REPORT FOR THE AFP grant 00-02-01-06, Faculty of Architecture, METU.

Kerkenes Homeline.GIF (135 bytes)previouscontentsnext

RESULTS: Geographical Information Systems


Figure 27

Figure 28

Figure 29

The aims of this facet of the Kerkenes Project are:

1. Construction of a database containing the graphic information that has been obtained over the northern end of the city and other selected areas.

2. The linking of graphic data to different data sets in tabular form.

3. The linking of graphic data to textual and pictorial data sets.

The Kerkenes Project has, since its inauguration in 1993, gathered huge amounts of data in different forms. GIS appears to provide the best possible platform on which these widely variant types of data set can be integrated, thereby allowing for the implementation hierarchically ordered levels of analysis. The advantages of selecting GIS as the platform are twofold:

1. By far the largest data sets are graphic, e.g. aerial (balloon) photography, survey grids and nets for spatial referencing, measured maps and plans, Global Positioning System (GPS) 3 dimensional data, geophysical data of sub-surface features (very largely buildings), detailed structural plans from limited excavation. Other data sets will be constructed in the future, including geological, geomorphological, soil cover, vegetation cover and microclimatic.

2. The nature of the data is, and will continue to be, very largely spatial. The types of analyses that we wish to apply to the data sets are to do with concepts that come from urban geography, city planning and cultural interactions with spaces.

Database Construction for GIS
Features are added to the database through digitising from a variety of imagery. Geomagnetic images provide the most detailed evidence available over large areas. The interpretation of the geomagnetic data has, in a few instances, been proven to be correct by precise experimentation in the form of excavated test trenches. (This is of particular importance with respect to roofed and unroofed spaces and the recognition of stone paved external areas).

Currently there are three levels or layers within the GIS database, (Figs 27, 28, 29) to which more will be added once an assessment of this first experimentation has been completed.

Level 1 large enclosed spaces (termed urban blocks)

Level 2 roofed structures (buildings)

Level 3 small roof units (rooms)

Tabulated data, in the form of attributes, cover a huge variety of subjects, spatial, material (stone, mud-brick etc), relational (e.g. position of doorways). Linked data can include other sections of the existing database, such as photographs, geophysical imagery, description in text form, slope and aspect maps.

Urban Blocks
· Graphical: Over the northern end of the city the large walled spaces, termed urban blocks, have been digitised from the geophysical data, and additional information has been added from measured maps (made by total station survey) and balloon photography.
· Tabulated: Tables of data in tabulated form have been added to each block. The attributes that have been included at this early stage represent an experimental selection, and do not embrace the full range of attribute categories. Further, The availability of data for each of the selected attributes varies from block to block, primarily because the clarity of information that can be derived from geomagnetic survey varies from place to place. This variation in the clarity of geophysical data occurs for a variety of reasons that range from the quality of data collection, the intensity of burning in the destruction of the city, underlying geology and the depth of overburden.
· Assessment: The data appears to be good and reliable and it should now be possible to make the first analytical trials using GIS.

Buildings or Structures
· Graphical: Within each urban block an attempt has been made to identify and digitise individual buildings and structures. This is not as straightforward as might be expected because it becomes necessary to introduce levels of interpretation. It is only as the process of database construction progresses that the levels of interpretation required become clear. It will perhaps become necessary, or at least prove to be useful, to rank different levels of interpretation according to some scale of confidence: a stage that has not yet been reached. Indeed, there appear to be different classes of interpretation, some of which are more difficult to evaluate than others.
· Tabulated: Each structure or building has a set of attributes in tabulated form. In the majority of instances, such as two-roomed structures and rows of small rooms, there ought not to be any obvious difficulties (provided that the geomagnetic image is clear). Some larger complexes pose questions of what to include and what to exclude from a "structure". It is not possible, except in the rarest of instances, to reconstruct the order in which structures were erected; thus there is no time dimension that requires incorporation into this section of the database.
· Assessment: The first experimentation has made clear that the identification of structures and buildings, and the determination (for the purposes under consideration here) of such attributes as walled spaces, roofed areas and external stone pavements requires the initial input of architect and/or archaeologist together with, on occasions, a geophysicist.

Rooms or Roofed Units, and Internal Open Spaces or Courtyards
· Graphical: Buildings or structures generally comprise more than one unit. Individual units are usually in the form of rooms, but we do have some evidence from open units, such as courtyards and storage areas, that are integral parts of buildings.
· Tabulated: Tables of data in tabulated form are added to each room. As with the buildings, the attributes that have been included at this early stage represent an experimental selection, and do not embrace the full range of attribute categories.
· Assessment: Where four walls can be seen the identification of rooms raises less doubt than the recognition of the parameters for multi-roomed structures. It might be worthwhile to begin by defining rooms, and then proceeding to buildings.

Kerkenes Homeline.GIF (135 bytes)previouscontentsnext