Imaging the distribution of bitumen contaminants in shallow coastal plain sands in southwestern Nigeria using electrical resistivity

Naturally occurring bitumen seeps migrating within the shallow subsurface contaminates soil and groundwater in communities within the eastern segment of the Dahomey Basin in southwestern Nigeria. Managing these contaminations require an understanding of the distribution of the bitumen seeps to isolate contaminated regions to avoid farming activities or drilling shallow drinking water wells within them. Several studies have assessed the shallow occurrence of bitumen in the region, focusing on their economic potentials but ignores the challenges with soil and groundwater contamination in these communities. This study focuses on using electrical resistivity imaging techniques with limited soil cores to delineate the distribution of bitumen seeps within the shallow subsurface in Imakun-Omi community in southwestern Nigeria. We used both 1D and 2D numerical and field approaches in this study. 13 vertical electrical soundings using a Schlumberger electrode array were first acquired to obtain the bulk resistivities and layer thicknesses. The 1D resistivity alongside well lithologic data served as prior data for a forward modelling study to assess the capability of resolving the bitumen seeps within electrically resistive coastal sands. Results of the forward modelling studies guided the acquisition of five 2D electrical resistivity profiles used to image the distribution of bitumen seeps using a Wenner electrode array with a unit electrode spacing of 2 m. Results of this study show that bitumen seeps with resistivities greater than 3000 Ohm-m are distributed within the top 4 m in the area and extends down to 10 m in some locations. Our numerical studies show that despite difficulties in imaging bitumen seeps within sediments overlain by higher electrical resistivity layers, reliable results can be derived, thus proving the versatility of electrical resistivity in delineating areas contaminated by bitumen seeps within shallow coastal sands.


Introduction
Environmental contamination from bituminous sand also refer to as tar sand or oil sand is a major concern globally (Grant et al. 2008;Rahman et al. 2013;Rooney et al. 2012). The source and mode of contamination differs in different regions and could either be of geogenic or anthropogenic origin (Gbadebo 2010;Hein 2006;Mossop 1980). Most of the reported contamination and concerns till date, mostly from Canada and Australia are focused on anthropogenic contamination facilitated by the exploration and production of oil from them (McKirdy et al. 2017;Timoney and Lee 2009). However, bitumen seeps; which are liquid or semisolid petroleum escaping from compromised subsurface reservoir at a low rate to the Earth's surface, may contaminate the soil and groundwater while migrating through them. Such contamination is considered to be of a geogenic origin (Ligate et al. 2022). As the bitumen seeps naturally through the shallow subsurface to the surface, it migrates through aquifers and soils and may be emplaced within them. This occurrence of bitumen seeps within soils and aquifer over time may cause the release of toxic chemical compounds such as PAHs, sulfides, and Naphthenic acids (NAs) through hydrological and biogeochemical processes such as leaching, sorption and desorption mechanisms (Norin and Strömvaix 2004;Roy et al. 2016). Due to chemical disequilibrium, the concentration of the chemical compounds released into the soil and groundwater system would increase over time making the system unfit for use (Loop and White 2001;Rivett et al. 2001;Zheng et al. 2015).
Geogenic soil and groundwater contamination by bitumen seeps have been observed in multiple communities within the Dahomey Basin in southwestern Nigeria (Gbadebo 2010). While the source of the seeps is yet a subject of research (Adeyemi et al. 2013), managing their impact on the soil and water resources requires delineating their spatial extent. Accessing clean drinking water in these regions would require targeting non-contaminated sections of the aquifers in the area hence, the need to understand the spatial distribution of bitumen in the shallow subsurface. While environmental contamination resulting from bitumen seeps has been the subject of some recent studies (Hewitt et al 2020;Rahman Talukder et al. 2013;Rooney et al. 2012;Timoney and Lee 2009), none has focused on the geogenic contamination as observed in Southwestern Nigeria. Most of the studies on shallow subsurface occurrence of bitumen have focused on its exploration and production potentials for oil neglecting the current environmental impacts on communities where these seeps occur.
