Age-depth model for uppermost Ndutu Beds constrains Middle Stone Age technology and climate-induced paleoenvironmental changes at Olduvai Gorge (Tanzania).

Olduvai Gorge in northern Tanzania is part of a globally important archeological and paleoanthropological World Heritage Site location critical to our understanding of modern human evolution. The Ndutu Beds in the upper part of the geological sequence at Olduvai Gorge represent the oldest unit to yield modern Homo sapiens skeletal material and Middle Stone Age technology. However, the timing of the deposition of the Ndutu Beds is poorly constrained at present, which limits our understanding of the paleoenvironments critical for contextualizing H. sapiens and related technologies in the Olduvai Basin. Using a suite of 15 luminescence ages of sedimentary core samples, combined with Bayesian statistics, this study provides a new higher-resolution age-depth model for the deposition of the uppermost Upper Ndutu and Naisiuiu Beds cored by the Olduvai Gorge Coring Project. The luminescence and modeled ages are presented as ±1 σ uncertainties. The Ndutu Beds intersected by the Olduvai Gorge Coring Project cores are dated to between 117.1 ± 17.9 and 45.3 ± 4.2 ka (between 125.9 ± 26.5 and 45.8 ± 8.2 ka modeled ages), while a probable overlying layer of Naisiusiu Beds dates to 23.7 ± 10.9 to 12.1 ± 1.7 ka (25.7 ± 18.9 ka and 12.0 ± 3.4 ka modeled age). Time-averaged accretion rates are derived during this time: (1) initially low rates (<5 cm ka-1) from the bottom of the core at 117.1 ± 17.9 ka up to 95.3 ± 11.1 ka (125.9 ± 26.5 to 95.5 ± 23.3 ka modeled ages); (2) the middle section spanning between 95.3 ± 11.1 and 62.7 ± 5.7 ka (95.5 ± 23.3 to 61.9 ± 10.4 ka modeled ages) with mean rates above 15 cm ka-1; and (3) the last 62.7 ± 5.7 ka (61.9 ± 10.4 ka modeled age) where the accretion rate reduces to below 5 cm ka-1. This reduction can be explained by the evolution of the gorge system that was likely driven by subsidence of the Olbalbal depression and changes in climate, particularly precipitation and resulting lake and base level changes. Older Upper Ndutu and Lower Ndutu Beds are contained within proto-gorges within the modern gorge system.

Olduvai's oldest Oldowan.

Previously, Olduvai Bed I excavations revealed Oldowan assemblages <1.85 Ma, mainly in the eastern gorge. New western gorge excavations locate a much older ∼2.0 Ma assemblage between the Coarse Feldspar Crystal Tuff (∼2.015 Ma) and Tuff IA (∼1.98 Ma) of Lower Bed I, predating the oldest eastern gorge DK assemblage below Tuff IB by ∼150 kyr. We characterize this newly discovered fossil and artifact assemblage, adding information on landscape and hominin resource use during the ∼2.3-2.0 Ma period, scarce in Oldowan sites. Assemblage lithics and bones, lithofacies boundaries, and phytolith samples were surveyed and mapped. Sedimentological facies analysis, tephrostratigraphic and sequence stratigraphic principles were applied to reconstruct paleoenvironments and sedimentary processes of sandy claystone (lake), sandstone (fluvial), and sandy diamictite (debris flow) as principal lithofacies. Artifacts, sized, weighed, categorized, were examined for petrography, retouch, and flake scar size. Taxonomic classifications and taphonomic descriptions of faunal remains were made, and phytoliths were categorized based on reference collections. Lithics are dominantly quartzite, mainly debitage and less frequently simple cores, retouched pieces, and percussors. Well-rounded spheroids and retouched flakes are rare. Identifiable taxa, Ceratotherium cf. simum (white rhinoceros) and Equus cf. oldowayensis (extinct zebra), accord with nearby open savanna grasslands, inferred from C3 grass, mixed and/or alternating with C4 grass-dominated phytolith assemblages. Palms, sedges, and dicots were also identified from phytoliths. Diatoms and sponge spicules imply nearby freshwater. The assemblage accumulated at the toe of a Ngorongoro Volcano-sourced fan-delta apron of stacked debris flows, fluvials, and tuffs, preserving fossil tree stumps and wooded grassland phytoliths farther upfan. It formed after the climax of Ngorongoro volcanic activity during a Paleolake Olduvai lowstand and was then buried and preserved by lacustrine clays, marking the first of two lake transgressions, signifying wetter climates. Orbital precessional lake cycles were superposed upon multimillennial (∼4.9 kyr) lake fluctuations.

