Investigation of past earthquakes in the Kopili Fault zone, NE India: New evidence of paleoliquefaction

The Kopili fault (KF) zone, one of the active faults in northeastern region (NER), has experienced large earthquakes in 1869 (M_w 7.4) and 1943 (M_w 7.2). In order to mitigate future occurrences of earthquakes in the KF, it is essential to understand its long-term seismic history and seismic hazard implications. Seismogenic liquefaction features were identified at three trench sites in the floodplain deposits of Kolong River, near KF. The liquefaction features include multiple sand dykes and sand sills and are direct response to liquefaction of saturated sediment induced during past seismic activity. A total of seven samples from marker horizons have been processed to constrain the chronology of liquefaction features using optically stimulated luminescence (OSL) dating technique. OSL age constraints from Trench 1 indicate two episodes of earthquake-induced liquefaction since perhaps AD 1692, one, possibly two episodes of liquefaction in Trench 2 since AD 1540 and Trench 3, suggest one liquefaction event during the past 1000–2000 years. From the present study, given the limited results of the dating available we can conclude that two earthquakes induced liquefaction in the vicinity of the Kopili fault zone during the past ~ 480 years. Additional excavations and dating of earthquake-induced liquefaction features are required to precisely evaluate the frequency of major earthquakes in the KF zone.


Introduction
One of the most useful secondary evidence in the category of seismically generated structures in the eld of Paleoseismology is liquefaction. Liquefaction of saturated cohesionless sediments is caused by increase of the pore-water pressure due to the propagation of the cyclic shear waves during seismic shaking (Youd 1973;Sims 1973Youd 1977Seed 1979;Obermier 2009). Liquefaction is de ned as transformation of a granular material from a solid to lique ed state due to increased pore-pressure. One of the main mechanisms, which produce liquefaction, is the cyclic shear stress/strain induced by earthquakes. Liquefaction occurs mostly in soft sedimentary sequences like ne and coarse sand/silt but has minor effect on sandy-gravels (Obermeier 1996;Owen and Moretti 2011). The liquefaction structures include sand dykes, sand blows, sand viens, pseudonodules, convolute bedding, load structure, diapirs, etc. The minimum earthquake magnitude able to cause liquefaction is M 5-6 (Allen, 1986; Obermeier and Pond 1999) and the factors that control the liquefaction are lithology, depth to the water table, epicentral distance, magnitude, earthquake duration, ground acceleration and amplitude of cycles (Seed, 1979;Obermeier 1996; Moretti et al., 1999). Several studies have used liquefaction features as proxies in various geological and tectonic settings to investigate the seismic history (e.g. Russ 1982;Obermeier 1998 The north-east region (NER) of India has experienced high-magnitude earthquakes due to the ongoing collision of the Indian Plate with the Eurasian plate to the North; and Burmese plate to the East (Ambraseys and Douglas 2004;Bilham 2004) and the continual tectonics can potentially lead to more seismic activity. Hence, comprehending the frequency of major earthquakes becomes a prerequisite by focusing at evidence of the paleoearthquakes in the tectonically active regions. In the Himalaya, Paleoseismic investigations have been carried by trenching across primary fault scarp (e.g. Anand (Bapat et al. 1983;Iyengar et al. 1999; Baro and Kumar 2017) suggested multiple seismic events to have occurred during the last millennium.
Where seismogenic faults are di cult to recognize, earthquake-induced liquefaction features have been the focus of a number of Paleoseismological studies (e.g. Amick et al. 1990;Obermeier 1998;Saucier 1989; Tuttle and Seeber 1991; Tuttle and Atkinson 2010). Several recent earthquakes, including the 1897 Shillong and the 1950 Assam, induced liquefaction but were not associated with surface rupture. These earthquakes raised the awareness that certain types of prehistoric earthquakes could not be identi ed by studying surface faults. Therefore liquefaction features are used increasingly in Paleoseismological investigations around the world, wherever sediments susceptible to liquefaction are present. Criteria for recognizing earthquake induced liquefaction features from other types of soft sediment deformation structures already have been described in several papers (e.g. Sims Kumar et al. (2016) carried out paleoseismological studies using seismogenic liquefaction in the oodplain deposits of Kopili and Kolong Rivers. They excavated trenches at two locations and observed several sand dykes. In addition, they suggest three intervals of liquefaction formations using radiocarbon and OSL dating techniques. The paleoseismological record, however, provides only part of the data necessary to fully understand seismic hazard in the region. Observations of liquefaction features in the NER (Sukhija et al. 1999;Rajendran et al. 2004;Reddy et al. 2009) and accounts of liquefaction during historical earthquakes suggest that a history of paleoliquefaction events can be gleaned from the geologic record that would shed light on the long-term behaviour of the KF.
Here we present new evidence of sand blows and sand dykes resulted from the paleoearthquakes centered at KF region. In this paper, we communicate the outcome of study carried out mainly in River/channel cutoffs of Kolong River, Assam to investigate the liquefaction elds by excavating abandoned ood plain of Brahmaputra River and age constrains on liquefaction features using OSL dating.

