Environmental Challenges 5 (2021) 100220

Contents lists available at ScienceDirect

Environmental Challenges
journal homepage: www.elsevier.com/locate/envc

Correlation between field electrical resistivity and geotechnical SPT blow

counts at tropical soils in Brazil
Juliana Teles Diniz Gonçalves a, Marco Antonio Barsottelli Botelho b, Sandro Lemos Machado a,
Leonides Guireli Netto c,d,∗
a
Department of Geotechnics, Federal University of Bahia (UFBA), Salvador, Bahia State, Brazil
b
Department of Applied Geophysics, Federal University of Bahia (UFBA), Salvador, Bahia State, Brazil
c
Geosciences and Exact Sciences Institute (IGCE), São Paulo State University (UNESP), Rio Claro, São Paulo State, Brazil
d
Institute for Technological Research of the State of São Paulo (IPT), São Paulo, São Paulo State, Brazil

a r t i c l e i n f o a b s t r a c t

Keywords: The joint application of geophysical investigations, non-invasive methods, and geotechnical investigations, inva-
Electrical resistivity tomography (ERT) sive methods, in order to interpret the geological features in subsurface has been increasingly accepted, however,
SPT the correlation between the values collected by the geophysical methods and the direct investigative techniques
Correlation
have been developed in recent years. The obtained linear relationships are site specific and complementary studies
Soil resistivity
are needed to establish their validation and limitations for each site. In this study, Electrical Resistivity Tomog-
Tropical soils
raphy (ERT) was used with Standard Penetration Test (SPT) for geotechnical investigation at three hillslope sites
with different tropical soil properties in Bahia, Brazil. The results from thirteen boreholes were analyzed against
data from 2D inverted sections. Linear fitting successfully correlates resistivity and NSPT for one borehole. Cor-
relations of electrical resistivity and NSPT values were also performed using soils with similar compositions and
physical characteristics, which showed good results. The coefficient of correlation (R) was 0.94. In addition,
results the fitting slope coefficient seems to be dependent of soil genesis (lithology, texture and clay content
variation) scription obtained in SPT surveys.

1. Introduction plication of non-destructive methods associated with destructive meth-

ods (such as drilling and collecting soil samples) allows interpret in de-
Subsurface investigation is one of the most important phase of any tail geological elements, such as the stratigraphy and the presence of
civil engineering construction or development activity. The geological structural elements, like faults and fractures (Barros and Pinto, 1997;
subsurface conditions can be extremely complex, variable and subject Lundström et al., 2009; Solberg et al., 2012; Bortolozo et al., 2018;
to change with time (Hossain et al., 2018). The fact is that conventional Di Maio et al., 2020).
direct sampling methods, such as SPT, can determine different soil prop- In geotechnical studies, the SPT blows count, NSPT (number of blows
erties. However, because it is a method of direct investigation, even us- per 30 cm of the sampler penetration into the soil), is used as an indi-
ing a vast mesh of instrumentation, there will be areas with no data cator of soil shear strength and stiffness. Although there are some defi-
information, since direct data are collected in a timely manner. In this ciencies, most of them are related to human errors and lack of standard-
way, several points of data acquisition take a long time to be executed ization in rods and sampler. SPT is the most dominant in-situ test for
to get the entire view of the subsurface through the borehole data, in soil investigation in Brazil and many other countries in South America
addition to increasing the cost of the investigation. (Coutinho and Schnaid, 2010).
Therefore, the use of indirect non-destructive geophysical methods, The physics of electrical current flow in the subsurface soil suggests
such as electrical resistivity tomography (ERT), induced polarization that some parameters such as grain size distribution, mineralogy, degree
(IP), seismic refraction and Multichannel Analysis of Surface Waves of saturation, porosity and cementation control both soil shear strength,
(MASW), offers (used in a combination with direct methods) the op- here evaluated employing NSPT , and electrical resistivity (𝜌). In coastal
portunity to overcome the limitations and inherent problems of con- areas, the salinity of the soil interstitial fluid and its distribution in sub-
ventional sites investigation methods (Cosenza et al., 2006; Akin et al., surface can be a serious obstacle in the acquirement of good correla-
2011; De Vita et al., 2012; Tucker et al., 2015; Kirar et al., 2016). The ap- tions between 𝜌 than NSTP. The ion exchange capacity of the soil clay

∗
Corresponding author at: Geosciences and Exact Sciences Institute (IGCE), São Paulo State University (UNESP), Rio Claro, São Paulo State, Brazil.
E-mail addresses: [email protected] (M.A.B. Botelho), [email protected] (S.L. Machado), [email protected] (L.G. Netto).

https://doi.org/10.1016/j.envc.2021.100220
Received 11 May 2021; Received in revised form 18 July 2021; Accepted 21 July 2021
2667-0100/© 2021 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

