Paleogene India-Eurasia collision constrained by observed plate rotation

When India collided with Eurasia and how this process varied along the strike of the Indian front have long been debated. An abrupt deceleration of India’s convergence at ca. 50 Ma has often been taken as a marker for initial collision, yet stratigraphic, sedimentologic, metamorphic, and paleomagnetic records along the suture yield inconsistent ages. A diachronous collision has been proposed, with suggested initiation in the west, center, or east, but carries large uncertainties. Resolving the Paleogene collision history is critical to understanding Tibetan Plateau tectonics, uplift, and associated climatic evolution such as the Asian monsoon. The Indian plate’s asymmetric distribution of continental and oceanic lithosphere along the convergent boundary makes its kinematics highly sensitive to collision-related forces and torques, implying that rotation history could encode information about collision processes. This study asks whether observed Cenozoic rotation of the Indian plate can constrain the timing and along-strike pattern (diachroneity) of the India–Eurasia collision and the subsequent coupling between the plates.

Prior work recognized a slowdown in India–Eurasia convergence around 50 Ma and inferred collision timing from diverse proxies, but results vary widely along strike with proposals of western, central, or eastern initiation. Paleomagnetic studies noted significant Cenozoic rotation of India, but lacked continuous quantitative rotation histories and geodynamic explanations. Multiple global plate reconstructions provide differing absolute reference frames (e.g., moving hotspots vs slab-fitting) but broadly similar relative motions. Reconstructions considered here (Müller et al., 2016, 2019; Torsvik et al., 2019; van Hinsbergen et al., 2021) consistently reveal two Paleogene rotation-rate peaks, though magnitudes and an additional ca. 60 Ma peak vary with reference frame. Geological observations across the Himalaya and Tibet, including sediment provenance shifts and high/ultrahigh-pressure metamorphism, suggest collision onset between ~55 and ~40 Ma, but with along-strike variability and debate over diachroneity. Alternative geodynamic scenarios include double subduction and arc–continent collision (Kohistan–Ladakh, West Burma) and pre-collisional extension of Greater India; their expected torque signatures can be compared to the rotation history.

The study combines observational kinematics and geodynamic modeling. Rotation-rate calculation: Using multiple recent global plate reconstructions, three reference points on the Indian continent are tracked back to 66 Ma. The azimuth of the great circle connecting northern and southern points defines plate orientation; rotation rate is computed as the azimuthal change normalized by the traveled distance, mitigating artifacts from variable plate speed. Four reconstructions (Müller et al. 2016, 2019; Torsvik et al. 2019; van Hinsbergen et al. 2021) are analyzed and averaged to identify robust rotation-rate features. Numerical modeling: Fully dynamic 3D spherical finite-element models (CitcomS) simulate subduction and collision in a free-slip shell (72° × 82° × 2890 km; finest resolution ~10×8×8 km near subduction). The inner domain includes parts of eastern Asia, India, Neo-Tethys, and Indian Ocean, initialized from reconstruction geometry with an initially straight, E–W-trending trench consistent with geological constraints. The overriding plate is fixed relative to the western boundary and deformable to the east, mimicking rigid western Eurasian blocks and weaker eastern boundaries influenced by Pacific and Neo-Tethys subduction. Rheology: Depth-, temperature-, strain-rate-, and composition-dependent viscosity with diffusion and dislocation creep plus pseudo-plastic yielding (Drucker–Prager), layered viscosity structure (lithosphere 10^22–10^23 Pa s; asthenosphere/transition zone 10^19–10^20 Pa s; lower mantle 10^22 Pa s), weak subduction interface (8×10^18 Pa s), and temperature/composition/phase-change-dependent density. Lithosphere is multi-layered for continents (upper/lower crust, mantle lithosphere) with cratonic vs non-cratonic Eurasia distinguished; oceanic lithosphere has crust and mantle layers. Initial thermal structure uses half-space cooling with representative ages. Collision scenarios: Three end-member geometries of Greater India along its northern margin are tested—west-bulging, middle-bulging, and east-bulging—leading to diachronous suturing from the initial contact location across the margin. Diagnostics: Time series of Indian plate rotation rates are extracted from each model and compared to observed rotation histories. Simplified torque calculations: First-order analytical estimates compute the evolution of resistive torque as the product of contact width (resistive force per unit length assumed constant) and lever arm length from the plate’s moving center of mass, exploring how bulge location controls torque and expected rotation-rate peaks.

