1 Introduction

GPS based geodetic studies suggest variation in the rate of convergence and deformation of the Indian plate along the Himalaya, across which about 50% of India–Eurasia convergence has been estimated to take place (Avouac and Tapponier 1993). The GPS results in the Kashmir Himalaya indicate arc-normal convergence of about 14 mm/year (Jade et al. 2014) and 5 mm/year of dextral motion on the Karakorum Fault system (Kundu et al. 2014). The convergence rate in the Garhwal Himalaya region has been estimated at about 14 mm/year (Banerjee and Bürgmann 2002; Jade 2004). Similarly, a convergence rate of 10–15 mm/year in the Kumaun Himalayan province has been approximated (Jade et al. 2014; Dumka et al. 2014a, b). The GPS derived convergence rate in the Nepal Himalaya has been calculated at about 18–20 mm/year (Bilham et al. 1997; Jouanne et al. 1999; Ader et al. 2012). Furthermore, this rate varies between 15 and 21 mm/year in the northeast Himalaya (Jade et al. 2007, 2017; Banerjee et al. 2008; Mukul et al. 2010; Stevens and Avouac 2015) (see Fig. 1). Based on the GPS derived slip and strain rate estimations, Bilham et al. (1997, 1998, 2001) proposed possibility of a great earthquake in the eastern Nepal Himalaya. Throughout the Himalaya, most of the crustal deformation is localized on the Main Himalayan Thrust (MHT) and its surface trace is known as the Himalayan Frontal Thrust (HFT) (Nakata 1989; Hauck et al. 1998; Cattin and Avouac 2000; Lavé and Avouac 2000; Ader et al. 2012). Based on the arc-parallel gravity anomaly analysis, Hetényi et al. (2016) have suggested disparate flexural geometry in the four segments of the Himalaya. According to Stevens and Avouac (2015), about 100 km width of the MHT is entirely locked along the Himalayan arc. On the whole, the MHT is locked to a depth of about 20 km from the surface over a width of about 115 km towards the western part of Nepal (Bettinelli et al. 2006; Zhao and Nelson 1993; Nelson 1996; Cattin and Avouac 2000) and the down-dip creeping rate is comparable with the geological slip rate (Pandey et al. 1995; Bilham et al. 1997; Larson et al. 1999; Jouanne et al. 1999, 2004).

Fig. 1
figure 1

a Convergence rate of various segments throughout the Himalaya (based on Banerjee and Bürgmann 2002; Jade 2004; Jade et al. 2007, 2014; Banerjee et al. 2008; Mukul et al. 2010; Ader et al. 2012; Dumka et al. 2014a, b; Stevens and Avouac 2015). b Shows the arc parallel gravity anomaly along the Himalaya (Hetényi et al. 2016). Four litho-tectonic units in Uttarakhand Himalaya (Sub-Himalaya, Lesser Himalaya, Higher Himalaya and Tethys Himalaya), demarcated by major thrust/fault systems. Dotted lines show subsidiary thrusts/faults. Details of subsidiary faults with geology are given in Fig. S1

Full size image

Due to the intense stress accumulation at the down-dip edge of the MHT, a substantial seismic activity has been observed along the front of the higher Himalaya (Pandey et al. 1999; Cattin and Avouac 2000; Arora et al. 2012). According to Ader et al. (2012), the width of the locked zone is estimated at 80–100 km in the central Nepal Himalaya, and that an earthquake of Mw ~ 8 is required to balance the measured slip deficit. Mugnier et al. (2013) suggested that a great earthquake could occur anytime at all along the front of the central Himalaya. After the 2015 Nepal earthquake, Avouac et al. (2015) showed that the west of this earthquake, the area between the 2015 rupture and the 1905 Kangra (Mw 7.8) event has been locked for over 500 years with a slip deficit of 10 m. Based on the receiver function study in the Uttarakhand Himalaya, Caldwell et al. (2013) indicated presence of ramp beneath the trace of the Munsiary Thrust. The trench investigations along the HFT in Kumaun Himalaya also reveal seismic inactivity during the last 700 years, giving indication to a great earthquake in this region (Rajendran et al. 2015). The tectonic activities are more common towards the zone of HFT-foreland side (Nakata 1982, 1989; Le Fort 1975; Valdiya 2003; Thakur 2004).