fraction forms a mobile cloud of additional ions around each clay par- In terms of regional geology, the metropolitan region of the city
ticle. These ions facilitate the electrical current flow. Thus, in clayey of Salvador is inserted in the Salvador-Esplanada belt (Paleoprotero-
soils, electrical resistivity is always expected to be lower than sandy zoic age). This structure is the result of the subdivision of the Itabuna-
soils (Zhdanov and Keller, 1994). However, sandy soils are expected to Salvador-Esplanada orogen, formed from the collision of the Gavião,
present higher shear strengths, in way that 𝜌 and NSPT are expected to Jequié, Itabuna-Salvador-Curaçá and Serrinha continental plates, which
vary inversely (Siddiqui and Osman, 2013). make up the São Francisco Craton. These blocks when colliding resulted
Numerous studies have been performed to investigate the correla- in highly deformed and fractured rocks (Souza et al., 2003). Recent stud-
tion between the soil electrical resistivity and the SPT blows count, ies show that the geographical development of the city of Salvador is
NSPT , such as Giao et al. (2003), Oh and Sun (2008), Folle et al. (2008), limited by the faults Salvador and Iguatemi, and the crystalline base-
Abidin et al. (2014) and Hatta et al. (2015). The obtained results varied ment, composed predominantly of granulites and amphibolites, is ex-
widely, from weak to strong correlations. Braga et al. (1999) present posed in a large part of the city (Barbosa et al., 2005).
the first attempt to correlate NSPT and 𝜌 in Brazil. They reported The study areas are covered mainly by the Phanerozoic sedimentary
a poor exponential correlation with a low (0.3) determination co- rocks of the Reconcavo Basin, Barreiras Group and by the quaternary
efficient at Corumbata´ı Formation, S˜ao Paulo, Brazil. According sedimentary deposits.
to Sudha et al. (2009) there is no universal relationship between Area I (12o 51’ 01.05’’S; 38o 17’ 51.16’’W) is located about 40 km
𝜌 and NSPT , and specific correlations must be obtained for differ- from Salvador. It is inserted in the Domain of Quaternary Deposits, con-
ent locations. Long et al. (2012), for example, showed good correla- taining fluvial and marine sediments, which are rich in organic matter
tions between electrical resistivity and geotechnical clay parameters. and are associated to a local mangrove ecosystem (see Fig. 1).
Devi et al. (2017) presented 𝜌 values that were linearly correlated with Areas II (12o 52’ 31.80’’S; 38o 28’ 40.40’’ W) and III (12o 57’ 40.80’’ S;
NSPT in the investigation of a bridge foundation site. The correlation per- 38o 30’ 07.44’’ W) are located in the city of Salvador, whose continen-
formed from the simple linear regression in this study, showed that the tal part is segmented by the Salvador’s Geologic Fault in two urban
fitted linear slopes were influenced by the lithology and the clay con- parts: lower (graben) and upper (horst). The city’s morphostructure con-
tent. Devil et al. (2017) obtained values of determination coefficient, templates many slopes which are occupied by an intense irregular dis-
r2 > 0.85, were found. tribution of houses with ineffective and/or inadequate urban planning
In this paper, Electrical Resistivity Tomography (ERT) has been used policies. Due to its occupation pattern, the city presents an historic of
in association with Standard Penetration Test (SPT) for geotechnical in- numerous slope failures and many high landslide susceptibility areas,
vestigation at three sites with different soil properties in Bahia, Brazil. including areas II and III themselves (Vieira et al., 2005).
Results from thirteen boreholes were analyzed against data from five In terms of local geology, area II is mainly constituted of residual
2D inverted sections. Values of 𝜌 and NSPT were fitted using least square soils and sedimentary rocks of the Pojuca Formation (Recôncavo Rift),
method in regression analysis and the behavior of the fitting coefficients which include an integrated arrangement of siltstones clayey lenses with
were analyzed considering the different soil types identified in field. shale and multiple interleaving of medium to fine sandstone. In response
to the weathering mechanisms, an expansive residual clayey soil is com-
monly found locally, which is regionally called “massapê”. Close to this
2. Definition of the study area and geological-geotechnical
area, more steep slopes are found, formed by a superficial layer of Bar-
characterization
reiras Formation, overlying the granulite basement. Massapê samples
are known to present significant amounts of montmorillonite, smectite,
The study areas were not chosen at random. The present study de-
vermiculite and illite, trilaminar minerals with high specific surface area
fined the areas of study considering the support of the non-invasive
and cation exchange capacity (CEC) (Machado and Presa, 2000).
method proposed in this work in situations of construction of civil engi-
Area III is located in a slope in the uplifted block (horst) of the Sal-
neering.
vador Geologic Fault, comprising granulite-gneiss rocks, which form the
The main criteria for choosing the study area were:
city’s crystalline basement, and the clayey residual soil derived from the
(1) application of non-invasive investigations methods during the exe- weathering of these metamorphic-plutonic rocks. The presence of boul-
cution of geotechnical works. In area I, for example, the construction ders and heavily fractured exposed rocks is common in such slopes.
of a road was carried out. This type of study is interesting, as it aims The geological map of the city of Salvador, illustrated in Fig. 1,
to prove the support of the geophysical method used in this type makes it clear that much of the urban area is on the crystalline base.
of work. After the underground investigation phase, the area would This is one of the factors of the high levels of susceptibility soil ero-
receive material for the construction of a landfill. sion when subjected to weathering and erosion in the city, especially
(2) use of non-invasive investigations methods as an alternative to re- in hillside areas (areas II and III of the present study). Considering the
search techniques in areas with difficult access to invasive methods, lithotypes of this substrate, the soils developed under these conditions
such as SPT. Drilling campaigns often make use of large numbers of can be divided into 3 types, according to the classification of Brazilian
equipment in the field. Areas II and III of the present work are quite Agricultural Research Corporation (Embrapa, 2018): latosols, argisols
challenging. Both areas are located on hillside areas, with consider- and neossols. Fig. 2 shows in more detail the main soils in the study
able slope angle. Documents from the city of Salvador showed that area, specially the 3 mentioned.
landslides often occur in these areas. In this scenario, which is quite The urban area of Salvador and the metropolitan region is mostly
common in other regions of the world, it is practically not feasible composed of latosols. This type of soil in the region is quite thick (a few
to apply a vast drilling network. Therefore, the application of geo- meters), with a clay texture, good permeability, and porosity. The or-
physical methods is quite interesting, as the present study intends to thogranulites, referring to the granulite terrains of the crystalline base-
demonstrate. ment, are found at the base. Argisols, in turn, are more present in cer-
tain areas of the metropolitan region, including study area I. This type
The three study areas are in the Metropolitan Region of Salvador of soil has a higher clay content than latosols, being heavier and more
(Fig. 1), a portion of the eastern coast of the state of Bahia which belongs fluid when in saturated conditions (Guimarães and Vale 2012).
to the Recôncavo sedimentary basin. The Recôncavo Basin located in the The definition of the SPT’s points considered some criteria. First one,
northeastern of Brazil, defined by important geological strucutures, such the objectives of the investigation: in area I, for example, a road con-
as the Aporá High to the north and northwest, the Barra fault system struction project is being carried out, so the SPT’s were concentrated
to the south, the Maragogipe Fault to the west and the Salvador fault in the region of the crest of the highway (central axis). Another impor-
system to the east (Silva et al., 2012), where the present study is located. tant criterion was the operational conditions in the field: as areas II and

2
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

Fig. 1. Geographic and geological localizations of study areas (modified from Souza et al., 2003).

