Integrated geophysical approach for Coal mine re in Jharia coal eld, India


 Jharia coalfield is one of India's largest and most productive coal fields, home of the oldest-burning coal seam fires in the world. This paper highlights the combined study comprising Surface Temperature measurement, Magnetic, and Electrical Resistivity Tomography (ERT) methods for delineation and mapping of coal fires with their depths, extension and propagation direction. Surface temperature mapping allows straight forward demarcation of coal fire areas. Magnetic data provide a map showing the earlier burned, currently burning and unburned patches based on the contrast between magnetic effects of the subsurface and fluctuations with temperatures exceeding or under the Curie temperature. The pyrolysis of coal is one of the main processes in coal fires, which results in very high conductivity. Thus, electrical resistivity tomography (ERT) is also a promising technique for the characterization of a coal fire-affected area. Most of the features of low and high resistivity associated with active fires and voids have been delimited well in the ERT sections that are generated from combine inversion of all arrays. These are also delineated well by the surface temperature mapping, and Magnetic methods. The field photographs, available coal seam and borehole lithology clearly show the suitability of the combined study for characterization of coal seam fire. A model of the coal fire, char formation and void formation resulting from coal seam fire has been established. Finally, a fire propagation model of the study area has been established based on the combined geophysical data that can be effectively used for mine hazard mitigation.


Introduction
Thousands of coal mines res are reported to be burning in at least 22 countries throughout the globe (Stracher, 2010). Coal res are common worldwide problems posing severe hazards to valuable energy sources, environments and human lives. China, United States, India, South Africa (Prakash and Gupta 1998;Stracher 2004;Mishra et al. 2011;Srivardhan et al. 2016), Indonesia, Russia, Poland and Australia  are severely victimized by coal res. The coal res may be caused by illegal mining activities, exothermic reactions, lighting, forest re and anthropogenic activities which can smoulder underground coal for several years and may continue for hundreds of years. The ignition of underground coal res activities is a chaotic process comprising numerous illde ned variable parameters, such as: greenhouse gases, e.g., methane (CH4), carbon dioxide (CO 2 ), (Schloemer 2006;Litschke et al. 2008;Schloemer et al. 2008), toxic gases, e.g., nitrogen (N 2 ), ammonia (NH 3 ), sulphur dioxide (SO 2 ), hydrogen, carbon monoxide (CO) (Kuenzer et al. 2007b), trace elements, e.g., mercury (Hg), selenium (Se), Arsenic (As), (Dai et al., 2012;, and various unsaturated hydrocarbons, sulphur, H2O vapour, asphalt vapour, organic, inorganic chlorides and hydrogen chlorides, etc. (Engle et al. 2011;Xie et al. 2011;Song and Kuenzer 2014;Gürdal et al. 2015;) which threaten the local environment and healthiness of public alive in the surrounding areas.
The delineation of a concealed underground coal re is of utmost importance to establish scienti c and administrative strategies to control and extinguish or rehabilitation of local establishment.
The coal res have been studied by many researchers, mainly using satellite remote sensing data over Jharia coal eld and its surroundings. Bhattacharya et al. (1991)) exploited aerial thermal data for the mapping of coal res and its depth of occurrence, in the Jharia coal eld, using linear heat ow equation. Saraf et al. (1995) and Prakash and Gupta (1998) established the effectiveness of space-borne thermal data (Landsat-5 TM) for the identi cation of coal res in the area. Prakash et al. (1997) carried out a comparative study of surface and subsurface res using Landsat TM. Agarwal et al. (2006) conducted coal mine res in the Jharia coal eld using NOAA/AVHRR data. Chatterjee (2006) established a technique of pixel integrated temperature modelling on Landsat TM thermal IR images for understanding the dynamics of the coal re. Mishra et al. (2011), empirically, appraised a scaled deviation between the temperature of the surface and underground coal res. They established it from the thermal band of Landsat ETM+ data and in-situ temperature observations using thermal imaging camera. Further, in recent years geophysical studies, mainly, magnetic (Vaish and Pal 2013Kumar et al. 2015;Pal et al. , 2017 , SP (Bharti et al. 2014;Singh et. , 2019Srivardhan et al. 2016, Mishra et al. 2018Kumar and Pal 2020;Srivardhan et al., 2016) and electrical resistivity (Gangopadhyay 2003;Mishra et al. 2018) have been carried out to map coal re in the Jharia coal eld and surroundings. In this paper, the rst time in India we have attempted an integrated geophysical approach comprising surface temperature measurement, magnetic, and electrical resistivity tomography (ERT) studies for detection of concealed subsurface coal re zones over Akashkinaree Colliery to impart a review for suitability of coal re study in Jharia coal eld using geophysical methods.

