The 2009 Bhutan and Assam felt earthquakes (M_w 6.3 and 5.1) at the Kopili fault in the northeast Himalaya region

Abstract

Seismotectonics of the two recent earthquakes, one M_w 6.3 in the Bhutan Himalaya on 21 September 2009 and the other M_w 5.1 in the Assam valley on 19 August 2009, are examined here. The recent seismicity and fault plane solutions of these two felt earthquakes suggest that both the events occurred on the Kopili fault zone, a known active fault zone in the Assam valley, about 300 km long and 50 km wide. The fault zone is transverse to the east–west Himalayan trend, and its intense seismicity indicates that it transgresses into the Himalaya. The geologically mapped curvilinear structure of the Main Central Thrust (MCT) in the Himalaya, where the epicentre of the Bhutan earthquake is located, is possibly caused by the transverse Kopili fault beneath the MCT. This intensely active fault zone may be vulnerable to an impending larger earthquake (M > 7.0) in the region.

1. Introduction

The northeastern Himalaya and the adjoining region, between 22–30° E latitude and 88–98° N longitude, are seismically and tectonically more complex than the eastern/central or western Himalaya segments (Kayal 2001, 2010). The region experiences several large/great earthquakes due to the Himalayan collision tectonics to the north, Indo-Burma (Myanmar), atypical subduction tectonics to the east, and due to intra-plate seismic activity in the Assam valley–Shillong plateau–Bengal basin area (Kayal 2008). The intra-plate seismic activity is fairly intense in this region due to the complex stress regime (Angelier and Baruah 2009).

The region has experienced two great earthquakes (M > 8.0), the 1897 Shillong earthquake M_s 8.7 (Oldham 1899), revised to M_s 8.0 (Ambraseys and Bilham 2003), and the instrumentally well recorded 1950 Assam-Tibet earthquake M_s 8.7 (Tandon 1954). Although the 1897 great earthquake was assigned to be a Himalayan earthquake (Seeber et al. 1981), it is now argued to be a shield intra-plate earthquake due to pop-up tectonics of the Shillong plateau (Bilham and England 2001, Kayal et al. 2006). The 1950 great earthquake occurred at the syntaxis zone, and is argued to have occurred by a strike slip mechanism compatible with the strike slip regime in the syntaxis zone (Ben Menahem et al. 1974, Armijo et al. 1989, Kayal 2010), rather than a pure thrust mechanism (Chen and Molnar 1990).

In addition to these great earthquakes, about 22 large earthquakes (M > 7.0) have been recorded in the region after the 1897 great earthquake (Kayal 1996, 2008). Most of the large earthquakes are recorded in the Indo-Burma region, in the inter-plate/dipping seismic zone, but four intra-plate large earthquakes occurred much to the east of the Indo-Burma subduction zone or much to the south of the Himalayan collision zone. These are the 1943 event (M 7.2) on the Kopili fault in the Assam valley, the 1930 event (M 7.1) on the Dhubri fault at the western boundary of the Shillong plateau, and another two in the Bengal basin (figure 1). In the Bengal basin, the 1923 event (M 7.5) is assigned to the Hinge zone and the 1918 event (M 7.6) to the Sylhet fault (figure 1) (Nandy 2001). Among the historical earthquakes, before 1897, though few records are available, the 1869 large earthquake M 7.5 that occurred on the south-eastern end of the Kopili fault is fairly well documented (figure 1) (Nandy 2001).

Figure 1. Tectonic map of the study region (modified from Kayal et al. 2006); MCT, Main Central Thrust; MBT, Main Boundary Thrust; DF, Dauki Fault; DT, Dapsi Thrust; BF, Brahmaputra Fault; SP, Shillong Plateau; MH, Mikir Hills; other features are named on the map. The local seismic broadband stations (blue triangles) used in this study are shown. The recent EHB located earthquakes M > 4.5 (1995–2007) within the rectangle area are shown by green solid circles, the past four large (M > 7.0) intra-plate earthquakes (solid yellow circles) and the two great earthquakes (larger yellow stars) are annotated with the year of occurrence. The historical 1869 large earthquake (M 7.5) is also annotated with a yellow star. The two felt earthquakes of 2009 are shown by red stars. The United States Geological Survey three fault plane solutions, the HRV centroid moment tensor (CMT) solution and the two solutions obtained in this study for the 21 September 2009 Bhutan earthquake are illustrated with the usual notation of beach balls of different colours. Two HRV CMT solutions of the past two earthquakes, 1995 and 2006, respectively, in the Bhutan Himalaya are shown by black beach balls. The fault plane solution of the 17 August 2009 Assam earthquake obtained in this study is also shown and annotated. A north–south cross section of the earthquakes that fall within the rectangle area is shown below; the red star and cluster of red solid circles below the curvilinear Main Central Thrust (MCT) indicate the 21 September 2009 Bhutan main shock and aftershocks, respectively (USGS reports). Inset: map of India showing the study region in a rectangular box. Available in colour online.

