Mapping the subsurface structure of the suevite deposit in the north of the Bosumtwi impact crater using electrical resistivity and seismic refraction tomographies

Electrical resistivity imaging and seismic refraction methods have been used to map the subsurface structure of the suevites north of the Bosumtwi impact crater in Ghana to determine the depth extent and in-situ resistivity and P-wave velocity of the suevites. Seven electrical resistivity and two seismic refraction profiles were surveyed. The lengths of the profiles varied between 160 and 600 m. The multi-electrode system was combined with roll along techniques for the resistivity data collection using the gradient array. The data was acquired in 2D where the electrodes separation of 4 m was the same as the geophones spacing, while the shooting interval was 8 m. The data was processed with Res2DInv and ReflexW for the resistivity and seismic refraction respectively. Electrical resistivity and seismic refraction tomographies identified the suevite deposits which were observed within 12 m depth. The resistivity of the northern Bosumtwi suevites varies between 1.56 and 25 Ωm, and the P-wave velocity ranges from 3 to 3,9 km/s. The results also showed that the subsurface is made up of either two or three layers: an unconsolidated topsoil, clayey soil, and fractured claystone. The seismic velocities and the electrical resistivities of the northern Bosumtwi suevites compare very well with those found for suevites at the Ries impact crater Germany and other impact craters.


Introduction
The Bosumtwi impact crater (Figure 1) is an important research site in Ghana. It is one of the world's youngest complex impact craters that was formed by a meteorite impact 1.07 0.05 Ma. The crater which is occupied by Lake Bosumtwi (Figure 1), has an average diameter of ≈ 10.5 km [W. Jones,1985;Junner,1940]. The lithology of the area consists of Proterozoic rocks (2.1 -2.2 Ma years) of the Birimian Supergroup that consists of lower greenschist facies metasediments, greywackes, schists, quartzites, phyllites, and a minor granitic component Leube, et al., 1990;Wright, 1986]. It is situated about 32 km southeast of Kumasi in the Ashanti Region of Ghana [Reimold, et al., 1998]. The meteorite that caused the Bosumtwi crater was an oblique impactor from the northeast [Aning, et al., 2013b].
The Bosumtwi suevites are impactites formed from a mixture of melted materials of country rock and meteorites caused by high temperatures and pressures. They have been thrown out and deposited in northern and southwestern parts of the Bosumtwi crater [Koeberl et al., 2007;Reimold et al., 1998]. Stoffler and Grieve [1994] have defined the suevites as polymictic clastic matrix breccia containing glass fragments, rocks, mineral and clasts component of impact melt exhibiting various stages of shock metamorphism. The suevites in the northern part are the interest of this study. The northern Bosumtwi suevites are situated withi 1°235 / ,1°024 / W and 6°332 / , 6°342 / N about 2.5 km from the lake shore outside of the rim (Figure 4). The suevites of Bosumtwi impact crater is formed by molten materials of country rock and meteorite, so the detailed study of suevites could reveal the rheology of country rock.
In addition, petrography and radiometric dating studies about suevites can prove the chemical composition of the meteorite as well as the age of impact crater. The lithology and subsurface structures of the impact crater can undergo serious modification as a result of shock, heat, and chemical alteration [Pilkington et al., 1992].
The energy of hypervelocity meteorite usually causes different effects on the surface target, such as deformation of the subsurface and the formation of impactites. It also induces physical property contrasts between the lithology  [Koeberl et al., 2007]. P wave velocity of suevites. In situ measurements of these physical properties as well as a continuous 2D imaging of suevitic deposits are an additional contribution to the African impact crater science.
Seismic refraction tomography uses first arrival travel times as raw data and is based on inversion techniques to image the subsurface by considering velocity gradient of 2 dimension pixels [Eppelbaum, 2014;Sandmeier, 1998;Wapenaar et al., 2005]. The subsurface deformations and the various strata were also delineated. In addition, the electrical resistivity and the P-wave velocity of the northern Bosumtwi suevites were determined.
In the Bosumtwi impact crater, Plado et al. [2000] revealed generally the difference in physical properties of pre-impact early Proterozoic metasediments (target rocks) and melt-rich suevites (impactites). It was found that the suevites have low densit ≈ 2040 kg/m 3 , high porosity ≈ 25 % and high magnetization (magnetic susceptibility ≈ 330 × 10 -6 relative to the target rocks with density, porosity as well as magnetic susceptibility SI. Also, a study which was done at the Ries impact crater in Germany by Ernstson (1974), has specifically found that resistivities of suevites range from about . The Chicxulub Scientific Drilling Project (CSDP) found that the ultrasonic P-wave velocity for the suevites in the Chicxulub impact structure (Mexico) ranges from with as mean porosity [Popov et al., 2014]. Additionally, suevites for the Chesapeake Bay impact structure (USA) have P-wave velocities between with aboutas mean porosity [Popov et al., 2014].

