| GISdevelopment.net ---> Asian GPS --->Proceedings --->2002 --->GPS: Emerging Trends |
Crustal Shortenning and tectonics of the NW Himalaya from GPS measurements
Paramesh Banerjee
Wadia Institute of Himalayan Geology, Dehra Dun, India
Paramesh Banerjee
Wadia Institute of Himalayan Geology, Dehra Dun, India
Abstract
Data from two permanent GPS station and more than 25 campaign mode GPS stations in the NW Himalayan region, were processed. The region under study covers the seismic gap area between the 1905 Kangra earthquake and 1934 Bihar earthquake. GPS sites distributed over the Higher Himalaya as well as in the Lesser Himalaya show crustal shortening is taking place at nearly 15-20mm/yr rate within a narrow zone of 100-150 km north of the Himalayan Frontal Fault. Model inversion of the GPS data shows that nearly 50km length of the frontal thrust system is locked with the advancing India plate along the detachment surface, at depths of 6-10km. Kangra, Simla and DehraDun region along the NW Himalayan front show lateral variation in configuration and geometry of the decollement surface. Both the Kangra and Dehra Dun blocks offer greater resistence to the convergence process and are accumulating strain through the locked portion of the detachment surface, which will eventually be released through future earthquakes. The in-between Nahan block slides relatively smoothly over the detachment surface, and shows only 10 km length of locking, thus have less potential for big earthquakes.
Introduction
The topography, geologic structure and the earthquakes of the Himalaya are a consequence of the collision and subsequent convergence of the northward moving Indian plate with the Eurasian plate. Since late Cretaceous time, an estimated 2000-3000 km of convergence is believed to have taken place (Molnar and Tapponier, 1977). Current rate of the Indian plate has been estimated to be around 50mm/yr (Demetes et al, 1994; Freymuller et al, 1996; Larsen et al, 1999). There is a broad agreement that at least 25-30% of the total convergence is being accommodated within the Himalayan arc. Most of the estimations of crustal shortening happening within the Himalaya are from various geological and seismic evidences (Leathers 1987, Baker et al, 1988, Molnar and Deng, 1984, Armijo et al, 1986, Molnar and Lyon Caen 1989, Avouvac et all, 1998) range between 8-18mm/yr (Powers et al, 1998). Recent GPS measurements in Nepal Himalaya (Bilham et al, 1997, Larsen et al, 1999, Burgmann et all, 1999) have produced a rate of 18mm/yr of surface contraction taking place across the Nepal Himalaya. Our GPS measurement (discussed in this paper) from the western Himalayan region have brought out contraction rate of 14-21mm/yr between Higher Himalayan sites and Bangalore (South India), and a lower rate of 7-14mm/yr between these sites and Delhi, the northern tip of the Indian shield (Fig.1).
Fig.1 Surface relief map of the Indian subcontinent showing velocity vectors (relative to IISC, Bangalore) of three permanent GPS stations aloong the Himalayan arc (Nadi, WIH2, and NAGA). Also shown the IGS sites used as fiducial stations for the current GPS data processing. Representative sites of Higher Himalaya and Lesser Himalaya shows the amount of shortening happening along the Himalayan arc in the study area.
Needless to say, the 2500km long Himalayan convergence arc is much more complicated, and needs more wide coverage of GPS geodetic measurements to bring out the along-the-arc variations, as well as more details of the convergence geometry and mechanism. We carried out GPS measurements covering a major portion of the western Himalayan sector between 75°E to 80°E. This covers the entire Kangra reentrant and DehraDun reentrant. Out of the four great earthquakes (1897, 1905, 1934 and 1950) that have struck the Himalayan foothills over the last 100 years, the 1905 Kangra event (Middlemiss, 1910, Molnar, 1987) has produced nearly 100 km long rupture zone in HP and possibly another smaller rupture in Dehra Dun region. The region between 1905 Kangra and 1934 Bihar events, the region under present study, has been identified as a seismic gap where no major earthquakes have occurred during last 200-300 years (Bilham et al 1995, Khattri, 1987,Seeber et al, 1981, Yeats et al 1992).
An important step in using GPS geodetic techniques for India-Tibet convergence studies came from the Nepal GPS geodetic campaigns. Bilham et al, (1997) and Kristine et al (1999) used six years of GPS data in Nepal Himalayas to constrain maximum surface contraction rate to 18±2 mm/yr with a corresponding slip rate of nearly 20mm/yr. They also reported an eastward motion between Lhasa and Nepal at nearly 11 mm/yr, thus confirming that lateral extrusion is indeed taking place to accommodate the indenting rigid Indian plate through strike-slip faulting (Molnar and Tapponier, 1995). Burgmann et al,( 1999) proposed a segmented fault model for the Nepal Himalaya suggesting along-the-arc variation in the convergence process. The fault models for the east and west Nepal dipping at different angles (3-8°), shows locking at different depths (15-25 km). At least 500 km long stretch of the fault system with an width of nearly 140km were said to be locked, accumulating 6-15m of potential slip, which would eventually be released through a future great earthquake.
Isoseismals of 1905 Kangra event (Middlemiss, 1910) indicate two zones of strong ground shaking, one with maximum intensity=X around Kangra region, and another with max. intensity of VII around DehraDun region, leaving the in between Simla sector relatively unharmed. The preferential rupturing during the 1905 Kangra earthquake, reported upliftment of Dehra Dun region (Gehlaut and Chandra,1997), the sinuous nature of the Main Boundary thrust, varying thickness of the Sub-Himalayan ranges and the very existence of the Kangra and Dehra Dun re-entrants, intervened by a lateral basement ramp (Powers et al, 1998), presence of evaporites in some parts of HP, transverse alignment of MBT-PT in Kulu-Bilaspur region and several transverse lineaments picked up in satellite imagery perhaps suggests segmentation of Himalayan frontal belt into different blocks. It would be interesting to know if such segmentation exists and if exist, how different blocks respond to the stress developed due to the Indian plate movement and how the built-up energy is distributed between aseismic creep and seismic events. In the region covering both the Kangra and Dehra Dun re-entrants, we established two permanent stations and nearly 50 episodic GPS stations (Fig.2). In this paper, we have presented results from repeat GPS measurements covering both the region, including sites in the Siwaliks, lesser Himalaya as well as Higher Himalaya, since 1996 till date. Through model inversion of the GPS data, we have explained the distribution of velocity vectors through a three-segment dislocation model, representing the three adjacent Himalayan segments, viz. The Kangra reentrant, the Simla block, and the DehraDun reentrant. These three blocks are detached from one another by strike slip movement, and have different configuration of the detachment surface.
