Introduction

Orogens are commonly characterized by an arcuate geometry in a map view. The formation of such curvatures (also referred to as oroclines1) has commonly been thought to involve the bending of quasi-linear belts around vertical axes2. Oroclinal bending plays a major role during the evolution of convergent plate boundaries, influencing paleogeography, magmatism, continental growth, and mountain-building processes3,4,5,6. This process is inevitably accompanied by orogenic-scale deformation, but how strain is partitioned and accommodated during oroclinal bending is not fully understood. Studies that focused on oroclines in eastern Australia and the central Mediterranean have suggested that flexural-slip mechanisms7 or orogen-perpendicular tear faults8 possibly played important roles in accommodating the large-scale deformation of oroclinal bending. In other oroclines (e.g., Cantabrian Orocline in Western Europe), it has been suggested that the inner hinge zone of the orocline may undergo contractional deformation and crustal thickening9,10. While these interpretations are supported by small-scale structural observations (e.g., ref. 11), crustal-scale evidence supporting the idea that oroclinal bending leads to contraction and crustal thickening has hitherto not been provided. One way to address this issue is to estimate paleo-crustal thickness based on LaN/YbN ratios in intermediate-felsic arc magmas12,13,14,15. Here we apply this method to investigate the crustal response to the development of the Mongolian Orocline in Central Asia.

The Mongolian Orocline constitutes a ~6000 km Andean-type subduction system that forms a U-shape geometry (Fig. 1a–c), thus representing one of the most contorted orogens on Earth16. The orocline was developed during the Permian and early Mesozoic, contemporaneously with the convergence of the Siberian and North China cratons and the subduction of the Mongol-Okhotsk Ocean17,18,19,20. In a map view, the Mongolian Orocline shows a relatively wider hinge zone (we use the words “wide” and “narrow” to describe the map-view width of tectonic units; Fig. 1a, c). Based on gravity data21,22, the crust in the inner hinge of the orocline around the Hangay Mountains is thicker (we use “thick” and “thin” in reference to crustal thickness; Fig. 1d). In this paper, we investigate whether these variations in the crustal architecture are possibly linked to the process of oroclinal bending. Based on geochronological and geochemical data from Permian and Triassic magmatic rocks (Fig. 1d), we suggest that the inner hinge zone of the Mongolian Orocline was subjected to contractional deformation and crustal thickening in response to oroclinal bending.

Fig. 1: Geological, magnetic, and crustal thickness maps around the Mongolian Orocline.
figure 1

a Simplified tectonic map of the central and eastern Central Asian Orogen Belt showing the geometry of the Mongolian Orocline (thin black dash line). The geology is based on the 2008 edition of the 1:2,500,000 geological map of central Asia and adjacent areas. MOO, Mongol-Okhotsk Orogen. b Total magnetic image of Mongolia (modified with permission from ref. 25, Geological Society of America) showing the geometry of the Mongolian Orocline (white dash line). c Tectonic map of Mongolia (modified with permission from ref. 25, Geological Society of America). d Present-day crustal thickness in Mongolia with contours in km (modified with permission from ref. 21, Elsevier and ref. 38, Elsevier). Sample locations are also shown.

Full size image

Geological setting

The Mongolian segment of the Central Asian Orogen Belt can be divided into the southern and northern domains, separated by the Main Mongolian Lineament23 (Fig. 1c). The southern domain comprises a series of late Paleozoic island-arc systems. The northern domain, which includes the Mongolian Orocline, comprises Precambrian microcontinental blocks, Neoproterozoic island arcs, and Paleozoic accretionary complexes23 (Fig. 1c). The amalgamation of the southern and northern domains likely occurred in the latest Paleozoic following the closure of the Paleo-Asian Ocean18,23,24. Permian to Jurassic tectonic evolution of the Mongolian segment of the Central Asian Orogen Belt was characterized by continuous subduction of the Mongol-Okhotsk oceanic plate, during which the Mongolian Orocline was developed16,19,20,25.

