ICAA

Field projects

Geophysical Survey at the Ancient City of Apollonia 2007 [1]

The geophysical survey was funded by the Packard Humanities Institute. The project was launched by the Albanian Rescue Archaeology Unit and it is a collaboration among ARAU (represented by Maria Grazia Amore), the Institute of Archaeology (represented by Vangjel Dimo), the National Archaeological Park of Apollonia (represented by Arjan Dimo), and the School of Environmental Sciences-University of East Anglia (represented by David Bescoby, who provided the technical expertise and the equipment). [2]
The magnetometer survey was undertaken in the low lying areas to the south-east of the walled city of Apollonia in June 2007, covering an area of 5.2 ha. The survey revealed dense, although badly disturbed areas of former urban occupation, including surviving building foundations and possible roadways. Several of the revealed building plans follow a north-west/south-east orientation, following the principal axes of the walled city, and suggesting the formal planning and layout of these lower urban areas. Lack of access to co-joining fields severely limited the coverage of the survey, making an archaeological interpretation difficult, although it appears the survey located the south-westerly limits of urban settlement, 200 m from the base of the Hill of Shtyllas. No definitive evidence for the existence of a sea port was uncovered, despite the proximity of the survey to former meanders of the Vjosa river. The survey however highlighted the extent of the settlement along the western flanks below the walled city, estimated at over 80 ha.
The aim of the current project was to investigate the application of geophysical survey (magnetometry) within these low lying areas, assessing the potential of the technique as a means of rapidly mapping surviving archaeological remains over large areas. The successful use of geophysical survey techniques at Apollonia is potentially of great importance to the ongoing management of the area, designated a national archaeological park, in light of a proposed road scheme that threatens to disrupt the immediate surroundings. Of particular importance is locating the former city port(s), alluded to in historical sources[3], but now indiscernible beneath the heavily farmed alluvial plain.

The wider landscape
The former city of Apollonia, is located upon a low hill and surrounding slopes that form the north western end of the Frakulla hills, a low elongated ridge aligned NNW-SSE marking an anticline. To the west fluvial sediments delivered by the Seman and Vjosa rivers has led to the formation of the Myseqè coastal plain, which now separates Apollonia from the Adriatic by over 10 km. The dynamic coastline is characterized by bars formed by longshore drift away from the deltaic river mouths, behind which lagoonal embayments have formed – eventually evolving into marshes. The plain continues to undergo subsidence as a result of neotectonic activity (a N-S fault line is present bordering the Adriatic coast), which accounts for the persistence of wetland areas throughout the Holocene (Fouache et al., 2001).
During the past 6 millennia, changes in the courses of the two rivers, particularly the Seman, have been numerous (until the agricultural drainage programme of the 1950s). The Vjosa river, running along the western margins of the Frakulla Hills has its present outlet several kilometers south of Apollonia, although remotely sensed data (see below) reveals traces of meandering palaeochannels c. 600 metres from Apollonia, immediately west of the Hill of Shtyllas. It his postulated that in antiquity, a large lagoon occupied the central part of the deltaic plain immediately west of Apollonia; the ancient Aous river flowing into this lagoon from the southeast. A former marsh area called the Gjoli lagoon, appears on early maps at this approximate location.

Geophysical survey
Within urban contexts throughout the Mediterranean, geophysical techniques have become important in mapping buried archaeological remains over spatially large areas. The use of magnetometry, in particular, has proved effective for the fast, high-resolution mapping of urban deposits. The technique works by detecting of very small fluctuations or distortions in the earth’s magnetic field caused by concentrations of buried magnetic material, such as archaeological remains. Here objects such as brick and tile (which contain enhanced magnetic properties as a result of their firing) effectively cause an ‘anomaly’ in the earth’s geomagnetic field. The position of these magnetic anomalies can be plotted to form a magnetic map, representative of buried archaeological deposits.

