Content uploaded by Jamon Van Den Hoek
Author content
All content in this area was uploaded by Jamon Van Den Hoek on Nov 07, 2015
Content may be subject to copyright.
INTRODUCTION
is chapter weighs the results of the remote sensing pro-
gram at Zeugma against topographical features of the
city discovered in the rescue excavations, and it makes use
of unpublished reports provided by two contractors who
carried out independent surveys: Stratascan Geophysical
and Specialist Survey Services and GSB Prospection.1 Geo-
physical survey at Zeugma commenced aer rescue exca-
vation had already begun. Nine surveys (Surveys A–I) were
conducted between zones of active excavation, and one of
these, Survey A, was tested by excavation towards the end
of the rescue project (g. ). Most survey areas, like the
areas of excavation, are now underwater and thus contain
information about an irrecoverable resource, but the data
collected provides vital points of connectivity for recon-
struction of the ancient city plan. For example, results of the
geophysical surveys corroborate archaeological evidence
for placing the terraced promontory that overlooked the
Euphrates River from the city center at the core of the Hel-
lenistic town. Unlike streets in other parts of the city, align-
ments in this sector match those on the opposite bank at
Apamea. e bridgehead was almost certainly here, on the
riverbank below the promontory at Zeugma. While Apa-
mea diminished in Roman times, Zeugma ourished, and
residential districts sprang up along the river’s west bank
on both sides of the promontory. e geophysical surveys
reveal dierent street and house alignments for these resi-
dential districts, and in some cases even specic boundar-
ies between zones of development. is data enhances our
ability to reconstruct the development of Zeugma’s city
plan and to chart functional and topographical relation-
ships between Apamea and Zeugma through time.
Also important are topographical connections sug-
gested by the geophysical data for parts of the city now be-
neath the Birecik reservoir and unthreatened parts of the
. .
Geophysics
Jamon Van Den Hoek and William Aylward
. .
Magnetometry
Ground-penetrating radar
Electrical resistivity
A
F
I
H
G
E
D
B
C
Trench 7
Trench 18
Trench 12
Trench 15
Trench 11 Trench 5 Trench 2
Trench 9
Trench 3
Trench boundary
meters
0200
380
390
370
360
350
400
410
420
400
380
390
370
Figure . Orientation plan, showing areas of geophysical survey and method employed.
.
050
meters
Scan depth = 0.01–0.70 meters
Scan depth = 0.70–1.55 meters
Area of low activity
Area of moderate activity
Area of high activity
Figures a (top) and b. Survey A. Ground-penetrating radar (GPR. Grid = -m intervals.
.
050
meters
Scan depth = 1.55–2.25 meters
Area of low activity
Area of moderate activity
Area of high activity
Figure c (top). Survey A. GPR. Grid = -m intervals.
Figure d (bottom). Features excavated in Trench within area of Survey A.
.
city available for further investigation. e authors hope
that the following discussion of methodology, data, and in-
terpretation will be useful as a case study for geophysical
survey both in the Euphrates Valley and in the context of
rescue archaeology, and that it will complement results of
geophysical survey at the now ooded site of Apamea, once
on the east bank of the Euphrates, and at sites proposed for
military installations on the west bank near Zeugma.2
METHODS
Overview
Methods used for geophysical survey during the archaeo-
logical rescue work at Zeugma in were magnetom-
etry, electrical resistivity, and ground-penetrating radar
(GPR). About , sq. m were surveyed, with two sur-
veys, I and E, covering some of the same ground with dif-
ferent methods.3 In general, the tripartite methodological
scheme for geophysics at Zeugma produced reliable
data that contribute to meaningful reconstruction of the
city’s urban topography. e three ground-based methods
for subsurface prospection applied at Zeugma are distin-
guished by the way geophysical signatures are measured.
Results for each method varied against the archaeological
prole for the site, which is covered by mature pistachio
orchards planted on deep colluvial deposits with abundant
broken ceramic and tile. ese deposits are on average
about m deep, but in some cases up to m deep. Excava-
tions have shown that walls of buried houses are normal-
ly preserved to about m. Streets are paved in stone, and
rooms are oen paved in mosaic. Fired tiles from collapsed
roofs appear in high frequency.