Precambrian microcontinents that occur around the Mongolian Orocline include the Buteel, Tuva-Mongolian, Tarvagatay, Zavkhan, Baydrag, Ereendavaa, and Idermeg blocks23 (Fig. 1c). Their present-day configuration delineates the U-shaped geometry of the Mongolian Orocline18,26. The core area of the Mongolian Orocline is occupied by Ediacaran–Carboniferous accretionary complexes, which include, from west to east, the Bayanhongor zone (Ediacaran–Cambrian), the Zag zone (Cambrian–Ordovician), and the Hangay–Hentey complex (Devonian–Carboniferous; Figs. 1c and 2)23. The presence of Ediacaran–Carboniferous accretionary complexes is a testimony to the subduction of the Mongol-Okhotsk Ocean, but the exact plate-tectonic configuration of this subduction zone is poorly understood.

Fig. 2: Geological map of the Hangay Mountains and adjacent areas, and an age profile across the research area (from west to east) showing progressively younger magmatic rocks towards the east.
figure 2

The geology is based on the 1990 edition of the 1:500, 000 geological map and ref. 25. For the location of the geological map, see Fig. 1c. For the classification criteria of Group I and Group II magmatic rocks, see the main text. Blue and yellow boxes in the geological map are the ages of Group I and Group II granitoids, respectively (for a summary of U–Pb ages, see Supplementary Data 4).

Full size image

Permian–Triassic volcanic and plutonic rocks widely occur around the Mongolian Orocline20,23. This magmatic belt has been considered to represent the magmatic arc associated with the subduction of the Mongol-Okhotsk Ocean20,27. In the inner hinge zone of the Mongolian Orocline, the Permian–Triassic magmatic belt is represented mainly by the Hangay Batholith27,28 (Fig. 2). Based on the spatial distribution, magmatic rocks of the Hangay Batholith are divided, in this study, into two groups. Group I occur within the Precambrian basement terranes (Tarvagatay, Zavkhan, and Baydrag blocks); Group II are located within accretionary wedge complexes (Bayanhongor zone, Zag zone, and Hangay–Hentey complex), or along the boundary between the accretionary wedge complexes and the Precambrian microcontinental blocks (Fig. 2).

Results

To conduct geochronological and geochemical analyses, we collected 65 samples from 21 granitoid plutons from the Hangay Batholith in the inner hinge zone of the Mongolian Orocline (Fig. 2 and Supplementary Fig. 1). Group I granitoids are dominated by quartz monzonite, monzonite, monzogranite, and syenogranite; Group II granitoids include mainly syenogranite and monzogranite. A detailed lithological description of analyzed samples is presented in Supplementary Note 1 and Supplementary Data 1. Analytical results can be found in the Supplementary Data 2 and 3.

Zircon U–Pb geochronology

Zircon U–Pb dating was conducted on 23 samples from 21 granitoid plutons. For each sample, 20–40 representative zircon grains were analyzed. These zircons show euhedral-subhedral crystal morphology, and have length/width ratios varying from 2:1 to 3:1 (Supplementary Fig. 2). These zircon characteristics, together with evidence of moderate to high Th/U ratios (0.2–3.5) and oscillatory zoning, are indicative of a magmatic origin29.

The zircon U–Pb data are presented in Supplementary Data 2 and Supplementary Fig. 3, and the age description of each sample is provided in Supplementary Note 2. Group I granitoids yield mainly Permian ages (297.7–256.7 Ma, Supplementary Data 2). Group II granitoids yield mainly Triassic ages (247.9–235.5 Ma, Supplementary Data 2). One pluton from the Zag zone (Group II) yields an older age of 256.5 ± 1.6 Ma (sample L18M24).

