Detection of resistive features using towed slingram electromagnetic induction instruments.

Archaeological Prospection
Archaeol. Prospect. 16, 103–109 (2009)
Published online 24 March 2009 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/arp.350
Detection of Resistive Features using
Towed Slingram Electromagnetic
Induction Instruments
JULIEN THIESSON1*, MICHEL DABAS2 AND SE¤BASTIEN FLAGEUL1
1
2
ABSTRACT
UMR 7619 Sisyphe, UPMC, Case 105 4 place Jusieu 75252 Paris cedex 05, France
Geocarta, 16 rue du sentier 75002 Paris, France
Slingram frequency domain electromagnetic (FDEM) instruments allow simultaneous measurement
of both magnetic susceptibility and electrical conductivity, which should justify their widespread use
in archaeological surveying. However, this is not the case and their application remains quite limited
due to: (i) a lack of knowledge about the role of coil orientation and spacing in terms of the detection
abilities for archaeological features (especially for resistive bodies); and (ii) a lack of instrumentation
specifically designed for shallow targets.We present here a test of a new version of the CS60 instrument (VCP coil configuration and 0.6 m intercoil spacing) for shallow depth resistive feature detection.
This experiment was undertaken on the Roman site of Vieil-Evreux where a complete series of control
resistivity and radar data was obtained. Detection of buried Roman walls was successful, in accordance with what can be expected from three-dimensional modelling.This confirms that the application
of this type of instrument in archaeological surveys merits to be extended significantly. Copyright
# 2009 John Wiley & Sons, Ltd.
Key words: Slingramelectromagneticinstrument; resistivitymapping; walldetection; susceptibility
mapping
Introduction
Among the different solutions that can be
adopted to measure and to map electrical
resistivity using electromagnetic induction
(EMI) methods, Slingram instruments consist
of a transmitter coil and a receiver coil separated
by a distance, L, the depth of investigation being
of the same order as L.
In archaeological surveying, as in other areas
of exploration geophysics, EMI was considered
in order to bypass the limitation of requiring
sufficiently good galvanic contacts for the d.c.
resistivity method. The first field experiments
* Correspondence to: J. Thiesson, UMR 7619 Sisyphe, UPMC,
Case 105 4 place Jusieu 75252 Paris cedex 05, France.
E-mail: julien.thiesson@upmc.fr
Copyright # 2009 John Wiley & Sons, Ltd.
showed that the observed responses depend
mainly on magnetic susceptibility (Musson,
1968). A series of field trials coupled with an
improvement of the theoretical approach (Tite
and Mullins, 1969, 1970) allowed the recognition
that both electrical conductivity (s) and magnetic
susceptibility (kph) could be measured simultaneously using frequency domain instruments.
As s corresponds to a 908 out-of-phase response
and kph to an in-phase response, it has been
possible to propose instrument characteristics in
accordance with archaeological surveying
requirements (Tabbagh, 1974). As this type of
instrument is also able to detect metallic objects
(Tabbagh and Verron, 1983), they constitute very
flexible tools for geophysical surveying in
archaeology (Tabbagh, 1986a). However, in spite
of the existence of both prototypes and commerReceived 4 February 2009
Accepted 12 February 2009
104
cial instruments, the most widespread being the
EM38 (Geonics Ltd), their use has remained
limited in archaeology because of:
(i) misunderstanding of the spatial response of
such instruments and consequently erroneous data interpretation;
(ii) uncertainties about the knowledge of the
depth of investigation (Dabas and Tabbagh,
2003), together with feature detection abilities;
(iii) difficulties to compensate for the thermal
drift of in-phase responses and mechanical
instabilities.
To overcome these issues, it is useful to first
recall the physical constraints and assumptions
that govern the interpretation of the Slingram
results.