Methodology
The survey was undertaken using a Scintrex caesium vapour magnetometer, offering measurement sensitivity of 0.01 nT (the strength of the earth’s magnetic field at the survey location is 47,800 nT). For added stability and ease of use, the magnetometer was mounted upon a wheeled cart, allowing the magnetometer to be pushed along survey transects during survey. The accurate positioning of the magnetometer was achieved using a differential GPS unit fitted to the cart, allowing the location of the cart to be logged with magnetometer readings.
Measurements were taken along parallel transects spaced one metre apart, and orientated N-S, or in some instances, E-W, depending upon field orientation and positions of field irrigation channels. The magentometer was configured to take readings every 0.2 seconds, which for an average walking speed of 1 metre/second typically resulted in a reading every 20 centimeters. The orientation of the magnetometer and survey cart was kept constant throughout the survey to minimise heading errors – handles at either end of the cart allowing it to be pushed from both ends (fig. 1).

Figure 1: Magnetometer survey at Apollonia.

The location of the fields surveyed is shown in fig.2. The underlying methodology was to commense the survey close to the foot of the Hill of Shtyllas; an area of known surface scatters of building materials and wall foundations revealed in drainage cuts. It is also an area close to a former meander of the Vjosa river, which in antiquity flowed northward of its present location. Adjoining fields to the north-west were then surveyed, taking in areas containing known surface scatters and surface finds.

Figure 2: Location of surveyed fields.

Results
Results in the form of grayscale magnetic maps of each field surveyed are shown in figs. 3 to 10. Here the dynamic range is -8 to +8 nT. The maps have been annotated to highlight anomalous magnetic features of archaeological interest. Here surviving building foundations are shown in red, whilst other features of interest mentioned in the text below are shown in yellow.
Description of results
Field 1: A narrow strip c. 70 x 16 meters, bordered by a large drainage channel to the east and a flush of thick wetland vegetation (fig. 3).

Figure 3: Magnetic map of Field 1.

The survey revealed the corner of a substantial building in the southern portion of the field, aligned to c. 45 degrees to the Southeast. To the north, the survey detected a series of more ephemeral magnetic anomalies which also seem to conform to this alignment. The small size of the survey area makes interpretation difficult, although these features may represent the remains of a roadway.
Field 2: Located west of Field 1 and covering 0.6 ha, this area revealed a complex pattern of magnetic anomalies, indicative of intense settlement, potentially spanning several phases (fig. 4).

Figure 4: Magnetic map of Field 2.

One clear set of wall foundations, seen as dark (negative) linear magnetic responses are recorded in the northern portion of the field, forming a large, square building. The lack of overall definition may well relate to later robbing of stone blocks. A slightly weaker set of similar anomalies can also be see in the southeast of the field, at it is tempting to suggest these are a continuation of the building recorded in Field 1; both building foundations in Field 2 follow this same alignment. The roadway hypothesized above would also fit in with the layout recorded. A further series of strong linear anomalies were recorded within this field and it is not clear whether these are of archaeological significance or relate to later land use, such as field drainage features predating the current layout.
Field 3a: The most easterly area to be surveyed in this block of fields (fig. 5) revealed another complex set of anomalies indicative of a continuation of the area of settlement recorded to the east. Again, clear anomalous features are hard to discern, reflecting the truncated nature of surviving remains and the undoubted quarrying/removal of stone from the fields. It is just possible to detect the faint outlines of a building foundations or wall lines which seem to follow the same basic alignment detected to the east. What is most interesting is the apparent edge of the settled area running approximately N-S through the field. It is not clear whether this represents the position of a former channel edge or some other form of demarcation, possibly related to the group of broad linear anomalies on a N-S/E-W alignment immediately to the west. It is not entirely clear what these represent, although they may relate to relatively recent agricultural engineering works.

Figure 5: Magnetic map of Field 3a.

Field 3b: This field, located to the north west of Field 3a (fig. 6), revealed very little of archaeological significance. Several faint linear anomalies can be seen which probably relate to agricultural activity. The data is also of poor quality, showing considerable striping and changing background levels. This is due to the overall unevenness of the terrain in this field and difficulties experienced in keeping the cart stable during traverses.

Figure 6: Magnetic map of Field 3b.

