ARTICLE IN PRESS

Journal of Archaeological Science xxx (2010) 1e10

Contents lists available at ScienceDirect

Journal of Archaeological Science

journal homepage: http://www.elsevier.com/locate/jas

Marine electrical resistivity tomography for shipwreck detection in very shallow

water: a case study from Agropoli (Salerno, southern Italy)
Salvatore Passaro*
Istituto per lAmbiente Marino Costiero (IAMC), Consiglio Nazionale delle Ricerche, Calata Porta di Massa, Porto di Napoli, 80133 Napoli, Italy

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

Article history: A marine electric resistivity survey was carried out over a submerged beach along the Agropoli shore
Received 31 July 2009 (Salerno, Italy) to detect buried objects of archaeological interest below the sandy seabed. We found
Received in revised form a shipwreck, a military vessel that probably sunk during the Salerno landing operations of the allied
24 January 2010
forces in the Second World War. Resistivity data provide information on the vertical and horizontal
Accepted 17 March 2010
extension of the shipwreck, which is characterized by very low calculated resistivity values (about 2e5
ohm m). Such values differ signicantly from the sand and the bedrock values (5e40 ohm m). Although
Keywords:
the presence of the shipwreck is clearly visible from geoelectric data, the joint application of electric,
Marine resistivity measurements
DEM
magnetic and multibeam bathymetric techniques reduces the ambiguities inherent in each method. As
Magnetics shown in the Electrical Resistivity Tomography (ERT) and conrmed by the Digital Elevation Model
Archaeological target detection (obtained from the processing of bathymetric data), the shipwreck extends more than 30 m in NEeSW
direction and it is about 13 m wide. The global extension of the relic is consistent with the magnetic data,
that are characterized by a magnetic anomaly with an amplitude of about 1800 nT and similar dimension,
as inferred from the estimation of source boundaries obtained from the computation of the analytic
signal. The results of our survey encourage the use of marine geoelectrical methods for the detection of
buried archaeological targets, particularly in locations where the use of seismic prospecting is not
effective (e.g. very shallow water with sandy sea-bottoms). The integration of different geophysical
methods allows to better dene the extension, depth and thickness of buried objects, suggesting that
such an approach is the most effective for underwater archaeological investigations.
2010 Elsevier Ltd. All rights reserved.

1. Introduction economic importance of archaeological ndings (Goodheada and

Aygen, 2007), coastal zone management and related planning
Over the last 20 years, geophysical surveys have become makes it necessary to implement preliminary studies over the
fundamental to detect and locate buried archaeological targets, intervention zone. Exposed submerged objects are easily detected
particularly on land. Buried man-made features and objects, such as by acoustic systems such as multibeam or side-scan sonar (Quinn
walls, metals, etc., usually produce geophysical anomalies which et al., 2000, 2005; Momber and Geen, 2000). These systems are
are recordable at the surface. Furthermore, the increasing care to highly effective and allow high-resolution mapping of the seaoor
a sustainable management of the coastal zone, makes the recog- in terms of morphological (multibeam) or backscatter (side-scan
nition of targets in very shallow water an essential objective for sonar) signatures. Conversely, the recognition of buried archaeo-
regions that have a particularly important archaeological past logical targets in marine environment is still not easy because of the
(Vallega, 2003; Fletcher et al., 2007). It is worth stressing that the difculties to acquire sub-surface data with very high resolution,
emplacement of coastal engineering structures (e.g. offshore both vertical and horizontal, as required for archaeological surveys.
breakwaters, rubbles, etc.) strongly inuences the long-shore Standard acoustic methods are difcult to be used in very shallow
dynamics of sediment transport and may result in irreversible water (depth <20 m) and, therefore, they are not particularly
burial of archaeological nds under a blanket of sediments. suitable for archaeological purposes. As a matter of fact, multi-
Therefore, it becomes essential to determine the possible presence channel reection seismic investigations provide fast and accurate
of archaeological targets prior to seabed developments. Due to the results when appropriate, specic instruments are used (e.g. SEA-
MAP-3D ultra-high-resolution acquisition system; see Muller et al.,
* Tel.: 39 81 5423831; fax: 39 81 5423888. 2009), while standard seismic data acquisition systems and data
E-mail address: [email protected] processing are difcult to carry out in very shallow water. Other

0305-4403/$ e see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2010.03.004

