Partitioning of India-Eurasia convergence in the Pamir-Hindu Kush from GPS measurements

The study investigates how India–Eurasia convergence is partitioned across the Pamir–Hindu Kush region and whether deformation mechanisms show scale dependence similar to Tibet. Continental collision zones exhibit lateral and vertical variations in rheology, leading to differing modes of strain localization from fault-scale elastic behavior to range-scale plastic wedge dynamics and plateau-scale viscous flow. The Pamir region (NW Pakistan, Afghanistan, Tajikistan, Kyrgyzstan) provides an opportunity to test whether viscous, distributed deformation can characterize regions smaller than the Tibetan Plateau (<1000 km). By collecting and analyzing GPS velocities, the authors aim to constrain regional kinematics, quantify slip rates on major faults, and assess the contribution of distributed deformation to the total India–Eurasia convergence.

The paper situates its work within established models of continental deformation across multiple length scales, including elastic dislocation models for faults (Okada, 1985), plastic wedge mechanics for mountain belts (e.g., Dahlen et al., 1984), and viscous thin-sheet models for plateaus (England and McKenzie, 1982; England and Houseman, 1986; England et al., 1985). Prior geodetic studies constrained deformation in Central Asia and Tibet (e.g., Reigber et al., 2001; Gan et al., 2007; Chen et al., 2004; Zhang et al., 2004), while geologic investigations documented slip on major regional faults (e.g., Burtman and Molnar, 1993; Burtman et al., 1996; Arrowsmith and Strecker, 1999; Lawrence et al., 1992) and mapped Quaternary activity (Ruleman et al., 2007). Comparisons to Tibet highlight that a substantial fraction of convergence can be accommodated by distributed deformation rather than large discrete faults, and that present-day lateral extrusion of rigid blocks is limited by low slip rates on bounding faults (Zhang et al., 2007; Kirby et al., 2007).

• GPS network: Installed a regional network of continuous GPS sites between 30–44°N and 60–76°E encompassing eastern Afghanistan, Tajikistan, NW Pakistan, Baluchistan, and western Kyrgyzstan. Two IGS sites (KIT3, POL2) supplemented the network; Kabul (KBUL) was occupied episodically (93 days in 2006 and 10 days in 2008). Site details and time series are provided in auxiliary material.
• Data processing: Daily site positions estimated with GAMIT (v10.3) in a loosely constrained global frame including 14 IGS sites. Velocities were determined using GLOBK in ITRF05 and transformed to Eurasia-fixed and India-fixed reference frames (poles given). Error ellipses correspond to 95% confidence.
• Fault compilation: Major active faults were digitized from published sources (e.g., Burtman and Molnar, 1993; Ruleman et al., 2007; Ambraseys and Bilham, 2003). For each major fault or fault system, the relative GPS velocity between geodetic sites bracketing the fault was used as an upper bound on the fault-parallel and fault-normal slip components, neglecting elastic strain accumulation and potential block rotations. This assumption is more robust for strike-slip faults than for dip-slip faults.
• Distributed deformation zones: In regions lacking clearly mapped major active faults (Hindu Kush and central Pamir), velocities were decomposed into components parallel and normal to the regional E–W topographic grain using GPS site pairs spanning the region.
• Frame reconciliation for Pamir extension: East–west extension across the central Pamir was estimated using the velocity difference between Tashkurgan (TASH; from Gan et al., 2007) and Khorog (MANM). A local transformation between the two Eurasia-fixed frames (this study and Gan et al., 2007) was computed using common sites (POL2 and SELE) to ensure consistency.
• Uncertainty treatment: Quoted uncertainties for slip-rate bounds derive from GPS velocity uncertainties only; they do not include systematic errors from the simplifying assumptions (e.g., ignoring elastic strain fields and rotations).

