1. Topographical Survey

1The Kephali Hill is an oval-shaped hill located at 35°18’32’’ N latitude and 25°31’43’’ E longitude with a maximum elevation of ca. 22.46 m above mean sea level. The two slopes of the hill are cut by valleys, forming the Bouffo and Avlaki-Kharkoma beaches respectively to the northwest and northeast, making the hill a prominent landscape feature. The overall aim of the topographical survey of the Kephali was the investigation and mapping of visible architectural and topographical features. The focus of the survey during the 2007 campaign was the mapping of the terracing system. Additionally, accuracy and precision of the topographical information, obtained through existing topographical maps of the site1, were improved by collecting as many Total Positioning System (TPS) measurements as possible. In 2008, mapping of terrace walls and collecting topographical information continued2. Moreover, a number of standing architectural remains were mapped and architectural descriptions of these remains were completed. All this information has been processed in a Geographical Information System (GIS) environment. Since starting the survey in 2007, about 2 hectares of the hill surface have been covered. An area of about 0.5 hectares remains to be done since parts of the hill are still covered by bushes or not visible from the current reference points used for TPS measurements.

2Mild hill slopes (< 10 degrees) appear to have had a favourable impact on wall preservation (fig. 2.1). Parts with relatively mild slopes (ca. 10 to 20 degrees) have only few visible remains among which those of the “Cyclopean wall” (fig. 1.6) and burial area to the north. The cliffy northeast side of the hill which carries narrow natural terraces possesses the steepest slope (ca. 40 degrees). At the moment, shrubs covering the limestone cliffs do not allow for verification on whether these terraces may have formed locales for potential rock shelter burials. Preliminary observations in the north-eastern area lead us to suggest that some of the limestone outcrops may have been used for quarrying3 but as yet no traces of tool marks can be made out.

1.1. Terracing System

3Artificial and natural terraces on the hill are formed either by dry-laid terrace walls with small stone packing or by surfacing bedrock rising or falling steeply. Besides substituting for terrace walls, it is possible that in some cases surfacing bedrock was used as a foundation for terrace walls. Future excavation and survey may illustrate this. Although terrace walls were reused and modified in later periods – and probably especially during the 20th century – the lower courses of many of the walls seem to have been constructed during the Bronze Age. Common benefits for horticultural and building activity aside, the terracing system on the Kephali could also have served as an element influencing functional organisation of the site. Preliminary observations suggest that the relatively long terrace wall to the north, SF039, may have formed the spatial separation between the settlement and the funerary area on the north side of the hill (see Parker Pearson 2003: 124). Note that another large wall, SF044, to the southwest of SF039 is likely to be the continuation of SF039 as the size and the material (limestone) of the blocks used in its construction are similar. SF039 is better preserved than SF044 and stands with a single facing of average-sized blocks (ca. 0.60 m edge size), mostly three courses high (fig 2.2).

4The longest terrace wall preserved in one piece (SF002) is located on top of the hill (length = ca. 105 m). It was partly cleaned and excavated during the 2008 campaign and its relation to the architectural remains in the area has been investigated4. Another terrace wall, SF047 (length = ca. 30 m), forms the continuation of SF002 at the north and turns east at a right angle. This wall is well-preserved with a single face exposed and blocks standing four or five courses high. Like SF002, the lower courses of SF047 may have been constructed during Minoan times but excavations are needed to confirm this. Terrace wall SF038, sloping down to the north of SF047, is very badly preserved. The limestone and size of the few blocks preserved in situ are similar to those of the blocks in the lower courses of SF002 and SF047. Hence, it is possible that SF002, SF047 and SF038 were part of a single Minoan construction.