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

2 S. Passaro / Journal of Archaeological Science xxx (2010) 1e10

mono-channel seismic systems have poor resolution (e.g., sparker tomography with data derived from acoustic and magnetic surveys:
systems; see Sheriff, 1977), or are scarcely penetrative, on coarse i.e., multibeam swath bathymetry for seaoor mapping and
grain sandy seaoor (e.g. 3.5 kHz sub-bottom, 2e7 kHz chirp magnetics. Such methods detect buried archaeological objects
modulated systems). In any case, reection seismic is still thought when they are characterized by a contrast in the magnetic
the most affordable prospecting method for the detection of buried susceptibility with respect to the surrounding sediments.
marine archeological targets (e.g. Belfer et al., 1998; Quinn et al.,
2002; Lafferty et al., 2006; Plets et al., 2009). 2. Geological setting
Electrical Resistivity Tomography (hereafter ERT) is widely used
for terrestrial archaeological purposes, since it is suitable for The Salerno Plain (Tyrrhenian Sea, Fig. 1) is a Plio-Pleistocenic
detecting walls, caves, etc., at different depths (Leucci et al., 2007, peri-Tyrrhenian basin extended between the Cilento Promontory
Compare et al., 2009). In particular, the method is used coupled to the south and the Sorrento Peninsula to the North. Multi-
with other techniques, e.g. with Ground Penetrating Radar (GPR) channel seismic proles and data from deep boreholes (see Sacchi
data in integrated geophysical surveys (Conyers and Goodman, et al., 1994) show that the total thickness of the sediments of the
1997; Pipan et al., 1999; Nuzzo et al., 2002, 2009; Leucci, 2006), Salerno Bay consists in more than 1700 m, with an average
and has been widely used for the discovery of anthropogenic subsidence rate of 24 cm/1 ka during the post MioceneeMiddle/
objects. In this sense, it is a useful tool for planning and conducting Upper Pleistocene time interval. The shelf morphology is charac-
archaeological excavation. Marine electrical resistivity methods terized by a considerable tectonic inuence, as shown by the
(Baumgartner, 1996; Baumgartner and Christensen, 1998; Yang shape of the shelf break. This is located at a distance of 2e12 km
et al., 2002) like their terrestrial equivalents, exploit the contrasts from the coast in the northern sector, while moves farther
in material properties to identify possible target (anomaly) loca- southward (15e25 km). Such a pattern of the shelf break is due to
tions. Recently, electrical methods have been particularly used to the presence of a complex fault system that deeply cuts the
clarify the spatial (and temporal) distribution of freshwater and substrate, thus controlling the shelf break position and the
saltwater in coastal settings (see Henderson et al., 2010, and shelfeslope system morphology (see Insinga et al., 2008, and
references therein). The marine electrical resistivity method is references therein). The Agropoli coastline is situated in the
known since 30 years, but its applications (and continuous southernmost sector of the Salerno Gulf, mainly formed by
measurements in particular) are actually rare, as conrmed by the Mesozoic carbonates and Miocene ysches, but Holocene marine
lack of references in literature. sandstones, together with beach sands, also crop out. A Holocene
This paper reports a multidisciplinary survey over a submerged high-stand depositional system, related to the rising up of the sea
beach along the Agropoli shore (Salerno, Italy), with sand at the level during the post-Wurmian transgressive phase, lies on the
sea-bottom, carried out by a research team of the IAMC-CNR of clinoform system. This high-stand interval is essentially made up
Naples. The survey has been conceived to explore the reliability of of mud, silt and clay, but tephra layers, mainly related to the
marine continuous electrical resistivity method as an alternative, or eruptions of the Mt. Vesuvio, are also present. Shallow biogenic
a complementary activity, to the seismic prospecting methods for gas facies trapped by mud layers have also been recorded in the
archaeological prospection in very shallow water. high-resolution single-channel seismic proles (Budillon et al.,
In order to obtain information on the vertical distribution of the 2005) from the area.
resistivity anomaly, the resistivity tomography was carried out The historical importance of Agropoli is testied by many
along proles. 3D imaging, based on the 2D tomographic inversion archaeological ndings, the more ancient dated 7th Century BC.
of the proles, permits to represent the rst 6e8 m of the sub- The area is located at a distance of 5 km from the important
surface in terms of calculated resistivity. In order to assess the GreekeRoman site of Paestum. Allied invasion of Italy during the
quality of the resistivity measurement, we compared electrical World War II, started in this coastline sector.

Fig. 1. Location map of the acquisition zone. Datum is WGS84, projection is UTM (Zone 33). A) the Campanian region. The survey area is enclosed in the small black-lled rectangle.
B) Detail of the survey area. The black-lled polygon represents the survey area, where a coastal defense structure is actually under construction.