Tephrochronology of Bed II, Olduvai Gorge, Tanzania, and placement of the Oldowan-Acheulean transition.

Tuffaceous marker beds, derived from volcanic products from the Ngorongoro Volcanic Highlands, help define a stratigraphic framework for the world-renowned fossil and stone tool record exposed at Olduvai Gorge, Tanzania. However, previous efforts to constrain this tuff record, especially for Olduvai Bed II, have been limited because of erosion, contamination, reworking, and the alteration of volcanic glass under saline-alkaline conditions. This paper applies previously defined geochemical and mineralogical "fingerprints" for several major Bed II marker tuffs, based on glass (where available) and phenocrysts more resistant to alteration (feldspar, hornblende, augite, and titanomagnetite), to tuffs from stratigraphic sections in the Olduvai Junction Area, including previously and recently excavated Acheulean and Oldowan sites (HWK EE (Locality (Loc) 42), EF-HR (Loc 12a), FLK (Loc 45), and MNK (Loc 88)). The Middle Bed II Bird Print Tuff (BPT) is found to be more compositionally variable than previously reported but is still valuable as a stratigraphic marker over short distances. The confirmation of blocks of Tuff IID in conglomerate helps constrain Upper Bed II stratigraphy at sites where in-situ tuffs are absent. This paper also compiles the results of published geochronological research, providing stratigraphic context and updating previously reported dates using a consistent 40Ar/39Ar reference standard age. The results of this work support the following paleoanthropologically relevant conclusions: 1) the early Acheulean site EF-HR (Loc 12a) is situated above the level of Hay's Tuff IIC, and thus sits in Upper rather than Middle Bed II, (2) the HWK EE (Loc 42) Oldowan site is constrained between Tuff IIA and Tuff IIB, just above the boundary between Lower and Middle Bed II, and 3) the Acheulean site at FLK W most likely lies within the Middle Augitic Sandstone, above Tuff IIB, similar to the placements by Leakey and Hay for the earliest Acheulean at Olduvai.

Bed II Sequence Stratigraphic context of EF-HR and HWK EE archaeological sites, and the Oldowan/Acheulean succession at Olduvai Gorge, Tanzania.

Archaeological excavations at EF-HR and HWK EE allow reassessment of Bed II stratigraphy within the Junction Area and eastern Olduvai Gorge. Application of Sequence Stratigraphic methods provides a time-stratigraphic framework enabling correlation of sedimentary units across facies boundaries, applicable even in those areas where conventional timelines, such as tephrostratigraphic markers, are absent, eroded, or reworked. Sequence Stratigraphically, Bed II subdivides into five major Sequences 1 to 5, all floored by major disconformities that incise deeply into the underlying succession, proving that simple "layer cake" stratigraphy is inappropriate. Previous establishment of the Lemuta Member has invalidated the use of Tuff IIA as the boundary between Lower and Middle Bed II, now redefined at the disconformity between Sequences 2 and 3, a lithostratigraphic contact underlying the succession containing the Lower, Middle, and Upper Augitic Sandstones. HWK EE site records Oldowan technology in the Lower Augitic Sandstone at the base of Sequence 3, within Middle Bed II. We suggest placement of recently reported Acheulean levels at FLK W within the Middle Augitic Sandstone, thus emphasizing that handaxes are yet to be found in earlier stratigraphic units of the Olduvai sequence. This would place a boundary between the Oldowan and Acheulean technologies at Olduvai in the Tuff IIB zone or earliest Middle Augitic Sandstone. A major disconformity between Sequences 3 and 4 at and near EF-HR cuts through the level of Tuff IIC, placing the main Acheulean EF-HR assemblage at the base of Sequence 4, within Upper rather than Middle Bed II. Sequence stratigraphic methods also yield a more highly resolved Bed II stratigraphic framework. Backwall and sidewall surveying of archaeological trenches at EF-HR and HWK EE permits definition of "Lake-parasequences" nested within the major Sequences that record downcutting of disconformities associated with lake regression, then sedimentation associated with lake transgression, capped finally by another erosional disconformity or hiatal paraconformity caused by the next lake withdrawal. On a relative time-scale rather than a vertical metre scale, the resulting Wheeler diagram framework provides a basis for recognizing time-equivalent depositional episodes and the position of time gaps at various scales. Relative timing of archaeological assemblage levels can then be differentiated at a millennial scale within this framework.