Details Of The Study Area
In India, Jhaia coal eld is famous for coal mine res. Here coal re was rst reported in 1916. About 40 million tons of coal have been consumed, and about 1.5 billion tons of coal are inaccessible due to re (Stracher 2010). Jharia will continue to burn until effective prevention, and extinguishment procedures are developed and employed, or the coal burns out. Considering the amount of coal that has burnt since 1916, it is estimated that if all the remaining inaccessible coals are left to burn at the same average rate, the res will continue for another 3,800 years (Stracher 2010). Approximately 77 coal mine res have been reported (Michalski 2004;Mishra et al. 2011). Government of India is determined for rehabilitation and resettlement of the people from 595 endangered sites and diversion of railway tracks and roads, making it easier to douse the underground res that have burned for a century and huge mine reserves of premium coal (http://timeso ndia.indiatimes.com/city/kolkata, 10 Feb 2016). Keeping the safety of people, Ministry of Railways, Government of India has decided to stop/divert the operation of goods and passenger tra c on Dhanbad-Chandrapura Railway line ( Figure 1) approximately 400 m from the study area since 15 June 2017 due to the Jharia coal eld Fire (http://www.newindianexpress.com/nation/2017/jun/12/coal-mine-re-railways-to-close-section-injharkhand-1615847.html). This is expected to cause huge revenue loss to Indian railways. The ery deposits of Jharia coal eld in India, are mainly valued because they are the only source of high-quality coking grade coal in the country. India spends Rs.300 billion rupees a year on importing that grade alone (http://in.reuters.com / article/india-coal-mines-modi-idINKBN0OG01R20150531). Government of India hopes to increase production at the blazing mine. The effects of the coal re and associated carbon emissions played a serious concern on climate change at the COP21 summit in Paris (http://www.cnbc.com/2015/12/02). Jharia coal eld covers areal spreads approximately 456 sq. Km (fox1930) and the basin extends 38 km in East-West direction and 18 km in North-South direction. The coal eld has 46 seams in Barakar formation, and 26 coal seams in Raniganj formation, and rests are local seams (Sengupta 1980). As per the quality, coal of Barakar formation is superior to Raniganj formation. It has a thick sedimentary sequence approximately 2900m associated with Carboniferous to Permian age. The coal eld is bounded by metamorphic rocks and the basin is mainly consists sandstone of Talchir, Barakar, barren measures and Raniganj formation. (CMPDI 1988). Figure 1 shows the generalized geological map of Jharia coal eld and location of Akashkinaree Colliery, Jharia Coal eld. Borehole litholog indicating coal seams with depth is shown in Fig.2.

Methodology
The present study deals with an integrated study comprising surface temperature measurement, magnetic and electrical resistivity tomography (ERT) methods for mapping of coal res with their depths, extension and propagation direction. The heat transmission from the underground combustion zone to the surface is a prolonged process, which involves convection and gas emissions through cracks/ ssures/vents. These are the main sources of surface thermal anomalies. This technique is coste cient and may provide high accuracy for delineation of sur cial location and extent. Although it is time-consuming; it could not be used for depth estimation of the combustion zone. It is also affected by air circulation and ambient conditions, for example, solar radiation, wind, and precipitation (Song and Kuenzer 2014).
