Archaeological Site Modeling

John C. Sanders, Head
Oriental Institute Computer Laboratory

Peggy M. Sanders
Archaeological Graphic Services


Archaeology and Data Acquisition

Archaeological projects amass large quantities of written documents, notes, and forms, as well as drawings (plans, sections, and sketches) and photographic images of the ancient sites, architecture, and artifacts that are recovered during survey and excavation. Effectively utilizing all this material in the pursuit of research goals has always been a major challenge confronting archaeologists. Further seasons of excavation will, of course, produce even more data to be integrated with current information. The sheer amount of data which must be processed and evaluated by project members quite often necessitates the adoption of new tools and strategies for interpretation and analysis.

Although these new tools are vital for future investigations, their use will only magnify an on-going responsibility of the field of archaeology. During excavation, archaeologists expose and destroy forever the extremely sensitive context of ancient artifacts and architecture. After all the elaborate research designing, planning, and study, conclusions about ancient societies are limited to how well archaeologists record the material they excavate. After a traditional excavation is completed, line drawings, photographs, and written descriptions are all that is left of the clues to an ancient site, household, temple, or burial.

Archaeology, Templates, and Computers

The field of archaeology still relies on the traditional plan and section drawings as its primary form of graphic documentation, the standard 'templates' that are used to represent the ancient past to the general public, to the scholarly community, and to pass such information on to generations of students. In the 21st century, the emerging technologies of computer-aided drafting (CAD) and surveying instrumentation, remote sensing/satellite imaging, digital scanning, Global Positioning Systems (GPS), photogrammetric mapping, and digital/video (multimedia) imaging are of critical importance for modelling archaeological data, and Geographic Information Systems (GIS) and Exploratory Data Analysis (EDA) software are of critical importance for archaeological analysis. New 'templates' will emerge using these tools, the construction of vector and raster-based three dimensional computer models of ancient sites and monuments, architcture, artifacts, point proveniences, and loci. The development of computer graphic databases instead of drawings means that archaeologists can view their data from any perspective or orientation, through time, under user control.

Archaeology and Computer Analysis

Computer databases and generated imaging are more than an aid in illustration. They help archaeologists comprehend ancient structures and sites. Computer models allow archaeologists to accurately measure and analyze ancients monuments, to extrapolate the original appearances of deteriorated monuments, or to recreate the forms of ruined domestic structures and the morphology of ancient settlements. Computer models of ancient buildings and structures can be used for "finite element analysis" of material and structural properties, for accelerated weathering studies to create digital images of artifacts and features after so many years have past given certain environmental conditions, and for fenestration/light studies of ancient structures and the daily movement of the sun throughout the year.

CAD and GIS

CAD systems emphasize computer graphics, whereas GIS technology places emphasis on building an intelligent database of artifacts, features, and their attributes, based on their spatial coordinates and complex interrelationships as objects in three-dimensions. GIS programs use spatially related data to perform a host of analyses, and they can create new spatial information from existing spatial data.

Topology is a branch of mathematics that deals with relationships among geometric objects. A topological data structure is the foundation for many of the analytical functions found in a vector-based GIS program, defining and managing the connectivity and contiguity among database features. Connectivity and contiguity are the types of relationships that are intuitively obvious to the human brain when viewed on a map but difficult to translate to the computer. Topology provides the means by which to accomplish this translation.

Vector-based GIS programs use a topological structure consisting of discrete points, line segments, and arcs to represent features and identify locations. Lines can also be joined to form polygons representing boundaries, such as features, settlements, vegetation zones, etc. Creating a vector-based topological structure involves defining and maintaining the relationships among all the lines and nodes making up linear segments and polygons of a GIS database (nodes are defined as the points representing the intersection of line segments). The relationships handled by topology include coordinate locations of all features as well as the polygons to the right and left of each line and the line segments extending in each direction from each of the nodes.

The beauty of topological structure comes in performing tasks such as updating a map to incorporate a change in the boundary of a feature. It is topology which ensures that when a feature boundary is moved, the change is automatically reflected in all connecting features and that the area and perimeter calculations are changed in the database accordingly. When a topological structure is combined with a relational database management system, GIS software automatically updates the relational database to acknowledge feature changes and thus keeps track of the full implications of these changes throughout the database. The intelligence implicit in the connectivity and contiguity provided by topology is of crucial importance to archaeological research because it can answer such queries as:

  • Is feature 1 simply near feature 2, or do they share a boundary?
  • What features are found inside Room 1?
  • What is the volume measurement for Room 1?
  • What features lie on the shortest route from point A to point B?

The ability to answer such descriptive questions does not come from a database of pre-defined characteristics on which a query is performed, but instead from a "real-time" analysis of the spatial locations and relationships of "objects" in a three-dimensional, GIS database.

In raster-based GIS programs a region is represented by a matrix of grid cells forming a sequence of rows and columns on the X,Y axis. Each cell or collection of cells has a numeric Z value that represents some database characteristic, such as elevation, soil type, land cover, or terrain slope. Values for the Z component are assigned to the grid cells using a variety of different methods, including presence/absence, highest/lowest values, average value, predominant value, etc. In the case of satellite-derived imagery, the Z component is a value for reflected sunlight from the surface of the earth. Thus, a satellite image is raster data consisting of a matrix of cells that capture areas of the earth's surface as picture elements (pixels). Raster-based GIS databases are particularly well suited for analysis and modelling applications not only because virtually any type of data can be encoded and stored in a "cell-based" format, but because these data can be accessed for univariate or multivariate analyses and because complex algorithms or decision models can be applied to them.

Recent trends in GIS data formatting have been from raster formats in the late 1970's, to vector formats in the 1980's, and now to integrated raster/vector formats in the 1990's. Grid cell databases and raster processing are having a resurgence for a couple of reasons:

  1. government agencies such as NASA and private companies such as SPOT Image Corporation have developed and made available very large raster-based, satellite image databases, and the kinds of analyses that require overlaying of multiple data formats are performed much more efficiently in a raster format.
  2. raster-based, satellite images are now providing the foundation for many of the vector-based databases being created around the world.

The integration of raster and vector data analyses of a single GIS database can produce a much more comprehensive spatial model than a GIS of either the vector or raster format alone. Under these conditions, vectors serve to abstract or emphasize features, while the raster pixels capture and display the intervening spatial detail important to any analysis. As an example, satellite imagery can be incorporated as a base map for presenting regional coverage, settlement pattern development, and other large scale themes. The ability to overlay vector information, such as boundary defintions, onto digital images provides a powerful capability for data extraction, verification, correction, and visualization. It allows for upgrading the quality of vector-based maps and plans by using raster pictures to update and refine errant vector data, or to create new vector elements.

Using the ancient Egyptian pyramid complex at Giza as an example, rather than relying on traditional profiles of the insides of the Khufu (Cheops), Khafre (Chephren), and Menkaure (Mycerinus) pyramids, a computer model of the Giza Plateau allows the viewer to move through the chambers and passages of a three-dimensional database of these monuments, a representation of the same structures in computer memory. A GIS-based computer model can aid in sophisticated predictive modelling of archaeological data, and can improve archaeological analysis by detecting and mapping distributions of artifacts and features within a structure, or throughout the entire Giza Plateau.