The contexts and early Acheulean archaeology of the EF-HR paleo-landscape (Olduvai Gorge, Tanzania).

Renewed fieldwork at the early Acheulean site of EF-HR (Olduvai Gorge, Tanzania) has included detailed stratigraphic studies of the sequence, extended excavations in the main site, and has placed eleven additional trenches within an area of nearly 1 km2, to sample the same stratigraphic interval as in the main trench across the broader paleo-landscape. Our new stratigraphic work suggests that EF-HR is positioned higher in the Bed II sequence than previously proposed, which has implications for the age of the site and its stratigraphic correlation to other Olduvai Middle Bed II sites. Geological research shows that the main EF-HR site was situated at the deepest part of an incised valley formed through river erosion. Archaeological excavations at the main site and nearby trenches have unearthed a large new assemblage, with more than 3000 fossils and artefacts, including a hundred handaxes in stratigraphic position. In addition, our test-trenching approach has detected conspicuous differences in the density of artefacts across the landscape, with a large cluster of archaeological material in and around the main trench, and less intense human activity at the same level in the more distant satellite trenches. All of these aspects are discussed in this paper in the light of site formation processes, behavioral contexts, and their implications for our understanding of the early Acheulean at Olduvai Gorge.

New excavations at the HWK EE site: Archaeology, paleoenvironment and site formation processes during late Oldowan times at Olduvai Gorge, Tanzania.

This paper reports the results of renewed fieldwork at the HWK EE site (Olduvai Gorge, Tanzania). HWK EE is positioned across the boundary between Lower and Middle Bed II, a crucial interval for studying the emergence of the Acheulean at Olduvai Gorge. Our excavations at HWK EE have produced one of the largest collections of fossils and artefacts from any Oldowan site, distributed across several archaeological units and a large excavation surface in four separate trenches that can be stratigraphically correlated. Here we present the main stratigraphic and archaeological units and discuss site formation processes. Results show a great density of fossils and stone tools vertically through two stratigraphic intervals (Lemuta and Lower Augitic Sandstone) and laterally across an area of around 300 m2, and highlight the confluence of biotic and abiotic agents in the formation of the assemblage. The large size and diversity of the assemblage, as well as its good preservation, qualify HWK EE as a reference site for the study of the late Oldowan at Olduvai Gorge and elsewhere in Africa. In addition, the description of the stratigraphic and archaeological sequence of HWK EE presented in this paper constitutes the foundation for further studies on hominin behavior and paleoecology in Lower and Middle Bed II.

In situ ∼2.0 Ma trees discovered as fossil rooted stumps, lowermost Bed I, Olduvai Gorge, Tanzania.