5Another terrace wall in the southeast part of the hill (SF027), located to the northeast of Zone 6 (grids BK89, 90, 91 and BL89, 90, 91) may also date to the Bronze Age. Although founded on steeply sloped ground (ca. 70 percent slope) and severely eroded, some of the average-sized blocks in the lower courses of this wall are still in situ. A small wall, SF056 (length = ca. 5m long), oriented SE-NW and located to the southeast of SF002, immediately west of SF027, is noteworthy as it may be the remains of a wall originally linking SF002 and SF027. If indeed SF002, SF056 and SF027 comprised a single wall surrounding the hill top at its west, north and (at a lower elevation) east sides, then this raises the possibility that at some stage, the settlement may have had an acropolis-like configuration partially or completely surrounded by a terrace wall5.

6At the moment, it has not yet been possible to identify remains of terrace walls in the southern area of the site because of vegetation. It remains to be seen whether the Cyclopean wall (figs 1.6-1.7) on the southwest corner is related to the terracing system higher up the hill or forms part of another structure. This wall is remarkable for the size of its blocks since, for example, the boulder used in its corner measures ca. 2.30 by 1.20 m.

1.2. Other Architectural Remains

7In the area to the north-northeast of terrace wall SF039, remains of nearly twenty, badly preserved walls were traced, mostly made of small stones. Large segments of these walls remain buried but eleven were mapped during the surveys. The walls seem mostly to comprise single lines of stones and are likely to be the remains of other “built tombs” like those excavated in nearby Zone 1.

8SF016 denotes the sharp elevation difference in the bedrock, forming the lower limit of the terrace where burial areas 1.11 and 1.12 are located6. The elevation difference varies along SF016, reaching a maximum height of ca. 2 m towards the middle, where the rock shelters explored during the 2008 campaign are located. The elevation difference created by the bedrock gradually dwindles down towards the east and west. Almost parallel to SF016, to the south, one can see the remains of terrace wall SF014. To the west of SF014, a few stones of a badly preserved terrace wall (SF015) are still in situ. Considering the similarity of size and material of the stones in this wall to those used in SF014, it may be suggested that SF014 and SF015 originally formed a single terrace wall. A relatively large Z-shaped feature (SF013), located on the terrace bordered by SF016 and SF014, is noteworthy as its size and configuration differ from the rest of the structures in this area. The feature labelled SF013 may actually form a series of walls that belonged to two different structures: the longer wall (length = 6.75 m) has an overall northwest-southeast orientation although slightly turning eastward halfway. The terrace bordered by SF014 to the north and SF039 to the south could also have been used as a burial area. It is possible that the architectural remains in this part of the site are better preserved since the mild slope limited erosion and may have resulted in burying the features.

9Among the other walls visible to the west and northwest of terrace wall SF044 on the north-western part of the hill, SF021 and SF045 are conspicuous due to the large size of their blocks. SF021 consists of two related walls perpendicular to one another forming a “Γ” shape. The alignment of the stones west of SF021 consists of two surfacing, average-sized blocks. It is perpendicular to SF021 and may be related to it. Visible remains of SF045 to the east of SF021 seem to have been constructed with blocks comparable in size to those of SF021. Two depressions on the same terrace (SF062 and SF063) together with a rectangular, U-shaped feature (SF041) – located in between the two grand terrace walls (SF039 and SF044) – may have been created during the 2nd World War. SF041 is a probable gun post and may have been constructed by reusing the stones of an ancient terrace wall that originally connected SF039 and SF0447.

10Near the terrace bordered by SF002 and SF047, to the north of the hill-top, there are five discernible architectural features. The northern parts of SF002 and SF047 form a right angle with two of these structures (SF040 and SF069), and parallel to two other walls (SF070 and SF071). A group of shrubs 10m east of SF071 covered an area of ca. 15m² and was removed on the last day of the 2008 campaign. Before the shrubs were cut, three boulders (on average 0.70x0.70x0.80m) of a possible building had already been observed as jutting out. After clearing, the remains of a large building with similar-sized boulders were discovered. The building will be further investigated during the 2009 campaign.

11The ashlar block to the south-south east of the hilltop (grid BM84) is worth noting. Although few walls are visible in the southern part of the hill, the ashlar block may suggest the presence of a structure of fine architecture. The area in the vicinity of the block will be investigated during future excavations.