Magnetometry
For magnetometry at Zeugma in , the surveyors used
a Geoscan FM uxgate magnetometer with two inde-
pendent uxgates spaced mm apart on opposite ends
of a vertical pole. Fluxgate magnetometers are composed
of a permeable nickel-iron alloy core that is magnetized by
a primary winding and the earth’s magnetic eld. Fluctua-
meters
010
Area of low activity
Area of moderate activity
Area of high activity
Figure . Survey B. GPR. Grid = -m intervals. Scan depth = .–. m.
.
tions in the earth’s magnetic eld are produced by objects
above and below ground with magnetic properties, and the
magnitude of these disruptions is measured in nanoTesla
(nT) or gamma. Both subsurface anomalies and “noise”
(random responses) caused by objects with magnetic prop-
erties in or around the survey can be detected by magne-
to m et r y.4
A magnetometer typically detects features within m
of the survey surface, and this means the method was not
ideal for the occasional deep (up to three-meter) colluvial
deposit at Zeugma. e terrain at Zeugma is also uneven,
sloped, and dotted with pistachio trees. In some cases this
dicult terrain inhibited data collection. Adequate detec-
tion of the city’s built environment depended on contrast
between magnetic responses from archaeological features
and surrounding colluvium. But in each of the four areas
tested with magnetometry, the surveyors encountered
broad zones of random magnetic response very close to
the survey surface, presumably caused by colluvium replete
with broken ceramic and tile. Noise was ltered in the data-
processing stage, but only to the detriment of meaningful
survey data.5 As a result, whereas in several cases streets
and buildings were sometimes perceptible in the processed
survey data, excavation at the site has shown that the fre-
quency of such features is normally much higher.6
Electrical Resistivity
Electrical resistivity depends on an object’s tendency to
conduct electricity.7 For electrical resistivity at Zeugma in
, surveyors used a Geoscan RM resistance meter
with a Twin-Probe arrangement that introduced cur-
rent into the ground via two electrodes, one current and
one potential. e potential dierence caused by the cur-
rent was measured by two potential inner electrodes. An
increased distance between the two potential inner elec-
trodes allowed for investigation at greater depths.
e extremely hot and arid conditions of the summer
of at Zeugma were a signicant drawback for the
electrical resistivity survey. Archaeological features are not
markedly distinguished from surrounding ll by electrical
resistivity in hot and dry conditions. Parched soil lacks
interstitial water between soil particles. Without the con-
ductive properties of water, a soil’s low electrolytic conduc-
tivity produces abnormally high resistance, and this masks
weaker signatures from archaeological features. On occa-
sion, topsoil soened by mechanized plowing impeded
robust electrical contact with the subsurface, and this pro-
duced spurious readings.8
Since electrical resistivity measurements of this type are
not aected by above-ground objects of any conductance,
0
meters
20
Area of low activity
Area of moderate activity
Area of high activity
Figure . Survey C. GPR. Grid = -m intervals. Scan depth = .–. m.
.
electrical resistivity systems are oen used instead of, or as
a complement to, magnetic methods. Such was the case in
Survey I, where surveyors used both electrical resistivity
and magnetometry.9
Ground-Penetrating Radar (GPR)
Surveyors at Zeugma applied ground-penetrating radar at
four locations on the site.10 Surveys consisted of -m-wide
traverses within an orthogonal grid at a rate of scans
per meter on a SIR system manufactured by Geo-
physical Survey Systems Inc.11 Data was collected with a
mid-range frequency (MHz) antenna, on or near the
ground, which emitted electromagnetic (radar) pulses into
the ground. Portions of the radar waves emitted by the
antenna were reected by subsurface objects at locations
of electric or magnetic discontinuities and were detected
by the receiving antenna on the surface, where the sig-
nal was amplied. Electric and magnetic discontinuities
produce abrupt changes in the pulse’s velocity, normally
due to changes in the soil type, interstitial water content,
underground cavities, and archaeological features. us,
reected signals are generally stronger when an object’s
properties are in contrast with its surroundings. Subsur-
face features were mapped in two dimensions based on the
amplitude and reection patterns of the waves.