GPS Data Acquisition and Processing
We established nearly 50 GPS stations in the region covering the Indian states of Himachal Pradesh, Uttaranchal, and adjoining border areas of Jammu & Kashmir, Punjab and Uttar Pradesh (Fig.2 ). Out of these 50 stations, velocity vectors are computed for 26 stations, as many stations could not be re-measured, and some bad measurements were rejected after computing the vectors. Most of the GPS stations are monumented by inserting Bevis Steel Pin into hard rock. Small punched circular holes on top face of the steel pin are used as the measurement points. In few cases where the monuments were to be established in soil covered area, 2ftx2ftx6ft concrete structures were constructed underground, with a Bevis steel pin protruding from the top face of it, which is flush with the ground. All the campaign mode stations were occupied for 4-5 days during each campaign. Many of the stations that were established during 1998 and 1999, were reoccupied in 2000. Two permanent GPS stations were established at Dehra Dun, (Uttaranchal), and Naddi (Himachal Pradesh), and had been operating since the end of 1998. Both the permanent stations are equipped with Trimble 4000SSi receivers, and Trimble Choke ring antennas. All the campaign data were collected using Trimble 4000 SSE/SSi receivers. GPS data collections were initiated during 1995. Gradually more and more stations were added to the network.
Fig.2 More than 50 GPS stations were established in the entire NW Himalaya (red stars).
Red diomonds are sites to be installed in future. Black stars are permanent stations at DehraDun
and Naddi (Dharamsala). Total 26 sites were reoccupied, and results are discussed in this paper.
The GPS data were processed using the GAMIT/GLOBK suite of software (Ver.9.95 ) . GAMIT and GLOBK are a comprehensive suite of programs developed by MIT, Scripps Institution of Oceanography, and Harvard University for analyzing GPS measurements primarily to study crustal deformation. GAMIT uses the GPS broadcast carrier phase and pseudorange observable to estimate three-dimensional relative positions of ground stations and satellite orbits, atmospheric zenith delays, and earth orientation parameters. GLOBK is a Kalman filter whose primary purpose is to combine various geodetic solutions such as GPS, VLBI, and SLR experiments. It accepts as data, or "quasi-observations" the estimates and covariance matrices for station coordinates, earth-orientation parameters, orbital parameters, and source positions generated from the analysis of the primary observations. The input solutions are generally performed with loose a priori uncertainties assigned to all global parameters, so that constraints can be uniformly applied in the combined solution.
In addition to the GPS data collected from our study area, we used data from IISC, KIT3, POL2 and LHAS IGS stations, as well as NAGA permanent station data from Nepal (Fig.1). These IGS sites were constrained to 5mm for horizontal, and 10mm for vertical coordinates. The a-pripori coordinates of the IGS sites were taken from global IGS solutions in ITRF97 reference frame. Two regional permanent stations, WIH2 and NADI, were similarly constrained tightly, after their accurate coordinates were obtained in ITRF97, from separate processing of a subset of the continuous station data. Fig. 3 shows time series for the entire span of the WIH2 continuously monitoring station. The tight constraining of the fiducial stations are useful only for resolving ambiguities, as loose solutions are used for finally combining daily solutions through GLOBK. In the primary processing using GAMIT, apart from the site coordinates, we also estimated parameters representing the effects of the atmosphere, Earth's rotation, and motions of the satellites. We included in our analysis, daily combined solutions of global IGS stations processed and archived as 'h-file's by SOPAC. Reference frame were defined by minimizing the departure from the a-priori ITRF97 values of the coordinates of a subset of IGS core stations. We used 24 IGS stations for position and velocity system stabilizations. This resulted in 2.4mm of RMS for position, and 1.6mm/yr of RMS for velocity stabilization. As UT1 and pole position are defined as single global parameters in GLOBK, we allowed Markov process for earth orientation parameters to allow day-to-day variations. We also allowed Markov process with tight constraints for satellite orbital parameters as well as radiation pressure parameters to absorb a given degree of mis-modelling of the satellite orbital motions. At the end, we have allowed large tolerance for uncertainties in station height component, effectively making the Kalman filtering process insensitive to height. After combining IGS site solutions with our regional solutions on a daily basis, total 735 daily solution files were combined to produce the final velocities of all the 63 sites, including both regional and global sites. We extracted the sites lying within our study area. In Fig. 4, we show few important regional site vectors relevant to India-Asia convergence process
Fig.3 Time series of the residual of WIH2 Permanent station positions (N, E and Height components) in ITRF97.
Data were processed using GAMIT 9.95 software
Dynamics of the convergence process
We plotted the velocity vectors relative to IISC in Fig. 4. Velocities at NAGA (Nepal) are also shown. The Higher Himalayan sites have systematically much higher southward velocities compared to the sites lying within Lesser Himalayan zone. When taken IISC as fixed, Higher Himalayan sites (UDAI, JIPA, KOTH, HARS, BHAT, BADR) have velocities ranging between 14.14 - 21.42 mm/yr, with an average of 17.08±1 mm/yr. We have excluded JOSH and GARU sites which showed large velocities, perhaps due to post seismic strain relaxation of 1999 Chamoli earthquake. When velocities are taken with JNUC fixed, higher Himalayan sites have velocities ranging between 7.34 -8.38mm/yr, with an average of 10.27±1.1 mm/yr. We found JNUC -IISC baseline to have a contraction rate of 4.41±0.36mm/yr in N-S direction and -5.24±0.76mm/yr along E-W direction. Recently Paul et al (in press) reported from GPS measurements mostly concentrated in South Indian region, including JNUC site and few others, that Indian shield is 'stable within 7x10-9/yr'. Considering the distance between IISC and JNUC sites being 1714071km, 7x10-9/yr strain rate is equivalent to nearly 12mm/yr N-S contraction. In fact, Paul et al reported JNUC to have 'southward velocity' of 3.7±1.5mm/yr. The E-W expansion between IISC and JNUC can be explained by the fact that India has a anti-clockwise rotation with the pole of rotation of the angular velocity vector (.44±0.03° Myr-1 ) with respect to Eurasia lie at 25.6N and 11±9E.
Fig.4 Velocity vectors from GPS measurements in NW Himalaya and NAGA (Nepal) relative to IISC, Bangalore, superimposed on the surface releif map of the western and central Himalaya. Blue solid circles are earthquake locations.
The lesser Himalayan sites are having almost negligible movements relative to JNUC. This is surprising considering that JNUC represent the northern tip of the Indian shield. Due to lack of any other GPS measurement data from the Indian shield area, a proper Indian reference frame could not be defined. So far, it has been assumed that the entire Indian peninsula can be treated as a rigid shield without much intra-plate movements. On the basis of the current data, it can not be told with certainty how much component of the observed Bangalore-Delhi shortening is due to tectonic activity within the Indian peninsula, and how much could be attributed to the error arising out of computing vectors in an improper reference frame.
