Two-dimensional resistivity imaging in the Kestelek boron area by VLF and DC resistivity methods

Journal of Applied Geophysics 82 (2012) 1–10

Contents lists available at SciVerse ScienceDirect

Journal of Applied Geophysics journal homepage: www.elsevier.com/locate/jappgeo

Two-dimensional resistivity imaging in the Kestelek boron area by VLF and DC resistivity methods Murat Bayrak ⁎, Leyla Şenel Istanbul University, Engineering Faculty, Department of Geophysical Engineering, Avcılar Campus, 34320 Istanbul, Turkey

a r t i c l e

i n f o

Article history: Received 3 October 2011 Accepted 28 March 2012 Available online 6 April 2012 Keywords: VLF Dipole–dipole Geoelectrical modeling Current density Boron deposit Kestelek

a b s t r a c t A VLF and DC resistivity investigation was conducted in the Kestelek area, western Turkey, to determine the twodimensional images of the boron deposits. The two-dimensional resistivity images were obtained by the inversion of tipper and resistivity data for VLF and DC resistivity methods, respectively. The VLF tipper data also were improved applying the Fraser and Karous & Hjelt (K&H) filtering to delineate the boundaries of the subsurface boron deposits. The main findings are: (1) moderate (>25 Ωm) and relatively high (>40 Ωm) resistivity zones in the two-dimensional models, which is mostly supported by the K&H real part of tipper as the negative current density peaks, may be interpreted as middle level of potatoes type colemanite and lower level of crystal type colemanite boron deposits inside the conductive units, respectively. (2) Transition from positive peaks (conductive zones) to negative peaks (resistive zones) in the K&H real part of tipper current density pseudosections may indicate the potential locations of the boron deposits. (3) Drilling well results obtained around two profiles of the study area are consistent with distribution of the resistive boron deposits in the two-dimensional resistivity models and K&H real part of tipper filtering images. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Turkey holds about a 72% share of the total boron deposits of the world. About 92% of boron minerals are exported and the rest is used domestically. Boron ore and its various minerals' demand is increasing around the world (Kar et al., 2006). Turkey has started research and development activities on boron. For this purpose, National Boron Research Institute, BOREN, was established in 2003 under the authority of Ministry of Energy and Natural Resources in order to organize and sponsor the scientific studies. The important borate minerals from a worldwide commercial standpoint are borax, ulexite, and colemanite. They are produced in a limited number of countries, dominated by Turkey and the United States, which together furnish about 90% of the world's borate supplies. The main borate districts of Turkey are Bigadiç, Kestelek, Sultançayır, Emet and Kırka (Kistler and Helvacı, 1994). Western Turkey is significant from the point of view of boron deposit exploration. Development of the boron potential of the Kestelek region was first initiated in the 1954 by Mineral Research and Exploration Institute of Turkey (MTA). In most of the studies particular attention was focused on stratigraphy and reservoir features. In contrast, two-dimensional resistivity imaging of the region, which could show resistivity characteristic of the boron deposits, to this day has not been sufficiently explored. This motivated us to image the ⁎ Corresponding author at: Istanbul University, Engineering Faculty, Department of Geophysical Engineering, Avcılar Campus, 34320 Istanbul, Turkey. Tel.: + 90 212 473 70 70/17835; fax: + 90 212 473 71 80. E-mail address: [email protected] (M. Bayrak). 0926-9851/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.jappgeo.2012.03.010

resistivity structure of the boron deposits in the region using VLF (Very Low Frequency) and dipole–dipole (DC) resistivity data. Boron is the fifth element of the periodic table and is the only electron-deficient non-metallic element. Boron transforms from a non-metal to a superconductor at about 160 GPa (Eremets et al., 2001). Boron is hard and the resistivity of the pure bulk boron is ~106 Ωcm for the rhombohedral structure (Buhro et al., 2003). Boron concentrations in rocks range from 5 ppm in basalts to 100 ppm in shales, and averages 10 ppm in the earth's crust overall (Woods, 1994). Total annual world consumption of borates is: insulation, fibreglass, and heat-resistant glass, 41% of boron consumption; ceramic and enamel frits and glazes, 13%; detergents, soaps, and personal care products, 12%; and micronutrients, 6% (Kar et al., 2006). Boron deposits can be detected by VLF and DC methods, since they produce strong variations in subsurface electrical resistivity. In the VLF and DC methods, resistivity is sensitive to high resistivity contrast in ore sites while VLF tipper parameters are sensitive to conductive structures of the subsurface. The ground resistivity is related to various geological parameters such as the mineral and fluid content, porosity and degree of water saturation in the rock (Carr, 1982). The application of VLF method for cost effective imaging and detection of near surface targets for prospection of ore deposits has been in use for over 10 years; some of examples include Eze et al. (2004), Frasheri et al. (1995), Hutchinson and Barta (2002), Ligas and Palmoba (2006), Liu et al. (2006), McCaffrey et al. (1995), Paterson and Ronka (1971), and Saydam (1981). Bayrak (2002) applied the VLF method to chrome area to inspect the effectiveness of the method. He indicates that the resistivity changes, Fraser, and K&H filtered data

