Geophysical investigations for stability and safety mitigation of regional crude-oil pipeline near abandoned coal mines


 This study aims for the protection of a crude-oil pipeline, buried at a shallow depth, against a probable environmental hazard and pilferage. Both surface and borehole geophysical techniques such as electrical resistivity tomography (ERT), ground penetrating radar (GPR), surface seismic refraction tomography (SRT), cross-hole seismic tomography (CST) and cross-hole seismic profiling (CSP) were used to map the vulnerable zones. Data were acquired using ERT, GPR and SRT along the pipeline for a length of 750 m, and across the pipeline for a length of 4096 m (over 16 profiles of ERT and SRT with a separation of 50 m) for high-resolution imaging of the near-surface features. Borehole techniques, based on six CSP and three CST, were carried out at potentially vulnerable locations up to a depth of 30 m to complement the surface mapping with high-resolution imaging of deeper features. The ERT results revealed the presence of voids or cavities below the pipeline. A major weak zone was identified at the central part of the study area extending significantly deep into the subsurface. CSP and CST results also confirmed the presence of weak zones below the pipeline. The integrated geophysical investigations helped to detect the old workings and a deformation zone in the overburden. These features near the pipeline produced instability leading to deformation in the overburden, and led to subsidence in close vicinity of the concerned area. The area for imminent subsidence, proposed based on the results of the present comprehensive geophysical investigations, was found critical for the pipeline.


Introduction
Shallow buried pipeline networks, deployed for transmission and distribution of economically vital products such as crude oil, gas and so on are occasionally subjected to instability because of volumetric changes in the near-surface soils due to failures in the form of cracking, sagging and buckling because of surface deformations. Probable causes could be natural forces, high groundwater fluctuations or any impending subsidence under the mining environment. In one such case, a segment (∼42 km) of 300-km long crude-oil pipeline in the eastern part of India, running over old abandoned coal mines, came under thorough investigation because of such associated subsurface problems. Poorly planned and unauthorised mining activities close to this segment led to ground subsidence (figure 1), and the formation of minor and major weak zones such as voids and sinkholes (figure 2), with significant (1) Pipeline passing over the area of large subsidence, (2) series of subsidence occurred along the pipeline, (3) subsidence occurred close to the pipeline, and (4) surface runoff recharging water into the subsidence area through feeder channel.
risk to the stability of the pipeline. Failure of remnant pillars in the shallow old workings (<50 m depth) is a potential cause of ground subsidence and sinkholes. Such ground subsidence led to an abnormal sag in the pipe line leading to opening of a joint in its structure. Subsequent leakage of crude oil caused concerns of an environmental hazard in the surroundings.
It is plausible to address such problems by comprehensive geo-investigations. Geological and geotechnical studies yield information based on direct observations and point tests, but they may not be adequate for comprehensively designing a solution. Thus, geophysical investigations that are capable of bringing out the hidden causes by mapping the subsurface in detail are of economic choice and are normally preferred for such near-surface high-resolution imaging (Wang 2021). Geologically, the subsurface around the troubled segment of pipeline is composed of soil with clay and sand, followed by alternating sandstone and coal. Presence of weak zones in the form of cavities and sinkholes brings out a contrast in the electrical properties of the soil (Maillol et al. 1999; Van-Schoor 2002;Cardarelli et al. 2006;Kim et al. 2007;Martínez-Pagán et al. 2013;Metwaly & Al Fouzan 2013;Bharti et al. 2016b;Das & Mohanty 2016). Several authors followed an integrated approach to delineate the subsurface weak zones (Drahor et al. 2015; Bernatek-Jakiel & Kondracka 2016; Drahor & Berge 2017;Cueto et al. 2018;Vargemezis et al. 2019). Electrical resistivity tomography (ERT) and ground penetrating radar (GPR) are potential tools for locating such weak zones (Ballard 1983;Yelf & Turner 1990;Fenner 1995;Carpenter et al. 1998;Dobecki & Upchurch 2006;Gómez-Ortiz & Martín-Crespo 2012;Carbonel et al. 2014;Simyrdanis et al. 2018;Mogren 2020). In addition, it would be useful to derive the engineering properties of the zones for planning suitable reinforcement measures (Singh et al. 2017a(Singh et al. , 2017b. This is better achieved by seismic techniques from surface and borehole. While surface techniques could yield information on the weak zones closer to the surface (Lankston 1989;Cohen & Donahue 1994;Sheehan et al. 2003;Cardarelli et al. 2010;Butchibabu et al. 2019b), borehole imaging is used to better resolve the deep-seated features (Miller & Steeples 1991;McDowell & Hope 1993;Mufti 1995;Rechtien et al. 1995;Flecha et al. 2004;Grandjean & Leparoux 2004;Inazaki et al. 2004;Cha et al. 2006;Debeglia et al. 2006;Xu & Butt 2006;Park et al. 2008;Balasubramaniam et al. 2013;Butchibabu et al. 2017Butchibabu et al. , 2019a. Similar studies have been carried out to successfully detect underground cavities, ground deformations, mine shafts, mine and sinkholes subsidence in several coal mines elsewhere (Fisher 1972;McCann et al. 1987;Marino 2000;Johnson 2003;Prakash et al. 2009;Gómez-Ortiz & Martín-Crespo 2012;Chatterjee et al. 2015;Sahu & Lokhande 2015).
The present study aims to map the subsurface voids and cavities in the vicinity of the pipeline by deploying surface geophysical investigative techniques, such as ERT, GPR and SRT, to assess the shallow subsurface conditions and borehole techniques, such as CSP and CST, were used for mapping the deep-seated anomalies. Integration of both surface and borehole results leads us to infer the possible hazards zones below the pipeline, and motivates us to adopt the remedial measures to enhance its stability.

