range: (1) an initial period of early Cretaceous cooling and slow exhumation over the early Cenozoic, supported by the exposure of a regional unconformity contact between Cretaceous and early Cenozoic sedimentary rock; (2) rapid Oligocene cooling in the eastern domain related to slip along the southern Qaidam thrust that subsequently propagates to the west; (3) extensive strongly rapid cooling since the early-middle Miocene (ca. 20 Ma) in mostly the eastern-central domains and significant uplift of the Eastern Kunlun Range; and (4) a final rapid pulse of cooling from the Late Miocene to present associated with the initiation of the left-slip Kunlun fault, and the dip-slip shortening at the western and eastern terminations of the fault. Early Cenozoic deformation was distributed along the northern Tibetan Plateau, and overprinting out-of-sequence deformation migrated to the south during the Miocene to present deformation in the Eastern Kunlun Range.

We thank Dr. Andrew Zuza and Dr. Xuan-hua Chen for inspiring discussion. This research was supported by grants from the Basic Science Center for Tibetan Plateau Earth System (CTPES, grant no. 41988101-01), the Second Tibetan Plateau Scientific Expedition and Research Program (grant no. 2019QZKK0708), National Natural Science Foundation of China (grants nos. 41702232, 41941016, and 41661134049), the National Key Research and Development Project of China (grant no. 2016YFC0600303), and the Chinese Academy of Sciences Strategic Priority Research Program (grant no. XDA20070301). We thank Science Editor Andrea Hampel, Associate Editor Alexander Rohrmann, Peter Haproff, and two anonymous reviewers for their constructive comments that led to significant improvement of the original draft of this paper.