ABSTRACT The electrical resistivity structure of the crustal and uppermost mantle beneath the Dharwar craton in the southern India are investigated by magnetotelluric method. In the present study, a northwest-southeast oriented 220 km long profile of 18 stations with a station spacing of ~10-15 km is used. The profile extends from Dharwar cratonic nucleus in the north to Billigiri Rangan charnockitic massifs in the south. Time series data are processed to get the apparent resistivity and phase in terms of the transfer function. The dominant geoelectric strike direction (N75°E) is measured in a period range of 0.01-1,000s. The data are rotated to N75°E strike direction. Two-dimensional modelling is carried out by using the non-linear conjugate gradient scheme (NLCG) for both apparent resistivity and phase. We also taken into consideration the results of Heat flow, gravity and magnetotelluric studies in the same region. Our modelling results show one conductor in the northern side of the profile and two distinct prominent conductors in the southern part of the profile. The mid-lower crust in the southern part of the profile shows the less resistive (<300 ohm-m) structure in the depth range of 20-50 km related with the Chitradurga shear zone and Billigiri Rangan charnockite massif.These zones can be interpreted as CO2 flushed terranes. Regional scale carbonation occurred in Late Archaean is associated with Chitradurga shear zone and in Late Proterozoic is associated with Salem-Attur shear zone. The CO2 rich fluids derived during that time might have exhausted in dehydration reactions. However, the later events such as the Indian plate passing over several hotspots and the metasomatized fluids associated with the Cretaceous-Tertiary magmatism in the region is the reason for observed low resistivity near Billigiri Rangan massif and surrounding regions in the south. Keywords: Magnetotellurics; Dharwar craton; Chitradurga shear zone; Billigiri Rangan charnockitic massif, fluids, conductivity 1. Introduction The Dharwar craton in the southern Indian shield region is one of the most significant cratonic regions of the world with exposes of the oldest lower crustal rocks found on the Earth. Broadly the southern Indian shield region can be divided into granite-greenstone terrain of Dharwar craton towards north and the region of granulite facies rocks of Southern Granulite Terrain (SGT) towards south. Prograde metamorphic transition took place in between along an approximately 60 km wide zone following an east-west trend (Naqvi and Rogers, 1987 and Newton, 1990 and references therein). Large scale charnockitic massifs (Coorg, Nilgiri, Billigiri Rangan and Shevroy hill massifs) that separate the amphibolite and granulite facies and the shear zones e.g., Chitradurga, Moyar, Bhavani, Salem-Attur, Palghat-Cauvery as well as several deep seated faults, some of which even reaches the Moho (Drury et al., 1984) are of interest to study in order to understand the deep crustal structure in this region. Observation of regional scale carbonated alteration of the crust in, near and along the shear zones and dense CO2 rich fluids in the deep seated crust would make magnetotelluric method a useful tool. The other interesting problem is to evaluate the deep crustal electrical resistivity properties of the region. It should be apt to note that the lower crust even in the cratons is not homogeneous (Jones, 1981). The magnetotelluric method is a way of determining the electrical conductivity distribution of the subsurface from measurements of natural transient electric and magnetic fields on the surface (Tikhonov, 1950; Cagniard, 1953). The magnetotelluric technique is widely used to determine the electrical resistivity properties of the subsurface (Jones, 1992; Chave and Jones, 2012 for a review). Magnetotelluric studies are rigorously employed to image the shallow and deep crustal structure (Azeez et al., 2015; Bologna et al., 2017; Brasse et al., 2009; Jones, 1992; Naganjaneyulu and Santosh, 2010a, b, 2011; Nagarjuna et al., 2017; Pina-Varas et al., 2014; Sarafian et al., 2018) and to delineate faults and shear zones (Adetunji et al., 2015; Naganjaneyulu et al., 2010a, b; Rao et al., 2014; Unsworth et al., 2000; Wu et al., 2002). A magnetotelluric study was carried out along a 220 km long profile from Hassan to Chamarajanagar across the two branches of Chitradurga shear zone (marked as ChSz L and ChSz R, Fig. 1) to study the deep crustal electrical resistivity characteristics. A synthesis of other geophysical and geological study results was also used in explaining the resistivity structure of the study area.2. Geological framework of the Dharwar craton In the age range of 3.6-2.6 Ga, the Indian shield is made up of a variety of Precambrian metamorphic terrains that reveal low to high grade crystalline rocks. A craton consists of old continental crust that was formed during Archaean and has attained stability by 2500 Ma. Some of the cratons expose large areas of greenstone-granite terranes which enable us to study the process of continental crust formation during the Archaean (Viswanatha and Ramakrishnan, 1981; Allen et al., 1986). The Dharwar craton in southern India conserve Archaean cratonic nuclei with a basement collection of linear schist belts in association with a widespread