- © 2016 by the Seismological Society of America
Deformed monuments can be used to help constrain the timing of an earthquake, its epicenter, and the extent of its damage. In the present work, we studied the ancient temples in the Chamba and Bharmour area from the seventh century onward in the Ravi River catchment of the Chamba district of western Himachal Pradesh, India, which lies in the Kashmir seismic gap of the northwest Himalaya. Many architectural features of the temples show earthquake‐induced deformation signatures that include tilting of the pillars and temple structures, fractures, and opening in the brick masonry. The Bharmour temples located 45 km north of Kangra show shear‐type deformation in the east–west direction. The Chamba temples, which lie 52 km northwest of Kangra and outside the rupture zone of 1905 Kangra earthquake, display shear or rigid body rotation in the north–south direction. The deformation features observed in the temples suggest that the Bharmour area lies within the 1905 Kangra earthquake meizoseismal zone, and that the 1555 Kashmir earthquake rupture zone extended southeast toward Chamba.
Early accounts of four great earthquakes documented by the Geological Survey of India are 1905 Kangra Mw 8 (Middlemiss, 1910), 1934 Bihar–Nepal Mw 8.1 (Dunn et al., 1939), 1950 Assam (now Arunachal) Mw 8.6 (Chen and Molnar, 1990) in the Himalaya, and 1897 Shillong (Assam) Mw>8 (Oldham, 1899) (Fig. 1a). There are historical accounts of other large earthquakes with Mw>7, including A.D. 1255 Nepal, A.D. 1344 Nepal, A.D. 1505 Tibet–Nepal border, A.D. 1555 Kashmir, A.D. 1803 Garhwal, and A.D. 1833 Nepal (Ambraseys and Jackson, 2003; Ambraseys and Douglas, 2004; Bollinger et al., 2016). The epicentral locations, meizoseismal areas, and rupture extents of the historic major and large earthquakes are near approximate and not well constrained. For example, the epicenter of the A.D. 1934 Bihar earthquake, magnitude Mw 8.1, was earlier thought to be located within the Main Boundary thrust (Dunn et al., 1939), but this was later revised indicating the epicentral location to the east of Kathmandu in the Lesser Himalaya (Ambraseys and Bilham, 2000). Similarly, the magnitude of the A.D. 1950 Assam (Arunachal) earthquake as well as the rupture extent and propagation is not well constrained. It is not feasible to assign precise epicenters to the past large earthquakes due to lack and poor resolution of the available seismic data. Attempts have been made in some cases to define the epicentral areas based on intensity data and damage scenario (Ambraseys and Jackson, 2003; Ambraseys and Douglas, 2004). Paleoseismological studies along the Himalayan front provide the evidence of past historical earthquakes and their sizes in terms of amount of slip along the faults (Lavé et al., 2005; Kumar et al., 2006, 2010; Sapkota et al., 2013).
Archaeoseismology has developed as an earthquake geology discipline to study ancient earthquakes through deformation traces left in archaeological monuments. It is used as a sensor to date unknown earthquake events that only left traces in the archaeological and possibly also in the geological record (Sintubin, 2011). The monuments serve as a link between instrumental and historical seismology (Caputo and Helly, 2008). In the Himalaya, a number of archaeological temples have been used to reveal the history of paleoearthquakes. Bilham et al. (2010) studied the tenth century Shiv temple in Srinagar Valley, India. Based on deformation features and photographs taken at different times, they suggested that the temple suffered European Macroseismic Scale (EMS) intensity VII due to an earthquake in 1828. Rajendran et al. (2013) used the archeological temples of the Garhwal and Kumaun to obtain spatial and temporal constraints on three significant earthquakes of A.D. 1255, A.D. 1505, and A.D. 1803, and they calibrated the epicentral location of 1803 earthquake. Thakur, Pandey, et al. (2010) identified two earthquake events, between <A.D. 1400 to <A.D. 850 and <A.D. 350, respectively, through radiocarbon and optically stimulated luminescence dating of archaeological deposits associated with damaged house structures at an archeological site at Khajnawar near Saharanpur, ∼12 km south of the Himalayan front.