Geological background of the Bosumtwi crater area
The geological stratigraphy around lake Bosumtwi consists of series of supracrustal rock types; namely metasedimentary as well as meta-volcanic rocks, belonging to the 2.1-2.2Ga Birimian supergroup. Both these sequences are intruded by mostly granitoides [Jones et al., 1981]. The Bosumtwi impact event excavated lower greenschist facies metasediments of the 2.1-2.2 Ga Birimian Supergroup [Jones et al., 1981;Leube et al., 1990]. The lithology of the country rocks that constituted the target rocks are chiefly made of meta-greywackes, shales, and phyllite of the Proterozoic Birimian Supergroup and some intrusion of granitoid [Ferriere et al., 2007;Junner, 1940). There are three types of impact breccias which appear at and around the crater; these are monomict lithic breccia, polymict lithic breccia and also suevites [Boamah and Koeberl, 2003;Reimold et al., 1998]

Data acquisition and processing
Seven profiles of lengths varying between 160 and 600 m were surveyed in the northern suevite deposit ( Figure 4).
Six of them were specifically in suevite deposits whilst one was outside the deposits to serve as a controlled experiment.

Electrical Resistivity Survey
During the resistivity survey, the ABEM LUND resistivity imaging system was employed for data collection. Four multi-core cables and 41 electrodes spaced 4m were used for each spread. The gradient array and GRAD 4L8 protocol  [Boamah & Koeberl, 2003].
were then selected to enhance the data density, lateral and vertical resolution during the data acquisition [Aning et al., 2013a;Stummer et al., 2004] Before each measurement, the electrode test was done to ensure that the electrodes were properly connected and planted as well. The computer-controlled multi-electrode system or electrical resistivity tomography (ERT) and roll along survey techniques were combined to scan across the survey line.

5
Mapping of Bosumtwi crater suevites Figure 4. Location of the study area relative to geological map of Ghana, adopted from [Duodu, 2009]. Seven profiles and three boreholes are indicated with blue line color, and black circle respectively.
RES2DINV software (www.geotomosoft.com) was used process the data. After removing bad data points, the raw data were inverted based on robust and least squares inversion in order to get the subsurface resistivity distribution that almost fits the measured data. The absolute errors varied between 1.38 to 9.4. The resistivity model sections underwent 5 to 7 iterations and were displayed with user-defined logarithmic contour intervals. Model sections were displayed including topography where topographic correction was done for each electrode-position.

Seismic Refraction Survey
In this work, a hammer of 6 kg was used as energy source to cause the vibration of the ground. This nondestructive source was therefore used to deliver 3 consecutive impacts on a rubber plate and the results were stacked.
At each shot-position, the trigger geophone was used to trigger a measurement. A spread of two cables was used simultaneously, and 12 geophones of 10 Hz were vertically planted in series into the ground along each cable with 4 m spacing making a total distance of 92 m, and 8 m shooting interval. On each section of 92 m, thirteen inline shot points and one offset point at 4 m from the first geophone were considered in order to increase the resolution of seismic refraction tomography. The roll-along technique was implemented by overlapping the last two geophones to extend horizontally the area to be covered by the survey ( Figure 5).
ReflexW software (www.sandmeier-geo.de) was used for processing the seismic data. Processing was done by picking first arrival travel times at first break of wiggles. In travel time analysis, after inserting shot zero travel time of all first arrival seismic traces, layer models were generated. However, as a result of detailed subsurface imaging in terms of vertical and lateral velocity gradient, seismic refraction tomography models were subsequently generated. Before the inversion process, maximum velocity variation of 200% was used in order to enable strong vertical velocity gradient. Detection of small-scale variation is ensured since space increment of 1 m was chosen while smoothing value in the horizontal direction was half the shot point spacing (4 m). Finally, 10 iterations were executed as well as topographic correction for each geophone position.