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

S. Passaro / Journal of Archaeological Science xxx (2010) 1e10 3

Table 1 compared to the sub-seaoor materials. On the other hand, the

Calendar of survey activities chosen conguration has two decisive advantages: 1) it is easier
Date/ Activity Shipwreck conditions and faster to carry out and 2) avoids problems such as the cable
period being snagged by obstacles on the seaoor. The marine cable was
8e9 June Multibeam survey (Reson Partially outcropping suspended at 1 m depth (Fig. 2). Proles reached a depth of about
2006 Seabat 8125) 6e8 m below the surface of the sea, inline spacing was 2 m. The
26e27 Magnetic survey (SEASPY Partially outcropping (less than it
dataset consists of 15 proles, 12 of which approximately parallel to
June Marine Magnetics Corporation) was during the multibeam survey)
2006 the coastline (roughly NeS oriented) and 3 tie-lines (roughly EeW).
24e26 Marine electrical resistivity Covered by sediments A total of more than 11 km of proles have been acquired, with
July (Sysca Pro Switch Iris a signicant overall coverage (5  5 m of grid cell size) of
2006 Instruments) 105,000 m2.
Spring, Diving Outcropping (also including the
2007 pivoting door)
The proles have been processed and interpolated to obtain
a plan view of the resistivity anomalies. Data processing was carried
out using the RES2Dinv software. The apparent resistivity values
3. Data and methods are calculated by using a forward modelling subroutine. In order to
convert apparent resistivity in real resistivity values, a non-linear
3.1. Positioning least-square optimisation technique has been used for the data
inversion. The technique is based on the smoothness constrained
The data acquisition (Table 1) was carried out in three different least-squares method (Loke and Barker, 1996; Loke and Dahlin,
surveys executed onboard of the Redsh vessel (Teknomar), 2002). At this point, the obtained tomographic proles were
equipped with a Fugro Seastar DGPS system. To geo-reference and interpolated by using the kriging algorithm, in order to produce
compare magnetics, electric resistivity and multibeam data, horizontal slices of the resistivity anomalies at different depth
a highly accurate positioning is necessary. Sub-meter accuracy is levels (Fig. 3). Final voxels was 5 m (x)  5 m (y)  1 m (on the z
available and highly reliable with GPS systems that are equipped for axis).
real-time kinematic corrections. During the surveys, occasional
losses of the standard NMEA GGA string transmission occurred, due
to failed connection between the cable and the acquisition unit. The
strings were removed manually by using the standard positioning
tool of the PDS2000 navigation software. In any case, Positional
(3D) Dilution Of Precision (PDOP), which is a standard parameter
used to classify GPS precision, resulted to be 0.9 on average. Such
a value of PDOP ensures a sub-metrical precision (for the on-keel
instruments) during the survey.

3.2. Marine electrical resistivity survey

The continuous marine resistivity data acquisition was carried

out using the IRIS-SYSCAL PRO switch System, equipped with a
submarine cable with graphite connectors. This system allows
a continuous transfer of data to a laptop computer by a serial or
a USB link, and a real-time display of the apparent resistivity pseudo-
section and of the voltage of each channel. Obtained measurements,
together with the GPS data, are automatically stored into a binary
le.
We used a towed mobile conguration with a 2 m, horizontally
spaced, multi-electrode marine cable (Fig. 2). A serial link between
the instrument and the central computer unit was used. Through
the Sysmar software (IRIS instruments), we were able to receive
and record real-time images of resistivity pseudo-sections. The
marine towed resistivity cables can be deployed in two different
congurations: either by dragging a cable behind a boat or by using
oating electrodes on the water surface (Loke, 1996). For this study,
we used a conguration with oating electrodes, which is prob-
lematic when there is a larger current-ow within the water layer

DGPS

M4 M3 M2 M1 A B N1 N2 N3 N4

1m
Sea surface Seafloor
Marine cable
Fig. 3. Electrical resistivity tomographic slices, obtained through the inversion of
Fig. 2. Schematic sketch of the array geometry used for the acquisition of continuous apparent resistivity pseudo-sections at 1e6 m depth. No-data values correspond to
resistivity proles. A and B are the Current electrodes, and M and N are the light grey in the color scale, while white values refer to measurements falling inside
measurement electrode pairs. the column water (r < 0.25 U m). Navigation patterns are shown in A.