The present study area includes the seventh century temples of the Chamba and Bharmour areas that lie northwest and north of the Kangra between the rupture zones of the A.D. 1905 Kangra and A.D. 1555 Kashmir earthquakes in the Kashmir seismic gap (Fig. 1b). In the Nepal Himalaya, the recurrence interval of great Mw>8 earthquakes is ∼650 yrs (Sapkota et al., 2013), and the recurrence interval for the A.D. 1905 Kangra earthquake is ∼350 yrs (Molnar, 1987; Yeats and Thakur, 1998). The town of Chamba and its old capital of Bharmour have a history that dates back to A.D. 700 with many old temples and monuments (Bharti, 2001). Its cultural history and heritage is much older than the estimated recurrence interval of the major earthquakes. We observed deformation signatures in the architectural features of the Chamba and Bharmour temples, and these features provide an opportunity to improve the estimates of the rupture extent of the major earthquakes in the adjoining regions. In the present study, the existing post‐seventh century temples of the Chamba region are used as a proxy to understand the paleoearthquake history of the area.
HISTORY OF CHAMBA STATE TEMPLES
Chamba, a hill kingdom, has a well‐documented early history since its establishment with ancient temples dating back to the tenth century (Hermann, 1955; Bharti, 2001; Haṇḍa, 2005). In the ancient history of Chamba, it is believed that the area was inhabited by certain Kolian tribes (Haṇḍa, 2005). In A.D. 500, the Rajputs founded Bharmour (old name Brahmpura) in the Budhil River Valley, a tributary of the Ravi River, 41 km east of the town of Chamba. Later one of their descendants Sahil Varman shifted the capital from Bharmour to the present location of Chamba. The Chamba kings built many temples at different places during their long reign. Most notable temples are the Lakshmi Narayan group of temples built between ninth and tenth centuries and the eighteenth century Chamunda Devi temple in Chamba. The Bharmour with its temples has a history ranging A.D. 700 to A.D. 1000, and the Chitrari temple was built ∼A.D. 680 (Bharti, 2001).
Based on the construction material, the temples of the Chamba area are broadly classified into two groups: (1) wooden temples and (2) stone temples. In the wooden temples, the structural elements including the walls and pillars are made of wood, and the roof is made of slate tiles with a broad circumference such as the Lakhna Devi temple at Bharmour and the Shiv Shakti temple at Chitrari (Hermann, 1955; Haṇḍa, 2005). The stone temples have a narrow circumference and greater height with a conical top that forms an apex, locally called the Shikhar (Nagar, 1990). The Lakshmi Narayan group and the Bharmour group of temples belong to this category. The stone temples are further classified into two groups: (1) small temples (height 3–4 m) made with small stone bricks with cementing material at their binding and (2) large temples (height >6 m) constructed of stone bricks interlocked without cementing material.
Deformation in Chamba Temples
In the town of Chamba, most of the temples retain their original form. Some temples show damage signatures. The damage to the temples may be due to aging, foundation subsidence, or an earthquake. In the cases of aging and subsidence, there is no regular pattern of damage, whereas in the case of earthquake produced deformation, consistency in deformation pattern is expected. The Chamba temples show a regular deformation pattern, which suggests earthquake‐induced deformation. In the Chamba–Bharmour area, seven groups of historic temples are studied (Fig. 2a,b) and their deformation patterns are recorded and summarized in Table 1. This study includes measurement of the tilt direction, the amount of inclination of the pillars as well as temple structures, and cracks on the building stones.
Lakshmi Narayan Temples
The Lakshmi Narayan group of stone temples is located in the western part of the town of Chamba (Fig. 3). The temples are of the tenth century of Shikhara style and were built by Raja Sahil Verman (Haṇḍa, 2005). In the Lakshmi Narayan complex, there are four large and two small temples arranged in a row from north to south (Fig. 3). The temples are largely made of carved stones with figures and motifs. In the small shrine enclosure on the outer wall of the Lakshmi Narayan temple, the space between the left and right pillars and the walls increases from top to bottom, and the right pillar shows a horizontal fracture in the middle (Fig. 4a). The tilted pillars in the north–south direction invoke shear deformation. Outside the periphery of the Lakshmi Narayan temple, a 1.5‐m‐tall stone pillar with a carved human face is tilted north at an angle of 8° (Fig. 4b), which can be attributed to rigid body rotation. In the Gauri Shankar temple, two large stone pillars are used for support of the structure. Variations in the width of the open space from bottom to top between the pillars and walls show a similar sense of tilt as that observed in the Lakshmi Narayan temple (Fig. 4c). The doorstep stone at entrance of the Gauri Shankar temple is broken and shows slight upwarping (Fig. 4d). On the southern side of the Lakshmi Narayan temple in the shrine enclosure, the right stone pillar has been replaced with a cement pillar, suggesting repair, and the left pillar shows a shear fracture in its middle (Fig. 4e). On the north sidewall of the Lakshmi Narayan temple, the left pillar of the shrine enclosure shows cracks and is clamped with steel strips (Fig. 4f). The space between the left pillar and the wall increases from top to bottom, suggesting tilt of pillar to the north.