The occurrence of bitumen seeps in soils and shallow aquifers in southwestern Nigeria is associated with the absence of an effective caprock that could seal the original hydrocarbon deposit in place within the subsurface reservoir (Eruteya et al. 2021;Ogungbemi et al. 2019;Omosanya et al. 2012). Hence, the hydrocarbon migrates upward to the shallow subsurface where processes including evaporation, reduction-oxidation, and biodegradation remove the lighter component leaving behind the heavier component-e.g., bitumen (Ahad et al 2018;Fennell and Arciszewski 2019;Korosi et al 2016). When the seeps arrive at the shallow subsurface, their interaction with air and porewater phases in the unsaturated zone (soil) and groundwater in the saturated zone result in adsorption to grains and desorption into the liquid phase via chemical disequilibrium releasing dense non-aqueous phase liquid (DNAPL) hydrocarbon over a long duration which contaminates the soil and groundwater (Scott et al. 2012). Understanding the level of such geogenic contamination is necessary to manage and remediate the contaminated soil and aquifers. Also, such understanding will be necessary to distinguish the naturally occurring contamination from the future impact that may be associated with the exploration of the shallow and deep bitumen deposits in Nigeria (McKirdy et al. 2017).
Delineating the spatial extent of bitumen seeps in the shallow subsurface can be done via core samples and crosscorrelation between sampled locations to extrapolate the spatial extent of the bituminous sand (Korosi et al 2016). This approach assumes continuity or discontinuities in unsampled regions which like other geologically heterogeneous systems may be misleading, whereas collecting more sediment cores to better constrain the deductions is invasive and expensive. However, geophysical techniques such as seismic, electromagnetics, ground penetrating radar and electrical resistivity have been successfully used to delineate such stratigraphic variabilities within the shallow subsurface (Adepelumi et al. 2006;Doro et al. 2013;Sharafeldin et al. 2019).
The use of electrical resistivity to image bitumen occurrence and distribution within the subsurface relies on an expected contrast in electrical properties of bitumen-bearing sediments and that of their surrounding (Blondel et al 2014). We hypothesize that due to the low conductivity of bitumen, sediments with a significant amount of bitumen seeps should have higher electrical resistivity. Hence, electrical resistivity measurements offer a proxy, yet effective approach to study the distribution of bitumen seeps within the environment. Omosanya et al. (2012) carried out electrical resistivity measurement of bitumen seep zones using Wenner and Schlumberger electrode configuration within the eastern Dahomey basin and discovered the bitumen to have relatively high electrical resistivity values compared to the surrounding lithologic units. Also, the effect of geological structure on bitumen seeps formation has been studied using electrical resistivity measurements, with results showing the lack of quality cap rocks and the presence of conduit faults as potential mechanisms for the seep development within the bitumen belt of Dahomey basin (Akinmosin et al. 2011). Other geophysical techniques employed in the exploration of bitumen seepages within the Dahomey basin include the use of well logging techniques and shallow seismic refraction and reflection techniques to determine the effect of stratigraphic control on bitumen seep formation (Enu 1990). Laboratory and field geophysical technique has also been applied to detect subsurface hydrocarbon contaminations (Blondel et al 2014). This combined method was employed to delineate the temporal and spatial change in the geoelectrical signatures from a hydrocarbon contaminated site.
