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

Journal of Human Evolution 63 (2012) 300e308

Contents lists available at SciVerse ScienceDirect

Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol

Fine resolution of early hominin time, Beds I and II, Olduvai Gorge, Tanzaniaq Ian G. Stanistreet Geology and Geophysics, School of Environmental Sciences, University of Liverpool, Liverpool L69 3GP, UK

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 December 2008 Accepted 15 March 2012 Available online 30 May 2012

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 15,000e20,000 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. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Pliocene Pleistocene Sequence stratigraphy Oldowan hominins Incised surface Lake parasequence Dansgaard-Oeschger events

Introduction

q For the Olduvai Landscape Paleoanthropology Project’s special issue of Journal of Human Evolution, "Five decades after Zinjanthropus and Homo habilis: Landscape Paleoanthropology of Plio-Pleistocene Olduvai Gorge, Tanzania." Guest edited by R.J. Blumenschine, F.T. Masao, I.G. Stanistreet, and C.C. Swisher, III. E-mail address: [email protected]. 0047-2484/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2012.03.001

Paleoenvironmental reconstruction of landscapes, on which early hominins such as Homo habilis and Paranthropus boisei were active, requires time planes defined at the finest possible resolution. At Olduvai an absolute time resolution has so far been established using a volcanic ash tephrostratigraphic framework, with numerical ages provided by dated tuffs (Hay, 1976; Walter

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

et al., 1992; Manega, 1993; Blumenschine et al., 2003). Such chronostratigraphic resolution of volcanic tuffs delimits intervals of about 15e30 ka duration but can be insufficient for defining precise time planes within the sedimentary intervals between when not adjacent to a tuff marker. Thus, there can be a lack of precision for reconstructing and determining the timing and duration of individual landscapes. A useful provision for hominin research in East Africa would be a more finely resolved time framework, on the order of thousands of years, that might complement the existing tephrostratigraphy. A Sequence Stratigraphic approach, potentially applicable to analysis of all lake-centered, hominin-yielding basins, was attempted in this study of the eastern Olduvai Basin. The approach is analogous to the Sequence Stratigraphic analyses that are more usually (Van Wagoner et al., 1988), but not exclusively (Shanley and McCabe, 1994), applied to marine basins. The pilot analysis at Olduvai was undertaken in Bed I and Lower Bed II, separated by the time marker Tuff IF, as a contribution to the Olduvai Landscape Paleoanthropology Project (OLAPP), dedicated to reconstructing the paleoenvironments and landscapes hosting Homo habilis and the associated Oldowan stone tool industry (Leakey, 1971). Tuff IF, sourced from the nearby, now extinct volcano Mt. Olmoti (Hay, 1976, 1990,1996; McHenry, 2005; McHenry et al., 2008; Stollhofen et al., 2008; Mollel et al., 2009), has been dated at just prior to 1.785 Ma (Blumenschine et al., 2003). The technique was further applied to the interval within Bed I between Tuffs IB and IC

301

(1.845
0.002 and 1.839
0.005 dating by Deino, reported in Blumenschine et al., 2003). Setting The approximately two million year history represented by the Olduvai Basin-fill (Fig. 1) comprises Beds IeIV, succeeded by the Masek, Ndutu, and Naisuisui beds (Fig. 2; Hay, 1976). Of these, Beds IeIV and the Masek Beds were all sourced directly from the eastern Volcanic Highlands (Fig. 1), aligned along the western margin of the Gregory Rift, the eastern arm of the African Rift System. Initially, during deposition of Beds I and II, dominant drainages and volcaniclastic fans were directed westwards away from the highlands into a fault-bound Rift Shoulder (¼“Rift Platform”) Basin (for definitions see Stollhofen et al., 1998; Ashley and Hay, 2002; Stollhofen and Stanistreet, 2012). Subsequently, as a response to further rift evolution, major drainages were directed northwesterly along the western side of the Volcanic Highlands (Hay, 1976), with additional sediment and tributary contributions from the area of the modern Serengeti. As faulting continued, drainages broke through the riftbounding faults towards the east, and the Olduvai River broke through the First Fault into the drainage sump of the Ol’balbal Depression during this time (Hay, 1976). This new direction of drainage, opposite to that of the earliest drainage, has been accompanied by incision of the Olduvai Gorge and exposure of the earlier fan deposits. The Ndutu and Naisuisui Beds (Hay, 1976) were

Figure 1. Location of Olduvai Gorge and its main paleogeographic features during Beds I and II deposition. An alluvial fan apron transported volcaniclastic material from then active volcano Mt. Olmoti towards the west, where it interacted with Paleolake Olduvai. Lettered sites HWK, MCK, RHC, etc. refer to Korongo (swahili ¼ gully) or Cliff site nomenclature of Leakey (1971). Loc. 6, Loc. 200 refer to Numbered Localities of Hay (1976).