At Chontra, the front of Sri Guru Nanak Sarain (rest house) in the western part of the town of Chamba, there are 3‐m‐high stone built Shiv temples of the Shikhar type of architecture. Of these three, the temple in the front is tilted S20°E at 6° (Fig. 5a), whereas the other two temples also show minor tilting in the same direction. About 150 m away from this locality near Sanatan Dhram Sanskrit Mahavidyalay, the Shiv temple is tilted N20°W at an angle of 10° (Fig. 5b). The tilting of the whole body structure of these temples invokes rigid body rotation.
The Kariya Shiv temple is situated 4 km southeast of the town of Chamba. The main body of the temple shows tilting to the south (Fig. 6a). In the upper part of the temple, the building stones show openings with displacements parallel to the layering of the bricks. Some bricks are broken in their middles. The top canopy of the temple has shown evidence of sliding in the south direction.
Chamunda Devi Temple
The Chamunda Devi temple lies about a kilometer east of the Lakshmi Narayan temples (Fig. 6b). A brass metal bell in the mandap entrance of the temple is incised with a date 2 April 1762, suggesting the construction date of the temple. The temple has a north Indian Nagara architectural style and has a rectangular layout on its outer perimeter. The temple is constructed on a 2‐m‐high stone platform. The temple does not show damage reconstruction or any earthquake‐induced deformation feature, suggesting that no major earthquake affected the area since A.D. 1762.
Deformation in Bharmour and Chitrari Temples
The Chitrari temple is located 15 km west of the Bharmour temple and 26 km east of the town of Chamba. It is a wooden temple of Shiv Shakti constructed during A.D. 680 (Haṇḍa, 2005). The original shrine of the Shakti Devi temple is surrounded by a gallery supported by 12 heavy wooden pillars. The entrance to the sanctum is decorated with a double row of standing pillars on each side. The temple shows earthquake‐induced deformation features. The entrance gate of the main temple is inclined toward N80°E at an angle of 5° and the supporting pillar also is inclined toward N83°E at an angle of 6° (Fig. 7a,b).
Bharmour Group of Temples
Bharmour, formerly known as Brahmpura, is situated in the Budhil Valley, 41 km to the east of Chamba. It lies north of the Dhauladhar range and 45 km north of the city of Kangra and ∼31 km from Dharamsala. Bharmour is famous for its group of temples “the Chaurasi temples complex (mandir).” The Bharmour temples were built between A.D. 700 and A.D. 1000. The stone temples have shikhar style architecture with stone carvings on the outer walls. The temples show some earthquake‐induced deformation features. The Narshing temple is situated on the north side of the temple periphery. At the entrance door of the Narshing temple, two stone pillars are tilted, as indicated by variations in the open space from bottom to top between the pillars and walls (Fig. 8a). The right pillar shows open space increasing from bottom to top, and the left pillar reveals open space increasing from top to bottom. The upper part of right pillar is tilted toward the N80°W direction with ∼8 cm of displacement. The increase in open space between wall and pillar from bottom to top in the right pillar is opposite to the left pillar, suggesting a shear motion. Inside the temple, there are opened cracks in the wall corner (Fig. 8b). In the nearby small temple, the right pillar shows a small displacement and the carved basal stone displays an inclined fracture (Fig. 8c). In the adjoining temples, the building stones show opening up parallel to the layering (Fig. 8d). There are also small ∼3‐m‐high temples in the complex. These temples show prominent deformation signatures. In one of the small temples, the top Chattra (roof top) is made of flat stone with small, vertical carved columnar stones at the center of the temple roof (Fig. 8e). Here, the Chattra is shifted from its original position. The bottom flat stone of the Chattra together with columnar stone are shifted from its original position in the west direction relative to the main temple structure, suggesting a shear motion.
LARGE EARTHQUAKES IN THE KASHMIR AND KANGRA AREA
The Chamba temples are located between Kangra and Kashmir within the Kashmir seismic gap. Chamba lies southeast of Kashmir Valley and Bharmour is located north of Kangra. There are archival and historical accounts of damaging earthquakes in the Kashmir Valley. Based on the damage accounts, some assessments have been made about the intensity and epicentral area of the A.D. 1555, A.D. 1828, and A.D. 1885 earthquakes in Kashmir (Oldham, 1883; Iyengar and Sharma, 1996; Iyengar et al., 1999; Bilham et al., 2010).