2

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

consistent well with the extension of chrome ore, which correspond to known chrome mineralization Dunit/Harzburgite type rock and serpentinized rock boundaries. Thus, VLF might be also effective for the boron ore prospection. Several researchers have also used DC resistivity method for ore areas located in the Mt. Milligan copper gold porphyry deposit, Canada (Oldenburg et al., 1997), in shallow coal mine workings (Johnson, 2003), in the East Pennine Coalfield, UK (Chambers et al., 2007), in uranium mining sides (Legault et al., 2008; Ramalho et al., 2009) and in Landusky gold mine, Montana, USA (Rucker et al., 2009). Doyle (1990) also extensively discussed the gold exploration by electrical and VLF methods. Our objective was to produce two-dimensional resistivity images of subsurface boron deposits by employing VLF and dipole–dipole (DC) resistivity methods, using advanced two-dimensional inversion techniques in Kestelek boron deposits. Subsequently, both VLF and DC two-dimensional modelling results were interpreted in terms of the geological features and the possible boundary of the Kestelek boron field of western Turkey. 2. Geological setting The Kestelek boron deposit is 27 km southeast of the town of Mustafa Kemalpaşa in the province of the Bursa (Fig. 1). It was discovered accidentally during a survey of lignite deposits for the Mineral Research and Exploration Institute of Turkey (MTA) in 1954. The known borates of the Turkey were deposited in the lacustrine environment during Miocene when the volcanic activity occurred since Tertiary to Quaternary (Helvacı et al., 2004). In Fig. 2a, generalized Playa lake sedimentation model showing borate deposits formation in Neogene basins of western Turkey is indicated (Helvacı, 2004). The Miocene sediments which contain the borates in the Kestelek area rest unconformably on a Palaeozoic and Mesozoic basement complex. The Miocene sequence contains from bottom to top: basement conglomerate and sandstones; claystone with lignite seams, marl, limestone, and tuff; agglomerates and volcanic rocks; the borate zone comprising clay, marl, limestone, tuff and borates; and limestones with thin clay and chert bands. These sediments were deposited during a tectonically stable period accompanied by extensive volcanic activity. A yellow to brown coarse-grained tuff unit, up to a few centimetres thick, occurs within the borate zone. This sequence is capped by loosely cemented Pliocene conglomerate, sandstone and limestone (Helvaci, 1994). The borate minerals occur interbedded with clay as nodules or masses and as thin layers of fibrous and euhedral crystals. Colemanite, ulexite, and probertite predominate and sparse hydroboracite is also present locally (Fig. 2b) (Helvaci and Alonso, 2000). The properties of borate minerals recorded from the Kestelek area are shown in Table 1 (Helvaci and Alonso, 2000; Koçak and Sözügüzel, 1989). Probertite, which forms in the same chemical environment as ulexite in the Kestelek deposit, indicates a period of higher temperature within the ephemeral lake. Calcite, dolomite, quartz, zeolite, smectite, illite, and chlorite are accessory minerals. Borate minerals in the Kestelek deposit, like the tuffs and clays in the borate zone, are characterized by very low concentrations of As, Sr and S relative to the other deposits in Turkey, and certain western U.S. borate deposits (Helvaci and Alonso, 2000). Volcanic rocks, tuffs and clays interbedded with the borates are the most likely sources for Ca, Mg, Na, Sr, B, As and S. In the Kestelek area, the extensive volcanic rock associations and tuff intercalations with the borates indicate that much of the sediment was derived from a volcanic terrain. Hydrothermal solutions, thermal springs and tuffs associated with local volcanic activity are thought to have been the source of the borates. The Kestelek deposit is characterized by high Ca (colemanite), very low Na, and very low chloride and sulphate (Helvaci, 1994). Smectite is the major clay mineral in the Kestelek borate mine, with 73 wt.% in the clay fraction (b2 μm). The Li2O content of the clay samples from the Kestelek mine varies between 0.15 and 0.17 wt.%. Whole-rock chemical

analyses of samples from the Kestelek mine waste dump show that they contain 0.22 wt.% Li2O. These lithium amounts indicate that the clays associated with borate deposits are potential lithium resources, and that they may be considered for economic use in the near future (Helvacı et al., 2004). 3. VLF and DC resistivity survey The VLF and DC resistivity survey, carried out during the summer of 2009, was made along two profiles across boron deposits in the study area, illustrated in Fig. 1a, b and c. The VLF data was achieved using the Scintrex EDA-OMNI instrument. The instrument is microprocessor controlled and facilitates automatic tuning; digital data capture and signal stacking. The radio magnetic field is recorded by the three orthogonal coils mounted in a cylindrical housing with a pre-amp signal circuit and the electric field is measured perpendicular to the magnetic field with two probes in contact with the ground. The VLF radio station transmitted from Germany (Rhauderfehn, DH038) at 23.4 kHz was used for two profiles. This radio station is quite stable. The DC resistivity profiling measurements were made with a microprocessor-controlled signal averaged METZ equipment using dipole–dipole electrode array. In the signal averaged system, two consecutive readings (in both the forward and reverse directions) are taken automatically at each measurement point and then compared with each other. When acceptable measurements were obtained these readings are averaged. So, the signal-to-noise rate is enhanced. Dipole–dipole array can be used effectively when the instrument has comparatively high sensitivity and very good noise rejection circuitry (Loke, 2011). The pseudosection plotting for two-dimensional DC resistivity survey is also an important tool for data quality estimation (Dahlin and Loke, 1998), however it gives a distorted picture of the subsurface because the shapes of the contours depend on the type of array used as well as the true subsurface (Loke, 2011). Figs. 3g and 4g show the measured DC apparent resistivity pseudosections for profile 1 and profile 2, respectively. Good quality data usually show a smooth variation of apparent resistivity values in the pseudosection (Loke, 2011). Some of well locations (sk73 and sk2) near the VLF and DC resistivity stations are shown in Fig. 1a. The lithological sections of these wells are also presented in Fig. 5a and b, respectively. 3.1. VLF method Powerful military radio transmitters operating in the 15–30 kHz in the world are sources of the VLF method. A theoretical background of this method extensively discussed in the literature (McNeill and Labson, 1991; Reynolds, 1997; Wright, 1988). In this method, horizontal electric (EX) and magnetic (HY), and vertical (HZ) magnetic field of electromagnetic components are measured. For a two dimensional structure, the x-direction can be considered to be the direction of the geological strike and preferably the direction to the VLF transmitter used. The y axis is the profile direction. Tipper is the ratio between the
 HZ vertical and horizontal magnetic fields scalar tipper ¼ TYsca ¼ H Y and depends on the near surface characteristics of the Earth. Vertical magnetic field (HZ) results from entirely lateral changes in electrical conductivity (Oskooi and Pedersen, 2005; Pedersen and Becken, 2005). In the presence of a conductor in the underground, the total VLF field is elliptically polarized (Smith and Ward, 1974). The tangent of the tilt angle is a good approximation to the real part of tipper (Re (HZ/HY) = tan αx100 in percent) and the ellipticity is a good approximation to the quadrature part of tipper (Qu (HZ/HY) = εx100 in percent) (Paterson and Ronka, 1971). The real and imaginary components are expressed as a percentage of the total VLF transmitter's primary field. Vertical component of the magnetic field decreases at sites far from conductors. The real part of tipper is sensitive to low resistivity