III are places with high declivity, it is often not feasible to carry out as if they were "connecting" the points. This type of arrangement makes
surveys at certain points. These conditions were desired in the present sense when it is intended to correlate the information of the investi-
project, since one of the main objectives is to show the help that geo- gations methods and can be observed in important geotechnical and
physical methods can offer to geotechnical investigations in engineering geophysical investigations works (Lundström et al., 2009; Groves et al.,
projects. 2011, Bortolozo et al., 2018).
Fig. 3 shows the layout of the study areas. In this scheme, the lo- In area I, as mentioned, 4 SPT’s were executed with 20 m spacing
cation of the 13 SPT’s and 5 ERT’s lines executed in the present study between them. The surveys are concentrated in the central region of the
were represented. All ERT lines were arranged close to the SPT’s points, highway installation. The ERT acquisition line was carried out in the

3
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

Fig. 2. Pedogenic map of the metropolitan region of the city of Salvador, Bahia (modified from IBGE, 2018).

central region of the road too. It should be noted that the topography of pler attached, samples were collected during drilling. Information on
the area was flat. the soil penetration index is determined by means of a standard weight,
In area II, 5 SPT’s and 2 ERT’s lines were executed at the top and bot- which falls at a standard distance to reach the ground. The standard
tom of the slope, practically parallel to each other. This condition was weight was a 63.5 kg (140 lb) hammer, which was dropped repeat-
imposed since there were irregular settlements in the area, which did edly from a height of 76 cm (30 inches) by driving the sampler into
not allow the ERT line to be installed perpendicular to the topography. the ground to a depth of 15 cm (6 inches).
In Brazil and other countries in South America, this type of situation The result of the penetration test is expressed by the resistance index
is common, which makes the present work quite challenging. Subse- to the dynamic penetration (NSPT ). The value of this index corresponds
quently, the layout of the ERT acquisition lines became interesting in to the number of strokes necessary for the 0.30 m spike of the sam-
this way, since they allowed the comparison with the results of area III, pler in the soil, after an initial spike of 0.15 m, called “seating drive”
where it was possible to execute the lines perpendicularly to the topog- (ABNT-NBR-6484, 2001). The interpreted results are used to estimate
raphy. the geotechnical engineering properties of the soil and, in the case of
Finally, in area III, 4 SPT’s were carried out. 2 ERT lines were ex- this study, were correlated with the electrical resistivity values of the
ecuted practically parallel to each other. Each of the ERT’s lines con- lithologies.
nected 2 SPT’s. In area III, as it was possible to execute the lines per-
pendicular to the topography, the topographic data becomes important
during the geophysical data processing step (Fig. 3). 3.2. Electrical resistivity tomography (ERT)

3. Material and methods The ERT technique obtains discrete images of the subsoil, through
various points at different depth levels. These data are interpolated and
3.1. Standard penetration test (SPT) provides continuous 2D images of the subsurface, making possible to an-
alyze the variation of the electrical resistivity laterally and with depth.
The SPT is a geotechnical investigation test widely used to obtain in- The analysis and interpretation of these tomograms allows to identify
formation about the stratigraphy of the subsoil, characterize the types of resistivity contrasts that are associated to lithological and water content
soils and their occurrence depths and the water level position, for exam- variations. Geoelectric measurements are based on the injection of a
ple. The SPT surveys consist of a repetitive process, carried out based on constant current into the ground by a resistivitymeter, through two cur-
three operations: drilling the hole, penetration testing and sampling. In rent electrodes and measuring the resulting voltage differences between
the present study, in addition to the information on the in-depth litholo- another two potential electrodes at the surface (Ward, 1990; Sass and
gies provided by the method, other information was desired: the soil Glade, 2008).
penetration resistance (ABNT-NBR-6484, 2001). The configuration of the electrodes used in ERT surveys allows the
For the investigation in the present study, the standards of Brazil- acquisition of data quickly compared to invasive techniques. The re-
ian Association of Technical Standards (ABNT-NBR-6484, 2001), were sult is a high data density, in the form of multi-electrode tomographic
adopted. In this way, the test consisted of drilling a borehole to the de- acquisition, improving subsurface coverage and achieving desirable res-
sired depth of investigation. Since the drilling rod has a split-spoon sam- olution levels.

4
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

Fig. 3. Sketch of study areas I, II and III (detail for the lines of acquisitions and SPT’s).

The location of survey lines was decided based on the position of the resolving complex structures than the pole‐pole array. Another study
boreholes’ performed previously in each area. Five ERT sections were conclusion is the pole‐dipole array gives somewhat less resolution than
measured in total. The inter-electrode spacing in ERT 1, ERT 4 and ERT the dipole‐dipole array but yields greater signal strength.
5 was 5.0 m. This spacing was defined considering the depth of investi- In a more recent study, which looked at the two main factors that
gation (shallow targets, less than 50 m, typical of geotechnical studies) affect data quality for measurements of direct current resistivity images
and the resolution of the investigation. By the same criteria and objec- for all common electrode arrays (pole-pole, pole-dipole, pole-bipole,
tives, in ERT 2 and ERT 3, this acquisition parameter was reduced to Wenner, Schlumberger, dipole-dipole, matrix 𝛾, Wenner-𝛽): electrode
2.5 m in order to increase the image resolution and better comparison spacing errors and errors correlated with the magnitude of the observed
with the SPT results, which are spaced 1 m deep. potential; Zhou and Dahlin (2003) showed that the magnitude of the
The choice of a particular ERT array is related to the expected re- spacing errors are quite different with these arrays, being largest for
sults. The geometry and resistivity of the investigated structures are im- dipole-dipole, Wenner-𝛽 and 𝛾-array surveys, for which a 10% in-line
portant factors that must be considered too (Martorana et al., 2009). spacing error may cause twice as large an error (>20%) in the observed
Sasaki (1992) in one of the pioneering studies on the factors that affect resistance or apparent resistivity, which in turn will produce some arte-
the resolution and accuracy of resistivity tomography, compared some facts in the inverted model.
commonly used ERT arrays, such as pole-pole, pole-dipole and dipole- The dipole-dipole array was utilized for all ERT sections. This array
dipole. This study showed the dipole‐dipole array is more suitable for enables studying the lateral variation of the electrical resistivity at dif-

5
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

Table 1
Data acquisition parameters for each ERT section.