A.D. 1555 Kashmir Earthquake
Most of the information about this earthquake comes from the historical records given in Tarik‐i‐Kashmir (Koul, 1710) and Tabqat‐i‐Akbari (Ahmad, 1911). In September 1555, a devastating earthquake occurred in the Srinagar Valley followed by a sequence of aftershocks that continued for several days (Hindushah, 1607; Haidar, 1620). This was a destructive earthquake in Kashmir, which ruined towns, killed several hundred people, and changed the course of the River Vesha, a tributary of the Jhelum (Anand, 1913). The damage belt extended toward the southeast from Srinagar in the Kashmir Valley, and its epicentral area lies southeast of the location of the A.D. 1885 earthquake (Iyengar et al., 1999). The A.D. 1555 earthquake was assigned intensity VII (Medvedev–Sponheuer–Karnik) and magnitude Mw 7.6 (Ambraseys and Jackson, 2003; Ambraseys and Douglas, 2004).
A.D. 1828 Kashmir Earthquake
On 26 June 1828, a strong earthquake shocked the Kashmir area. About 1000 persons lost their lives and 1200 houses were destroyed. In some places, the earthquake produced fissures up to 13 m long and 1.3 m wide on the ground. There were ∼20 pre‐ and postearthquake shocks occurred in the region (Vigne, 1844). Bilham et al. (2010) studied the tenth century Shiv temple Pandrethan in the Srinagar Valley, Kashmir, and based upon existing deformation features and photographs taken at different times, they suggested EMS intensity VII for the A.D. 1828 earthquake in the Kashmir region.
A.D. 1885 Kashmir Earthquake
On 29 May 1885, an earthquake occurred near Wullar Lake, 35 km northwest of Srinagar (Jones, 1885; Iyengar et al., 1999). The epicentral area of this earthquake was probably near Baramulla. The earthquake caused total destruction in the districts of Sopore and Baramulla, with the number of causalities reported varying from 3000 (Bamzai, 1962) to 3500 (Lawrence, 1895). The destruction by the earthquake at Srinagar was not high as compared to at Baramulla. At Pattan, fissures 1.0 m wide (Neve, 1928) to 7 m wide and 1.60 km in length at Baramulla (Jones, 1885) were reported. The destructive effect of this earthquake was considerable over an area of 1000 km2 (Jones, 1885). Based on historical reports, Ahmad et al. (2014) estimated a 54‐km‐long coseismic surface rupture with a damage region spread over an estimated area of 5000 km2. Ambraseys and Douglas (2004) assigned magnitude Mw 6.3 to the 1885 A.D earthquake.
A.D. 1905 Kangra Earthquake
The 4 April 1905 Kangra earthquake, magnitude Mw∼8, killed 18,815 people and strongly affected two different areas, one extending from Kangra to Kulu‐Lahaul and the other around Dehradun (Fig. 1b). The Kangra area felt a maximum intensity of X (Rossi–Forrel), and Dehradun had a maximum intensity of VIII. The maximum damage was concentrated around Kangra–Dharamsala and the area to its east including Mandi and Kulu‐Lahaul (Middlemiss, 1910). Reevaluation of seismograms of the A.D. 1905 event gave a revised magnitude of Mw 7.8 (Ambraseys and Bilham, 2000). The Kangra earthquake is thought to have triggered the Dehradun event with deep (∼40 km) focus (Hough et al., 2005). The Bajreshwari (Kangra) temple has a well‐recorded history of about 1000 years. Mahmud Ghaznavi, who ruled Afghanistan and parts of Iran and northwest India, invaded Nagarkot (now Kangra) in A.D. 1008 and plundered the temple wealth. The main temple at Kangra Bhawan was totally collapsed, and the Kangra fort wall was tilted to the north by the 1905 Kangra earthquake, as documented in photographs (Fig. 9a,b) by Middlemiss (1910). The Bakloh cantt, 21 km southwest of Chamba and 51 km northwest of Kangra, was a training center of Gurkha riflemen where no damage was reported. North of the Dhauladhar range in the Ravi catchment area of the state of Chamba, great damage was reported from the Bharmour area, whereas the state capital of Chamba remained unaffected (Negi, 1963). The rupture parameters of the A.D. 1905 earthquake estimated through remeasurements of historical triangulation points in the area indicate a 100×55 km rupture, extending toward the southeast and dipping northeast with ∼4 m slip. The southern extent of this rupture is interpreted to coincide with the Jawalamukhi thrust (Wallace et al., 2005). The instrumental‐recorded microseismicity obtained through the Wadia Institute of Himalayan Geology network suggests a ramp beneath the Dhauladhar range and a flat base beneath the Chamba sequence (Thakur et al., 2000; Kumar et al., 2009). Malik et al. (2015) identified a right‐lateral strike‐slip fault, the Kangra Valley fault, located 65 km north of the Himalayan front in the Kangra reentrant and proposed it as a surface rupture of the 1905 Kangra earthquake with the epicenter in the Kangra Valley.
