Tuff fingerprinting and correlations between OGCP cores and outcrops for Pre-Bed I and Beds I/II at Olduvai Gorge, Tanzania

doi:10.1016/j.palaeo.2020.109630

Palaeogeography, Palaeoclimatology, Palaeoecology

Volume 548, 15 June 2020, 109630

https://doi.org/10.1016/j.palaeo.2020.109630 Get rights and content

Highlights

•
Fingerprints of tuffs in Olduvai paleolake cores link them to outcrop exposures.
•
Olduvai Bed I marker tuffs identified in cores using mineral compositions
•
New tuffs unknown from outcrop exposures identified in Olduvai Bed II in core
•
New rhyolitic tuffs of Ngorongoro Formation recorded in lower core

Abstract

Sediment cores retrieved from the Pleistocene Olduvai Basin by the Olduvai Gorge Coring Project (OGCP) provide a high resolution record of tuffs and other volcaniclastic deposits, together with a lacustrine sedimentary record full of paleoenvironmental indicators. Correlating tuffs between the cores and outcrops at Olduvai, where these tuffs are identified at paleoanthropologically important sites, is critical for applying the new paleoenvironmental data to the conditions under which hominins lived. Tuffs and other volcaniclastic deposits from three cores were analyzed for mineral assemblages and glass and mineral major element compositions (feldspar, augite, hornblende, titanomagnetite, and glass where possible) to compare to published geochemical fingerprint data, based on marker tuffs from outcrop equivalents at Olduvai Gorge. In combination with stratigraphic position, these mineralogical and geochemical data were used to correlate between the cores and outcrops, providing direct temporal tie-lines between the cores and sites of paleoanthropological interest. Direct correlations are most certain for Olduvai Bed I, where all major tuff markers from outcrop are identified for one or more of the three core sites, and for the upper part of the underlying Ngorongoro Formation, which includes the Coarse Feldspar Crystal Tuff (CFCT) and Naabi ignimbrites exposed in the oldest Pleistocene exposures of the Western Gorge. Also characterized were the mineral and glass compositions of tuffs and ignimbrites pre-dating the oldest exposed outcrop units, extending our record of explosive events from the Ngorongoro Volcano. While no specific correlations can be confirmed between individual Bed II tuffs in the cores and in outcrops, correlations are possible between the cores themselves (using newly identified tuff compositions), and some potential correlations (non-unique, based on individual mineral phases) between core and outcrop can be used in conjunction with other stratigraphic tools to help constrain the intervals in question.

Introduction

The palaeoanthropologically-important Pleistocene strata of Olduvai Gorge, Tanzania contain abundant pyroclastic material derived from the adjacent Ngorongoro Volcanic Highlands, preserved in primary and reworked tuffs and as a component of various other volcaniclastic deposits (e.g., diamictites, conglomerates, sandstones derived from volcanic materials). A tephrostratigraphic framework based on these volcanic units helps correlate between different sites within the Olduvai Basin. Recent cores recovered by the Olduvai Gorge Coring Project (OGCP) from the Olduvai paleolake depocenter (Fig. 1) contain many of the same units, hosting a high-resolution paleoclimatic record. This study aims to characterize the tuffs and volcaniclastic sandstones of the OGCP cores in terms of their stratigraphic position, mineral assemblages, and mineral and glass compositions, and compare these against the mineralogical and geochemical “fingerprints” previously established for the Olduvai tuffs in outcrop. This work is therefore essential to establish correlations between the cores and the Olduvai outcrop stratigraphy, thus relating the paleoclimatic record derived from the cores to the classic Olduvai hominin sites. These sites document changes in hominin species (Homo habilis, Homo erectus, Paranthopus boisei, and Homo sapiens) and stone tool technologies (Oldowan, Acheulean, and Middle Stone Age), together with a record of the paleoecological context (e.g. Leakey, 1951, Leakey, 1965, Leakey, 1966, Leakey, 1971; Leakey et al., 1964). This study also documents the changes in volcanic compositions over time, which tracks the evolution of the Ngorongoro Volcanic Highlands through the Pleistocene.

