1Geoarchaeology is a multi-proxy approach where geographical and geoscientific concepts and methods are applied to Prehistory, Archaeology and History (Rapp and Hill, 1998). Geoarchaeology consists in using methods and concepts of the Earth Sciences for archaeological research purposes. However, to elucidate environmental contextual issues, geoarchaeologists must be more than casual practitioners of applied science (Butzer, 1982; Fouache, 2006; Fouache and Rasse, 2007). Indeed, if archaeological excavation emerged in the 18th Century with a systematic analysis of the material excavated—notably in Herculaneum (Italy)—,stratigraphic excavation that applied environmental evolution data for the first time ever did not become established until the end of the 19th Century. Finally, to better understand environmental changes, particularly throughout the historical period, geomorphological research became an essential preliminary to the study of all archaeological sites in the 1980s.
- 1 Abbreviations and acronyms used in the article are listed in Annex 1.
2The Geographic Information System (GIS)1 is a digital support capable of integrating, storing, editing, analyzing, sharing, and displaying geographically referenced information (Marble et al., 1984). GIS is well adapted to share all the information provided by different disciplines from Human and Social Sciences and from Earth Sciences. In an extended sense, GIS is a tool that allows users to create interactive queries, analyze the spatial information, edit data, create maps and present the results of all these operations for archaeological and geoarchaeological studies (Kvamme, 1999; Fletcher, 2008). This development took place in the 1970s when several methods became available: computer cartography and Computer-Aided Drafting, the linking of computer-drawn maps with relational databases, quantitative spatial analyses and their mapped by-products, views and uses of three-dimensional terrain models (Digital Elevation Models), remote sensing and image processing applications in regional simulation and modeling exercises (Kvamme, 1999). Nowadays—far from being limited to produce aesthetically pleasing cartographic material— GIS plays an important key role in archaeology and enables dynamic viewing of morphological activity. This paper presents the methods and the results derived from several case studies from Albania (Korça Basin) and Greece (ancient Methoni harbour and Thessaloniki Plain) during the Holocene—the last 10000 years (Ghilardi, 2006; 2007).
3Until the 1990s, archaeological studies were essentially based on two-dimensional (2D) cartographic representation developed on a local (in situ) scale (from 0.1 to 10 km2). Computer cartography and computer-aided drafting helped to make within-site geoarchaeological studies, a rather limited technique compared to GIS. For example, vector outlines showing the locations of walls, pits, middens, ditches, post holes, etc., are generally colour coded by feature type, cultural affiliation or temporal period: various artefact distributions were similarly portrayed (Kvamme, 1999). Using C.A.D., ground observations, chart interpretations (topography, geology, etc.) aerial photographs and satellite image treatments can all be combined into environmental maps (geomorphological and vegetation maps, pedological charts, etc.). Until recently, different layers corresponding to points, lines, and polygons were created using Adobe Illustrator© software. This method lacked the possibility to associate graphic elements with geographic coordinates and to access dynamic geodatabases. These limitations are now addressed using GIS.
4The use of the GIS in archaeology is essential:
At the site level (from 0.1 to 10 km2), extensive data about excavation and surface mappings of artifacts, topography and other features are collected. It is necessary to efficiently manage these data and address fundamental research and spatial analysis questions (Kvamme, 1999). Three-dimensional GIS allows deposits, features, and artifacts to be visualized in their proper 3D contexts (Katsianis et al., 2008). A volume may be rotated, sliced, diced, or "exploded" to yield virtually any possible view of internal relationships. These systems allow better understanding of complex deposits and greatly help in the interpretation of intrasite spatial relationships, site structure, and formation processes (Kvamme, 1999).
At the regional scale (areas of more than 10 km2), GIS is frequently used to analyse the spatial distribution of settlements using statistical methods (Kvamme, 1999; Anschuetz et al., 2001). Archaeological predictive modelling—one of the earliest applications made possible by GIS—continues to grow in importance as a tool for cultural resource management and planning (Kvamme, 1999; Fry et al., 2004). GIS can support other information derived from:
3D modelling of present and past environments (relief, hydrology, shorelines, vegetation cover, etc.) and of their evolution.
the cross comparison of environmental, palaeoenvironmental and archaeological data. For example, GIS can be used to quantify changes in water volume of ancient reservoirs caused by the rise or fall of the water level (Desruelles and Cosandey, 2005).