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

4 S. Passaro / Journal of Archaeological Science xxx (2010) 1e10

3.3. Multibeam swath bathymetry 10 m by using the Kriging algorithm. The analytic signal (Nabighian,
1972; Roest et al., 1992) helped us in the identication of the
Multibeam data were collected with a Seabat 8125 (Reson- horizontal boundary and in the determination of the exact location
Thales) echo-sounder, characterized by 240 beams, 120 of swath of the magnetic sources. The analytic signal amplitude, A(x,y), is
coverage and a pulse frequency of 455 kHz. The system was cali- calculated by taking the square root of the sum of the squares of
brated with an SVP15 sound velocity probe (Navitronics), and each of the three directional rst derivatives of the magnetic eld, T,
completed by an OCTANS VRU1000 (gyrocompass plus motion as follows:
sensor), that allowed a real-time cinematic correction of the vessel
asset, thus enabling a correct swath re-positioning at every 0.1 s. jAx; yj dt=dx2 dt=dy2 dt=dz21=2
The bathymetric data consist of 70 proles, roughly parallel to the The calculation, based on the Hilbert transform, was carried out by
coastline (NS direction). The multibeam data records were pro- using the appropriate U.S.G.S. FORTRAN routine (Phillips et al.,
cessed with a replay of the swaths, according the IHO standards 1993). In the absence of high-frequency noise, the local peaks are
(IHO, 1998). Additionally, erroneous measurements have been interpreted as corners of source, and provide an accurate horizontal
removed, and the records exported as xyz triplets. Subsequently, locating for the magnetic sources.
they were converted from scattered measurement to an equally
spaced matrix with a grid cell size of 20  20 cm, where, at each
squared cell, we associate the average value of beams that falls 4. Results
inside the cell itself. We completed the grid by simply associating
a no-data value to empty cells. Note that multibeam footprint The electrical resistivity tomographic slices (Fig. 3), obtained
dimension depends on technical specics of the instrument and through the inversion of apparent resistivity pseudo-sections, show
depth, according the equation: two differently increasing resistivity trends. The rst with depth
and the second in the EeW direction. The increase of the resistivity
Nf D  tgSO=NB values with depth is due both to the passage from the water column
(r z 0.25 U m) to sandy and ysch sediments (4e30 U m) (that
where: Nf Nadir footprint, D depth, SO Swath Opening (in contain water as well) than to the decrease of water content in
degrees), NB number of beams in the used array. In our case, the sediments. The variations in EeW direction are ascribable to the
average value of the footprint size was 3.5 cm, which is signicantly prograding nature of the beach morphology, that makes the higher
smaller than the resolution that is possible to obtain by using resistivity pixels appear at shallower depth (2 m) in the eastern
a DGPS positioning system. To reach such a resolution (in the order sector compared to the western one. The resistivity values follow
of 3e4 cm of grid cell size), a more accurate positioning system is the same, almost homogeneous, trends, with the exception of the
required (e.g., an RTK dynamic positioning system). The obtained northernmost sector of the study. Here, a low resistivity anomaly
grid was characterized, anyway, by a 20  20 cm grid cell size. Such occurs. This anomaly starts from 2 m and is clearly visible at least
result conrms the sub-metrical precision of the positioning of our down to 5 m of depth.
survey, as shown by the absence of mismatches between contig- The inverted proles conrm these general trends for resistivity
uous multibeam swaths acquired over the internal structures of the (Fig. 4). The low resistivity anomaly detected is characterized by
shipwreck. Swath opening, which is 3.5  depth for the used reson very low values (roughly the same of the water column) and by
seabat 8125 multibeam, ranged 3e10 m during the survey. The a sharp vertical boundary. These boundaries separate the anomaly
total coverage amounts to 100,000 square meters. area by the normal sedimentary contest, which is characterized, in
this point, by relatively high resistivity values (25e30 U m). A very
3.4. Magnetics small stripe occupies the lateral vertical transition between the
resistive and the conductive bodies. Four sections are characterized
The marine magnetic method is widely applied to discover by the presence of this anomaly. We have tried to map the
buried targets (e.g., Quinn et al., 2002; Boyce et al., 2004; Missiaen conductive area by simply picking the planar coordinates corre-
and Feller, 2008). Nevertheless several problems could arise from sponding to the steepest gradients on the proles (indicated by
the presence of anthropic manufacts, such as pipes. Geomagnetic arrows in the proles of Fig. 4), hence dening and mapping an area
surveys were carried out using the SeaSpy magnetometer (Marine having an extension of about 25/30 m along NEeSW and 13e15 m
Magnetics Corporation), with an absolute accuracy of 0.2 nT. About along NWeSE (Fig. 4).
8500 m of trackline magnetic data were acquired at an average The bathymetric range founded in the area varies between
sampling rate of 1 s. The sensor was towed at a minimum distance 0.98 m and 3.54 m, and shows the presence of an NS trending
of 20 m from the ship, and the cruise speed did not exceed 5 knots. relative high and three depressions of about 1 m. The shaded relief
The magnetometer was towed at 1 m depth below the sea level by of the northernmost area of the survey shows several NEeSW and
using two oating buoys. The eld measurements were acquired NWeSE almost linear trending features, emerging for 5 cm on
during a relatively quiet magnetic period. A total of 13 magnetic average from the seaoor (Fig. 5C). Magnetic data show a non-
proles (6 roughly in the NS direction and 7 in the EW) were magnetic response of the seaoor for the whole acquisition area,
recorded. Data have ensured a coverage of 130,000 square meters, with the exception of a 1800 nT anomaly detected in the northern
with a 10  10 m grid cell size. The magnetic data were pre-pro- sector (Fig. 6A,C). Planar source boundaries of this anomaly are
cessed as follows: 1) Navigation and Layback correction, consisting identied by the computation of the analytic signal (Fig. 6Band D),
in a re-positioning of each single measurement made by accounting and shows a circular shape having a diameter of 40 m.
for the distance between the DGPS antenna and the magnetometer
tow-sh (20 m) 2) Diurnal variation, carried out taking into account 5. Discussion
data referred to the LAquila (Italy, 42 N; 12 E) Geomagnetic
Observatory of the INGV. And 3) De-noising, consisting in a tie-line The low resistivity calculated anomaly located to the northern
levelling, done on the basis of the computation of shifts on cross- sector of the survey area roughly coincides with the bathymetric
over check values. Cross-over shifts errors resulted to be <4 nT on anomaly of the seaoor (Fig. 5) and with the 1800 nT anomaly
average. Pre-processed data were then gridded to an interval of detected in the residual magnetic anomaly eld map (Fig. 6), and