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

3

Fig. 1. (a) Geological map of Kestelek boron deposit and locations of VLF and DC profiles with drilling log (modified after Koçak and Sözügüzel, 1989). Photographs showing VLF and DC measuring location of (b) profile 2. (c) Profile 1. Solid thin lines indicate the location where each photo was taken.

bodies. On the other hand, the quadrature part of tipper is sensitive to the variation of the earth electrical properties (Santos et al., 2006a). The apparent resistivity ρa and the phase angle ϕ between horizontal electric and magnetic field of the subsurface may be defined as a function of frequency (ƒ) as ρa ¼

  1  EX 2 in Ω m 5f HY 

ð1Þ

(Cagniard, 1953) −1

and ϕ ¼ tan

  ImðEX =HY Þ in degree: ReðEX =HY Þ

ð2Þ

Phase is 45° for homogeneous earth condition. However, in general, phase is greater than 45° for a conductive deeper layer and less than 45° for a resistive deeper layer (Arcone, 1979). The skin depth describes the penetration depth of VLF method and expresses as rffiffiffi ρ in meters; d ≈ 503 f

ð3Þ

where ρ is the resistivity in Ω m and ƒ is the frequency in Hz. Skin depth depends on the frequency and resistivity of the subsurface. VLF tipper data can be improved by applying Fraser (1969) and Karous and Hjelt (1983) linear filters that are the two techniques generally employed in tipper data processing. The Fraser filter converts the zero-crossing points of real part of tipper to peaks that show a correct location of the conductive zones under the peak along the profile. The shape of positive peaks changes with depth of conductors. Karous and Hjelt (1983) developed a statistical linear filter, based on Fraser (1969) and linear field theory of Bendat and Piersol (1968), to solve the integral equation for the current distribution, assumed to be located in a thin horizontal sheet of varying current density, situated everywhere at a depth equal to the distance between the measurement stations. By calculating the inverse filter at various depths (e.g., Δx, 2Δx, 3Δx), one can study the variation of current densities with depth. Thus, this filter technique provides relative current density (Ia) pseudosections which are derived from the real part of VLF tipper data at consecutive stations (H− 3, H− 2, etc.). The filter can be calculated as:

Ia ðΔx=2Þ ¼

2πð0:102H−3 −0:059H−2 þ 0:561H−1 −0:561H1 þ 0:059H2 −0:102H3 Δz

ð4Þ

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

4

Fig. 2. a) Generalized Playa lake sedimentation model showing borate deposits formation in Neogene basins of western Turkey and b) representation of the interpreted borate distribution (colemanite, ulexite, probertite) in the Kestelek area (Helvacı, 2004).

where Δz is the assumed thickness of the current sheet, Δx is the distance between the data points and also the depth to the current sheet, location of the calculated current density is beneath the center point of the six data points (Gürer et al., 2009; Karous and Hjelt, 1983). The filtered VLF tipper data can provide the interpreter with a useful standing point for numerical modelling. According to Benson et al.

(1997), lower values of K&H relative current density values may correspond to higher values of resistivity. Also, Ogilvy and Lee (1991) concluded that this filter can be used as a diagnostic tool to image the main characteristics of conductors. The detailed theoretical background of Fraser filter and K&H filter is reported by Fraser (1969) and Karous and Hjelt (1983). 3.2. DC resistivity method

Table 1 Properties of borate minerals recorded from the Kestelek borate deposit (Helvaci and Alonso, 2000; Koçak and Sözügüzel, 1989). Borate minerals

Oxide formula

Crystal system

%B2O3

Hardness

Specific gravity

Colemanite Ulexite Probertite Hydroboracite

Ca2B6O11.5H2O NaCaB5O9.8H2O NaCaB5O9.5H2O CaMgB6O11.6H2O

Monoclinal Trigonal Monoclinal Monoclinal

50.8 43 49.6 50.5

4 2.5 3.5 3

2.42 1.96 2.14 2.16

Two-dimensional DC resistivity measurements are commonly used to reveal the true electrical resistivity properties of the subsurface. DC resistivity method is based on measuring the electrical resistivity distribution of the subsurface using DC current (I, mA) transmitted into the ground, by two electrodes (A and B), and measuring the potential difference (ΔV, mV) between a second pair of electrodes (M and N). In homogenous ground, the depth of penetration is proportional to the separation between the electrodes, and varying the

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

electrode separation provides information about the stratification of the ground (Dahlin, 2001). Several electrode arrays are available for subsurface resistivity studies. Most surveys are carried out using conventional arrays such as the Wenner, Schlumberger, dipole– dipole and pole–pole arrays. Each type of array has its advantages and disadvantages, and the general characteristics of these arrays are generally well known (Dahlin and Zhou, 2004; Loke, 2011; Loke et al., 2010; Seaton and Burbey, 2002). They can be selected depending on the geology, the planned depth of investigation, required horizontal and vertical resolution and field feasibility (Nguyen et al., 2007; Reynolds, 1997). The type of array can also influence the final resistivity image, as each array presents different sensitivities, depth of investigation and resolution power (Nguyen et al., 2005). In this study, dipole–dipole array is used for determining the two-dimensional images of the boron deposits in the study area. Sasaki (2006) suggested that dipole–dipole array is more suitable for resolving complex structures when the instrumental accuracy is high. In the case of the dipole–dipole array, the apparent resistivity is calculated as: ρa ¼ πnðn þ 1Þðn þ 2ÞaR:

ð5Þ

The imaging resolution of dipole–dipole better than other arrays, particularly for the location of vertical and dipping structures (Dahlin and Zhou, 2004). Also, earth models generated from the dipole–dipole surveys consistently indicated more geologic detail and greater depth of investigation than the other arrays, although it is more susceptible to noise contamination (Dahlin and Zhou, 2004; Seaton and Burbey, 2002). The reader is referred to Keller and Frischknecht (1966), Koefoed (1979), Parasnis (1962), and Zhdanov and Keller (1994) for more detailed theoretical background of DC resistivity method.