5To create the GIS, various data sources are used, integrated with the main steps presented below.
- 2 Hatt, Transverse Mercator 3 degrees, Hellenic Geodetic Reference System (H.G.R.S.) 87 and U.T.M. E (...)
6The georeferencing phase of a cartographical study can be difficult in countries that do not use a single cartographic projection system to serve as a unique referential. In Greece for example, four systems are in use since the beginning of the 20th Century2 that can not be converted into each other. Polynomial equation (Ghilardi, 2006) and/or freeware (software) can help significantly to convert geographic coordinates. It is now crucial to use a single international reference for GIS such as the World Geodetic System (W.G.S.) 84 cartographic projection.
7The common definition of a D.E.M. can be presented as follows: a Digital Elevation Model is the digital image of altitudes for a topographical surface set in a geographical marker and a 3D representation of the territory without vegetation or buildings (Hubert, 2001; Ghilardi, 2006). Two methods of D.E.M are in usage depending on the community: the first employs the digitalisation of points on contour lines in order to create a Triangular Irregular Network (T.I.N.) type D.E.M.: points make up the mesh of the digital elevation modelling in which all the points are linked together by lines forming flat triangles that never intersect. These triangles are contiguous by their sides and form a continuous surface in space (Hubert, 2001). Raster D.E.M. has a lower quality of representation but file created by the GIS—which uses mass points and provides a smooth view in 2D—is smaller. The topographic data for the derivation of the D.E.M. can be obtained from several sources: contour lines (reported on maps), S.O.N.A.R. records, S.R.T.M. (Shuttle Radar Topography Mission) data and D.G.P.S. (Differential Global Positioning System) surveys:
8The georeferenced topographic maps have often the major drawback of presenting an "artificialised" topography due to the numerous anthropogenic installations (construction of roads, railway tracks). Such installations usually imply the excavation of materials in very high quantity and/or the accumulation of the excavated materials over significant thickness to produce more rectilinear layouts and milder gradients in favour of establishing communication routes. Before GIS, contour lines on topographic maps were digitalized using lines. Today, GIS contour lines are deduced from a grid of points that gives a much better modelling of the landscape (Ghilardi, 2006). To create more realistic palaeo-topographic reconstructions throughout the different periods of the site's occupation, the contour lines must be re-interpreted manually in the GIS whilst ensuring that the overall aspect of map contour lines is respected as much as possible (Ghilardi, 2006; Ghilardi et al., 2007).
9In addition to terrestrial data, it is appropriate to complete the D.E.M. in marine environment to produce an overall topographical view of the concerned areas, both above and below sea level. Bathymetric data provide particularly precious information concerning the topography of the seabed in areas recently affected by the last post-glacial sea-level rise. Bathymetric points, produced using S.O.N.A.R. technique, can be included to the GIS and added to the D.E.M. (Ghilardi, 2006). In addition, L.I.D.A.R. technique is currently employed in the framework of shallow bathymetric surveys (Li, 2005). Photogrammetry and L.I.D.A.R. data complement each other: photogrammetry is more accurate in the x and y direction while L.I.D.A.R. is more accurate in the z direction.Integration of S.R.T.M. data.
10Conventional topographic mapping technologies have produced maps of uneven quality—some with astounding accuracy, some far less adequate. Most industrial countries maintain national cartographic databases. The map products derived from these databases vary greatly in scale and resolution, and are often referenced with country-specific data and are thus inconsistent across national boundaries. The Shuttle Radar Topography Mission produced elevation data on a near-global scale and generated the most complete high-resolution digital topographic database of Earth (Farr, 2007; Rabus et al., 2003). The new S.R.T.M. D.E.Ms. have probably had the largest impact on studies of regions in the developing world for which reliable, high-resolution digital topography was not previously available. With relatively few exceptions, a nearly complete topographic coverage is now available for most of the nonpolar world and provides a foundation for a new analysis of diverse landscapes (Farr et al., 2007).
11G.P.S. is an excellent data collection tool for creating and maintaining a GIS. It provides accurate positions for point, line, and polygon features. By verifying the location of previously recorded sites, G.P.S. can be used for inspecting, maintaining and updating GIS data. G.P.S. provides a tool for validating features, updating attributes and collecting new features. Differential correction techniques are used to enhance the quality of location data gathered using G.P.S. receivers. The underlying premise of D.G.P.S. is that any two receivers that are relatively close together will experience similar atmospheric errors.