302

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

deposited and remained as terraces at deepening levels within the gorge as incision proceeded up until the present day. Beds I and II in the eastern Olduvai Basin (Hay, 1976) were deposited as both primary volcanic but mostly secondary volcaniclastic materials on a subaerial fan or apron that extended 25 km westwards from the then active Mt. Olmoti volcano (Fig. 1). These Olduvai Fan sediments were interleaved with Paleolake Olduvai clays (Hay and Kyser, 2001) in the lower fan area (in the junction area of Leakey, 1971), an area referred to as the lake margin. Salinealkaline Paleolake Olduvai usually varied from about 8e15 km across (Hay, 1976) but occasionally almost totally dried out, as at the time of Tuff IF emplacement (Stollhofen et al., 2008). Characteristically within the study interval of Bed I and Lower Bed II, smectitic clays were deposited during lake flooding to form waxy claystones. During highstands in restricted sub-basins, these consisted almost entirely of the neoformed magnesian smectite end-member (stevensite) deposits, termed butter claystones. Lake withdrawal resulted in abandoned lake plains in which meteoric (fresh rainwater) conditions produced calcareous soils (Bennett et al., 2012) under the prevailing semi-arid conditions. Soil profiles were bioturbated by insect burrows, particularly termites, and root and rootlet systems, which could extend to form root mats. In rare cases, silicification of plant tissues has preserved and revealed the rhizomes of the subaerial portions of the plant (Bamford et al., 2008), capping the most complete soil profiles. Methods During nine field seasons from 2000 to 2009, measured sections for facies determination and stratigraphic delineation were derived from OLAPP trenches at 10 of the 14 sites investigated, with total backwalls or, less commonly, sidewalls mapped down to the millimeter scale. In addition, any natural outcrop between trenches was mapped to the centimeter scale to aid correlation within site. Up to 14, more commonly fewer than six, trenches at any one site were mapped, with spacings of 30e50 m, but one or two representative trenches were chosen from which to construct the more broad-scale

correlation diagrams. At the remaining four sites (Loc. 200, Loc. 6, DK, RHC) natural outcrop was excellent, allowing direct measurement to the centimeter scale, extending to mapping of facies panels 100 m across where this was appropriate or advantageous, and data are taken from the most complete sections. Facies, sedimentary structure, paleocurrent, plant, bone, and hominin trace information were mapped when exposed on backwall or natural outcrop. Archeological step and level number information and trench orientation were recorded on the trench wall map, as well as any relevant or important archeological information that could be projected onto the adjacent wall stratigraphy. This facilitated comparison with complementary archeological records. Incised disconformity surfaces Uppermost Bed I and Lower Bed II (Figs. 2 and 3) could immediately be partitioned by fundamental Sequence Stratigraphic boundaries: two incisional surfaces of regional extent, classically termed Type I disconformity, were generated after base lake-level was rapidly and pronouncedly lowered. At such times river and stream systems eroded down into the underlying volcanosedimentary sequence, and at the start of the subsequent lakelevel increase (transgression) river sediments were deposited in resulting topographic lows or thalwegs. The younger of these surfaces was originally identified by Hay (1976), lying immediately beneath the Lower Augitic Sandstone and defining the top of Lower Bed II. Subsequently recognized during OLAPP excavations, an older Type I incisional surface within Lower Bed II was deeply (>5 m) incised by sandy and gravelly river systems (Figs. 2 and 4). In the lake marginal area, incision resulted in a valley, whose subsequent fill was dominated by earthy sediments or siliceous earths sensu stricto (Hay, 1976), contrasting markedly with the wholly lacustrine waxy clays that underly the incision surface (Figs. 2 and 3). The incision surface also marks that time in the sequence history when augitic grains flooded into the basin, indicating a fundamental change in source volcanic style. Thus the post-incisional earthy claystone dominated sequences, the Lemuta Member, and the

Figure 2. Detailed measured stratigraphy of uppermost Bed I and Lower Bed II in the eastern Olduvai Gorge, showing location of major incision surfaces and identity of lakeparasequences. Circles ¼ ignimbrite; vvv ¼ tuff; inverted triangles ¼ diamictitic mudflow deposits; lines ¼ sheetflood deposits; dots ¼ volcaniclastic sandstones/reworked tuffs; dashes ¼ claystones, some bioturbated; thin dotted layers ¼ lake turbidites; irregular lumps with lozenge ornament ¼ paleosol carbonates.

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

303

Figure 3. Sequence Stratigraphic Wheeler Diagram of the same time interval as that of Figure 2. Horizontally distinct units are parasequences (4000e5000 years period average), each generated by a single advance and withdrawal of Paleolake Olduvai. Key colors are as follows: yellowþorange spots ¼ conglomerate; yellow ¼ volcaniclastic fluvial sandstone; pink ¼ volcaniclastic mudflow deposits; green ¼ lacustrine claystones modified by soil processes; bluish-green ¼ lacustrine claystones with turbidites; blue ¼ lacustrine dolomite; white ¼ earths and earthy claystones; VVV pattern ¼ primary tuffs.

Augitic Sandstones are all rich in augite, in contrast with the preincisional sequences. The recognition of this surface immediately resolves a previous correlation problem originally noted by Mary Leakey (Leakey, 1971; Hay, 1976): how to correlate earthy and waxy clay dominated Lower Bed II sediments at the HWK and FLK locations (Fig. 1; the reader is directed to Leakey [1971] for details of the Korongo system used to name places at Olduvai; e.g., MNK ¼ Mary Nicol Korongo). The solution is that the earthy clay dominated sediments are younger than the waxy clay dominated and are limited to the incised valley-fill. From the abundance of fossil crocodile remains and taphonomic evidence of crocodile toothmarked bone (for examples see Njau and Blumenschine, 2006) excavated from the earthy valley-fill sediments, the valley has been

named Crocodile paleo-Valley. From the facies perspective the earthy clays were deposited in wetland settings dominated by freshwater input (Deocampo et al., 2002; Liutkus and Ashley, 2003). The current sequence analysis (Fig. 3) now reveals that wetlands filling Crocodile Valley were fed by river systems that flowed from the volcanic highlands through fluvial paleo-gorges incised into the Olduvai Fan surface during the same lake lowstand. The setting might be compared to the way the modern Munge River within Ngorongoro caldera feeds valley-filling wetlands downstream at the Hippo Pools area. An older disconformity, subsequently filled and covered by Tuff IF volcanic surges (Stollhofen et al., 2008), is not deeply incised and would classify as a Type II unconformity in accordance with Sequence Stratigraphic terminology. Using the Type I and II disconformities (Figs. 2 and 3), the broadscale framework begins to emerge for an Oldowan Sequence Stratigraphy. However, a much finer time scale can be resolved within that overall framework, with an architecture provided by much higher frequency cyclical lake-level changes. Time slices or lake-parasequences of the Paleolake Olduvai margin

Figure 4. The whole of Lower Bed II and uppermost Bed I exposed at KK. The Crocodile paleo-Valley incision surface is indicated, demarcating pre- and post-incisional Lower Bed II. Darker incision-fill sediments reflect augite content. Other formal and informal stratigraphic units are labeled.