5

along profile 1 and profile 2 (Figs. 3e and 4e) have an RMS misfit of 2.16 and 2.94, respectively. The dipole–dipole (DC) resistivity data were inverted using the 2-D non-linear robust (L1-norm) inversion method (Loke et al., 2003), which is less sensitive to the noise levels and produce fewer artefacts in the inverted resistivity models (Dahlin and Zhou, 2004). The finite element, which uses the standard first-order triangular elements (Silvester and Ferrari, 1996), is used to calculate the apparent resistivity values (Loke et al., 2003). Robust (L1-norm) inversion method attempts to minimise the sum of the absolute values of the spatial changes in the model resistivity (Loke et al., 2003). The robust inversion (L1-norm) generally gives better imaging results than the smoothness-constrained least-squares (L2-norm) inversion (Loke and Dahlin, 2002), with both noise-free and noise-contaminated data (Dahlin and Zhou, 2004). Additionally, Loke et al. (2003) and Olayinka and Yaramancı (2000) made a comparison between inversion results with smooth and robust inversion, and they pointed out that robust inversion (L1-norm) method gives significantly better results where there are sharp boundaries. Seaton and Burbey (2002) have shown that the resistivity earth model resulting from the robust inversion has more linear and relatively sharper resistivity zone boundaries. The better results can also be obtained by using narrower model cells for dipole–dipole array which is more sensitive to large resistivity variations near the surface (Loke, 2011). Hence, the width of model cells was set at half the unit electrode spacing for profile 1 and profile 2. The initial models were a 100 Ωm half space. The final dipole–dipole (DC) resistivity models obtained along profile 1 and profile 2 (Figs. 3f and 4f) have an RMS of 1.72 and 2.2, respectively. The RMS values for DC models were computed after 5 iterations, with very good fits between the measured and inverted data (Figs. 3g, h and 4g, h). DC resistivity models have not changed much with more iteration. In the measured DC pseudosections for two profiles, there are reasonable agreements between apparent resistivities and modelled twodimensional structures (Figs. 3f, g and 4f, g).

4. 2-D inversion of the VLF tipper and DC resistivity data 5. Results and discussion In order to reveal the near-surface parts of the area, twodimensional inversion of the VLF tipper and dipole–dipole (DC) dataset was undertaken. The quantitative interpretation of single frequency VLF tipper data was examined by several authors (Beamish, 1994, 2000; Chouteau et al., 1996; Kaikkonen and Sharma, 1998), who showed that quite detailed information about the subsurface resistivity distribution can be obtained from regularized inversion of that data. In this work, two-dimensional inversion of the VLF tipper data was undertaken using a 2-D regularized inversion approach based on Sasaki (1994, 2001). The effectiveness of the inversion is discussed by Santos et al. (2006a,b), and they suggest that, the useful 2-D resistivity models can be obtained, inverting single frequency tipper data, when an adequate value for the background resistivity is available. The background average resistivity can be collected using the VLF resistivity data or electrical resistivity methods. Beamish (2000) has demonstrated that when single frequency VLF data are collected at a high lateral density (1 to 5 m), the measurements can be used to infer the main elements of the subsurface resistivity distribution. In our work VLF data are collected at a high lateral density and we have background resistivity at each area, so it can be expect valuable 2-D resistivity models for the boron deposit in the Kestelek area. The final VLF two-dimensional resistivity model was determined by comparing tipper parameters calculated from models with estimates provided by the field data. RMS misfit is the quality of the fitting between the observed and the synthetic data of the model. The low RMS can indicate a trustworthy geoelectrical model. The VLF tipper data were inverted using a 10 Ωm half space as initial models for profile 1 and profile 2. A floor error of % 2 for the VLF real and imaginary tipper components was assumed for all models. The final VLF resistivity models obtained

5.1. Profile 1 We performed VLF and dipole–dipole (DC) resistivity measurements on the same profile in the northeastern part of the boron deposit in the Kestelek area, providing an opportunity for determining the geophysical responses on the boron deposits (Fig. 1a, c). The elevation of this profile is ~ 49 m. A drilling well (sk73) at the 165th m distance of this profile with depths maximum to 49 m were obtained, revealing a boron deposits in the Kestelek area (Koçak and Sözügüzel, 1989). VLF radio station, site interval and profile length were 23.4 kHz, 6 m and 202 m, respectively. DC resistivity data were collected in the field using dipole–dipole array on a 20 electrode (190 m length) with the dipole length (a) of 10 m and maximum dipole spacing (n) of 6. Values of n spacing greater than 6 can cause the signal strength to diminish rapidly below the background noise levels and the resolution of the resistivity meter (Seaton and Burbey, 2002). The two-dimensional resistivity models obtained by the inversion of VLF tipper and dipole–dipole (DC) resistivity data for profile 1 are shown in Fig. 3e and f, respectively. In georesistivity models (Fig. 3e and f), marl banded clay deposits at 15–25th m, 40–55th m, 100– 150th m and 165–195th m are represented by low resistivity values (b8 Ωm). Conductive structures (b8 Ωm) are more evident and widespread in the 2-D dipole–dipole (DC) resistivity model. Fraser filtering of real part of VLF tipper indicates the positive peaks on around these conductive zones (Fig. 3c). The blue colored area at 50–75th m (A anomaly) of the profile 1 indicates a moderate resistive zone (>25 Ωm), extend to maximum depth of 7 m, overlaying the more conductive (b8 Ωm) units of the marl banded clay. This moderate resistive zone may be interpreted as middle level of potatoes type