Seismotectonics
The contiguous distribution and frequency of large intraplate earthquakes in India are poorly understood. The subduction of Indian plate under the Eurasian plate in the north brought Himalayan mountain ranges into picture while the eastern collision-subduction zone caused the existence of the Indo-Burma ranges ( Fig. 1). As a result, the NER has a complex tectonic setting with a history of past large to great earthquakes (Ambraseys and (Fig. 1). The 300 km long and 50 km wide KF is seismically very active due to its transverse tectonics which makes the region at risk for impending large earthquakes (Kayal et al. 2012). The KF which separates Shillong Plateau from the Mikir hills, is one of the most important tectonic features of the NER. The KF zone is bounded by the Himalayan collision in north, the Indo-Burmese subduction zone to the south, the syntaxis zone in the east, and the SP in the west (Fig. 1

Methods And Research Approach
Sites that lique ed during large modern earthquakes and historic earthquakes furnish good target for paleoliquefaction studies because liquefaction often reoccur where susceptible sediments are present. Therefore we have carried out investigation for liquefaction features along Kolong River near Namgaon, Satargaon and Nampani near KF Zone in the eastern part of SP where ground failure and sand vents were reported during 1869 Cachhar and 1943 Hajoi earthquakes. Overbanks of the Kolong River were examined, during November and December 2016, for the liquefaction and other soft-sediment deformation structures, to assess spatiotemporal extent of strong shaking in KF zone (Figs. 3, 4, and 5). We found, and have studied, many earthquake-induced liquefaction features along Kolong River. The information from sites is gathered from local knowledge by oral communication. The methods employed for identifying earthquake induced liquefaction include reconnaissance of open ground and River/stream cut-offs, investigation of sand-blow/sand dyke features by making trenches, documentation of liquefaction features by logging and river exposures.
The strata here are susceptible to liquefaction, because the riverine sand beds with con ning clay layers are better situated for increased pore pressure, under shallow water table conditions. Identi cation of liquefaction features in the eld in Brahmaputra ood plain is not always suitable due to experience of monsoonal ooding and erosion by River during high ows. Consequently, during progressive excavation we observed liquefaction features viz., several sand dykes (Figs. 3-5) as well as multiple dykes rising from the sand reservoir, mainly along and adjoining areas of the Kolong River, a tributary of Brahmaputra and in distal parts of the alluvial fans. Samples for optically stimulated luminescence (OSL) age determinations were collected from different stratigraphic level to obtain maximum and minimum or contemporaneous ages of the liquefaction feature. As it is found that both sites have uvial dominated deposits which hindered to trace the presence of organic samples. The samples were collected in PVC tubes of 20 cm long and 2.5 cm diameter and OSL dating was carried out by Wadia Institute of Himalayan Geology (WIHG), Dehradun (Table 1). Several studies have shown that in a given site recurring liquefaction corresponds to multiple earthquakes occurring at different time intervals (Youd 1977; Saucier 1989; Tuttle and Seeber1991; Obermeier 1998; Sims and Garvin 1995). The method to differentiate subsequent events of liquefaction is based on the stratigraphic criteria and cross-cutting relationships.  Table 1 and illustrated on Figs. 3, 4 and 5. Along Kolong River, we found twelve liquefaction features in three trenches at three sites Namgaon (NG; Trench 1) Nampani (NP; Trench 2) and Satargaon (SG; Trench 3) (Fig. 2). Liquefaction features include sand dykes up to 23 cm wide and six sand blow deposits. The documented liquefaction features and their age constraints are discussed below.