As noted earlier, all previous studies attempting to map bitumen distribution within the shallow subsurface in the Dahomey Basin of southwestern Nigeria have focused on the potential for exploring them as non-conventional sources for oil production and/or as sources of asphalt (Adeyemi et al. 2013;Agagu 1985;Akinmosin et al. 2011;Anukwu et al. 2016). Hence, these studies focused on large-scale structures and distributions which are also at relatively greater depths. However, bitumen seeps have been observed outcropping on the surface in several locations with the dark brown color of the soil and groundwater in areas where they outcrop ( Figs. 1 and 2). To understand and better manage the environmental impact of these naturally occurring contaminations, detailed knowledge of the distribution of the bitumen seeps, most especially within the top 5 m is required. Hence, this study focuses on using geophysical techniques to delineate naturally occurring bitumen seeps within the shallow subsurface which contaminates soil and water resources in parts of southwestern Nigeria by employing 1D and 2D electrical resistivity measurements. A conceptual model of the near surface aquifer architecture in Imakun-Omi area in southwestern Nigeria is also developed. This study extends the application of the well-established electrical resistivity geophysical method to delineate shallow subsurface distribution of bitumen seeps within the shallow coastal plain sands in southwestern Nigeria and also isolate regions within the shallow aquifer that may not be impacted by bitumen seeps based on measured electrical resistivity contrast. As the bitumen seeps are emplaced in saturated and unsaturated coastal plain sands which typically have high electrical resistivity values, discriminating areas within the coastal plain sand with bitumen seeps is not trivial. This is because the highly resistive unsaturated coastal plain sands mask the imaging of the resistivity perturbations in underlying low resistivity saturated sand units. Hence, using a forward calculation with prior estimate of the bulk resistivities via VES, we assessed the suitability of resistivity method to delineate the bitumen seeps.

Study site
This study was conducted at Imakun-Omi Community in the Dahomey basin of southwestern Nigeria located within Latitudes 6°26′00′′ to 6°27′30′′ and Longitudes 4°24′00′′ to 4°26′00′′ (Fig. 1). The region and the adjacent communities are characterized by lowlands with few ridges' characteristics of the tropical rain forest of southwestern Nigeria (Faniran and Jeje 1993;Odunuga et al. 2013). Temperatures in the area are relatively high during the dry season with highs of about 33 °C while low temperatures are around 26 °C experienced during the rains, especially between July and August (Akintola 1986). Rainfall distribution in the region varies from about 1000 mm in the western part to about 2000 mm in the eastern part; and the high rainfall promotes perennial tree growth with varying height (Oguntunde et al. 2011). The area is well drained by NE-SW trending rivers such as the Lekki, Ogun, Shasha, Oba and Opeki Rivers.
The study area falls within the coastal area of the onshore eastern Dahomey basin of Nigeria (Fig. 1). The lithostratigraphic units which consist of Late-Cretaceous to Quaternary sedimentary sequence according to Enu (1985), are summarized in Table 1. The surface sediments in the area consist mainly of Pleistocene-Oligocene coastal plain sands.
These tertiary sediments are some of the youngest within the Dahomey basin (Agagu 1985;Falufosi and Osinowo 2021). At the study site, bitumen was observed seeping out at the surface within the coastal plain sand. This was observed as a dark-brownish, mostly linear coloration in the typically whitish-grey beach-like sands (Fig. 2a). Surface and groundwater in the area were also observed to be contaminated by the bitumen seeps with the water having a dark brown coloration (Fig. 2b). While a few groundwater wells within Imakun-Omi Community gave clean water year-round, most show seasonal contamination with a dark brown color during the rainy season when the water table rises significantly (oral communication with Imakun-Omi community members).

Methods
For this study, we used the classical electrical resistivity method which involves the injection of electrical current between a pair of source electrodes and measuring the resulting potential difference. Theoretical details of this widely used technique are discussed in Binley and Slater (2020) and Everett (2013). We first acquired 1D electrical resistivity data via vertical electrical sounding to determine bulk resistivity distributions at the study site. We then used the bulk electrical resistivity estimates coupled with ground truthing information from well lithology logs as prior data to conduct a resistivity forward modeling to assess the capability of electrical resistivity to image bitumen distribution within the heterogeneous coastal plain sands at the site testing different geometry and soil-bitumen distribution scenarios. This was necessary as imaging the subsurface bitumen seeps through resistivity survey was complicated within the coastal plain sand terrain of the study area. The coastal plain sand units are highly resistive within the unsaturated section and this layer inhibits current flow thus limiting the application of resistivity imaging. This justified the need for forward numerical exercise to ascertain the suitability and limitations of ERI in delineating the distributions of the bitumen seeps within the study area.

