IOP PUBLISHING JOURNAL OF GEOPHYSICS AND ENGINEERING

J. Geophys. Eng. 10 (2013) 035013 (10pp) doi:10.1088/1742-2132/10/3/035013

2D and 3D resistivity tomography of the

Suı́mo garnet-bearing dyke, Lisbon
Volcanic Complex, Portugal: a case study
M A Khalil 1 , F M Santos 1 , M Cachão 2,3 , P E Fonseca 2,3 and J Mata 2,3
1
Universidade de Lisboa—IDL, Campo Grande, Ed. C8, 1749-016 Lisboa, Portugal
2
Universidade de Lisboa, Faculdade de Ciências, Departamento de Geologia, Campo Grande C6,
1749-016 Lisboa, Portugal
3
Centro de Geologia da Universidade de Lisboa, Faculdade de Ciências, Campo Grande C6,
1749-016 Lisboa, Portugal
E-mail: khalil250@hotmail.com

Received 27 July 2012

Accepted for publication 28 February 2013
Published 6 June 2013
Online at stacks.iop.org/JGE/10/035013

Abstract
Upper Cretaceous volcanic activity in the West Iberian Margin produced a singular
garnet-bearing basaltic dyke occurring on the northern slope of the Suı́mo hill near Lisbon.
2D and 3D resistivity tomography have been carried out to elucidate the properties of the dyke
regarding the horizontal and vertical subsurface extension after historical mining. 2D
large-scale resistivity profiles (GA-1 and GA-2) were measured. Ten sub-parallel 2D
resistivity sections were measured in small scale to perform a 3D resistivity model. All
measurements were carried out using a pole–dipole array. The basaltic dyke, nowadays not
cropping out, is detected as a low resistivity layer, extending to a depth of at least 40 m. The
width of the dyke ranges from 150 to about 40 m to the south. A high resistivity limestone unit
bounds the dyke from east and west sides. Some karstification features can be easily observed
in the limestone unit, forming very high resistivity zones. The obtained results allow the
prediction that for sampling of in situ basaltic rock the extraction of an ≈20 m thick layer of
loose material would be necessary. The central garnet-rich zone of the dyke would be only
reached at depths higher than ≈45 m.
Keywords: geological heritage, basaltic garnet-bearing dyke, Suı́mo, 2D and 3D geoelectrical
resistivity methods
(Some figures may appear in colour only in the online journal)

Introduction 3.1 Ma (Ferreira and Macedo 1979), related intrusives namely

the tectono-magmatic annular structure of Sintra have been
At the West Iberian Margin, Lisbon and its surroundings dated from 86.8 ± 2.5 Ma (K–Ar; Mahmoudi (1991)) to
were locus of important Upper Cretaceous volcanic activity 93.8 ± 3.9 Ma (40Ar/39Ar; Miranda et al (2009)).
(Lisbon Volcanic Complex—LVC), whose effusive, explosive A very peculiar occurrence of the LVC is a basaltic
and hypabyssal remnants spread over more than 200 km2,
dyke mapped at the northern slope of the Suı́mo hill,
mainly at north and west of the Portugal capital (e.g. Alves
et al (1980), Rock (1982), Palácios (1985)). The alkaline near the village of Belas, between Lisbon and Sintra. Its
LVC fossilized a karst morphology previously developed on interest results from the occurrence of garnet, amphibole
Cenomanian rudist limestones and predated the formation and clinopyroxene megacrysts set in an aphanitic basaltic
of a Paleogene conglomerate. Despite the available isotopic groundmass (see Choffat (1914), Palácios (1985)). Despite
age (K–Ar method) for the basaltic volcanism being 72.6 ± the dyke not cropping out nowadays, small fragments of the

1742-2132/13/035013+10$33.00 © 2013 Sinopec Geophysical Research Institute Printed in the UK 1