We have recorded two strongly felt earthquakes in 2009 in this region, one M_w 6.3 on 21 September, which occurred on the Main Central Thrust (MCT) in the Bhutan Himalaya, and the other M_w 5.1 on 19 August 2009 in the Assam valley on the Kopili fault, respectively (figure 1). These two events occurred almost within a month, and are reported as shallow focus (depth ∼ 10 km) earthquakes in the United States Geological Survey (USGS) reports. The northeast India region is well equipped with about 25 permanent broadband seismic stations since 2001. We have studied the focal mechanisms of these two events by waveform inversion, and examined the recent seismicity. We focus our discussion on the seismotectonics of these two strongly felt earthquakes with a background of known tectonics of the two different tectonic domains in close proximity.

2 Bhutan Himalaya earthquakes

The Bhutan Himalaya has no record of great or large earthquakes (M > 7.0) during the past 200 years (Ambraseys and Jackson 2003). The recent seismicity of the Bhutan Himalaya during the last 100 years, as reported in the International Seismological Centre (ISC; http://www.isc.ac.uk) catalogue is, however, low compared to its adjoining Himalayan segments in the west. The low seismicity has been attributed to a lower convergence rate, and it is also suggested that the India–Eurasia convergence is largely accommodated by the pop-up tectonics of the Shillong plateau to the south, and the Bhutan Himalaya lies in the shadow zone with less seismicity (Gahalaut et al. 2010).

The 21 September 2009 strongly felt earthquake M_w 6.3 provides an insight into the seismotectonics of the Bhutan Himalaya. We have examined the recent seismicity and a north–south cross section in this area that includes the Shillong plateau, Assam valley and the Bhutan Himalaya; the EHB (Engdahl et al. 1998) relocated events (M > 4.5) since 1995 are considered (figure 1). The section shows that the Shillong plateau earthquakes are mostly confined within a depth of 40 km and the seismic activity is bounded by the two major boundary faults, the Dapsi thrust (DT) and the Brahmaputra fault (BF), which has been also observed by local broadband network data (Kayal et al. 2006). The north dipping DT, a conjugate of the Dauki fault (DF), is identified to be an active thrust that demarcates the southern boundary of the Shillong plateau activity and also truncates the maximum isoseismal of the 1897 great Shillong earthquake along this thrust (Kayal and De 1991, Kayal 2001). The Bhutan Himalaya earthquakes are found to be much shallower (∼10 km) at the MCT zone, particularly the 2009 main shock and its aftershocks that occurred on the north–south curvilinear segment of the MCT. The earthquakes further north of the MCT are deeper, down to 50 km (figure 1).

2.1 The September 2009 felt earthquake (M_w 6.3)

The 21 September 2009 earthquake M_w 6.3 was widely felt; 11 casualties, more than 18 injuries and about 1100 damaged houses were reported in Bhutan (figure 2). The maximum intensity reached VII+ in the epicentre area (Dowchu 2010, personal communication). The epicentre of the event was given at latitude 27.34° N and longitude 91.41° E, and depth ∼10 km (USGS report; http://earthquake.usgs.gov). The tremor was well felt in the Bhutan Himalaya and in the adjoining northeast India region including Sikkim, Assam, Arunachal Pradesh, Shillong plateau and Bangladesh; minor damage to a few houses in the Guwahati city was also reported in the local newspapers.

Figure 2. Damaged houses in the meizoseismal area of the 21 September 2009 earthquake, Narang village, Bhutan Himalaya (courtesy: Dowchu 2010, personal communication).

