1 Introduction

Intraplate earthquakes occur in the plate interior but are possibly driven by plate boundary forces (Wdowinski 1998 and references therein). A component of the net forces acting in the plate boundary may be transferred towards the plate interior, which slowly deforms the intraplate region (Zoback et al. 1989). However, local structural heterogeneities generally play a significant role to re-distribute this strain. Several hundred to thousands of years are required to accumulate high stress in intraplate regions in order to generate a large magnitude earthquake (Kanamori and Brodsky 2004). However, in the last two to three decades, certain areas within the intraplate or tectonically stable regions have experienced rapid increases in seismicity due to anthropogenic activities such as mining, construction of large gravity dams, injection of fluid into the subsurface to increase formation permeability, and injection of water for enhanced geothermal systems (McGarr et al. 2002 and references therein). It is seen that such activities significantly alter the pre-existing subsurface conditions in multiple ways, e.g., by increasing pore fluid pressure in the fault zone, by reducing frictional strength through chemical alteration and other mechanisms (Sibson 1974, 1977a, b). As a result, earthquake nucleates in pre-existing fault(s)/fractures well before the normal or expected recurrence time. Study of the subsurface stress regime, rate of deformation of the crust, localised zone(s) of high-stress concentration, etc., are essential to understand the seismic potential of such regions. In addition, study of the deformation features, microstructural variations preserved in rocks during earthquakes, composition and properties of fault zone materials and pseudotachylites may provide critical constraints in understanding the mechanics of earthquake and faulting processes. Structural analysis of the intact host rock and the fault zone rocks may provide new insights into the subsurface kinematics of the region.

Scientific drilling along active faults provide rare opportunities to sample fault core and to study the internal structure and deformation mechanism(s) of the fault zone (Reches and Ito 2007). In recent times, scientific drilling has been carried out in a number of active faults worldwide, e.g., San Andreas Fault Zone Observatory at Depth (SAFOD), Taiwan Chelungpu Fault Drilling Project (TCDP), Nojima Fault drilling in SW Japan, Deep Fault Drilling Project (DFDP) in the Alpine Fault in New Zealand and Gulf of Corinth in Greece. These scientific drilling programmes have brought forth valuable information on deep fault zone rocks from studies on recovered core samples (Ando 2001; Lin et al. 2001; Ma et al. 2006; Schleicher et al. 2006; Bradbury et al. 2007; Yeh et al. 2007; Zoback 2007; Wu et al. 2008; Louis et al. 2008; Lockner et al. 2009; Zoback et al. 2010; Holdsworth et al. 2011; Boullier 2011; Taylor et al. 2011; Carpenter et al. 2012; Sutherland et al. 2012; Townend et al. 2013; Warr et al. 2014; Toy et al. 2015, 2017; Allen et al. 2017). However, the delineation of the subsurface fault zones and acquisition of cores from the fault zone(s) and their surrounding host rocks are often quite complicated. Geophysical well-logging plays an important role in the characterization of fault damage zones at depth. Generally, physical and mechanical properties of fault zone rocks are significantly different from the intact host rock formation. Measurements in multiple scientific boreholes have shown that fault zones are generally characterised by low density, low resistivity, low P- and S-wave velocities, high porosity, high Poisson's ratio and low Young's modulus (Ben-Zion and Sammis 2003; Boness and Zoback 2004; Hickman and Zoback 2004; Ito and Kiguchi 2005; Faulkner et al. 2006; Ikari et al. 2009; Carpenter et al. 2009, 2014; Jeppson et al. 2010; Zoback et al. 2011; Townend et al. 2013; Jeppson and Tobin 2015; Allen et al. 2017; Goswami et al. 2019). Additionally, in many cases, significant rotation in subsurface stress orientation is observed in proximity to the fault damage zones (Barton and Zoback 1994; Brudy et al. 1997; Wu et al. 2007; Lin et al. 2010; Goswami et al. 2020). Therefore, downhole geophysical measurements provide important information for distinguishing between fault damage zones and the surrounding intact host rocks, which may provide a basis for sidewall coring and/or supporting further critical investigations on the cores/cuttings obtained from such zones of anomalous rock properties (Misra et al. 2017; Goswami et al. 2020). Additionally, modern high-resolution borehole image logs provide critical information about the orientation of subsurface structural features that could not be obtained otherwise from cores, as in most cases, cores are generally not oriented or cores are missing due to poor recovery in highly fractured zones.

Cores/cuttings retrieved through scientific drilling from seismogenic depths provide valuable information about the composition, texture, internal structure, fracture patterns, deformation mechanism of the fault zone rocks, which play an important role in controlling fault movements and also explaining the role of fluids in activating preexisting faults (Evans and Chester 1995; Caine et al. 1996; Evans et al. 1997; Vrolijk et al. 1999; Faulkner and Rutter 2003; Wibberley et al. 2008; Bradbury et al. 2015). For example, fault zone materials from SAFOD reveal that fault rocks are comprised of serpentinite-bearing clay gouge of low frictional strength, which is responsible for active creeping in a ~200 km long section of the San Andreas Fault and also explains the long-lasting debate on the stress-heat flow paradox (Lachenbruch and Sass 1980; Hickman and Zoback 2004; Scholz et al. 2006; Lockner et al. 2011; Bradbury et al. 2011). Detailed geological analyses indicate that the amorphous materials observed in the gouge is a result of comminution of grains due to active creeping and authigenic clay growth, not of melt origin (Janssen et al. 2014). On the other hand, the Chelungpu fault zone contains fault breccias and pseudotachylites (Ma et al. 1999, 2001; Yeh et al. 2007). Unlike SAFOD, the amorphous materials present in the clay gouge of the principal slip zone (PSZ) of Chelungpu fault are of melt origin, formed by a rise in temperature during seismic slip (Janssen et al. 2014). On the other hand, mineralogical and structural analysis of fault core in the Alpine Fault (DFDP) reveals PSZ gouge and cataclasite-series of rocks formed due to elastico-frictional deformation within the fault zone (Toy et al. 2015). Sutherland et al. (2012) provided a detailed account of the formations of cataclasites and mylonites in the presence of fluids within the fault zone. Thus, mineralogical and structural studies of fault zone rocks from different regions help in characterizing the internal structure of the fault zones and mechanism of earthquake genesis in a region.

