The Practicality of Resistivity Method for Recognizing Vertically Distributed Anomalies Using Simulated Models

Abstract


Introduction
At any construction site, ground is expected to be variable in its mechanical and physical properties vertically and/or horizontally.In complex geology sites, for example, heterogeneous glacial soil, site investigation can be a challenging task (Simons et al., 2002).Brownfield sites, might reveal near surface buried and unidentified structures (e.g.buried foundations, buried tanks).Site investigation, therefore, is a primary goal before any redevelopment process.It can be performed using cost-effective, timesaving and non-invasive geophysical methods such as resistivity method which then followed by traditional direct testing methods (e.g.cone penetrating test).
The resistivity method has been applied and studied in different site investigation studies due to its financial aspects, rapid deployment and ability to reconstruct subsurface variations (Loke, 2021).It is used to recognize different soil layers (Soliman et al., 2021).The resistivity method was utilized to map the shape and thickness of weakness zones from the hosting material at the University of Anbar, Iraq (Salman et al., 2020), and for the State Company for Glass and Refractories in Al-Ramadi city, Iraq (Salman and Al-Rahim, 2023).Adesola et al. (2017) performed the resistivity method to reconstruct the material and the groundwater table underneath a collapsed pavement of a road in Nigeria.It has been applied for investigating the near surface concrete buried foundations were found at depth, about 2 to 3 m, Oslo Harbour, Norway (Lysdahl et al., 2017).Numerical studies have provided the electrical resistivity method to eliminate different buried foundations as a potential obstacles in brownfield sites, (Eissa, 2022).For a contaminated groundwater site, caused by a dumpsite in Nigeria, the resistivity method was able to detect the contaminated zone, which was determined based on the low resistivity values and a hydrochemical test (Olaseeni et al., 2018).
The resistivity method can be applied by injecting an electrical current via a pair of metal electrodes and the potential difference is then measured by another pair of electrodes.Based on the arrangement of these four electrodes, different electrode configurations can be obtained.Szalai and Szarka (2008) summareized more than a hundred electrode configurations, however, few are widely used, for example, Wenner, Sclumberger and dipole-dipole configurations.For buried foundation studies, the electrical resistivity method is reviewed and the dipole-dipole configuration was found the most deployed one (Eissa, 2021).Each of the electrode configurations has advantages and limitations if compared with other configurations (Loke, 2021).Therefore, comparison studies were conducted to compare certain configurations ability in determining a known subsurface target.Several resistivity profiles were collected over a single buried wall foundation that had been built using clay bricks; where Wenner and dipole-dipole configurations were found better than pole-pole and pole-dipole configuration (Eissa et al., 2019).
In this study, a comparison is attempted to image two horizontal targets that located at different depths but centered on the same vertical line (i.e.vertically distributed).These two targets not only have different resistivity values; than that of the hosting background, but also have different sizes.This theme was tested to simulate an expected complex underground condition in the brownfield sites investigations, for example, a water-filled storage tank underneath a paved ground.The aims of this study, therefore, can be achieved through the following methodology; 1) to utilize a forward modeling software package to generate different resistivity data sets, 2) to compare three different electrode configurations (Wenner (W), dipole-dipole (DD), and Wenner-Schlumberger (WS)), 3) to investigate six different cases of the high and low resistivity targets, 4) to apply the inversion process for all the obtained data sets, and 5) to evaluate the tested electrode configurations based on a qualitative comparison.

Generating the Synthetic Datasets
The resistivity datasets were generated using RES2DMOD ver.3.01 (Loke and Barker, 1996), this numerical modeling software package has been utilized to generate synthetic datasets in other studies, see for example (Abdullah et al., 2018;Eissa, 2022).The finite difference algorithm was applied in the data generation.The total length of each tested configuration was 15 m, where 61 electrodes were applied with an electrode spacing of 0.25 m.The model cells between any two adjacent electrodes were sized down by using four nodes, in purpose to increase the generated models resolution.
The resistivity values of the background hosting material was set up to 100 ῼ.m.The resistivity values of the high resistivity target (H), and the low resistivity target (L) were 400, and 20 ῼ.m, respectively.The targets were modelled in two different lengths, 1m and 0.5m parallel to the survey profile (i.e. on the x direction).Furthermore, the targets' position was swapped, where the high resistivity target was positioned shallower, after that, the low resistivity target was positioned to be the shallower.Ultimately, six different cases can be obtained.First, the high resistivity target located closer to the ground surface and underneath it the low resistivity target was positioned, and both of the targets have the same length (1m), this case is named as H and L case 1, as shown in Fig. 1a.In the second case, the targets have the same positions but the high resistivity target (H) has a shorter length (0.5m), this case is named as H and L case 2, as shown in Fig. 1b and c, on the other hand, illustrates the third case, where the low resistivity target has shorter length to be 0.5 m and the high resistivity target has its original length 1 m.
In the second tested arrangement, the right column in the Fig. 1, the low resistivity target was positioned closer to the ground surface, and the high resistivity target underneath it.When the targets have the same length (1m) the case is given the name L and H case 1, when the low resistivity target is shorter (0.5m) the case is named L and H case 2, and the name L and H case 3 is given when the high resistivity target is shorter (0.5m).These three cases are presented in Fig. 1d through f.