The fine-scale time beats of Sequence Stratigraphic units are orchestrated by the cycles of rising (transgression) and falling (regression) base levels within a basin. Classically these relate to sea-level changes, usually of Milankovitch frequencies, but within the enclosed Olduvai Basin they are due to water-level changes of paleo-Lake Olduvai. It is for this reason that the fundamental timerock unit generated by a single rise and fall of base level has been specified here a lake-parasequence. At Olduvai, lake-parasequences are most strikingly defined in the lake marginal setting, developed mainly below the Crocodile Valley Type I incision (see locations FLK to MCK, Fig. 3); however, this parasequential style returns within the Lemuta Member at the top of Lower Bed II. Lake-level highstands resulted in deposition of lacustrine “waxy” (Mg-smectite-rich; Hay, 1976) clay units 0.5e1.5 m thick or thicker. Complementary lower lake-levels after lake withdrawal are represented by minor disconformities and

304

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

subaerial calcareous soil profiles of rainfall isotopic character, precipitated in the subsurface near the top part of those units (Figs. 2 and 3). In the lake marginal area of the pre-incisional sequence, Figure 2 records at least six such parasequences. Actual lake withdrawal surfaces may be marked, as well as erosive surfaces, by horizon(s) marked by grass rootlet mats (Bamford et al., 2008), or hoof- and/or foot-printed to trampled surfaces. The parasequential nature of the waxy claystoneecalcareous soil couplets is supported by analysis of their character from archaeological trenches that encountered the Lemuta Member at TK at the top of Lower Bed II (e.g., Fig. 3: Location TK). There, each waxy claystoneepaleosol couplet is separated by a more minor Type II disconformity, with the resulting topography filled and covered by a volcaniclastic sandstone unit, representing, respectively, erosion during lake lowstand and subaerial deposition during the start of the next lake transgression. Subsequent lake flooding covered the area and deposited the overlying waxy clay unit prior to the next lake withdrawal and soil formation. This demonstrates the regressive-transgressive nature of each lake marginal couplet. The same cyclic character is further supported by d13C and d18O isotopic analysis (Bennett et al., 2012) of accretionary carbonate bodies within several paleosols of successive lake marginal parasequential couplets. Time series d13C and d18O isotope evolution paths show that the paleosols formed largely during the transgressive half of each lake cycle, with ultimate precipitation of strontianite and strontian dolomite derived directly from lacustrine groundwaters following lake inundation and prior to the eventual deposition of the next waxy clay unit. Above the Type I disconformity that defines the Crocodile paleoValley, during the time of earthy deposition, parasequences are erosionally based and fluvially dominated, comprising gravelly and sandy braided and meandering river systems in proximal gorgelike settings (see MK location in Fig. 3) and gravelly to sandy earths, passing up into earthy claystones and then waxy claystones (see VEK to MCK locations in Fig. 3) in the junction area. Above this there is a gradational change to parasequences within the Lemuta Member that return to the style closer to pre-incisional lakeparasequences. Tracing parasequential time slices eastwards onto the fan and westwards into the lake Higher on the Olduvai Alluvial Fan, east of the lake margin area and beyond the highest flood-levels of Paleolake Olduvai, is evidence of up to six pre-Crocodile Valley incision parasequences (Fig. 3). They exhibit facies characteristics very different from those close to Paleolake Olduvai. During aggradational sedimentary phases associated with lake transgression, clayey or sandy mudflows, sourcing from both Mt. Olmoti itself and reworked from higher areas of the fan, deposited their loads to form units 0.5e1.75 m thick. Lake withdrawal phases are represented either by shallow erosion of the fan surface, subsequently filled by 0.3e1.5 m gravelly sand shallow braided stream deposits, or in interfluve areas by w0.5 m deep vegetation root and termite bioturbated soils, biological elements that would have helped stabilize the fan surface. Rarely in areas close to watercourses, calcareous soil profiles also developed, most notably shown at the DK, but also at the MK, location (Fig. 3). Parasequential couplets, defined in the fan area, can be correlated with their lake marginal counterparts. The Paleolake Olduvai area, in which waxy clay units (0.4e0.7 m) were deposited west of the lake marginal setting, was bounded on the eastern side during intermediate lake levels of the pre-incision by the FLK Fault (Figs. 1 and 3). The lake transgressed across the fault during each cycle to deposit waxy clays over a broad