12The different shapes (points, lines, polygons) are georeferenced and connected with databases. Regarding the present and the past environments, stratigraphic, sedimentological, palynological and/or chronological (14C datings) information can be collected. The archaeological databases can integrate information concerning the architecture, the function and the dating of buildings constituting the archaeological sites. The cross-comparison of these informations into the GIS allows palaeo-landscapes (hydrology and vegetation, in particular) and palaeo-topographies reconstruction.
- 3 The thickness of post-Neolithic sediments (peat deposits at the location of present dried up lake (...)
13The Korça Basin, located in southern Albania, is a plain at 818 m surrounded by high mountain ranges which culminate at 2028 m. The nortwestern part of this basin was occupied by Maliq Lake until drainage works in the 1950s. Probably due to climatic variability and, since the second half of the Holocene, to anthropogenic forest clearances in the catchment area (Bordon et al., in press), the surface of the palaeo-lake varied between a minimum of 40 km2 during periods of low level to a maximum of 80 km2 (Fouache et al., 2001). From the Early Neolithic period (around 9000 B.P.) to the Early Iron Age (2300 B.P.), and especially during the Middle Bronze Age (around 4500 B.P.), the nearby lake shore was occupied by several settlements like Maliq (Prendi, 1966) or Sovjan (Touchais et al., 2005). These settlements were studied by a French-Albanian archaeological team to elaborate a model of human implantation around the palaeo-lake Maliq. To perform surveys, palaeogeographical reconstructions of the palaeo-lake were established using GIS and D.E.M. taking into account archaeological, geological and new palaeo-environmental data3. Then, geological, palaeo-environmental and archaeological records have been included to a GIS and connected to the S.R.T.M. topographic data controlled with D.G.P.S. measurements. Figure 1 presents a 3D modelling of four stages of palaeogeographic reconstruction of the Maliq Lake along the Holocene period.
Figure 1. Four palaeogeographical reconstructions of palaeo-lake Maliq
a: Last Glacial Times; b: Early Neolithic; c: Middle Bronze Age; d: Roman Times. The four lake dwelling sites (Sovjan, Maliq A, Maliq B and Maliq C) discovered by the archaeological team are on the nearby reconstructed lake shores.
14The reconstruction of the extension of the palaeo-lake during high levels, together with the knowledge of the thickness of the sediment (accumulation of colluvial deposits) covering settlements allowed us to design a predictive map of the potential archaeological layers for the Neolithic, the Bronze Age and the Iron Age (Fig. 2). Since the lake level rised between the Neolithic and the Iron Age, the increase of the extension was taken into account to determine potential areas where sites could be fossilized. The preliminary results of the prospecting carried out in August 2007 confirmed the predictive map: lacustrine sites were actually found in the areas designated by the GIS-based predictions.
Figure 2. Predictive map for surface archaeological surveys of the Korça Basin and thickness of post-Neolithic sediments. The thickness of sediments covering archaeological layers is inferred from boreholes studies.
15The Thessaloniki Plain is the largest deltaic complex in Greece, covering an area of approximately 2000 km² (Fig. 3). This vast deltaic complex presents a flat relief-topography and originates mainly from the coalescence of alluvial deposits from Aliakmon and Axios Rivers, over the past 6000 years (Ghilardi, 2007; Ghilardi et al., 2008a; 2008b). .The palaeo-environmental study allowed reconstructing the landscape evolution for six millennia (Ghilardi, 2007). Based on chronostratigraphic sequence (14C A.M.S. datings performed on marine shells and peat episodes), derived from borehole analysis, this important work for the area highlighted the rapid infilling of a shallow bay from the Neolithic period. Up to a maximum depth of 11 meters, eight boreholes recorded deltaic sediments, ranging from marine environment (the lower part) to lagoonal deposition (the middle part) and finally to fluvial deposits (upper part); the microfaunal helped in differentiating the different environmental conditions. Subsequently, sedimentological analysis helped in classifying the grain-size distribution (clays, silts, sands, coarse sands) and in identifying the contribution of the different drainage-basins. The rather flat appearance of deltaic areas does not reflect a lack of morphological processes. The three-dimensional display of minor relief forms (deltaic lobes, debris flow, alluvial fans…) often transpires to be difficult to implement due to the inaccuracy of available cartographic documents and also due to the fact that research scales are often oversized (Ghilardi, 2006). The different landforms (former levees, alluvial fans, etc.) are identified on satellite image as false colour composite objects. To obtain altimetric information, high-resolution topographic data derived from S.R.T.M. surveys are added in a GIS and superimposed on the satellite imagery. Subsequently, topographic information is linked to the palaeo-environmental results derived from borehole stratigraphy. This combination allows a spatial interpretation and a palaeogeographic reconstruction of the whole area, including location of settlements (see Fig. 4 for a palaeogeographic reconstruction over the last 6000 years).