J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

megacrysts and small blocks of basaltic rock hosting those Details of resistivity and ERT theory, different types of arrays,
xenocrysts can still be found in ancient quarry open pits. and procedures of data acquisition and field survey are found
Suı́mo references as an old quarry for semi-precious in many text books, such as Telford et al (2001) and Reynolds
stones (garnets) date back at least to the beginning of (1998).
the 1st century, as reported by Gaius Plinius Secundus In recent years, there has been a widespread increase
(23–79), better known as Pliny the Elder, in his monumental in the use of 2D and 3D electrical imaging (tomography)
work Naturalis Historia. Since then and until the beginning of covering a large spectrum of hydrogeological, archaeological,
the 20th century more than 60 classic authors have explicitly environmental and engineering applications (e.g. Wynn and
referred to it as an example of roman mining for semi- Grosz (2000), Madsen et al (2001), Manheim et al (2001),
precious minerals in the Iberian Peninsula. Garnets continued Dahlin et al (2002), Tonkov and Loke (2006), Forquet and
to be explored throughout Middle-Ages (Azevedo 1916). The French (2012)).
ancient mining activity and the subsequent infilling of the open
pits by loose materials dictates that the garnet-bearing dyke is
no longer outcropping nowadays.
Data acquisition
All this justifies that more than 20 centuries after
In the Suı́mo hill the 2D resistivity data were collected in
the first written reference, Suı́mo is still the object of
surveys carried out in two campaigns. The first was for
both geological and archaeological interest (Cachão et al
reconnaissance of the subsurface geological situation and to
2010, Cardoso et al 2011). To better understand the structure of
test the previous assumptions about the existence and extension
the Suı́mo magmatic body and to constrain the depth reached
of the Suı́mo garnet-bearing dyke and karstification features
by mining activity into the still recognized open pits of the
in the calcareous deposits in the area. During this campaign
Suı́mo mine, two-dimensional (2D) and three-dimensional
two 2D resistivity profiles (GA-1 and GA-2) were measured
(3D) non-destructive electrical resistivity tomography (ERT)
were conducted across this dyke structure to address (figure 1). GA-1 is a 360 m length profile measured using
the feasibility of a future joint geological/archaeological 72 electrodes and with an off-set distance of 5 m between
excavation. each two successive electrodes. GA-2 is 210 m in length
and was measured using 42 electrodes with 5 m off-set
between electrodes. A one direction pole–dipole array was
ERT methodological background used for deeper detection. A pole–dipole array is a single
probe arrangement proposed by Eve and Keys (1956), in
The basic concept of resistivity measurements is the injection
which two current electrodes are separated by a very long
of electric current into the earth via two current electrodes
distance and the resistivity measurements are carried out in
and measuring the potential difference via two potential
the neighbourhood of one of the current electrodes. Due to the
electrodes. Many electrode arrays are being used either for
vertical electrical sounding or horizontal profiling. The vertical good horizontal coverage of this array, it is an attractive array
electrical sounding is a one-dimensional technique based on for multi-electrode resistivity meter systems with a relatively
the concept of a homogeneous, isotropic, horizontal layered small number of nodes. This array is probably more sensitive
earth model, where the subsurface is restricted to a number of to vertical structures.
horizontal layers and does not take into account lateral changes Guided by the results of GA-1 and GA-2, an attempt was
in the layer resistivity. made to carry out a 3D resistivity image of the study area.
A more confident model of the subsurface is a 2D model For the achievement of such an objective, ten relatively
where the resistivity changes in the vertical direction, as well parallel 2D resistivity profiles were carried out with 6 m
as in the horizontal direction along the survey line, assuming interval distance as shown in figure 1. The lengths of the
that resistivity does not change perpendicularly to the survey profiles range from 175 m in P-1 to 85 m in P-10. A
line. If this is not the case, theoretically, a 3D resistivity survey pole–dipole array was used as well. The current electrode
and inversion should be much more accurate. B—infinity electrode—was a unique point for all measured
The growing of computing power in the 1970s led to profiles. It is located about 1000 m away in the northwestern
developing new modelling tools (Dey and Morrison 1979), and direction of the study area. The data have been collected using
the appearance of electrical imaging concepts (Lytle and Dines the Syscal-pro, ten-channel resistivity meter manufactured by
1978), which emerged in parallel with advances in biomedical IRIS instruments.
imaging. The emergence of imaging concepts prompted the
development of multi-electrode measurement systems in the Data processing
1980s. The term ‘ERT’ is widely used to describe the inversion
process together with images of the results. Recently, imaging A—2D modelling
techniques have been developed to study 3D and 4D systems,
coupled with very recent advances in the modelling and The acquired 2D resistivity data have been inverted using
inversion technique. RES2DINV (Geotomo software 2009). The inversion routines
The electrical imaging technique combines surface are based on the smoothness-constrained least-squares method
profiling with vertical sounding to produce a 2D or 3D image (Sasaki 1989, 1992, Loke and Barker 1995, 1996a, 1996b) and
of the subsurface resistivity. For 2D and 3D cases, the finite- the forward resistivity calculations were executed by applying
difference and finite-element methods are the most versatile. an algorithm based on a finite element method. The first