In Uttarakhand, the rate of convergence varies between 14 and 12 mm/year in the Garhwal and Kumaun Himalaya respectively (Jade 2004, 2014). The horizontal shortening of < 7 mm/year between IISC (Peninsular India) and GBPK (Lesser Kumaun Himalaya) reveals that the frontal part of the Himalaya is presently locked with the Indian Plate and that the Higher Himalaya is converging towards the Lesser Himalaya at a rate of about 9–16 mm/year (Dumka et al. 2014a, b). The campaign mode study in the Lesser Kumaun Himalaya (Ponraj et al. 2010, 2011) also indicates the accommodation of the maximum crustal strain near the Main Central Thrust (MCT).

Our current study attempts to understand the deformation and strain pattern in the Uttarakhand state as a whole (Fig. 1) by analyzing the GPS results of 15 new campaign mode sites in addition to published datasets (e.g., Banerjee and Bürgmann 2002; Ponraj et al. 2010, 2011; Jade et al. 2014; Dumka 2011; Dumka et al. 2014a, b). Based on the combined data, we recognized the zones of maximum crustal deformation and strain. The active nature of faults and uplift rate of river terraces in the study area are linked with the higher deformation rates. Finally, the achieved information is correlated with the earthquake distribution of the region and this may be useful for future hazard management of this hill state.

2 Tectonic setting

Tectonically one of the most active Himalayan segments, the Uttarakhand Himalaya constitutes four litho-tectonic units from south to north; Sub-Himalaya, Lesser Himalaya, Higher Himalaya and Tethys Himalaya, which are delimited by the Himalayan Frontal Thrust (HFT), Main Boundary Thrust (MBT), Main Central Thrust (MCT) and Trans Himadri Fault/South Tibetan Detachment Fault (THF/STDF), respectively (see Fig. 1). Besides these intracrustal boundary thrusts, a number of local thrust systems lie in between, the major ones being the Ramgarh Thrust (RT), South Almora Thrust (SAT), North Almora Thrust (NAT) and Munsiary Thrust (MT) from south to north (Gansser 1964; Jhingran et al. 1976; Valdiya 1979,1999; Nakata 1989; Thakur 1992; Godin 2003). Also, the faults/thrusts, oblique or parallel to the main intracrustal boundary thrusts, are predominantly oriented NW/NNW-SE/SSE (Middlemiss 1890a, b; Gansser 1964; Valdiya 1976; Valdiya and Kotlia 2001) and many of these were formed during the Quaternary period (i.e., Kotli Fault, Bhimtal-Naukuchiatal Fault, Banlekh Fault, Ladhiya Fault, Bisoriya Fault, Bhikiyasain Fault, Dulam Fault, Garampani Fault, Betalghat–Basgaon Fault, Kosi Fault, Alaknanda Fault, Chaukhutia Fault, Nandprayag Fault (See Fig. S1) (Valdiya 1976, 1993, 2003; Nakata 1982; Kotlia et al. 1997, 2000, 2008, 2010; Valdiya and Kotlia 2001; Kotlia and Rawat 2004; Thakur 2004; Rawat and Kotlia 2006; Sati et al. 2007; Mehta and Sanwal 2011; Kotlia and Joshi 2013; Rana et al. 2013; Joshi and Kotlia 2014) and these are considered to be active and thus responsible for damming the ancient rivers/streams, resulting in formation of palaeolakes (Kotlia et al. 1997, 2008, 2010; Valdiya and Kotlia 2001). The studies on the uplift due to such active faults/thrusts can be well corroborated with the seismic activity (e.g, Arora and Mahashabde 1987; Shrikande et al. 2006; Lyubushin et al. 2010; Roy and Mondal 2012).