2. GIS Applications

12Geographical Information Systems (GIS) undoubtedly form one of the most essential tools in contemporary archaeological research. Generally, GIS-based techniques are used for data analysis at the post-excavation stage and only rarely from the pre-excavation stage preparations onwards and in actual fieldwork8. Employing GIS, however, to conduct a wide spectrum of applications within a single digital environment at the different stages of an archaeological project has several advantages.

13Prior to the first excavation campaign in 2007, a spatial database was created using the GIS software of Arc-GIS^TM. Topographical and archaeological information previously collected within the research group at the UCL9 formed the foundation of this database. Information was organised in three different datasets according to their content and, subsequently in different classes, according to their content and geometry (point, line or polygon) selected for the representation: archaeological information (comprising classes of object_point.shp, object_polygon.shp, feature_point.shp), interpretations (building_point.shp, room_polygon.shp) and landscape features (coastline.shp, caves.shp). Subsequently, all the information in the existing topographical maps of the site was digitised and a Digital Terrain Model (DTM)10 of the Kephali Hill was created (fig. 2.3). The DTM was used to obtain a good understanding of the morphology of the hill and its environs. The model created a certain familiarity with the topography of both the site and the modern settlement around it and assisted the team in planning the excavation campaign.

14During the campaigns of 2007 and 2008, spatial data were collected using different methods and different formats: TPS measurements, hand drawings, levelling instruments, pictures taken by hand-held digital cameras, aerial pictures, geophysical survey and textual information collected with the field laptop computers. This data was processed and integrated within the spatial database as follows:

• On a daily basis, TPS measurements were downloaded as Leica IDEX files and opened as a text file. From these text files, Excel files were created which were in turn imported into the GIS environment. Coordinate information from these files was used to plot the TPS measurements as points.
• Hand drawings and sketches of the archaeologists were scanned, compressed into JPEG-format and geo referenced in the GIS environment using the reference points indicated on the drawings. Information in these images (including the elevations collected with the levelling instruments) was digitised in the corresponding layers. For instance, in a drawing where only artefacts and their elevations were illustrated, those artefacts which were to be represented with their shape in the spatial database were digitised in “object_polygon.shp”, other artefacts were digitised in “object_point.shp” and elevations were digitised in “elevation_point.shp”.
• Orthogonal pictures (JPEG) taken with hand-held digital cameras were used to map small areas (i.e. ca. 20 m²) with dense artefact concentrations (such as pottery deposits). These pictures were rectified and relevant information was digitised.
• Orthogonal aerial pictures11 (JPEG) were also rectified and used as a tool to check completeness and accuracy of the architectural drawings within the GIS environment.
• Results of the geophysical survey were received as a GRID file which is a raster data storage format native to ESRI12 and hence could directly be integrated into the ArcGIS database.
• After the campaign ended information in the MS Access database 13, collected with the laptop computers on site, was linked to the spatial database (see below).

15Thanks to the variety of methods used, on-site data collection was flexible, accurate, fast and efficient. One of the most significant benefits of using an integrated, spatial database, updated daily during the excavation campaign, was the “reflexivity” it provided (see Conolly & Lake 2006: 37). This said, trench supervisors were provided with updated maps of their trench regularly during the excavations which helped the data recording process and provided a representation of progress in the field. Information stored in the Access database was linked to the spatial database during the post-excavation stage and interactive maps were created. The aim was to explore detailed textual information collected on site in a cartographical environment provided by the GIS software. The first trials of these interactive maps have been uploaded on the local area network of Université Catholique de Louvain and are used by the trench supervisors and auxiliary disciplines. Users can open these maps using the free software package ArcReader14 and query the data therein (fig. 2.4). These maps provide a means to visually and textually explore the data collected and detect potential errors.

16At the post-excavation stage, the GIS database was also used to prepare archaeological plans for presentations and publications. Trench supervisors actively participated in the planning process by checking the plans of their respective zones and providing further relevant information. Moreover, the team had the chance to try different cartographic alternatives provided by the GIS software used in order to find the best way in which to represent the collected information. Note that all the trench plans in this volume are created with GIS software.