Study area
Coal mining activities began in this area about 300 years back in a random manner and standard coal mining practices such as board and pillar were adopted during the late 19th century (Srivastava & Mitra 1995;Lahiri-Dutt 2003). The sizes of the panel and coal pillar vary around 200 m, and between 20 and 24 m, respectively. The average thickness of the coal seam varies between 4 and 6 m and the overburden thickness ranges from ∼15 to 50 m (Singh & Yadav 1995). The occurrence of high-grade coal at a shallow level led to illegal extraction, and caused serious concern to society (Lahiri-Dutt 1999). Combustion of coal due to assimilation of oxygen into the underground methane in the coal seams was documented in the area (Martha et al. 2005;Guha & Kumar 2012). These unauthorised mining activities associated with coal fires resulted in the development of mine voids, land subsidence, sinkholes, cavities, potholes, etc., and posed threats to the environment with respect to health and safety issues (Kuenzer & Stracher 2012;Saini et al. 2016;Butchibabu et al. 2019b). These illegal coal mining operations caused severe horizontal and vertical deformations in the ground and created foundation problems in nearby structures such as buildings, road/rail links, pipelines, etc. (Can et al. 2012).
A number of cases of premature collapse of small pillars during the early part of mining in this coal belt were reported, and consequently extensive roof-falls, collapses of entrances and water logging forced government agencies to declare the areas unsafe for human habitation (CMPDIL Tech. Report 1988;Areeparampil 1996;DGMS Interim Report 1997). The Government of India rehabilitated 60000 people affected by mining induced subsidence and designated 60.55 km 2 of land as severely damaged due to subsidence, abandoned quarries and spoil dump (Chari 1989;Areeparampil 1996). The study region is the second highest urbanisation area in the state of West Bengal. It attracted human settlement due to its long history of mining, and equivocally faced serious threats of subsidence, loss of human lives and property, etc. (Lahiri-Dutt 1999).
The study area has lateral dimensions of 750 × 150 m, lies close to the Kajora coalfield of West Bengal, India and is characterised by an undulating ground surface, sloping gently toward the SW direction. In the area of concern, significant ground subsidence (∼8-10 m, figure 1) and large numbers of sinkholes with diameters ranging from ∼2 to 10 m and up to 30 m deep were found in the vicinity of pipeline (figure 2). The probable causes of the formation of sinkholes and ground subsidence were shallow depth cover, weak overburden, geological discontinuities and dislocation of rocks. Lowering of groundwater and rainfall further aggravated their occurrences. Two types of subsidence, such as sinkholes representing collapse of cover as well as trough formation, were observed in the area. The land is still not stable and events are active either in the form of troughs or sinkhole subsidence, which might have been caused by subsurface cavities formed in the area. Earlier incidences of sag in pipes and breakout at joints were reported. A sinkhole with the dimensions of 10 m diameter and 30 m deep occurred very close to the pipeline in 2015 and was subsequently filled with fly ash (figure 2a). During the borehole drilling for cross-hole tests, one full-length (∼3 m) drill bit was suddenly dropped at a 13 m depth, possibly in a void or abandoned mine gallery. This borehole casing was repeatedly going down or sinking. To stop the sinking, we tied two sticks to the borehole casing (figure 3) during the present study. Groundwater level was identified at a depth of 10 m from the ground surface in an open well (figure 3) close (∼5 m) to the pipeline alignment.