area. By contrast, water courses transported sand down the fan and directly into the lake during lake withdrawal phases, and the sand was redistributed within the lake by density current flows that deposited successive thin (1e2 cm) turbidite layers interbedded with the waxy clay deposits. The parasequential units so defined can also be correlated with their lake marginal counterparts. In contrast, at lake highstand pure chemical and allochemical carbonates were deposited as dolomite layers. Thicknesses of correlated parasequences vary, particularly in the junction area within the lake margin and adjacent lake area. It was realized that changes in parasequence thickness coincided with changes in thickness of and facies types within interleaved tuffs, notably Tuff IF. (Stollhofen and Stanistreet, 2012). Further, sudden changes of this type coincide with the position of synsedimentary faults, the FLK, KK, and Long K faults, and the parasequential succession can be used to help constrain the incremental evolution of those faults. In addition, artifact and bone assemblages recovered from individual fault-affected parasequences allow the reconstruction of how paleoenvironmental and paleoecological variation across a fault compartment provided a localized control on hominin activity (Blumenschine et al., 2012a). Thus, within each of fan, lake margin, and permanent lake settings parasequences can be traced, each representing a cycle of lake flooding and withdrawal. Correlating these upward or downward from well-marked volcanic ash time marker layers (Tuff IF, IIA, and Nge’ju Tuff of McHenry [2005]) allows individual time slices to be correlated throughout the eastern and potentially the entire Olduvai Basin. It is on this finely resolved time-rock framework that the Wheeler Diagram of Figure 3 is based. Additionally, this Sequence Stratigraphic framework can be used to develop a time scale for the study interval, complementary to that of the existing tephrostratigraphy. A first step is to investigate how much time is represented by an individual parasequence. Time framework, lake-parasequence period, and landscape reconstruction precision In Figure 3 the date of Tuff IF is taken as just prior to 1.785 Ma (Blumenschine et al., 2003). The date of the “eolian” Tuff IIA is more problematic, but I apply the date of 1.72
0.03 Ma, which represents the weighted mean biotite age (Manega, 1993). This date differs slightly from the date of 1.74
0.03 Ma chosen by Ashley (2007) for Tuff IIA, which does not fall into the age range of Tuff IIA delimited by Manega (1993) himself. On the basis of a 1.72
0.03 Ma date for Tuff IIA, the time span in between the two markers amounts to 65,000 years of Earth history. During that period a minimum of 13 parasequences can be counted, a minimum because of uncertainties related to potentially “missing” parasequential stratigraphy associated with incision surfaces, particularly the Crocodile Valley incision event. Additionally, parasequences recorded within the Crocodile incised valley infill are confined to channels and might be missed in line of exposure. Thus, each parasequence represents a maximum average period of <5000 years. Previously, paleoanthropological analysis of hominin sites in East Africa might identify a single synchronous surface for paleoenvironmental landscape reconstruction in a unit of the magnitude of Lower Bed II, an interval of w65,000 years, and would need to be related to a single, adjacent tuff layer. This analysis reveals 13 time slices, each constrained to thousands of years. Within each parasequence at least two landscapes can be drawn, one with the lake transgressed, the other with the lake withdrawn, and placed to within several thousand years during each parasequential period, between Tuff IF and Tuff IIA. Paleoanthropological occurrences can then be placed in their synchronously appropriate landscape and related to contemporary occurrences.

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

Comparison with the East African post-glacial lake record indicates a 4000e5000 year period Classic marine parasequences relate to Milankovitch cyclic forcing (frequency of w23, 41, and 100 kyr). During the PlioPleistocene these can be characteristically associated with pronounced glacial/interglacial sea-level (Van Wagoner et al., 1988) and sea temperature (e.g., Shackleton et al., 1995) changes. Clearly, individual lake-parasequences recorded in the Olduvai Basin (<5000 years) are of far too high a frequency to be of a Milankovitch period (>20,000 years). In order to identify a lake-level variation analogous with those identified in Paleolake Olduvai, the record of post-glacial East African lake variation (Gasse, 2000) has been considered. From dating of pronounced rises and falls of lake level, drying and wetting cycles are documented. In particular, the water-level variations of Lakes Abhe, Turkana, and Albert show sub-10,000 year cycles, generating paleosols during pronounced dry intervals in a manner analogous to the Olduvai parasequential cycles. The mean period of these cycles, recorded during the last 30 ka, is about 3900 years, like the Olduvai parasequences, relating to sub-Milankovitch climatic variation on a period of thousands of years. A recalibration of such cycles generated during the last 10 ka reveals that a period can be determined at about 4200 years. This is thought to represent the order of average periodicity represented by the Olduvai lake-parasequences, a conjecture that can be tested lower in the Olduvai basinal sequence as follows.

305

Figure 5 reproduces a stratigraphic profile of the Tuff IB to Tuff IC interval through the FLK, FLK N, and FLK NN sites, which shows the FLK balk section recorded by Leakey (1971) correlated laterally to 13 OLAPP trenches. The base of Tuff IC is seen to sit upon a Type I incised disconformity that forms a sequence boundary within Upper Bed I as it was defined by Hay (1976). From the difference between the ages of these tuffs, this interval might simplistically be measured as 6000 years; however, the combined errors are
7000 years, which allows an interval that might vary from very short to 13,000 years. Between Tuff IB and Tuff IC two parasequences can be easily identified. However, there is claystone of a parasequence only partially preserved below the incised surface, and this must have been completed to allow Tuff IC to be laid down subaerially above, implying lake withdrawal that allowed the development of the Zinjanthropus surface. Thus, close to three parasequential periods are represented between the tuffs. At the outside, 13,000 yr/3 allows only a maximum average of 4333 years per lakeparasequence. This tests the hypothesis that parasequential PlioPleistocene lake level transgressions and withdrawals were of a very similar average cyclic period to those described from the Holocene East African lake record by Gasse (2000) at 4200 years. This does not deny, however, that the Tuff IB to Tuff IC lakeparasequences could be shorter in duration. Continuing analysis of sequences between other well dated tuffs should allow further determination and refinement of lake-parasequence periodicity.

Stadial/interstadial climatic events as a possible control on lake-parasequential period Parasequential period within the Tuff IB to Tuff IC interval OLAPP has excavated trenches within Upper Bed I at locations throughout the Olduvai Basin. These include the western basin at Naisuisui around Trench 64 where a Homo habilis mandible was described (Blumenschine et al., 2003). Trenches have similarly established the landscape (Blumenschine et al., 2012b) that contextualizes the FLK Zinjanthropus site of Leakey (1971). At the Naisuisui site dating undertaken by Deino (reported in Blumenschine et al., 2003) revealed more precise ages for Tuff IB (1.845
0.002 Ma) and Tuff IC (1.839
0.005 Ma), and McHenry (2005) confirmed correlation of these tuffs into the FLK area. Thus the FLK trenches are tightly constrained chronostratigraphically.