Results
All the synthetic and numerically produced resistivity datasets were inverted using the RES2DINVx64 ver.4.03, Geotomo software.Normal mesh with finite difference method were chosen.Due to the straight boundaries of the subsurface generated targets, robust inversion algorithm was selected (Claerbout and Muir, 1973).This algorithm was used by many researchers; such as (Zhou and Dahlin, 2003).

Resistivity Models of the Dipole-Dipole Configuration
With regard to the H and L cases, 1, 2, and 3, the inverted resistivity models are represented in Fig. 2a through c, respectively.The dipole-dipole successfully eliminated the high resistivity target (H) and the low resistivity target (L) from the hosting ground.The high resistivity target (H) was represented in a reasonable size when compared with its original size in the forward modelling step.The low resistivity target (L) seems to be dilated downward about more than double as its in original size.For the (H and L) case 3, although the (L) has a small length on the survey profile direction (0.5m), it was imaged and dilated in a similar response as (H and L) case 1 and case 2, where the length of the target is 1m.The hosting background material between the two investigated targets was barely reconstructed and was partly imaged with high resistivity value (as much as the high resistivity target) and partly imaged with low resistivity value (as low as the low resistivity target).For the L and H cases, 1 through 3, the dipole-dipole configuration was able to recognize the H and L targets' anomalies as can be seen in Fig. 3 a through c.In the L and H case 1, the low resistivity target (L) was imaged in a size closer to its true size in the forward model.The high resistivity target (H) was represented in a dilated anomaly much wider and much deeper than its actual geometry.For the L and H case 2, the L target was reconstructed in reasonable size (close to its original dimensions).The H target, in case 2, not only imaged in stretched anomaly but also it was represented with a higher resistivity value compared with those in cases 1 and 3.This might be related to the smaller size of the low resistivity target (L) target (0.5m) which does not mask the deeper target completely.In the third case, the result is the same to that represented in the case 1.
The low resistivity hosting background was poorly eliminated from the above high resistivity and the below low resistivity targets.

Resistivity Models of the Wenner Configuration
The images obtained from the inversion datasets of the Wenner configuration, for the H and L cases 1, 2, and 3 are illustrated in Fig. 4 b and c, respectively.In the H and L case 1, the Wenner was able to image only the H target but it was poor in determining its lower boundary.The low resistivity target (L) was imaged in a much deeper location than its original place and stretched several times in the horizontal direction.For case 2, a similar result to case 1 was obtained, however, less stretching in the horizontal direction.In case 3, the H target appeared stretched downward more than the original geometry of the target.The low resistivity target, on the other hand, was not detectable in the Wenner configuration and it was totally missed.
The low resistivity background between the high resistivity and low resistivity anomalies was noticeably recognized, although not to its original location, especially in case 1 and case 2. The obtained resistivity images from the Wenner configuration, for L and H cases, are represented in Fig. 5 In all three cases, only the low resistivity target (L) was imaged with a reasonable match with the target's geometry in the forward model.The high resistivity target (H) was not detectable using the Wenner configuration, where the target was imaged in a resistivity value as that of the background.

Resistivity Models of the Wenner-Schlumberger Configuration
The Wenner-Schlumberger configuration displays both the H and the L targets clearly as shown in Fig. 6 a, b, and c, for the H and L cases 1, 2 and, 3, respectively.The H target, in case 1, was reconstructed with a reasonable size compared with the actual size in the forward modeling.The L target appears stretched in the horizontal direction and downward.In the second tested case, similar results to those represented in case 1 were obtained.Case 3 showed slightly different results, if compared with those in case 1 and case 2, where the lower boundary of the high resistivity target was located deeper than its original location.In addition, the low resistivity target appeared smaller than those in the cases 1 and 2.
Fig. 6 also shows the response of the low resistivity background that lay between the investigated high and low resistivity targets.Although it appeared smaller than its actual height, it was represented slightly well in case 2.  The obtained resistivity images of the Wenner-Schlumberger configuration generated from L and H theme is represented in Fig. 7.For all the three investigated cases, 1, 2, and 3, only the low resistivity target was recognized, whilst the high resistivity target disappeared in the resistivity model.

