Taphonomic Interpretation of the Developed Oldowan Site of Garba IV (Melka Kunture, Ethiopia) through a GIS Application

Article excerpt


Melka Kunture is a valley site formed mainly of fluvial sediments, extending for over 5 km along the banks of the Awash (FIGURE 1). Its deposits attain a maximum depth of 100 m, and are interspersed with tuff and cinerite providing important chronological markers. The site, discovered in 1963 by hydrogeologist Gerard Dekker, was explored systematically by a French-Ethiopian expedition directed by Jean Chavaillon from 1965 to 1982, and then again from 1993 to 1995. Since 1999, an Italian expedition directed by one of the authors (MP) has been collaborating with J. Chavaillon to create an open-air museum on the site and publish the abundant geological, archaeological and paleontological data gathered so far (Berthelet et al. 2001).


Of the over 70 archaeological levels identified so far, about 20 have been partially investigated, while 8 have been extensively excavated; the latter are referable to the Oldowan (Karre I and Gombore I), the Developed Oldowan (Garba IV, Gombore I[gamma]), several phases of the Acheulean (Garba XII J, Gombore II, Garba I) and the Middle Stone Age (Garba III), with ages ranging between 1,700,000 and 200,000 years old (Chavaillon et al. 1979).

The site of Garba IV is one of the most interesting, especially extensive and with abundant lithic and palaeontological finds spanning a broad chronological range (Piperno & Bulgarelli 1974-1975). The site, excavated from 1973 to 1982 (FIGURE 2), consists of five late Oldowan (about 1,500,000 years) archaeological layers (from the top down: C, D, E, F, G). A right hemimandible attributed to a 3-5-year-old Homo erectus child was found in level E. An area of about 100 sq. m at level D (FIGURE 3), explored systematically, yielded an extremely dense concentration of nearly 10,000 artefacts and over 2,700 faunal remains, mostly of bovids, equids, suids, hippopotamuses, elephants, giraffes and a primate resembling the present-day gelada (baboon) (Geraads 1979).


The lithic tools are made from volcanic rocks (obsidian, basalt, lava, trachyte, tuff). Most of the flake tools are made of obsidian, while the pebble tools are mainly of other volcanic rocks. This lithic assemblage is characterized by a high percentage of flakes and fragments, mono- and bidirectional choppers, heavy scrapers and polyhedrons. It also includes broken and battered pebbles, cores and core fragments, some retouched tools (side-scrapers, denticulates and notches), a few bifacial tools and two cleavers (FIGURE 4).

Computerized archaeological investigations

Intra-site spatial analysis makes a fundamental contribution to the interpretation of prehistoric deposits, especially in extensively investigated sites. Since the spatial distribution of archaeological evidence often reflects the functional organization of a site, statistical analyses aimed at the classification of finds and the highlighting of specific associations can be extremely useful. In a recent essay on the study of artefacts, Djindjian (2001) discusses once again the most commonly employed quantitative methods for the analysis of spatial data. He points out that intra-site analysis is still a relatively new method and hence open to considerable improvement, especially in the selection of data used in spatial distribution models.

In the interpretation of `social structures', considered as-functional areas into which a given archaeological surface is subdivided, special attention should be given to the position of artefacts and the distribution of faunal and paleobotanical remains. Spatial technology, and especially GIS, are particularly effective in correlating the spatial distribution of evidence with analytical studies of each individual find.

It should be stressed that, although many computer applications for the recording and processing of excavation data are generically called `GIS', a true GIS is distinct from a cartography or mapping application, not only because it allows the integration of spatial and non-spatial data, but first and foremost because it processes, manipulates and visualizes information differently. …