ERT Area Electrode Array Electrode spacing (m) Section Length (m) RMS%

01 I Dipole- 5.0 210.0 7.8

02 II Dipole 2.5 42.5 4.5
03 II 2.5 42.5 3.9
04 III 5.0 70.0 6.1
05 III 5.0 60.0 5.7

ferent depths. This arrangement has the current (AB) and the potential on a cell scheme. The 2-D model used by the inversion program is deter-
(MN) dipoles, which are most commonly placed linearly on the ground. mined automatically and its consists of a number of rectangular blocks
The space between A and B is equal to the space between M and N both loosely tied to the distribution of the data points in the pseudo section
of which remain unchanged during data acquisition. (Loke and Barker, 1996; Loke, 2013). A forward modeling subroutine
Because it is a traditional electrode array, with numerous application is used to calculate the subsurface apparent resistivity and a non-linear
cases, the DD array has known properties, such as good lateral imaging least-squares optimization technique. The pseudo-sections of apparent
capability, desirable in a study that aims to delimit geological layers resistivity are converted to a subsurface electrical resistivity distribution
(Moreira et al., 2016). The parameters used, such as the length of the using a quasi-Newton optimization technique.
acquisition line, considered factors of the study area. Area II, for exam- In order to analyze the quality of each ERT profile, the RMS (Root
ple, had adverse factors, such as irregular settlements that made data Mean Square) error between the measured and calculated model were
acquisition difficult. All parameters were established to choose the best used as an inversion quality factor (Sudha et al., 2009) and is given as:
configuration for a good resolution and reliability of the survey. Details √
√ 𝑁 ( )
of data acquisition parameters for each section are given in Table 1. √1 ∑ 𝜌𝑖𝑚𝑒𝑠 − 𝜌𝑖𝑐𝑎𝑙𝑐
The SYSCAL PRO resistivimeter (Iris Instruments) was used for the 𝑅𝑀𝑆 = √ (1)
𝑁 𝑖=1 𝜌𝑖𝑚𝑒𝑥
ER measurements. This device combines transmitter and receiver in a
single module for environmental and engineering geophysical studies, where 𝜌imes and 𝜌icalc are the measured and the calculated apparent
powered by a 12 V battery. It is a multi-channel instrument (10 channels resistivity at ith data point, respectively and N is the total number
receiver), which allows multiple potential dipoles to be used simultane- of points. The RMS percent errors varied between 3.9% to 7.8% (see
ously to obtain the resistivity measurements. Table 1).
The electrical resistivity values and their respective depths were plot-
3.3. Characteristics of the study areas: field work
ted in 2D sections using a free open-source interactive data-visualization
Area I, located on a road, was chosen for investigation because pre- tool. The Data Explorer Software allowed to explore the data in highly
vious geotechnical projects occurred there. As it is an area with the pres- visual and interactive ways. The software was used to define a unique
ence of soft soils, the geotechnical work consisted of removing this mate- color/resistivity scale for all ERTs.
rial followed by a process of compacting a new, more resistant material
(artificial fill). This scenario makes the application of geophysical inves- 3.5. Correlations between resistivity and NSPT
tigations associated with geotechnical investigations more interesting,
since in this case it is known what type of material is being inserted in To investigate the correlation and relationship between geotechni-
the first layers, facilitating the correlation between the data. The SPT cal and geophysical parameters the least square method was applied.
surveys and the ERT lines were carried out along the main axis of the First, a qualitative analysis is carried out between the electrical resis-
artificial fill used to build the road. tivity values obtained by the ERT measurements and the lithology
Areas II and III had the most critical and challenging scenario for found in SPT’s surveys. At each borehole, linear regressions were con-
the project. Both areas are located near the fault of Salvador in areas ducted employing the resistivity values derived from ERT sections at the
of hillsides. The historical survey carried out evidenced the report of same depth where NSPT were acquired.
countless landslides in these areas over the years. Another characteris- Further, data from nearby boreholes with the same soil composition
tic of these areas that hindered the execution of the works is the fact that (and position relative to water table were gathered at each study area,
both are located close to the historic center of Salvador. This neighbor- making possible to establish correlations for each soil according to their
hood is made up of streets and architectural monuments from the time of geological and geotechnical properties.
Colonial Brazil, such as museums, shops, cultural centers and numerous According to Sudha et al. (2009), the relationships between those
churches. In addition to the difficulty of accessing the area for carrying variables are site-specific and once they are well establish for a particu-
out large drilling campaigns, the use of geophysical methods is less in- lar location, soil strength can be determined from ERT results. Further-
vasive to the environment, being interesting due to the cultural issues of more, analysis were performed trying to understand the behavior of the
these study areas. Areas II and III still have a very common social prob- linear fitting coefficients and their sensibility to lithology variations and
lem: the presence of irregular settlements. In developing countries, like clay content, as observed by Sudha et al. (2009), Groves et al. (2011) and
Brazil and other countries in South America, the presence of unregulated Devi et al. (2017). In this case the fitting parameters were analyzed
buildings in hillside areas is quite common. against the average NSPT and average resistivity.
The geological conditions in the two areas are different. While in
area II residual soil and sedimentary rocks of the Pojuca Formation are 4. Results and discussion
found, in area III there is the presence of clayey residual soil derived
from the weathering of these metamorphic-plutonic rocks, allowing to 4.1. Geotechnical investigations
apply and analyze the methodology proposed by the present study in
different geological contexts. Thirteen SPT probes were performed: four SPT’s in area I, five SPT’s
in area II and four SPT’s in area III. The results are shown in Fig. 4 below.
3.4. Data processing and inversion
The SP-56, SP-60, SP-62 and SP-64 were executed in area I along
The resistivity inversion was achieved using RES2Dinv. This soft- central axis of the road. The SPT’s reached maximum depths close to
ware performs the smoothing inversion process using algorithms based 11 m and it was possible to identify four different lithologies in the

6
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

Fig. 4. Results of SPT soundings and the corresponding geological interpretations.

area. In the first 3 surveys it was possible to identify the artificial fill ever, in SP-56 and SP-60 this material has a layer of approximately
used to build the road, composed of a clayey silt with fine sand. NSPT 1 m thick with a material of less resistance to blows, classified as sandy
values for this lithology were between 1 and 6, consistent with this silt with gravel. Area I also has (identified by the SP-64 survey), inter-
type of material. The next layer has a predominance of material clas- calated layers of soft clay (silty clay with organic matter) at different
sified as saprolite (sandy silt with gravel and rock fragments). How- depths.

