LAUSR

The Central Asian Orogenic Belt constitutes the Kazakhstan and Mongolian oroclinal collages located between the Tarim–North China collage (TNC) to the south and the Siberian craton to the north (Fig. 1a) [1]. While the oroclinal bending of the western Kazakhstan collage is widely accepted by the paleomagnetic and tectonic community, the oroclinal bending model of the eastern Mongolian collage is disputed [2,3]. Despite differences, all studies report: (i) difference in Late Paleozoic paleolatitudes between Siberia and the peri-Siberian terranes accreted during Cambrian–Ordovician [4,5], the Amuria and the TNC; (ii) Permian consumption of the Paleo-Asian Ocean (PAO) and (iii) Permian–Triassic counterclockwise rotations of the latter two blocks during scissor-like closure of the Mongol OkhotskOcean (MOO) to thenorth [6]. Recently acquired [2,3] and unpublished paleomagnetic data show a consistent evolution for the PAO Ordovician strata of the Mongolian Altai Wedge (MAW), theDevonian andCarboniferous oceanic rocks of theTrans-Altai Zone (TAZ), the northerly Proterozoic Lake Zone (LZ) and the Mongolian Precambrian blocks (MB) constituting the Amuria (Fig. 1a). Their joint paleomagnetic record can be grouped into five direction groups A1, A2, B, C and D, all due to successive magnetic overprinting phases (Fig. 1b). The magnetizations are carried by secondary minerals such as titanohematites for A1 and A2, titanomagnetite and magnetite for B components, and low unblocking temperatures minerals such as maghemite or sulfides for the C and D components. Coexistence of normal and reversed polarities of A1 and A2 magnetizations locate remagnetizations before or after the Kiaman reversal, i.e. Middle Carboniferous or Late Permian–Early Triassic. The second time span coincides with 270–231 Ma K–Ar ages from TAZ andMAW (Fig. 1c). Since a small part of the B components is syn-folding, the remagnetization is associated with the late Triassic–Jurassic shortening [2] while post-tectonic C and D components were acquired after the late Jurassic tightening. Assuming that the A1, A2 and B overprints occurred when the PAO units were tilted by ∼35◦ to the southwest, the corrected APWP of the MAW and TAZ overlap with the APWP of North China, implying that they have a common post-Permian evolution. The A1–A2 and the A2–B APWP reveal successive counterclockwise rotations by ≤140◦ (Fig. 1d). The C–D path involves a slight clockwise rotation associated with a northward drift of the TNC, PAO units and Amuria. These data can be compared with new structural and geochronological data from the MAW and TAZ which show that their Devonian to late Carboniferous N–S trending accretionary fabric was reworked by almost orthogonal E–W trending sinistral shear zones (Fig. 1e), asymmetrical ‘S-shaped’ folds (Fig. 1f) and conjugate but namely NE–SW trending sinistral shear zones (Fig. 1g) associated with E–W folding of late Permian to Jurassic successions.The collision of theTAZ and MAWwas estimated by 280–260MaU– Pb zircon and monazite ages from a N–S trending pegmatite dyke swarm (Fig. 1e) indicating initial pure shear shortening of the PAO units [7]. These dykes and hosting units were later asymmetrically folded and transposed by sinistral shear zones forming a structural and geophysical corridor between the TNC and MB [8]. The 40Ar/39Ar and K–Ar data show three peaks at ∼270–260, 250–230 and 220–160 Ma [9–11] that are all related to: (i) mid-Permian shortening and pegmatite dyke intrusion and (ii) Triassic sinistral shearing associated with the shortening and anticlockwise rotation of pure shear dominated ‘S’-shaped lithons (Fig. 1d lower-left inset) and development of conjugate late Triassic–Jurassic discrete shear zones [9], respectively. The above timescales are mirrored by three-stage tectonic evolution of the Solonker suture marked by the formation or the Permian ophiolite mélange where greenschist metamorphism and deformation were dated at 269–255 Ma [12,13], Mid-Triassic NNE-directed thrusting and folding and Late Jurassic SEand NW-directed thrusting [14]. All these data can be interpreted as a result of: (i) Permian exhumation of