The Koyna seismogenic zone in the Deccan trap province, western India (figure 1), characterized by shallow and recurrent water-reservoir triggered seismicity during the past five decades, provides an excellent site to study potential brittle deformation in Archean granitoids through scientific drilling. Since the impoundment of the Koyna reservoir in 1962, the region has experienced the largest triggered earthquake of magnitude M~6.3 in 1967, 22 earthquakes of magnitude 5.0 ≥ M ≥ 5.9, more than 200 earthquakes of magnitude M≥3.5 and several thousand smaller earthquakes (Gupta 1992, 2002, 2011; Gupta et al. 2015, 2016). The seismic activity was enhanced after the impoundment of the Warna Reservoir, located ~20 km to the south of Koyna Dam. Notable features constrained from seismological studies are as follows: (i) the entire seismic activity is restricted within a relatively small volume, covering an area of ~20 km × 30 km to the south of the Koyna Dam and extending southward to the vicinity of the Warna Dam, (ii) seismicity extends to a depth of ~10 km, with the great majority of the earthquakes in depth range 2–8 km, (iii) earthquake hypocentres are located in the granitic basement rocks underlying the Deccan basalt, and (iv) the frequency of earthquakes is modulated by the changes in the water levels of the Koyna and Warna reservoirs, particularly during the monsoon and post-monsoon seasons (Gupta 1992, 2002; Talwani 1997a, b; Gupta et al. 2015). The 1967 M~6.3 earthquake formed a NNE–SSW trending surface rupture zone, known as the Donichawadi fissure zone, and mapped a few km southeastward of the Koyna Dam along a length of ~4 km and width of ~200 m (GSI 1968; Misra et al. 2017). Recent seismological studies show a concentration of seismicity along the trend of the fissure zone (Gupta et al. 2017; Shashidhar et al. 2017). Since earthquakes occur predominantly in the granitic basement rock, detailed investigations on basement rock samples may provide valuable insights for understanding subsurface processes associated with the recurrent seismicity in the region.

Figure 1
figure 1

Map of Koyna seismogenic zone showing the distribution of cored boreholes KBH-1 through KBH-9 and borehole KFD1. The seismicity data, taken from Shashidhar et al. (2017) is also plotted. The Koyna (Shivajisagar) and Warna artificial water reservoirs and the trace of the Donichawadi Fault are also shown. The Western Ghats escarpment divides the region into a low-lying western part called the Konkan coastal plains (light shaded region) and an elevated region in the eastern part (dark shaded region). Inset shows the location of Koyna in the Deccan Traps province (shaded green) on the outline map of India.

Full size image

In the present study, we carry out geological investigations on the core samples of intact as well as fault zone rocks recovered from depths down to 1.5 km in the granitic basement through scientific drilling at four sites in the Koyna seismogenic zone. Based on our investigations, we report: (i) the mineralogical and structural features of the basement granitoids and deformation sequences for the first time, (ii) signatures of brittle deformation and their possible association with recurrent seismic activity at depth, (iii) composition of chemically altered fault zone rocks, (iv) evidence of fluid channelization at depth, and (v) fractures and their relationship with the pre-existing anisotropy, and discuss the implications of the acquired data towards better understanding of the recurrent triggered earthquakes in the region.

2 Geologic setup

The rocks exposed in Koyna and the surrounding region of western India consist of Deccan flood basalts that erupted as a consequence of the passage of Indian plate over the Reunion hotspot about 65 my ago (Morgan 1972; Cox and Hawkesworth 1985; Beane et al. 1986; Courtillot et al. 1986). Duncan and Pyle (1988) and Courtillot et al. (1986) suggested a narrow span of ~1 my for the eruption of the lavas with a mean age of 65 Ma. However, other studies have reported a larger time span of a few million years for the volcanism (Venkatesan et al. 1986; Courtillot et al. 1988). Typically, the flood basalt formation comprises a thick succession of lava flows, with the individual flows varying in thickness from a few metres to several tens of metres. The Deccan flood basalts exposed in different parts of western and central India have been studied extensively to characterize the lava flows from geological, geochemical and geochronological aspects (Cox and Hawkesworth 1985; Beane et al. 1986; Devey and Lightfoot 1986; Venkatesan et al. 1986; Duncan and Pyle 1988; Khadri et al. 1988; Mahoney 1988; Mitchell and Widdowson 1991; Seth 1999, 2007). However, the thickness of the Deccan Traps and the characteristics of the underlying basement rocks have remained a subject of debate due to the lack of deep drilling in the region. The thickness of the Deccan Traps is believed to vary from a few thousand metres in the western parts to a few tens of metres in the eastern parts of the flood basalt province. Drilling through the Deccan Traps at Killari, near the southeastern margin of the province, revealed a trap thickness of 338 m (Gupta and Dwivedy 1996). In contrast, continuous vertical successions of basaltic flows with thickness of more than 1200 m are observed along some road sections in the Western Ghats (Bondre et al. 2004). The Mahad–Mahabaleshwar section in the Western Ghats provides a relatively uniform, undeformed flood basalt sequence. Najafi et al. (1981) have reported 47 flows of Deccan Traps forming a 1240 m thick horizontal sequence along this section. Geophysical studies carried out during the past few decades have also attempted to characterise the thickness of Deccan Traps. On the basis of deep resistivity sounding data collected along Solapur–Karad–Guhagar traverse, Kailasam et al. (1976) inferred the thickness of the Deccan Traps in the range 370–1000 m. Deep seismic sounding studies carried out along a profile from Guhagar on the west coast to Chorochi in the east and passing through Koyna and Karad revealed variations in the thickness of the Deccan trap from 0.4 km in the east to 1.5 km near the west coast (Kaila et al. 1981). Trap thickness was interpreted to be ~0.4 km near Chorochi, ~0.6 km near Karad, 1–1.1 km near Koyna and then decreasing to ~0.7 km between Alore and Chiplun. However, information on the rock formations below the Deccan Traps was meager, limited to interpretations from geophysical data only.