Lake-parasequences do not have the sort of rigid, orbitally forced periodicities of their marine counterparts and thus their periods estimated here are always quoted as averages. In order to identify climatic cyclic events that might force such a high frequency record, it is necessary to examine the sorts of changes that occur within a single glacialeinterglacial cycle. The best studied and most highly resolved such record is that which follows the penultimate Eemian Interglacial period during the last 125 ka. In cores from the Arabian Sea, Schulz et al. (1998) relate monsoonally driven organic-rich/organic-poor sedimentary cycles within the time interval 10e110 ka to the recurrence of warm Heinrich Interstadial events, representing the maximal development of

Figure 5. Stratigraphic profile of the Tuff IB to Tuff IC interval in the FLK, FLK N, and FLK NN. Two and a half parasequences are contained between the tuffs but close to three lakeparasequential periods were completed. If such periods average about 4200 years, the resulting time can only just be fitted into the time interval, taking into account errors. However, 4200 years is an average period and the exact periods can be less. Trþnumber indicates OLAPP trench numbers. VVVV ¼ Tuff units; dashes ¼ claystones; dots ¼ volcaniclastic sandstones.

306

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

a series of Stadial/Interstadial climatic cycles: Dansgaard-Oeschger (DeO) events, registered in and identified from the Greenland ice record (Johnsen et al., 1992; Dansgaard et al., 1993). Using a variety of both sedimentary and faunal/floral indicators, Scholz et al. (2007) identify cycles within the same time interval and of similar periodicity in borehole cores from Lake Malawi, correlating with lakes Bosumtwi and Tanganyika. Several tropical African megadroughts relate to the same cyclicity between 135 and 90 ka and are marked by paleosol formation in the area of what is presently the floor of Lake Malawi (Cohen et al., 2007). Brown et al. (2007) use Zr:Ti plots against age to represent the Lake Malawi cycles and relate them directly to Dansgaard-Oeschger events 1e14. The sedimentary cycles recorded in the Lake Malawi and Arabian Sea borehole cores are related to DeO Events 1e14 (Brown et al., 2007) and 1e24 (Schulz et al., 1998), respectively. The longer time range records 24 cyclic DeO Events during a time interval of 94,000 years, representing an average recurrence of 3920 years, which is remarkably close to the average periodicities estimated for the lakeparasequences at Olduvai. If the same types of DeO events control the longevities of lake-parasequences, then the record demonstrates that this polar effect was registered at least as early as 1.8 Ma. Just as marine parasequences represent the sedimentary record of GlacialeInterglacial alternations, so lake-parasequences may represent the sedimentary record of StadialeInterstadial rhythms.

Implications of the incision surfaces for previous paleoenvironmental and cyclic paleoclimatic models contextualizing Olduvai hominin settings The identification of major disconformity surfaces previously unrecognized within the Lower Bed II stratigraphy has considerable implications for existing models of paleoanthropologically significant paleoenvironmental settings at Olduvai. The more general interpretations of Hay (1976) are relatively little affected because they involve more general correlations concerning overall time intervals between successive tuffs. Figure 2 and derived Figure 3 can be viewed as a detailed stratigraphy set within his broadscale Bed I and Bed II framework. As previously mentioned, the position of the Crocodile Valley incision explains difficulties of correlation encountered by Leakey (1971) between her Levels 1 and 2 at HWK E and the Deinotherium horizon at FLK N. However, because there was no recognition of the Crocodile Valley incision, subsequent interpretations (Ashley and Hay, 2002; Ashley et al., 2009) misplace units between the middle and lower fan and the lake margin area. For example, the Lower Bed II earthy claystonerich sequence of the junction area above the incision surface has been wrongly correlated with diamictite-rich parasequences of the middle to lower fan situated below the incision surface (Fig. 3). This led to the mistaken conclusion that the earthy claystone sequences are spring-fed wetlands at the toe of the fan system. Instead, reference to Figure 3 indicates that they are river-fed wetlands contained within the Crocodile Valley, fed by the same river system during its aggradational phase, which had already caused the incision itself. The more proximal fluvial characteristics of the river system that fed the valley are preserved at Site MK (Fig. 3). Whereas some hominin sites in East Africa may have centered on spring related wetlands, such settings should not be promoted as a common or necessarily usual focus for hominin activity (Ashley et al., 2009). The important hominin sites at Olduvai have so far appeared to be related to river-fed wetland or riparian settings, a good example of which is provided by the paleolandscape derived from the sub-Tuff IC Zinjanthropus incised surface portrayed in Figure 5, described and interpreted in more detail in Blumenschine et al. (2012b).

A second consequence related to the Crocodile Valley incision surface is that the number of cycles has been undercounted. Although only three cycles between Tuffs IF and IIA were counted by Ashley (2007) and Ashley et al. (2009), at least 12 are present. Thus, cycles with average 4000e5000 year period cycles have been misidentified as ones of Milankovitch frequency. Nevertheless, orbitally forced climatic cycles of Milankovitch period do register in a different way within the new stratigraphic framework at Olduvai portrayed in Figure 3, and these will now be considered.