tonalite-trondhjemite-graniodiorite (TTG) gneiss complex, and intruded by younger granites (Bhaskar Rao and Naqvi, 1978; Swami Nath and Ramakrishnan, 1981; Anil Kumar et al., 1996). The western Dharwar craton is bordered to south by the Pan-African-Pandyan mobile belt (Southern Granulite Terrain) (Fig. 1) and to the north by end-cretaceous Deccan traps (Ray et al., 2008). The Moyar-Bhavani shear zone is mark the boundary between the western Dharwar craton and the SGT. The metamorphic grade increases from green schist facies in the north to granulite facies in the south (Halls et al., 2007). In the reconstructions of the initial supercontinent "Ur" is significantly figured out in the Dharwar craton (Rogers, 1996; Rogers and Santosh, 2003, 2004). The northern part is additionally sub-divided into the eastern Dharwar craton (EDC) and the western Dharwar craton (WDC). The eastern and western blocks of the Dharwar craton is separated by the Chitradurga shear zone (Ramakrishnan, 2003) and then it extends via the western margin of the Billigiri Rangan hills. The WDC begin to develop at about 3.5 Ga (Bhaskar Rao et al., 1992; Rogers, 1996). Other argument is that the closepet granite batholith is the boundary between the WDC and EDC which consists of several plutons that extends Deccan traps in the north to granulite terrane in the south with metamorphism in the transition zone (Bhaskar Rao et al., 1992; Moyen et al., 2001). The oldest (3.36 Ga) crustal nuclei is formed in the southern side of the WDC consists as a green stone belts near Hassan (Taylor et al., 1984; Peucat et al., 1995). The WDC was formed by the Chitradurga schist belt, Bababudan group and Shimoga schist belt. The Neoarchaean dioritic–granitic basement, with linear schist belts are categorized by the EDC (Chadwick et al., 1996, 2000; Nutman et al., 1996). Based on the seismic reflection/refraction study the Dharwar craton is divided into two tectonic blocks with distinct Moho depths vary from 41 km in the west and 34 km in the east (Kaila et al., 1979; Roy Chaudhury and Hargraves 1981).3. Previous geophysical studies of Dharwar craton A brief summary of various geophysical studies carried out in the Dharwar craton are presented here. One of the oldest geophysical study utilizes deep seismic sounding method (Kalia et al., 1979) that investigates the crustal structure along Kavali - Udipi profile. Over 15 distinct blocks were delineated and many of the faults delineated are extending up to Moho. The crustal thickness increases from about 34 km in the east to about 41 km in the west (Kaila et al., 1979). The gravity model along Kavali-Udipi profile (further north of the study region, not shown in Fig. 1) image high density (2.85-2.90 gm/cc) structures near the Chitradurga shear zone (Singh et al., 2004). Less dense structures were identified in the western Dharwar craton as compared to eastern Dharwar craton (Singh et al., 2004). Recent gravity model imaged 40 km thick crust in this region (Niraj et al., 2013). Gupta et al (2003) estimated crustal thickness in the western Dharwar craton is from 40 km to 60 km with an average value of 45 km whereas the crustal thickness in the eastern Dharwar craton is 35 km. More recent studies delineated crustal thickness of 42–46 km, 38–42 km and 48–52 km for northern, central and southern parts respectively in the western Dharwar craton (beneath profile AA' in Fig. 1). The crustal thickness in the eastern Dharwar craton is 34-38 km (profile BB' in Fig. 1). The minimum lower crustal shear wave velocity is ?3.95 km/s. These studies have also delineated anomalously high velocity material in the depth range of 50-100 km. The heat flow values are in the range of 25-40 mW/m2 (Roy et al., 2007). The heat flow at the base of the crust (< 20 mW/m2) and the temperature at the Moho depth are well below 450°C indicating a cold crust (Roy and Rao, 2003). The lithosphere thickness is not well estimated for the region. The value ranges from 80 – 300 km in various studies (Table 1).4. Magnetotelluric Data acquisition and processing MT data were collected at 18 stations in the Dharwar craton region with a station spacing of 10-15 km along a traverse from Hassan in the northwest to Chamarajanagar in the southeast (Fig. 1). The data were collected with wide band magnetotelluric systems in the period range of 0.01–10,000 s. Magnetic field components were measured using induction coil magnetometers and telluric field measurements with the help of Cu–CuSO4 porous pots. The electrode separation area is 100 X 100 m. The data are subjected to robust processing (Chave and Thomson, 2004) and tensor decomposition (McNeice and Jones, 2001) before deriving a 2-D model. For strike evaluation the lengths of the arrows are scaled by standard error in each site in that period band. The strike results are shown in Fig. 2 for five representative period bands for individual stations in the period range 0.01-1000 s. During the various trials, with various period bands, we observed that the shallow and the upper mantle structures on edges of the profile are showing approximate north-south (or east-west considering the 90 ambiguity in determining strike) orientations. The middle bands which roughly correspond to the crustal structure are showing 45° to 85°. The geoelectric strike direction chosen here is 75° based on the