6

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

Fig. 3. Profile 1 in the Kestelek boron deposit area. (a) VLF apparent resistivity in Ω m. (b) Phase in degrees. (c) Real part of tipper (solid rectangle) and Fraser filtering of real part of tipper (solid circle). (d) Pseudosection of current density (%) using real part of tipper. (e) The two-dimensional resistivity model obtained by the inversion of VLF tipper data. (f) The two-dimensional resistivity model obtained by the inversion of dipole–dipole (DC) resistivity data. (g) Measured apparent resistivity pseudosection for dipole–dipole array. (h) Calculated apparent resistivity pseudosection derived from the two-dimensional DC resistivity model.

colemanite boron deposits inside the conductive units of the marl banded clay. The area between 150–165th m (B anomaly) of the profile 1 indicates a relatively high resistive zone (>40 Ωm), extend to maximum depth of 20 m, beneath the more conductive (b8 Ωm) units of the marl banded clay. The principal borate mineral is colemanite with ulexite, probertite and sparse hydroboracite in the Kestelek boron deposit (Helvaci and Alonso, 2000). Ulexite is less hard

and less dense than the colemanite (Table 1), therefore, relatively high resistive zones (>40 Ωm) in the georesistivity models may be interpreted as moderate amount of lower level of crystal type colemanite boron deposits with higher amount of ulexite and probertite inside the conductive units of the marl banded clay under profile 1 as indicated by Koçak and Sözügüzel (1989). The consistency, in terms of the resistive and conductive zones, between VLF

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

7

Fig. 4. Profile 2 in the Kestelek boron deposit area. (a) VLF apparent resistivity in Ω m. (b) Phase in degrees. (c) Real part of tipper (solid rectangle) and Fraser filtering of real part of tipper (solid circle). (d) Pseudosection of current density (%) using real part of tipper. (e) The two-dimensional resistivity model obtained by the inversion of VLF tipper data. (f) The two-dimensional resistivity model obtained by the inversion of dipole–dipole (DC) resistivity data. (g) Measured apparent resistivity pseudosection for dipole–dipole array. (h) Calculated apparent resistivity pseudosection derived from the two-dimensional DC resistivity model.

8

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

mode. Fischer et al. (1983) and Gürer et al. (2009) reported that, VLF resistivity and phase become stronger in H-polarization mode in the presence of high resistivity contrast with the adjacent or lower layers. So we can expect strong VLF resistivity and phase responses in transverse magnetic (TM) mode. Also, the high phase values (>45°) at 50–75th m (A anomaly) and the low phase values (b45°) at 150–165th m (B anomaly) is a good example of response of a deeper conductive layer and a deeper resistive layer, respectively as supported by two-dimensional VLF and dipole–dipole (DC) resistivity models (Fig. 3e and f). The K&H real part current density pseudosection (Fig. 3d) shows a negative inflection (peaks going from negative to positive) in around A and B anomalies that may indicate a transition from resistor (probable boron ore) to a good conductor region (marl banded clay) as supported by two-dimensional VLF and dipole–dipole (DC) resistivity models (Fig. 3e and f). In accordance with the Benson et al. (1997), resistive zones as shown in the two-dimensional models (A and B anomalies) are consistent with K&H lower relative current density anomalies. The lateral boundaries of the probable boron ore may be determined quickly in the area by following these inflections. 5.2. Profile 2

Fig. 5. The lithological sections of the study area (Koçak and Sözügüzel, 1989) for boron deposit drill holes sk73 and sk2 (locations as shown in Fig. 1a).

and dipole–dipole (DC) two-dimensional resistivity models for profile 1 is an interesting finding. The depth penetration of VLF is much deeper than that of the dipole–dipole (DC) resistivity measurements. Therefore, VLF is affected by a larger volume of the subsurface, so that the likelihood of influences by very resistive boron deposits is increased, resulting in different resistivity ranges of the VLF and DC models. The depth of boron ores in the geoelectrical models is well-matched with the drilling well obtained by Koçak and Sözügüzel (1989) (Fig. 5a). These resistive zones are observed with the characteristic VLF resistivity responses with high apparent resistivity (~500 Ωm), and with high (>45°) and low (b45°) phase values (Fig. 3a and b). The direction of VLF horizontal electric field (23.4 kHz, Germany) for this profile is approximately perpendicular to the boron deposits describing the H-polarization

Profile 2 is placed at the southwestern part of the boron deposit in the Kestelek area (Fig. 1a, b). VLF and dipole–dipole (DC) resistivity responses on the boron deposits were obtained and the results were correlated with each other. The elevation of this profile is ~25 m. A drilling well (sk2) at the 24th m distance of this profile with depths maximum to 41 m were obtained, revealing a boron deposits in the Kestelek area (Koçak and Sözügüzel, 1989). VLF measurements were collected on the 80 m profile length with site spacing of 4 m using 23.4 kHz frequency transmitted from Germany. DC resistivity data were collected in the field using dipole–dipole array on a 20 electrode (76 m length) with the dipole length (a) of 4 m and maximum dipole spacing (n) of 6. It is often not advisable to go beyond 6, due to the resulting very low signal-to-noise ratios (Dahlin and Zhou, 2004). The two-dimensional resistivity models obtained by the inversion of VLF tipper and dipole–dipole (DC) resistivity data for profile 2 are shown in Fig. 4e and f, respectively. In 2-D VLF resistivity model (Fig. 4e), clay–marl–tuff deposits at 12–18th m and 40–76th meters are represented by low resistivity values (b8 Ωm). Very conductive zones (b4 Ωm) in the DC resistivity model (Fig. 4f) may be associated with a higher amount of the Li2O content of the clay from the Kestelek as indicated by Helvacı et al. (2004). Fraser filtering of real part of VLF tipper doesn't show significant positive peaks on around these conductive zones (Fig. 4c). The blue colored area at 0–10th m (AA anomaly), 12–20th m (BB anomaly), 20–24th m (CC anomaly), 25–34th m (DD anomaly), 42–48th m (EE anomaly) and 58–68th m (FF anomaly) of the profile 2 may indicate a relatively high resistive zones (>40 Ωm), extend to maximum depth of 20 m, in the very conductive (b4 Ωm) units of the clay–marl–tuff (Fig. 4e and f). The colemanite is the most common calcium borate mineral in the Kestelek deposit with the form of nodules. It coexists with ulexite, probertite and hydroboracite in the Kestelek area (Helvacı et al., 2004). Colemanite is harder and more dense the other borate minerals (Table 1), therefore, relatively high resistive zones (>40 Ωm) in the georesistivity models may signify higher amount of middle level of crystal type colemanite boron deposits with lower amount of ulexite and probertite inside the conductive units of the clay–marl–tuff under profile 2 as indicated by Koçak and Sözügüzel (1989). By comparing VLF and dipole–dipole resulting images on this profile, the consistency, in terms of the resistive and conductive zones, between two-dimensional VLF and dipole–dipole (DC) resistivity models is also an interesting finding. In addition, due to the nature of the VLF and the measurable field limitations in the DC method (such as the relatively short length electrode array), the depth penetration of VLF, which is influenced by a larger volume of the very resistive