Relating Olduvai Sequence Stratigraphy to a Milankovitch time framework The Sequence Stratigraphic analysis of Beds I and II allows the paleoenvironmental variation registered by successive incisions and changes of intervening parasequence style to reflect major paleoclimatic variation. This can be correlated with the glacial/ interglacial periodicity registered in East Africa through the forcing of the monsoonal circulation (Gasse, 2000). Glacial maxima are associated with drier phases or interpluvials and interglacial maxima with wetter phases or pluvials (Clemens et al., 1991). The resulting absolute time framework (Fig. 6) applies separate isotopic (Shackleton et al., 1995) and blown sediment flux (DeMenocal and Bloemendal, 1995) climatic variation indicators, and importantly is independent of and complementary to the radiometric tuff date record, allowing the relative parasequential time record to be structured within a more detailed time frame. The record of Figure 3 identifies two well defined and datable arid phases, the first during Tuff IF emplacement when Paleolake Olduvai reduced to minor ephemeral playas at the basin center (Stollhofen et al., 2008) and the second during the deposition of the

Figure 6. Incision surfaces and maximum lake flooding events fitted to climatic cycles, also registered by change in parasequence style and constrained by their contained tuff time markers. The cycles are ultimately forced by Milankovitch orbital variation and were measured by isotope temperature record (3.5 to 4.5 d18O pdb; Shackleton et al., 1995) and terrigenous flux (1.0e2.0 g/cm2/kyr) record of grains blown offshore (DeMenocal and Bloemendal, 1995). Arid and humid phases relate respectively to glacial and interglacial events, when monsoonal circulations were either suppressed or promoted. By this approach, absolute time constraints on the Oldowan framework and contained parasequences are considerably more finely resolved by a climatic framework independent of but complementary to the existing stratigraphy.

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

Lemuta Member, including Tuff IIA deposition (Hay, 1976). Between those arid phases another interval of overall pronounced lake withdrawal is marked by the incision of the Crocodile paleo-Valley. Two major wet phases intervened, the first when lake transgression achieved a maximum immediately after Tuff IF (Figs. 3 and 6) and the second associated with the filling of the incised Crocodile paleo-Valley, which was fed fluvially via paleo-gorges cut headward into the Olduvai Fan. These two pronouncedly wet phases were separated by a semi-arid phase when the Olduvai Fan was particularly aggradationally active. This was associated with volcaniclastic reworking of tephra which was erupted from Mt. Olmoti at the time of Tuff IF, including reworking (Fig. 3) of voluminous ignimbrites deposited at the base of the volcano and onto the upper Olduvai Fan (Hay, 1976; Stollhofen et al., 2008). In addition, a wet phase is recorded in a transgressive maximum that immediately preceded Tuff IF prior to the formation of the grassland occurrence (Bamford et al., 2008), and the Lemuta aridity crisis was succeeded by another transgressive maximum prior to the incision surface upon which the Lower Augitic Sandstone was laid down (Hay, 1976). Thus, the Tuff IF and Lemuta arid phases can be closely dated by the age of Tuff IF and less well by Tuff IIA, allowing equivalent glacial maxima to be tentatively specified from the oceanic glacial sea temperature record (Fig. 6). Tuff IF dates just prior to 1.785 Ma (Blumenschine et al., 2003), while the date taken for Tuff IIA is 1.72 Ma (Manega, 1993). Figure 6 places these ages of aridity in the context of the oceanic isotopic glacial/interglacial record (Shackleton et al., 1995), and the correlative glacials, associated with minimal monsoonal circulation, were at 1.787 Ma and 1.700 Ma (marked “Tuff IF aridity” and “Lemuta aridity” in Figure 6). The only significant glacial phase between those two is that at 1.745 Ma, which is identified as the reason for the lake-level drop that caused the relatively deep incision of Crocodile Valley. Intervening wetter periods can also then be dated. The lake transgressive maximum of uppermost Bed I (Bamford et al., 2008) just prior to Tuff IF would date to 1.802 Ma. The pluvial phase that filled the fluvial paleo-gorges that fed the earthy fill of Crocodile paleo-Valley would date to between 1.715 and 1.735 Ma, and the post-Lemuta transgressive sequence might date to 1.680 Ma. On that basis the glacial most likely to have allowed the erosion of the major Type I incision surface preceding Lower Augitic Sandstone deposition would tentatively date to 1.65 Ma.

Benefits to paleoanthropology and new challenges and implications of Sequence Stratigraphy at Olduvai and elsewhere Identification of time-rock units of the order of 4000 years average provides a finely resolved time framework in which accurate reconstructions of contemporaneous paleoenvironments and landscapes can be drawn within intervals of several thousand years. However, they also allow the contextualization of classically known (Leakey, 1971) and recently discovered (OLAPP) occurrences at Olduvai of considerable paleoanthropological significance. For instance, FLK N Level 1 and the Deinotherium find recorded by Leakey (1971) can now be shown to have a time separation of about one lake-parasequential period. On the other hand, the elephantid and giraffid carcasses at VEK and a rhinocerid carcass at KK discovered by OLAPP, all of which are associated with stone artifacts, are seen to occur on a contemporaneous landscape (Fig. 3). On the other hand, Leakey (1971) recorded Levels 1 and 2 at HWK E as apparently closely related spatially, within 4 m of section of one another. However, Level 1 is pre-incisional and Level 2 is postincisional and they are therefore in fact separated by 10