2
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

Figure 1. Location map of the 2D resistivity sections GA-1 and GA-2, and the ten profiles (P-1 to P-10) arranged in parallel. The profile
layout and length was constrained by topography, dense vegetation and bushes. The satellite vertical image was taken from Google Earth
(2012).

Figure 2. 2D resistivity profile GA-1.

step in the inversion process is dividing the subsurface into expressed in terms of the RMS error. Figures 2 and 3 show
a large number of small rectangles, which resistivity values of the final models of profile GA-1 and GA-2 with topography,
the model rectangles should be estimated directing toward respectively.
minimizing the difference between the calculated and the Figure 4 shows the 2D inverted resistivity models of the
observed apparent resistivity values. The quality of the fit is ten parallel profiles. The measured 2D apparent resistivity data

3
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

Figure 3. 2D resistivity profile GA-2.

sets of all profiles are combined into a 3D data set to be inverted where Wd is a diagonal matrix, consisting of the reciprocal
in a 3D manner. All profiles have the same logarithmic colour of data standard deviations, and C is the roughness operator
scale that ranges from 7.49 to 552  m. The RMS errors of as defined by Sasaki (1994). The parameter λ is a Lagrange
the 2D inversion range from 1.6 in profiles P-6 and P-7 to 7.5 multiplier, and is used to control the balance between data
in profile P-8. fit and model smoothness. The value of this parameter is
decreased by a factor of 0.6, from an initial value of 0.8,
B—3D modelling till the minimum value of 0.01 during the inversion.
The normal equations (3) were solved by means of
The 3D resistivity inversion code is based on the finite element a preconditioned conjugate gradient method. The model
technique and regularization method proposed by Monteiro parameters were then updated by adding the vector p. The
Santos and Sultan (2008). It was applied upon the measured ten iteration procedure continued until the misfit is reduced to an
parallel resistivity profiles. In this work only a brief description acceptable level. The misfit between data and model responses
of the method is presented. is given by the average absolute error:
The inverse problem involves nonlinear relationships 

N 
between the model response and the model parameters, thus 100   yci − yob
2

requiring an iterative procedure. The logarithms of the model Misfit (%) = i

, (4)
N i=1 yob
i
resistivity and of the apparent resistivity are used as model
parameters and data set, respectively. The scheme adopted in where N represents the number of data points.
this study is based on Sasaki (1994, 2001). The problem is The ten relatively parallel 2D resistivity sections (figure 4)
linearized as were inverted using the 3D algorithm described above to have
a 3D resistivity model of the area as shown in figure 5.
Jp = d, (1)
where p is the vector containing the corrections to the model Interpretation
parameters p, d = yc – yob is the vector of the differences
between the model responses and the measured (observed) The 2D resistivity section GA-1 (figure 2) allowed the
data, and J is the derivative matrix (Jacobian) containing demonstration of a noticeable low resistivity geoelectric zone,
the derivatives of the model responses with respect to the extending horizontally for a distance of 170 m approximately