Tectonics of the area
Both the Kangra and Dehra Dun re-entrants and intermediate Nahan salient presents and interesting tectonic problem. The 1905 Kangra earthquake created a preferential ground damage pattern with maximum intensity isoseismals concentrated in Kangra and DehraDun region, leaving the intermediate Simla sector (Nahan salient) relatively less damaged (Middlemiss, 1910). It is not clear whether this was the result of two separate ruptures at Kangra and Dehra Dun (Molnar, 1987). The Main Boundary Thrust (MBT) is sinuous in nature, and the width of the Sub-Himalaya varies from nearly 80km in Kangra, to nearly absent in the Nahan salient, and again widens to nearly 30km in DehraDun region. Main Frontal Thrust (MFT) is the southern-most of the known Himalayan fault system (Nakata, 1989). The MFT is almost always parallel to the main Himalayan arc, and is not always exposed to the surface. The Sub-Himalaya is confined between the MBT and the MFT. A gentle, north dipping detachment surface separates the overlying allochthonous sedimentary strata, where convergence is accommodated through anticlinal growth and surface shortening (Powers et al, 1998, Lillie et al 1987) within the Sub-Himalaya. Presence of evaporites in the decollement level reduces the strength of the rocks, and thrusting can propagate far into the foreland, as has been seen in the Potwar plateau (Lillie et al, 1987, Baker et al, 1988). Evaporites were not found to be present at the decollement in the Kangra re-entrant region (Powers et al, 1998). Also, there is no proof that evaporites are not present within the Nahan Salient. Some patches of salt are exposed in the adjacent region between Kangra reentrant and Nahan salient region (Srikantia and Sharma, 1976). Powers et al (1998) studied the seismic and bore hole data in the Kangra re-entrant region and suggested that the decollement in the Kangra re-entrant has a greater strength of the overlying strata, and the decollement dips at nearly 2.5°N, with a depth varying between 4-7 km, from south to north. In the Dehra Dun re-entrant, the dip of the decollement is 6°N, and decollement depth varies between 3.5-6km.
INDEPTH seismic profiling has shown the existence of a discrete fault plane dipping at 9°, between 27.7° and 29° N in Nepal, where it reaches a depth of nearly 45 km. (Makovsky et al, 1996, Zhao et al, 1993). More recently, Wesnousky et al (1999) measured terrace deposit upliftment in DehraDun region, and inferred 6.9±1.8mm/yr vertical upliftment rate, which, assuming a 30° dipping frontal thrust, translates into horizontal shortening of nearly 12mm/yr.
Modeling of the GPS data:
As seen in Fig(4), we find that the sites lying within the Kangra re-entrant area have almost negligible movement relative to the JNUC site, which represent the northern tip of the Indian shield. Exceptions are three sites on the eastern border of the Kangra re-entrant, which shows a noticeable amount of movement directed across the general Himalayan arc trend. The sites to the north of the re-entrant, that are located in the Higher Himalayan regions, Udaipur, Kothi, and Jispa, have across the strike velocities of 7.34±0.6, 12.04±1, and 14.58± 1.1 mm/yr, relative to JNUC. The sites lying to the south of HFF, like MOND, QASI, SABA in the Dehra Dun region, and BOHA, MANU in the Kangra region have almost negligible movements relative to the JNUC site. We can assume the above rates for higher Himalayan sites as the surface contraction rates taking place within the Himalayan frontal zone. The Higher Himalayan sites in the eastern part of our network, in the Garhwal Himalayan region, e.g. BADR, JOSH, BHAT and HARS have rates of 8.38±1.37, 17.48±.99, 9.71±1.1 and 9.6±1.16 mm/yr respectively. The shortening rate increases with northward distance of the measurement sites from the MBT. We don't have any site further north to see where the shortening rate saturates. This would give us the total shortening taking place across the Himalayan system. The maximum shortening rate that we found for the northern-most site is 14.5mm/yr, at Jispa. The abnormally high value for the Joshimath (JOSH) site is perhaps due to post-seismic stress relaxation after the 1998 Chamoli earthquake.
The differential movement between the Indian shield, Lesser Himalayan sites and Higher Himalayan sites are caused by aseismic slip occurring at the detachment surface. GPS measurements in Nepal (Bilham et al, 1997) indicated existence of a locked >120 km wide detachment fault. The elastic strain accumulation is eventually released by a southward travelling dislocation along the previously unstressed low-angle Himalayan Frontal Thrust (Bilham et al, 1998, Brune, 1996). Assuming that the southern tip of the southward moving Himalayan thrust system is locked with the converging Indian plate along the detachment surface, we try to quantify the depth and geometry of the locked system, as well as to demarcate boundaries of the Himalayan thrust segments, through the dislocation model inversion of the GPS data. We assume that dislocation can be characterized by slip on rectangular dislocation plane buried in an homogeneous, isotropic, elastic half space (Okada, 1985). We used a constrained non-linear optimization algorithm to estimate the fault geometry, viz. Length, depth, width, dip, strike, location, and rates of strike slip and dip-slip movement that best fit the GPS data (Arnadittir & Segall 1994, Burgmann et al, 1997). Guided by the existent geological data, and results from the Nepal GPS studies, we applied constraints on the depth of the fault, range of the fault strike, and allowing dip-slip movement, or strike-slip movement, or both together, to find the minimum of Weighted Residual Sum of Squares (WRSS) given by:
WRSS= (dobs-dmod)T.cov -1. (dobs-dmod) …..(1)
Where dobs and dmod are observed and modeled displacements respectively, and cov is the covariance matrix (Burgmann et al, 1999).
Fig.5 Dislocation model obtained from the GPS data. Blue arrows are observed velocity vectors and yellow arrows are predicted vectors. Red circles show earthquake locations. Black lined rectangles represent modelled dislocation fault.
See text for details.
We evaluated models of increasing complexity, and the best fitting model is shown in Fig (5). Our modeled fault system consist of three adjacent fault segments with varying widths, dip and depths, that best represent the morphological and geological characteristics of the study area like the sinuous nature of the MBT and existence of the Kangra and Dehra Dun re-entrants intervened by the lateral structural ramp in the Nahan Salient. The best-fit model have only dip-slip motion (18mm/yr) for the eastern and western blocks, whereas the in-between block have some strike-slip motion (10 mm/yr), which is in consistence with the fact that the MBT-PT alignment becomes almost transverse to the Himalayan arc on the eastern part of the Kangra re-entrant. We assumed a sufficiently large width of the faults, and the lateral extents of the two boundary faults were sufficiently large to avoid the edge effect . In consistence with the geological findings as discussed above, the Kangra block has the most gentle slope of the decollement (6°), where as the other two blocks have 10.5° and 10° dipping detachment surface. We found the depth to the locked portion of the fault to vary from 6 km in the Kangra block to almost negligible value of 2km in the central block, and 10 km in the Dehra Dun block. The black front line of the rectangular fault model shows the locking line at the detachment surface. With the applied dip angels of 6.5°, 10.5° and 10°, the detachment surface projects to 50km, 10km and 55km further south, which approximately match with the surface exposures of the Himalayan Frontal Thrusts. Nearly 55km length of the detachment surface is locked both in Kangra and Dehra Dun segments, whereas the length of that in the central sector is only 10 km.