3 Methodology and data

The GPS data from 15 campaign mode stations together with previously published results from 43 sites (e.g., Banerjee and Bürgmann 2002; Ponraj et al. 2010, 2011; Jade et al. 2014; Dumka et al. 2014a, b) (see Table 1) were analyzed. The raw GPS data from 15 sites were converted into the RINEX (Receiver Independent Exchange) format using the TEQC program (translation, editing and quality control), distributed by UNAVCO (University NAVSTAR Consortium) (Estey and Meertens 1999). This program provides a synopsis containing L1/L2 tracking status for each SV session’s start and end time, data logging interval, list of satellite observed and missing observations, clock drift rate/gaps, number of cycle slips and session length. These parameters are most important factors for quality inspection of the GPS data and are useful for data processing. To improve the uncertainties in relative positioning measurements, both L1 and L2 frequencies of several satellites must be present in the data (Blewitt 1989; Dong and Bock 1989). The multipath report is important parameter for assessment of the site-specific errors while the session length of data is important for reduction in the RMS values (Snay et al. 2002; Satirapod et al. 2001). The dataset was post-processed using GAMIT-GLOBK-GLORG software, developed by MIT, USA (King and Bock 1998; Herring et al. 2010) to obtain constrained solution (H-) files of parameter estimates and covariance (Herring et al. 2006). We selected a set of sites with applied constraint to minimize the adjustment of the network (Herring et al. 2006).

Table 1 Velocities of all the GPS sites with respect to the Indian plate (Mahesh et al. 2012) and tectonic position of the GPS sites
Full size table

In GAMIT, Zenith tropospheric delay for each station was estimated by a linear model with stochastic constraint for signal delay due to the troposphere. The basic input for the GAMIT are the observation files of permanent/IGS stations in the RINEX format, orbital file (sp3 files) or g-files and the global navigation files. The orbital and IGS-RINEX files are available at Scripps Orbit and Permanent Array Centre (SOPAC) and brdc files are assessable at Continuously Operating Reference Stations. After combining the daily solutions, velocities of all the stations were estimated in the ITRF08 reference frame. To account for temporal correlation in the time series, we used random walk noise at the level of 0.75 mm year−1/2 (Mao et al. 1999). For improved understanding of the deformation pattern, velocities of all the sites were calculated in the Indian Reference Frame (Mahesh et al. 2012) (see Fig. 2). However, to avoid any discrepancy due to combining results of different reference frames, velrot tool of the GLOBK (Herring et al. 2010) was used for conversion and all the velocities were estimated in the ITRF08 reference frame. For calculation of deformation rate at each site, Euler pole of the Indian plate (e.g., Ader et al. 2012; Mahesh et al. 2012; Jade et al. 2014) was used.

Fig. 2
figure 2

Velocities of all the sites in the Indian reference frame, showing deformation pattern in the region. Blue arrows indicate velocities of 15 new sites and black arrows the velocities obtained by previous studies (e.g., Ponraj et al. 2010, 2011; Jade et al. 2014; Dumka et al. 2014a, b). (Color figure online)

Full size image

The strain fields of the area were calculated following a method proposed by Pesci and Teza (2007) and Teza et al. (2008). This method provides a clear-cut calculation of major and minor strain axes. The strain field was computed on the nodes of 8 × 8 km grid via Least Square (LS) method which is based on rescaling of the covariance matrix of velocity data by weighting function which takes into account the distances between the grid node and the GPS stations (e.g., Shen et al. 1996; Pesci et al. 2009). Errors are rescaled by means of weight function, exp (δd/d0) (Pesci and Teza 2007), where d is distance between the knot and d0 is smoothing constraint (Shen et al. 1996; Pesci and Teza 2007; Pesci et al. 2009). Smoothening constraint was considered as three times the grid spacing (Livieratos 1980; Shen and Jackson 2000). If the data points are widely distributed, the local strain can be estimated at each knot of the grid (or in a defined point) using weighting strategy (see Shen and Jackson 2000; Pesci and Teza 2007; Pesci et al. 2009).

A significance test was conducted for various grid intervals, i e., 5, 8, 12, 15, 20, 25 km and finally we chose 8 km grid interval as it is significant for strain analysis of study area (Fig. S2). The significance test detects the points which can be considered as actual representative of a local strain at selected scale. The significance test indicates the spatial distribution of the GPS points around grid point at distance lower or equal to the scale factor. If there are three GPS points in 120° area around the grid point at distance lower or equal to the scale factor, then it has high significance. Based on this, we categorized the whole region into three zones, most significant, mid-significant and low-significant (e.g., Pesci and Teza 2007). In the present study, we used the dataset from the most and mid-significant areas only.