Scientific drilling studies in the Koyna region, undertaken by the Ministry of Earth Sciences, India, in 2012, provide a rare opportunity to gain access to the entire thickness of Deccan flood basalt formation and the underlying basement rocks. Vertical core drilling was carried out at nine sites in the vicinity of the Koyna seismogenic zone during the preparatory phase of investigations to study the reservoir-triggered seismicity in the region. The borehole sites are located along the periphery of the seismogenic zone, extending from Rasati (close to the Koyna Dam) in the north to Udgiri (south of the Warna Reservoir) in the south and cover both eastern and western parts of the Western Ghats escarpment (figure 1). The first direct information on trap thickness in this region was obtained from borehole KBH-1 at Rasati, near Koyna where the thickness was found to be 932.5 m (Roy et al. 2013). All the nine boreholes penetrated the Deccan basaltic pile and continued a few hundred metres into the underlying granitic basement (Gupta et al. 2017). Preliminary studies on the core samples of Deccan basalt formation, including the physical characteristics of the lava flows, were reported from two boreholes KBH-1 and KBH-7 by Sinha et al. (2017) and Misra et al. (2017), and information on the rock types comprising the basement granitoids from four boreholes KBH-1, KBH-5, KBH-6 and KBH-7 were reported by Misra et al. (2017). The cores reveal a relatively sharp contact between Deccan basalt formation and the granitic basement, with the contact zone devoid of infra-trappean sediments but consisting of less than a meter to a few meters thickness of rock representing a mish-mash resulting from the contact of hot lava flows emplaced over the uppermost weathered portions of the granitic basement. The thickness of Deccan Traps in the Koyna seismogenic zone is found to vary in the range of 412–1251 m. The boreholes to the west of the Western Ghats escarpment reveal trap thicknesses in the range of 412–518 m, in contrast to the boreholes on the eastern side, which show significantly larger thicknesses, 775–1251 m. In spite of the large surface relief (order 1000 m) and the significant variations in trap thickness, the undulations in the basement topography are relatively small, generally within 100 m. The basalt-basement contact is relatively sharp, characterized by cohesive material intermixed with clasts of both basalt and granite within dark-coloured ferruginous cement. The contact zones in individual boreholes vary in thickness, 2 m in KBH-1, 45 cm in KBH-5, 11 m in KBH-6 and 90 cm in KBH-7.

The basement cores from boreholes KBH-1, KBH-5, KBH-6 and KBH-7, recovered from the northern, western, southern and eastern parts of the study region, respectively, are considered as representative of the basement granitoids in the Koyna seismogenic zone. Also, these boreholes provide access to substantial thickness of the basement granitoids, 588, 406, 715 and 251 m, respectively. The location, total vertical depth, and thicknesses of Deccan basalt and granitic basement in the individual boreholes are listed in table 1. Lithologs prepared from the study on cores show the spatial variations in composition of the granitoids in the region (figure 2). In general, the basement rocks are composed of granite, granite-gneiss and migmatitic gneiss with localised occurrences of amphibolites. Mesoscopic and microscopic studies have shown that the basement section represented in borehole KBH-1 is dominantly composed of granite gneiss (~60%), migmatitic gneiss (~25%), granite (~10%), amphibolite and other mafics (~5%). Granites exhibit prominent igneous textures with interlocking arrangement between grains. Granite-gneisses show well developed gneissosity in most cases (figure 3a). The basement section represented in borehole KBH-5 is dominantly composed of migmatitic gneiss including amphibolite (~65%), granite gneiss (~25%), and granite (~10%). Prominent ductile deformation bandings are present within the gneisses (figure 3b). The basement section represented in borehole KBH-6 is dominantly composed of granite (~55%), amphibolite and other mafics (~20%), granite gneiss (~15%), and migmatitic gneiss (~10%). Prominent deformation bandings are found in granite gneisses (figure 3c). The basement section represented in borehole KBH-7 is dominantly composed of migmatitic gneiss (~60%), granite gneiss (~20%), granite (~15%), amphibolite and mafic dyke (~5%). Igneous textures are predominantly present within these granites and granite gneisses (figure 3d).

Table 1 List of boreholes of the present study, along with the thicknesses of basalt and granitic basement rock as observed from studies on core samples.
Full size table
Figure 2
figure 2

Lithologs prepared from cores of granitic basement section in boreholes KBH-1, KBH-5, KBH-6 and KBH-7. These lithologs represent the variation in basement rock types in the Koyna seismogenic zone. Major rock types are granite, granite gneiss and migmatitic gneisses. In borehole KBH-7, a mafic dyke body that cross-cuts the basement section has been observed. Depth intervals marked as A, B, C and D represent the scale of deformation (SQD), with C and D representing the fault zones within the basement cores.