The underground coal combustion and rise in temperature of overlying formation result in signi cant changes in magnetic susceptibility and thermo-remanent magnetization (Shao et al. 2014). Magnetism rises gradually with the increase of temperature till the Curie temperature. The overlying formation drops its magnetization and becomes paramagnetic due to heating above the Curie temperature. Afterwards, magnetism intensi es signi cantly due to cooling of the overlying formation below the Curie temperature (Shao et al. 2014;).
Powell and Scho eld (1939) examined the electrical resistivity of carbon and graphite at high temperatures. They observed that electrical resistivity decreases with thermal conductivity, similar to graphitization. Further, Northrop (1975) observed that in normal state the coal's resistivity varies from ~100 to ~ 3000 Ωm. In the pyrolysis process for temperatures greater than ~650 0 C, the resistivity of coal decreases approximately to 1 Ωm (Northrop 1975;Duba, 1977Duba, , 1983. This huge decrease in resistivity is due to the higher carbon content of char. Based on these studies, Duba (1977) recommended electrical resistivity technique for characterization of underground concealed coal seam res.
The electrical resistivity tomography (ERT) has been used for detection and mapping of coal seam re, its depth, extension, combustion centre, and the resulted cave or rubble zone. In the present study, ERT data have been collected using 61-channels 64 electrodes FlashRES-Universal ERT data acquisition system (ZZ Resistivity Imaging Pty Ltd, Australia). The acquisition involves imposing direct current into the ground using two current electrodes, and the potentials are measured at remaining 62 electrodes. This results in 61 potential differences with respect to a reference electrode (Fig.3). The choice of the reference electrode is dynamic and is always considered to be three spacing away from the position of the current electrodes through which the current is being injected (Zhe et al. 2007). The ERT data of multiple arrays have been collected in-one-go which support effective and faster data acquisition than the conventional systems (Zhe et al. 2007). The program considers both the quality and current threshold in noise analysis. Quality is de ned as the standard deviation of the ratio of voltage to current and is expressed in percentage. The histogram of quality of the collected data versus frequency and average current values for all electrode pairs versus each electrode position is used to optimize the maximum number of data points for analysis considering current threshold and quality factor. Bharti et al. (2016a) have analyzed ERT data collected for cavity detection over a coal re-affected area using 3 different combinations of quality factor and current threshold, i.e., (i) 5% and 60 mA, (ii) 12% and 40 mA, and (iii) 17% and 10mA. They argued that the best result is obtained with minimum RMS error, with quality factor 5% and current threshold of 60mA. Further, Bharti et al. (2016b) analyzed ERT data collected for cavity detection using a xed quality factor 5% with variable current thresholds (50mA, 100mA and 150mA). This study reveals that the best result is obtained with a minimum RMS error when the current threshold is 150 mA followed by100mA. Generally, relatively noisy and inconsistent data are sieved out at higher current thresholds and decreasing order of quality factor.
In the present study, the collected data have been processed with quality factor 5% and the current threshold of 100mA in FlashRES Universal survey data checking program (FlashRES Universal, User manual). These thresholds correspond to 94% (pro le AA / ), 95% (pro le BB / ) and 93% (pro le CC / ) of the acquired data covering a good amount of consistent data points ( Figure 1). We have acquired data using Schlumberger, Wenner, Dipole-Dipole and Gradient arrays. Further, combined inversion of all ltered data collected by all arrays have been performed in a 2.5D resistivity tomography program developed by Zhou and Greenhalgh, (1999, 2000, 2001. Zhou and Greenhalgh (2000) and Dahlin and Zhou (2006) argued that the inversion of combined data (all arrays) would further optimize the resolution capability and signal-to-noise ratio. De la Vega et al. (2003), Stummer et al. (2004) and Athanasiou et al. (2007) argued that combined inversion provide the relative bene ts associated with all the arrays and generate more realistic 2D subsurface resistivity section. The colour legend for all 2D electrical resistivity tomography sections are shown individually to highlights various anomalous resistivity variations associated with subsurface features (Bharti et al. 2016a, b;Shao et al. 2016;Gautam, and Biswas 2016;Srivastava et al. 2020).