307

parasequences, counterintuitively representing an intervening period on the order of 40,000 years. The refinement of time sequences in which hominin fossils and stone artifacts can be contained and considered brings new opportunities and challenges to landscape paleoanthropology at Olduvai. Previous work by OLAPP has focused on examination of landscape patterning in Oldowan hominin activity traces within time-stratigraphic units delimited by volcanic tuffs. Here, the potential advantages over traditional site-based interpretations of hominin land use gained by examining activity traces at nearbasin-wide scales (e.g., Blumenschine et al., 2012a) are compromised by the potentially great, Milankovitch-scale environmental variability within the target stratigraphic interval. For the whole of lowermost Bed II, OLAPP’s most intensively investigated interval, the estimated 80,000-year duration of the interval witnessed two Milankovitch 41,000 year cycles. It proved possible to detect predicted patterning in the transport of quartzite for stone tool manufacture in the eastern Olduvai Basin during this interval (Blumenschine et al., 2008) in large part because the expected environmental changes did not affect the fixed source of quartzite at the inselberg, Naibor Soit. To do so, however, required at that time 11 field seasons devoted in major part to excavation of almost 8200 stone artifacts from 100 trenches across lowermost Bed II eastern basin. More recently, Blumenschine et al. (2012a) demonstrated behavioral patterning in eastern basin distribution of stone artifacts for the interval within lowermost Bed II preceding the major incision event that created Crocodile Valley. This temporal refinement to a duration of an estimated 20,000 years came at the cost of reducing artifact sample sizes by more than 50%, resulting in sample size inadequacies for some of the geographic locales across the eastern basin. Sample size problems are exacerbated by the new parasequence framework. For example, for the first parasequence of lowermost Bed II, the number of artifacts recovered by OLAPP over 12 seasons of excavation is reduced by another 60%. It will therefore be necessary to increase sample sizes through further excavation in order to take advantage of the current ability to examine landscape-scale variability in activity traces within and among single-parasequence intervals of about 4000 years. Sequence Stratigraphic analyses such as those undertaken here in the Olduvai Basin should be applicable to other hominin-yielding lake-centered basins, not only in East Africa but elsewhere worldwide. Because the lake-parasequences and sequences are climatically driven, the former possibly by Stadial/Interstadial DansgaardOeschger events, the latter by Milankovitch-scale variation, there should also be the potential in East Africa, wherever tephrostratigraphic absolute frameworks are sufficient, to correlate sequences and even parasequences from one hominin-yielding basin to others. This has been achieved by Gasse (2000) for lake-parasequences in Holocene basins of lakes Abhé, Turkana, Albert, Tanganyika, and Malawi and by Scholz et al. (2007) for lakes Malawi, Bosumtwi, and Tanganyika. There seems to be no reason why such regional correlations might not now be possible for the Pleistocene and Pliocene. Acknowledgments I would like to thank the Tanzania Commission for Science and Technology, the Tanzania Antiquities Department of the Ministry of Natural Resources and Tourism, and the Ngorongoro Conservation Area Authority for granting permission to conduct research at Olduvai Gorge. The late Richard Hay together with Harald Stollhofen, Lindsay McHenry, Robert Blumenschine, Fidelis Masao, Jackson Njau, Jim Marshall, Lis Rushworth, Charles Peters, Peter Andrews, Marion Bamford, Rosa Albert, and many other members of the Olduvai Landscape and Paleoanthropology Project (OLAPP)

308

I.G. Stanistreet / Journal of Human Evolution 63 (2012) 300e308

provided stimulating scientific discussions and invaluable support in the field. I am also grateful for financial contribution to the research, provided by NSF Grant BCS-0109027; Wenner-Gren Foundation for Anthropological Research; National Geographic Committee for Research and Exploration; The Leakey Foundation; and the Center for Human Evolutionary Studies, Rutgers University.

References Ashley, G.M., 2007. Orbital rhythms, monsoons, and playa lake response, Olduvai Basin equatorial East Africa (ca. 1.85e1.74 Ma). Geology 35, 1091e1094. Ashley, G.M., Hay, R.L., 2002. Sedimentation patterns in a Plio-Pleistocene volcaniclastic rift-margin basin, Olduvai Gorge, Tanzania. In: Renaut, R.W., Ashley, G.M. (Eds.), Sedimentation in Continental Rifts. Society for Sedimentary Geology Special Publication, vol. 73, pp. 107e122. Ashley, G.M., Tactikos, J.C., Owen, R.B., 2009. Hominin use of springs and wetlands: paleoclimate and archeological records from Olduvai Gorge (w1.79e1.74 Ma). Palaeogeogr. Palaeoclimatol. Palaeoecol. 272, 1e16. Bamford, M.K., Stanistreet, I.G., Stollhofen, H., Albert, R.M., 2008. Late Pliocene grassland from Olduvai Gorge, Tanzania. Palaeogeogr. Palaeoclimatol. Palaeoecol. 257, 280e293. Bennett, C.E., Marshall, J.D., Stanistreet, I.G., 2012. Carbonate horizons, paleosols and lake flooding cycles: Beds I & II of Olduvai Gorge, Tanzania. J. Hum. Evol 63, 328e341. Blumenschine, R.J., Peters, C.R., Masao, F.T., Clarke, R.J., Deino, A.L., Hay, R.L., Swisher, C.C., Stanistreet, I.G., Ashley, G.M., McHenry, L.J., Sikes, N.E., Van der Merwe, N.J., Tactikos, J.C., Cushing, A.E., Deocampo, D.M., Njau, J.K., Ebert, J.I., 2003. Late Pliocene Homo and Hominid Land Use from Western Olduvai Gorge, Tanzania. Science 299, 1217e1221. Blumenschine, R.J., Masao, F.T., Tactikos, J.C., Ebert, J.I., 2008. Effects of distance from stone source on landscape-scale variation in Oldowan artifact assemblages in the Paleo-Olduvai Basin, Tanzania. J. Archaeol. Sci. 35, 76e86. Blumenschine, R.J., Masao, F.T., Stollhofen, H., Stanistreet, I.G., Bamford, M.K., Albert, R.M., Njau, J.K., Prassack, K.A., 2012a. Landscape distribution of Oldowan stone artifact assemblages across the fault compartments of the eastern Olduvai Lake Basin during early lowermost Bed II times. J. Hum. Evol. 63, 384e394. Blumenschine, R.J., Stanistreet, I.G., Njau, J.K., Bamford, M.K., Masao, F.T., Albert, R.M., Stollhofen, H., Andrews, P., Prassack, K.A., McHenry, L.J., FernandezJalvo, Y., Camilli, E.L., Ebert, J.I., 2012b. Environments and hominin activities across the FLK Peninsula during Zinjanthropus times (1.84 Ma), Olduvai Gorge, Tanzania. J. Hum. Evol. 63, 364e383. Brown, E.T., Johnson, T.C., Scholz, C.A., Cohen, A.S., King, J., 2007. Abrupt change in tropical african climate linked to the bipolar seesaw over the past 55,000 years. Geophys. Res. Lett. 34, L20702. doi:10.1029/2007GL031240. Clemens, S., Prell, W., Murray, D., Shimmield, G., Weedon, G., 1991. Forcing mechanism of the Indian Ocean monsoon. Nature 353, 720e725. Cohen, A.S., Stone, J., Beuning, K., Park, L., Reinthal, P., Dettman, D., Scholz, C.A., Johnson, T., King, J.W., Talbot, M., Brown, E., Ivory, S., 2007. Ecological Consequences of early late-pleistocene megadroughts in Tropical Africa. Proc. Natl. Acad. Sci. 104, 16422e16427. Dansgaard, W., Johnsen, S.J., Clausen, H.B., Dahl-Jensen, D., Gundestrup, N.S., Hammer, C.U., Hvidberg, C.S., Steffensen, J.P., Sveinbjörnsdottir, A.E., Jouzel, J., 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364, 218e220. DeMenocal, P.B., Bloemendal, J., 1995. Plio-Pleistocene climatic variability in subtropical Africa and the paleoenvironment of hominid evolution: a combined data-model approach. In: Vrba, E.S., Denton, G.H., Partridge, T.C., Burckle, L.H. (Eds.), Paleoclimate and Evolution, with Emphasis on Human Origins. Yale University Press, New Haven, pp. 262e288. Deocampo, D.M., Blumenschine, R.J., Ashley, G.M., 2002. Wetland Diagenesis and Traces of Early Hominids, Olduvai Gorge, Tanzania. Quat. Res. 57, 271e281.