Section snippets

Background

Tuffaceous units are used at Olduvai as chronostratigraphic time-planes throughout the Olduvai Basin, in some cases marking the boundaries between various Beds of the Olduvai Beds (e.g. Tuff IF at the boundary between Beds I and II). Such correlations have been based on physical mapping (e.g. Hay, 1976), and more recently based on mineral and glass compositional “fingerprints” (e.g. McHenry, 2005, McHenry, 2012; McHenry et al., 2016; Habermann et al., 2016; McHenry and Stanistreet, 2018). Fig. 2

Core sampling

Vertical boreholes OGCP-1A, 2A, and 3A were drilled at three different sites in the Olduvai Basin (Fig. 1) targeting the purported depocenters for the paleolake during different stratigraphic intervals, and a 23° inclined borehole was additionally drilled (3B) for paleomagnetic analysis. Cores were retrieved in 3 m intervals (6 cm core diameters), and the core segments were then cut into ~150 cm increments. For example, for the 15th interval collected for core 2A, the core would be broken down

Results

The results comprise major and minor element compositions for all major phenocrysts present within each sample, where possible from 15 to 17 grains or shards of each. Minor or non-target phenocrysts may be represented by a smaller number of grains analyzed, or a simple observation of presence/absence (thus the EPMA analysis was used only to confirm the identification of the phase). Table 1 tabulates the presence and qualitative relative abundances of each phase. Notes on the mineral

Proposed correlations

The EPMA results for the target phases analyzed for each sample were plotted against a database of published and unpublished analyses of tuffs from Olduvai outcrops (McHenry, 2005, McHenry, 2012; McHenry et al., 2016, Habermann et al., 2016); McHenry and Stanistreet, 2018), and against other tuffs from the same general interval within the cores, to look for similarities. Such similarities, or distinctive properties, were noted for each phase in each sample. Correlations are proposed for when:

Discussion

Fig. 5 shows the proposed tephra correlations between the three cores and a composite outcrop section, with the major tuffs indicated. Many tuffs recognized in outcrop do not appear to have equivalents in any of the cores.

While the Bed I tuff record from the OGCP cores match the outcrop record well (with all major marker beds identified in at least one core, and some marker beds identified in all cores), the correspondence between the cores and Bed II was disappointing. None of the major Bed II

Conclusions

In this study we identified all of the major Upper Bed I tuffs in one or more of the three Cores 1A, 2A, 3A, using mineral assemblages and compositions, allowing for direct correlations between the paleoenvironmental records in the cores and outcrops in the Olduvai Basin. No specific correlations can be made between Bed II tuffs in the cores and outcrops based on mineral assemblage and composition alone, due in part to the poor preservation of these tuffs in lacustrine sediments. New tuff

Declaration of competing interest

The authors avow that there are no conflicts of interest associated with this manuscript.

Acknowledgements

The authors would like to thank the Stone Age Institute for funding the Olduvai Gorge Coring Project (OGCP), with support from the Kaman Foundation, the Gordon and Ann Getty Foundation, the John Templeton Foundation, the Fred Maytag Foundation, Henry and Glenda Corning, and Kay and Frank Woods. Additional funding for the tephrostratigraphic work and core-to-outcrop correlations came from the National Science Foundation (BCS grant #1623884 to Njau and McHenry). We thank the Tanzanian Commission

References (31)