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

S. Passaro / Journal of Archaeological Science xxx (2010) 1e10 5

Fig. 4. Navigation map and picking of the coordinates (latitude and longitude) corresponding with low resistivity anomaly in inverted resistivity proles. Asterisks in the A frame
correspond to the vertical stripes detected in four proles (indicated arrows). The union of these points (lled polygon in leftmost frame) denes an area having an extension of
about 25/30 m along NEeSW, and 13e15 m along NWeSE. Datum is WGS84, projection is UTM (Zone 33).

correspond to the presence of an iron-made shipwreck, that is Photos taken during the bathymetric survey (Fig. 7A) and during
probably a landing military vessel sunk during the World War II, diving operations (Fig. 7C) suggest that the wreck could belong to
in the course of the Salerno operations. Such hypothesis was the LCT (Land Craft, Tank) class of amphibian naval vehicles. This
conrmed by successive diving carried out on the wreck (Fig. 7). is based on the presence of a well preserved pivoting door

Fig. 5. DEM (Digital Elevation Model). A: Navigation map; B: Image map of DEM, C: Shaded relief of the frame enclosed in the small rectangle in B. Linear trending features
emerging about 5 cm from the seaoor are visible. D: Comparison between DEM (dashed line) and resistivity (lled polygon) boundary anomalies. Datum is WGS84, projection is
UTM (Zone 33). Vertical datum (for depths) is the mean sea level.

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

6 S. Passaro / Journal of Archaeological Science xxx (2010) 1e10

Fig. 6. Magnetic data. A: residual magnetic anomalies and navigation patterns. B: Analytic signal. C: zoomed area of the frame indicated by the arrow in A. D: zoomed area of the
frame indicated by the arrow in B. E: comparison between Analytic Signal (dashed line) and resistivity (lled polygon) derived boundaries of the anomaly. Datum is WGS84,
projection is UTM (Zone 33).

detected on the SW side of the shipwreck and by the extension of The uppermost coverage of the vessel is absent, and a part of the
the wreck (Fig. 7C). Such hypothesis was also suggested by the internal structure of the vehicle is exposed. Small emergencies, that
technicians of the Soprintendenza ai Beni Archeologici e Culturali are visible on the DEM in correspondence of the shipwreck, are
di Salerno, Benevento, Avellino e Caserta, which is the local related to the exposure of this internal structure, that rises about
organization responsible for the safeguard of cultural heritage. 5 cm above the seaoor (Figs. 5C and 7A). This fact might be related
The northern shore of Agropoli was subjected to the allied inva- to the proximity of the relic to the coast that enabled the local
sion (Operation Avalanche) by the U.S. 5th Army starting on 9 inhabitants to strip the uppermost coverage of the craft during
September, 1943. The german mobile defenses in this zone the years successive to the sinking, since they used it as an iron
included a railway battery mounting 132-mm guns, kept on reservoir. Although the target was already clearly visible from
a track just north of Agropoli. The observation post placed on the resistivity data, a multiple approach strategy helps to reduce the
dominating terrain features enabled to re on the gulf, the bea- ambiguities inherent in each method and provides additional
ches, and the plain. The allied attack started without a prelimi- information. The approach benets, in particular, of coupling
nary naval or aerial bombardment, in the attempt to produce different data that rely on various techniques for the detection of
a (failed) surprise effect, and many US and British vehicles were archaeological targets (De Domenico et al., 2006; Negri and Leucci,
sunk. Other relicts, including a rare amphibian Sherman Double 2006; Leucci et al., 2007; Piro et al., 2007; Di Fiore and Chianese,
Drive Tank, have been found both onshore and (mainly) offshore 2008). The proles extracted from the resistivity horizontal slices
over the past 20 years all along the Salerno Gulf. at different depths, the bathymetric prole extracted by DEM and

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

S. Passaro / Journal of Archaeological Science xxx (2010) 1e10 7

Fig. 7. Photos of the relic detected during the survey. A) Wreck internal structures, acquired by the vessel; B) An image of an LCT (Land Craft, Tank) MKV (the shipwreck detected in
Agropoli should belong to this class of military vehicles, used during the avalanche operation during the WWII). C) Photo acquired during a diving more than 1 year later. The
pivoting door is almost completely emerged from the surrounding sediments.