Conclusion
As seen if Fig (5), the seismic activities are confined mostly to the northern bound of the locked portion of the detachment surface. Further north, the shortening is accommodated through asiesmic slip. It also appears that though we don't have sites further north to prove, our northern most site like JIPA perhaps already reaches the northern boundary where most of the shortening is talking place. It is also evident from our analysis that a sizable portion of the Frontal Thrust system is locked which will eventually be released through major seismic strain relief, though the dimension of the locked portion is not as great as was found in Nepal. Variation from the Nepal GPS studies, and also lateral variation within our local study area brings out the more pronouncedly the existence and role of the segmented nature of the Himalayan frontal thrust system. It is possible that ruptures caused by great earthquakes are limited to these segments. The Kangra, Nahan and Dehra Dun blocks response differently to the convergence process. The Kangra block, offers the maximum resistance to the advancing Indian plate along the detachment surface, and resist the southward progression of the overlying allocthonous sedimentary strata. The detachment surface has a gentler slope and is locked at nearly 6km depth. The central Nahan block has the least strength, and offers the least resistance to the convergence process. The frontal thrust system are easily carried further south, with a negligible locking depth of 2km, and length of 10 km. The Dehra Dun falls into in-between category between these two extreme cases. However, due to steeper dip angle of the detachment surface, and greater locking depth, both the Dehra Dun and Kangra region have almost same length of the locked portion, and same degree of earthquake potential.
Acknowledgement
The studies were carried out under a funding received from the DST, Govt. of India,, sponsored project. The paper is published with the permission of the Director, WIHG. I gratefully acknowledge participation in GPS field data acquisition campaigns by Sri CP Dabral, Dr.SatyaPrakash, Sri Sanjay Sood, and others. Helps and suggestions received from Dr. Roland Burgmann, UC-Berkeley, USA, and Dr. Roberts King, MIT, Boston, USA, are gratefully acknowledged.
References
Data from two permanent GPS station and more than 25 campaign mode GPS stations in the NW Himalayan region, were processed. The region under study covers the seismic gap area between the 1905 Kangra earthquake and 1934 Bihar earthquake. GPS sites distributed over the Higher Himalaya as well as in the Lesser Himalaya show crustal shortening is taking place at nearly 15-20mm/yr rate within a narrow zone of 100-150 km north of the Himalayan Frontal Fault. Model inversion of the GPS data shows that nearly 50km length of the frontal thrust system is locked with the advancing India plate along the detachment surface, at depths of 6-10km. Kangra, Simla and DehraDun region along the NW Himalayan front show lateral variation in configuration and geometry of the decollement surface. Both the Kangra and Dehra Dun blocks offer greater resistence to the convergence process and are accumulating strain through the locked portion of the detachment surface, which will eventually be released through future earthquakes. The in-between Nahan block slides relatively smoothly over the detachment surface, and shows only 10 km length of locking, thus have less potential for big earthquakes.
Introduction
The topography, geologic structure and the earthquakes of the Himalaya are a consequence of the collision and subsequent convergence of the northward moving Indian plate with the Eurasian plate. Since late Cretaceous time, an estimated 2000-3000 km of convergence is believed to have taken place (Molnar and Tapponier, 1977). Current rate of the Indian plate has been estimated to be around 50mm/yr (Demetes et al, 1994; Freymuller et al, 1996; Larsen et al, 1999). There is a broad agreement that at least 25-30% of the total convergence is being accommodated within the Himalayan arc. Most of the estimations of crustal shortening happening within the Himalaya are from various geological and seismic evidences (Leathers 1987, Baker et al, 1988, Molnar and Deng, 1984, Armijo et al, 1986, Molnar and Lyon Caen 1989, Avouvac et all, 1998) range between 8-18mm/yr (Powers et al, 1998). Recent GPS measurements in Nepal Himalaya (Bilham et al, 1997, Larsen et al, 1999, Burgmann et all, 1999) have produced a rate of 18mm/yr of surface contraction taking place across the Nepal Himalaya. Our GPS measurement (discussed in this paper) from the western Himalayan region have brought out contraction rate of 14-21mm/yr between Higher Himalayan sites and Bangalore (South India), and a lower rate of 7-14mm/yr between these sites and Delhi, the northern tip of the Indian shield (Fig.1).
Fig.1 Surface relief map of the Indian subcontinent showing velocity vectors (relative to IISC, Bangalore) of three permanent GPS stations aloong the Himalayan arc (Nadi, WIH2, and NAGA). Also shown the IGS sites used as fiducial stations for the current GPS data processing. Representative sites of Higher Himalaya and Lesser Himalaya shows the amount of shortening happening along the Himalayan arc in the study area.
Needless to say, the 2500km long Himalayan convergence arc is much more complicated, and needs more wide coverage of GPS geodetic measurements to bring out the along-the-arc variations, as well as more details of the convergence geometry and mechanism. We carried out GPS measurements covering a major portion of the western Himalayan sector between 75°E to 80°E. This covers the entire Kangra reentrant and DehraDun reentrant. Out of the four great earthquakes (1897, 1905, 1934 and 1950) that have struck the Himalayan foothills over the last 100 years, the 1905 Kangra event (Middlemiss, 1910, Molnar, 1987) has produced nearly 100 km long rupture zone in HP and possibly another smaller rupture in Dehra Dun region. The region between 1905 Kangra and 1934 Bihar events, the region under present study, has been identified as a seismic gap where no major earthquakes have occurred during last 200-300 years (Bilham et al 1995, Khattri, 1987,Seeber et al, 1981, Yeats et al 1992).
An important step in using GPS geodetic techniques for India-Tibet convergence studies came from the Nepal GPS geodetic campaigns. Bilham et al, (1997) and Kristine et al (1999) used six years of GPS data in Nepal Himalayas to constrain maximum surface contraction rate to 18±2 mm/yr with a corresponding slip rate of nearly 20mm/yr. They also reported an eastward motion between Lhasa and Nepal at nearly 11 mm/yr, thus confirming that lateral extrusion is indeed taking place to accommodate the indenting rigid Indian plate through strike-slip faulting (Molnar and Tapponier, 1995). Burgmann et al,( 1999) proposed a segmented fault model for the Nepal Himalaya suggesting along-the-arc variation in the convergence process. The fault models for the east and west Nepal dipping at different angles (3-8°), shows locking at different depths (15-25 km). At least 500 km long stretch of the fault system with an width of nearly 140km were said to be locked, accumulating 6-15m of potential slip, which would eventually be released through a future great earthquake.