The discovery of fossil rooted tree stumps in lowermost Lower Bed I from the western Olduvai Basin, Tanzania, age-bracketed by the Naabi Ignimbrite (2.038 ± 0.005 Ma) and Tuff IA (1.88 ± 0.05 Ma), provides the first direct, in situ, and to date oldest evidence of living trees at Olduvai Gorge. The tree relicts occur in an interval dominated by low-viscosity mass flow and braided fluvial sediments, deposited at the toe of a largely Ngorongoro Volcano-sourced volcaniclastic fan apron that comprised a widely spaced network of ephemeral braided streams draining northward into the Olduvai Basin. Preservation of the trees occurred through their engulfment by mass flows, post-mortem mold formation resulting from differential decay of woody tissues, and subsequent fluvially-related sediment infill, calcite precipitation, and cast formation. Rhizolith preservation was triggered by the interaction of root-induced organic and inorganic processes to form rhizocretionary calcareous root casts. Phytolith analyses were carried out to complete the paleoenvironmental reconstruction. They imply a pronounced seasonality and indicate a wooded landscape with grasses, shrubs, and sedges growing nearby, comparable to the low, open riverine woodland (unit 4c) along the Garusi River and tributaries in the Laetoli area. Among the tree stump cluster were found outsized lithic clasts and those consisting of quartzite were identified as Oldowan stone tool artifacts. In the context of hominin activity, the identification of wooded grassland in association with nearby freshwater drainages and Oldowan artifacts significantly extends our paleoenvironmental purview on the basal parts of Lower Bed I, and highlights the hitherto underrated role of the yet poorly explored western Olduvai Gorge area as a potential ecologically attractive setting and habitat for early hominins.

Fine resolution of early hominin time, Beds I and II, Olduvai Gorge, Tanzania.

Reconstructing paleoenvironments and landscapes within lake-centered, hominin-yielding basinal sequences requires a resolution of time-rock units finer than but complementary to that provided by the present tephrostratigraphy. Although indispensable in providing an absolute time frame at Olduvai, the average 15,000-20,000 year intervals between successive tuff units lack the time resolution to construct a sufficiently contemporary paleolandscape within sedimentary intervals away from the interleaved tuffs. Such control is essential to construct valid paleogeographies in which to contextualize contemporaneous paleoanthropological sites and the traces of hominin land use they contain. Within Beds I and II of the Olduvai Basin a Sequence Stratigraphic analysis has achieved a relative time framework in which time-rock units, "lake-parasequences," each generated by a major advance and withdrawal of the lake system, are recognizable for average periods of about 4000 years duration. Within each of these time slices at least two paleogeographic landscapes are identifiable, reducing the time constraints of an individual landscape reconstruction to a few thousand years. Within the sedimentary succession both highly incised and less incised unconformities are identifiable to provide sequence boundaries. Within each sequence the higher frequency lake-parasequences can be identified by (1) a disconformable base, (2) accretion of sediment during lake transgression and at maximum, (3) a disconformable top caused by lake withdrawal, and (4) a soil profile generated beneath that disconformable land surface. Individual lake-parasequences can be recognized in lake marginal and fan settings, and their imprint can also be seen in the lake setting where, for example, maximum flooding might be marked by a layer of dolomite. Lower Bed II parasequences represent time intervals of <5000 years, while parasequential periods between Tuffs IB and IC in Bed I are <4300 years. Analogous Holocene lake-level changes of the same order in East Africa have a period close to 4200 years. The estimated period is close to that defined by Stadial/Interstadial Dansgaard-Oeschger Events recorded in the Greenland Ice record, which force cycles of period similar to lake-parasequences, both in the Arabian Sea and Lake Malawi. Lake-parasequences not only aid construction of landscapes, they also allow contextualization of individual paleoanthropological occurrences. For instance, a parasequence lies between Leakey's Level 1 and her butchered Deinotherium occurrence at FLK N. However, elephantid and giraffid skeletons associated with stone artifacts at VEK, uncovered by OLAPP excavations, are situated on the same land surface as a possibly butchered rhinocerid at KK. To complement the existing absolute radiometric time framework, relative Sequence Stratigraphic techniques might be applied to any lake-centered, hominin-yielding basinal sequence, not only those found within East Africa. Because they are climatically controlled, and might plausibly be related to globally driven Dansgaard-Oeschger Events, lake-parasequences and their associated sequences might be correlatable between various East African basins in the Plio-Pleistocene in the same way as they presently are for the Holocene.