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

9

characteristic strong VLF resistivity and phase responses, which are compatible with the 2-D resistivity models, with high apparent resistivity (>500 Ωm), and high (>45°) and low (b45°) phase values along two survey profiles. The high (>45°) and the low (b45°) phase values are a good example of response of a deeper conductive layer and a deeper resistive layer, respectively as supported by VLF and DC two-dimensional resistivity models. 4) By comparing VLF and dipole–dipole (DC) resulting images, the consistency, in terms of the resistive and conductive zones, between VLF and DC two-dimensional resistivity models is an interesting finding. The depth of boron ores in the geoelectrical models is also well-matched with the drilling wells. 5) The K&H real part current density pseudosections show that the lateral boundaries of the boron deposits may be determined in the area by following negative inflections that may indicate a transition from resistive boron ores to a good conductor zones. 6) Finally, our results suggest that combination of VLF and DC resistivity methods can be used as a relatively fast and inexpensive tool for two-dimensional resistivity imaging of buried boron deposits. We believe that the combination of VLF and DC methods overcomes the limitation of each method when it used alone.

boron deposits, is much deeper than that of the dipole–dipole (DC) resistivity measurements. Therefore, some resistive anomalies, extending to maximum 20 m depth, in the VLF model cannot be identified by DC resistivity method (Fig. 4e and f). The depth of boron ores in the geoelectrical models is also well-matched with the drilling well obtained by Koçak and Sözügüzel (1989) (Fig. 5b). A very high apparent resistivity (~900 Ωm) values, extending from 28 m to 80 m distances, are observed in the VLF resistivity responses with very high phase values (>45°) (Fig. 4a and b). The direction of horizontal electric field for used radio station is approximately perpendicular to boron deposits, so the observed VLF responses for this profile are close to the TM mode and strong VLF resistivity and especially phase responses can be expected (Fischer et al., 1983; Gürer et al., 2009). Very high VLF resistivity and phase responses in H-polarization mode also can be enhanced by very high resistivity contrast between resistors and surrounding medium (Fischer et al., 1983; Gürer et al., 2009). Very high phase values (>45°) also may indicate that the thickness of conductive layer is bigger than the penetration depth of the VLF signals. Also, the high phase values (>45°) at these distances is a good example of response of a deeper conductive layer, as supported by two-dimensional VLF and dipole–dipole (DC) resistivity models (Fig. 4e and f). The K&H real part current density pseudosection (Fig. 4d) shows a negative inflection (peaks going from negative to positive) in around AA and DD anomaly that may indicate a transition from resistor (probable boron ore) to a good conductor region (clay–marl–tuff) as supported by two-dimensional VLF and dipole– dipole (DC) resistivity models (Fig. 4e and f). In accordance with the Benson et al. (1997), resistive zones as shown in the two-dimensional models (AA, DD, EE and FF anomalies) are consistent with K&H lower relative current density anomalies (Fig. 4d, e and f). The possible lateral locations of boron deposits can be found quickly in the area by following these inflections.

The authors wish to thank the anonymous reviewers, and the Editor Klaus Holliger, for their constructive contributions and improvement of the manuscript. We extend our gratitude to M. Nihat Yaman, Habib Göz and Adnan Kadri from ETİ Kestelek boron works and Musa Kırcı from ETİ mine works general management. Many thanks to F.A. Monteiro Santos from Lisbon University-Portugal for supplying the code.

6. Conclusions

References

Although, Turkey has the largest borate reserves in the world, the two-dimensional resistivity structure of the boron deposits wasn't explored previously. We believe our study fills this gap. For this purpose, 2-D geoelectrical resistivity structures of the Kestelek boron area have been revealed using VLF and DC resistivity methods. The main results and finding of this study can be summarized as follows. 1) The results of the 2-D resistivity modeling indicate that boron deposits produce strong variations in subsurface electrical resistivity. Moderate (>25 Ωm) and relatively high (>40 Ωm) resistive zones in the georesistivity models, may be interpreted as moderate amount of middle level of potatoes type and lower level of crystal type colemanite boron deposits with higher amount of ulexite and probertite inside the conductive units of the marl banded clay under profile 1. On the other hand, relatively high resistive zones (>40 Ωm) may signify higher amount of middle level of crystal type colemanite boron deposits with lower amount of ulexite and probertite inside the conductive units of the clay–marl–tuff under profile 2. Some deeper resistive anomalies imaged in the VLF resistivity models cannot be identified by DC resistivity method due to the limitations of DC measurable fieldwork. 2) Very conductive structures (b4 Ωm), which could not be identified by VLF method, are well defined by dipole–dipole (DC) resistivity method. These very conductive zones (b4 Ωm) may be associated with a higher amount of the Li2O content of the clay from the Kestelek boron area. Also, Fraser filtering of real part of VLF tipper doesn't show significant anomalies on around conductive zones. 3) The direction of VLF radio station (23.4 kHz) along two profiles is approximately perpendicular to the boron deposits describing the H-polarization mode. Accordingly, we obtained quite good