7
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

The SP-201, SP-202, SP-204, SP-205 and SP-206 executed in area Table 2
II obtained possibly the best results for correlation. This is because the Layers description and their respective resistivity range.
contact between the different lithologies is well marked and, unlike that Resistivity Range (Ωm) Layer Description
found in the other study areas, in these SPT’s there are few thin layers,
0–15 Shale or Soft clay with organic matter
which facilitates the correlation with the ERT data. The first layer was
15–45 Massapê to shale rock
classified as a sandy silt with clay (landfill layer), and it is found in 45–75 Unconsolidated sediments
the five SPT surveys in the area. This material presented NSPT values 75–200 Sandy silt with clay to clayey silt
varying between 12 and 18. The layer below composed by saprolite 200–2000 Granulite rock alteration
>2000 Granulite rock
(sandy silt with gravel and rock fragments) has higher values of NSPT ,
varying between 18 and 35. In area II, an expansive residual clayey soil
was also found at greater depths. This material, known regionally as
Massapê soil, reached maximum blows values of 17.
Massapê soil is a vertisol commonly found in the coastal region of areas were close. The surface layers (up to 3 m deep) did not present
northeastern Brazil. In tropical regions, which have well-defined dry and similar lithologies and pedogenetic elements, however, the electrical re-
humid seasons, it forms through the decomposition of granite, gneiss sistivity values for these materials (shale or soft clay with organic mat-
and in some cases limestone (Silva et al., 2018). This vertisol is widely ter and sandy silt with clay) showed similar electrical resistivity values.
studied in the region due to its good fertility for several crops. It is com- Some lithologies, such as sandy silt with clay and fragments of sandstone
pound as clayey to very clayey soil with a clay content like the mont- found in area II, for example, and clayey silt with rock fragments found
morillonite group, which presents characteristics of contraction and ex- in area III, have a similar composition, and presented similar electrical
pansion as a function of its humidity content. Because of this, in the wet resistivity values.
season, the massapê soil has a sticky consistency and in the dry period Resistivity distribution at Area I (Fig. 5) indicates the unsaturated
it becomes rigid (Lima et al., 2015; Lago et al., 2018). This vertisol is near surface soil represented by 45–75 Ωm. Below 10 m, the resistiv-
important due to the large presence in Brazil and other countries, such ity increases with the presence of granulite fragments/alteration (75–
as Australia and India (Dudal and Eswaran, 1988; FAO, 2014). 2000 Ωm) until it reaches the granulite rock (>2000 Ωm), featuring a
In area III, two layers can be interpreted: the first, with NSPT values high resistivity zone. The information is consistent with SPT’s surveys
less than 5, was classified as sandy silt with organic matter, represent- carried out in the area. Some SPT’s (SP 56, SP 60, SP 62 and SP 64) did
ing the landfill layer. The layer below presented NSPT values greater not reach depths greater than 15 m. The most in-depth information was
than 15 and refers to saprolite (clayey silt with rock fragments). Note obtained through ERT’s carried out throughout the area.
that SPT 15 was carried out at a point with a higher topography com- Results from ERT 02 and ERT 03 (Fig. 6) indicates that the resistiv-
pared to other geological surveys carried out in the area. In this case, as ity distribution at Area II decreases with depth. At the top, there is an
the investigation depth of SPT 15 reached almost 5 m, it was not pos- unconsolidated near surface layer (45–75 Ωm) with localized high re-
sible to reach the saprolite layer at this point. In the three study areas, sistivity inhomogeneities, which are due to the presence of sandstones
the ERT lines were defined from the execution of SPT’s. fragments (75–200 Ωm) at 4.0 m (ERT 02) and 2.0 m (ERT 03). Below
9.0 m, the resistivity decreases with the increase in clay content until it
4.2. Electrical resistivity tomography results reaches the shale rock (<15 Ωm).
Note that the electrical resistivity values found for Area II are
The results of the geophysical surveys used in this study were ana- lower and vary less than the values found in Area I. This type of be-
lyzed individually for each area (I, II and III) and subsequently inter- havior was expected and desired in the present study, since the litholo-
preted to obtain the resistivity values of the soils and rocks prospected gies found in the two areas are different. While Area I present values
by the SPT surveys. The inverted resistivity-depth models are shown in consistent with the sandy material used for grounding the area, Area
Figs. 5 and 6. II is a region where residual soils and sedimentary rocks from the Pojuca
The resistivity range were established with different colors for each Formation are found, which include an integrated arrangement of silt-
layer and were summarized in (Table 2). It was possible to prepare this stones clayey lenses with shale and multiple intercalations of medium to
table since the 3 areas are located on granulitic terrains. Thus, the elec- fine sandstone. This type of lithology traditionally has lower electrical
trical resistivity values found for the rock and the saprolites in the 3 resistivity values.

Fig. 5. Resistivity distribution at Area I (ERT 01).

8
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

Fig. 6. Resistivity distribution at Area II (ERT 02 and 03) and Area III (ERT 04 and 05).

Table 3
Soil properties and fitting parameters for each layer according to tactile-visual observations.