R.J. Blumenschine et al.
Environments and hominin activities across the FLK Peninsula during Zinjanthropus times (1.84 Ma), Olduvai Gorge, Tanzania
J. Hum. Evol.
(2012)
A.L. Deino
⁴⁰Ar/³⁹Ar dating of Bed I, Olduvai Gorge, Tanzania, and the chronology of early Pleistocene climate change
J. Hum. Evol.
(2012)
J.M. Habermann et al.
Discrimination, correlation, and provenance of Bed I tephrostratigraphic markers, Olduvai Gorge, Tanzania, based on multivariate analyses of phenocryst compositions
Sediment. Geol.
(2016)
L.J. McHenry
Element mobility during zeolitic and argillic alteration of volcanic ash in a closed-basin lacustrine environment: Case study Olduvai Gorge, Tanzania
Chem. Geol.
(2009)
L.J. McHenry
A revised stratigraphic framework for Olduvai Gorge Bed I based on tuff geochemistry
J. Hum. Evol.
(2012)
L.J. McHenry et al.
Tephrochronology of Bed II, Olduvai Gorge, Tanzania, and placement of the Oldowan–Acheulean transition
J. Hum. Evol.
(2018)
L.J. McHenry et al.
Compositional and textural correlations between Olduvai Gorge Bed I tephra and volcanic sources in the Ngorongoro volcanic highlands, Tanzania
Quat. Int.
(2008)
L.J. McHenry et al.
Geochemical “fingerprints” for Olduvai Gorge Bed II tuffs and implications for the Oldowan–Acheulean transition
Quat. Res.
(2016)
I.G. Stanistreet et al.
Bed II Sequence Stratigraphic context of EF-HR and HWK EE archaeological sites, and the Oldowan/Acheulean succession at Olduvai Gorge, Tanzania
J. Hum. Evol.
(2018)
H. Stollhofen et al.
Plio-Pleistocene syn-sedimentary fault compartments underpin lake margin paleoenvironmental mosaic, Olduvai Gorge, Tanzania

H. Stollhofen et al.

Fingerprinting facies of the Tuff IF marker, catastrophe for early hominin palaeoecology, Olduvai Gorge, Tanzania

Palaeogeography, Palaeoclimatology, Palaeoecology

(2008)

G.M. Ashley et al.

Sedimentation patterns in a Plio-Pleistocene volcaniclastic rift-margin basin Olduvai Gorge, Tanzania

C.J. Campisano

The Hominin Sites and Paleolakes Drilling Project: high-resolution paleoclimate records from the East African Rift System and their implications for understanding the environmental context of hominin evolution

A. Cohen

The Hominin Sites and Paleolakes Drilling Project: inferring the environmental context of human evolution from eastern African rift lake deposits

Sci. Drill.

(2016)

S.M. Greenwood

Mineralogy, Petrology, and Geochemistry of Pleistocene Volcanics at Embagai Caldera and Natron Basin, Tanzania: Potential Constraints on the Stratigraphy of Olduvai Gorge

(2014)

Cited by (15)