Three moment tensor solutions are given by the USGS and a centroid moment tensor (CMT) solution by Harvard (HRV) (figure 1). All four solutions are compatible with M_w 6.1–6.3 and depth ∼8–10 km, inferring that the event occurred on a shallow north dipping plane. The north dipping plane is inferred to the plane of detachment as the fault plane for this event. The plane of detachment is defined as the interface between the gently dipping Indian basement and the Himalayan sedimentary wedge. The seismotectonic model that the Himalayan earthquakes occur by thrust faulting on the plane of detachment is the most widely accepted tectonic model as envisaged by (Seeber et al. 1981 , Ni and Barazangi, 1984), and the Himalayan earthquake solutions are mostly biased to this model. Kayal (2001, 2010), however, argued that this model fits fairly well in the western Himalaya, but not in the eastern or northeastern Himalaya.

We have reanalysed the teleseismic waveforms of about 30 global digital seismic stations and obtained a solution of thrust faulting with a strike slip component for this event (figure 1). We infer that the east dipping north–south nodal plane is the fault plane that is compatible with the north–south trending curvilinear segment of the MCT where the main shock occurred at a shallow depth (figure 1). A similar solution is reported in the HRV catalogue (http://www.seismology.harvard.edu) for the M_w 5.4 event on 26 February 2006, near to the 2009 main shock epicentre; this event possibly also occurred on the same seismogenic structure. Another similar solution is observed for the M_w 5.4 event that occurred on 17 February 1995 within 50 km of the main event (epicentre and CMT solution are shown in figure 1).

Further, we have used the waveform data of five broadband seismic stations of the local network in the northeast India region, which is running to the immediate south of the epicentre of the 2009 Bhutan earthquake, and we obtained a pure thrust faulting mechanism by inversion (figure 1). In this pure thrust faulting solution, we also infer the east dipping north–south nodal plane as the fault plane, comparable to that obtained by the teleseismic data of the 30 global stations (figure 1). The moment magnitude M_w is found to be 6.2 in both the solutions. Thus, the two solutions obtained in this study are unbiased and compatible. These two solutions are also compatible with the HRV CMT solutions of the past two events (M > 5.0) in the immediate vicinity.

3 Assam valley earthquakes

We focus our observations on the Kopili fault earthquakes in the Assam valley. The Kopili fault zone, approximately 300 km long and 50 km wide, separates the Shillong plateau and the Mikir massif by strike slip movement (figure 1), and it is identified as the most active fault in the Assam valley area (Kayal et al. 2006, Bhattacharya et al. 2008, 2010). The Shillong-Mikir massif are believed to be part of the Indian shield, and transported to the east by the Dauki fault (Evans 1964) (figure 1). Further, the Mikir Hills massif is a fragmented part of the Shillong massif and separated by the Kopili fault. A recent seismicity map was prepared by relocation of the earthquakes recorded by the local networks in the northeast India region (figure 3). The seismicity trends clearly indicate that the Shillong plateau and the Kopili fault zone are the two main intra-plate earthquake source zones in the area. It is further noted that the Kopili fault transgresses into the Himalaya (figure 3). No local network is presently running in the Bhutan Himalaya that would help to incorporate its local earthquake data in the seismicity map of figure 3, but the global network data in the Bhutan Himalaya clearly shows that the Kopili fault active zone is extending up to the MCT zone (figures 1 and 3). Using the local broadband seismic network data, Kayal et al. (2006) made a detailed study of the seismotectonics of the Kopili fault zone and reported that the earthquakes are generated by strike slip faulting on this northeast dipping fault.

Figure 3. Recent seismicity map prepared by the relocated events using local network data (1993–1999) in northeast India showing high seismic activity in the Shillong plateau and intense seismicity along the northwest–southeast Kopili fault zone (modified from Bhattacharya et al. 2008). The two felt earthquakes of 2009 are shown by red stars. Available in colour online.

3.1 The August 2009 earthquake (M_w 5.1)

The 19 August 2009 earthquake (M_w 5.1) occurred in the Assam valley, epicentre 26.56° N latitude, 92.48° E longitude with a focal depth ∼10 km (USGS report), this earthquake was well felt in the region. The event is well recorded by the local broadband seismic stations in the area. A well constrained fault plane solution is obtained by waveform inversion using the broadband seismograms. The estimated seismic moment is compatible with the moment magnitude M_w 5.1. A right lateral strike-slip solution is obtained; the beach ball represents the fault plane solution (figure 1). The north-northwest oriented nodal plane is the inferred fault plane, which is comparable with the Kopili fault. A similar solution is reported in the HRV catalogue for the M_w 5.4 event that occurred at the northern end of the Kopili fault on 23 February 2006 (figure 1).