Area Soil Description Fitting slope Linear coef. R W.T Position NSPT P.R

I Clayey Silt with medium to fine sand (Landfill Layer) 0.38 -0.24 0.63 Above 1< NSPT <6 -
Silty clay with organic matter (Soft Clay) 0.092 1.51 0.64 Above 6< NSPT < 10 -
Sandy silt with gravel 0.069 6.19 0.83 Below 9< NSPT < 17 -
Sandy silt with gravel and rock fragments (Saprolite) 0.0062 11.4 0.97 Below 19< NSPT < 27 G
II Sandy silt with clay (Landfill Layer) 0.76 -39.0 0.77 Above 12< NSPT < 18 -
Sandy silt with clay and fragments of sandstone 0.20 -0.19 0.9 Above 18< NSPT < 35 -
Silt clay and weathered shale (Massapê and Saprolite) -0.14 27.8 0.82 Below 18< NSPT < 27 S
III Sandy silt with organic matter (Landfill Layer) 0.078 -0.28 0.95 Above NSPT < 5 -
Clayey silt with rock fragments (Saprolite) 0.047 1.85 0.95 Below NSPT > 15 G

W.T Position = layer position regarding the Water Table; P.R = Parental rock; G = Granulite; S = Shale

The Pojuca Formation has well-known and well-marked lithologies, At Area III, due to similarities in relationship obtained in terms of
which makes its correlation between ERT’s (ERT 02 and ERT 03) and NSPT x Resistivity for the subsurface soil, all borehole data were gathered
SPT’s (SP 201 to 206) surveys in the study area interesting. Note that the in (Fig. 7 H). Therefore, the generalized site-specific relationship for this
electrical resistivity values between the SPT’s do not change much. study area can be written as y = 0.04x + 1.94 and the coefficient of
At area III, ERT 04 and ERT 05 (Fig. 6) indicates that the resistiv- correlation is 0.97.
ity distribution is mainly constituted by an unconsolidated near surface It was possible to observe the influence of salinity variations in satu-
layer (45–75 Ωm) which presents low resistivity zones (15–45 Ωm) re- rating fluid at the relationship between the electrical parameter and the
lated to water saturation or sewage leak. The resistivity increases with soil strength as reported by Sudha et al. (2009). The coefficients of cor-
depth due to the presence of granulite rock alteration (75–2000 Ωm) relation were significantly reduced at partially to fully saturated layers
until it reaches the granulite rock (>2000 Ωm). with brackish water from the mangrove ecosystem (Fig. 7 A).
The Fig. 7 (E) indicates a negative correlation for the residual
4.3. Resistivity correlation with geotechnical data soil profile from shale. This distinct behavior can be explained by
parental rock mineralogy which has high trilaminar minerals concentra-
Fig. 7 shows linear relationships between the NSPT and resistivity tion and properties that facilitates easy flow of electric current around
for each layer at the three study areas. Table 3 summarizes the soil de- each clay particle, as Cation Exchange Capacity (CEC) and Specific
scriptions (with increasing depth), fitting parameters and coefficients of Surface.
correlation (R), layer position regarding the water table, NSPT and, when As reported by Sudha et al. (2009) and Devi et al. (2017), the coef-
applicable, parental rock for the layers of each area. ficients of linear relationships were also related with the lithology and

9
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

Fig. 7. Linear relationship between NSPT and Resistivity obtained for: (a) Materials above the water table at Area I (b) Materials below the water table at Area I; (c)
Landfill Layer at Area II, (d) Sandy silt with clay and fragments of sandstone at Area II; (e) Massapê and Saprolite layer at Area II; (f) Saprolite layer at Area III; (g)
Landfill Layer at Area III and (h) All SPT’s joined together at Area III.