Discussion
The electrical resistivity method was tested to eliminate near-surface that horizontally extended but arranged in vertical sequence anomalies, as potential underground conditions in redevelopment brownfield sites.Synthetic models were generated using the RES2DMOD (a forward modeling software) (Loke and Barker, 1996).Three popular electrode configurations (Wenner, dipole-dipole, and Wenner-Schlumberger) were utilized to generate 2D electrical resistivity datasets over the tested targets.
The results show that the dipole-dipole electrode configuration was better than Wenner-Schlumberger to discriminate the investigated targets, as has been noted by other researcher, (Abdullah et al., 2018).The dipole-dipole was also better than the Wenner configuration to discriminate the targets from the hosting background and reconstruct the target dimensions.That might be related to the fact that the dipole-dipole configuration generates better resolution than the Wenner configuration (Loke, 2021).Similar findings have been found as shown by (Eissa et al., 2019).
The importance of the targets arrangement and targets dimensions was also highlighted.It appears when the low resistivity and small target (0.5 m) locates closer to the ground surface can have a limited masking phenomenon and make it possible for the longer (1 m) and deeper target to be detected and eliminated from the containing background, only dipole-dipole could image the deeper and longer anomaly where the Wenner and Wenner-Schlumberger could not.Therefore, not only multiple profiles of resistivity data should be collected, as expressed by (Lysdahl et al., 2017) but also different electrode configurations should be utilized, as (Eissa et al., 2019) had performed in their study.Electrode configurations, therefore, should be wisely chosen and 2D resistivity datasets should be carefully interpreted and supported by other datasets such as direct drilling data.

Conclusions
Three common electrode configurations in resistivity surveys were attempted to detect two different near-surface targets.Those targets have different resistivity values, different dimensions, and are horizontally extended but vertically distributed, in total, six different themes were investigated.
From the obtained inverted images, it can be concluded that the vertical distribution of the subsurface targets can have a noticeable effects on the overall resistivity model.The high resistivity target, almost, can be detected only if it is positioned shallower than the low resistivity target.On the other hand, for the low resistivity target, it was detectable in most of the modelled cases even when it has been located deeper (i.e. at the bottom of the vertical distribution of the tested targets).
In the perspective of the attempted electrode configurations, the dipole-dipole seems to be the optimum in recognizing the targets in all the investigated cases, followed by the Wenner-Schlumberger.The dipole-dipole, therefore, is proposed to perform resistivity surveys for brownfield site investigations.

Fig. 1 .
Fig. 1.Concentrated slices on the horizontal but vertically distributed targets obtained from the numerical forward modeling using the RES2DMOD, all dimensions are in metre, H is the high resistivity target, and L indicates the low resistivity targets

Fig. 2 .
Fig. 2. Resistivity inverted images using DD configuration, a) H and L case 1, b) H and L case 2, and c) H and L case 3, H is the high resistivity target, L is low resistivity target, see Fig. 1 for more details about their arrangement and dimensions

Fig. 3 .
Fig. 3. Resistivity inverted images using DD configuration, a) L and H case 1, b) L and H case 2, and c) L and H case 3, H is the high resistivity target, L is low resistivity target, see Fig. 1 for more details about their arrangement and dimensions

Fig. 4 .
Fig. 4. Resistivity inverted images using W configuration, a) H and L case 1, b) H and L case 2, and c) H and L case 3, H is the high resistivity target, L is low resistivity target, see Fig. 1 for more details about their arrangement and dimensions

Fig. 5 .
Fig.5.Resistivity inverted images using W configuration, a) L and H case 1, b) L and H case 2, and c) L and H case 3, H is the high resistivity target, L is low resistivity target, see Fig. 1 for more details about their arrangement and dimensions

Fig. 6 .
Fig. 6.Resistivity inverted images using WS configuration, a) H and L case 1, b) H and L case 2, and c) H and L case 3, H is the high resistivity target, L is low resistivity target, see Fig. 1 for more details about their arrangement and dimensions

Fig. 7 .
Fig. 7. Resistivity inverted images using WS configuration, a) L and H case 1, b) L and H case 2, and c) L and H case 3, H is the high resistivity target, L is low resistivity target, see Fig. 1 for more details about their arrangement and dimensions