Full size image
Figure 3
figure 3

Photographs of borehole core samples from the Koyna seismogenic zone showing the sharp contacts between Deccan basalt and granitic basement rock at different sites and deformation signatures of ductile and brittle nature preserved in the granitic basement rock. (a) Core sample from KBH-1 at 1060.00 m depth representative of S1–2 segregated bandings within quartzofeldspathic gneiss. (b) Core sample from KBH-5 at 857.40 m depth showing segregated layerings, which are mainly S1–2 fabrics. A near-vertical fault surface displacing S1–2 fabrics along which secondary precipitations are found. (c) Core sample is from KBH-6 at 805.75 m depth exhibiting the same S1–2 fabrics within granite gneiss. Interconnected fractures mostly follow the S1-2 fabrics orientation. These fractures are filled up with secondary siliceous precipitations. (d) Core sample from KBH-7 at 1346.25 m depth representing interlayerings of granite and quartzofeldspathic gneisses with prominent S1–2 fabrics. (e) Core sample showing sharp contact between Deccan basalt and the underlying granite at 932.5 m depth in borehole KBH-1. Plagioclase laths are prominent within Deccan basalt in the upper portion, while clasts of quartzofeldspathic aggregates are entrapped within the weathered surficial part of granite in the lower part of the sample. (f) Core sample collected from 936 m depth in borehole KBH-1 showing anastomosing fractures and ferruginous secondary precipitations. (g) Core sample collected from 1159 m depth in borehole KBH-1 showing a fault surface containing downdip slickenlines on a slickenside. (h) Core sample showing the sharp contact between Deccan basalt and underlying granitic basement rock at 499.70 m depth in borehole KBH-5. Plagioclase laths are present within Deccan basalt in the upper part of the sample, whereas large quartzofeldspathic clasts entrapped within fine matrix are present in the lower part. (i) Core sample from 500.50 m depth in borehole KBH-5 showing prominent downdip slickenlines on a well-developed vertical fault surface. (j) Core sample of sheared granite-gneiss from 500.85 m depth in borehole KBH-5, showing lensoid clasts of host rock entrapped within fine-grained gouge matrix of same material within a fractured zone. (k) Core sample from 732.00 m depth in borehole KBH-5 showing a near-vertical master fault with associated tensile fractures. A narrow fault zone with brecciated angular fragments of host rock is seen. Angular clasts of host rock are entrapped within fine-grained gouge matrix. The dark black portions along these fractures represent extremely fine grained, pseudotachylite melts injected from the master fault into the associated tensile fractures. (l) Core sample showing the sharp contact between Deccan basalt and the underlying granitic basement rock at 775.04 m depth in borehole KBH-6. The basement rock below this contact is highly fractured and shattered. (m) Core sample from 897.60 m depth in borehole KBH-6 showing a thin layer of pseudotachylite along the fault surface. Downdip slickenlines are prominently seen on this fault surface. (n) Core sample of a fault surface at 948.50 m depth in borehole KBH-6 showing a slickenside with prominently developed slickenlines that formed due to friction during movement. (o) Core sample showing highly fractured and brecciated granite rock at 1044.75 m depth in borehole KBH-6. Fragmented clasts of host rock are entrapped within fine matrix of gouge material. Anastomosing fracture network filled with greenish silicification from chanellized water is also observed. (p) Core sample showing the direct contact of Deccan basalt and the underlying granitic basement rock at 1251.23 m depth in borehole KBH-7. Plagioclase laths are seen in the first basaltic flow. (q) Core sample from 1253.72 m depth in borehole KBH-7 showing fractured quartzofeldspathic layers within granite gneiss. Fractures are filled up with ferruginous precipitations. The mafic layers are mostly weathered and altered in this section. (r) Core sample from 1476.70 m depth in borehole KBH-7 showing a mafic dyke (amphibolite) body in a near-vertical orientation and having direct contact with the basement rock.

Full size image

This scientific drilling has also established that the ~65 Ma Deccan flood basalt lies directly over a Precambrian granite-gneiss basement. U–Pb dating of zircons from a granodiorite sample from 999 m depth representing the TTG composition and a monzogranite sample from 958 m depth in the borehole KBH-1 yield ages of 2710±63 Ma and 2700±49 Ma, respectively (Rao et al. 2017). Hf isotopic data indicate their emplacement as crust building juvenile magma at ~2700 Ma.

3 Faults and fractures in the basement

Seismic slip along a fault surface causes brittle deformations and highly fractured rocks. The fracture density generally decreases with distance away from the fault core or principal slip zone (Sibson 1986; Paschier and Trouw 2005). Fault core or principal slip zone is frequently surrounded by a highly fractured fault damage zone. Due to progressive seismic activities, fractures generally extend along length and width based on the energy released during an earthquake. These fractures are commonly restricted within a particular zone, identified here as fault zone.

Brittle deformation features of varying intensity are well preserved in the basement granitoids of the Koyna region. To identify the fault zones in the core and provide a better representation of the intensity of brittle deformation, we propose a scale of deformation termed as Sample Quality Designation (SQD). The SQD classification is primarily based on the length of available core samples, which is shown in figure 4. Cores are classified into four categories: Category-A for core length >50 cm, Category-B for 25–49 cm, Category-C for 5–24 cm and Category-D for <5 cm. Physical observations on cores suggest a gradual increase in fracture density from SQD A to SQD D. Core samples from SQD A and B are highly intact and show comparatively lower fracture density than SQD C and D. SQD C and D are considered as highly fractured zones and may be associated with fault damage zones. The SQD generally agrees with the Rock Quality Designation (RQD) (Misra et al. 2017) and therefore provides additional constraints to delineate the extension of fault zones in the Koyna seismogenic zone. The changes in SQD of the basement granitoids with depth are indicated along with the lithologs of boreholes KBH-1, KBH-5, KBH-6 and KBH-7 (figure 2). Thus, SQD helps to classify core samples based on the intensity of brittle deformation and fracture density.

Figure 4
figure 4

Classification of core samples of basement granitoids by sample quality designation (SQD), an index of deformation based on sample length. (a) Cylindrical cores from KBH-7 showing interlayerings of granite, granite gneiss and migmatitic gneiss with SQD category A. (b) Core samples from KBH-5 representing sheared granite gneiss with SQD category B. (c) Core samples from KBH-6 demonstrating fractured granite and granite gneiss with SQD category C. (d) Core samples of KBH-6 representing highly fractured, brecciated and shattered granite and granite gneiss with SQD category D. Physical studies on cores indicate the SQD of categories C and D are from highly fractured and fault zone related rocks at different depths.

Full size image

The depth sections in the basement granitods with SQD C and D in the KBH-1, KBH-5, KBH-6 and KBH-7 cores correspond to zones of intensive faulting and fracturing (figure 2). The fractures in all the four boreholes show dominant dip angles in a narrow range, 30°–45° with the vertical borehole axis. Intensity of fracturing varies both within individual boreholes as well as between different boreholes. Both shear and extensional fractures are preserved in these core sections. Fractures are characterized by displacement of pre-existing features and secondary precipitations. Most fractures are fissures (Mode-I) and/or shear fractures (Mode-II). Fractures are interconnected and form an anastomosing network, which may facilitate fluid flow at depth. Secondary mineralization, silicification, cementation and reddish-brown iron staining are commonly found within SQD C and D zones (figures 5i and 6a, b).