model parameters Ji j = ∂yci /∂ p j . In this work the Jacobian (from x = 100 to 270 m). This zone, corresponding to a
components were estimated using the algorithm presented by layer extending down to about 20 m, at most, from the
Loke and Barker (1996a) and by assuming that each block is ground surface, has a resistivity lower than 22  m and
embedded in a homogeneous half-space. The Jacobian values is surrounded from both sides by a higher resistivity layer
were subsequently updated using Broyden’s (1965) method. which extends to the maximum measured depth (≈45 m).
The minimization of an appropriate objective function This higher resistivity (>60  m) layer corresponds to the
allows the estimation of the corrections to the model Urgonian limestones. The contrast between these two layers
parameters in each iteration. The objective function to be produces significant geoelectrical discontinuities as is clearly
minimized is observed at about 105 m from the start of the section. It resulted
 2  2 in high-condensed resistivity contours extending vertically,
Q = Wd(d − Jp) + λ Cp . (2) and is interpreted as reflecting the sharp contact between the
Minimization of this function yields the normal equations (see surrounding limestone and the zone mapped as basaltic dyke.
details in Sasaki (1994)): In this, the above referred low resistivity layer is underlain
 T T  by a zone of equivalent x extent and higher resistivity (up
J Wd Wd J + λCT C p = JT WdT Wdd, (3)
to 60  m) subdivided by a central vertical zone of lower

4
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

Figure 4. 2D resistivity models obtained for profiles P-1 to P-10.

5
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

Figure 5. 3D inversion model showing the x, y, z resistivity distribution in the area. Resistivity slices were taken at depths of 3, 5, 8, 12.5,
17.5, 22.5, 27.5, 32.5 and 42.5 m from the ground surface.