Results And Discussions
Initially, surface temperature and magnetic data are collected at about 820 locations along 13 pro les at 5m measurement spacing. The pro le length varies from 200 to 500m. Each pro le is almost parallel and separated approximately at 25m line spacing, which covers an area of about 300m width and 500m length. The pro les are selected on the basis of existing plan and space over the study area for demarcation and mapping of underground coal seam res. A base station is selected over a non-coal re area. Repeated base station readings are collected at 30-minute interval to correct drift or diurnal variation. To achieve stable readings of surface temperature at each station were dug a small hole to eliminate the effect of top surface soil.
In-situ surface temperature measurement is carried out using Fluke 51 series-II Digital thermometer. The variation in the temperature data is about 5 0 C. Drift corrected surface temperature anomaly map (Fig.4) shows variations from ~90 to 22 0 C. Three prominent (TH1, TH2 and TH3) zones of high thermal anomalies are identi ed with black dotted lines which are possibly associated with re in coal seam XVIC, XVIB and L1/L2, respectively. In addition, some more temperature anomaly zones (TH4, TH5) are identi ed in Fig.4. TH5 (denoted by white dashed lines) could be associated with burned and cooling of XVIA. The thermal anomalies are not following exactly as the anomaly might be related with a thermal vent and could not relate therefore to the exact horizontal position of the burning front itself (Prakash et al. 1995;Karaoulis et al. 2014).
The total magnetic eld intensity (TMI) data are corrected for diurnal variations which is about 30nT over the study area. This collected data are adjusted according to the International Geomagnetic Reference Field (IGRF) in Geosoft Oasis Montaj. Further, reduced to North Magnetic Pole (RTP) anomaly map has been generated in Geosoft for removal of lateral shifts and asymmetries caused by the local orientation of the magnetic anomaly Pal et al. , 2017. The RTP anomaly map and its residual anomaly map are shown in Fig.5a and 5b. It is observed that RTP anomaly varies from ~866 to -637nT, whereas residual of RTP anomaly varies from ~162 to -283nT. From residual of RTP anomaly map, three prominent (RTPL1, RTPL2 and RTPL3) zones of negative magnetic anomalies have been demarcated that are possibly associated with underground re in coal seam XVIC, XVIB and L1/L2. RTPH1 could be associated with burned and cooling of XVIA.
Different sets of Euler's depth solutions have been calculated for structural index (SI) = 0 (represents contact, fault), 1 (represents thin sheet edge/ sill), 2 (represents vertical pipe and horizontal cylinder), and 3 (represents sphere) for the magnetic data (Reid et al. 1990) using Euler deconvolution technique. It nds probable tightest cluster of the solution associated with different complex coal re source geometry and edges of coal re front using various window sizes (5x5, 10x10, 15x15) and depth tolerances (5%, 10%, and 15%). Only a limited number of solutions are found in SI = 0, but a larger number of clustered solutions are observed in the higher order of SI (SI= 1, 2 and 3) except some additional clustered solutions of deeper depth. Reid et al. (1990) discussed that the depth estimation is more precise for higher SI. Thompson (1982) argued that most of the subsurface structures are better represented by higher indices, but lower indices are generally directly related to the depth to the top of structures. Relatively better clustering observed for a window size of 10x10 and depth tolerance 10% with SI=3 for magnetic data. The Euler's depth solutions of RTP anomaly (SI = 3.0) is overlain on residuals of RTP anomaly (Fig.6a). Three clustering patterns EM1, EM2 and EM3 have been identi ed from Euler's depth solutions, which could be associated with re in coal seam XVIC, XVIB and L1/L2 respectively (Fig.6a).