Vertical electrical sounding
We acquired thirteen vertical electrical resistivity soundings (Fig. 1b) using a Campus Ohmega resistivity meter while Garmin's global positioning system (GPS) was used to record the geographic coordinates of each VES station. All 13 resistivity soundings had a maximum current electrode spacing (AB) of 200 m and were acquired using the Schlumberger array. In the Schlumberger array, the measuring electrodes are placed closely together such that the ratio of the measured voltage to the current electrode separation is approximately equal to the voltage gradient at the midpoint of the current spread (Oladunjoye and Jekayinfa 2015). Apparent resistivities from the field data were plotted against the half current electrode separation spacing and the data were adjusted to minimize inversion errors using IPI2win software (Nwachukwu et al. 2019). From the resultant curves of these plots, qualitative information such as the resistivity of subsurface lithologic units, the depth and thickness of each units and the curve signatures were deduced. The qualitative interpretation of each VES curves were improved upon using the JOINTEM software, an automatic iterative computer program employed for carrying out an iterative inversion processes until the root mean square (RMS) error is lower than 10% (Pirttijärvi 2010).

Numerical simulation of electrical resistivity imaging
To optimize field Electrical Resistivity Imaging (ERI) of the distribution of bitumen seeps within the study area, we first conducted a conceptual and forward resistivity modelling of two-dimensional electrical resistivity distribution using the finite element based forward solution, R2 resistivity code implemented in ResIPY (Blanchy et al 2020). The parameters used in the model generation were based on prior information acquired from subsurface lithologic sequence derived from wells constructed within the study site, as well as resistivity and thickness of subsurface geologic units derived from VES surveys. The manipulation of the geoelectric parameters (Table 2) obtained from thirteen VES data acquired in February and July 2020 respectively  (Table 3) for the calculation of numerical resistivity data based on the Wenner array, which were later inverted using the ResIPY resistivity inversion program (Boyd et al. 2019). For the inversion the regularized inversion method with linear filtering was employed with the inversion process converging with final RMS misfit below 1.5% after 2 iterations. Topography data were not incorporated into the resistivity models since topographic changes are negligible within the coastal plain terrain of the study area.
Based on the available prior information, conceptual resistivity models were developed for the numerical simulation. The model assumed four (4) subsurface resistivity units (Fig. 3) along a resistivity profile acquired using the Wenner array with a minimum electrode spacing of 'a' = 2 m. The conceptual model imitates anticipated position of bitumen units in subsurface geologic units. Based on available Prior information from Table 2, the lithologic unit and corresponding resistivity values used for the numerical modelling exercise are; topsoil/Unsaturated sand (6000 Ωm), Contaminated sand aquifer (1000 Ωm), clean sand aquifer (80 Ωm) and the suspected bitumen units (39000 Ωm) as highlighted on Table 3. A 5% gaussian noise representing instrumental and possible background noise was added to the synthetic model to simulate field conditions.