6
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

resistivity (down to 22  m) extending to depths of 45 m. matching with the 2D resistivity profiles P-7 and P-8 as well.
Appreciably, three very high resistivity zones (about 3000  They obviously show the shallow high resistivity zone under
m) could be observed inside the limestone layer. These zones x = 60 and 80 m.
may correspond to some karst caves. Accordingly, 2D ERT and 3D inversion models are in
The same geoelectric zones could be detected in the a considerable degree of agreement regarding the spatial
second profile, GA-2 (figure 3). Here the thickness of the distribution and magnitude of resistivity in the study area.
low resistivity layer (lower than 60  m) is thinner than Both can provide a general comprehensive resistivity image
that determined in GA-1. It is directly underlain by the high either in the vertical or horizontal direction.
resistivity layer (from 60 to 300  m), which is thicker in this The subsurface domain of the 3D model was divided into
profile. The sharp contact and caves could be detected as well. 3576 cells. The initial resistivity is 35  m, which is the average
Examination of the ten 2D resistivity profiles shows the resistivity of all measured data points. The misfit of the model
same features discussed earlier, concerning the existence and (RMS) is 23.3; this high ratio could be attributed to the nature
extension of the low resistivity layer in the middle of the of the measured data, where the available 2D data form a
sections and the surrounded high resistivity layers from both series of survey lines measured only in one direction. However,
sides. A new observation which could be detected here is the in most cases, 2D techniques without remote electrodes,
sharp contrast in resistivity in both sides of the low resistivity applied in parallel lines, are probably adequate and are simpler
layer in the majority of the profiles especially those close to the logistically. Such parallel and possibly crossing lines can be
mine’s open pit, particularly, profiles P-1, P-2, P-3, P-4 and inverted using 3D techniques in complex areas (Dahlin et al
P-5. This may outline the vertical and horizontal extension 2002). However, it is always expected to have a lower quality
of the contact between the basaltic dyke and the surrounding 3D model (high misfit of the model) than that produced
with a complete 3D survey (e.g. a square grid or a cross-
limestone.
diagonal survey). This is because the 3D inversion accounts
The 3D model shows a high resistivity zone (more than
for the effects that depend on the electrode array type and
400  m) in the western and northwestern corner of the area,
layout direction, because the electrode array characteristics are
extending from the surface to a depth of 209.5 m (a.m.s.l.).
incorporated in the Jacobian matrix employed in the inversion
This high resistivity zone is conformable with the shallow
process (Dahlin et al 2002). However, the 3D inversion of
high resistivity zones in 2D resistivity profiles P-1, P-2, P-3,
2D parallel resistivity lines could reveal major resistivity
P-4, P-5 and P-6, which begins from x = 100 and x = 120
variations across the survey lines. The geological sketch
to the end of the mentioned profiles. Another high resistivity
diagram (figure 7) is a combined geological interpretation of
zone appears in the opposite southeastern direction at a depth the 2D resistivity cross sections represented in figures 2 and
of 219.5 m (a.m.s.l.) and extends vertically to the depth of 4. The first and second geological profiles represent the third
194.5 m (a.m.s.l.). This high resistivity zone appears in the and first resistivity profiles (P-3 and P-1) in figure 4. The third
majority of 2D sections beneath x = 40 m and at a similar geological profile represents the geological situation to the
depth. The majority of the area has a low resistivity (less than north of the measured resistivity profiles. This profile crosses
32.45  m). The resistivity of this zone begins with a very the old excavated zone of the dyke, which is characterized by
low value on the surface and gradually increases to a depth of low topography and dense vegetation in the middle.
219.5 m (a.m.s.l.) and decreases again to the depth of 194.5 m
(a.m.s.l.). This low resistivity zone is completely in agreement
Discussion and conclusion
with the low resistivity zone, in blue colour, covering the most
part of 2D sections. 2D and 3D resistivity tomography have been carried out in
A very low resistivity zone (less than 6  m) in the eastern northern slope of Suı́mo hill, where an Upper Cretaceous
direction of the area extends vertically from the shallowest alkaline basaltic rock belonging to the LVC hosts garnet
level of 234 m (a.m.s.l.) to the deepest level. As the depth xenocrysts, justifying the very ancient (at least since the
increases, the low resistivity zone migrates gradually to the beginning of the 1st century B. C.) mining activity in the
west direction. area. Two large-scale 2D resistivity profiles (GA-1 and GA-2)
Figure 6 shows the 3D resistivity inversion as a sequence were measured to generally outline the basaltic dyke. A more
of vertical sections. There is a clear matching between the detailed survey was also performed through ten parallel 2D
first resistivity model (figure 6(A)) and the first 2D resistivity resistivity sections and a 3D resistivity model was constructed.
profile (P-1) in figure 4, where the low resistivity zone in the This study confirms the dyke verticality, an aspect already
middle is bounded by high resistivity zones from both sides suggested by the geological map (Almeida 1991) where
of the section with the same magnitude of resistivity. The the dyke is represented as a rectilineous structure. A high
second resistivity model (figure 6(B)) is located at y = 10 m, resistivity limestone unit bounds the dyke from east and west
and between 2D resistivity profiles P-2 and P-3 (figure 4). It sides. Some karstification features can be easily observed in
shows a good matching as well. The third resistivity model the limestone unit, forming very high resistivity zones. The
(figure 6(C)) is comparable with P-5, which is in the same width of the dyke ranges from 150 m in the north direction
location approximately. This resistivity model shows the same (resistivity section GA-1) to about 40 m to the south (resistivity
shallow high resistivity zone in the northwestern direction and section GA-2).
the wide low resistivity zone in the middle. The fourth and In what has been considered the present-day dyke
fifth resistivity models (figures 6(D) and (E)) show a good expression, i.e. the relatively low resistivity sector in vertical

7
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

(E )

(D )

(C )

(B )

(A )

Figure 6. 3D resistivity inversion as a sequence of vertical sections at y = 1, 10, 20, 30 and 40 m.

contact with the high resistivity limestones, this study allows corresponds to the infilling, by loose material, of the open
the identification of three distinct zones. A superficial one pits left by mining activity. This layer, extending down to
characterized by very low resistivity (up to 20  m) which 20 m, is underlain by a zone of equivalent extent characterized