Figure 5
figure 5figure 5

Photomicrographs of basement granitoids representing mineralogy and deformation signatures of both ductile and brittle events in the Koyna seismogenic zone. (a) Granite gneiss showing biotite (Bt) defined gneissosity (S1) under microscope. (b) Garnet (Grt)–biotite (Bt)–muscovite (Ms) defined fabric within granite-gneiss representing both S1 and S2. (c) Alternate layerings of recrystallized quartz (Qz)–feldspars (Fsp) rich felsic and pyroxene (Py)–biotite (Bt)–epidote (Ep) rich mafic layers representing gneissosity of S1 within granite gneiss. (d) High D2 strained rock showing kinked and recrystallized muscovite (Ms) along the S2 fabric. (e) Chlorite defined fabric within granite schist indicating S2 fabric that formed with dynamically recrystallised quartz grains. (f) Quartz (Qz) ribbon within high D2 strained rock surrounded by dynamically recrystallised quartz grains. The lower dark portion indicates fracture network that is overprinted on D2 structures and these fractures are filled up with ferruginous material. (g) Cataclasite of granite rock where clasts of quartzofeldspathic aggregates are embedded within fine-grained quartz-rich matrix. (h) Narrow fault zone within granite, which developed during D3 event. This fractured zone is filled up with gouge material. (i) Fault breccia observed within granite wherein angular fragments of quartz and plagioclase feldspars (Pl) are embedded within matrix of gouge and siliceous precipitations. (j) Set of trans-granular fractures (Fr) developed during D3 deformation. (k) Narrow band of fault gouge transecting coarse-grained granite following a micro-fault zone. (l) Photomicrograph of a pseudotachylite vein that intruded along a tensile fracture within the basement of granite gneiss at 732.50 m depth in borehole KBH-5.

Full size image
Figure 6
figure 6

SEM-BSE images of fault rocks and pseudotachylite veins sampled from basement granites and EDS analyses to characterize the fault zone materials. (a) SEM-BSE image of a fault breccia representing a domain similar to figure 5(i). Angular fragments of quartzofeldspathic material are entrapped within fine siliceous and ferrugenious matrix. The section represents 1354.15 m depth in borehole KBH-7. (b) SEM-BSE image of a fault zone rock at 511.04 m depth in KBH-5, representing repeated brittle fracturing of rocks along which siliceous and ferruginous precipitations occurred. (c) SEM-BSE image of a pseudotachylite vein (representing a domain similar to figure 5l), sharply transecting the gneissosity of the country rock at a high angle at 732.50 m depth in KBH-5. (d) SEM-BSE image of the same pseudotachylite vein under very high magnification, revealing fibrous nature of the pseudotachylite material that formed due to sudden cooling of melt produced during seismic activity. (e) Electron image of fault rock collected from the similar domain of figure 6(a) showing locations of spectra 23–32 at which EDS analysis has been done. Dark coloured areas represent the clasts and light coloured areas represent cemented gouge/matrix. The data are given in table 2. (f) Electron image of the same pseudotachylite vein as shown in figure 6(c). Locations of spectra 1–6, at which EDS analysis has been done, are shown and the data are given in table 3.

Full size image

4 Mineralogy, microstructures and deformation sequence

The basement granitoids are dominantly composed of granites, granite-gneiss and migmatitic gneisses. Structural investigations of these cores reveal both the ductile and brittle deformation features, which are represented here as D1, D2 and D3. The D1, D2 and D3 deformation events are characterized on the basis of associated mineralogy and microstructures, as discussed below.

4.1 Mineralogy and microstructures of D1

Alternate segregated layerings of felsic and mafic minerals are identified on mesoscopic to microscopic scales. Felsic components are dominantly composed of quartz, plagioclase and orthoclase feldspars (majority are microcline), whereas the mafic components are mostly composed of hornblende, biotite, clinopyroxene, occasionally muscovite and intermittently garnet (figure 5a–c). Gneissosity is strongly defined by alternate bandings of quartz + plagioclase feldspar + orthoclase feldspar, and clinopyroxene + hornblende + biotite rich minerals. These are represented as alternate layerings and are labelled in figure 5(a–f), as S1. Annealed mosaic textures of quartz and feldspar indicate static recrystallization (Passchier and Trouw 2005). Unstrained feldspars and low strained quartz are present following the S1 (figure 5a). Here, S1 has been identified as the first developed structural fabric due to preferred alignment of metamorphic minerals. No pre-existing data is available to correlate with the metamorphic event. The first fabrics (S1) were overprinted and transposed by the later high strain S2 fabrics (figure 5d–f).

4.2 Mineralogy and microstructures of D2

The D2 represents strong shearing event and the related planar fabric is termed as S2. Chlorite, epidote and recrystallized muscovite are aligned along this fabric. Microstructures of quartz show bulging and sub-grain formation within and along grain margins, indicating dynamic recrystallization during this deformation (Passchier and Trouw 2005). Significantly, quartz and feldspar represent variable deformation signatures. In most cases, quartz grains are dynamically recrystallized to form serrated grain boundaries and fine-grained groundmass. Occasionally, feldspars are present as ‘augen’ shaped clasts surrounded by fine quartz groundmass and form core-mantle structure (Passchier and Trouw 2005). Basement granitoids in KBH-5 are intensely sheared and at places form schists and mylonites. Schistosity within granitoids generally forms due to preferred alignment of micaceous minerals which were developed during metamorphic reactions. Occurrences of kinked geometry within recrystallized muscovite and stabilized chlorite following S2 fabrics are strongly sheared during this event (figure 5d, e). The shearing is responsible for strong alignment of stretched, boudinaged, quartz ribbons following the S2 (figure 5f). Structurally the S2 transposes all pre-existing fabrics (i.e., S1) following the identical or acute angular orientation. The textural differences between these two different sets of fabrics indicate a deformation gap in between S1 and S2. Therefore, two separate events of deformations, D1 and D2, are inferred.

4.3 Mineralogy and microstructures of D3

Significant evidences of brittle deformation such as fault breccia, fault gouge, anastomosing fracture network, pulverized rocks, secondary precipitations along tensile fractures, pseudotachylites and slickenlines on slickensides are observed in the basement granitiods of the seismogenic zone (figure 3e–r). Fault breccias show angular clasts of parent rocks, which are embedded within fine-grained gouge matrix (figure 5i, k). Sets of fractures are present within fault zone rocks at several depths (figure 5h, k). Narrow band of fault gouge material is observed in zones of intense shearing (figure 5g, j, k). The presence of fault breccias and pseudotachylite veins are the direct manifestations of brittle deformation (Sibson 1986; Lin 1996, 2008; Janssen et al. 2010). Fault rocks occur with friable gouge material within which clasts of quartzofeldspathic minerals are embedded (figure 3j, k, o, q). Incidences of slickenlines with slickensides are the direct indication of frictional sliding along adjoining fault surfaces (Sibson 1975; Lin 2008). Most of these slickensides are unaltered and show shining lustre within fault zone rocks (figure 3g, i, m, n). Reddish iron staining or ferruginous precipitations are commonly found within anostomosing fractured network (figure 3f, j, o, q). In most cases, e.g., figure 5(l), occurrences of dark black coloured, fine-grained, compact material with tapering ends are present as injected veins from the master fault (figure 3k). These are pseudotachylites that were likely formed due to frictional heating along adjoining master faults during previous earthquakes and may contain a wealth of information about the fault dynamics and earthquake processes (Lin 2008, 2011). All these evidences of brittle deformation are overprinted on the pre-existing textural and structural features S1 and S2. Therefore, a separate deformation event D3 has been identified. A detailed account of these features is presented in the subsequent sections.