Whole-rock major and trace elements

Forty-eight samples from Group I granitoids and seventeen samples from Group II were analyzed for major and trace elements. Samples from Group I have a wide range of SiO2 (54.43–76.62 wt%), K2O (1.33–6.24 wt%), and Na2O (3.26–6.63 wt%). Plotted on a TAS diagram, the samples appear mainly in the monzodiorite and granite fields (Fig. 3a). Most samples belong to the high-K calc-alkaline and shoshonitic series (Fig. 3b). The rocks exhibit a transitionary trend from metaluminous to weakly peraluminous composition with low A/CNK values (0.75–1.08; Fig. 3c). Trace-element compositions of Group I granitoids show variable Sr and Ba contents, but negative Nb, Ta, P, and Ti anomalies and positive Rb, Th, and K anomalies (Supplementary Fig. 4). Based on the Sr, Ba, and Eu contents, we divide Group I granitoids into three sub-groups (Group IA, Group IB, and Group IC). Group IA is characterized by depleted Sr and Ba and negative Eu anomalies, indicating feldspar fractionation (Supplementary Fig. 4a, b). A few samples from Group IA show concave-down MREEs distribution that likely results from fractional crystallization of amphibole (Supplementary Fig. 4a). Group IB granitoids show weak negative Sr, Ba, and Eu anomalies (Supplementary Fig. 4c, d). In contrast, Group IC granitoids show positive Sr and Ba anomalies, and no obvious negative Eu anomalies, indicating limited feldspar fractionation (Supplementary Fig. 4e, f).

Fig. 3: Geochemical and geochronological diagrams.
figure 3

a TAS diagram60. Abbreviations: PG Peridot-gabbro, MD Monzo-diorite. b SiO2 versus K2O diagram61. c A/CNK versus A/NK diagram of felsic and intermediate rocks62. d Age probability diagram for igneous samples with geochemical data around the inner hinge of the Mongolian Orocline. e Age probability diagram for igneous samples after filtering (see text). f LaN/YbN (and corresponding crustal thickness) versus age for the samples shown in Fig. 3e (for the data, see Supplementary Data 4). The red line with two sigma error bars represent the average trend of LaN/YbN ratio, which is calculated by a Monte Carlo analysis with a weighted bootstrap resampling approach41. Given that data are not available between 285 and 270 Ma, a dotted line is used to trace the trend. The yellow rectangle shows the present-day crustal thickness around the inner hinge of the Mongolian Orocline21 (see Fig. 1d).

Full size image

Group II granitoids are characterized by higher SiO2 contents (60.56–72.77 wt%). On the TAS diagram, most samples are plotted in the granite and monzonite fields (Fig. 3a). All samples of Group II belong to the high-K calc-alkaline and shoshonitic series (Fig. 3b), and show metaluminous to weakly peraluminous characteristics with slightly low A/CNK (0.84–1.11; Fig. 3c). The chondrite-normalized REE patterns show that all samples from Group II are enriched in LREEs, and have Eu anomalies (δEu = 0.41–0.74) associated with plagioclase and/or K-feldspar fractionation (Supplementary Fig. 4g). In the primitive mantle-normalized spidergram, Group II granitoids show negative Ba, Nb, Ta, P, and Ti anomalies and positive Rb, Th, and K anomalies (Supplementary Fig. 4h).

Discussion

Migration of arc magmatism

New petrological and geochemical data support the previous interpretation that the Permian–Triassic magmatic belt around the Mongolian Orocline likely represents an arc associated with the subduction of the Mongol-Okhotsk Ocean20,27 (Fig. 1c). The lithology of our samples (monzonite, quartz monzonite, granodiorite, monzogranite, and syenogranite), as well as other lithologies (gabbro and gabbrodiorite) of the Hangay Batholith reported in previous studies27,28, are typical rock associations in active marginal arcs (e.g., Coastal and Sierra Nevada batholiths)30,31,32. Further evidence supporting arc magmatism is the calc-alkaline to shoshonite-series composition, the enrichment in large-ion lithophile elements (LILEs; Rb, Ba, Th, and K), and the depletion in high field-strength elements (HFSEs; Nb, Ta, P, and Ti; Supplementary Fig. 4)33,34. Plotted on tectonic discrimination diagrams of Th/Yb–Nb/Yb and Rb–Y+Nb (Supplementary Fig. 5), samples from the Hangay Batholith fall mainly in the fields of continental arc and volcanic arc granite35,36.