To measure s using EMI instruments (or its
inverse electrical resistivity, r, both representing
at the macroscopic scale electric charge displacement) without mixing with the dielectric permittivity e (that represents at the macroscopic
scale the electrical polarizability of a material),
one must respect first the condition s >> ev
(where v ¼ 2pf is the angular frequency, f being
the frequency). Given the range of values of soil
conductivity and permittivity, this implies that
f < 300 kHz. As the intercoil spacing, L, must stay
in the range of both the depth of investigation
and the lateral resolution, it cannot be very
different from 1 m in the majority of archaeological applications. Therefore, the low induction number (LIN) condition, L2smv << 1 (where
m is the magnetic permeability), is always
respected in archaeological prospection. Under
this condition the s response, the secondary field
generated by eddy currents, is proportional to the
induction number and 908 out of phase with the
primary field. Consequently this response does
not overshadow the in-phase component of
the secondary field generated by magnetic
susceptibility, as is the case for deeper target
investigations with larger intercoil spacing, for
example in mining geophysics. The simultaneous
measurement of both properties by taking into
account the phase of the responses is thus
possible (and far better than the measurement
of the modulus of the secondary field, which
Copyright # 2009 John Wiley & Sons, Ltd.
J. Thiesson et al.
would be difficult to interpret due to the
cumulative effect of the two independent
properties).
Another consequence of the LIN condition is
that the depth of investigation and the lateral
resolution depend both on L and on coil
orientation but not on frequency (Frischknecht
et al., 1991). The choice of the coil orientation
must be based on the objectives of archaeological
surveying, which is first to characterize archaeological features that are three-dimensional and
may correspond to both conductivity and/or
susceptibility contrasts. According to these
objectives, the definition of the best coil orientation has been studied by forward numerical
modelling (Tabbagh, 1986b). This study concluded that the best orientation is the perpendicular one, PERP, followed by the vertical
coplanar (horizontal dipoles) one, VCP. No later
field trial refuted this conclusion.
These results contradict the driving choice of
most instrument manufacturers that tend to
favour depth of investigation for soil conductivity layering, thus the HCP (horizontal coplanar, vertical dipoles) configuration (Geonics Ltd,
1996).
In the present paper, to contribute to a better
understanding of Slingram instrument abilities
we present a field trial of resistive targets
detection. We chose the walls of a Roman
building, buried in silt on a plateau, a conductive
context. This case corresponds to an unfavourable test situation because EMI methods are more
sensitive to conductive than resistive targets.
Test site, control data and apparatus
The Roman city of Gisacum is located on a
plateau in the sedimentary Paris Basin at Vieil–
Evreux, 7 km east from the modern city of Evreux
(Eure, France). It is formed by a large polygonal
worship place of more than 100 ha, which
surrounds a series of public religious buildings
including temples, thermae and fana (Dabas
et al., 2005). It has been surveyed using electrical
and magnetic methods (Aubry, 2003). Now, as it
constitutes an archaeological reserve, tests of new
instruments and surveying techniques can be
Archaeol. Prospect. 16, 103–109 (2009)
DOI: 10.1002/arp
Detection of Resistive Features Using Towed Slingram
easily performed over grassland plots where
tillage has stopped.
The test area, 40 by 40 m2, covers the central
part of the west fanum that comprises in its last
state (second century AD) a 13-m-sided square
stony wall cella (square central room with sacred
statue) surrounded by a rectangular 72 m T 36 m
outer wall. The resistivity mapping was achieved
105
using the ARP1 multidepth instrument. It has a
‘vol de canards’ array configuration with three
different receiver dipoles corresponding to three
different depths of investigation of about 0.5 m,
1 m and 2 m (Panissod et al., 1997). Figure 1a–c,
shows the apparent resistivity maps corresponding to these depths of investigation. The cella
wall generates a marked anomaly that can be
Figure 1. ARP1 (seewww.geocarta.net) apparent resistivitymapsonthe cellaofthewestern fanumat Vieil-Evreux (Eure,France):
(a) 0.5 m dipole spacing, (b) 1m dipole spacing and (c) 2 m dipole spacing.
Copyright # 2009 John Wiley & Sons, Ltd.
Archaeol. Prospect. 16, 103–109 (2009)
DOI: 10.1002/arp
J. Thiesson et al.
106
Figure 2. Ground-penetrating radar time-slice of the reflected
amplitude at 4 ns (one-way travel time) over the fanum west
cella (Radar Pulse Ekko1000 with a 225 MHz central frequency
antenna and a 0.2 m horizontal distance step).