5 SEM and EDS studies on fault zone rocks

Detailed studies on fault zone rocks have been performed using Focused Ion Beam – Scanning Electron Microscope (FIB-SEM) and Energy Dispersive X-ray fluorescence system (EDS). Fault breccia (figure 5i) and a pseudotachylite vein (figure 5l) observed in the granitic rocks were selected for detailed analysis to study the textural and elemental composition of fault zone materials.

The Back Scattered Electron (BSE) images obtained from SEM are shown in figure 6(a–f). Spectra were focused on selected points of fault breccia and the pseudotachylite vein, which are shown in electron images in figure 6(e, f). Fault breccia shows angular fragments of parent rock entrapped within a siliceous and ferruginous matrix (figures 5i, j, k, 6a). The host cataclasite is transected by a narrow brecciated zone rich in ferruginous matrix (figure 6b). Both the cataclasite and the narrow brecciated zone are transected by transgranular fractures. The fault breccia is characterized by dense anastomosing fracture network with ferruginous precipitation. The BSE image of figure 6(c) corresponds to the photomicrograph of figure 5(l), which represents a pseudotachylite vein at a depth of 732.55 m in borehole KBH-5. Glassy texture of this pseudotachylite vein material is shown in figure 6(d), revealing both the amorphous and fibrous textures. Amorphous/fibrous textures of pseudotachylites are identified by both the polarized and electron microscopes. These findings may be significant in further detailed studies to characterize seismic slip (Lin 2008; Lin et al. 2015; Lin and Nishiwaki 2019).

The BSE images of the fault breccia and pseudotachylite vein are shown in figure 6(e and f). Spectra were focused at several points on these two images and analysed by using X-ray diffraction in FIB-SEM. Detailed elemental percentage at these points are presented in tables 2 and 3. The dataset represents a significant difference in the composition between clasts and matrix within the fault breccia, especially in the fault gouge material. The elemental studies on fault breccias also reveal compositional differences between host rock and secondary precipitations. Angular fragmented clasts of quartzofeldspathic aggregates are entrapped in the fine-grained matrix. Fine-grained matrices are formed by chemical alterations and secondary precipitations within the fault zone. Fragmented clasts present within fault zone materials (fault breccias and gouge) have different elemental wt.% when compared with the matrix (table 2). Secondary precipitations present in the fractures of fault breccia are also shown in terms of elemental wt.%. The clasts have wt.% ranges of O: 44.58–53.09, Al: 0.05–10.51, Si: 30.42–46.46, Ca: 0.04–0.61, Fe: 0.19–0.34, Na: 0.00–9.07, and K: 0.00–12.84. In contrast, the fine-grained matrix has different elemental wt.% ranges of O: 44.42–45.20, Al: 12.36–13.65, Si: 17.85–18.26, Ca: 15.12–15.46, Fe: 7.74–9.58, Na: 0.00–0.11, and K: 0.00–0.33. Hence, the matrix is rich in Ca and Fe when compared with the fragmented clasts. Enrichment of these elements within fine-grained matrix groundmass could be related with secondary mineralization and ferruginous precipitations along fractures in the brecciated rocks (Lin 2008; Lin et al. 2015; Moore and Lockner 2013; Lin and Nishiwaki 2019).

Compositional and textural studies on pseudotachylites can provide significant information about earthquakes and are therefore often known as ‘earthquake fossils’ (Lin 2008). Physically, pseudotachylite veins are dark black in colour with fine-grained appearances along the fault surfaces (figure 3k, m). These veins are associated with fault breccias and fault gouge, prominent slickenlines on slickensides and show shining lustre. Elemental wt.% of the pseudotachylite vein is represented by O: 45.58–49.54, Na: 0.00–0.44, Mg: 0.35–14.34, Al: 8.37–10.37, Si: 16.24–30.48, K: 0.04–12.47, Ca: 0.00–0.42, and Fe: 0.37–13.26. The wt.% of Fe and Ca represent the composition of unaltered pseudotachylite material; no secondary precipitation or alteration are observed within this vein. The wt.% of Al, Si and O are higher than other elements, and the wt.% of Fe, Mg and K show a wide range (table 3). Pseudotachylites are commonly formed due to sudden cooling of a melt. The melt is generally produced because of frictional heating along faults during earthquakes (Sibson 1973, 1975; Rice 2006; Lin 2008). As a result, composition of pseudotachylites are directly dependant on the mineralogy of the host rocks adjacent to faults (Lin 2008; Jiang et al. 2015). Here, pseudotachylites show amorphous and fibrous textures (figure 6d), and characterized by glassy material that is rich in Si and O, while the presence of Al, Fe, Mg and K indicate presence of feldspars and other mafic contents within the host granites and gneisses.

6 Discussion

Scientific deep drilling in the Koyna seismogenic zone provides a rare opportunity to sample and study the granitic basement rock below the Deccan Traps. The recurrent seismic activity in the region over the past 55 years is concentrated within the granitic basement rocks buried under a kilometer thickness of flood basalts and extends to a depth of ~10 km from the surface (Gupta et al. 2015). The core samples of basement granitoids down to 1.5 km depth, recovered from boreholes drilled in close proximity to the epicentral region, therefore, make a good case to study the deformation signatures in general and co-seismic deformation and related structural features in particular. Petrography and microstructural studies carried out in the present study show that structural features and deformation signatures observed in the cores of basement granitoids preserve significant information about brittle deformation.