Geology of the study area
Geologically, the Raniganj coal fields comprise rocks mainly from the Gondwana group of the Early Permian to Lower Cretaceous (Guha & Kumar 2012). The Gondwana group was stratigraphically subdivided into the upper and lower Gondwana series. The lower Gondwana was again divided into three sub-series; these were the Panchet, Domodar and Talchir Formations, respectively. The lower Gondwana group (permo-carboniferous age) of rocks contains valuable resources of coal and is covered by a thick blanket of alluvium near to the Durgapur area (Das 1970). The study area marked on the geological map of the Raniganj coal fields (figure 4) is known for several environmental hazards such as subsidence, coal fires, sinkholes, cavities, etc. related to shallow coal workings. Recent studies have established the extension of the Raniganj Gondwana basin around Durgapur to form a new sub-basin known as the Durgapur sub-basin, and is separated by a subsurface ridge around Andal and Muchipara (Ghosh et al. 1993). The Raniganj coal belt comprises sedimentary rock formations with recent alluvial and lateritic deposits. Lateritic soils in the area appear to be a sheet rock on the surface and borehole drilling indicates red lateritic soil up to 4 m, followed by anthropogenic stowing sand up to 17 m and broken pieces of sandstone up to a depth of 30 m (figure 3). The laterite deposition is composed of numerous quartz grains and these quartzose laterites pass into lateritic conglomerates, lateritic gravels, etc. The principal lithological units of lateritic types and sandy gravel beds are laterites, quartzose laterite, lateritic gravels and conglomerates. Figure 5 shows the location of the study area including the layout of surface geophysical lines and borehole sites.

Data acquisition and processing
The area of present concern comprises abandoned coal mines. Drill data for the vertical section show three main layers: the top one is predominantly characterised by lateritic soil, the second one is made up from loose sand and there is a bottom third layer of sandstone layer. GPR and ERT techniques were used for mapping the subsurface cavities and weak zones. The weak zones in the strata of coal were later confirmed by the SRT technique. Borehole investigations were carried out to resolve the deeper information, and simultaneously provided the detailed information of the strata as well as associated weak zones.
148 Figure 3. Site conditions of the study area. (a) Lateritic soil is exposed close to the pipeline, (b) open well close to the pipeline illustrates the ground water level at a 10 m depth from surface, (c) borehole drilling has exposed anthropogenic stowed sand and (d) sinking of borehole casing under its own weight and sticks tied to the borehole collar to arrest sinking.