Isoseismals of 1905 Kangra event (Middlemiss, 1910) indicate two zones of strong ground shaking, one with maximum intensity=X around Kangra region, and another with max. intensity of VII around DehraDun region, leaving the in between Simla sector relatively unharmed. The preferential rupturing during the 1905 Kangra earthquake, reported upliftment of Dehra Dun region (Gehlaut and Chandra,1997), the sinuous nature of the Main Boundary thrust, varying thickness of the Sub-Himalayan ranges and the very existence of the Kangra and Dehra Dun re-entrants, intervened by a lateral basement ramp (Powers et al, 1998), presence of evaporites in some parts of HP, transverse alignment of MBT-PT in Kulu-Bilaspur region and several transverse lineaments picked up in satellite imagery perhaps suggests segmentation of Himalayan frontal belt into different blocks. It would be interesting to know if such segmentation exists and if exist, how different blocks respond to the stress developed due to the Indian plate movement and how the built-up energy is distributed between aseismic creep and seismic events. In the region covering both the Kangra and Dehra Dun re-entrants, we established two permanent stations and nearly 50 episodic GPS stations (Fig.2). In this paper, we have presented results from repeat GPS measurements covering both the region, including sites in the Siwaliks, lesser Himalaya as well as Higher Himalaya, since 1996 till date. Through model inversion of the GPS data, we have explained the distribution of velocity vectors through a three-segment dislocation model, representing the three adjacent Himalayan segments, viz. The Kangra reentrant, the Simla block, and the DehraDun reentrant. These three blocks are detached from one another by strike slip movement, and have different configuration of the detachment surface.
GPS Data Acquisition and Processing
We established nearly 50 GPS stations in the region covering the Indian states of Himachal Pradesh, Uttaranchal, and adjoining border areas of Jammu & Kashmir, Punjab and Uttar Pradesh (Fig.2 ). Out of these 50 stations, velocity vectors are computed for 26 stations, as many stations could not be re-measured, and some bad measurements were rejected after computing the vectors. Most of the GPS stations are monumented by inserting Bevis Steel Pin into hard rock. Small punched circular holes on top face of the steel pin are used as the measurement points. In few cases where the monuments were to be established in soil covered area, 2ftx2ftx6ft concrete structures were constructed underground, with a Bevis steel pin protruding from the top face of it, which is flush with the ground. All the campaign mode stations were occupied for 4-5 days during each campaign. Many of the stations that were established during 1998 and 1999, were reoccupied in 2000. Two permanent GPS stations were established at Dehra Dun, (Uttaranchal), and Naddi (Himachal Pradesh), and had been operating since the end of 1998. Both the permanent stations are equipped with Trimble 4000SSi receivers, and Trimble Choke ring antennas. All the campaign data were collected using Trimble 4000 SSE/SSi receivers. GPS data collections were initiated during 1995. Gradually more and more stations were added to the network.
Fig.2 More than 50 GPS stations were established in the entire NW Himalaya (red stars).
Red diomonds are sites to be installed in future. Black stars are permanent stations at DehraDun
and Naddi (Dharamsala). Total 26 sites were reoccupied, and results are discussed in this paper.
The GPS data were processed using the GAMIT/GLOBK suite of software (Ver.9.95 ) . GAMIT and GLOBK are a comprehensive suite of programs developed by MIT, Scripps Institution of Oceanography, and Harvard University for analyzing GPS measurements primarily to study crustal deformation. GAMIT uses the GPS broadcast carrier phase and pseudorange observable to estimate three-dimensional relative positions of ground stations and satellite orbits, atmospheric zenith delays, and earth orientation parameters. GLOBK is a Kalman filter whose primary purpose is to combine various geodetic solutions such as GPS, VLBI, and SLR experiments. It accepts as data, or "quasi-observations" the estimates and covariance matrices for station coordinates, earth-orientation parameters, orbital parameters, and source positions generated from the analysis of the primary observations. The input solutions are generally performed with loose a priori uncertainties assigned to all global parameters, so that constraints can be uniformly applied in the combined solution.
In addition to the GPS data collected from our study area, we used data from IISC, KIT3, POL2 and LHAS IGS stations, as well as NAGA permanent station data from Nepal (Fig.1). These IGS sites were constrained to 5mm for horizontal, and 10mm for vertical coordinates. The a-pripori coordinates of the IGS sites were taken from global IGS solutions in ITRF97 reference frame. Two regional permanent stations, WIH2 and NADI, were similarly constrained tightly, after their accurate coordinates were obtained in ITRF97, from separate processing of a subset of the continuous station data. Fig. 3 shows time series for the entire span of the WIH2 continuously monitoring station. The tight constraining of the fiducial stations are useful only for resolving ambiguities, as loose solutions are used for finally combining daily solutions through GLOBK. In the primary processing using GAMIT, apart from the site coordinates, we also estimated parameters representing the effects of the atmosphere, Earth's rotation, and motions of the satellites. We included in our analysis, daily combined solutions of global IGS stations processed and archived as 'h-file's by SOPAC. Reference frame were defined by minimizing the departure from the a-priori ITRF97 values of the coordinates of a subset of IGS core stations. We used 24 IGS stations for position and velocity system stabilizations. This resulted in 2.4mm of RMS for position, and 1.6mm/yr of RMS for velocity stabilization. As UT1 and pole position are defined as single global parameters in GLOBK, we allowed Markov process for earth orientation parameters to allow day-to-day variations. We also allowed Markov process with tight constraints for satellite orbital parameters as well as radiation pressure parameters to absorb a given degree of mis-modelling of the satellite orbital motions. At the end, we have allowed large tolerance for uncertainties in station height component, effectively making the Kalman filtering process insensitive to height. After combining IGS site solutions with our regional solutions on a daily basis, total 735 daily solution files were combined to produce the final velocities of all the 63 sites, including both regional and global sites. We extracted the sites lying within our study area. In Fig. 4, we show few important regional site vectors relevant to India-Asia convergence process
Fig.3 Time series of the residual of WIH2 Permanent station positions (N, E and Height components) in ITRF97.