Acknowledgments

Arcone, S.A., 1979. Resolution studies in airborne resistivity surveying at VLF. Geophysics 44, 937–946. Bayrak, M., 2002. Exploration of chrome ore in Southwestern Turkey by VLF-EM. Journal of the Balkan Geophysical Society 5, 35–46. Beamish, D., 1994. Two-dimensional, regularised inversion of VLF data. Journal of Applied Geophysics 32, 357–374. Beamish, D., 2000. Quantitative 2D VLF data interpretation. Journal of Applied Geophysics 45, 33–47. Bendat, J.S., Piersol, A.G., 1968. Measurement and Analysis of Random Data. Wiley, New York. Benson, A.K., Payne, K.L., Stubben, M.A., 1997. Mapping groundwater contamination using resistivity and VLF geophysical method — a case study. Geophysics 62, 80–86. Buhro, W.E., Chang, R.P.H., Lu, J.G., Ruoff, R.S., 2003. Electrical and mechanical properties of boron and metal-boride nanowires, and devices built from them. NSF Nanoscale Science and Engineering Grantees Conference, Arlington, Virginia. Cagniard, L., 1953. Basic theory of the magneto-telluric method of geophysical prospecting. Geophysics 18, 605–635. Carr, C., 1982. Handbook on Soil Resistivity Surveying. Centre for American Archaeology Press, Illinois, p. 676. Chambers, J.E., Wilkinson, P.B., Weller, A.L., Meldrum, P.I., Ogilvy, R.D., Caunt, S., 2007. Mineshaft imaging using surface and crosshole 3D electrical resistivity tomography: a case history from the East Pennine Coalfield, UK. Journal of Applied Geophysics 62, 324–337. Chouteau, M., Zhang, P., Chapellier, D., 1996. Computation of apparent resistivity profiles from VLF-EM data using linear filtering. Geophysical Prospecting 44, 215–232. Dahlin, T., 2001. The development of DC resistivity imaging techniques. Computers and Geosciences 27, 1019–1029. Dahlin, T., Loke, M.H., 1998. Resolution of 2D Wenner resistivity imaging as assessed by numerical modelling. Journal of Applied Geophysics 38, 237–249. Dahlin, T., Zhou, B., 2004. A numerical comparison of 2D resistivity imaging with 10 electrode arrays. Geophysical Prospecting 52, 379–398. Doyle, H.A., 1990. Geophysical exploration for gold — a review. Geophysics 55, 134–146. Eremets, M.I., Struzhkin, V.V., Mao, H.K., Hemley, R.J., 2001. Superconductivity in boron. Science 293, 272–274. Eze, C.L., Mamah, L.I., Israel-Cookey, C., 2004. Very low frequency electromagnetic (VLF-EM) response from a lead sulphide lode in the Abakaliki lead/zinc field Nigeria. International Journal of Applied Earth Observation and Geoinformation 5, 159–163. Fischer, G., Le Quang, B.V., Müller, I., 1983. VLF ground surveys, a powerful tool for the study of shallow two-dimensional structures. Geophysical Prospecting 31, 977–991. Fraser, D.C., 1969. Contouring of VLF-EM data. Geophysics 34, 958–967.

10

M. Bayrak, L. Şenel / Journal of Applied Geophysics 82 (2012) 1–10

Frasheri, A., Lubonja, L., Alikaj, P., 1995. On the application of geophysics in the exploration for copper and chrome ores in Albania. Geophysical Prospecting 43, 743–757. Gürer, A., Bayrak, M., Gürer, Ö.F., 2009. A VLF survey using current gathering phenomena for tracing buried faults of Fethiye–Burdur Fault Zone, Turkey. Journal of Applied Geophysics 68, 437–447. Helvaci, C., 1994. Mineral assemblages and formation of the Kestelek and Sultançayırı, borate deposits, Western Turkey. Proceedings of 29th International Geological Congress, Kyoto, Part A, pp. 245–264. Helvacı, C., 2004. Türkiye Borat Yatakları: Jeolojik Konumu, Ekonomik Önemi ve Bor Politikası. 5. Endüstriyel Hammaddeler Sempozyumu, 13–14 Mayıs, İzmir, Türkiye. Helvaci, C., Alonso, R.N., 2000. Borate deposits of Turkey and Argentina, a summary and geological comparison. Turkish Journal of Earth Sciences 9, 1–27. Helvacı, C., Mordoğan, H., Çolak, M., Gündoğan, İ., 2004. Presence and distribution of lithium in borate deposits and some recent lake waters of west-central Turkey. International Geology Review 46, 177–190. Hutchinson, P.J., Barta, L.S., 2002. VLF surveying to delineate longwall mine-induced fractures. The Leading Edge 21, 491–493. Johnson, W.J., 2003. Case histories of dc resistivity measurements to map shallow coal mine workings. The Leading Edge 22, 571–573. Kaikkonen, P., Sharma, S.P., 1998. 2-D nonlinear joint inversion of VLF and VLF-R data using simulated annealing. Journal of Applied Geophysics 39, 155–176. Kar, Y., Şen, N., Demirbaş, A., 2006. Boron minerals in Turkey, their application areas and importance for the country's economy. Minerals & Energy — Raw Materials Report 20, 2–10. Karous, M., Hjelt, S.E., 1983. Linear filtering of VLF dip angle measurements. Geophysical Prospecting 31, 782–794. Keller, G.V., Frischknecht, F.C., 1966. Electrical Methods in Geophysical Prospecting. Pergamon Press, Oxford. Kistler, R.B., Helvacı, C., 1994. Boron and borates, In: Carr, Donald D. (Ed.), Industrial Minerals and Rocks, 6th Edition. Society of Mining, Metallurgy and Exploration, Inc., pp. 171–186. Koçak, N., Sözügüzel, H., 1989. Kestelek Neojen Havzasının Jeolojik Etüdü. Etibank Maden Arama Dairesi Başkanlığı, Ankara. Koefoed, O., 1979. Geosounding Principles, 1. Resistivity Sounding Measurements. Elsevier, Amsterdam. Legault, J.M., Carriere, D., Petrie, L., 2008. Synthetic model testing and distributed acquisition dc resistivity results over an unconformity uranium target from the Athabasca Basin, northern Saskatchewan. The Leading Edge 27, 46–51. Ligas, P., Palmoba, M., 2006. An integrated application of geological–geophysical methodologies as a cost-efficient tool in improving the estimation of clay deposit potential: case study from South-Central Sardinia (Italy). Ore Geology Reviews 29, 162–175. Liu, H., Liu, J., Yu, C., Ye, J., Zeng, Q., 2006. Integrated geological and geophysical exploration for concealed ores beneath cover in the Chaihulanzi goldfield, northern China. Geophysical Prospecting 54, 605–621. Loke, M.H., 2011. Tutorial: 2-D and 3-D electrical imaging surveys. 150 pp., accessed Dec. 2011 at http://www.geoelectrical.com/downloads.php. Loke, M.H., Dahlin, T., 2002. A comparison of the Gauss–Newton and quasi-Newton methods in resistivity imaging inversion. Journal of Applied Geophysics 49, 149–162. Loke, M.H., Acworth, I., Dahlin, T., 2003. A comparison of smooth and blocky inversion methods in 2D electrical imaging surveys. Exploration Geophysics 34, 182–187. Loke, M.H., Wilkinson, P.B., Chambers, J.E., 2010. Fast computation of optimized electrode arrays for 2D resistivity surveys. Computers and Geosciences 36, 1414–1426. McCaffrey, R.J., McElroy, W.J., Leslie, A.G., 1995. Exploration of a lignite-bearing basin in Northern Ireland using ground magnetic and VLF-EM methods. Geophysics 60, 408–412. McNeill, J.D., Labson, V.F., 1991. Geological mapping using VLF radio fields. In: Nabighian, M.N. (Ed.), Electromagnetic Methods in Applied Geophysics, V2, Application, Part B. SEG, Tulsa, pp. 521–640.