Plio-Pleistocene synsedimentary fault compartments, foundation for the eastern Olduvai Basin paleoenvironmental mosaic, Tanzania.

Normal faults displacing Upper Bed I and Lower Bed II strata of the Plio-Pleistocene Lake Olduvai were studied on the basis of facies and thickness changes as well as diversion of transport directions across them in order to establish criteria for their synsedimentary activity. Decompacted differential thicknesses across faults were then used to calculate average fault slip rates of 0.05-0.47 mm/yr for the Tuff IE/IF interval (Upper Bed I) and 0.01-0.13 mm/yr for the Tuff IF/IIA section (Lower Bed II). Considering fault recurrence intervals of ~1000 years, fault scarp heights potentially achieved average values of 0.05-0.47 m and a maximum value of 5.4 m during Upper Bed I, which dropped to average values of 0.01-0.13 m and a localized maximum of 0.72 m during Lower Bed II deposition. Synsedimentary faults were of importance to the form and paleoecology of landscapes utilized by early hominins, most traceably and provably Homo habilis as illustrated by the recurrent density and compositional pattern of Oldowan stone artifact assemblage variation across them. Two potential relationship factors are: (1) fault scarp topographies controlled sediment distribution, surface, and subsurface hydrology, and thus vegetation, so that a resulting mosaic of microenvironments and paleoecologies provided a variety of opportunities for omnivorous hominins; and (2) they ensured that the most voluminous and violent pyroclastic flows from the Mt. Olmoti volcano were dammed and conduited away from the Olduvai Basin depocenter, when otherwise a single or set of ignimbrite flows might have filled and devastated the topography that contained the central lake body. In addition, hydraulically active faults may have conduited groundwater, supporting freshwater springs and wetlands and favoring growth of trees.

Carbonate horizons, paleosols, and lake flooding cycles: Beds I and II of Olduvai Gorge, Tanzania.

This study documents the petrology and stable isotope geochemistry of carbonates from six horizons from Beds I and II of Olduvai Gorge, Tanzania. The studied succession, immediately below and above Tuff IF, consists of interbedded waxy and earthy claystones with discrete carbonate horizons and thin sandstones. The succession was deposited in response to repeated flooding and withdrawal of a saline-alkaline lake. The carbonates and their overlying disconformities are important because they help define the surfaces on which hominin activity took place and allow very high-resolution correlation of geographically separated levels of hominin exploitation. The range of different carbonates includes unambiguous land-surface and pedogenic features including calcified rootmat horizons, rhizocretions, and micritic nodules, together with less determinate sparry calcite nodules. Stellate nodules are interpreted as pseudomorphs after sulfate-roses. The carbonate nodules are synsedimentary features, truncated by fluvial and other erosional surfaces. The isotopic composition of the carbonates is variable with δ(18)O ranging from -7.0‰ to -4.3‰, and δ(13)C from -8.5‰ to -1.6‰. A covariant increase in δ(13)C and δ(18)O repeats in each carbonate horizon and in individual nodules (inner to outer layers): it reflects the evolution of synsedimentary groundwaters. At times of low lake level, the carbonates started to precipitate from meteoric waters with low isotopic values and continued to form as lake levels rose and the waters became increasingly saline. Some of the samples have a last-stage cement of strontium rich dolomite, which supports late-stage flooding by the saline-alkaline lake. Previous studies of carbonate horizons from Olduvai have interpreted carbon isotope values in terms of changes in C(3) and C(4) plants that colonized the land surface. This study demonstrates that in some instances the isotope values from carbonates deposited in these lake marginal settings reflect changes in hydrology rather than vegetation.

Landscape distribution of Oldowan stone artifact assemblages across the fault compartments of the eastern Olduvai Lake Basin during early lowermost Bed II times.