Data were processed using GAMIT 9.95 software
Dynamics of the convergence process
We plotted the velocity vectors relative to IISC in Fig. 4. Velocities at NAGA (Nepal) are also shown. The Higher Himalayan sites have systematically much higher southward velocities compared to the sites lying within Lesser Himalayan zone. When taken IISC as fixed, Higher Himalayan sites (UDAI, JIPA, KOTH, HARS, BHAT, BADR) have velocities ranging between 14.14 - 21.42 mm/yr, with an average of 17.08±1 mm/yr. We have excluded JOSH and GARU sites which showed large velocities, perhaps due to post seismic strain relaxation of 1999 Chamoli earthquake. When velocities are taken with JNUC fixed, higher Himalayan sites have velocities ranging between 7.34 -8.38mm/yr, with an average of 10.27±1.1 mm/yr. We found JNUC -IISC baseline to have a contraction rate of 4.41±0.36mm/yr in N-S direction and -5.24±0.76mm/yr along E-W direction. Recently Paul et al (in press) reported from GPS measurements mostly concentrated in South Indian region, including JNUC site and few others, that Indian shield is 'stable within 7x10-9/yr'. Considering the distance between IISC and JNUC sites being 1714071km, 7x10-9/yr strain rate is equivalent to nearly 12mm/yr N-S contraction. In fact, Paul et al reported JNUC to have 'southward velocity' of 3.7±1.5mm/yr. The E-W expansion between IISC and JNUC can be explained by the fact that India has a anti-clockwise rotation with the pole of rotation of the angular velocity vector (.44±0.03° Myr-1 ) with respect to Eurasia lie at 25.6N and 11±9E.
Fig.4 Velocity vectors from GPS measurements in NW Himalaya and NAGA (Nepal) relative to IISC, Bangalore, superimposed on the surface releif map of the western and central Himalaya. Blue solid circles are earthquake locations.
The lesser Himalayan sites are having almost negligible movements relative to JNUC. This is surprising considering that JNUC represent the northern tip of the Indian shield. Due to lack of any other GPS measurement data from the Indian shield area, a proper Indian reference frame could not be defined. So far, it has been assumed that the entire Indian peninsula can be treated as a rigid shield without much intra-plate movements. On the basis of the current data, it can not be told with certainty how much component of the observed Bangalore-Delhi shortening is due to tectonic activity within the Indian peninsula, and how much could be attributed to the error arising out of computing vectors in an improper reference frame.
Tectonics of the area
Both the Kangra and Dehra Dun re-entrants and intermediate Nahan salient presents and interesting tectonic problem. The 1905 Kangra earthquake created a preferential ground damage pattern with maximum intensity isoseismals concentrated in Kangra and DehraDun region, leaving the intermediate Simla sector (Nahan salient) relatively less damaged (Middlemiss, 1910). It is not clear whether this was the result of two separate ruptures at Kangra and Dehra Dun (Molnar, 1987). The Main Boundary Thrust (MBT) is sinuous in nature, and the width of the Sub-Himalaya varies from nearly 80km in Kangra, to nearly absent in the Nahan salient, and again widens to nearly 30km in DehraDun region. Main Frontal Thrust (MFT) is the southern-most of the known Himalayan fault system (Nakata, 1989). The MFT is almost always parallel to the main Himalayan arc, and is not always exposed to the surface. The Sub-Himalaya is confined between the MBT and the MFT. A gentle, north dipping detachment surface separates the overlying allochthonous sedimentary strata, where convergence is accommodated through anticlinal growth and surface shortening (Powers et al, 1998, Lillie et al 1987) within the Sub-Himalaya. Presence of evaporites in the decollement level reduces the strength of the rocks, and thrusting can propagate far into the foreland, as has been seen in the Potwar plateau (Lillie et al, 1987, Baker et al, 1988). Evaporites were not found to be present at the decollement in the Kangra re-entrant region (Powers et al, 1998). Also, there is no proof that evaporites are not present within the Nahan Salient. Some patches of salt are exposed in the adjacent region between Kangra reentrant and Nahan salient region (Srikantia and Sharma, 1976). Powers et al (1998) studied the seismic and bore hole data in the Kangra re-entrant region and suggested that the decollement in the Kangra re-entrant has a greater strength of the overlying strata, and the decollement dips at nearly 2.5°N, with a depth varying between 4-7 km, from south to north. In the Dehra Dun re-entrant, the dip of the decollement is 6°N, and decollement depth varies between 3.5-6km.
INDEPTH seismic profiling has shown the existence of a discrete fault plane dipping at 9°, between 27.7° and 29° N in Nepal, where it reaches a depth of nearly 45 km. (Makovsky et al, 1996, Zhao et al, 1993). More recently, Wesnousky et al (1999) measured terrace deposit upliftment in DehraDun region, and inferred 6.9±1.8mm/yr vertical upliftment rate, which, assuming a 30° dipping frontal thrust, translates into horizontal shortening of nearly 12mm/yr.
Modeling of the GPS data:
As seen in Fig(4), we find that the sites lying within the Kangra re-entrant area have almost negligible movement relative to the JNUC site, which represent the northern tip of the Indian shield. Exceptions are three sites on the eastern border of the Kangra re-entrant, which shows a noticeable amount of movement directed across the general Himalayan arc trend. The sites to the north of the re-entrant, that are located in the Higher Himalayan regions, Udaipur, Kothi, and Jispa, have across the strike velocities of 7.34±0.6, 12.04±1, and 14.58± 1.1 mm/yr, relative to JNUC. The sites lying to the south of HFF, like MOND, QASI, SABA in the Dehra Dun region, and BOHA, MANU in the Kangra region have almost negligible movements relative to the JNUC site. We can assume the above rates for higher Himalayan sites as the surface contraction rates taking place within the Himalayan frontal zone. The Higher Himalayan sites in the eastern part of our network, in the Garhwal Himalayan region, e.g. BADR, JOSH, BHAT and HARS have rates of 8.38±1.37, 17.48±.99, 9.71±1.1 and 9.6±1.16 mm/yr respectively. The shortening rate increases with northward distance of the measurement sites from the MBT. We don't have any site further north to see where the shortening rate saturates. This would give us the total shortening taking place across the Himalayan system. The maximum shortening rate that we found for the northern-most site is 14.5mm/yr, at Jispa. The abnormally high value for the Joshimath (JOSH) site is perhaps due to post-seismic stress relaxation after the 1998 Chamoli earthquake.
