Nancy L. Barrickman1, Patrick Lewis2, Lee R. Berger3, and Steven E. Churchill1

 

1Department of Biological Anthropology and Anatomy, Box 3170, Sands Bldg., Duke University, Durham, NC 27710

 

2Department of Biology, Pfeiffer University, Misenheimer, NC  28109,

 

3Bernard Price Institute for Palaeontology, University of the Witwatersrand, Johannesburg 2050, South Africa

 

Introduction

 

South African breccia caves often have complex depositional histories, and determining the site formation processes can be difficult.  Disentangling the spatial associations between the geological features, fossil remains, and artifacts is imperative to correctly dating the site, deciphering cave taphonomy, inferring hominin activity, and reconstructing the paleoenvironment.  One such breccia cave - Plovers Lake - is crucial to understanding human occupation of the Witwatersrand area during the Middle Stone Age.  The MSA levels of Plovers Lake have produced thousands of faunal remains, hundreds of lithics, and several hominin fossils.  In addition, there are extensive calcium carbonate flowstone layers within the deposits that sealed and protected the deposits from subsequent disturbance by biotic and physical agents.  The site is one of only six MSA sites to produce hominin remains, so understanding the spatial relationships between the different types of data is crucial to understanding the evolution and adaptation of hominins from that time period.  Here we report on the use of Geographic Information Systems (GIS) to store, retrieve, analyze, and display excavation and geologic data as an aid to interpreting site formation processes.

 

 

A laser theodolite was used to provenience all of the fossil remains, lithics, and rocks >10cm.  The cave walls and roof were mapped used the same electronic surveying equipment.  In addition, the calcium carbonate flowstone layers were mapped as they were revealed during excavation, and prior to their removal.  All of the data were stored on Microsoft Access databases.  The information on these databases was expanded by the various researchers on the project (working on hominin and faunal remains, cutmarks and taphonomy, and lithics) as their analyses proceeded.

 

We used ArcView 9.0, a GIS software package that includes programs such as ArcCatalog and ArcScene, to create a three dimensional model of the cave and the finds uncovered during the excavation (see Figures 1 and 2).  We created a local coordinate system (i.e. we did not need to specify a large-scale map projection for a project of this scale) on the ArcView program.  With this information on the mapping scheme, ArcView directly converted the three-point provenience data stored on the Access databases into spatial renderings or “shapefiles”.  In addition to representing the spatial location of these finds, the software stored attribute data that the researchers on the project had determined from their analyses.

 

We reconstructed the cave walls, roof and flowstone layers by creating surfaces for each using the Triangulated Irregular Network (TIN).  ArcView constructs a TIN by triangulating a set of vertices, which are the points that we mapped using the theodolite.  The vertices are connected with a series of edges to form a network of triangles.  Within this matrix, the faunal and hominin remains, lithics, and recent artifacts were plotted using ArcScene.  The ArcMap program does not easily handle three-dimensional data of this sort.  It has difficulty rendering a surface, such as the cave roof or the flowstone layers, that overlies a series of data points, such as the fossil and lithic finds.  The surface cannot share the same x and y dimensions as the other points, even though they may have different z dimensions (see Nigro et al. 2003).  However, ArcScene allows for this sort of layering within three-dimensional space.

 

 

Figure 1.  Three dimensional model of Plovers Lake, view from above facing northwest

Figure 2.  Three dimensional model of Plovers Lake, view on level facing south southeast

 

References:

Nigro JD, Ungar PS, de Ruiter DJ, and Berger LR (2003) Developing a Geographic Information System (GIS) for mapping and analysing fossil deposits at Swartkrans, Gauteng Province, South Africa. Journal Of Archaeological Science 30:317-324

Last updated by J. Rhodes 20, April 2006