the proles extracted by the analytic signal (the inversion algo- literature (Quinn, 2006). The internal structures of the wreck are
rithm applied to magnetic data) are compared, in order to enhance clearly visible on DEM. The presence of the shipwreck is high-
the presence of common features. Two anomalies, marked as AN1 lighted by all the geophysical equipments used, but in the time gap
and AN2 (in Fig. 8), appear in the proles extracted by ERT slices between different geophysical surveys (JuneeNovember 2006), the
data, but AN1 is unrelated to an intense magnetic anomaly, as in the wreck has been subjected to phases of partial and/or almost
case of AN2 (which corresponds to the shipwreck). In addition, AN2 complete, burial by sand (Table 1).
shows a closer correlation between the bathymetric data and the Surf zone dynamic studies demonstrate that the presence of
resistivity anomaly. Further excavations in correspondences of the long-shore and rip currents, their reciprocal reports and their
AN1 anomaly detected the presence of lens of blast material. mutual balancing (due to the prevailing sea, the interaction
The correlation between resistivity and bathymetric proles between waves, refraction phenomenon and their construction/
which crosses the shipwreck shows the difference between the disruption mutual reports, tides, etc.) may lead to the emplace-
electric response of the shipwreck itself and the adjacent litholo- ment of baretrough pairs similar to what happens in our case.
gies (Fig. 9). The horizontal axis shows the distance (in meters) Lateral bar continuity can be interrupted by the presence of
from the starting point. It is also clearly visible a resistivity peak sectors in which erosional necks, generally roughly orthogonal to
that has signicantly higher values. This peak corresponds to the coastline are emplaced (typically, the strongest part of a rip
a lateral lithologic variation, due to the presence of poorly covered current is directed from the beach line to the sandy bar opening).
bedrock that outcrops a few meters northward of the relic. It is interesting to note that the seaoor shapes (ripples) associ-
Finally, three sedimentary depressions, roughly 1 m in depth, ated to a NeS movement of the current was present at the
are clearly visible on DEM. Two of them formed near a brush of moment of the survey (Fig. 10D). The emplacement of
a coastal protection structure, currently under construction. Their a submerged coastal engineering structure, which was under
formation may be due to along-shore changes of sediment transfer construction, inuenced the sediment pathway. An increase of the
rate that are induced by the presence of this construction (Figs. 5 rip current in correspondence of the cut made on the
and 10). Two sandy bars are also visible (the changes of sediment submerged breakwater, left to preserve the ship, may be a result
transport always produces gap/accumulation pairs). The shipwreck of such activity. One year later, the pivoting door of the vessel,
outcrops in the northernmost depression, and its presence modies which was not visible at the moment of the survey on DEM, was
the normal sediment transfer development, as already known in almost completely emerged.

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

8 S. Passaro / Journal of Archaeological Science xxx (2010) 1e10

Fig. 8. Comparison among proles extracted from DEM (A), marine resistivity slices at 3 m (B), 4 m (C) and 5 m (D), and from the analytic signal (E). The bold line in A, B, C, D and E
maps is the prole location. Color scale 1 is relative to A, 2 to B, C and D, 3 to E. Extracted proles are overlaid in the F frame (SN direction). Two common anomalies, marked as AN1
and AN2, are revealed. See text for details.

Fig. 9. Correlation between the inverted prole L8_N and the corresponding bathymetric prole extracted from DEM. Positioning of the prole is reported in B.

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

S. Passaro / Journal of Archaeological Science xxx (2010) 1e10 9

Fig. 10. A) An example of typical patterns of long-shore and rip currents. B) Typical baretrough bathymetric prole. C) Zone of the emplacement of the coastal defense submerged
structure, actually under construction, and inferred patterns of the old system of rip currents. Actually, the northernmost rip current (enclosed in the rectangle in C and highlighted
in D) may be faster, due to the emersion of the pivoting door of the shipwreck (see also Fig. 7).