The differential movement between the Indian shield, Lesser Himalayan sites and Higher Himalayan sites are caused by aseismic slip occurring at the detachment surface. GPS measurements in Nepal (Bilham et al, 1997) indicated existence of a locked >120 km wide detachment fault. The elastic strain accumulation is eventually released by a southward travelling dislocation along the previously unstressed low-angle Himalayan Frontal Thrust (Bilham et al, 1998, Brune, 1996). Assuming that the southern tip of the southward moving Himalayan thrust system is locked with the converging Indian plate along the detachment surface, we try to quantify the depth and geometry of the locked system, as well as to demarcate boundaries of the Himalayan thrust segments, through the dislocation model inversion of the GPS data. We assume that dislocation can be characterized by slip on rectangular dislocation plane buried in an homogeneous, isotropic, elastic half space (Okada, 1985). We used a constrained non-linear optimization algorithm to estimate the fault geometry, viz. Length, depth, width, dip, strike, location, and rates of strike slip and dip-slip movement that best fit the GPS data (Arnadittir & Segall 1994, Burgmann et al, 1997). Guided by the existent geological data, and results from the Nepal GPS studies, we applied constraints on the depth of the fault, range of the fault strike, and allowing dip-slip movement, or strike-slip movement, or both together, to find the minimum of Weighted Residual Sum of Squares (WRSS) given by:
Fig.5 Dislocation model obtained from the GPS data. Blue arrows are observed velocity vectors and yellow arrows are predicted vectors. Red circles show earthquake locations. Black lined rectangles represent modelled dislocation fault.
See text for details.
We evaluated models of increasing complexity, and the best fitting model is shown in Fig (5). Our modeled fault system consist of three adjacent fault segments with varying widths, dip and depths, that best represent the morphological and geological characteristics of the study area like the sinuous nature of the MBT and existence of the Kangra and Dehra Dun re-entrants intervened by the lateral structural ramp in the Nahan Salient. The best-fit model have only dip-slip motion (18mm/yr) for the eastern and western blocks, whereas the in-between block have some strike-slip motion (10 mm/yr), which is in consistence with the fact that the MBT-PT alignment becomes almost transverse to the Himalayan arc on the eastern part of the Kangra re-entrant. We assumed a sufficiently large width of the faults, and the lateral extents of the two boundary faults were sufficiently large to avoid the edge effect . In consistence with the geological findings as discussed above, the Kangra block has the most gentle slope of the decollement (6°), where as the other two blocks have 10.5° and 10° dipping detachment surface. We found the depth to the locked portion of the fault to vary from 6 km in the Kangra block to almost negligible value of 2km in the central block, and 10 km in the Dehra Dun block. The black front line of the rectangular fault model shows the locking line at the detachment surface. With the applied dip angels of 6.5°, 10.5° and 10°, the detachment surface projects to 50km, 10km and 55km further south, which approximately match with the surface exposures of the Himalayan Frontal Thrusts. Nearly 55km length of the detachment surface is locked both in Kangra and Dehra Dun segments, whereas the length of that in the central sector is only 10 km.
Conclusion
As seen if Fig (5), the seismic activities are confined mostly to the northern bound of the locked portion of the detachment surface. Further north, the shortening is accommodated through asiesmic slip. It also appears that though we don't have sites further north to prove, our northern most site like JIPA perhaps already reaches the northern boundary where most of the shortening is talking place. It is also evident from our analysis that a sizable portion of the Frontal Thrust system is locked which will eventually be released through major seismic strain relief, though the dimension of the locked portion is not as great as was found in Nepal. Variation from the Nepal GPS studies, and also lateral variation within our local study area brings out the more pronouncedly the existence and role of the segmented nature of the Himalayan frontal thrust system. It is possible that ruptures caused by great earthquakes are limited to these segments. The Kangra, Nahan and Dehra Dun blocks response differently to the convergence process. The Kangra block, offers the maximum resistance to the advancing Indian plate along the detachment surface, and resist the southward progression of the overlying allocthonous sedimentary strata. The detachment surface has a gentler slope and is locked at nearly 6km depth. The central Nahan block has the least strength, and offers the least resistance to the convergence process. The frontal thrust system are easily carried further south, with a negligible locking depth of 2km, and length of 10 km. The Dehra Dun falls into in-between category between these two extreme cases. However, due to steeper dip angle of the detachment surface, and greater locking depth, both the Dehra Dun and Kangra region have almost same length of the locked portion, and same degree of earthquake potential.
Acknowledgement
The studies were carried out under a funding received from the DST, Govt. of India,, sponsored project. The paper is published with the permission of the Director, WIHG. I gratefully acknowledge participation in GPS field data acquisition campaigns by Sri CP Dabral, Dr.SatyaPrakash, Sri Sanjay Sood, and others. Helps and suggestions received from Dr. Roland Burgmann, UC-Berkeley, USA, and Dr. Roberts King, MIT, Boston, USA, are gratefully acknowledged.
References
- Armijo, R., Tapponnier, P., Mercier, J. L., and T.-L., H., 1986, Quaternary extension in southern Tibet; field observations and tectonic implications: J. Geophys. Res., v. 91, p. 13,803-13,872.
- Avouac, J. P., and Tapponnier, P., 1993, Kinematic model of active deformation in central Asia: Geophys. Res. Lett., v. 20, p. 895-898.
- Avouac, J.P., J. Lave, R. Cattin, M.R. Pandey, and R.P. Tandulkar, 1998. Active mountain building in the Himalaya of Nepal. EOS Trans. AGU, v.79, pp.204.
- Bilham, R., 2001, Slow tilt reversal of the Lesser Himalaya between 1862 and 1992 at 78°E, and bounds to the southeast rupture of the 1905 Kangra earthquake: Gephys. J. Int., v. 144, p. 1-23.
- Bilham, R., Blume, F., Bendick, R., and Gaur, V. K., 1998, Geodetic constraints on the translation and deformation of India: Implications for future great Himalayan earthquake: Curr. Sci., v. 74, no. 3, p. 213-229.
- Bilham, R., Larson, K., Freymueller, J., and authors, a., 1997, GPS measurements of present-day convergence across the Nepal Himalaya: Nature, v. 386, no. 6620, p. 61-64.
- Bilhar, R. 1995. Current Science, v.69, pp.155-187.
- Bürgmann, R., Larson, K., and Bilham, R., 1999, Model inversion of GPS and leveling measurements across the Himalaya: Implications for earthquake hazards and future geodetic networks: Himalayan Geology, v. 20, p. 59-72.
- Bürgmann, R., Segall, P., Lisowski, M., and Svarc, J., 1997, Postseismic strain following the 1989 Loma Prieta earthquake from GPS and leveling measurements: J. Geophys. Res., v. 102, p. 4933-4955. Cattin, R., and Avouac, J. P., 2000, Modeling the mountain building and the seismic cycle in the Himalaya of Nepal: J. Geophys. Res., v. 105, p. check.
- Chen, Z., Burchfiel, B. C., Liu, Y., King, R. W., Royden, L. H., Tang, W., Wang, E., Zhao, J., and Zhang, X., 2000, Global Positioning System measurements from eastern Tibet and their implications for India/Eurasia intercontinental deformation: J. Geophys. Res., v. 105, p. 16215-16227.
- DeMets, C., Gordon, R. G., Argus, D. F., and Stein, D., 1994, Effect of recent revisions to the geomagnetic reversal time scale and estimates of current plate motions: Geophys. Res. Lett., v. 21, p. 2191-2194.