GPR has several advantages over electrical resistivity
methods and magnetometry. GPR data is relatively easy
to interpret, and GPR surveys cover more ground in less
time. Whereas electrical resistivity measurements must be
taken at small intervals in order to achieve high-resolution
data, GPR hardware is most oen hand-towed or pulled by
a vehicle over the survey area, with data collected at a high-
er rate. Unlike electrical resistivity methods, GPR is most
eective in dry, nonconductive soil types, like those found
at Zeugma, because saturated media impede a radar wave’s
ability to pass through a given medium. In addition, GPR
scans can be targeted for specic depths by changes to the
pulse frequency, with the scan resolution generally dimin-
ishing in quality for deeper scans. GPR thus makes pos-
sible three-dimensional graphic representations of a survey
area, with readings shown relative to depth underground.
Deeper scans yield data at lower resolution, but the over-
all eect of detecting superimposed habitation levels and
building phases is an especially important advantage for
archaeological prospection.
020
meters
Area of low activity
Area of moderate activity
Area of high activity
Figure . Survey D. Grid = -m intervals. Scan depth = .–. m.
.
Data Display
Illustrations selected for this chapter include grayscale
display for magnetronomy and electrical resistivity data,
and color timeslice plots for GPR data. For grayscale dis-
play, the full range of values is subdivided into intervals,
and each interval is assigned a shade of gray between black
and white. For electrical resistivity, darker shades represent
stronger magnetic responses and lighter shades represent
weaker ones. For magnetometry, stronger shades of black
and white indicate positive and negative magnetic uctua-
tions, respectively. For the GPR data, timeslice plots depict
data retrieved from dierent depths in the same survey
area. e delay recorded between the time a pulse is sent
and received is called the timeslice window. Weak reec-
tions in a timeslice window are shown in dark blue or
green, whereas stronger reections appear in brighter col-
ors, such as light green, yellow, orange, red, and white (in
order of intensity, with red and white being most intense).
e surveyors produced four timeslice plots per area, and
we have selected the plots that best inform on the archaeol-
ogy of Zeugma for presentation in this chapter.
5.0
-5.0
50
0
meters
nT
Figure . Survey E. Magnetometry.
SURVEY RESULTS AND
INTERPRETATION
GPR Survey
Survey A
Survey A was the only survey area tested with excavation
(Trench ). Dierent anomalies were detected in scans
conducted at dierent depths (gs. a–c). Most meaning-
ful were a shallow scan set to .–. m, a mid-range
scan set to .–. m, and a deep scan set to .–. m.
Indications of a large building across scan depths led to the
decision to excavate.
e shallow scan detected a group of anomalies at the
center of the survey area, linear trends to the east and west
of this, and moderate activity at the southwestern corner
of the survey area. With the benet of excavation, these
anomalies are now understood as a large concentration of
ancient robbing activity that brought detritus near the sur-
face above the southeast corner of a large masonry build-
ing, ceramic and mortar hydraulic installations to the east
and west of this (e.g., contexts and ), and Cistern
with associated waterworks to the southwest. e
mid-range scan picked up some of the anomalies detected
.
5.0
-5.0
nT
0
meters
30
Figure . Survey F. Magnetometry.
on the shallow scan, but with weaker signals. is sug-
gests more substantial construction or debris, or both, at a
deeper level, but contiguous with overlying levels. e deep
scan had a weaker response across the entire survey area,
but a few small areas of strong response appear in the same
place on the mid-range scan. Some of these areas of high
response were deep masonry piers on wall found by
the excavators.
ere is an exceptional level of correlation between the
GPR results for Survey A and the excavation ndings in
Trench (g. d). Most of the excavated structures in
Trench can be located on at least one of the GPR timeslice
plots. A notable exception is Trench ’s wall , a large
feature oriented east-west in the southeast part of the
trench.12 A signicant correlation is the stone and tile pave-
ment at the center of the trench, which turned out to be a
much larger anomaly than suggested by the GPR data.
Survey B
Survey B was conducted to shed light on the unexcavated
gap between Trenches and .13 e scan displayed here
shows the response from a depth of .–. m (g. ).
Wa lls discovered in Trenches and were presumed to con-
tinue into the survey area, and the survey results conrmed
this (g. ). Anomalies consistent with the continuations
of walls found in Trench were detected at the western
edge of the survey area, and similar signatures detected at
the southeast corner of the survey area appeared to be con-
tinuations of walls in Trench . Data from Survey B suggest
the presence of an additional room on the north side of the
building excavated in Trench . As in the case of Survey A,
the complex response in this area may be caused by super-
imposed layers of building material or collapsed debris. A
mosaic pavement should not be ruled out, because rooms
discovered in Trench were paved in this technique. Sig-
natures for walls on the east side of the survey area appear
to have an orientation consistent with walls discovered in
Trenches and . In this case, the survey results are espe-
cially useful for connecting archaeological features in these
adjacent trenches.