8
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

Figure 7. Sketch diagram showing the geological interpretation of the garnet-bearing dyke of Suı́mo area. β—basaltic type dyke with
central slag dump; 1—slag dump (scoria); 2—contact metamorphism area. The values in metres indicate elevations above mean sea level.

by higher resistivity (up to 60  m) subdivided by a central the region of higher velocity flow, which in a dyke usually
vertical zone of lower resistivity (down to 22  m) extending corresponds to the central part, where phenocrysts and
to depths of 45 m. This central zone of lower resistivity may xenocrysts tend to concentrate by flow differentiation (e.g.
suggest that at depths the mining activity was restricted to the Ross (1986), Chistyakova and Latypov (2010)). All these
central part of the dyke, where the garnet megacrystals are offer an explanation to the hypothetical concentration of garnet
probably concentrated. xenocrysts at the central part of the dyke, which, after mining,
Flow regimes characterized by viscous resistance to flow, may have been filled in by loose material producing the
like those in basaltic magmas, are characterized by low registered present day low resistivity vertical zone.
Reynolds numbers and tend to be laminar, which means Based on this study any excavation to obtain in situ dyke
that in a dyke the flow velocity is higher in the central samples need to remove approximately 20 m of loose slag
zone (Best and Christiansen 2001). In addition, in a flowing dump material, while the central garnet-rich zone of the dyke
magma with suspended crystals, these tend to migrate into may probably be reached at depths higher than some 45 m.