• Observed rotation history: The Indian plate rotated ~15.8° counterclockwise since the early Cenozoic. Two prominent Paleogene rotation-rate peaks are identified across reconstructions: ca. 52–44 Ma and ca. 33–20 Ma. Peak rates reach ~0.006°/km (first) and ~0.010°/km (second), exceeding the mean background ~0.003°/km.
• Robustness across reconstructions: Despite differing reference frames (moving hotspots vs slab-fitting), the two-peak pattern persists; averaging four reconstructions preserves timing and magnitudes similar to Müller et al. (2019). A ca. 60 Ma peak appears in some models but is not universal and is not analyzed further.
• West-bulging model: Produces two rotation-rate peaks with magnitudes 0.0067°/km (first) and 0.0095°/km (second), closely matching observations. Peaks postdate the initial and complete collisions by ~8 Myr, respectively. Mechanism: During initial diachronous collision, increasing collisional contact width boosts resistive force faster than lever arm shortens, increasing torque and yielding the first peak; as suturing completes, lever arm reduction temporarily lowers torque; after full collision, increased gravitational potential energy, upper-plate deformation, and interplate coupling raise resistive force, producing the second peak.
• Middle-bulging model: Also yields two peaks at 0.0049°/km and 0.0080°/km, lagging initial and complete collisions by ~6 Myr. First-peak amplitude is lower due to shorter lever arm; second-peak amplitude reflects E–W asymmetry in resistive and slab-pull forces.
• East-bulging model: Produces only one peak (~0.0061°/km), significantly below observed first-peak magnitude; initial collision east of the center of mass induces a transient clockwise torque that suppresses an early counterclockwise rotation peak. This scenario fits observations worst and is least plausible.
• Collision timing inferred from rotation: Given the modeled lag of 5–10 Myr between collision stages and rotation peaks, the initial India–Eurasia collision likely began at 55 ± 5 Ma along the western–central margin, with suturing propagating eastward to a complete collision by 40 ± 5 Ma.
• Geological consistency: These ages align with sediment provenance shifts and metamorphic ages in southern and western Himalaya (e.g., ~59–51 Ma onset indicators; UHP eclogites peaking ~47–46 Ma implying earlier collision), less constrained eastern Himalaya metamorphism initiating ~40 Ma, and late Eocene–early Oligocene India–Burma collision. The second rotation peak (33–20 Ma) coincides with strong plate coupling evidenced by southeastern Tibet extrusion/rotation and South China Sea spreading (~33–16 Ma).
• Broader kinematic context: Similar rotation-rate peaks in the Cretaceous correlate with Gondwana breakup events and LIP emplacement (Comei/Bunbury ~132 Ma; Morondava ~90 Ma; Deccan ~65 Ma), suggesting plume-induced torques and diachronous collision torques both drive anomalous plate rotations.
• Alternative scenarios: Double subduction and late arc–continent collisions after ~60 Ma predict different rotation signatures inconsistent with the observed Cenozoic double-peak pattern; earlier arc collision (<~60 Ma) remains possible but effects may be obscured by late Cretaceous plume dynamics.

By linking rotation-rate peaks of the Indian plate to torque evolution during collision, the study provides a kinematic diagnostic for constraining the timing and along-strike progression of the India–Eurasia collision. The presence of two counterclockwise rotation peaks, their magnitudes, and their temporal offsets relative to modeled collision stages support a diachronous initial collision starting in the western–central Greater India margin around 55 ± 5 Ma and culminating in complete suturing by 40 ± 5 Ma. This reconciles conflicting stratigraphic and metamorphic age estimates along the suture by framing them within an along-strike propagating collision. The second peak (33–20 Ma) reflects post-collisional strengthening of plate coupling, consistent with Himalayan uplift, increased gravitational potential energy, large-scale extrusion and rotation of Indochina/Sibumasu blocks in SE Tibet, and contemporaneous South China Sea spreading. Comparison across bulge geometries indicates west- or middle-bulging Greater India shapes best reproduce observations, whereas an east-bulging configuration is inconsistent with the early peak. The rotation-based constraints challenge scenarios invoking late arc–continent collision or dominant double subduction into the Cenozoic, which would impart different torque histories, though earlier arc collision remains compatible. Overall, the findings underscore plate rotation history as a powerful integrative constraint on abnormal plate kinematics and the forces/torques driving them, advancing understanding of Tibetan Plateau tectonics and associated climatic implications.

The study introduces plate rotation-rate history as a quantitative constraint on the Paleogene India–Eurasia collision. Analyses of multiple reconstructions reveal two robust counterclockwise rotation-rate peaks (52–44 Ma and 33–20 Ma). Fully dynamic 3D geodynamic models and simplified torque analyses show that the first peak records a diachronous initial collision beginning in the western–central margin, and the second peak reflects enhanced coupling after complete suturing. Integrating modeled lag times with observations yields collision timing of 55 ± 5 Ma (initial) and 40 ± 5 Ma (complete), consistent with key geological records along the suture and with pronounced Oligocene–Miocene deformation in SE Tibet and South China Sea spreading. These results provide a coherent framework for the early tectonic evolution of the Tibetan Plateau and demonstrate the diagnostic value of rotation histories for interpreting abnormal plate kinematics. Future research could refine absolute reference frames and uncertainties in reconstructions, incorporate additional geological data to better resolve along-strike diachroneity, explore more complex multi-collision and double-subduction scenarios with data-assimilation, and test rheological parameter sensitivity to further constrain plate coupling evolution.

• Dependence on plate reconstruction reference frames: Differences between moving hotspot and slab-fitting frames affect absolute motions and the robustness of early peaks (e.g., ca. 60 Ma), though the two-peak Paleogene pattern is consistent across models.
• Model simplifications: Initial straight E–W trench geometry, simplified boundary conditions (rigid west, weaker east), and rotation of the model domain may not capture all natural complexities.
• Rheological assumptions: Parameterized viscosity structure, weak interface layer, reduced activation energy to account for slab weakening, and simplified continental rheology (temperature effect simplified) introduce uncertainty.
• Torque analysis idealizations: The analytical torque calculation assumes constant resistive force per unit length, simplified center-of-mass motion, and neglects some dynamic feedbacks; it provides first-order trends rather than precise predictions.
• Scenario coverage: End-member bulge geometries explore diachroneity but cannot fully encompass potential complexities (e.g., multiple collision events, double subduction interactions, pre-collisional extension magnitude) though authors argue these do not change first-order results.
• Geological constraints uneven: Eastern Himalaya and Myanmar records are less complete, adding uncertainty to along-strike comparisons and inferred diachroneity.