January 2005
Paleoenvironments of the Earliest Toolmakers
Jay Quade, University of Arizona, Desert Laboratory, Tucson AZ
I was invited to join the Gona Palaeoanthropology Research Project (GPRP) as geologist in 1999, when the project was just starting. The project consists of a multidisciplinary team of scientists currently undertaking field and laboratory research on the earliest hominid fossils and archaeological materials discovered from Gona, Afar, Ethiopia (Fig. 1). When I first arrived next to nothing was known about the geology at Gona, which encompasses about 120 km2 of deeply eroded badlands. After five years of work we now know that the period 6 to 0.5 Ma is nearly continuously represented by deposits and fossils. Hominid remains are proving quite common: parts of over twenty hominids have been found to date, including a number of remains from the last two million years.

Figure 1. Geology and major localities in the Gona project area. The
Busidima Formation remains unmapped outside the Gona project area.
This month I published a paper summarizing what we now know about the geology of Gona, with particular emphasis on the geologic history of the last 2.6 million years. The deposits spanning 2.6 to <0.5 ma belong to the newly designated Busidima Formation. The Busidima Formation was deposited by the ancestors to the modern Awash River that flows through the area today. A combination of tuff dating and tephrostratigraphy provides the age control for the deposits.
Our results show that the study area experienced repeated deep dissection and backfilling by the Awash River, starting between 2.92 and ~2.7 Ma and continuing through the top of the record at <0.5 Ma. Each backfilling cycle is 10-20 m in thickness, and fines upward from well-rounded volcanic conglomerates at the base to rhizolith-rich sand, bedded silt, and capping paleosols at the top. The ancestral Awash during this period was dominantly meandering, as it is today, and paleoflow direction, although dispersed by the meandering pattern, was toward the north-northeast. Smaller channels tributary to the axial Awash system are also extensively exposed in the Busidima Formation. Compared to the axial-system conglomerates, the tributary channels transported finer, less mature volcanic clasts mixed with abundant carbonate nodules reworked from adjacent badlands. These channels and interbedded mature, often thick (>2m) paleosols, characterize much of the upper part of the Busidima Formation, as the axial system shifted south, out of the study area.
Stone artifacts (Oldowan) at the oldest archeological sites (2.6 to 2.0) are only associated with the axial Awash system, in the bedded silts or capping paleosols of the fining upwards sequences (Fig. 2). Raw material for the implements were the rounded cobbles from the channels, but manufacture and use of the tools was always away from the channel bars, on the nearby sandy banks and silt-dominated floodplains. Archeological sites higher in the record (<1.7 Ma) and containing Acheulian artifacts occur in similar axial river contexts, as well as along tributary channels further removed from artifact raw material sources.

Figure 2. Fining upward sequence and summary of the position of Oldowan
implements with respect to Awash floodplain topography and ecosystems
of the modern Awash.
Mature paleosols in the Hadar and Busidima Formations are mostly pale to dark-brown vertisols typified by Ca-smectite, abundant clay slickensides, pseudo-anticlinal and vertical fracturing, and carbonate nodules. These types of soils are common in the region today, demonstrating that the paleoclimate over the past 3.4 Ma has always been semi-arid and strongly seasonal.
Carbon isotopic results from pedogenic carbonates in the vertisols allow reconstruction of the proportion of C3 plants (trees and shrubs) to C4 plants (grasses) through time (Fig. 3). δ13C results from the Hadar Formation prior to 2.9 Ma range from -9.3 to -4.1 ?, indicating a dominantly forested environment but with locally substantial (40%) grasses on the Awash floodplain. δ13C values from soil carbonate in the upper Kada Hadar Member (2.7-1.5 Ma) increase (-6.5 to -2.7?) in floodplain paleosols, indicating 39-66% grass. Vertisols of the Busidima Member (<1.5 Ma) formed on gently sloping alluvial fans adjacent to the Awash floodplain display even more positive δ13C values (up to -1.8?), showing that extensive grasslands grew on the margins of the active Awash floodplain, as they do today.

Figure 3. δ13C (PDB) values of soil carbonate in the Hadar and
usidima Formations (sub-divisions shown at right). Percent grassland calculated
assuming δ13C values of -12 and +2? of soil carbonate associated
with end-member C3 and C4 vegetation, respectively.
Quade, J., Levin, N., Semaw, S., Simpson, S., Rogers, M., Stout, D., 2004, Paleoenvironments of the earliest toolmakers. Geological Society of America Bulletin v. 116, p. 1529-1544.
For more information contact:
Jay Quade (jquade@geo.arizona.edu)