Basement core specimens from boreholes, KBH-1, KBH-5, KBH-6 and KBH-7 provide compelling evidences to infer the sequence of deformation in the region. The basement rock is a complex of granite, granite-gneiss and migmatitic gneiss. Strongly developed segregation bandings are observed within the gneissic domains. The mineral assemblage of pyroxene + hornblende + biotite ± garnet ± muscovite following the S1 fabric indicates amphibolite facies condition, which prevailed during D1 event, whereas the assemblage chlorite + epidote + recrystallized muscovite and dynamically recrystallized quartz minerals within the S2 fabric suggest that greenschist facies prevailed during D2 event. The occurrence of mylonites, S-C fabrics within muscovite-rich granite-schists, recrystallized quartz-rich groundmass, serrated quartz grain boundaries indicate high strain condition during D2 deformation. The pre-existing D1 and D2 features are overprinted by cataclasite, fault breccias and fault gouge. These latter brittle deformation features are identified as D3 event. The information available so far is inadequate to determine whether the D3 deformation event could be the manifestation of the recurrent seismicity over the past 55 years in the area. Association of pseudotachylites with slickenlines on slickensides preserved within the fault zones support frictional deformation. Cores recovered from all the four boreholes show clear evidences of faulting, fracturing and superposition of breccias with transgranular fractures and fault gouge, which are indicative of multiple episodes of brittle deformation. Cataclasites with multiple deformations are also found locally away from the fault zones. Frequent occurrences of microcracks, fault-breccias and cataclasites in drill core samples confirm that the faults were reactivated in a brittle regime (Ismat and Mitra 2001; Passchier and Trouw 2005; Matin and Misra 2009).

Basement rocks record evidence of intensive fractures and fissures which were developed during D3 deformation. In the majority of the boreholes, fractures generally follow the pre-existing weak planes such as foliations although other orientations are also observed. The fault and fracture zones (SQD category C and D) are dominantly characterized by intense shattering and brecciation. Occurences of slickenlines with slickensides and pseudotachylites are mostly associated with intense shattering and brecciation. Well-developed fractures in the basement granitoids show prominent striations with shining lustre on the slickensides (figure 3g, i, j, m, n). Very little or no geochemical alterations are observed on such fault and fracture surfaces. In fact, most of these faults and fractures are filled up with secondary precipitation, which provide evidences for the presence of water at depth.

In this study, the angles of fractures and foliations measured with respect to the vertical borehole axis are plotted in figure 7. Fractures and foliations in the KBH-1 and KBH-5 borehole cores show nearly identical orientations (figure 7a–d). In case of borehole KBH-6, only a few fractures follow the foliation, whereas most of the fractures have different angles which are near vertical (figure 7e, f). Intensive shattering and brecciation of cores in borehole KBH-6 are directly related with near-vertical faults/fractures. Notably, pseudotachylite veins are also associated with these sub-vertical faults and fractures (figure 3k, m, n). Additionally, these near-vertical faults and fractures are largely unaltered, containing lustrous slickensides with prominent downdip slickenlines. In borehole KBH-7, fracture angles show a wide range of 0°–90°. Here, they have identical orientations with the foliations, however, there are other orientations which represent localized faulting and fracturing activities (figure 7g, h). Importantly, the borehole KBH-7 is located in the proximity of the NNE–SSW trending and steeply dipping Donichawadi fault zone. The wide range of fracture angles in borehole KBH-7 may be associated with the vertical movements in the Donichawadi fault zone. Filling of fractures and fissures with secondary precipitations suggests fluid-filled tensile fractures and fissures which are closely associated with this near-vertical fault zone. Occurences of prominent downdip slickenlines on shining slickensides with thin layers of psedutachylites are the strong evidences of frictional shearing which are mostly unaltered. In conclusion, it can be inferred that the D3 deformation zone present within the basement rocks is an ongoing process of vertical movement within an extensional fault zone. The frictioned surfaces preserve the evidences of elastico-frictional deformation, and, tensile openings are filled up with secondary precipitations within the same fault zone.

Figure 7
figure 7

Rose diagrams representing the distribution of angles of foliations (N=794) and fractures (N=1241) measured with respect to the vertical borehole axis (BA) in KBH-1, KBH-5, KBH-6 and KBH-7. Note that in most cases, natural fractures and fabrics show similar range of orientations except in borehole KBH-6 in which near-vertical fractures dominate. These near-vertical fractures likely formed due to faulting of rocks in close association with a near-vertical fault.

Full size image

A recent study by Goswami et al. (2020) has determined fracture dip amounts and orientations from borehole image logs acquired in a 3014-m deep research borehole, called the Koyna Pilot Borehole KFD1 drilled within the Koyna seismogenic zone (figure 1). Downhole geophysical measurements confirmed that the borehole KFD1 had penetrated fault damage zones at multiple depth intervals in the granitic basement below 2100 m (Goswami et al. 2019). The borehole KFD1, located in the proximity of the exploratory borehole KBH-7, passed through 1247 m thick Deccan Trap and continued ~1767 m into the underlying granitic basement rocks (Roy 2017). The borehole images were oriented with respect to the true north and as many as 2226 shear fractures were picked manually in the 1767 m long basement section. The dip amount of the fractures shows a wide range from sub-horizontal (≤10°) to subvertical (≥80°) and the majority of the fracture strike azimuths show a range from NNE to NNW (figure 8) (Goswami et al. 2020). Significantly, the trend of the fractures obtained from borehole images is consistent with the surface fissures mapped in the Donichawadi fissure zone by GSI (1968) and in the surrounding areas by Misra et al. (2017). The study carried out by Goswami et al. (2020) provided a sound basis for correlating subsurface data from basement granitoids with the surface manifestations in the Koyna region.

Figure 8
figure 8

Rose diagrams showing the distributions of dip amounts and strike azimuths of shear fractures identified from borehole images in the 1767 m long granitic basement section in Koyna Pilot Borehole KFD1. The borehole images are oriented with respect to the true north and 2226 shear fractures between 1247 and 3014 m depth are studied.