reliability in the strike direction with low chi-square errors and then the data is rotated to 75°. The two magnetotelluric responses - transverse electric (TE) mode, in which the electric field is parallel to the geoelectric strike, and the transverse magnetic (TM) mode, in which the electric field is perpendicular to the strike were considered in the inversion. We observe very high apparent resistivity values in the range of about 6,000 ohm-m to 20,000 ohm-m at short periods (Fig. 3). The 2-D inversion is carried out using the non-linear conjugate gradient inversion algorithm of Rodi and Mackie (2001) which provides a minimum structured model required for the observed data. The model mesh consists of 85 horizontal and 179 vertical elements. The apparent resistivity and phase error floors used are 20% and 8% for TE mode and 16% and 7% for TM mode to get the model. Since TM data are usually less sensitive to three dimensional effects, more emphasis is given to TM data over TE data and more weightage is given to phase over resistivity data to avoid static shift affects, if any, following established methods discussed in literature (for example Wannamaker et al., 1984, 1987, 1989; Echternacht et al., 1997; Brasse et al., 2009). In the present study the model roughness value of 10 was chosen as smoothing parameter for the model.The MT model (Fig. 4) shows the upper crust is highly resistive (>10,000 ohm m) and this resistive character extends to a depth of about 13-20 km. The lower crust also shows high resistivity (>10,000 ohm m) in the north side of the profile up to centre of the profile and the resistive features extends to a depth of about 50 km. In the southern side of the profile the mid- lower crust are less resistive (<300 ohm m). Three conductors A, B and C are also present beneath the study region. The conductive nature is also supported by high sensitivity values of more than 0.1 obtained in the linear sensitivity analysis (Fig. 5). The robustness of these features are tested by placing resistors at the place of conductors and by carrying out forward model response calculations. Examples of nonlinear sensitivity analysis results are shown in Fig. 6. Further the model is inverted and the conductive features appear again. The deviations in the observed responses thus confirms that the features representing the model are robust.5. Discussion and interpretation Discussion and interpretation of the resistivity model (shown in Fig. 4) in terms of geometry and subsurface geological conditions is provided here. The highly resistive upper crust (>10,000 ohm m) extends up to 13-20 km which can be correlated with the tonanlite-trondhjemite-granodiorite (TTG) crust. In the northern part even the mid-lower crust is highly resistive. The immediate inference that we get is the resistivity properties are representing Dharwar cratonic nucleus. Before going further on discussing geoelectric structure it would be apt to review the causes for high conductivity anomalies in the crust. By invoking thin graphite films (Glover and Vine, 1992), partial melts (Hermance, 1979; Schilling et al., 1997), presence of fluids (Hyndman and Shearer, 1989) and/or due to a combination of more than one of the above can possibly explain high electrical conductivity in the deep crust. Partial melt as a causative source is ruled out considering low heat flow values (generally less than 50 mW/m2) in this region. This region can be considered as a stable region and hence connectivity between thin graphite films is likely to generate conductive zones. However, we have excluded this option considering the upliftments that the region witnessed and a few earthquakes that occurred in surrounding regions (Coimbatore and near PCSZ, Fig. 1). The instabilities can break the continuity of the graphite films (Jones, 1992; Wannamaker, 2000). Fluids can be other alternative.An examination of available Bouguer gravity maps (Subrahmanyam and Verma, 1982) suggests northern part of the profile has gravity values of -80 to -90 mgal with an exception between stations 1 and 2 where the gravity value ranges from -90 to -100 mgal. The central and the southern parts show gravity values of -80 to -110 m gal. This is an indication of low density structures in the central and southern parts. A less resistive local structure is observed near Hassan in the north of the profile – marked as ‘A’ (Fig.4) – between stations 1 and 2. This ~ 200 ohm-m conductive feature has a top at about 10 km depth and extends up to a depth of 20 km. The conductor ‘A’ is observed near to the margin of greenstone formations. Geoelectric structure on the southern side is showing more conductive in nature from a depth of 10 km to about 50 km. There appear two conductors – B in the upper crust associated with Chitradurga shearzone Right branch; C in the mid-lower crust – and the deep crust is showing less than 300 ohm-m resistivity of the profile. The extension of feature B is limited to upper crustal depths. The extension of the feature C to the south of station 18 on the edge of the profiles is considered as not a well resolved one. Supracrustal rocks becomes less abundant as we move towards south of the study region. The boundary between the amphibolites and granulite facies marked by large scale charnockitic massifs include Coorg, Nilgiri, Billigiri