Figure 3. 3D view of Thessaloniki Plain using S.R.T.M. data.
Superimposition of the archaeological settlements and hydrographical network with the SRTM data. The topography of the Thessaloniki Plain varies between 0 and 10 meters from the actual shoreline to the north, close to Ancient Pella (a maximum length of 32 km), and between 0 and 10 meters to the west, close to the Neolithic settlement of Nea Nikomedia. The city of Methoni is located along the Pierian coast on the meridional border of the delta (Ghilardi et al., 2007). Red dots indicate Neolithic settlements, green dot indicates the capital Pella, light pink dot indicates the ancient settlement of Methoni (Sites A and B correspond to the sites identified by Hatzopolous et al., 1990). The dots circled in black colour are described in this article.
Figure 4. Palaeogeographic reconstruction of Thessaloniki Plain from Neolithic period to the present-day.
Panel 4a: the actual plain of Thessaloniki is occupied by a large marine gulf circa 4000 B.C. This period corresponds to the maximum shoreline extension during the last post glacial sea level rise. Panel 4b: in 2500 B.C., the bay starts to be infilled by terrestrial deposits coming from Aliakmon and Axios rivers mainly. The rapid growth of their respective deltas created some levee gradually transformed into natural dams and lagoon—brackish environments around the margins of the bay. Panel 4c: the novel feature of the plain is the appearance of a lake, confined to the western part of the bay, around 1600 B.C. In the area of the Ancient Pella, at these times, shallow marine conditions appear. Panel 4d: around the 4th century B.C. the Aliakmon and Axios deltas grew. The probable narrowing of the bay is from this epoch: the junction between the two main rivers draining the plain is not efficient, but there is a very small strait which permits the passage of boats until Pella. Panel 4e: gradual silting up of the harbour of Pella around 300 A.D. and the lacustrine occupation. Panel 4f: morphology of the plain nowadays.
- 4 for the D.E.M, we chose a series of topographic maps scaled to 1:5000. The digitalization of point (...)
- 5 1770 bathymetric points have been produced using S.O.N.A.R. The recorded sector, corresponding to (...)
16The ancient settlement of Methoni was an important harbour closely affiliated with the Athenian Alliance (5th Century B.C.). According to historical manuscripts, the urban settlement was distant from the harbour even though neither the distance nor the potential location of the harbour are documented. Using the D.E.M4 (digitalization of points on contour lines, integration of bathymetric surveys: the different shape files were integrated in a GIS) key landforms were identified indicative of the infrastructure of the ancient harbour (natural bays fossilized by intense sediment transfers in a deltaic context; Ghilardi, 2007). In addition to terrestrial data, a D.E.M. in marine environment was performed to produce an overall topographical view of the Methoni region both above and below sea level. Bathymetric data5 enabled completion of this marine D.E.M. and precised the topography of the Methoni bay:
The three-dimensional view of landscapes revealed signs of the intense morphological activity. In the North of the archaeological settlement, there is a sector in which contour lines are represented in a concentric manner, representing a mild and regular gradient. The hypothesis of the presence of an alluvial fan can be made. On the digital elevation modelling, slope transfer activity (transfer of sediments along slopes that have not been transported by river flow) is visible along the former active cliff of Methoni. Indeed, where the escarpment meets the low zone (made up of deltaic sediments), we observe that the contour lines are "disharmonic", showing no concentricity. This is a telltale sign of an impermanent runoff that has been subjected to irregular phases of material transfer along slopes.
Today, we propose two candidate sites for the ancient harbour infrastructure away from the city (Fig. 5): two natural bays that remained unfilled by sediments after the classical period (Ghilardi, 2007). Further palaeo-environmental investigations, based on boreholes analysis and chronostratigraphic sequence could help significantly in reconstructing the sedimentary history along the Pierian coastline. Archaeological excavations in the two former bays will provide important results to confirm or not the presence of these harbour infrastructures.