2D resistivity and seismic refraction models
The base-line passed between BH1 and BH3 and consisted of two successive profiles 1 and 2 separated by 14 m, through which is the Asisiriwa Nyameani road. Each profile was about 600 m long. The electrical resistivity and seismic refraction methods were solely implemented on the base line. Figure 6 (a) shows the resistivity model section of profile 1 which was surveyed from North-East to South-West.
The electrical resistivity tomography of profile 1 shows that the subsurface is formed by 3 layers. The depth of investigation of this profile as well as other electrical profiles was 25 m. Boundary B delineates the top layer thickness which is ≤ 4 m and has slightly high resistivity values from around 60 -400 Ω . This layer is interpreted to be unconsolidated soil with some moisture content. It is followed by a low resistivity layer of a thickness ranging between 4 and 10 m which are dominated by clayey soil, as confirmed by Boamah and Koeberl (2003).
Emmanuel Habimana et al. At the locations 32 -64 , 128 -160 , 256 -320 and 512 -544 along the model section, the subsurface is found to have a low resistivity distribution (blue formation) which could represent the suevites. The resistivities range between 1.56 -25 Ω , and they occur at depths 20 . The suevites are brecciated materials which are more saturated than the surrounding soil.
The highest resistivities of 6000 -25000 Ω , in the third layer start from around ≤ 20 m; it may be a zone of the bedrock of claystone. The resistivity distribution within the basement rock (claystone) is found to be highly changing laterally. This is interpreted to be due to the fractured and shattered basement rocks as a result of the meteorite impact. Fractured claystone zones can be found beyond 10 m depth, at the locations 170 -220 , 300 -350 , and 450 -500 along the model section. velocities of 600 / therefore representing the moist, unconsolidated soils in this layer. The second layer with a velocity of about 2000 / has a thickness varying from 4 -7 and is interpreted as clayey soil with some moisture content. The third layer was also delineated with a high velocity of about 2400 / . The lithology of this layer is dominated by consolidated country rocks of claystone similar to the results of the resistivity survey.
The seismic refraction data was further interpreted in terms of tomography in order to know how the lithology of the subsurface is controlling the velocity gradient laterally and vertically. Thus, Figure 6 (c) represents the tomography model section of profile 1, which revealed a high-velocity contrast (3000 -3900 / ) within the second layer. It was therefore interpreted as suevite deposits which were found at about 10 m depth. The high-velocity contrast within tomography models is the result of the suevites which are more saturated than the host sediments. Figure 7 (a) is the resistivity image of the second profile. It was surveyed from South-West to North-East. It is subdivided into three layers. The first layer of slightly high resistivity values of 60.00 -400.00 Ω represents unconsolidated soil with moisture content. This first layer is delineated by boundary B, where its thickness is ≤ 4 .
It is followed by a zone of the lowest resistivity 1.56 -24.00 Ω that is embedded in the second layer. This zone is interpreted as suevite deposits embedded in a clayey soil. Its thickness varies between 4 and 12 m.
The third layer starts from about 12 m with high resistivity values ranging from 400 -25000 Ω indicating the presence of a bedrock which is mainly composed of claystone. At depths ≥ 12 , the resistivity distribution is highly changing laterally; this shows that the basement rock is fractured and shattered. Fractured bedrock zones are imaged below the following distances along the model section: 160 -200, 280 -380 , and 530 -560 of the profile. These findings have good correlations with [Boamah and Koeberl, 2003] results from drill cores of BH1 and BH3. Moreover, the resistivities of the northern Bosumtwi suevites are well correlated with those of the Ries suevites in Germany [Ernstson, 1974;Pilkington et al., 1992].
The seismic refraction was also executed on profile 2. The layer model has also revealed three layers as shown in Figure 7 (b). The first layer has a thickness ≤ 3 with a P-wave average velocity of 600 / . It is then followed by a second layer of thickness varying between 4 -8 with an average velocity of 1900 / . The second layer, therefore, overlies the third layer with a P-wave velocity of 2300 / velocity. Thus, the low velocity of the first layer of profile 2 represents the presence of unconsolidated soil near the surface. The lithology of the second layer could be formed by clayey soil with little moisture content, while the third layer might be a fractured and shattered country rock of claystone.
The tomography which complements the layer models of profile 2 was further generated. Figure 7 (c) also imaged suevitic distribution within the subsurface. There is also a high-velocity contrast of 3000 -3900 / within the second layer. This high-velocity contrast occurs from the beginning of the profile up to about 380 m and around the end of the profile. The suevites were found within 10 m depth. The seismic and resistivity profiles did not cover the entire length extent of the suevite deposits.
The resistivity survey was further carried out along the Nyameani-Boamadumase road from South-West to North-East ( Figure 8). From the Bosumtwi geological map, this place has no suevites and was therefore chosen for comparison with other profiles which have been carried out in the suevite deposits. The survey was conducted along the road covering a distance of 280 m.
The model section of profile 3 indicates two layers within the subsurface separated by a boundary B; a little highly resistive near-surface is followed by a less resistive layer. The thickness of the high resistivity layer is 12 m and the resistivity is 300 . It is interpreted as dry metasediments comprising phyllite, metagreywackes, shales, and schist. The less resistive layer, whose resistivity is ≤ 300 Ω could be composed of soils with high moisture content. Since the resistivity model results do not show any less resistive zone near the surface, it reveals that this location does not host suevites. Therefore the resistivity results agree with the geological findings.
Profile 4 has a length of 160 m and was surveyed heading south ( Figure 9). Profiles 4 and 2 cross each other at 16 and 22 m respectively.
The model section of profile 4 indicates an image of nearly uniform low resistivity (≤ 300 Ω ) distribution within the subsurface. This probably indicates the presence of clayey soil. The model section does not show any subsurface strata; this means that this profile is likely to be in the same direction as subsurface layers.
In addition to the base-line, profile 5 ( Figure 10) was also measured moving southwards. Profiles 5 and 1 cross each other at 20 and 108 m respectively. The model section of profile 5 does not show a significant difference in resistivity distribution within the subsurface. It could be due to unevenly distribution of clayey soil.
The low resistivity zones around 4 m depth at the beginning and the end of profile5 were not well resolved.