• Total convergence: 29 ± 1 mm/yr between the NW Indian plate and Asia across the region.
• Partitioning on major structures (upper bounds, ignoring distributed strain/rotations): • Thrusting north of the Peshawar Basin: 13 ± 1 mm/yr. • Shortening in the Alai–South Tien Shan: 12 ± 2 mm/yr. • Sinistral shear on the northern Chaman–Gardiz–Konar system: 18 ± 1 mm/yr (with 5.4 ± 2 mm/yr across Gardiz+Mokur between QTAG–KBUL; up to 18.1 ± 1 mm/yr total across Gardiz–Mokur–Konar between KBUL–NCEG). • Sinistral shear on the Darvaz–Karakul Fault (DKF): 11 ± 2 mm/yr (geodetic bound 11.4 ± 2 mm/yr between MANM–SHTZ; consistent with geologic 10–15 mm/yr). • Slip rates on Herat and Talas–Ferghana faults: small, <2 mm/yr (geodetic for TFF 0.4 ± 2 mm/yr OSHK–POL2).
• Distributed deformation (regions without clear major active faults): • Hindu Kush and central Pamir: shortening not attributable to known faults of 16 ± 2 mm/yr, with central Pamir E–W extension of 9 ± 2 mm/yr (8.8 ± 2 from TASH–MANM). • South of Khorog: N–S shortening 13.2 ± 1 mm/yr with 9.9 ± 1 mm/yr sinistral shear (NCEG–MANM). • Central and northern Pamir: 16.2 ± 1.6 mm/yr shortening between MANM–GARM; total shortening NCEG–GARM is 31.8 ± 1.5 mm/yr but likely includes margin faults. • Tajik Depression: 6.2 ± 1 mm/yr shortening between MANM–SHTZ; additional 6.1 ± 1 mm/yr between SHTZ–KIT3 across the SW Gissar Range. • Central Alai and northernmost Pamir: 11.8 ± 2 mm/yr shortening between MANM–OSHK (cf. 13 ± 4 mm/yr east Alai from Reigber et al., 2001; ≥6 mm/yr geologic along central MPT).
• Additional kinematic constraints: • Herat Fault region: present-day 6.42 mm/yr sinistral shear KBUL–SHTZ (opposite to mapped dextral sense), with 5.1 ± 2 mm/yr shortening, suggesting lack of active dextral slip on the Herat Fault. • Karachi–Peshawar convergence is low (<3 mm/yr), implying Himalayan convergence localizes north of the Peshawar Basin.
• Overall: The diversity of strain styles (localized on major faults vs. distributed within plateaus) and measured rates indicate significant mechanical heterogeneity and show the Pamir behaves more like Tibet than a simple linear orogenic belt.

The GPS-derived velocity field reveals four kinematic regimes: (1) high relative rates (>10 mm/yr) across major faults (e.g., northern Chaman system), (2) high rates in regions lacking major through-going faults (e.g., central Pamir), (3) low rates (<5 mm/yr) across some mapped major faults (e.g., Talas–Ferghana, Herat), and (4) low rates within broadly deforming regions without obvious major faults (e.g., Tajik Depression). These regimes mirror the Himalayan–Tibetan system, where large faults girdle high, low-relief plateaux and interior deformation is distributed among many shorter, lower-rate structures. The sum of slip on major faults does not explain the full India–Eurasia convergence; a substantial fraction is accommodated by distributed deformation. Low present-day slip rates on bounding faults argue against rapid lateral extrusion of rigid blocks. Collectively, the results support scale-dependent continental rheology that includes viscous behavior even at the Pamir’s smaller length scale, and underscore the role of mechanical heterogeneity in governing strain localization.

This study provides new geodetic constraints on how India–Eurasia convergence is partitioned across the Pamir–Hindu Kush. Using a regional GPS network and consistent reference frames, the authors establish upper bounds on slip rates of major faults and quantify significant distributed shortening and E–W extension within the Pamir. The findings demonstrate that, despite its smaller size, the Pamir shares key kinematic and dynamic traits with Tibet, with both localized fault slip and pervasive distributed deformation contributing to convergence. These observations bolster scale-dependent rheological models incorporating viscous behavior for continental lithosphere. Future work should densify the geodetic network to resolve block rotations and elastic strain fields, integrate InSAR with GPS for improved spatial coverage, refine active fault mapping within the plateau interiors, and couple geodetic constraints with geodynamic modeling to better isolate contributions of individual structures versus distributed strain.

• Slip-rate estimates are upper bounds derived from bracketing site velocities; they neglect elastic strain accumulation and fault geometry effects, particularly important for dip-slip faults.
• Possible block rotations are ignored due to sparse station coverage, which can bias fault-parallel/normal decompositions.
• Uncertainties reported reflect GPS velocity errors only and do not include systematic errors from the simplifying assumptions.
• The network density is insufficient to isolate shortening within the Pamir from that along its margins in several transects (e.g., NCEG–GARM).
• Frame reconciliation for extension across the Pamir relies on transforming velocities from Gan et al. (2007) using common sites; residual frame inconsistencies may exist.
• One key station (KBUL) was occupied episodically rather than continuously, potentially increasing velocity uncertainty.