A compilation of newly obtained zircon U–Pb ages and published geochronological data (total 145 ages; Fig. 2 and Supplementary Data 4) shows that magmatism in the Hangay Batholith lasted from ~298 Ma to ~220 Ma. Spatially, magmatic rocks from this batholith become progressively younger from west to east, with Permian plutons intruding mainly into the Precambrian basement blocks (Group I), and Triassic granitoids predominantly intruding into the Neoproterozoic to Carboniferous accretionary complexes (Group II; Fig. 2). This spatio-temporal distribution suggests that from the Early Permian (~298 Ma) to the Late Triassic (~220 Ma), the magmatic arc migrated towards the core of the Mongolian Orocline.

Crustal thickness reconstruction based on LaN/YbN proxy

Available geophysical data show that the present-day crust around the inner hinge of the Mongolian Orocline (Hangay Mountains) is thicker than the surrounding area21,22 (Fig. 1d). This thicker crust (with a maximum value of ~60 km) provides an isostatic support for the high topography of the Hangay Mountains37. Given that there is no evidence of Cenozoic deformation that could have led to crustal thickening in this area, the present-day thick crust has been attributed to pre-Cenozoic accretionary or collisional tectonics associated with the consumption of the Mongol-Okhotsk Ocean38. One way to test this hypothesis is to investigate spatio-temporal changes of LaN/YbN ratios in intermediate to felsic arc-related rocks. This ratio can be used as a proxy for crustal thickness of supra-subduction zones13,14,15.

To assess spatial and temporal variation of LaN/YbN ratios in arc magma around the inner hinge of the Mongolian Orocline, we compiled a geochemical database based on new and published data from the Hangay Batholith (Supplementary Data 4). The LaN/YbN, crustal thickness proxy, is based on the assumption that magmatic rocks originating from thicker arc crust generally possess higher LaN/YbN ratios than those in thinner arc crust13. This is explained by garnet and/or amphibole fractionation during magma differentiation in a deeper crustal environment >~1.0 GPa14. However, the change in LaN/YbN ratio is not only controlled by the fractionation of high-pressure minerals but also influenced by other factors, such as the degree of magma differentiation and the composition of source rocks13,14. Highly fractionated granites could involve the fractionation of accessory minerals, such as monazite and apatite39, which are rich in LREEs; their fractionation can thus cause low LaN/YbN ratios in parental magma. If such, crustal thickness that is calculated based on LaN/YbN ratios, can be underestimated. Moreover, partial melting of mature sedimentary rocks could produce peraluminous high-Si granites, which may have high LaN/YbN ratios due to the presence of peritectic garnet, thus mimicking the information of the magma differentiation from thicker arc crust40. Therefore, prior to reconstructing crustal thickness from LaN/YbN ratios, it is necessary to filter the geochemical data.

Following the data filtering methods for crustal thickness reconstruction12,14, we excluded data with SiO2 contents of <55 wt% and >70 wt% and MgO >4 wt%, thus eliminating mantle-derived mafic rocks and high-Si granitoids whose LaN/YbN ratios may not be indicative of crustal thickness (Supplementary Data 4; Fig. 3d). Given that all granitoids of the Hangay Batholith exhibit metaluminous to weakly peraluminous characteristics (Fig. 3c), we can exclude the possibility that mature sedimentary rocks contributed significantly to the magma source. The filtering reduces the number of geochemical data from 129 to 85 (Fig. 3d, e). The remaining data, which spatially cover the whole research area (Supplementary Fig. 6), have an age range of 298–230 Ma (Supplementary Data 4; Fig. 3e). These data show that LaN/YbN ratios from the Hangay Batholith generally increase from Permian to Triassic (Fig. 3f). We note, however, that no data exist for the period 285–270 Ma and the dataset for 240–230 Ma is relatively small. To calculate the crustal thickness, we use the empirical regression: LaN/YbN = (0.98 ± 0.19)exp(0.047 ± 0.005)H, where H is the crustal thickness14. The results show an increase in crustal thickness from the Permian to the Triassic (Fig. 3f). To evaluate the average trend and to minimize sample bias, we conducted a Monte Carlo analysis with a weighted bootstrap resampling approach41,42. Based on these results, we infer an increase in average crustal thickness from ~50 km at ~298 Ma to ~65 km at ~230 Ma (Fig. 3f).