Field acquisition of electrical resistivity imaging data
Using the same minimum electrode spacing of 2 m as in the numerical study, field ERI data were acquired with the Wenner array technique using the Campus Ohmega resistivity meter during the break in the rainy periods of July 2020. Based on the work of Furman et al. (2003), the Wenner array shows a reasonable sensitivity to resistivity perturbation in the subsurface when deployed independently, and it is suitable for rapid deployment where resources might be limited. This is important for our study as the 2D data were acquired manually due to lack of appropriate resistivity meter for an automatic multi-electrode data acquisition. For this study, we acquired five 2D electrical resistivity profiles of length 150 m each with 'a' = 2 m (Fig. 1b). To ensure quality and  consistency of acquired field data, and to reduce data error, resistivity measurements at each point were taken in at least three cycles and quality data were obtained with the observational errors being less than 1%. Also, good contact was ensured at the current electrodes by driving the electrodes sufficiently deep into the subsurface and proper fastening of current and potential cables to their respective electrodes. Furthermore, contact resistance at the base of each electrode was reduced through the addition of saline solution (a mixture of salt, water, and lateritic mud) to the contact point between the electrode and the ground surface. For the inversion of the field data, the ResIPY program used for the inversion of the numerical modelling was employed. The resulting inversion results were constrained using the information available from well logs within the study site for thickness and the modelled background resistivities determined for each geo-electric layer from the numerical studies (Table 3).
In order to ensure that the regularized inversion solution of the field resistivity data converges within 10 iterations, a normalized inversion error filter set to filter inversion error outside the range of − 3 to + 3% was employed according to the suggestions of Binley et al. (1995).

Well installation and lithologic logging
Three wells A, B and C were installed along ERT profile 04 (Fig. 1b) within the study area to a uniform depth of 15 m. The wells were located at a region within the study area suspected of significant bitumen accumulation and also accessible with drilling equipment. The wells were logged during drilling at an interval of 1 m through the retrieval of drill cuttings done at the well head before the cuttings made it to the ditch to minimize cross-contaminations by sediments from different depth intervals. The sediments retrieved were subjected to physical examination to determine sediment types, degree of saturation, nature of saturating fluid and color in order to classify the lithologic changes with depth within the study site. The information from the well logging exercise was used to create a lithologic profile for the study site which served as ground truthing data for the interpretation of the 2D electrical resistivity data acquired within the study site as direct comparison can be made between change in subsurface resistivity and change in lithologic conditions.

Results
Selected results of the vertical electrical soundings are presented in Fig. 4, while others including a summary of the interpreted electrical resistivity, layer thickness, curve types and lithology are presented in Appendix A and B. Generally, the VES results show 4 to 5 distinct resistivity layers with varying combinations of sounding curves K, H and Q type. Resistivities varies widely for each layer and ranges from 853 to 25501 Ωm for the tops soil, 781 to 12,223 Ωm Generally, resistivity in excess of 3000 Ωm was suspected to signify the presence of bitumen in the subsurface lithologic unit. The topsoil/unsaturated sand unit and the aquifer units extends to the depth of about 6 and 28 m respectively, while the depth of the underlying layer is unknown. The results from numerical study of the electrical resistivity imaging of bitumen distribution within the shallow subsurface (depth ≤ 15 m) are displayed as Fig. 5. The conceptual models deployed for this study simulated three scenarios which include (i) subsurface layers modelled without the presence of the bitumen contaminants (Fig. 5a), in order to create a background subsurface resistivity model to which contaminated subsurface resistivity models can be compared, (ii) modelling of the subsurface layers with bitumen presence in the topsoil unit (Fig. 5b), and (iii) modelling of the subsurface layers with bitumen contamination of the aquifer units (Fig. 5c). Also, Fig. 5a, shows the forward and inverted resistivity model of the subsurface in the absence of bitumen contamination. The results of the numerical modelling of bitumen buried within the topsoil layer is shown in Fig. 5b, while those of bitumen buried within the contaminated aquifer is shown in Fig. 5c.
Based on the initial conceptual models, simulated bitumen seeps with resistivity of 39,000 Ωm located around 20 m, 70-80 m and 120-130 m at depth of 1-3 m within the resistive unsaturated layer (6000 Ωm) underlain by a less resistive contaminated aquifer (1000 Ωm at depth of 4-8 m) were well captured in the inverted model (Fig. 5b). On the contrary, when the simulated bitumen seeps were located in the less resistive contaminated aquifer (1000 Ωm) overlain by the more resistive unsaturated sand (6000 Ωm), they could not be resolved in the resistivity model (Fig. 5c) as there seems to be masking of the bitumen contaminant resistivity by the overlying high resistivity layer.
