Kinematics of deformation in the Tibetan Plateau and its margins constrained by GPS measurements

As the most prominent example of large-scale continental deformation, Tibetan Plateau offers an ideal natural laboratory for quantifying such deformation and understanding the relevant dynamic processes. GPS provides a powerful means to directly measure the kinematics of present-day deformation. Our synthesis of GPS velocities from 553 stations in Tibetan Plateau and its margins quantitatively show that most of the relative India/Eurasia motion has been accommodated primarily by crustal shortening along the margins, strike-slip and normal faulting in the plateau interior, and clockwise rotation around the eastern end of Himalaya. The eastward extrusion of Tibetan Plateau out of India’s northward pass is carried out by roughly eastward flow of crustal material rather than by rigid block rotation. To the east, the eastward flow of crustal material causes shortening across the eastern margin of the plateau and clockwise rotations where resistance to such flow may be weak. The present-day tectonics in the Tibetan Plateau is best described as deformation of a continuous medium, at least when averaged over distances of ~100 km. Introduction The Tibetan Plateau, the world’s largest and highest plateau, has been growing and evolving since the collision and subsequent penetration of the India with Eurasia 50 million years ago (Molnar and Tapponnier, 1975; Rowley, 1996). How the Tibetan Plateau deforms in response to the collision, however, remains enigmatic and subject to debate, with rigid plates or blocks, continuous deformation of the entire lithosphere, and flow in the lower crust providing keys to its understanding (Tapponnier et al., 1982, 2001; England and Houseman, 1986; Molnar et al., 1993; England and Molnar, 1997; Royden et al., 1997; Holt et al., 2000; Flesch et al., 2001). Unfortunately, much of the region is remote, and a complete kinematic description of deformation over the entire Tibetan Plateau has not been available until recently. Dynamic models intended to explain deformation of the Tibetan Plateau need to be tested in terms of kinematics. Global Positioning System (GPS), developed in last decade with the successful operation of International GPS Service (IGS), provides a powerful means to directly measure the kinematic pattern of present-day crustal deformation in remote and large-scaled region like the Tibet (Chen et al., 2001; Wang et al., 2001; Chen et al., 2003; Calais et al., 2003; Zhang et al., 2004). We, in this paper, synthesis GPS studies in the Tibetan Plateau and its margins (Paul et al., 2001; Wang et al., 2001; Banerjee and Burgmann, 2002) to show in which ways the collision between India and Eurasia is accommodated and to shed new insights into the dynamics of its contemporary tectonic deformation. We designate this paper to the decadal anniversary of IGS for its excellent service to the researches of geodynamics. Data and data process Significant advancement for the monitoring of crustal deformation in the Tibet Plateau was accomplished in 1998 when the Crustal Motion Observation Network of China (CMONOC) was established. The principal data used for this study come from the CMONOC collected during 1998 and 2002, including 25 continuous stations, 56 annually observed stations with an occupation of at least 7 days (~168 hours’ data collection) in each survey, and 961 regional stations observed in 1999 and 2001 with an occupation of at least 3 days (~72 hours’ data collection) in each survey. The data were processed in four steps (Shen et al., 2000, 2001). First, we put the observation data together to solve for the daily loosely-constrained station coordinates and satellite orbits using the GAMIT software (King and Bock, 1995). Second, we combined the regional daily solution with the loosely constrained global solutions of ~80 IGS tracking stations produced at the Scripps Orbital and Position Analysis Center (Bock et al., 1997) using the GLOBK software (Herring, 1995). The merged daily solution includes the loosely constrained station coordinates, polar motion and satellite orbit parameters, and the variance-covariances matrix. Third, we estimated station positions and velocities in the ITRF2000 reference frame using the QOCA software (Dong et al., 1998). The QOCA modeling of the data was done through sequential Kalman filtering, allowing adjustment for global translation and rotation of each daily solution. In the last step, we transformed the velocity solution to a Eurasia-fixed reference frame using the angular velocity of Eurasia with respect to the ITRF deduced from 11 IGS stations (NYAL, ONSA, HERS, WSRT, KOSG, WTZR, VILL, GLSV, IRKT, TIXI) in the stable Eurasia plate [Shen et al., 2000; 2001]. Besides the CMONOC data, we collected three additional data sets of station velocities from Paul et al. [2001], Wang et al. [2001], and Banerjee and Burgmann (2002) to increase the coverage and station density of the India, Himalayan and