Field 4a: The small area surveyed 40 metres to the west of Field 3 (fig. 7) revealed very few magnetic anomalies, comprising a few discrete spikes caused by brick and tile fragments in the plough soil. The traces of the irrigation channels that cut that cross the middle of the area in an east-west direction were also detected.

Figure 7: Magnetic map of Field 4a.

Field 4b: This area revealed a large spread of magnetic anomalies indicative of settlement and the remains of stone build structures (fig. 8). The size of the anomalies are not as great at that recorded in the fields 400 metres to the south east and may represent a different type/phase of settlement, or indeed variations in burial depth. It is also difficult to detect any alignment trends amongst the short linear anomalies. Interpretation is compounded by the gaps in the survey caused by the regular irrigation channel cuts running east-west (seen as blank areas within the survey). One feature of particular interest is the discontinuous line of strong positive anomalies running approximately N-S, possibly representing the remains of a large boundary. To the west there is also a much more ephemeral linear anomaly running for nearly 100 m which is orientated to the same 45 degree NW-SE alignment recorded in settlement to the south east. This may indicate the position of a former road or trackway.

Figure 8: Magnetic map of Field 4b.

Field 5: The areas surveyed within Field 5 are shown in fig. 9. Here the coverage is more continuous, giving a potentially more coherent picture of surviving archaeological remains. Again, the undulating, corrugated nature of the terrain between regular irrigation channels running east/west has led to a degree of striping, particularly in the southern sections, tending to mask the most subtle anomalies. In general, it can be seen that the northern portion of the surveyed area contains a spread of small anomalous responses indicative of low level settlement activity and the presence of cultural material, such as brick and tile fragments within the ploughzone. One of the most interesting features is the faint traces of a road or trackway running approximately north-south, extending c. 230 metres. This is faintest in the south, becoming more discernable towards the top of the field, where two parallel positive anomalies, possibly forming ditches, can be seen. There are several building remains revealed within Field 5; the first is a small, faintly recorded set of anomalies along the west of the roadline in the central portion of the surveyed area. This is suggestive of a structure of containing little stone/brick and tile in its construction and possibly represents the remains of one of a series of wooden/reed dwelling said to have existed in this area prior to the agricultural engineering works of the communist era. Here the recorded anomalies represent a habitation area of enhanced magnetic material and possible post pits/slots relating to the structure. It is possible that the trackway recorded is also of this date. In the northern section of the surveyed area, two, possibly related building foundations of much more robust construction were recorded, seen as a complex of large magnetic signatures. The inferred solid nature of the construction is reminiscent of Roman tombs or mosollia recorded along settlement margins at Butrint. To the east of these buildings the survey recorded a number of dark, curvi-linear anomalies, which may relate to further, more ephemeral structures or possibly walled or fenced animal enclosures.
Field 6: This small field, 450 metres north of Field 2 contained a substantial surface scatter of material and was surveyed to confirm the presence of a possible temple structure (fig. 10). The survey revealed the presence of building foundations, although the area surveyed was too small to discern any discrete structures. The surrounding fields were heavily cropped at this time.

Figure 9: Magnetic map of Field 5.

Figure 10: Magnetic map of Field 6.