Figure 3. The CS60 andits cart in use.This figure is available in
colour online at www.interscience.wiley.com/journal/arp
must take care to pull with a constant speed along
a profile.
interpreted by the presence of a wall and its
foundation, having its top just below the topsoil
and a significant extent in depth until about
0.8 m. A central smaller resistive feature can be
observed inside the cella, another shallow small
resistive feature appears 20 m to the east, and
part of the outer wall can be observed in the
northern part of the test area. The cella internal
structure was also surveyed by GPR using a
225 MHz antenna, the 0.3 m depth amplitude
slice is presented on Figure 2; it confirms the
existence of the thin internal central feature.
The CS60 instrument has a 27 960 Hz frequency, a 0.6 m intercoil spacing and is normally
used in VCP (horizontal dipoles) coil configuration; when necessary it can be used in HCP
(vertical dipoles) coil configuration to increase
the investigation ability for deep layers. Its
characteristics, coil orientation and spacing, were
originally designed for irrigated salted soil
studies (Job et al., 1995) and not for archaeology
(for which PERP configuration would be better).
Measurements can be achieved either point by
point or by continuously pulling the apparatus
on a cart (Figure 3). In the first case, the coil axes
are 0.05 m above the ground surface, and 0.07 m
in the second case. The acquired data are
transferred via a microcontroller to a PC at a
constant time step that corresponds to a distance
step of approximately 0.02 m. Thus, the operator
Copyright # 2009 John Wiley & Sons, Ltd.
Modelling control
A wall corresponding to those remains existing
in the west fanum can be modelled by a twodimensional feature of 300 V.m resistivity,
10 105 (SI) susceptibility, 0.6 m thickness and
0.45 m width embedded in a 25 V.m and
20 105 (SI) layer and covered by a 50 V.m,
40 105 (SI) and 0.2 m overburden topsoil. We
compared the different responses of three coil
configurations (PERP, HCP and VCP) for different intercoil spacings from 0.4 m to 1 m (Figure 4).
Generally speaking, it can be observed that on the
one hand, the wall is clearly detected for any
spacing and coil configuration, on the other hand,
that shorter spacings produce stronger
anomalies. This can by explained by the stronger
relative contribution of the wall response to the
total response, compared with the surrounding
medium. Then, discussing the shapes of the
anomalies, only the VCP configuration gives a
symmetric and simple shape. In fact, the PERP
configuration is clearly asymmetric whereas the
HCP tends to suggest two resistive bodies when
there is only one. According to these results, the
VCP configurations have to be preferred for
towed survey. For all coil configurations, the
susceptibility contrast is too weak to create any
observable anomaly in the apparent susceptibility.
Archaeol. Prospect. 16, 103–109 (2009)
DOI: 10.1002/arp
Detection of Resistive Features Using Towed Slingram
107
Figure 4. Calculated in-line apparent resistivity variation above a two-dimensional wall measured with a Slingram: (a) PERP,
(b) HCP an (c) VCP configured instrument for different values of the intercoil spacing from 0.4 m to1m.
Copyright # 2009 John Wiley & Sons, Ltd.
Archaeol. Prospect. 16, 103–109 (2009)
DOI: 10.1002/arp
J. Thiesson et al.
108
appears, this is also the case for the thin resistive
internal feature. The small feature to the east and
the outer wall of the fanum are less marked than
on d.c. ra maps but recognizable. The noise that
can be observed results partly from the errors in
restored positions. In fact, a true distance
measurement by coding wheel or by Doppler
radar would be better than a regular time step.
The presence of many small metallic (mainly iron
resulting from World War II bombing) pieces in
the topsoil generates outlier values that exist on
both the ra and kpha maps and have been
observed previously in the magnetic field map
(Aubry, 2003).
Figure 5. CS60 apparent magnetic susceptibility map of the
test area.
Field Results
The data have been acquired by pulling the
instrument on N–S profiles (1 m apart) following
a zigzag route. The kpha map, Figure 5, shows that
the cella corresponds to a slightly less magnetic
area, suggesting the existence of a floor. By
comparison the zone just north of the cella
appears more magnetic. On the ra map (Figure 6)
the range of values is more limited than in the d.c.
ra prospecting but the wall of the cella clearly
Figure 6. CS60 apparent electrical resistivity map of the test
area.
Copyright # 2009 John Wiley & Sons, Ltd.
Conclusions
In archaeological surveying, frequency domain
Slingram EMI instruments are recognized as
light and flexible tools able to map both kph and s,
but commonly not considered for resistive
feature mapping. The test presented here
indicates that, without considering the abilities
of multidepth electrical prospecting, they can
succeed in resistive feature detection if suitable
coil orientation and spacing, in accordance with
the expected depth of the features, are respected.