central Tibetan regions. Paul et al. (2001)’ s velocity data (13 stations in India and the Himalaya) are in an India-fixed reference frame, whereas Wang et al. (2001)’s velocity data (41 stations distributed in India, the Himalaya, and central Tibet) and Banerjee and Burgmann’ (2002)’s (24 stations in the western Himalaya) are in a Eurasia-fixed reference frame, which differs slightly from the above Eurasia-fixed reference frame we employed. As each of the additional velocity data sets has some common stations with the CMONOC data set, we chose common stations for the three data sets, to transform them to the Eurasia-fixed reference frame of the CMONOC data set, by minimizing the velocity differences of the common stations in the corresponding reference frames. After the transformation, the maximum difference of the velocities for each common station in different data sets is less than 2.9 mm/yr and 2.6 mm/yr for the east and north components, respectively, which are within the 2 standard deviations of the velocity components. Thus, we calculated the weighted average of the velocity components for the common stations and estimated their standard deviations. We finally obtain velocities for 553 stations in the Tibetan Plateau and its margins, that provide adequate coverage to interpret the magnitude and style of deformation despite of void areas in northwestern corner of the plateau (Fig. 1). Figure 1. GPS velocity vectors (mm/yr) in the Tibetan Plateau and its margin with respect to the stable Eurasia, The ellipses denote the region of 1-sigma error. The polygons show locations of the profiles, and GPS stations covered by each profile. Shortenings across the Tibetan Plateau and its margins The Tibetan Plateau and its margins, including the Himalaya, the Altyn Tagh and the Qilian Shan, undergo substantial shortening. We draw four profiles (A-A’, B-B’, C-C’ and D-D’) across the plateau along the N20°E direction, the inferred India-Eurasia convergence direction (Sella et al., 2002), to calculate the shortening across different parts of the plateau. The total shortenings between India and Tarim in this direction are 28±2.5, 33±2.0, and 34±3.0 mm/yr along profiles D-D’, C-C’ and B-B’ respectively (Fig. 2). The shortening seems to be slightly larger than 34±4.0 mm/yr, if the Shilong Hill is regarded as part of the Himalaya, between the India and Gobi Alashan along profile A-A’ (Figs. 1&2). Taking 36-40 mm/yr as total relative motion between India and Eurasia, the eastern Tibetan Plateau and its margins (Profiles A-A’ and B-B’) accommodates 85-94% of the total motion, whereas western Tibet (Profiles C-C’ and D-D’) absorbs 70-91% of total convergence and the rests are taken up by shortening across the Tianshan in the north (Abdrakhmatov et al. 1996; Reigber et al., 2001). Figure 2. Velocity components parallel to N20oE vs. distance in km along each of 4 profiles in Fig. 1. Squares, diamonds, crosses and triangles show profile A-A', B-B', C-C' and D-D' respectively. Partitions of the shortenings in the Himalaya, the northern margin, and the plateau interior The N20oE convergence across the western Himalaya is 16±2.5 mm/yr along profile D-D’ (80°-84°E), slightly less but approximating previous geological (Lavé and Avouac, 2001) and geodetic (Bilham et al., 1996; Larson et al., 1999; Paul et al., 2001; Banerjee and Bürgmann, 2002) findings. To the east, the shortening rates across the Himalaya are 15±3.0 and 14±3.0 mm/yr along profiles C-C’ and B-B’ respectively (Fig. 2). The shortening along profile A-A’ is not well constrained because of the poor station coverage and the influence by rotation of crustal material around the eastern Himalaya syntaxis (Fig. 1). We estimate 15-20 mm/yr shortening rate across the Himalaya along profile A-A’. The northern margin of Tibet absorbs only slow convergence: 5.3 ± 1.0 and 6.2 ± 1.5 mm/yr parallel to N20°E on profiles C-C’ and B-B’ (Figs. 2 and 3), which includes left-lateral strike slip at 5.6 ± 1.6 and 5.0 ± 2.0 mm/yr parallel to the Altyn Tagh fault and 2.9 ± 1.8 and 3.2 ± 1.5 mm/yr of convergence perpendicular to the margin of Tibet (Fig. 3). These strike-slip and convergence rates along the eastern third of the northern plateau margin (profiles B-B’ and C-C’) are consistent with geological (Working Group on the Altyn Tagh fault, 1993) and other geodetic (Bendick et al., 2000; Wang et al., 2001; Shen et al., 2001) results. Shortening occurs at 6.0 ± 1.5 and 5.5 ± 1.8 mm/yr across the northeastern edge of the Tibetan Plateau perpendicular to the western and eastern Qilian Shan respectively (Fig. 3). The amounts of shortening must be accommodated by the plateau interior are 11.3±5.0, 14±3.0, 12.7±3.0, and 10±3.0 mm/yr along profile A-A’, B-B’, C-C’, and D-D’ respectively (Figs. 2and 3). These velocity profiles show general feature of linear gradients except profile A-A’ that may be significantly affected by rotations around the eastern Himalaya syntaxis (Figs. 1 & 3). However, views on how the shortening in the plateau inter