Results of the field electrical resistivity tomography showing the measured apparent resistivity pseudosection, inverted resistivity section, model sensitivity and normalized inversion error plot for profile 1 is shown in Fig. 6. A model fit with root mean square (RMS) error of 1.13% was achieved using all the measured data while the model shows good sensitivity to considerably high resistivity zones to depth of about 12 m. The inverted resistivity section for all the five measured profiles (Fig. 7) show resistivity signatures corresponding to a similar lithology expected to be sand units' characteristics of the study site as observed from lithological information in constructed wells. Variations in resistivity signature within the inverted section were interpreted to result from degree of sand pore spaces saturation with air, water or bitumen. Water saturated layers give resistivity generally below 1000 Ωm with presence of bitumen resulting in high resistivity values in excess of 3000 Ωm. The unsaturated sand layers which extends to a depth of about 3 m show higher accumulation of bitumen units due to available air-filled pore space with resistivity values in excess of 20,000 Ωm. Lastly, information from the lithologic logging of wells A, B and C, (Fig. 8) shows that the study area is underlain to a depth of 15 m by quaternary sediments which are made up of alluvium or coastal plain sands as shown on Table 1. The sediments however shows varying physical properties in terms of coloration and saturating fluids, in well A, white sands unsaturated to a depth of about 2.5 m was observed, the unsaturated sands are followed by saturated white sands to a depth of about 3.5 m after which dark-brown sands saturated with brownish water was observed to a depth of about 5.5 m. The dark-brown sands are underlain by a light-brown sand unit which is also saturated with brownish water to a depth of about 8.5 m, this is followed by a dark-grey sand unit which extends to a depth of 11.5 m and saturated with a slightly clean water. The last unit in the well lithologic sequence consists of a brownish-white sand which is saturated by clean water and is expected to extend beyond 15 m where the well terminated. Wells B and C shows similar lithologic sequence to well A howbeit there appears to be difference in thickness as well as depth of observation of the various lithologic units as depicted on Fig. 8. The presence of bitumen was observed on the unsaturated sand units as shown on Fig. 2a, whereas the impact of bitumen contamination is believed to be pronounced between the depth of 3.5 m to about 8.5 m within the dark-brown and light-brown sand units with the influence of contamination decreasing with depth until about 11.5 m where the light-brown sand units which appears to be the clean water aquifer within the study site was encountered.

Discussions
Seeping bitumen migrating through subsurface lithologic units exhibit anomalous electrical resistivity signatures which can be assessed to delineate their presence and distribution in the subsurface. Bitumen is a mixture of several hydrocarbon compounds which are known to exhibit relatively high resistivity in contrast to their host'slithologic units (Blondel et al. 2014). This has been observed with seeping bitumen present within the Dahomey sedimentary Fig. 4 Vertical electrical sounding results showing the fitted apparent resistivity curves and inverted resistivity, thickness and depth at the righthand corner of each for a VES station 1, b VES station 5, c VES station 7, and d VES station 13 respectively basin of south western Nigeria exhibiting resistivity values in excess of 3000 Ω-m (Eruteya et al. 2021). In some cases, bitumen resistivity values drops relatively compared to that of the surrounding host rock particularly in near surface or surface environment where the bitumen has degraded to strains with relatively low electrical resistivity (Atekwana et al. 2000;Cassiani et al. 2014). The effect of hydrocarbon degradation on its electrical resistivity properties is Fig. 5 Conceptual, forward and inverted resistivity model sections for subsurface resistivity perturbations a in the absence of bitumen contamination, b for bitumen contamination in the unsaturated sand and c for bitumen units in the contaminated aquifer important when assessing the distribution of hydrocarbon rich contaminants such as bitumen in the environment using ERI techniques. However, in situ bitumen seeps migrating through the subsurface as observed within the site of this study are distributed in the shallow subsurface and at depth below the water table where the effect of degradation is minimal due to lack of exposure to atmospheric conditions. This implies that the electrical resistivity signatures of the seeping bitumen could be preserved in place.