Namgaon Site
The liquefaction features were developed in alluvial sediments and best observed at excavation area (Fig. 2). In ~ 1.8 long and ~ 2 m deep trench 1 along Kolong River at Namgaon (N 26°12'50.5"; E 92°26'41.7"), ve dykes (D1, D2, D3, D4 and D5) in a radial pattern and a small sand blow (SB) are observed (Fig. 3A). These dykes range from 25 to 55 cm height and 5 to 20 cm wide. Two generation of main dykes D1 and D2 indicate that they may be originating from the same source and two dyke outlets (D3 & D4) are originating from the D1 dyke (Fig. 3B). They are composed predominately of ne to medium sand. In addition, there is a second generation of sand dyke D2 composed of ne sand that crosscuts the older D3 dyke. The stratigraphic position and crosscutting relation of D3 dyke shows that it is older than D2 dyke and clearly indicates their formation in two distinct events which might have occurred within a short period of time. The D1 dyke intruding into host silt/clay layer is seen to terminate in to the related sand blow (Fig. 3B). This sand dyke (D1) appears to be a compound feature that was utilized during repeated provides maximum age for D2, D3 and D4 dykes and AD 1848-1885 provides maximum age for D2 dyke (Fig. 3B).

Nampani site
Trench 2 near Nampani village (N 26°21'58"; E 92°46'38.7"), revealed a 3 m long and 2 m deep sedimentary section consisting of white sand layers, brown sand, and clay (Fig. 4A). Here we observed three generations of sand blows and related sand dykes. The oldest sand blow 3 (Unit 6 ) occurs low in the section and is up to ~ 5-15 cm thick, composed of ne brown sand, and fed by two sand dykes, ranging from 8 to 10 cm in width (Fig. 4D). The sand blow immediately overlies a soil developed in clay. Dyke 2 is intruding on to the surface and light yellow in color and attains an height of 60 cm and 5 cm width (Fig. 4D). A younger sand blow 2 (Unit 4), occurring in the middle of the section is about 10-30 cm thick, composed of very ne sand containing a few small clasts, and of limited lateral extent and is fed by at least one sand dyke (Dyke 1) (Fig. 4C). The youngest sand blow 1 (Unit 2), occurring high in the section is only 50 cm thick and consists of white ne sand (Fig. 4B).

Satargaon site
At Satargaon site (N 26°23'12.5"; E 92°47'43.1), we observed two sand blows (Sand blow 1 (SB1) and sand blow (SB2)) and a dyke 2 m long and 1.2 m deep trench (Fig. 5A). The section consists of alternate layers of clay, silt and sand. The SB2 (Unit 5) is ~ 10 cm thick and SB1 is 25 cm thick and composed of ne silt and brown sand (Fig. 5B). One main sand dyke, 5 cm wide, fed the SB1. The dyke composed of ne light yellow color sand, crosscutting unit 6 and is terminated in Unit 5 (SB2). Flame like intrusion of clay deposit is also observed (Unit 4) and clasts of clay are also seen in the sand blow. Sample collected from within the sand blow 1 yielded OSL date of AD 1057-1211 (883 ± 77) ( Table 1). It provides contemporaneous age constraint for the event.