9
J. Geophys. Eng. 10 (2013) 035013 M A Khalil et al

Acknowledgments Loke M H and Barker R D 1996a Rapid least-squares inversion of

apparent resistivity pseudosections using a quasi-Newton
The field work was partially funded by FCT and FEDER method Geophys. Prospect. 44 131–52
Loke M H and Barker R D 1996b Practical techniques for 3D
through the project Pest-OE/CTE/UI0263/2011. resistivity surveys and data inversion Geophys. Prospect.
44 499–523
References Lytle R J and Dines K A 1978 An impedance camera: a system for
determining the spatial variation of electrical conductivity
Almeida F M (coordinator) 1991 Carta geológica de Portugal à Lawrence Livermore National Laboratory Report UCRL
escala 1/50 000. Folha 34A, Sintra. Serviços Geológicos de 52413
Portugal Madsen J A, Brown L, McKenna T, Snyder S, Krantz D,
Alves C A M, Rodrigues B, Serralheiro A and Faria F 1980 O Manheim F, Haeni F-P, White E and Ullman W 2001
complexo basáltico de Lisboa Comum. Serv. Geol. Port. Geophysical characterization of fresh and saline water
66 111–34 distribution in a coastal estuarine setting Proc. Symp. of the
Azevedo P 1916 As pedras preciosas de Lisboa (Belas) na história Application of Geophysics to Engineering and Environmental
Archeol. Port. 23 158–202 Problems (SAGEEP) (4–7 March, Denver, CO)
Best M and Christiansen E H 2001 Igneous Petrology (Oxford: Mahmoudi A 1991 Quelques intrusions alcalines et basiques du
Blackwell) p 458 Cretacé superieur au Portugal PhD Thesis Université Nancy I
Broyden 1965 A class of methods for solving nonlinear Manheim F T, Krantz D E, Snyder D S, Bratton J F, White E A
simultaneous equations Math. Comput. 19 577–93 and Madsen J A 2001 Streaming resistivity surveys and core
Cachão M, Fonseca P E, Galopim de Carvalho R, Neto de drilling define groundwater discharge into coastal bays of the
Carvalho C, Oliveira R, Fonseca M M and Mata J 2010 A mina Delmarva Peninsula Geological Society of America Annual
de granadas do Monte Suı́mo: de Plı́nio-o-Velho e Paul Choffat Meeting vol 33 p A42
à actualidade 8th Congresso Nacional Geologia e-Terra vol 18 Miranda R, Valadares V, Terrinha P, Mata J, Azevedo M R,
p4 Gaspar M, Kullberg J C and Ribeiro C 2009 Age constraints on
Cardoso J L, Guerra A and Fabião C 2011 Alguns aspectos da the late cretaceous alkaline magmatism on the West Iberian
mineração romana na Estremadura e Alto Alentejo Margin Cretaceous Res. 30 575–86
ed J L Cardoso and M Almargo-Gorbea Lucius Cornelius Monteiro Santos F A and Sultan A S 2008 On the 3D inversion of
Bocchus—Escritor lusitano da Idade da Prata da Literatura vertical electrical soundings: application to the south Ismailia
Latina (Lisbon: Academia Portuguesa da História/Real area-Cairo desert road, Cairo, Egypt J. Appl. Geophys.
Academia de la Historia) pp 169–88 65 97–110
Chistyakova S and Latypov R 2010 On the development of internal Palácios T 1985 Petrologia do complexo vulcânico de Lisboa PhD
chemical zonation in small mafic dykes Geol. Mag. 147 1–12 Thesis University of Lisbon p 260
Choffat P 1914 Les mines de grenats du Suı́mo Comunicações da Reynolds J M 1998 An Introduction to Applied and Environmental
Comissão do Serviço Geológico de Portugal vol 10 (Lisbon: Geophysics ISBN 0-471-96802-1 (New York: Wiley) p 778
Avante) pp 186–98 Rock N M S 1982 The late cretaceous alkaline igneous province in
Dahlin T, Bernstone C and Loke M H 2002 A 3D resistivity the Iberian Peninsule, and its tectonic significance Lithos
investigation of a contaminated site at Lernacken, Sweden 15 111–31
Geophys. 67 1692–700 Ross M E 1986 Flow differentiation, phenocryst alignment, and
Dey A and Morrison H F 1979 Resistivity modelling for arbitrarily compositional trends within a dolerite dike at Rockport,
shaped two-dimensional structures Geophys. Prospect. Massachusetts Geol. Soc. Am. Bull. 97 232–40
27 106–36 Sasaki Y 1989 Two-dimensional joint inversion of magnetotelluric
Eve A S and Keys D A 1956 Applied Geophysics in Search for and dipole–dipole resistivity data Geophysics 54 254–62
Mineral, Revised Edition (Cambridge: Cambridge University Sasaki Y 1992 Resolution of resistivity tomography inferred from
Press) numerical simulation Geophys. Prospect. 40 453–64
Ferreira M R P and Macedo C R 1979 K–Ar ages of the Sasaki Y 1994 3D resistivity inversion using the finite element
Permian-Mesozoic basaltic activity in Portugal Europ. Col. method Geophysics 59 1839–48
Geochron., Cosmochron. and Isotope Geology (Norway) Sasaki Y 2001 Full 3D inversion of electromagnetic data on PC
Abstracts VI pp 26–7 J. Appl. Geophys. 46 45–54
Forquet N and French H K 2012 Application of 2D surface ERT to Telford W M, Geldart L P, Sheriff R E and Keys D A 2001 Applied
on-site wastewater treatment survey J. Appl. Geophys. Geophysics (New York: Cambridge University Press) pp 1–860
80 144–50 Tonkov N and Loke M H 2006 A resistivity survey of a burial
Geotomo Software (2009)—RES2DINV ver. 3.58 2009 Rapid 2-D mound in the ‘Valley of the Thracian Kings’ Archaeol.
resistivity & IP inversion using the least-squares method Prospect. 13 129–36
Malaysia www.geoelectrical.com Wynn J C and Grosz A E 2000 Induced-polarization—a tool for
Google 2012 Google Earth www.google.com/earth2012 mapping titanium-bearing placers, hidden metallic objects,
Loke M H and Barker R D 1995 Least-squares deconvolution of urban waste on and beneath the seafloor J. Environ. Eng.
apparent resistivity pseudosections Geophysics 60 1682–90 Geophys. 5 27–35

10