Emmanuel Habimana et al.
However, the low resistivity area with resistivity (2 -25 Ω ) could be due to the presence of suevites. If this low resistivity area at the base is a suevite deposit, then it will be the deepest suevite deposit found in the Bosumtwi crater area.
The resistivity model section of profile 6 is also 160 m in length and indicates three layers ( Figure 11).
The first layer is slightly resistive due to the dry unconsolidated soil. It is then followed by a less resistive second layer which is likely dominated by clay and moisture towards the end and unevenly intrudes the first layer. The third layer is also slightly resistive; this region could possibly be the fractured claystone bedrock. At the beginning and at the end of the profile around 8 m depth, there are low resistivity zones (6 -25 Ω ), which are likely to be suevite deposits. Profile 7 covers 160 m and was surveyed westwards (Figure 12). Profiles 6 and 7 cross each other at 100 and 20 m respectively.

9
Mapping of Bosumtwi crater suevites    The model section, Figure 12, shows that the subsurface is formed by two layers. The first layer is less resistive, with a resistivity value less than 300 Ω . It is a result of unconsolidated soil with some amount of moisture. The second layer with slightly high resistivity between 300 and 1000 Ω starts from about 15 m depth. This layer could represent fractured claystone containing some moisture. At the crossing point of profiles 6 and 7 (around 32 m along profile 7), there is a low resistivity zone (6 -25 Ω ) which could be suevite deposits.

Conclusion
The seismic refraction and electrical resistivity results showed the suevite deposits and mapped out the subsurface strata as well as the fractured zones. Suevite deposits were found at depths ≤ 12 with a low resistivity range of 1.56 -25 Ω along all the profiles.
Generally, three layers were identified on the baseline; the top layer with slightly high resistivity (60 -400 Ω ) has a thickness ≤ 4 . It was found that this resistivity is controlled by unconsolidated soil and moist clay. The low resistivity regions ≤ 25 Ω which are mostly dominated by suevite deposits embedded in clayey soils was observed in the second layer of profiles 1 and 2. The very high and varying resistivities (6000 -25000 Ω ) were found in the third layer of both profiles beyond 12 m depth. The high resistivity is the result of the bedrock of claystone, while the lateral variation is controlled by fractured zones. Fractured zones were identified in the third layer of the resistivity model sections with slightly low resistivity.
Two layers were mapped out for profiles 3, 4, 5 and 6 whilst profile 7 was a three-layer model. Profile 3 was outside the demarcated suevites area and no suevites deposits were found on its resistivity model section.
The seismic refraction also delineated 3 layers of the subsurface and mapped out the suevite deposits. The nearsurface layer with thickness ≤ 4 has a P-wave velocity of about 600 / . The thickness of the second layer varies from 4 -8 m and the P-wave velocity varies between 1900 and 2000 / . This second layer is dominated by clayey soils. The third layer can be found beneath 12 m depth. Its P-wave velocity varies from 2300 to 2400 / . It represents a fractured and shattered basement of claystone.
The high-velocity contrast, 3000 -3900 / that was observed in the tomography model could be as a result of the suevite deposits and was found at depths ≤ 10 . The resistivity and P-wave velocity values of northern Bosumtwi suevites are in good correlation with an ultrasonic P-wave velocity of suevites in Chicxulub impact structure (Mexico) and Chesapeake Bay impact structure (USA) [Ernstson, 1974;Popov et al., 2014]. The study area of this project is covered by cocoa trees, and the surface is covered by cocoa leaves; thus, evaporation of the subsurface water is limited. Therefore, the suevites are likely characterized by saturated pores.