Chronostratigraphy and age modeling of Pleistocene drill cores from the Olduvai Basin, Tanzania (Olduvai Gorge Coring Project)
2021, Palaeogeography, Palaeoclimatology, Palaeoecology
Citation Excerpt :
Sample weighted-mean ages after application of the outlier-deletion protocol described above are given in Table S9. Several of the tuffs and the basalts dated directly from core material have also been correlated to outcrop (Table S8) based on tephra geochemistry (McHenry et al., 2020) or unique lithology combined with stratigraphic position (Bed I Basalt). Seven of these units have both new and published 40Ar/39Ar ages from outcrop that supplement core dating results (Table S9).
The Olduvai Gorge Coring Project drilled a total of 611.72 m of core (575.48 m recovered) of mostly fluvio-lacustrine and fan-delta volcaniclastic Pleistocene strata at three sites in the Olduvai Basin, Tanzania, in 2014. We have developed a chronostratigraphic framework for three of the cores based on ⁴⁰Ar/³⁹Ar dating of core and outcrop volcanic and volcaniclastic units, core paleomagnetic stratigraphy, and tephrochemical correlation between cores and from core to outcrop. This framework is then used to constrain Bayesian stratigraphic age models which permit age estimates for desired core levels with realistic confidence intervals. The age models reveal that the deepest core level reached at 245 mbs is ~2.24 Ma, ~210 kyr older than the oldest strata exposed at Olduvai Gorge. Strata net accretion rates in this early phase of basin history were relatively rapid (57–69 cm/kyr), but decreased within ~250 kyr to ~15 cm/kyr in Lower Bed I. Rates rebounded partially in Upper Bed I, but subsequently declined to <10 cm/kyr by Middle to Upper Pleistocene. The age models also provide new estimates for the basal contacts of upper Olduvai Gorge stratigraphic units that have been previously difficult to calibrate: Bed III at 1.14 ± 0.05 (95% confidence interval), Bed IV at 0.93 ± 0.08, Masek at 0.82 ± 0.06, and Ndutu at 0.50 ± 0.04 Ma. Finally, based on recently acquired seismic imaging identifying basement another 135 m beneath the bottom of the deepest core, extrapolation of net accretion rates suggests that sedimentation began at this site in the Olduvai Basin at ~2.5 Ma.
New excavations in the MNK Skull site, and the last appearance of the Oldowan and Homo habilis at Olduvai Gorge, Tanzania
2021, Journal of Anthropological Archaeology
MNK Skull is one of the most significant archaeological sites in Olduvai Gorge, particularly due to the previous discovery of human fossils referred to in the paper where the Homo habilis taxon was originally defined. An important archaeological assemblage is contained in the same horizon as the hominin fossils, constituting the last evidence of both Homo habilis remains and handaxe-free tool kits in the Olduvai Gorge sequence. Our excavations at the site are the first to be conducted since the original work in the 1960s, and sought to refine the archaeological context wherein the Homo habilis remains were discovered. Chronostratigraphic results place the MNK Skull sequence in Middle Bed II prior to deposition of Tuff IIB. The assemblage was deposited near the shoreline, as Palaeolake Olduvai withdrew into the basinal depocentre, and fossils and stone tools were subjected to significant post-depositional processes. The assemblage was affected by mudflow deposits that buried and preserved the assemblage but also entrained surficial bone and lithic elements into the flow. Rather than an occupation site as originally interpreted, the assemblage is better understood as a background deposit, possibly accumulated on an unconformity surface over a long period of time. The stone tool assemblage is typical of the Oldowan, with no technological elements announcing the appearance of the Acheulean, which is well attested to across the Olduvai sequence in post-Tuff IIB times. Our results highlight that, with an approximate age of circa 1.67 Ma, MNK Skull stands as a key site to understand the late Oldowan and the disappearance of Homo habilis in East Africa.
Olduvai's oldest Oldowan
2021, Journal of Human Evolution
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.
The Olduvai Gorge Coring Project: Drilling high resolution palaeoclimatic and palaeoenvironmental archives to constrain hominin evolution
2021, Palaeogeography, Palaeoclimatology, Palaeoecology
Volcano-sedimentary cores recovered from Pleistocene Palaeolake Olduvai by the Olduvai Gorge Coring Project (OGCP) provide a high-resolution record for reconstructing climatic and environmental contexts of hominin evolution. Approximately 612 m were recovered from four cores from three drill sites across the basin depocentre through scientific drilling, and these have yielded unprecedented data that substantially extend the Olduvai record both temporally and spatially. Results from multiproxy analyses based on sedimentological, mineralogical, isotopic, and geochemical measurements, along with analysis of organic matter and microfossils, backed up by new ⁴⁰Ar/³⁹Ar dating, palaeomagnetic reversal analysis and tuff fingerprinting, provide detailed palaeoecological and palaeoenvironmental data, and stratigraphic resolution adequate to provide context for the palaeoanthropological records from outcrops. In this Special Issue we present the first phase reporting of the core results, which establishes a new stratigraphic and palaeogeographic framework upon which palaeoenvironmental and palaeoanthropological data can be contextualized. The cores revealed thick intercalated lakebeds below the Bed I Basalt and the even older Naabi Ignimbrite, more than doubling the thickness of the known Olduvai stratigraphy. Seismic studies suggest that the sedimentary sequence contained in the lake's depositional sump extends even further back in time, perhaps back to ~2.5 Ma, nearly 500 kyr older than the deepest strata exposed in outcrop. From stratal characteristics, we deduce that the palaeolake was deeper, more permanent, and longer-lived within the lake sump than previously thought, at least until the Masek Beds (before ~0.82 Ma), although frequent phases of total or nearly total lake emptying/drying were detected, and the lake depocentre was at times filled by volcaniclastic fan progradation during periods of intense volcanic activity in the neighboring Ngorongoro Volcanic Highlands. Furthermore, biogeochemical evidence supports our interpretation of the lake history, and documents abrupt transitions in terrestrial vegetation, aquatic biota, and lake dynamics on a Milankovitch and sub-Milankovitch scale, which potentially exerted complex influences on hominin-exploited palaeolandscapes. This work revises the longstanding view of the basin history and transforms the scientific debate about the environmental conditions under which hominins evolved in the Olduvai Basin.