Electrical resistivity tomography (ERT)
ERT data were acquired using a dipole-dipole array laid out at a 2-m distance from the pipeline that was buried at a depth of 2-3 m in the ground. The array measured 744 m in the NW-SE direction along the pipeline (AL-1) with electrode spacings of 8 m, in roll-along mode with an overlap of 24 electrodes (see Table 1 and figure 5). In addition, 16 ERT profiles were also gathered with a 3-m-spaced 48-eletrode array orthogonal to the AL-1, each of 141 m in length (R1-R16) in the NE-SW direction. These lines were laid at 50m intervals. The ERT line along the pipeline was profiled to map the deeper horizons (up to 50 m), while the across ones aimed to decipher shallower features close to the pipeline, as the subsurface was expected to be criss-crossed with irregularly formed tunnels and the size of the troublesome weak zones was expected to be ≥2 m. The assumption of abandoned mine workings is that they are more or less internally homogeneous with sharp boundaries. We used the blocky inversion method (l 1 -norm) in RES2DINV processing software (Claerbout & Muir 1973) as the resistivity contrasts were expected to be significant in subsurface features. As the data were relatively free of noise, a damping factor of 0.05 was used for ERT inversion. Vertical to horizontal flatness, the filter was set to 1.0 to reduce the elongation of features that were neither along the horizontal nor vertical directions. Nodes between adjacent electrodes were set at four to optimise apparent resistivity values for the chosen electrode spacing. A Jacobian matrix calculation was applied using a quasi-Newton approximation, and the width of the model blocks was set equivalent to the electrode spacing. This inversion parameter file was read several times and minor values were changed for achieving the minimum ABS error. Thus, the ABS errors were reduced to 5-10% during the processing for the along (AL-1) and 16 across (R1-R16) profiles.

Ground penetrating radar (GPR)
Step frequency GPR (SFGPR, Kong & By 1995) was used for subsurface profiling along the pipeline over a length of 700 m (AL-1, figure 5) covering the span by ERT spread.   Measurements were made in static mode using a 2-m long dipole antenna at every 0.5-m step, in the frequency range of 10-200 MHz (Table 1). This bandwidth was chosen on the basis of resonant ground response and signal power. Common move-out profiling was done to obtain velocity of the subsurface. The radar wave velocity arrived from the moveout in the study area was 0.11 m/ns. With the selected resonant frequency range and the medium velocity, it would be feasible to resolve weak zones as small as 1 m in the top 10 m. The computed velocity values were used in ASYST data processing software to convert the time section into a depthsection (RADARGRAM, figure 6b). A radargram is colorcoded imagery depending on the amplitude of the reflected signals.

Seismic refraction tomography (SRT)
The SRT data were acquired for a length of 700 m along the AL-1. In addition, 16 lines (S1-S16) were laid out (orthogonal to AL-1) as was done for R1-R16, with a spacing of 50 m. These 16 lines, each 115-m long, were added to a 24-channel seismic array with 5-m geophone (f n = 10 Hz), the spacing was adopted for profiling using a 16-lb sledgehammer (Table 1). Each profile was gathered with nine shot points (forward, middle and reverse) for a good data coverage along the profile and multiple stacking at every shot location was done to enhance the signal to noise ratio. With this acquisition setup and the signal acquired, data were analysed to resolve features of at least 2-3 m dimensions in the top 10 m, based on the combined manual and automated processing techniques. First arrival times were manually picked up for all the 24 channels for each shot position and used in manual and automatic processing based on conventional techniques (intercept-time and cross-over distance) as well as an automated Δt-V technique (Gibson et al. 1979;Gebrande & Miller 1985). The initial model for inversion was generated based on velocity inputs of the weathering layers and inverted using wave path eikonal tomography (WET) algorithm (Schuster & Quintus-Bosz 1993) based on tracing the curved ray path of seismic wave propagation. The final 2D velocity tomogram for a P-wave was reconstructed by generating velocity contours against depth based on the WET output with an average RMS error of 2.5%.