4 Results

4.1 Deformation analysis

Average deformation rate for the Sub Himalayan sites is around 1.5 ± 1 mm/year (with 1 sigma uncertainty) which increases to 11.3 ± 1.6 mm/year for the Higher Himalayan sites. The average deformation rate in the Lesser Himalayan sites is computed at 2.8 mm/year. However, average rate of the GPS sites in the Inner lesser Himalaya is higher (5.2 ± 1.2 mm/year), compared to that of 1.7 ± 1.3 mm/year in the outer Lesser Himalayan sites (Fig. 3a). The average deformation rate of GPS sites along the MCT zone is observed as 8.7 ± 1.7 mm/year, higher than that of the Lesser Himalaya. This implies that at present most of the Himalayan shortening is mainly concentrated near the NAT, MT and MCT in the Uttarakhand Himalaya. This may suggest that the MHT is locked until the base of the Outer Lesser Himalaya and most of the slip is being accommodated on the root of the hinterland faults.

Fig. 3
figure 3

a Deformation rate from GPS sites (arc-perpendicular distance from the HFT) in different tectonic units of Uttarakhand Himalaya. An abrupt increase in deformation rate is clearly seen in vicinity of the MCT and maximum in the higher Himalaya. The letters in the top of figure are I, Indo-Gangetic plain; S, Sub-Himalaya; L, Lesser Himalaya; M, MCT zone; H, Higher Himalaya; T, Tethys Himalaya. b Depth-wise distribution of earthquakes during 2007–2015 in the Kumaun Himalaya (redrawn from Negi et al. 2017) indicating focussed activity near the MHT ramp and thrust faults above. (Color figure online)

Full size image

The seismicity pattern of the region suggests that small and moderate earthquakes occurred on the mid-crustal ramp of the MHT (Seeber and Armbruster 1981; Molnar 1990; Avouac 2003; Gahalaut 2008; Negi et al. 2017) (Fig. 3b). We also conclude that the maximum amount of deformation has occurred in the same vicinity. The distribution pattern of earthquakes (ISC, International Seismological Centre, http://www.isc.ac.uk) in the Uttarakhand Himalaya indicates a high concentration in the Higher Himalaya, proving extreme accumulation of the stress at the down-dip edge of the locked MHT.

4.2 Strain analysis

The estimated site velocities with respect to the Indian plate were used as an input for the strain analysis. The most crucial finding of the present study is identification of the zones of strain accumulation (Fig. 4a) and these are categorized into High Strain Zones (HSZ) and Moderate Strain Zones (MSZ) based on the magnitude of strain. The zones with strain value greater than 0.3 micro-strain/year are considered as HSZ while less than 0.3 micro-strain/year are considered as MSZ. Accordingly, two (HSZ1, HSZ2) and one (MSZ1) were identified in the study area (Fig. 4a). The distribution pattern indicates lateral variation in the strain value in the study region. Strain rate of 0.4–0.45 micro-strain/year is viewed in the vicinity of the MCT towards the eastern and western part but in the middle part of the study area, the strain rate appears less than 0.2 micro-strain/year. To witness the variation in strain with respect to the MHT, we have drawn three arc normal profiles of strain rate in the region (Fig. 4b). The maximum strain rate of ≥ 0.45 micro-strain/year has been detected at a distance of 80 and 110 km from the HFT along Profile I and III, respectively. However, the profile II indicates maximum strain at a distance of 190 km from the HFT. The presence of a ramp is proposed at a distance of 100–130 km north of the MHT (see also Caldwell et al. 2013; Negi et al. 2017). The maximum strain of profile I and III can be associated with the ramp of the MHT but the low strain rate close to the ramp in profile II may be due to other factors, which have to be investigated in the future.