Discussion
The magnetometry survey undertaken has proved successful in mapping surviving archaeological deposits upon the margins of the alluvial plain flanking the walled city. The complex nature of the recorded anomalies over areas of dense settlement suggests that underlying deposits have been disturbed through the robbing of building material and later ploughing. It is also likely that these areas represent several phases of occupation and reuse. Within Fields 1 and 2, building foundations appear to follow a coherent alignment of 45 degrees north east, which appear to agree with the principal orientation of buildings within the walled city. However, a more detailed analysis of the urban topography is required to properly investigate the nature of planned elements of the settlement, their phasing and alignments to existing road systems etc. Judging from the survey results (Field 6) and the location of surface scatters of material further north, it is estimated that this belt of urban settlement is spread over an area c. 80 ha in size, which is not unreasonable for a city of the size described in antiquity, with reputedly over 50,000 citizens at its height.
The location of the city port or ports at Apollonia has never been satisfactorily resolved. The favored location has been close to the Hill of Shtyllas, where a large looping meander of a former course of the Vjosa river can be clearly identified from remotely sensed satellite imagery; the difference in sedimentary infill of the palaeo-channel resulting in detectable differences in surface reflectance as well as vegetation cover. The southern most portion of the survey (Fields 1 and 2) indicated dense occupation north of the course of the ancient river, although it is not clear whether any of the features detected relate to structures associated with quayside areas. The overall lack of access to fields within this critical area, along with the presence of dissecting drainage channels prevented any clearer assessment. It is interesting to note that it is at this location that a smaller (now canalized) river or stream flowing down from the higher ground to the east is likely to have converged with the Vjosa. This confluence may well have afforded further opportunity of moorings. Substantial canalization and other hydrological engineering are also likely to have been undertaken in antiquity to create or enhance locations suitable for docking ships.
Perhaps the most significant factor dictating the position of the port is the location in antiquity of a large lagoon to the immediate west, providing access to the Adriatic through a gap in the gravel spit formations. Assuming this low energy body of water, fed from the south by the ancient Vjosa remained stable for some periods in antiquity, its location would potentially provide a large natural harbour area. The utilisation of such natural features is well attested in the ancient world. It might well be that it was along the eastern margins of this former lagoon that a network of jetties were constructed, forming a port of the size reported in antiquity. Fig. 11 shows palaeofeatures, such as the former course of the ancient Vjosa river visible upon a specially enhanced Landat 5TM composite image, along with the postulated shoreline of the former lagoon.

Figure 11: Magnetic map of Field 6.

Conclusions
The large areas of buried remains interpreted as former settlement detected on the alluvial margins southwest of Apollonia, suggests that the lower slopes to the west of the walled city are likely to conceal further remains of the former settlement. The detection of certain coherent building alignments within the survey, which appear to relate to former road systems is suggestive of the organised, planned urban settlement, although the exact date is unclear.
Although the survey did not conclusively reveal the location of the ancient port(s) at Apollonia, the results go some way to delimit the southern edge of settlement, which must have come close to a former course of the Vjosa river. A more complete survey of this area is required, including a geoarchaeological investigation of the former river course and lagoonal margins.
It can also be concluded that the use of magnetometry is an effective geophysical technique for the mapping of buried remains within the depositional environment encountered at Apollonia.

References
Fouache, E. G., Gruda, S. Mucaj & P. Nikolli (2001). Recent geomorphological evolution of the deltas of the rivers Seman and Vjosa, Albania. Earth Surface Processes and Landforms 26, 793-802.
Mathers, S., D. S. Brew & R. S. Arthurton (1999). Rapid Holocene evolution and neotectonics of the Albanian Adriatic coastline. Journal of Coastal Research 15, 345-354.
___________________________________________________________

[1]This report was written by David Bescoby, and edited and completed with the ancient sources references by Maria Grazia Amore.

[2]Many thanks to Professor Richard Hodges, Professor Muzafer Korkuti, and Marin Haxhimihali for their support. Special thanks also go to Surja Lela (IoA) and to John Papadopoulos and the UCLA team for their hospitality.

[3]From ancient sources it is known that Apollonia was located on favorable land between two rivers (Dio Cassius, 41.45; Strabo, 7.5.8: Strabo specified that Apollonia was situated on the Aous river), the Seman (ancient Apsos) and the Vjosë (ancient Aous). From geological studies (the most recent one undertaken by Eric Fouache) has emerged the conclusion that the beds of both rivers were closer to Apollonia than today, as was the coastline. The bed of the Aous river was also wider and deeper, and was navigable by ships. Indirect information about this can be found again in sources of the early Roman period, that tell about battles that took place at Apollonia between Rome and the Illyrian queen Teuta, or between Caesar and Pompey (Polybius, 2.11.6; Caesar, B. Civ. 3.13; Plutarch, Vit. Caes. 38.2). The port probably was not far from the hill of Shtyllas: a popular legend even says that the solitary column of the temple was used to moor the ships. Fouache has suggested that the port was on the Aous river, which probably issued into a lagoon.

Content on this page requires a newer version of Adobe Flash Player.

ICAA

Latest News

Field projects