Spatially, Group I granitoids of the Hangay Batholith, with dominant Permian ages, were emplaced mainly into Precambrian microcontinents (Fig. 2), so the LaN/YbN ratios of these granitoids constrain the crustal thickness of these microcontinents. During this period, the accretionary complexes of both Zag zone and Hangay–Hentey complex were located at a forearc position (Fig. 2). The crustal thickness of these accretionary complexes during this period is unconstrained, but was likely thinner than the adjacent arc crust of the Precambrian microcontinents. This interpretation is consistent with modern supra-subduction zones, where the maximum crustal thickness coincides with the magmatic arc43. The emplacement of Group II granitoids allows us to reconstruct the crustal thickness of the accretionary complexes in the Triassic, which is even larger than the Permian crustal thickness of the Precambrian microcontinents (Supplementary Fig. 6). This means that the accretionary complexes were thickened during the period from Permian to Triassic. We therefore conclude that, while the arc front migrated towards the core of the Mongolian Orocline, from the Permian to the Triassic, the crust in the inner hinge zone of the orocline was subjected to thickening.

Crustal response of oroclinal bending

The development of the Mongolian Orocline was accompanied by continuous subduction and the closure of the Mongol-Okhotsk Ocean20 (Fig. 4a, b). This subduction system was approximately quasi-linear in the latest Carboniferous17,19,44. By the end of the Triassic, the angle between two limbs of the orocline was ~30°19, indicating that significant bending took place during the Permian–Triassic. Plate tectonic reconstructions have suggested a diachronous closure of the Mongol-Okhotsk Ocean, with the western portion of the Mongol-Okhotsk Orogen recording an earlier closure relative to the eastern portion (modern coordinate)20,45. Based on paleomagnetic data, which show overlapping Late Jurassic paleomagnetic poles from the two limbs of the orocline, it has been suggested that the closure of the Mongol-Okhotsk Ocean was completed by the Late Jurassic19. However, the onset of the diachronous closure of this ocean is not well constrained but is likely after ~220 Ma based on the presence of arc-related rocks (dated ~220 Ma) in the Hangay Batholith as documented above. Considering that the arc around the inner hinge of the Mongolian Orocline migrated towards the core of the orocline during the Permian–Triassic, and the crust became progressively thicker, we suggest that the processes of arc migration, crustal thickening, and oroclinal bending were interlinked.

Fig. 4: Schematic illustrations showing the tectonic evolution of the Mongolian Orocline.
figure 4

a, b Tectonic reconstruction of the Mongol-Okhotsk subduction system during the Early Permian and Middle–Late Triassic (modified with permission from ref. 63, Elsevier, ref. 64, Wiley, and ref. 65, Elsevier). c, d Cartoons that illustrate the spatial migration of arc magmatism, the map-view widening of the supra-subduction system, and crustal thickening in response to inner hinge contraction. Note that the Permian Group I of the Hangay Batholith was emplaced mainly into Precambrian microcontinents (Fig. 4c), whereas the Triassic Group II occurred farther east within the Paleozoic accretionary complexes (Fig. 4d).

Full size image

In a map view, the inner hinge zone of the Mongolian Orocline is wider than the limbs, as recognized, for example, by the spatial distribution of the Baydrag–Zavkhan block, the Bayanhongor Zone, and the Zag Zone, which all show widening from the southwestern limb of the orocline toward the hinge (Figs. 1c and 2). Numerical models of oroclinal bending have shown that a wider hinge zone can be produced by progressive oroclinal bending46. We suggest that the inner hinge zone of the Mongolian Orocline, which represents the supra-subduction zone of the subducting Mongol-Okhotsk Ocean, became wider during oroclinal bending, thus forcing the trench to migrate oceanward. Accordingly, the arc retreated oceanward passively (Fig. 4c, d), with earlier (Permian) magmatism emplaced mainly into Precambrian microcontinents (Fig. 4c) and later (Triassic) granitoids intruded into the Paleozoic accretionary complexes (Fig. 4d). This behavior is not dissimilar to arc migration driven by slab rollback around the inner hinge of other oroclines (e.g., the Calabrian Orocline and the curved Banda arc system)47,48, with the difference that trench/arc migration in the Mongolian Orocline may have been strongly influenced by the progressive widening and thickening of the overriding orogen.