Timing, Sources and Magnitude of Paleoearthquakes
The stratigraphic criteria, geometry and cross-cutting relations of deformational features provide evidence for repeated liquefaction events to seismic activity. The deformational features found in the present study, such as sandblows, sand dikes and ame structures, are correlated to earthquake induced mechanism because their morphology indicates sudden application of a large upward-directed hydraulic force of short duration (Obermeier 1996)  In the present study along Kolong River basin, from the structural, stratigraphic, and age relations at sites suggest that the paleoliquefaction features probably formed as the result of ve earthquake events (  (Fig. 6). Evidence for Event I is recorded in Trenches 1 and 2 (Figs. 3 and  4). According to the OSL ages measured at this site (Table 1), the Event I is constrained by both maximum and contemporaneous ages (Fig. 6). The maximum age (AD 1848-1885) of Event I is provided by youngest liquefaction features in Trench 1 (Fig. 3) at Namgaon site. This event is also inferred from contemporaneous age (AD 1889(AD -1915 in trench 2 and is recorded in youngest sandblow. Evidence for Event II liquefaction which took place sometime during the period AD 1782-1826 has been recorded in Trench 1 (Fig. 6). Event III has been inferred from the age constraints of sand blow in Trench 1 (Figs. 3 & 6) at Namgaon site and Sand blow 2 in Trench 2 (Figs. 4 & 6) at Nampani site. Two contemporaneous OSL ages AD 1692-1770 and AD 1640-1716 points to an earthquake in 17th and 18th century. Present study also provide Event IV in Trench 2 and the contemporaneous OSL age 1540-1626 points to an earthquake in the 16th and 17th century (Fig. 6). Event V occurring at AD 1057-1211 is also identi ed from Trench 3 at Satargaon site (Fig. 6). about 900 year BP at the Kakotigaon along Kolong River which is near by present study sites. This data suggest that the liquefaction Event V (AD 1057-1211) at Satargaon site in the present study may correspond to 900 year BP, but only one age precludes this conclusion (Fig. 6). Dating of more paleoliquefaction features at additional sites in the KF would greatly improve our determination of the timing and the likely seismic source.
The contiguous spreading of liquefaction elds (~ 41 km) along the Kolong River indicate local earthquakes of approximately M > = 6.0 using empirical relation of Ambraseys (1988) for moment magnitude to distance of farthest liquefaction to epicentre. If all the liquefaction events resulted from the KF, then according to Ambraseys (1988) the earthquakes were approximately M − 6.5. The present study liquefaction elds recorded at least ve local earthquakes in 1000 years, with a conjectural average recurrence interval of approximately 200 years for M 5.5 to 6.5 events.
In the past, the KF has experienced several earthquakes of magnitude 4.5 to 6.2, three of magnitude 6 to 7 and two earthquakes of magnitude greater than 7 viz., 1869 Cachhar and 1943 Hajoi earthaquakes. This region has been identi ed as a probable region for future generating large magnitude earthquakes (Kayal et al. 2006(Kayal et al. , 2010(Kayal et al. , 2012.    Namgaon   ame like intrusion features. SB2 (Unit 5) is ~10 cm thick and SB1 is 25 cm thick and composed of ne silt and brown sand. One main sand dyke, 5 cm wide, fed the SB1. The dyke composed of ne light yellow sand, crosscutting unit 6 and terminated in Unit 5 (SB2). Flame like intrusion of clay deposit is also observed (Unit 4) and clasts of clay in the sand blow. Sample collected from within the sand blow 1 yielded OSL date of AD 1057-1211 (883±77) provides contemporaneous age constraint for the event. Lithology of trench section consists of Unit 1: clay, Unit 2: silt, Unit 3: brown silt-sand, Unit 4: clay, Unit 5: ne sand and Unit 6: silt-clay Figure 6 The occurrence of earthquakes depicted in trench sediments in Kopili region, is stacked up for possible correlation of events. Kumar et al. (2016) dated events are compared with our ndings and there is some consonance with the events. Black and red thick bar ones are present and previous (Kumar et al. 2016) ndings of paleoearthquakes in the Kopili fault region