Full size image

A structural correlation can be established between the boreholes KFD-1 and KBH-7 taking advantage of their proximity. The basalt thickness in the two boreholes are quite similar, 1247 m and 1251 m, respectively. Study of KBH-7 core samples shows that majority of the fractures dip in the range 30°–80° with a mean of 48° (figure 7g). Thus, dip amounts of the inclined fractures observed in borehole KFD1 are consistent with the fracture angles measured in the KBH-7 basement core. However, the unoriented KBH-7 cores do not allow determination of strike azimuths. Since the boreholes KFD1 and KBH-7 are located close to each other and the dip amounts of the fractures are also consistent; therefore we can infer that the subsurface fracture strike azimuths at the KBH-7 site may be similar to those at the KFD1 borehole site. Additionally, both the boreholes are located in the vicinity of the Donichawadi fissure zone, which is a surface manifestation of the 1967 M 6.3 Koyna earthquake. The en-echelon fractures, oblique or diagonal tensile cracks and mole cracks associated with the Donichawadi fissure zone are oriented in the range N10oW to N25oE, which are very consistent with the strike azimuths of the shear fractures observed in the borehole KFD1. A study of structural orientations of these fissures shows that the principal orientation of the fissure zone is 25°/89°-E (Misra et al. 2017). A soil-gas helium study across the Donichawadi fissure zone (Gupta et al. 1999) and later online gas monitoring study during the drilling of the Koyna pilot borehole KFD1 (Podugu et al. 2019) have shown similar high levels of helium concentration (up to 12.8 ppmv), which confirm that the fissure is the surface manifestation of a subsurface fault/fracture zone and that the fault zone has not healed during the past 50 years. The consistency of the strike azimuths of the subsurface fault(s)/fractures with the surface fissures and the conductive nature of fault zone suggest that the interconnected fractures provide pathways for fluids to enter the fault zone(s). This could trigger earthquakes either by increasing fluid pressure in the fault zone or by decreasing the frictional strength of the fault gouge in the region.

During earthquake occurrence, up to 90% of the released energy is used to overcome the frictional strength of the fault (Kanamori and Brodsky 2004; Scholz 2002 and references therein; Rice 2006). Friction generates high temperature at the interface between the two sliding blocks causing the rock to melt and get injected into the fractures/veins adjacent to the principal slip zone. Cooling of such melts form pseudotachylites, which therefore provides information about seismic activity and energy budget of the earthquake (Sibson 1975, 1986; Lin 1996, 2008, 2011; Cowan 1999; Nielsen et al. 2008; Beeler et al. 2016; Aubry et al. 2018). In the present study, pseudotachylites are observed at multiple depths along veins/fractures. The FIB-SEM and EDS analyses carried out on fault rocks and pseudotachylites yield significantly different elemental composition. The clasts present in the brecciated section of the fault zone are mostly quartz and feldspars embedded within a fine-grained matrix. Compositional analysis of clasts, matrix and pseudotachylites show substantial differences in wt.% of Fe, Ca and Al between host rock and the secondary products. Higher percentage of Fe, Ca and Al are observed within the matrix when compared with the host rock. On the other hand, secondary products are mostly formed by siliceous and ferruginous precipitations in a fault zone (Sibson 1975, 1986; Lin 2008, 2011; Warr et al. 2014; Lin and Nishiwaki 2019). Presence of calcite crystals and sericitized feldspars along fractures are responsible for high wt.% of Ca and Al within the fault zone (table 2). The EDS studies on pseudotachylite veins show a large variation of elemental wt.% for Fe, Mg and K, whereas the wt.% of Al remains high throughout the domain, pointing to its source from feldspar rich country rock adjacent to the pseudotachylite veins. Amorphous/fibrous textures and compositional variations of pseudotachylites are related with the mechanism of shearing during friction (Lin 2008, 2011; Warr et al. 2014; Lin and Nishiwaki 2019). This compositional variation is similar to the mineralogy of the host rock. Amorphous/fibrous textures and compositional variations of pseudotachylites are related to the total increase in temperature during friction (Lin 2008, 2011; Warr et al. 2014; Lin and Nishiwaki 2019). Age dating of the pseudotachylites could shed more light on their possible association with seismic activity in the region.

7 Conclusions

  1. (1)

    The basement rock in the Koyna seismogenic zone comprises granite and granite-gneiss inter-layered with variable proportions of migmatitic gneiss, typically representative of the Peninsular Gneiss of the Dharwar Craton in south India. The metamorphic mineral assemblages of these basement granitoids indicate amphibolite to greenschist facies metamorphism.

  2. (2)

    The basement rocks record three successive deformation events, D1, D2 and D3. While D1 and D2 are associated with ductile deformation, D3 is associated with brittle or frictional deformation and could be a result of past seismic activity in the region.

  3. (3)

    The study on drill cores down to 1.5 km depth in the area has brought out the relationship between the fractures and the foliation in the fault zone rocks, suggesting variable angular relation both within a borehole column as well as between different boreholes. Fractures do not necessarily follow pre-existing weak planes during their development. Also, fault brecciation and shattering of rocks are observed along near-vertical fractures. Vertical fractures differ significantly from the mean angle of foliations observed in the basement granitoids, which indicates that local stress kinematics play a major role in their development.

  4. (4)

    Presence of prominent dip-slip slickenlines on near-vertical slickensides indicates vertical movement of a fault along its surface(s). Most of these fractures are tensile in nature, which implies an extensional faulting mechanism at depth.

  5. (5)

    Occurrences of fault breccia and fault gouge along with pseudotachylites, slickenlines and slickensides indicate frictional deformation within the fault zone(s). The presence of different sets of fractures and their dispositions point to multiple episodes of brittle deformations within the fault zone(s). Most of these fractures are filled with ferruginous and/or siliceous precipitation that indicate the presence of water at depth.

  6. (6)

    FIB-SEM and EDS studies on fault gouge and pseudotachylite veins confirm secondary precipitation following the anastomosing fracture networks. The secondary minerals are unusually rich in Fe and Ca, consistent with the host rock. Composition of pseudotachylite shows higher wt.% of O, Al and Si than the composition of parent rock-forming minerals, although at places, pseudotachylite has higher Fe, Mg and K content.

Table 2 EDS analysis of a fault breccia recovered from KBH-7 basement granite rock showing wt.% of elements. The spectrum points shown in electron image of figure 6(e) are studied under FIB-SEM.
Full size table
Table 3 EDS analysis on a pseudotachylite vein recovered from KBH-5 basement granite-gneiss showing wt.% of elements. The spectrum points shown in electron image of figure 6(f) were studied under FIB-SEM.
Full size table