Survey C
Survey C was designed to complement excavation results
in nearby Trench .14 e scan displayed here shows the
response from a depth of .–. m (g. ). e signatures
for walls in the northeast part of the survey suggest that
the house discovered at the northwest part of Trench (the
House of the Pelta Mosaic) continued into Survey Area C
(g. ). A rather large anomaly, at least three meters across,
along the southwest part of the survey is similar in com-
plexity to the response detected on the west side of Survey
B. e houses in Trench were covered with thick depos-
its of destruction debris, including burned mud-brick and
roof tiles, and an accumulation of this material inside a
room would be consistent with the signature of this anom-
aly. Likewise, given the large quantity of mosaic discovered
.
0 50
meters
5.0
-5.0
nT
Figure . Survey G. Magnetometry.
in Trench , it is conceivable that the anomaly in Survey
C could indicate the presence of a large mosaic pavement.
Survey D
Survey D was designed to complement excavation results
in Trenches and .15 Abundant rock on the survey sur-
face introduced noise into the survey results, and in some
cases this masked weaker responses from archaeological
features. Nonetheless, many of the same features show up
on multiple timeslice plots, and this suggests deep con-
structions or deposits. e scan displayed here shows the
response from a depth of .–. m (g. ). e major-
ity of the anomalies detected in the survey conform to one
of the rectilinear alignments evident in Trenches and ,
and these probably belong to either walls or drains (g. ).
At the southwest part of the scan, a large anomaly, about
. m wide, has the same signature as the large anomalies
detected in Survey B and C. A large mosaic pavement or
a room lled with burnt collapsed building debris would
not be inconsistent with the archaeological discoveries in
Trench . ere are no obvious correlations between the
data from Survey D and Trench .
Magnetometry Survey
Survey E
Survey E is the largest of the magnetometric survey areas
(g. ). A substantial linear trend across the entire survey
area is most probably a street paved in limestone.16 e sig-
nature may be enhanced by red materials, such as drain
pipes, under or alongside the street.17 e presumed street
is about m wide, and it can be traced from the south-
west corner of the survey area to its center, where the trend
turns slightly to the east. e change in orientation is sig-
nicant for reconstruction of the city plan. Buildings and
streets excavated to the north and east of this survey area
(e.g., in Trenches , , and ) have a dierent orientation
than buildings discovered to the west (e.g., in Trenches ,
, , and ).18 It is therefore conceivable that the bend in
the anomaly represents a juncture between districts of the
city with dierent building orientations (g. ).
In the very north part of the survey area, another lin-
ear anomaly is parallel to the northeast stretch of the street
through the middle of the survey area. is anomaly is
fainter, but the orientation and alignment are consistent
with the signature for a street.19 e parallel streets in the
.
200
40
ohms
50
meters
0
Figure . Survey I. Electrical resistivity.
northeast half of Survey E measure m apart on center.
ere are a number of very faint rectilinear anomalies be-
tween these presumed streets, but nothing strong enough
to suggest additional streets or buildings.
e paved street discovered in Trench is oriented
southeast to northwest — perpendicular to the anoma-
lies in Survey E. e monumental building in Trench
is parallel to the anomaly on the north side of Survey E.
Framed against these surrounding structures, the anoma-
lies in Survey E take on the distinct appearance of a large
city block, conceivably an open plaza or agora, framed on
at least three sides by streets and monumental buildings
(g. ). In rather stark contrast to most of the modern
surface topography along the banks of the Euphrates, this
part of the site is relatively at. e substantial terrace walls
discovered in Trenches and were probably installed as
part of the design for a large open plaza on the terrace to
the southwest.
e signatures for streets in the northeast part of Sur-
vey E are the rst substantial evidence for the size and
orientation of city blocks at Zeugma. is can be weighed
against results of geophysical survey on the opposite bank
of the Euphrates at Apamea, Zeugma’s counterpart in the
Hellenistic period. City blocks at Apamea have a uniform
orientation, and this is consistent with the city’s location
on a oodplain and its foundation as a Seleucid colony.