10
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

• An association between Electrical Resistivity Tomography (ERT) and

Standard Penetration Test (SPT) have been used at three sites with
different soil matrix and geological environments. Based on the 2D
inversions of ERT data, the resistivity range of different layers has
been established for each area and analyzed against results from
borehole data. Linear fitting successfully correlates resistivity and
NSPT for one borehole each time as well as when similar soils are
joined together in a site-specific relationship.
• The behavior analysis of the coefficients of linear relationship con-
cluded that only the fitting slope is sensible to lithology, texture and
clay content variation. The only negative fitting slope value rep-
resents the transition between the massapˆe soil to shale rock and
shows the influence of a soil with high trilaminar mineral content in
fitting slope values. In addition, it was possible to observe that the
Fig. 8. Fitting slope x average resistivity.
coefficient of correlation obtained the lowest values for layers with
salinity variations, behavior predicted by Sudha et al. (2009).
• It is important to emphasize the relevance of geophysical investi-
clay content of each study area. In this study, the coefficients’ behavior
gations associated with classical invasive techniques commonly ap-
was analyzed separately.
plied in Brazil and other countries in South America. However, the
Regarding the Table 3, no correlation were found between linear
methodology of the present study is applicable in other countries,
coefficient and lithology or clay content variation. Also, no correlation
which have other geological, pedological and climatic characteris-
between this coefficient and average NSPT or average resistivity were
tics.
established.
Although no granulometric analysis was performed on this research,
Declaration of Competing Interest
visual tactile descriptions from borehole data indicated that layers with
a higher coarse content had low fitting slope values. In the other hand,
The authors declare that they have no known competing financial
higher fitting slope represented soils with an expressive fine content.
interests or personal relationships that could have appeared to influence
Also considering the fitting slope, positive values indicate residual
the work reported in this paper.
soil transition to granulite rock and negative values indicates residual
soil transition to shale rock. The lowest value corresponds to the mas-
References
sapˆe and saprolite layer at Area II, highlighting the trilaminar minerals
content influence in the correlation. Abidin, M.H.Z., Saad, R., Ahmad, F., Wijeyesekera, D.C., Baharuddin, M.F.T., 2014.
No correlation between fitting slope and average NSPT was found, Correlation analysis between field electrical resistivity value (ERV) and ba-
as represented in Fig. 8. However, a correlation between fitting slope sic geotechnical properties (BGP). Soil Mech. Found. Eng. 51 (3), 117–125.
doi:10.1007/s11204-014-9264-x.
and average resistivity was established and after a curve fit, using Least ABNT-NBR-6484. Method for standard penetration test for soils. Technical Report; ABNT;
Square Method, this relationship can be observed in Fig. 8. The coeffi- Rio de Janeiro; 2001.
cient of correlation (R) was 0.94. Akin, M.K., Kramer, S.L., Topal, T., 2011. Empirical correlations of shear wave velocity
(Vs) and penetration resistance (SPT-N) for different soils in an earthquake-prone area
This curve behavior can be interpreted for situations where we have (Erbaa-Turkey). Eng. Geol. 119, 1–17. doi:10.1016/j.enggeo.2011.01.007.
average resistivity equals to zero and tending to infinity. In the first case, Barbosa, J.S.F, Corrêa-Gomes, L.C, Dominguez, J.M.L, Cruz, S.A.S, Souza, J.S, 2005. Pet-
the fitting slope limit value would be equals to 1.01. In the other hand, rografia e Litogeoquimica das Rochas da Parte Oeste do alto de Salvador, Bahia. Rev.
Bras. Geoc. 35 (4 - Suplemento), 9–22.
for an average resistivity with very high values, the fitting slope would Barros, J.M.C., Pinto, C.S., 1997. Estimation of maximum shear modulus of Brazilian trop-
be zero. ical soils from standard penetration test. In: Proceeedings of the International Confer-
ence on Soil Mechanics and Foundation Engineering, 14, Hamburgo , 1, pp. 29–30.
Bortolozo, C., Benessiuti, M.M., Andrade, M., Lavalle, L., Mendes, R., Simões, S.,
5. Conclusions Mendes, T., Pampuch, L., 2018. Combined analysis of electrical and electromag-
netic methods with geotechnical soundings and soil characterization as applied
to a landslide study in Campos do Jordão City, Brazil. J. Appl. Geophys. 161.
The use of geophysical methods to obtain geotechnical information doi:10.1016/j.jappgeo.2018.11.017.
from an area proved to be an excellent aid tool. The use of this indirect Braga, A., Malagutti Filho, W., Dourado, J., Chang, H.K., 1999. Correlation of elec-
method alone is not sufficient for the complete geological-geotechnical trical resistivity and induced polarization data with geotechnical survey stan-
dard penetration test measurements. J. Environ. Eng. Geophys. 4 (2), 123–130.
characterization, however, as shown in the present study, the commonly doi:10.4133/JEEG4.2.123.
used invasive geotechnical investigations (such as SPT) have certain lim- Cosenza, P., Marmet, E., Rejiba, F., Cui, Y.J., Tabbagh, A., Charlery, Y., 2006. Correlations
itations, the main one being the one-off character of the investigation. between geotechnical and electrical data: a case study at Garchy in France. J. Appl.
Geophys. 60, 165–178. doi:10.1016/j.jappgeo.2006.02.003.
Geological surveys obtain information in depth basically from the Coutinho, R.Q., Schnaid, F., 2010. Regional report for South America. Proceeedings of the
point where they were carried out. The geophysical methods in turn 2nd International Symposium on Cone Penetration Testing CPT10.
can supply and assist the understanding of the geological-geotechnical De Vita, P., Di Maio, R., Piegari, E., 2012. A study of the correlation between
electrical resistivity and matric suction for unsaturated ash-fall pyroclastic soils
context of the area, since the measured ERT’s provide electrical resis- in the Campania Region (southern Italy). Environ. Earth Sci. 67 (3), 787–798.
tivity values at various points in depth. These, when interpolated in doi:10.1007/s12665-012-1531-4.
dedicated software, present information that can be associated with the Devi, A., Israil, M., Anbalagan, R., Gupta, P.K., 2017. Subsurface soil characterization
using geoelectrical and geotechnical investigation at bridge site in Uttarakhand Hi-
presence of the groundwater level, clay content and depth of the bedrock
malayan region. J. Appl. Geophys. 144, 78–85. doi:10.1016/j.jappgeo.2017.07.005.
top, for example. Di Maio, R., Paola, C., Forte, G., Piegari, E., Pirone, M., Santo, A., Urci-
Another advantage of applying these investigation methods together, uoli, G., 2020. An integrated geological, geotechnical and geophysical approach
to identify predisposing factors for flowslide occurrence. Eng. Geol. 105473.
as shown in the present study, the electrical resistivity tomography
doi:10.1016/j.enggeo.2019.105473.
method showed great potential for extrapolating data from SPT surveys, Dudal, R., Eswaran, E., 1988. Distribution, properties and classification of Vertisols. In:
remedying the deficiency in the complete structural characterization in Wilding, L.P., Puentes, R. (Eds.), Vertisols: Their Distribution, Properties, Classifica-
civil engineering projects. tion and Management. Texas A&M University Printing Centre, College Station, Texas.
Elsevier, Amsterdam, pp. 1–22.
Final considerations, therefore, can be distributed according to the EMBRAPA – Brazilian Agricultural Research Corporation., 2018. Sistema Brasileiro de
following topics: classificação de Solos. EMBRAPA Solos, Brasília, p. 356 5a ed.