Nguyen, F., Garambois, S., Jongmans, D., Pirard, E., Loke, M.H., 2005. Image processing of 2D resistivity data for imaging faults. Journal of Applied Geophysics 57, 260–277. Nguyen, F., Garambois, S., Chardon, D., Hermitte, D., Bellier, O., Jongmas, D., 2007. Subsurface electrical imaging of anisotropic formations affected by a slow active reverse fault, Provence, France. Journal of Applied Geophysics 62, 338–353. Ogilvy, R.D., Lee, A.C., 1991. Interpretation of VLF-EM in-phase data using current density pseudosections. Geophysical Prospecting 39, 567–580. Olayinka, A.I., Yaramancı, U., 2000. Use of block inversion in the 2-D interpretation of apparent resistivity data and its comparison with smooth inversion. Journal of Applied Geophysics 45, 63–81. Oldenburg, D.W., Li, Y., Ellis, R.G., 1997. Inversion of geophysical data over a copper gold porphyry deposit: a case history for Mt. Milligan. Geophysics 62, 1419–1431. Oskooi, B., Pedersen, L.B., 2005. Comparison between VLF and RMT methods. A combined tool for mapping conductivity changes in the sedimentary cover. Journal of Applied Geophysics 57, 227–241. Parasnis, D.S., 1962. Principles of Applied Geophysics. Methuen&Co, London. Paterson, N.R., Ronka, V., 1971. Five years of surveying with the very low frequency electromagnetic method. Geoexploration 9, 7–26. Pedersen, B.L., Becken, M., 2005. Equivalent images derived from very-low-frequency (VLF) profile data. Geophysics 70, G43–G50. Ramalho, E., Carvalho, J., Barbosa, S., Santos, F.A.M., 2009. Using geophysical methods to characterize an abandoned uranium mining site, Portugal. Journal of Applied Geophysics 67, 14–33. Reynolds, J.M., 1997. An Introduction to Applied and Environmental Geophysics. John Wiley and Sons Ltd., Chichester. Rucker, D.F., Glaser, D.R., Osborne, T., Maehl, W.C., 2009. Electrical resistivity characterization of a reclaimed gold mine to delineate acid rock drainage pathways. Mine Water and the Environment 28, 146–157. Santos, F.A.M., Almedia, E.P., Gomes, M., Pina, A., 2006a. Hydrogeological investigation in Santiago Island (Cabo Verde) using magnetotellurics and VLF methods. Journal of African Earth Sciences 45, 421–430. Santos, F.A.M., Mateus, A., Figueiras, J., Gonçalves, M.A., 2006b. Mapping groundwater contamination around a landfill facility using the VLF-EM method — a case study. Journal of Applied Geophysics 60, 115–125. Sasaki, Y., 1994. 3-D resistivity inversion using the finite element method. Geophysics 59, 1839–1848. Sasaki, Y., 2001. Full 3-D inversion of electromagnetic data on PC. Journal of Applied Geophysics 46, 45–54. Sasaki, Y., 2006. Resolution of resistivity tomography inferred from numerical simulation. Geophysical Prospecting 40, 453–463. Saydam, A.S., 1981. Very low-frequency electromagnetic interpretation using tilt angle and ellipticity measurements. Geophysics 46, 1594–1605. Seaton, W.J., Burbey, T.J., 2002. Evaluation of two-dimensional resistivity methods in a fractured crystalline–rock terrane. Journal of Applied Geophysics 51, 21–41. Silvester, P.P., Ferrari, R.L., 1996. Finite Elements for Electrical Engineers, 3rd ed. Cambridge University Press, Cambridge, UK. Smith, B.D., Ward, S.H., 1974. On the computation of polarization ellipse parameters. Geophysics 39, 867–869. Woods, W.G., 1994. An introduction to boron: history, sources, uses, and chemistry. Environmental Health Perspectives 102, 5–11. Wright, J.L., 1988. VLF Interpretation Manual. EDA instrument, Toronto, Ontario, Canada. Zhdanov, M.S., Keller, G.V., 1994. The Geoelectrical Methods in Geophysical Exploration. Elsevier, Amsterdam.