e evidence from Survey E, complemented by the results
of rescue excavations in , shows that city blocks at
Zeugma have at least two, and probably more, orientations,
each unique to a particular district of the city. Buildings to
the east (e.g., in Trenches and ) and to the west (e.g., in
Trenches , , and ) of the buildings and streets clustered
around Survey E have independent orientations. Variant
building orientations are consistent with Zeugma’s location
on the undulating topography of the Euphrates’ west bank
and the city’s growth in ts and starts from Hellenistic into
Roman imperial times. Of all the building orientations
known from Zeugma, the orientation clustered around
Survey E is the only one that matches Apamea’s. e uni-
form building orientation on both sides of the Euphrates
on this particular axis suggests that this area is the likeliest
spot for not only Zeugma’s earliest Hellenistic settlement
but also the bridgehead that linked the two cities.
Survey F
Several anomalies were detected in this survey area, most
with a near north-south or east-west orientation (g. ).
ese anomalies are consistent with the suspected signa-
ture for walls in the other areas of magnetometric survey
at Zeugma, and the features detected in Survey F probably
.
050
meters
Drain
Possible feature
Street
Trench 7
Trench 18
Trench 12
Access Road
SURVEY E
SURVEY A
Possible
feature
Trench 3
Figure . Survey E with extrapolation and Trenches , , and .
belong to structures on the terrace above the street dis-
covered in Trench (g. ).20 A sizable anomaly oriented
southeast to northwest is indicated by two signals on the
same alignment, and this has a similar appearance to the
streets discovered in Survey E. Factors inhibiting results in
this area were small survey size and a metal fence on the
northeast side of the survey area that introduced consider-
able magnetic interference.21
Survey G
Readings in this western and central part of this survey
area were largely clouded by earth-moving machinery
parked nearby (g. ). e magnetic responses were quite
weak, and it is dicult to connect survey results with any
excavated features in Trenches , , or (g. ).
Survey H
No anomalies of archaeological interest were able to be
detected, due to the high quantity of magnetic gravel on
the survey surface area.
Electrical Resistivity Survey
Survey I
Survey I’s electrical resistivity measurements allowed for
comparison of results with the magnetometry survey on
the same spot (Survey E). Despite the arid soil conditions,
adequate electrical contact was achieved and acceptable
data obtained (g. ). e application of two methods to
the same area provides supplemental information for inter-
preting data. e electrical resistivity results support the
magnetometry-based interpretation of the anomalies in
the survey area as a major road, approximately m wide,
oriented southwest to northeast, with a change in orienta-
tion that may signal a boundary between zones of inhabita-
tion in the ancient city (g. ).
.
050
meters
Street
Mosaic?
Trench 11
Trench 5
SURVEY G
SURVEY F
SURVEY B
Trench 14
Figure . Surveys B, F, and G with extrapolation and Trenches , , and .
Schematic plan of Trench aer Abadie-Reynal et al. , g. ..
.
050
meters
Mosaic?
Mosaic?
Trench 2
Trench 9
SURVEY C
SURVEY D
Figure . Surveys C and D with extrapolation and Trenches and .
.
050
meters
Trench 12
SURVEY I
Street
Figure . Survey I with extrapolation and Trench .
.
NOTES
. e unpublished geophysical reports do not present results in
the context of the excavation ndings. Results of these geophysi-
cal surveys were not utilized in published interim reports on the
archaeological work of .
. For geophysical survey at Apamea, see Abadie-Reynal et al. ,
–; , –; , –; , –. For geophysi-
cal survey at Zeugma’s military installations, see Hartmann and
Speidel, volume , and , –.
. Stratascan Geophysical and Specialist Survey Services con-
ducted GPR surveys in four areas (here renamed A, B, C, and
D) between July and August , (Barker and Mercer
). Between September and , , GSB Prospection
(S. Ovenden-Wilson, C. Stephens, and A. Shields) conducted
magnetometry surveys in four areas (here renamed E, F, G, and
H) and an electrical resistivity survey (here renamed Survey I).
. For uxgate magnetometry in archaeology, see Ganey and
Gater , –; Kvamme , ; Sharma , –, -.
. For the signicance of data ltering, see Ganey and Gater ,
–; Kvamme , .