11
J.T.D. Gonçalves, M.A.B. Botelho, S.L. Machado et al. Environmental Challenges 5 (2021) 100220

FAO, 2014. World Reference Base for Soil Resources 2014: International Soil Classification Moreira, C.A., Lapola, M.M., Carrara, A., 2016. Comparative analyzes among electrical
System for Naming Soils and Creating Legends for Soil Maps. Food and Agriculture resistivity tomography arrays in the characterization of flow structure in free aquifer.
Organization of the United Nations, p. 191 Rome: FAO(World Soil Resources Reports Geofis. Int. 55 (2), 119–129.
106). Oh, S., Sun, C.G., 2008. Combined analysis of electrical resistivity and geotechnical SPT
Folle, D., Costa, J.F.C.L., Renard, D., Koppe, J.C., Zingano, A.C., 2008. 3-D soil- blow counts for the safety assessment of fill dam. Environ. Geol. 54 (1), 31–42.
resistance maps in the presence of a strong vertical trend. Eng. Geol. 102, 46–53. doi:10.1007/s00254-007-0790-y.
doi:10.1016/j.enggeo.2008.07.006. Sasaki, Y., 1992. Resolution of resistivity tomography inferred from numerical simulation.
Giao, P.H., Chung, S.G., Kim, D.Y., Tanaka, H., 2003. Electrical imaging and laboratory Geophys. Prospect. 40, 453–454. doi:10.1111/j.1365-2478.1992.tb00536.x.
resistivity testing for geotechnical investigation of Pusan clay deposits. J. Appl. Geo- Sass, O., Bell, R., Glade, T., 2008. Comparison of GPR, 2-D resistivity and traditional
phys. 52, 157–175. doi:10.1016/S0926-9851(03)00002-8. techniques for the subsurface exploration of the O¨schingen landslide, swabian alb
Groves, P., Cascante, G., Dundas, D., Chatterji, P.K., 2011. Use of geophysical methods for (Germany). Geomorphology 93, 89–103. doi:10.1016/j.geomorph.2006.12.019.
soil profile evaluation. Can. Geotech. J. 48, 1364–1377. doi:10.1139/t11-044. Siddiqui, F.I., Osman, S.B.A.B.S., 2013. Simple and multiple regression models for relation-
Guimarães, T., Vale, R.M.C., 2012. Suscetibilidade a erosão do solo na cidade de Salvador ship between electrical resistivity and various soil properties for soil characterization.
– Bahia. Anais 9º Simpósio Nacional de Geomorfologia (SINAGEO). UFRJ: Rio de Environ. Earth Sci. 70 (1), 259–267. doi:10.1007/s12665-012-2122-0.
Janeiro. Silva, D.R., Mizusaki, A., Milani, E., Pimentel, M., 2012. Depositional ages of Paleozoic
Hatta, K.A., Osman, S., Azahar, S.B., 2015. Correlation of electrical resistivity and SPT- and Mesozoic pre-rift supersequences of the Recôncavo basin in northeastern Brazil:
N value from standard penetration test (SPT) of sandy soil. Appl. Mech. Mater. 785, A Rb–Sr radiometric study of sedimentary rocks. J. S. Am. Earth Sci. 37, 13–24.
702–706. doi:10.4028/www.scientific.net/AMM.785.702. doi:10.1016/j.jsames.2011.12.006.
Hossain, S., Kibria, G., Khan, S., 2018. Site Investigation Using Resistivity Imaging. CRC Silva, P.S., Campos, J.E.G., Cunha, L.S., Mancini, L.H., 2018. Relationships
Press. of stable isotopes, water-rock interaction and salinization in fractured
IBGE – Brazilian Institute of Geographic and Statistical. 2018. Mapa de Solos da Folha aquifers, Petrolina region, Pernambuco State. REM Int. Eng. J 71 (1), 19–25.
SD.24 – Salvador. Access in https://dados.gov.br/dataset/cren_pedologiasd24 doi:10.1590/0370-44672016710178.
Kirar, B., Maheshwari, B., Muley, P., 2016. Correlation between shear wave velocity (Vs) Solberg, I.L., Hansen, L., Rønning, J.S., Haugen, E.D., Dalsegg, E., Tønnesen, J.F., 2012.
and SPT resistance (N) for Roorkee region. Int. J. Geosynth. Ground Eng. 2 (1), 1–11. Combined geophysical and geotechnical approach to ground investigations and haz-
doi:10.1007/s40891-016-0047-5. ard zonation of a quick clay area, mid Norway. Bull. Eng. Geol. Environ. 71.
Lago, A.C., Costa, I.S.L, da Cunha, F.G., Sousa, F.R.F.R., 2018. Geophysical investigation doi:10.1007/s10064-011-0363-x.
using the GPR method: a case study on the contamination of lead in the Santo Amaro Souza, J.D., de Melo, R.C., Kosin, M., 2003. Mapa geológico do estado da Bahia. Versão
town, Bahia, Brazil. J. Geol. Surv. Braz. 1, 61–67. doi:10.29396/jgsb.2018.v1.n2.1. 1.1. Salvador: CPRM. Escala 1:1.000.000.
Lima, G.K., Corrêa, M.M., de Souza, V.S.J., 2015. Mineralogia de vertissolo ebânico do Sudha, K., Israil, M., Mittal, S., Rai, J., 2009. Soil characterization using electrical re-
nordeste brasileiro. XXXV Congresso Brasileiro de Ciência do Solo, Natal – RN. sistivity tomography and geotechnical investigation. J. Appl. Geophys. 67, 74–79.
Loke, M., 2013. RES2DINV version 3.71, Software User’s Manual. Geotomo Software Sdn doi:10.1016/j.jappgeo.2008.09.012.
Bsd.. Tucker, S.E., Briaud, J.L, Hurlebaus, S., Everett, M.E., Arjwech, R., 2015. Electrical resis-
Loke, M., Barker, R., 1996. Rapid least-square inversion of apparent resistivity tivity and induced polarization imaging for unknown bridge foundations. J. Geotech.
pseudo sections by a quasi-newton method. Geophys. Prospect. 44, 131–152. Geoenviron. 141 (5), 04015008. doi:10.1061/(ASCE)GT.1943-5606.0001268.
doi:10.1111/j.1365-2478.1996.tb00142.x. Vieira, C., Pereira, A., Dias, V., Cerqueira Neto, J., Silva, F., Ferreira, L., 2005. Caracter-
Long, M., Donohue, S., L’Heureux, J.S., Solberg, I.L., Ronning, J.S., Limacher, R., iza¸c˜ao geol´ogico-geot´ecnica das encostas urbanas do Munic´ıpio de Salvador. IV
O’Connor, P., Sauvin, G., Romoen, M., Lecomte, I., 2012. Relationship between elec- COBRAE - Conferˆencia Brasileira sobre Estabilidade de Encostas, pp. 153–164.
trical resistivity and basic geotechnical parameters for marine clays. Can. Geotech. J. Ward, S.H., 1990. Resistivity and induced polarization methods. In: Ward, Stanley H. (Ed.),
49 (10), 1158–1168. doi:10.1139/t2012-080. USA Investigations in Geophysics, no 5. Geotechnical and Environmental Geophysics.
Lundström, K., Larsson, R., Dahlin, T., 2009. Mapping of quick clay forma- Society of Exploration Geophysicists, pp. 147–189 Ed.v. I.
tions using geotechnical and geophysical methods. Landslides 6, 1–15. Zhdanov, M., Keller, G., 1994. The Geoelectrical methods in geophysical exploration. El-
doi:10.1007/s10346-009-0144-9. sevier, Amsterdam.
Machado, S., Presa, E., 2000. Monitoring a creep process on an expansive Brazilian soil. Zhou, B., Dahlin, T., 2003. Properties and effects of measurement errors
In: Proceedings of the ISRM International Symposium. Melbourne, Australia. on 2D resistivity imaging surveying. Near Surf. Geophys. 1, 105–117.
Martorana, R., Fiandaca, G., Casas Ponsati, A., Cosentino, P.L., 2009. Comparative tests on doi:10.3997/1873-0604.2003001.
different multi-electrode arrays using models in near-surface geophysics. J. Geophys.
Eng. (6) doi:10.1088/1742-2132/6/1/001, –1.

12