The challenge of imaging the distribution of this seeping bitumen contaminants using ERI arises from the relatively high electrical resistivity of the coastal plain sand units hosting the bitumen seeps within the study site. Preliminary 1D electrical resistivity measurements at the study site (Table 2) shows that the unsaturated coastal plain sand units have resistivity values ranging from a few hundred ohm-m to values in excess of 10,000 Ω-m. This implies poor resistivity contrast between the contaminating bitumen seeps and the host coastal plain sand units. To overcome this limitation and efficiently image the distribution of seeping bitumen contaminants within the coastal plain sand units, results from 1D electrical resistivity soundings were combined with that of 2D numerical and field resistivity imaging to delineate the distribution of emplaced bitumen seeps within the first 15 m of the subsurface.
Initially, results from 1D electrical resistivity were combined with information from well-logs to determine the geo-electric subsurface parameters of the study site (Fig. 9), drawing parallel from earlier resistivity studies on shallow subsurface bitumen bearing units within the Dahomey basin (Omosanya et al. 2012;Ogunlana et al. 2019). The VES results delineated 4-5 litho-resistivity units interpreted as; (i) Topsoil/Unsaturated sands which is in tandem with the unsaturated white sand unit on the well lithologic log. (ii) Contaminated aquifer which corresponds to the depth between 2.5 and 11.5 m on the lithologic log consisting of the saturated white sand to the dark-grey sand units. (iii) Clean aquifer which corresponds to the brownishwhite sand units below the depth of 11.5 m on the lithologic profile. (iv) Clayey sand or Bitumen-bearing sands which is suspected to exist at depth beyond the scope of the well logs (Fig. 9).
The geo-electric parameters were employed in 2D numerical resistivity modelling and eventually on constraining the inversion of field ERI data. From the numerical study, the distribution of seeping bitumen contaminants within the unsaturated coastal plain sand units was vividly delineated (Fig. 5b). Conversely, the efficiency of the ERI to delineate bitumen contaminants within the saturated coastal plain sand aquifer is greatly reduced (Fig. 5c). The bitumen contaminants within the saturated units were masked in the resulting inverted resistivity section, as they are represented as mere distortions in the observed resistivity image. The masking of the bitumen contaminants within the saturated sands is Fig. 9 Comparison of results from VES station 7 with observed lithological information from well A suspected to result from the suppression of the bitumen induced resistivity perturbations within the saturated units by the highly resistive overlying unsaturated Coastal Plain sands. This resistivity suppression is related to the possible layer suppression observed when the target of an ERI, in this case the seeping bitumen does not produce sufficient contrast and geometry when compared to its surrounding (Reynold 2011).
The results of the constrained inversion of field ERI data corresponds with that from the numerical exercise. The study shows 3 main geo-electric layers which corresponds to different resistivity zones in response to water and bitumen saturation within the available pore spaces of the coastal plain sands (Fig. 10). Zone 'U' with average thickness of about 2 m represents the unsaturated white sand units with vacant pore spaces available for more bitumen accumulation. Zone 'cAq' is the unconfined saturated aquifer unit consisting of the white, dark-brown, light-brown and the dark-grey sand units suspected to be contaminated with the presence of bitumen. Lastly, zone 'Aq' represent the clean brownish-white sand aquifer unit with little or no bitumen contamination and potentially the only region suitable for groundwater exploration within the study site. Other Sub-Zone observed are 'Bi' which represents suspected accumulation of bitumen unit within the pore spaces of the unsaturated white sands and Zone 'm-Bi' which shows suspected masked bitumen accumulation within the saturated sand or aquifer unit. The inversion of the ERI field data was constrained with lithologic thickness from available well-logs and background values for each geo electric layer as observed from preliminary 1D resistivity survey and 2D numerical modelling within the study site (Tables 2 and 3). Thus, interpretation of the bitumen contaminants distribution from the field ERI data was assumed as regions with resistivity above 3000 Ω-m (Loke 2018) within the unsaturated sand and contaminated aquifer units, while anomalous high resistivity zones within the clean aquifer is suspected to represent bitumen seep contamination within the inverted resistivity sections. Increase in the observed background resistivity of the ERI geo-electric layers is suspected to be proportional to the degree of bitumen saturation.