The evidence of hinge zone widening and thickening supports the idea that Permian to Triassic oroclinal bending was accommodated by contraction in the inner hinge zone of the Mongolian Orocline. The main contribution to the thickened crust was unlikely related to new crustal growth (e.g., magma underplating), but more likely associated with the redistribution of crustal materials around the inner hinge zone of the orocline. Simultaneously with contraction at the inner hinge zone of the orocline, rift systems were developed in the West Siberian Basin49,50 (Figs. 1a and 4b). It is possible that this extensional tectonism was also linked to the development of the orocline, accommodating the rotation of the northern limb of the orocline (Siberian Craton) relative to the southern limb (Fig. 4b).

Our results demonstrate that the inner hinge of oroclines may undergo crustal-scale contraction, which would affect the orogenic architecture and the spatial and temporal evolution of arc magmatism. This deformation pattern seems independent of the geodynamic mechanism of oroclinal bending. In the case of the Mongolian Orocline, previous studies have demonstrated that large-scale bending was driven either by a rapid slab rollback along the two limbs of the orocline or by the convergence and rotation of the Siberian and North China cratons20,25. In other oroclines (e.g., Cantabrian Orocline), it has been suggested that oroclinal bending was driven by orogen-parallel shortening, which led to crustal thickening and thinning at the inner and outer zones of the orocline, respectively, as well as post-subduction magmatism4,51. It seems plausible that crustal-scale contraction around the inner hinge zones of oroclines represents an end-member mechanism to accommodate the large-scale bending of orogenic systems. The formation of tight oroclines possibly involves a combination of this mechanism and other modes of deformation, such as flexural slip and orogen-perpendicular tearing7,8.

The thicker crust surrounding the inner hinge zone of the Mongolian Orocline spatially corresponds to the relatively elevated topography of the Hangay Mountains (Fig. 1d and Supplementary Fig. 7). The uplift history of these mountains is relatively poorly constrained, but the spatial relationship suggests a potential genetic link between them, with the high topography possibly isostatically supported by the thicker crust. This phenomenon of heightened topography is mirrored around the hinge of modern curved convergent plate boundaries, such as the Bolivian Orocline and the curved Cascadia subduction zone3,52. The variation in topography along these curved convergent plate boundaries may ultimately influence the differing climate and environment observed along the mountain range, as demonstrated by the modern Andes53.

Methods

U–Pb geochronology

U–Pb dating of zircons was conducted by LA–ICP–MS at the Key Laboratory of Mineralogy and Metallogeny, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. The laser size was set to 29 μm at a frequency of 6 Hz. Detailed operation processes can be found in ref. 54. The external standards Zircon 91500 and NIST 610 were used to calibrate U–Pb isotopic and elemental analyses. ICPMSDataCal 10.9 was used to carry out quantitative calibration for U–Pb analysis55,56. Concordia diagrams and age calculation were done by Isoplot 4.1557.

Whole-rock geochemistry

Whole-rock major elements were analyzed by X-ray fluorescence (XRF) at the State Key Laboratory of Isotope Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Sciences (SKLIG GIG CAS). The detailed operation processes are described in ref. 58. The analytical precision for major elements was between 1% and 5%. Whole-rock trace elements were analyzed by Agilent 7700e ICP-MS at the Wuhan Sample Solution Analytical Technology Co., Ltd., Hubei Province (China). The rock standards AGV-2, BHVO-2, BCR-2 and RGM-2 were used to calibrate trace element concentrations. Detailed analytical processes are documented in ref. 55. The analytical precision for trace elements was generally better than 5%.