Fig. 4
figure 4

a Calculated strain rate tensor indicating high strain zones (HSZ1 and HSZ2) and moderate strain zones (MSZ1). The zones with strain value > 0.3 micro-strain/year are considered as HSZ while < 0.3 micro-strain/year are considered as MSZ. The distribution pattern indicates lateral variation in the strain value which is in accordance to the Mahendragar–Dehradun fault based on the gravity analysis (Hetényi et al. 2016). b Lateral variation is clearly seen in the 3 profiles starting from the HFT (surface trace of MHT). The profiles I and III point to maximum strain of 0.45 micro-strain/year, at a distance of 80 and 110 km from the HFT. Profile II in the middle of the study area reveals maximum strain of 0.30 micro-strain/year at a distance of 190 km from the HFT

Full size image

The calculated strain indicates that most of the region is in the compressional regime as expected because of the tectonics of the Indian-Eurasia collision. The maximum compressional strain is calculated in the vicinity of the MT and MCT in both the Garhwal and Kumaun regions (Fig. 5). Variation in the trend of the compression has been observed in the study area, which is perhaps controlled by the strike of the major Himalayan fault systems. The earthquake fault plane solution suggests WNW-ESE to NNW-SSE orientation of the compressional axis (e.g., Stevens and Avouac 2015; Mahesh et al. 2015; Negi et al. 2017). The world stress map (Heidbach et al. 2016) for the study area supports the orientation of the principle axis of deformation (Fig. S3). The orientation of extensional strain tensor is E-W to WNW-ESE which we believe is due to the lateral expansion along the major hinterland thrust systems and activity along the transverse faults. In the study area, the NAT and MBT are characterized by the presence of numerous transverse faults, as mentioned in the text. Further, Kayal (2014) and Kayal et al. (2003) have also indicated presence of the N–S fault, transverse to the MCT (see Fig. S1).

Fig. 5
figure 5

Calculated strain rate tensor indicating domination of compressional strain in the area as expected because of the tectonics of the Indian–Eurasia collision. Blue and red tensors represent extension and compression, respectively. (Color figure online)

Full size image

5 Discussion

Analysis of present dataset suggests that the Inner Lesser Himalaya and MCT zone accommodate maximum amount of shortening of the India-Eurasia convergence in Uttarakhand Himalaya. The occurrence of maximum seismicity in this part supports our observation. Based on the previous studies a mid-crustal ramp is situated nearly beneath the MCT (see also Molnar 1990; Srivastava and Mitra 1994; Avouac 2003; Yin 2006; Wobus et al. 2006; Caldwell et al. 2013, Negi et al. 2017) in our study region (see Fig. 3b).

Correlation between the deformation pattern and seismicity indicate a direct relation among them (see Fig. 3b). The arc-perpendicular deformation rate, which is less than 4 mm/year in the Outer Lesser Himalaya reaches up to 12 mm/year in the Inner Lesser Himalaya and Higher Himalaya. Therefore, the deformation in this part may be correlated with the ramp structure along the MHT, and suggests that a small part of stress accumulation around the down-dip end is expressed in the form of current seismic activity (Bilham et al. 1998; Banerjee and Bürgmann 2002; Bollinger et al. 2004). The earthquake catalog suggests that Mw 6.3 (Uttarkashi, 1991) and Mw 6.6 (Chamoli, 1999) are the highest magnitude earthquakes in this part during the last three decades. The last great earthquake in this part was probably during the thirteenth century (Rajendran and Rajendran 2011; Rajendran et al. 2015; Jayangondaperumal et al. 2017). In this scenario, the present day deformation rate, if we apply over last 700 year, is sufficient to generate a great earthquake. Based on these circumstances, Nath et al. (2008) have indicated a 10% probability of Mw 8.5 earthquake in the next 50 years in this part of the Himalaya.