Cross-hole seismic profiling (CSP)
The CSP survey was carried out in 12 boreholes (figure 5): six each located on either side of the pipeline. The locations of the boreholes were chosen based on the surface geophysical test results, wherein the potential weak zones were indicated. The 12 holes were drilled 30-m deep for profiling, juxtaposed on either side of the pipeline to measure the variation in V P and V S until the 30 m depth. Seismic data were acquired at every 0.5-m depth interval from bottom to top (surface) of the borehole. The P-wave borehole sparker and S-wave borehole source were used for generation of P-and Swaves. A pneumatically clamped system was deployed as the source for the generation of P and horizontally polarised SHwaves. A hydrophone (f n = 10 Hz) was used as a P-wave receiver and a borehole clamped tri-axial geophone was used as an S-wave receiver. Boreholes were prepared as per the standard of ASTM D 4428/D 4428M 2000. P-wave profiling was carried out between all the six pairs (figure 5) lying on either side of the pipe, while S-wave profiling was done across three pairs (B5-B6, B7-B8 and B15-B16) due to limited signal strength and adherence to the ASTM standards, restricting the borehole separation to 5 m only. P-and S-wave velocity profiles were generated from travel times across the corresponding pair of boreholes.

Cross-hole seismic tomography (CST)
The locations for CST borehole pairs B3-B4, B11-B12 and B17-B18 (Table 2,   12-channel hydrophone chain was used to receive them. Data were acquired from bottom to surface at 1-m spacing using a digital signal enhancement seismograph (Terraloc MK6) with a sampling interval of 25 µs and 4096 samples. The borehole seismic survey was carried out to evaluate the likely engineering properties of the weak zones based on V P and V S to derive as much information as possible in correlation with the surface geophysical test results. All the tomogram results for the three locations were analysed independently, and correlated with the results of other techniques adopted for this site. A comparative analysis of the results, evaluation of the subsurface conditions along the pipeline and interpretation along with combined visualisation is discussed in section 5.