Changing depocentre environments of Palaeolake Olduvai and carbonates as marker horizons for hiatuses and lake-level extremes
2020, Palaeogeography, Palaeoclimatology, Palaeoecology
Citation Excerpt :
Primary volcanic products were also commonly deposited at various times within the basin, including basaltic lava flows and pyroclastic flow, surge and airfall tuffs (Stollhofen et al., 2008; Blumenschine et al., 2012b, 2012a). Particular Tuff Markers have been distinguished (Fig. 2), including the Naabi Ignimbrite, Coarse Feldspar Crystal Tuff (CFCT), Tuff IA, all derived from Ngorongoro Volcano, and Tuffs IB, IC, ID, IE, Nge'ju and IF, derived from Olmoti Volcano (McHenry, 2012; McHenry et al., 2020a). Tephrostratigraphic markers and unconformity surfaces were preferentially used by Hay (1976) to subdivide the Plio-Pleistocene succession of Olduvai Gorge into seven Formations (Beds I–IV, Masek Beds, Ndutu Beds, and Naisiusiu Beds), overlying a crystal-rich tuff termed the Naabi Ignimbrite (Fig. 2).
Primary carbonate and marl layers and limestone nodular horizons were intersected in OGCP Boreholes 1A, 2A, 3A, 3B, drilled into the depocentre of Palaeolake Olduvai. The various carbonate types were analysed, employing petrographic (including cathodo-luminescence), stable isotope, and sequence stratigraphic techniques, and recorded important information concerning lake evolution.
Primary carbonate and marl layers are preserved at the top of lake deepening cycles (lake-parasequences), marking maximum flooding, followed by lake withdrawal, and then fluvial erosion, leading to the next depositional cycle. Presence of pelagites and a meromict lake favouring anoxic conditions, show it to be not a playa lake, but a deeper fault-bound lake, akin to present day Lake Eyasi. Carbonate and marl layers were exclusively deposited when claystone facies show a geochemical magnesium anomaly associated with a phase of basaltic volcanism affecting the basin, marked by mafic tuffs and basaltic lava flows. Calcium was partitioned into evaporites, such as gypsum/anhydrite, together with the nodular horizons and carbonate soil profiles in the sequence. The resulting high Mg²⁺/Ca²⁺ ratio in the lake waters promoted dolomite deposition and replacement, and the formation of aragonite.
The nodular horizons yield rainfall isotope values and mark times when the lake was empty and the lake-bed fell under meteoric conditions, precipitating limestone nodules just below the sediment surface under pedogenic vadose, groundwater interface, and groundwater phreatic conditions. The nodular limestone horizons sit below erosional/hiatal surfaces and show vadose micritic and fibrous types to phreatic sparry and other accretionary type textures. Enterolithic to chickenwire textured nodules indicate the pseudomorphing of anhydrite nodules that form beneath salt marshes.
Thus, the two carbonate types, primary layers and nodular horizons, provided depth gauges, respectively for the extremes of lake expansion and emptying/drying out. In the case of nodular horizons, the maturity of the carbonate profile gives an indication of the magnitude of the hiatal time gap represented, ~ 1 kyr to 2 kyr for a single horizon to ~ 3 kyr to 9 kyr for more mature soil profiles. Emptying/drying out episodes of Palaeolake Olduvai were not uncommon, and often short-lived. A spectrum of hiatal disconformities in the Olduvai Basin vary from multimillennial hiatuses to mega-disconformities of ~ 40 kyr and ~ 75 kyr.
Palaeovegetation changes recorded in Palaeolake Olduvai OGCP Core 2A (2.09–2.12 Ma) Naibor Soit Formation Olduvai Gorge, Tanzania
2020, Palaeogeography, Palaeoclimatology, Palaeoecology
For five decades Olduvai Gorge has been a key site to reconstruct and understand the relationship between environmental and landscape conditions and use of affordances by early African hominin populations. Following the first Olduvai Gorge Coring project (OGCP) during 2014, a multiproxy microbiological analysis, which includes phytoliths, pollen, diatoms, sponge spicules and chrysophyte cysts, was undertaken on samples collected from various borehole cores. The aim of the study is to better understand palaeoenvironmental and palaeovegetation conditions and changes through time and their relationship to hominin presence and evolution. This study details the first palaeobotanical and palaeoenvironmental study of Borehole 2A at Olduvai Gorge. It represents the as yet oldest known sedimentary sequence in the Olduvai Basin for a portion of the pre-Bed I Naibor Soit Formation; this is a unit that is not accessible in any natural exposures in Olduvai Gorge, and has only recently been encountered by drilling. Here we present the results from a particularly phytolith-prone portion between ~2.09 Ma and 2.12 Ma.
Phytolith results indicate a savannah environment dominated by grasses, where Poaceae were a key component and where the C₃ Pooideae grasses were mostly dominant, alternating with C₄ grasses. Oscillations between grass subfamilies, C₃ Pooideae, C₄ Chloridoideae, and C₄ Panicoideae to lesser degrees, indicate five substantial climatic shifts, varying between more humid and arid conditions. Associated with phytoliths, freshwater indicators such as diatoms, sponge spicules and chrysophyte cysts were also identified, suggesting the presence of wetlands in the lake catchment area. Pollen is extremely rare in the sediments but when present, comprises fungal spores and Poaceae pollen, thus supporting the wetland and grassland reconstructions, respectively. These results offer for the first time, a whole picture of the palaeovegetation and associated palaeoenvironments for this pre-Bed I period. Together with previous results from other areas and chronological periods, they improve our understanding of the evolution and adaptation of early hominins and their close relationship to the surrounding landscape.