The Himalaya, which is considered as a result of compressive tectonics, is mainly believed to sustain the compressional strain. Also, several studies carried out in the different segments of Himalaya support this hypothesis. Based on the strain analysis, our work provides a prominent signal of compressional strain in Uttarakhand Himalaya (Figs. 4, 5). A maximum strain of ≥ 0.45 micro-strain/year is observed near the vicinity of the MCT, including parts of the Lesser and Higher Himalaya. The HSZ1 and HSZ2 point to maximum compressional strain of 0.45 micro-strain/year in the study area. An earthquake of M 6.3 in 1991 has already been triggered in the HSZ2. Tectonically, these zones encompass the MCT, MT and NAT. Therefore, we suggest that the seismic activity along these faults systems is related to the maximum strain. Mahesh et al. (2015) correlated seismic activity in this part of Himalaya with the MT and few other transverse faults. The MSZ1 (Fig. 4) that falls in the middle of the study area reveals maximum strain of 0.20 micro-strain/year. The seismic data of this part show occurrences of numerous micro earthquakes (Pathak et al. 2015) and has previously experienced an earthquake of M 6.6 at Chamoli in the year 1999. We believe that the lateral variation in the strain value within the Uttarakhand Himalaya is an indicator of segmented behavior of the Himalayan arc. The seismicity pattern and fault plane solution provide evidence of along-strike variation in the region (Kayal 2001). Along-strike variation in the Moho depth between Garhwal and NW Himalaya has been reported by Caldwell et al. (2013). A gravity-based study (Hetényi et al. 2016) has also identified one such structural segment in the Uttarakhand Himalaya, indicating lateral variation due to inherited structure of the Indian plate.

The seismicity pattern indicates that the earthquakes are more concentrated near the MCT towards the west while, towards the east, these are widely distributed between the THF and MT in the study area. Another important trend is that the earthquakes are closely concentrated in the middle of the study area. We presume that this kind of pattern indicates variation in the locking length of the MHT, and that this is possibly due to the presence of transverse structural features and heterogeneity in the under-thrusting Indian plate (Valdiya 1976; Dasgupta et al. 1987). One such structure is known as Delhi–Haridwar ridge based on the seismicity analysis (Arora and Mahashabde 1987; Lyubushin et al. 2010; Roy and Mondal 2012) and Mahendragar–Dehradun Fault based on the gravity analysis (Hetényi et al. 2016) in the Uttarakhand Himalaya.

Although it is a well known fact that earthquakes result from strain buildup of longer time period and GPS results are based on only a few years of data, a comparison between strain accumulation and earthquake distribution suggests a good correlation in the Uttarakhand Himalaya. A number of studies in the different parts of world point out high strain rate in the low earthquake activity areas and vice versa. Analysis of the earthquake dataset (1970–2015) indicates higher concentration in the MSZ1 and lower concentration of earthquakes in the HSZ1 and HSZ2. This clearly infers that some amount of accumulated strain in the MSZ is getting released in form of the earthquakes through the MCT, MT and some local faults. Therefore, based on the comparison between seismicity and strain rate, we can conclude that the seismicity in the HSZ is directly associated with the locked segment of the MHT, whereas part of the seismicity in the MSZ is associated with the motion along the localized faults. In a detailed study of earthquake pattern of Kumaun and Garhwal region, Mahesh et al. (2015) observed a number of localized zones of stress accumulation related to the normal/strike-slip fault motion. The occurrence of a few strike-slip (component of normal/reverse) earthquakes together with the majority of thrust earthquakes (Stevens and Avouac 2015) indicates activity of the local faults (Figs. S3 and S4).

6 Conclusions

On the basis of new GPS data and a compiled dataset, we sum up with the following results on the deformation and strain pattern of Uttarakhand Himalaya.

  1. 1.

    Maximum amount of deformation near the traces of hinterland faults (mainly MCT) suggests excessive accumulation of stress at the down-dip edge of the locked MHT.

  2. 2.

    Most of the Uttarakhand Himalaya encompasses compressional strain and a maximum of 0.45 micro-strain/year is suggested in parts of the Inner Lesser Himalaya and Higher Himalaya.

  3. 3.

    Most crucial finding of the present study is the detection of strain accumulation, and this is categorized into High Strain Zones (HSZ) and Moderate Strain Zones (MSZ1) based on the magnitude of strain.

  4. 4.

    Variation in the strain value suggests along-strike behaviour of the Himalayan arc within Uttarakhand Himalaya and prevailing seismicity and gravity pattern is in agreement with our GPS results.

  5. 5.

    The comparison between seismicity and strain rate suggests that the seismicity in the HSZ is directly associated with the locked segment of the MHT, whereas a part of seismicity in the MSZ is associated with the motion along the localized faults.