6. Conclusions Baumgartner, F., Christensen, N.B., 1998. Analysis and application of a non-
conventional underwater geoelectrical method in Lake Geneva, Switzerland.
Geophysical Prospecting 46, 527e541.
The marine electrical resistivity measurements represent a valid Belfer, I., Bruner, I., Keydar, S., Kravtsov, A., Landa, E., 1998. Detection of shallow
alternative to seismiceacoustic prospecting for the detection of objects using refracted and diffracted seismic waves. Journal of Applied
archaeological targets in very shallow water, particularly in marine Geophysics 38, 155e168.
Boyce, J., Reinhardt, E.G., Raban, A., Pozza, M.R., 2004. Magnetic survey of
sectors with sand at the seabed. This method can be used to outline a submerged Roman harbour, Caesarea Maritima, Israel. The International
conductive and resistive bodies, like metallic objects and wood, Journal of Nautical Archaeology 33 (1), 122e136.
respectively. In this study, we were able to detect and provide Budillon, F., Violante, C., Conforti, A., Esposito, E., Insinga, D., Iorio, M., Pordo, S.,
2005. Event beds in the recent prodelta stratigraphic record of the small ood-
morphological information of a military vessel. The use of prone Bonea Stream (Amal Coast, Southern Italy). Marine Geology 222e223,
combined geophysical techniques ensures an accurate assessment 419e441.
of the anomalies of the survey area and helps to reduce the ambi- Compare, V., Cozzolino, M., Mauriello, P., Patella, D., 2009. Three-dimensional
Resistivity Probability Tomography at the prehistoric site of Grotta Reali (Mol-
guities inherent in each method. We found a remarkable consis- ise, Italy). Archaeological Prospection 16, 53e63.
tency between the anomalies emerged from the different datasets. Conyers, L.B., Goodman, D., 1997. Ground-penetrating RadardAn Introduction for
The DEM obtained from bathymetry claried the role of the ship- Archaeologists. Alta Mira Press, Walnut Creek.
De Domenico, D., Giannino, F., Leucci, G., Bottari, C., 2006. Integrated geophysical
wreck in the along-shore sediment transfer rate alteration, while surveys at the archaeological site of Tindari (Sicily, Italy). Journal of Archaeo-
the inverted magnetic data well circumscribed the shipwreck area. logical Science 33, 961e970.
A close correlation between the anomalies detected by the used Di Fiore, B., Chianese, D., 2008. Electric and magnetic tomographic approach for the
geophysical investigation of an unexplored area in the archaeological site of
methods (magnetic, marine resistivity and bathymetry) has been
Pompeii (Southern Italy). Journal of Archaeological Science 35, 14e25.
found only in the sector where the shipwreck is present. Marine Fletcher, S., Johnson, D., Hewett, T., 2007. Coastal management in the Solent: an
resistivity applications are not yet commonly used in geophysical introduction. Marine Policy 31, 585e590.
marine exploration, however the initial results of our surveys are Goodheada, T., Aygen, Z., 2007. Heritage management plans and integrated coastal
management. Marine Policy 31, 607e610.
promising to promote such a method as a valuable tool for the Henderson, R.D., Day-Lewis, F.D., Abarca, E., Harvey, C.F., Karam, H.N., Liu, L.,
detection of archaeological targets in very shallow waters. The Lane, J.W.J., 2010. Marine electrical resistivity imaging of submarine
method can be also used for sedimentological purposes. groundwater discharge: sensitivity analysis and application in Waquoit Bay,
Massachusetts, USA. Hydrogeology Journal 18, 173e185.
IHO Special Publ. N 44, 1998. Standards for Hydrographic Surveys, fourth ed.
Acknowledgements International Hydrographic Bureau, Monaco.
Insinga, D., Molisso, F., Lubritto, C., Sacchi, M., Passariello, I., Morra, V., 2008. The
proximal marine record of SommaeVesuvius volcanic activity in the Naples and
The author is especially grateful to Francesca Budillon (IAMC- Salerno bays, Eastern Tyrrhenian Sea, during the last 3 kyrs. Journal of Volca-
CNR), F. Cammarano (UC Berkeley), C. Iorio (ULB) and C. Leggieri nology and Geothermal Research 177, 170e186.
(Soprintendenza ai Beni Archeologici e Culturali di Salerno, Lafferty, B., Quinn, R., Breen, C., 2006. A side-scan sonar and high-resolution Chirp
sub-bottom prole study of the natural and anthropogenic sedimentary record
Benevento, Avellino e Caserta) for the precious discussion and of Lower Lough Erne, northwestern Ireland. Journal of Archaeological Science
suggestions, and to J. Boyce and another anonymous reviewer that 33 (6), 756e766.
greatly helped to improve early drafts of this paper. Leucci, G., 2006. Contribution of ground-penetrating radar and electrical resistivity
tomography to identify the cavity and fractures under the main church in
Botrugno (Lecce, Italy). Journal of Archaeological Science 33 (9), 1194e1204.
References Leucci, G., Greco, F., De Giorni, L., Maceri, R., 2007. Three-dimensional image of
seismic refraction tomography and electrical resistivity tomography survey in
Baumgartner, F., 1996. A new method for geoelectrical investigations underwater. the castle of Occhiola` (Sicily, Italy). Journal of Archaeological Science 34,
Geophysical Prospecting 44, 71e98. 233e242.