View all citing articles on Scopus

View full text

Tuff fingerprinting and correlations between OGCP cores and outcrops for Pre-Bed I and Beds I/II at Olduvai Gorge, Tanzania

Highlights

Abstract

Introduction

Section snippets

Background

Core sampling

Results

Proposed correlations

Discussion

Conclusions

Declaration of competing interest

Acknowledgements

J. Hum. Evol.

J. Hum. Evol.

Sediment. Geol.

Chem. Geol.

J. Hum. Evol.

J. Hum. Evol.

Quat. Int.

Quat. Res.

J. Hum. Evol.

Palaeogeography, Palaeoclimatology, Palaeoecology

Sedimentation patterns in a Plio-Pleistocene volcaniclastic rift-margin basin Olduvai Gorge, Tanzania

The Hominin Sites and Paleolakes Drilling Project: high-resolution paleoclimate records from the East African Rift System and their implications for understanding the environmental context of hominin evolution

The Hominin Sites and Paleolakes Drilling Project: inferring the environmental context of human evolution from eastern African rift lake deposits

Sci. Drill.

Mineralogy, Petrology, and Geochemistry of Pleistocene Volcanics at Embagai Caldera and Natron Basin, Tanzania: Potential Constraints on the Stratigraphy of Olduvai Gorge