Surface techniques
The top soil layer along the pipeline is characterised by highly weathered laterite along with sand and clay. Such a medium is known to exhibit significantly low range of resistivity under saturated conditions. However, the geophysical surveys were carried out in the month of January 2016, when there was no seasonal rain nor any large water bodies to influence the presence of shallow water table. Thus, the near-surface conditions were reasonably dry and therefore of higher resistivity. Several low resistivity (<20 Ωm) pockets, numbered 1 to 9, were mapped within the top 20 m depth (ERT2D section, figure 6a). There were no other pronounced low resistivity zones elsewhere along the AL-1 in the top 20 m; more specifically, in the 9-17 m depth range. These low resistivity pockets could be the manifestation of water-saturated or clay-filled cavities left over due to unauthorised mining. These pockets are located almost equidistant within the surrounding highresistivity strata, which is represented by sandstone and coal. Moderate resistive (∼330 Ωm) pockets in-between low resistive areas are presumably the locations of pillars used during excavation.
The up-arching 'gaps' in the GPR radargram (1-9) are in close agreement with the locations of those low resistivity pockets. The clear absence of any high-amplitude reflection is an indicator of a medium that is capable of attenuating electrical and electromagnetic signals. That is probably the impact of saturated soil with higher conductivity trapped in the abandoned void. The high-resistivity values below 15 m and lack of high-amplitude reflections below this depth are indicative of no significant contrast in the resistivity or dielectric permittivity caused by gradual change in rock character from weathered to hard rock mass. The amplitudes of radargram reflections decreased below a 12 m depth, whereas the resistivity values and P-wave velocities increased at this depth, which is inversely related to GPR. Thus, the intermittent presence of water or water-saturated rock mass, like excavated rooms, probably filled with acidic water and was surrounded by more compact and relatively dry rock mass. A drop in resistivity in the ERT at a 310-360 m distance and the coinciding multiple up-arching reflections in the same stretch of radargram are probably caused by the effect of filled well or shaft-like feature. The SRT shows the presence of a three-layered rock mass scenario with the layer apparently dipping at the end. Beyond 550 m along the profile, three prominent low velocity zones are identified. One of them is a dump zone, while the other two (V P = 400-600 m s -1 ) indicate the presence of a sinkhole (figure 2a) and cavities.
The overall resistivity distribution is in the range of 2-3900 Ωm along the 16 ERT profiles (R1 to R16) across AL-1 ( figure 7a and b). The overall scenario can be simplified by a three-layered subsurface with highly resistive (∼1350 Ωm) lateritic soil at the top (0-3 m) followed by low resistive (∼350 Ωm) anthropogenic stowing sand (3-11 m) and high-resistivity (≥1500 Ωm) sandstone at the bottom (11-25 m). Conspicuous low resistivity pockets (see the rectangular boxes in figure 7a and b) indicate the presence of a shallow level features (∼6-15 Ωm) of lateral dimensions of 6-9 m mapped in the depth range of 9 to 17 m. SRT across AL-1 lines also accounts for similar subsurface character as identified by ERT profiles with three-layered strata conditions ( figure 8a and b). They can be classified into three major velocity zones representing lateritic soils (400-1100 m s -1 ), moderate compacted and/or saturated sand (1100-2000 m s -1 ; Telford et al. 1990) and sandstone (≥2000 m s -1 ) generally in the depth ranges of 0-10, 10-20 and 20-30m, respectively.
Combined visualisation of all the 2D resistivity sections reveals a series of low resistivity pockets and is suggestive of an elongated subsurface feature, probably cavities that might be extending like a tunnel type feature. The abandoned mines have possibly undergone marked changes due to failures or deformation of pillars and collapse of the roof whose effects might have been extended up to the surface by way of stressed arching. This phenomenon subsequently affects the surface topography and occasionally leads to surface subsidence. SRT sections are not indicative of any significant contrast in the subsurface velocity coinciding with ERT features. Hence, the discrete but significant features of ERT cannot be directly correlated to any inferred subsurface feature from SRT. This is most likely the effect of size of the subsurface feature vis-à-vis the velocity variation with depth. Scattered low velocity zones observed in the SRT sections could be due to uncompacted zones in the sand or voids in the background of highly weathered rock. However, distinct low resistivity pockets lie in the relatively low velocity zones. Thus, it can be construed that low resistivity features at less than 7 m in depth could be influencing the velocity distribution in their surroundings, however, distinct contrast in velocities is not prominent. Therefore, such low resistivity features might appear as single shallow low velocity anomalies in SRT tomograms.