Figure 5. Proposal of location of two port sites for the city of Methoni (3D view of the sector).
If two sites of occupation have been identified for the ancient city of Methoni—Sites A (archaic and classical periods) and B (Roman Times) (Hatzopoulos et al., 1990)—the locations of the respective harbours are still unknown. Two natural bays that remained unfilled by sediments during historical times have the potential to be those ancient harbour sites.
17Over the last decades, archaeologists and historians have faced the necessity to reconstruct ancient settlements history not only through the study of the material excavated, but also with the use of palaeo-environmental parameters. For this reason, geographers were invited to collaborate and include their results in georeferenced maps allowing a spatial interpretation of the laboratory analyses. This paper describes several powerful methods to infer the evolution of landscapes in the context of such multi-disciplinary/geoarchaeological programmes.
18GIS is now the main digital support for scientists from various disciplines to reconstruct landscape around ancient settlements. The layers created in a digital format can have topics developed in Human and Social Sciences (Archaeology, Geography, History) as well as in Earth Sciences (Geology, Geochemistry, etc.). The main aim is to develop techniques and tools for multidisciplinary programmes dealing with the historical reconstructions of the landscape frequented by the Human societies since the last glacial period (circa 17500 BP).
19When combined with Digital Elevation Models, GIS represents an essential preliminary step for all geoarchaeological research. Information concerning relief forms provides insight into the morphological evolution of landscapes and gives a basis for selecting potential sites for future excavation campaigns. Today, the three-dimensional reconstruction of environments is the best available method to produce a common reference. Dynamic and three-dimensional thematic maps using the Digital Elevation Model as a reference document must be used in the framework of multidisciplinary programs. The gain in time and resources is also substantial.
20One of the limits encountered in the geomatic approach for geoarchaeology is the choice of the geographic scale of study: archaeologists focus on small structures (walls, etc.) or on simple pottery shards (sometimes no more than 10 cm in length) while geographers and specialists of Earth Sciences (Geology, etc.) employ different working scales which can be extended to hundreds of squared kilometres. Therefore, GIS can be used with difficulties by the different disciplines and need to be well adapted at a spatial level. Other problems can be observed within a discipline: source documents can be more or less reliable, for example it is still difficult to georeference maps older than the beginning of the 20th Century, and to adapt archaeological charts without spatial references in a GIS.
21Perspectives for the use of GIS in geoarchaeological studies seem limitless and encompass: surface microtopography surveys, mapped surface finds, data from test pits and excavations, and many multispectral and geophysical remote sensing data. All applications combined in one place, should yield tremendous potential for understanding site content, organization and structure. Multimedia presentations could offer video, sound, photographs, drawings and animated 3D views. In doing so, free Internet-based Software, such as Google Earth© and Geoportail©, which use 3D views could be implemented with additional data. Indeed, palaeo-environmental results provided by a large amount of international scientific programmes could be added and sea level rise since the last glacial period could be modelled, allowing not only 3D landscape reconstruction but also 4D modelling that relates long term evolution of shorelines displacement.
22As presented in this article, geoarchaeological studies offer now an unprecedented level of integration between disciplines to visualize a shoreline and its displacement. Over the last 20 000 years, humans had to constantly adapt their lifestyle according to the displacement of the shoreline. Given the current threats and uncertainties related to climate change, it is predictable and desirable that many disciplines will adopt similar integrated approach to model their favourite object of research. More generally, GIS offers a tremendous opportunity for scientific outreach and its international common databases are now ready to be shared for new purposes and adapted to create new usages beyond scientific communities.
Acknowledgements are addressed to the CNRS for financial support through the ECLIPSE project “Variations climatiques et dynamique des écosystèmes au Sud des Balkans au cours du dernier cycle climatique”, coordinated by A.M. Lézine and E. Fouache. Special thanks to the members of the Franco-Albanian cooperative project including the French School of Athens (Greece) and the Archaeological Museum of Korçë (Albania). Fruitful remarks from Theodoros Paraschou (Anafi) and David Psomiadis (University of Dimokritos, Athens, Greece) were highly appreciated. Finally, special thanks to the Ecole Pratique des Hautes Etudes (Paris-France) for technical support.