The density and composition of Oldowan stone artifact assemblages deposited during the first ca. 20,000 years of lowermost Bed II times show a recurrent pattern of variation across recognized synsedimentary faults that compartmentalized landscapes of the eastern Olduvai Lake Basin. When active, the faults created minor topographic relief. The upthrown fault footwalls accumulated assemblages with relatively high densities of artifacts, including types retaining potential usefulness, particularly volcanic flaked pieces, manuports, pounded pieces, and split cobbles. Values for these assemblage characteristics decline toward the lower-lying hangingwall of the fault compartments, accompanied by an increase in the proportionate weight of artifact assemblages comprising quartzite, particularly flaking shatter and potentially useful detached pieces. Values reverse once again at faults, either on the downthrown, hangingwall side or on the upthrown side. The patterns are stronger for the volcanic components of the artifact assemblages than for the quartzite components, reflecting the additional influence of distance from the local source on quartzite assemblage characteristics reported previously. The landscape distributions of artifact assemblages are consistent with a landscape-fault model in which minor fault-induced topographic relief at times created a mosaic of vegetation environments repeated within each of the three fault compartments of the lake margin and distal alluvial fan. The fault-compartmentalized landscape model is currently supported only by sediment thickness and facies changes across synsedimentary faults, but it provides predictions for spatial variation in the cover abundance of trees, freshwater reservoirs and associated distributions of resources and hazards associated with stone artifact use and discard that can be tested if sample sizes of key paleoenvironmental indicators are increased.

Environments and hominin activities across the FLK Peninsula during Zinjanthropus times (1.84 Ma), Olduvai Gorge, Tanzania.

We establish through 13 excavations the landscape context and nature of hominin activities across the Zinjanthropus land surface from which the Leakeys recovered the FLK 22 and FLK NN 1 paleoanthropological assemblages. The land surface was created by fluvial incision of the eastern margin of paleo-Lake Olduvai following a major lake withdrawal. Erosion was uneven, leaving a peninsula bounded by a river channel, the FLK Fault, and a freshwater wetland. This FLK Peninsula supported groves of trees that attracted hominins and carnivores, and that preserved the dense concentrations of carcass remains and stone tools they left behind, including those at FLK 22. Some carcasses appear to have been acquired at the ecotone of the Peninsula and Wetland, where another dense artifact and bone assemblage accumulated. A lesser topographic high at the edge of a Typha marsh in the Wetland was the site of FLK NN 1 and a scatter of large stone tools used possibly for rootstock processing. Our landscape reconstruction delimits the vegetation mosaic indicated by previous work and provides a topographical explanation for the existence of FLK 22 and FLK NN 1. Both are unexpected if the FLK area was the flat, featureless lake margin terrain typical of lake basins similar to paleo-Olduvai. The results show that the Leakeys' sites were not isolated occupation floors but rather parts of a land surface utilized intensively by hominins. Although commonly considered to have been home bases, their likely high predation risk, evidenced by large carnivore feeding traces and the remains of four hominin individuals, suggests visits to them were brief and limited to feeding. Finally, stratigraphic observations confirm that FLK NN 3 accumulated on an older land surface, refuting the hypothesis that the OH 8 foot found there is the same individual as the OH 35 leg from FLK 22.

Late Pliocene Homo and hominid land use from Western Olduvai Gorge, Tanzania.

Excavation in the previously little-explored western portion of Olduvai Gorge indicates that hominid land use of the eastern paleobasin extended at least episodically to the west. Finds included a dentally complete Homo maxilla (OH 65) with lower face, Oldowan stone artifacts, and butchery-marked bones dated to be between 1.84 and 1.79 million years old. The hominid shows strong affinities to the KNM ER 1470 cranium from Kenya (Homo rudolfensis), a morphotype previously unrecognized at Olduvai. ER 1470 and OH 65 can be accommodated in the H. habilis holotype, casting doubt on H. rudolfensis as a biologically valid taxon.