To improve the limitations of the ERI in delineating the distribution of the bitumen seeps within the saturated units (Fig. 5c), the ERI results are combined with the available 1D resistivity sounding results. Unsaturated regions suspected of hosting bitumen units observed on the ERT profiles have similarly high resistivity values on corresponding overlapping VES stations as shown at VES 07 on ERT profile 03 (Fig. 11). This combined interpretation shows that regions with bitumen accumulation within the unsaturated region; as indicated by high resistivity value from 1 and 2D resistivity inversion results, could be underlain by deep sited bitumen bearing units as observe at VES station 07. This implies bitumen seeps distributed within the unsaturated sand units must have migrated through the water table from underlying bitumen bearing lithologic units. Finally, the results from the numerical exercise based on prior data acquired through 1D resistivity survey show that the predicted resistivity response for the assumed bitumen blubs in the numerical exercise are consistent with field observation from the study area, although variations in field dimensions of bitumen units occurred. Also, observed geo-electric layer parameters show high degree of consistency.

Conclusions
Electrical resistivity measurements involving both 1D vertical electrical sounding coupled with well lithologic profiling and 2D numerical and field ERI were used in this study to delineate the distribution of bitumen seeps contaminating coastal plain sands within the shallow subsurface in the Dahomey Basin of Southwestern Nigeria. The coastal plain sands are characteristically highly resistive geologic material with this property reducing resistivity contrast between seeping bitumen and host Coastal Plain Sand units, thus making delineating of the later using ERI technique difficult. To overcome this limitation, we combined both numerical and field approaches to improve the experimental design and model interpretation in this study. From the results obtained, we identified zones with resistivities more than 3000 Ohm-m within the top 15 m which were interpreted as areas contaminated by the bitumen seeps with higher resistivities observed related to the degree of bitumen induration within the sands.
Numerical 2D resistivity results acquired from 1D resistivity studies coupled with lithological information from wells guided the acquisition and interpretation of 2D subsurface resistivity distribution data at the study site. The bitumen units within the unsaturated topsoil sands were imaged clearly with the 2D ERT, thus suggesting the reliability of electrical resistivity in delineating the distribution of bitumen seeps within the shallow subsurface. Results from field resistivity imaging data shows acceptable correlation between resistivity response of bitumen units to those from the numerical models, with subsurface images from field resistivity data delineating suspected bitumen units within the coastal plain sands. Also, combine interpretation of the VES and ERT generated subsurface geoelectrical parameters show that shallow bitumen units within the unsaturated sands delineated through ERT survey are overlying deeper bitumen units observed on the VES results, thus suggesting a connection between deep sited bitumen beds and near surface bitumen units. This study also shows that VES resistivity results can provide prior data for a rigorous numerical simulation study that can be used to improve field data acquisition and constrain the interpretation of 2D ERT for investigating complex systems.
As evident from numerical and field results, the combined interpretation of the VES (1D) and both numerical and field ERT (2D) improved the understanding of bitumen seeps distribution in the near surface environment, thus resistivity measurements may be employed to investigate near surface bitumen distribution within coastal plain sand terrain, however preliminary geophysical measurements coupled with numerical data interpretation will improve field data acquisition process and interpretation.

Appendix A
Summary of the estimated electrical resistivity, layer thickness, curve types and lithological interpretations from the 13 VES Stations.