4 Discussion and conclusions

It is reported that the Himalayan earthquakes are greatly influenced by transverse structures in the eastern and northeastern region (Mukhopadhyay 1984, Dasgupta et al. 1987, Kayal 2001). Unlike in the western Himalaya where the earthquakes mostly occur on the plane of detachment and fairly fit with the envisaged seismotectonic model (Seeber et al. 1981), the eastern or northeastern Himalayan earthquakes do not fit into this model (Kayal 2001, 2010). Long transverse structures across the eastern and northeastern Himalayas, like those of the East Patna fault and Tista lineament in the eastern Himalaya, and the Kopili fault and Dhubri/Jamuna fault in the northeastern Himalaya play a major role in generating earthquakes in the foredeep and foothills region (Mukhopadhyay 1984, Dasgupta et al. 1987). It is further interesting to note that the long lineaments or faults that transect the Himalaya cause a curvilinear structure on the MCT; the Tista and Gangtok lineaments caused such a structure in the Sikkim Himalaya (De and Kayal 2004), and the Kopili fault in the Bhutan Himalaya (figure 1). In the Bhutan Himalaya we also observe another curvilinear structure on the MCT along the Jumuna river/lineament. Although the MCT is believed to be seismically dormant (Ni and Barajangi 1984), the north–south segments of such curvilinear structures on the MCT possibly reflect seismogenic transverse structures transecting them, and the zone of intersection causes shallow earthquakes by strike-slip faulting (De and Kayal 2004).

The 21 September 2009 strong earthquake (M_w 6.3) and its aftershocks in the Bhutan Himalaya occurred on such a north–south segment of the curvilinear MCT (figure 1). Several studies have reported that the 300 km long Kopili fault transgresses into the Bhutan Himalaya up to the MCT (Nandy 2001, Bhattacharya et al. 2002, Kayal et al. 2006). The two fault plane solutions of the main event and the seismic cross section of the main shock and aftershocks clearly indicate that the source zone is below the north–south segment of the curvilinear MCT (figure 1). We believe that the main shock and the aftershocks occurred on the intersecting Kopili fault below this north–south curvilinear segment of the MCT in the Bhutan Himalaya.

The 19 August 2009 earthquake in the foredeep Assam valley, about 100 km south of the Bhutan Himalaya earthquake, also occurred at a similar depth (∼10 km) with a right lateral strike-slip fault mechanism on the Kopili fault. The two fault plane solutions that are obtained for the 21 September Bhutan earthquake in the present study are compatible with the 19 August Assam valley earthquake solution. The 21 September Bhutan earthquake solution obtained by the teleseismic data also shows a right lateral strike-slip movement. The 19 August Assam valley earthquake on the Kopili fault possibly triggered the 21 September Bhutan Himalaya earthquake at the northern end of the fault at a similar depth and with a similar source mechanism. Kayal et al. (2006) and Bhattacharya et al. (2010) strongly argued that the Kopili fault is intensively active in the region, and is vulnerable to an impending large earthquake in the northeast India/Himalaya region. The 19 August 2009 event could have been the foreshock for the 21 September 2009 Bhutan earthquake. A larger and significant question is whether both the events could be foreshocks for a larger impending earthquake (M > 7.0) in the region.

Although, the Bhutan Himalaya falls in the Himalayan collision zone and the Assam valley in the foredeep/intra-plate zone, the long transverse Kopili fault zone links these two tectonic zones, and this gigantic transverse structure is capable of generating larger earthquakes in the Bhutan Himalaya as well as in the Assam valley. The 10 January 1869 Cachar earthquake (M 7.5, depth ∼50 km, intensity VIII+) occurred at the southeastern end of the Kopili fault causing severe damage in the northeast India region (Nandy 2001). The 23 October 1943 earthquake (M 7.2) in the Assam valley occurred almost at the centre of the Kopili fault zone (figure 1). Present seismicity recorded by the local network shows intense activity along the Kopili fault zone (figure 3). We thus conclude that the gigantic Kopili transverse structure that intersects the MCT is seismically most active in the region and is vulnerable to an impending larger earthquake (M > 7.0).

Acknowledgements

This research is undertaken under the Integrated Long Term Programme (ILTP), an Indo–Russian collaboration, supported by the Department of Science and Technology, New Delhi, India and by the Russian Academy of Sciences, Moscow, Russia. We express our sincere thanks to all the heads of the institutes concerned for their kind support of this research.