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004
 ARTICLE IN PRESS

10 S. Passaro / Journal of Archaeological Science xxx (2010) 1e10

Loke, M.H., 1996. Tutorial: 2-D and 3-D Electrical Imaging Surveys. Geotomo Soft- Piro, S., Peloso, D., Gabrielli, R., 2007. Integrated geophysical and topographical
ware, Penang, Malaysia. investigation in the territory of Ancient Tarquinia (Viterbo, central Italy).
Loke, M.H., Barker, R.D., 1996. Rapid least-squares inversion of apparent resistivity Archaeological Prospection 14, 191e201.
pseudosections by a quasi-Newton method. Geophysical Prospecting 44, Plets, R.M.K., Dix, J.K., Adams, J.R., Bull, J.M., Henstock, T.J., Gutowski, M., Best, A.I.,
131e152. 2009. The use of a high-resolution 3D Chirp sub-bottom proler for the
Loke, M.H., Dahlin, T., 2002. A comparison of the Gauss-Newton and quasi-Newton reconstruction of the shallow water archaeological site of the Grace Dieu
methods in resistivity imaging inversion. Journal of Applied Geophysics 49, (1439), River Hamble, UK. Journal of Archaeological Science 36, 408e418.
149e162. Quinn, R., 2006. The role of scour in shipwreck site formation processes and the
Missiaen, T., Feller, P., 2008. Very-high-resolution seismic and magnetic investiga- preservation of wreck-associated scour signatures in the sedimentary record-
tions of a chemical munition dumpsite in the Baltic Sea. Journal of Applied evidence from seabed and sub-surface data. Journal of Archaeological Science
Geophysics 65, 142e154. 33, 1419e1432.
Momber, G., Geen, M., 2000. The application of the Submetrix ISIS 100 Swath Quinn, R., Cooper, A.J.A.G., Williams, B., 2000. Marine geophysical investigation of
Bathymetry system to the management of underwater sites. The International the inshore coastal waters of Northern Ireland. International Journal of Nautical
Journal of Nautical Archaeology 29.1, 154e162. Archaeology 29, 294e298.
Muller, C., Woelz, S., Ersoy, Y., Boyce, J.I., Jokisch, T., Wendt, G., Rabbel, W., 2009. Quinn, R., Breen, C., Forsythe, W., Barton, K., Rooney, S., OHara, D., 2002. Integrated
Ultra high-resolution marine 2De3D seismic investigation of the Limantepe/ geophysical surveys of The French Frigate La Surveillante (1797), Bantry Bay,
Karantina Island archaeological Site (Urla, Turkey). Journal of Applied Co. Cork, Ireland. Journal of Archaeological Science 29, 413e422.
Geophysics 68 (1), 124e134. Quinn, R., Dean, M., Lawrence, M., Liscoe, S., Boland, D., 2005. Backscatter responses
Nabighian, M.N., 1972. The analytic signal of two-dimensional magnetic bodies with and resolution considerations in archaeological side-scan sonar surveys:
polygonal cross-section: its properties and use for automated anomaly inter- a control experiment. Journal of Archaeological Science 32 (8), 1252e1264.
pretation. Geophysics 37, 507e517. Roest, W.R., Verhoef, J., Pilkinton, M., 1992. 3-D analytic signal. Geophysics 57,
Negri, S., Leucci, G., 2006. Geophysical investigation of the Temple of Apollo 116e125.
(Hierapolis, Turkey). Journal of Archaeological Science 33, 1505e1513. Sacchi, M., Infuso, S., Marsella, E., 1994. Late PlioceneeEarly Pleistocene compres-
Nuzzo, L., Leucci, G., Negri, S., Carrozzo, M.T., Quarta, T., 2002. Application of 3D sional tectonics in offshore Campania (eastern Tyrrhenian Sea). Bollettino di
visualization techniques in the analysis of GPR data for archaeology. Annals of Geosica Teorica e Applicata 36, 141e144.
Geophysics 45 (2), 321e337. Sheriff, R.E., 1977. Limitations on resolution of seismic reections and geologic
Nuzzo, L., Leucci, G., Negri, S., 2009. GPR, ERT and magnetic investigations inside the detail derivable from them. In: Payton, C.E. (Ed.), Seismic Stratigraphy e
Martyrium of St Philip, Hierapolis, Turkey. Archaeological Prospection 16, 1e16. Applications to Hydrocarbon Exploration. American Association of Petroleum
Phillips, J.D., Duval, J.S., Ambrosiak, R.A., 1993. National geophysical data grids: Geologists Memoir 26, 3e14.
gamma-ray, gravity, magnetic, and topographic data for the conterminous Vallega, A., 2003. The coastal cultural heritage facing coastal management. Journal
United States. U.S. Geological Survey Digital Data Series DDS-9, 1 CD-ROM. of Cultural Heritage 4, 5e24.
Pipan, M., Baradello, L., Forte, E., Prizzon, A., Finetti, I., 1999. 2-D and 3-D processing Yang, C.H., You, J.I., Lin, C.P., 2002. Delineating lake bottom structure by resistivity
and interpretation of multi-fold ground penetrating radar data: a case history image proling on water surface. Terrestrial Atmospheric and Oceanic Sciences
from an archaeological site. Journal of Applied Geophysics 41, 271e292. 13 (1), 39e52.

Please cite this article in press as: Passaro, S., Marine electrical resistivity tomography for shipwreck detection in very shallow water: a..., J.
Archaeol. Sci. (2010), doi:10.1016/j.jas.2010.03.004