Whereas EM4 could be associated with burned and cooling of XVIA (Fig.6a). Histogram of Euler's depth solutions of the RTP anomaly is shown in Fig.6b. It is observed that about 20%, 40%, 15%, 9%, 11&, 3% and 2% of Euler's depth solutions using magnetic data occurred in the depth range of i) 10-15m (seam L1 and L2 Fig.2), ii) 15-25m (seam XVIA, Fig. 2) and iii) 25 -30m (L3, Fig.2), iv) 30 -37m (seam XVIB Fig.2), v) 37-46m (seam XVIC and L4, Fig.2), vi) 47 -50m (seam L5, L6, Fig. 2), and vii) 50 -53m (seam L7, Fig. 2). It is observed that maximum solutions are occurring in 15-25m associated with seam XVIA, followed by the Euler's depth solutions of 10-15m associated with seam L1 and L2. It indicates that re intensity in seam XVIA is comparatively highest because of thick coal than the thin coal in L1/L1. Euler's depth solutions corresponding to depth ranges 25 -30m (seam L3), 30 -37m (seam XVIB), 37-46m (seam XVIC and L4), 47 -50m (seam L5, L6) and 50 -53m (seam L7) are decreasing because of insu cient oxygen supply to the deeper coal-seams. A schematic model of re propagation over the Akashkinaree Colliery, Jharia coal eld (Fig.13) has been generated based on the resistivity tomography sections of the lines AA / , BB / and CC / . Four highly conductive (RESL1, RESL2, RESL3 and RESL4) zones and one high resistive (RESH1) zone have been identi ed . Along pro le AA / , two prominent low resistivity anomalies of ~20 Ωm (RESL1) and ~10 Ωm (RESL3) associated with active coal res have been delineated at depths of ~25 (RD 50m) and ~ 35m (RD 160m), which correspond with the coal seam XVIC and XVIB (Fig.2), respectively. RESL1 corresponds with TH1, RTPL1, and EM1, whereas RESL3 corresponds with TH2, RTPL2, and EM2. Further, a prominent high resistivity anomaly of ~500 Ωm (RESH1) at a depth of ~35m and RD 240m, has been delineated as a void due to complete combustion of coal seam XVIA and different local seams (Figs.2 and L3). RESH1 could be correlated with TH5, RTPH1, and EM4. Along the pro le BB / , a prominent low resistive signature of ~10 Ωm (RESL3) has been delineated at a depth of ~40m and RD 175m, due to active coal re in coal seam XVIB and local seams (Figs.2 and L3). A high resistivity anomaly of ~700 Ωm at RD 250m associated with a void at a depth of ~20m have been delineated corresponding to the complete combustion of coal seam XVIA and local seams. In addition, a relatively high resistivity anomaly of ~320 Ωm at RD 75m associated with a void at a depth of ~42m have been delineated corresponding to the partial combustion of coal seam XVIC and local (L3) seams (Figs.2 ). Similarly, a low resistive zone with resistivity of ~10 Ωm (Res-L3, ~258 0 C of Fig.13f) due to active coal re in coal seam XVIB at depths of 3 0m at RD145m has been delineated along pro le CC / . Two high resistive zones of ~300 and ~400 Ωm resistivity at depths of ~30 and ~35 m with RD50 and RD220m have been delineated as voids due to complete combustion of coal seams XVIC and XVIA, respectively. The low resistive zones RESL4 and RESL2 at a depth of ~9 to 15m have been observed in all three pro les, which are possibly due to burning of the different thin local coal seam at shallow depth (Figs.2 and 13). RESL4 could be correlated with TH3, RTPL3, and EM3. The low resistive zone Res-L2 is not correlated well with the surface temperature, magnetic data. In addition, a number of small low resistivity patches near the surface are observed which may not be directly related to coal res. These are possibly cracks and small potholes lled with moist soil/debris. These potholes are generated from secondary fractures during complete combustion phase.
The moist soil/debris in near-surface cracks/ ssures/potholes generates vapour/smoke due to underneath coal re activities, which results in relatively low resistivity anomaly (Shao, 2016). The voids (RESH1) with resistivity range ~300 to 700 Ωm have been demarcated in all the lines AA / , BB / and CC / . Although, the air-lled cavities indicate extremely high resistivity values. Previous studies (Ezersky 2008 andBharti et al. 2016a) established based on synthetic modeling that the air eld cavities are appeared to be relatively low resistive cavity in the inverted resistivity tomography section due to the assorted and multifarious subsurface assorted coal mine environment. Shao et al. (2016) examined electrical resistivity variations of different coal-bearing rocks with very low resistivity at ~700 0 C temperature. Moreover, the very low resistivity at the burning centre of a coal reaffected area performs a crucial role in decreasing resistivity of the overburden sedimentary rocks and rock-forming minerals. Some micro-fractures occurred on exceeding thermal stress beyond the critical strength of the rock. Subsequently, they join each other resulting with macro cracks, which collectively further lead to the formation of major cracks, fractures, potholes, and ssures enhancing the electrical resistivity.