DISCUSSION AND CONCLUSION
In the post‐2005 Mw 7.6 Kashmir earthquake field survey, Jayangondaperumal and Thakur (2008) observed a horizontal shear sense displacement of the roof of a small mosque with the total collapse of walls in a village near Poonch, ∼100 km southeast of the epicenter of the 2005 Kashmir earthquake. In the 2011 Mw 6.9 Sikkim earthquake at a distance of 60 km south of the epicenter, Thakur et al. (2012) observed shear displacement displayed by opening of spaces within the masonary bricks in a well‐constructed hotel. In the Chamba and Bharmour regions, the temples are only tilted without collapse. This is only possible if the temples were shaken with an intensity of IX and VIII. Shaking at the higher intensity XI or X would cause collapse of the structure (Bilham et al., 2010). If the structure drifted more than the weight of the roof, then the deformation would tend to bring down the entire structure. The Chamba gazetteer was first published in A.D. 1910 and later significant contributions were added on different aspects including history, geology, and archeology by several workers. All the earlier accounts are revised in the last Chamba gazetteer by Negi (1963), the district collector. In the district gazetteer, there is no record of a large earthquake causing major destruction in the Chamba area since the inception of the town of Chamba, whereas Bharmour is mentioned as having suffered great damage to property in the 1905 Kangra earthquake. These accounts further corroborate that Chamba was outside the rupture area of the 1905 Kangra earthquake. The damage survey derived isoseismals (Middlemiss, 1910) and rupture estimated through geodetic measurements (Wallace et al., 2005) indicate the rupture propagated from Kangra to the southeast toward Mandi. According to their interpretation, the northern extent of the A.D. 1905 Kangra earthquake rupture lies beneath the Dhauladhar range and the southern extent nearly coincides with the Jawalamukhi thrust. Deformation features observed in the Bharmour temples suggest that the area lies in the Kangra rupture zone, implying a northward shift of the rupture area as inferred by earlier workers (Wallace et al., 2005). We propose that the northern extent of the rupture extends north of Dhauladhar toward the Bharmour area, and the southern extent coincides with the Medlicott Wadia thrust (MWT) (Thakur, Jayangondaperumal, and Malik, 2010; Thakur and Jayangondaperumal, 2015). This interpretation is corroborated with the mapping of an active fault rupture designated as the Kangra Valley fault (Malik et al., 2015), which is an imbricate (Gharo thrust) of the MWT (Thakur, Jayangondaperumal, and Malik, 2010).
Chamba is 52 km northwest of Kangra and 200 km southeast of Srinagar. The architectural artifacts of the Chamba temples show shear deformation that cannot be attributed to the Kangra earthquake. The most likely candidate is one of the Kashmir earthquakes. The A.D. 1555 Mw 7.6 earthquake (Ambraseys and Douglas, 2004) was the most destructive known event in the Kashmir Valley, with damage extending from Srinagar toward the southeast. Assuming that the 1555 epicenter lies around Srinagar as inferred from the historical record and that the rupture propagated ∼80 km to the southeast, similar in length to that of 2005 Kashmir earthquake of the same magnitude, imply that the Chamba temples were near the eastern extent of the 1555 rupture. In Hajipur paleoseismology trench on the Beas river flood plain, Malik et al. (2010) reported an earthquake event occurred between A.D. 1500 and A.D. 1600 that may correspond to the A.D. 1555 Kashmir earthquake. The Hajipur is located on the Sub‐Himalayan front at nearly the same longitude as that of Chamba, which further corroborates the 1555 rupture extending to the Chamba area.
DATA AND RESOURCES
The data on deformation features in the temples were collected in the fields at Chamba, Chitrari, and Bharmour areas by the authors. All other data, observations, and interpretations referred in our article are based on publications of different workers.
The study was funded by the Department of Science and Technology, New Delhi sponsored project. We thank the director of Wadia Institute of Himalayan Geology (WIHG) for providing the facility and administrative support. M. J. acknowledges S. S. Bhakuni, WIHG, and director of National Centre for Earth Science Studies (NCESS) for their support. We also acknowledge the reviewers for their useful comments in improving the article.