Rangan, Shevroy hill massifs. This boundary preserves the record of metamorphism that occurred in the transition zone (Bhaskar Rao et al., 1992). Several quarries near this boundary display textures indicating formation of gneiss to charnockitic massifs along minor shear zones (Janardhan et al., 1982). Apart from the minor shear zones, the major shear zones that immediately visible near our study region in the south are Chitradurga, Moyar, Bhavani, Salem-Attur and Palghat Cauvery Shear/suture zones (Fig. 1). All along the Salem-Attur shear zone and further west of it in Billigiri Rangan hills the background grey hornblende-biotite tonalite gneiss transformed into brown orthopyroxene charnockites (Buhl, 1987; Khan, 1989). Fluid inclusion studies have shown influx of prograde dehydrating fluids dominantly CO2 along channel ways to produce charnockitic structures and thus the charnockitic massifs can be interpreted by CO2 flushed terranes (Harris et al., 1982; Percival et al., 1992; Wickham et al., 1994 for a review). In the southern Indian shield region and in southern Norway open system orthopyroxene forming reactions are observed (Newton, 1990). Further east of the study region near Salem-Attur shear zone a conductive feature is delineated and interpreted to be caused by fluids (Harinarayana et al., 2003; Naganjaneyulu, 2004). Regional scale carbonated alteration of the crust along the margins and in the shear zones is observed both in southern India as well as in southern Africa (Chadwick et al., 1985; Faure and Harris, 1991; van Schalkwyk and van Reenen, 1992). When we compare the timing of carbonation it is clear that there exists at least two distinct-events. The older event associates with the Chitadurga shear zone in Late Archean (Chadwick et al., 1985) and the younger event in Southern Granulite Terrain and in Sri Lanka during Late Proterozoic (Wickham et al., 1994 and references there in). Thus it is evident that there exist fluids due to carbonation – intermittent, localized high CO2 fluxes associated with shear zones and even 100 km away from shear zones in the study region (see Baratov et al., 1984; Newton 1990b; van Schalkwyk and van Reenen, 1992; Dahlgren et al., 1993; Wickham et al., 1994). Regional carbonate metasomatism is directly linked to CO2 rich magmatism (Wickham et al., 1994). The elevated 87Sr/86Sr ratios in carbonatites (0.705-0.706) indicate the long term existence of enriched mantle material. This enriched lithosphere could have formed during subduction or under thrusting events and may be a potential source of CO2 rich magmas. Dense CO2 rich fluid inclusions in nearby Nilgiri hill massifs suggest that these fluids were carbonic in metamorphism of the charnockites (Srikantappa, 1987). Out of various processes (magmatic, metamorphic, mantle-derived, or crust-derived) such fluids could have resulted from an outgassing mantle hotspot, reaction of hydrous minerals and graphite in uplift and decompression of granulite-facies rocks, release of CO2 from deep crustal fluid inclusions by deformation during a metamorphic episode, sub crustal origin of transporting fluids (Cameron, 1988; Burrows and Spooner, 1987; Kerrich, 1990; Wickham et al., 1994). Fluids cannot be in molten state for such a long period. The Indian plate passed over several hotspots (Kerguelen, Crozet, Marion and Reunion) in the Indian ocean (Chatterjee et al., 2013 for a review). Deep seismic studies in the Dharwar craton imaged a number of deep seated faults (Kaila et al., 1979). So it is possible that the kimberlite magma has broken a possible fault which has formerly controlled the emplacement of kimberlite ~1 Ga before (Chalapathi Rao et al., 2016). This offers an explanation that the conductive features/zones – A, B and C – represent zones of metasomatized by fluids associated with the Cretaceous-Tertiary magmatism in the region.6. Summary The region beneath the Dharwar cratonic nucleus and surrounding charnockitic massifs was investigated by using magnetotelluric data obtained at 18 stations. The upper crust is highly resistive (>10,000 ohm-m) all along the profile. This is correlated with the tonanlite-trondhjemite-granodiorite (TTG) crust. The mid-lower crust in the northern part of the profile shows high resistive nature representing Dharwar cratonic nucleus part. The southern side of the study region is showing conductive nature for mid-lower crust. The charnockitic massifs in the southern side can be interpreted as CO2 flushed terranes. Regional scale carbonation occurred in Late Archaean and Late Proterozoic. The former event is associated with Chitradurga shear zone where as the latter event is associated with the Salem-Attur shear zone. This regional carbonate metasomatism is directly linked to CO2 rich magmatism. The fluids might have resulted from any of these reasons: outgassing mantle hotspot, release of CO2 from deep crustal fluid inclusions by deformation during a metamorphic episode, sub crustal origin of transporting fluids etc., The fluids thus released might not have in molten state. The later events such as Indian plate passing over several hotspots thus creating zones metasomatized by fluids (associated with the Cretaceous-Tertiary magmatism) in the region is the reason for observed low resistivity near Billigiri Rangan massif and surrounding regions in the south.