A conceptual model (modi ed after Bharti et al. 2016a) for coal combustion, char and void formation resulting from coal seam re is shown in Fig.14. The underground coal seam re activities are chaotic process involving numerous ill-de ned functions, which lead to the formation of char from the pyrolysis process at a temperature of ~650 0 C. The resulted volatile + O 2, and char + O 2 create an exothermic environment which generates a huge amount of heat developing micro fractures. Further, enormous heats create numerous primary and secondary fractures/cracks/ ssures, in the overlying rock strata which are exposed to the surface. Numerous primary and secondary fractures/cracks/ ssures lead to high O 2 circulation for complete combustion of char and associated volatiles with enhanced cavities caused by volume reduction due to the transformation of coal to char and further to ashes mixed with rock debris.

Conclusions
Delineation and mapping of coal mine res have been carried out by combined study comprising surface temperature measurement, magnetic, and electrical resistivity tomography methods over a re-affected area around the Akashkinaree Colliery, Jharia coal eld, India. Three prominent zones have been demarcated based on combined observations, surface temperature (TH1, TH2 and TH3), and RTP of magnetic data (RTPL1, RTPL2 and RTPL3), which are related with underground res in coal seam XVIC, XVIB and L1/L2. A zone associated with burned and cooling of coal seam XVIA has also been delineated separately with the surface temperature (TH5) and RTP of magnetic data (RTPH1). The clustering patterns of Euler's depth solutions for magnetic (Fig.6a) signatures are also corroborated well with the different coal seams. The varying range of SI values with similar clustering pattern indicate complex nature of possible re source geometry of spherical zones or re propagation along different inclined/horizontally extended fractured planes/vertical cylindrical opening through potholes, which are generated in multiphase coal seam burning associated with various complex coal re activities.
Resistivity tomography sections generated by the inversion of combined data (all arrays) clearly delineate active coal seam res and associated voids compared to an individual array (Wenner, Schlumberger, Gradient, Dipole-Dipole). The schematic model (generated by inversion of combined arrays in lines AA / , BB / and CC / ) of re propagation delineates four highly conductive zones (RESL1, RESL2, RESL3 and RESL4) and one highly resistive zone (RESH1). RESL1 is associated with re in coal seam XVIC, which corresponds with TH1, RTPL1 and EM1. Res-L3 is associated with re in coal seam XVIB corresponds with TH2, RTPL2, and EM2. Res-L4 associated with res in local coal seam-L2 corresponds with TH3, RTPL3, and EM3. Res-H1 is resulted due to burning of coal seam XVIA leading to void formation that corresponds with TH5, RTPH1, and EM4. The depth range of the estimated Euler's solutions is well correlated with the depth estimated using ERT, which are also consistent with different coal seam present in the Borehole. The re propagation model established based on combined geophysical data can be effectively used for mine hazard mitigation planning around the study area.
Declarations Kumar S, Pal S K, Vaish J and Shalivahan S (2015) Utilization of magnetic gradient method for coal re mapping of Chatabad area, a part of Jharia Coal eld, India; J. Eng. Geol. Special Publication 170-176. Litschke T, Wiegand J, Schloemer S, Gielisch H and Bandelow F K (2008)      61 Channel ERT data-acquisition eld setup using 64 electrodes, of which two as current electrodes(A and B) and one as common reference electrode (M) and 61 potentials reletave to M on remaing electrodes.
The choice of the reference electrode is dynamic (Zhe et al., 2007).
Page 24/34           A conceptual model of coal combustion, char formation and void formation resulting from coal seam re.