- England, P. C., and Molnar, P., 1997a, Active deformation of Asia: From kinematics to dynamics: Science, v. 278, p. 647-650.
- -, 1997b, The field of crustal velocity in Asia calculated from Quaternary rates of slip on faults: Geophys. J. Int., v. 130, p. 551-582.
- Freymueller, J., Bilham, R., Burgmann, R., Larson, K.M., Paul, J., and Gaur, V. 1996, Global Positioning System measurements of Indian plate motion and convergence across the Lesser Himalaya, Geophys. Res. Lett., v.23, pp.3107-3110.
- Gahalaut, V.K. and Chander, R., 1997. On interseismic elevation changes and strain accumulation for great thrust earthquakes in the Nepal Himalaya. Geophys. Res. Lett., v.24, pp.1011-1014.
- Holt, W. E., Chamot-Rooke, N., Le Pichon, X., and Haines, J., 2000, The velocity field in Asia inferred from Quaternary fault slip rates and GPS observations: J. Geophys. Res., v. 105, p. 19185-19209.
- Jouanne, F., Mugnier, J. L., Pandey, M. R., Gamond, J. F., Le Fort, P., Serrurier, L., Vigny, C., and Avouac, J. P., 1999, Oblique convergence in the Himalayas of western Nepal deduced from preliminary results of GPS measurements: Geophys. Res. Lett., v. 26, no. 13, p. 1933-1936.
- Khattri, K. N., 1987, Great earthquakes, seismicity gaps, and potential for earthquake disaster along the Himalaya plate boundary: Tectonophysics, v. 138, p. 79-92.
- Kogan, M. G., Steblov, G. M., King, R. W., Herring, T. A., Frolov, D. I., Egorov, S. G., Levin, V. Y., Lerner-Lam, A., and Jones, A., 2000, Geodetic constraints on the relative motion and rigidity of Eurasia and North America: Geophys. Res. Lett., v. 27, p. 2041-2044.
- Larson, K., Bürgmann, R., Bilham, R., and Freymueller, J. T., 1999, Kinematics of the India-Eurasia collision zone from GPS measurements: J. Geophys. Res., v. 104, p. 1077-1093.
- Lavé, J., and Avouac, J.-P., 2000, Active folding of fluvial terraces across the Siwalik Hills, Himalayas of central Nepal: J. Geophys. Res., v. 105, p. 5735-5770.
- Middlemiss, C.S., 1910. The Kangra Earthquake of 4th April, 1905. Mem. Geol. Surv. India, v.38, pp.1-409.
- Molnar, P., and Lyon-Caen, H., 1989, Fault-plane solutions of earthquakes and active tectonics of the Tibetan plateau and its margins: Geophys. J. Intl., v. 99, p. 123-153.
- Molnar, P., B.C Burchfiel, L. K'unangyi, and Z. Ziyun, 1987. Geomorphic evidence of active faulting in the Altyn Tagh and northern Tibet and qualitative estimates of its contribution to the convergence of India and Eurasia. Geology, v.15, pp249-253.
- Molnar, P, and Q. Deng, 1984. Faulting associated with large earthquakes and the average rate of deformation in central and eastern Asia. J. Geophys. Res., v.89, pp.6203-6228.
- Molnar, P. and Tapponier P., 1977, Relation of the tectonics of eastern China to the India-Eurasia collision: application of slip line field theory to large scale continental tectonics. Geology 5/4, p.212-216.
- Murphy, M. A., Yin, A., Kapp, P., Harrison, T. M., Lin, D., and Jinghui, G., 2000, Southward propagation of the Karakorum fault system, southwest Tibet: Timing and magnitude of slip: Geology, v. 28, p. 451-454.
- Okada, Y., 1985, Surface deformation due to shear and tensile faults in a half-space: Bull. Seism. Soc. Am., v. 75, p. 1135-1154.
- Pandey, M. R., Tandukar, R. P., Avouac, J. P., Vergne, J., and Héritier, T., 1999, Seismotectonics of the Nepal Himalaya from a local seismic network: J. Asian Earth Sci., v. 17, p. 703-712.
- Paul, J., Bürgmann, R., Gaur, V. K., Bilham, R., Larson, K., Ananda, M. B., Jade, S., Mukul, M., Anupama, T. S., Satyal, G., and Kumar, D., 2001, The motion and active deformation of India: Geophys. Res. Lett., v. 28, p. 647-651.
- Powers, P. M., Lillie, R. J., and Yeats, R. S., 1998, Structure and shortening of the Kangra and Dehra Dun reentrants, Sub-Himalaya, India: Geol. Soc. Am. Bull., v. 110, p. 1010-1027.
- Qi, W., Peizhen, Z., Zhijun, N., Freymueller, J. T., Bilham, R., Larson, K. M., Xi'an, L., Yanxi, L., Yang, Z., and Chen, Q., 2001, Present-day crustal deformation in China constrained by Global Positioning System (GPS) measurements: Science, v. to be submitted to.
- Royden, L. H., Burchfiel, B. C., King, R. W., Wang, E., Chen, Z., Shen, F., and Liu, Y., 1997, Surface deformation and lower crustal flow in eastern Tibet: Science, v. 276, p. 788-790.
- Seeber, L., Ambruster, J.G. and Quittmeyer, R.C., 1981. Seismicity and continental subduction in the Himalayan arc. In: Zagros-Hindu Kush-Himalaya Geodynamics Evolution. Eds. H K Gupta and F Delany. Am. Geophysical Union series, 3, 215-242.
- Seeber, L., and Pecher, A., 1998, Strain partitioning along the Himalayan arc and the Nanga Parbat antiform: Geology, v. 26, p. 791-794.
- Shen, Z.-K., Zhao, C., Yin, A., Li, Y., Jackson, D. D., Fang, P., and Dong, D., 2000, Contemporary crustal deformation in east Asia constrained by Global Positioning System measurements: J. Geophys. Res., v. 105, p. 5721-5734.
- Wesnousky, S. G., Kumar, S., Mohindra, R., and Thakur, V. C., 1999, Uplift and convergence along the Himalayan frontal thrust: Tectonics, v. 18, no. 6, p. 967-976.
- Yeats, R.S., T. Nakata, A. Farah, M. Fort, M.A. Mirza, M.R. Pandey, and R.S. Stein, 1992. The Himalayan frontal fault system. Ann. Tectonicae, v.6 suppl, pp. 85-98.
- Yeats, R. S., and Lillie, R. J., 1991, Contemporary tectonics of the Himalayan frontal fault system: folds, blind thrusts and the 1905 Kangra earthquake: J. Struct. Geol., v. 13, no. 2, p. 215-225.