. Noise was also encountered from ferromagnetic gravel on the
surface of Survey H, earth-moving machinery parked west of
Survey G, vehicles parked south of Survey E, and a wire fence to
the east of Survey F.
. For resistivity surveys in archaeology, see Ganey and Gater
, –; Kvamme , –; Sharma , –.
. Open or recently backlled excavation trenches near the survey
areas may have also interfered with the data, although to what
degree is uncertain.
. Most objects that can be detected through magnetism can also
be detected through electrical resistivity surveys because of the
relationship between electricity and magnetism as described by
Maxwell’s equations: cf. Scollar , .
. For GPR in archaeology, see Ganey and Gater , –;
Kvamme , –; Sharma , –.
. GPR plots were produced with Radan soware. A ltering algo-
rithm was applied to reduce noise and improve the clarity.
. See the discussion of Trench by Aylward, this volume.
. Excavations in Trenches and were ongoing at the time of the
su r v e y.
. Excavations in Trench were ongoing at the time of the survey.
Part of the survey was obstructed by a rst-aid station set up by
the excavators.
. Excavations in Trench had been completed at the time of the
su r v e y.
. GSB Prospection , : “Such a response is normally charac-
teristic of a ditch but such an interpretation is unlikely given the
wider archaeological context. e alignment of the anomaly and
its width suggests that it may represent a road. Such an inter-
pretation appears to t with the topography and known layout
of the site. However, it is not clear as to why a road should give
such a response. Assuming it is paved limestone it should be a
negative anomaly or a quiet band of data. e latter, however, is
unlikely to be visible due to the surface noise. It is possible that
the anomaly indicates some change along the edge of the road or
that the road may be at least partially constructed of, or associ-
ated with, red material such as crushed brick or tile.”
. E.g., Kvamme , ; Sharma , –.
. With the exception of Trench , features in these trenches are
discussed in the chapter by Tobin, this volume. For the paved
street in Trench , see Early , –.
. GSB Prospection , .
. See Abadie-Reynal , , –.
. GSB Prospection , : “e limited size of the survey area
makes it dicult to identify patterns and formulate any precise
interpretation.”
BIBLIOGRAPHY
Abadie-Reynal, C., et al. . “Mission archéologique de Zeugma:
Rapport sur la campagne de prospection .” Anatolia Antiqua
:–.
———. . “Mission de Zeugma-moyenne vallée de l’Euphrate.”
Anatolia Antiqua :–.
———. . “Zeugma-moyenne vallée de l’Euphrate: Rapport pré-
liminaire de la campagne de fouilles de .” Anatolia Antiqua
:–.
———. . “Zeugma-moyenne vallée de l’Euphrate: Rapport pré-
liminaire de la campagne de fouilles de .” Anatolia Antiqua
:–.
———. . “Zeugma: Rapport préliminaire des campagnes de
fouilles de .” Anatolia Antiqua :–.
Barker, P.P., and E.J.F. Mercer. . “A Report for Oxford Archaeo-
logical Unit on a Geophysical Survey Carried Out at Zeugma,
Turkey, July/August ,” Job Ref. No. . Unpublished.
Early, R., et al. . “Rescue Work by the Packard Humanities Insti-
tute.” In Zeugma: Interim Reports, edited by J.H. Humphrey, –
. JRA Suppl. . Portsmouth: Journal of Roman Archaeology.
Ganey, C., and J. Gater. . Revealing the Buried Past: Geophysics
for Archaeologists. Stroud: Tempus.
GSB Prospection. . “Geophysical Survey Report /: Zeug-
ma, South East Turkey.” Bradford: GSB Prospection. Unpub-
lished.
Hartmann, M., M.A. Speidel, and M. Drahor. . “e Roman
Army at Zeugma: Recent Research Results.” In Zeugma: Interim
Reports, edited by J.H. Humphrey, –. JRA Suppl. . Ports-
mouth: Journal of Roman Archaeology.
Kvamme, K.L. . “Geophysical Surveys as Landscape Archaeol-
ogy.” AmerAnt :–.
Scollar, Irwin. . Archaeological Prospecting and Remote Sensing.
Cambridge: Cambridge University Press.
Sharma, P.V. . Environmental and Engineering Geophysics. Cam-
bridge: Cambridge University Press.