The advantages and limitations of these instruments that can (and must) be assessed using
three-dimensional modelling are different, and
better for archaeological applications, from those
expected from using only one-dimensional
modelling.
The present example is one of the first
published results where a Slingram instrument
is used for resistive archaeological feature
mapping and other experiments must be undertaken to improve these conclusions. However,
commercially available Slingram instruments
exist, some of them can be used in PERP or
VCP coil configurations. They are light, of low
power consumption and can be transported
easily everywhere. The fact that, in addition to
their ability to simultaneously measure magnetic
susceptibility and electrical conductivity, resistive feature detection is possible, leads us to
consider a future significant increase of their
application in archaeological surveying.
Archaeol. Prospect. 16, 103–109 (2009)
DOI: 10.1002/arp
Detection of Resistive Features Using Towed Slingram
Acknowledgements
This experiment was performed within the
framework of the Agence Nationale de la
Recherche (ANR), number ANR-06-BLAN-0049
Celtecophys contract that aims to improve the
capabilities of Slingram instruments in order to
design, build and test a new combined magnetic/
electromagnetic multisensor instrument for
extensive in-field surveys.
References
Aubry L. 2003. Acquisition traitement et restitution des
données d’une reconnaissance archéologique: la ville
gallo-romaine du Viel-Evreux. Thèse Université
Pierre et Marie Curie, Paris, 259 pp. [in French]
Dabas M, Tabbagh A. 2003. A comparison of EMI
and DC methods used in soil mapping – theoretical considerations for precision agriculture. In
Precision Agriculture, Stafford J, Werner A (eds).
Academic Publisher: Wageningen; 12–129.
Dabas M, Guyart L, Lepert T. 2005. Gisacum revisité. Dossiers archéologie et sciences des origines 308:
52–61. [in French]
Frischknecht FC, Labson VF, Spies BR, Anderson
WL. 1991. Profiling methods using small sources.
In Electromagnetic Methods in Applied Geophysics,
Vol. 2, Nabighiam MN (ed.). Society of Exploration Geophysicists: Tulsa, OK; 15–270.
Geonics Limited. 1996. Application of ‘Dipole–
Dipole’ Electromagnetic Systems for Geological
Depth Sounding. Technical Note TN-31 (http://
www.geonics.com/html/technicalnotes.html).
Copyright # 2009 John Wiley & Sons, Ltd.
109
Job JO, Tabbagh A, Hachicha M. 1995. Détermination par méthode électromagnétique de la concentration en sel d’un sol irrigué. Canadian Journal
of Soil Science 75: 463–469. [in French]
Musson CR. 1968. A geophysical survey at South
Cadbury Castle, Somerset, using the soil conductivity anomaly detector. Prospezioni Archeologiche
3: 101–104.
Panissod C, Dabas M, Jolivet A, Tabbagh A. 1997.
A novel mobile multipole system (MUCEP) for
shallow (0–3 m) geoelectrical investigation: the
‘vol de canards’ array. Geophysical Prospecting
45: 983–1002.
Tabbagh A. 1974. Définition des caractéristiques
d’un appareil électromagnétique classique
adapté à la prospection archéologique. Prospezioni
Archeologiche 9: 21–33. [in French]
Tabbagh A. 1986a. Applications and advantages
of the Slingram electromagnetic method for
archaeological prospecting. Geophysics 51: 576–
584.
Tabbagh A. 1986b. What is the best coil orientation
in the Slingram electromagnetic prospecting
method? Archaeometry 28: 185–196.
Tabbagh A, Verron G. 1983. Etude par Prospection
électromagnétique de trois sites à dépôts de l’Age
du Bronze. Bulletin de la Société préhistorique de
France 80: 375–389. [in French]
Tite MS, Mullins CE. 1969. Electromagnetic prospecting, a preliminary investigation. Prospezioni
Archeologiche 4: 95–102.
Tite MS, Mullins CE. 1970. Electromagnetic
prospecting on archaeological sites using a
soil conductivity meter. Archaeometry 12: 97–
104.
Archaeol. Prospect. 16, 103–109 (2009)
DOI: 10.1002/arp