Borehole techniques
The CSP results (figure 9) show an increase in P-wave velocity, which is comparable with that of SRT lines S5-S6, S9-S10 and S15-S16 ( figure 8a and b). However, the S-wave velocity shows a reversal in trend beyond 9 m in depth (figure 9). Decrease in the S-wave velocity is attributed to the effect of ground water, which leads to dilation of the subsurface horizon mixed with the collapsed fragile roof of the underground galleries and uncompacted saturated sand that might have been stowed. Thus, such intriguing velocity variation at 9-17-m depths is identified as the dilatant zone (Kwasniewski & Rodríguez-Oitabén 2012; figure 9). Such velocity reversals indicate impending deformation of strata around the minedout zone. P-wave velocity varies from 1500 to 3000 m s -1 up to a depth of 30 m. The highest velocity of 3000 m s -1 was measured at a depth of ∼16 m and might be associated with a boundary between the second layer and bottom (sandstone) layer. However, the presence of a low velocity (∼1000-1500 m s -1 ) zone in the 5-10 m depth range (figure 10) probably causes voids in top lateritic soil. They are isolated anomalies and, unlike with other weak zones, are found elsewhere due to abandoned mine galleries. Presence of low resistivity pockets in ERT profiles (figure 7a and b) might be the resultant effect of such isolated pockets.   CST was conducted between a pair of holes at three locations (figure 5) affirming the trend of velocity profiles obtained by other techniques. The overall distribution of the seismic P-wave velocity (400-3400 m s -1 ) up to a depth of 30 m in the three tomograms (figure 11) has an embedded low velocity (∼1200 m s -1 ) pocket between 5 and 10 m in depth. This once again reaffirms the low velocity found in SRT and CSP (figures 8 and 9), which were identified as voids in the top lateritic soil. The effect of dilatant zone is not prominently visible in the P-wave tomograms as the contours demonstrate increasing velocity, so it is unlikely to be the S-wave, which would have shown a velocity reversal at the same depth range. The estimated poison's ratio between 9 and 17 m depths in the dilatation zone is computed as 0.4-0.5 (Table 3), indicative of soils deformed elastically at small strains and no significant volumetric change occurred in the medium (Suwal & Kuwano 2012). Such deformation is related to the dilatancy of the rock mass (Kwasniewski & Rodríguez-Oitabén 2012) observed in the respective depth range. Most likely, this dilatancy phase of the rock might have facilitated subsidence of the overburden through deformation.
Thus, the integrated geophysical investigations comprising ERT, GPR, SRT, CSP and CST exposed an intrinsic material property of the subsurface along and across the pipelines, ascertaining the locations of the weak zones. The features were mainly confined to the top 20 m of the subsurface, confirmed by both the surface and borehole geophysical techniques. The ERT section along the pipeline (figure 6a) demarcates a series of low resistivity (∼6 Ωm) anomalies in the background of higher resistivity (∼160 Ωm). This was detected because of a higher contrast in resistivity between the 158 Figure 11. Travel-time tomograms, obtained through cross-hole seismic tomography investigations, illustrate the variation of P-wave velocity up to a depth of 30 m. Note the zone of abandoned mine gallery marked by solid-line box identified in B3-B4. Low velocity cavities and feeder channels at shallow depths, extending from the surface to 10 m depth. Δ indicates a shallow level low velocity pocket also mapped in corresponding SRT sections S4 and S16. R indicates a corresponding low velocity depression on S4 and low resistivity zone on R4. water filled, saturated sand filled gallery and the surrounding high resistive sandstones. A radargram also produced hyperbolic reflections around the same locations (1-9, figure 6b), although similar congruence could not be seen prominently in the seismic sections due to the nature of moderate velocity changes.
Results from borehole techniques were in reasonably good agreement with that of the surface techniques mutually validating the overall geophysical findings. An interesting feature noted among them is the extension of a possible feeder channel up to ∼12 m in depth. The computed Poisson's ratio clearly points toward the state of the soil/rock in the early stages of deformation (Kwasniewski & Rodríguez-Oitabén 2012;Yu et al. 2020). The prolonged effect of the deformation facilitated subsidence below the pipeline and nearby surrounding regions, straining the pipe at locations where the ground subsides beyond normal because of developed local weak zones in the vicinity. Based on the results of comprehensive geophysical investigations, we recommended the area for future subsidence in close vicinity to the pipeline. Sinkhole subsidence was found later on in an area with diameter of 8 m and depth of 10 m near the CSP borehole B13-B14. It was further filled with gravel and compacted subsequently by running over it with a roller. Thus, the pipe comes under greater risk of damage due to the ground instability, and this has raised concern about its safety.

Conclusions
The integrated geophysical approach involving surface and borehole techniques clearly brought out the near surface and moderately deep-seated weak zones in a less ambiguous manner. The surface investigations were planned in strategic coordination with the borehole investigations to probe the maximum subsurface information concerning the weak zones below the pipeline, including identification of unknown mined-out regimes. Comprehensive geophysical investigations identified the possible subsidence areas in the study area and this later was confirmed by visual observation. Testing of soil samples at different depths could have enabled better quantification of the soil modulus, alongside borehole velocity values and Poisson's ratio, and empirical relationships between these parameters could be established for this site. Nevertheless, the complementing results ascertained the status of the subsurface below the pipeline allowing the authorities to prepare a systematic ground restoration design for the safety and stability of the pipeline.