Gasse, F., 2000. Hydrological changes in the African tropics since the last Glacial Maximum. Quat. Sci. Rev. 19, 189e211. Hay, R.L., 1976. Geology of the Olduvai Gorge. University of California Press, Berkeley, California, 203 pp. Hay, R.L., 1990. Olduvai Gorge: a case history in the interpretation of hominid paleoenvironments in East Africa. In: Geological Society of America, Special Publication, vol. 242, pp. 23e37. Hay, R.L., 1996. Stratigraphy and Lake-Margin Palaeoenvironments of Lowermost Bed II in Olduvai Gorge. Kaupia 6, 223e230. Hay, R.L., Kyser, T.K., 2001. Chemical sedimentology and paleoenvironmental history of Lake Olduvai, a Pliocene lake in northern Tanzania. Geol. Soc. Am. 113, 1505e1521. Johnsen, S.J., Clausen, H.B., Dansgaard, W., Fuhrer, K., Gundestrup, N., Hammer, C.U., Iversen, P., Jouzel, J., Stauffer, B., Steffensen, J.P., 1992. Irregular glacial interstadials recorded in a new Greenland ice core. Nature 359, 311e313. Leakey, M.D., 1971. Olduvai Gorge 3. Excavations in Beds I and II. Cambridge University Press, Cambridge, 306 pp. Liutkus, C.M., Ashley, G.M., 2003. Facies model of a semiarid freshwater wetland, Olduvai Gorge, Tanzania. J. Sediment. Res. 73, 691e705. Manega, P.C., 1993. Geochronology, geochemistry and isotopic study of the PlioPleistocene Hominid sites and the Ngorongoro volcanic highland in Northern Tanzania. Ph.D. Dissertation, University of Colorado, Boulder, 382 p. McHenry, L.J., 2005. Phenocryst composition as a tool for correlating fresh and altered tephra, Bed I, Olduvai Gorge, Tanzania. Stratigraphy 2, 101e115. McHenry, L.J., Mollel, G.F., Swisher, C.C., 2008. Compositional and textural correlations between Olduvai Gorge Bed I tephra and volcanic sources in the Ngorongoro Volcanic Highlands, Tanzania. Quat. Int. 178, 306e319. Mollel, G.F., Swisher III, C.C., McHenry, L.J., Feigenson, M.D., Carr, M.J., 2009. Petrogenesis of basalt-trachyte lavas from Olmoti Crater, Tanzania. J. Afr. Earth Sci. 54, 127e143. Njau, J.K., Blumenschine, R.J., 2006. A diagnosis of crocodile feeding traces on larger mammal bone, with fossil examples from the Plio-Pleistocene Olduvai Basin, Tanzania. J. Hum. Evol. 50, 142e162. Schulz, H., von Rad, U., Erlenkeuser, H., 1998. Correlation between Arabian Sea and Greenland climate oscillations of the past 110,000 years. Nature 393, 54e57. Scholz, C.A., Johnson, T.C., Cohen, A.S., King, J.W., Peck, J.A., Overpeck, J.T., Talbot, M.R., Brown, E.T., Kalindekafe, L., Amoako, P.Y.O., Lyons, R.P., Shanahan, T.M., Castaneda, I.S., Heil, C.W., Forman, S.L., McHargue, L.R., Beuning, K.R., Gomez, J., Pierson, J., 2007. East African megadroughts between 135e75 kyr ago and bearing on early-modern human origins. Proc. Natl. Acad. Sci. 104, 16416e16421. Shackleton, N.J., Hall, M.A., Pate, D., 1995. Pliocene stable isotope stratigraphy of Site 846. Proc. Ocean Drilling Program, Sci. Results 138, 337e355. Shanley, K.W., McCabe, P.J., 1994. Perspectives on the sequence stratigraphy of continental strata. Am. Assoc. Petrol. Geol. Bull. 78, 544e568. Stollhofen, H., Stanistreet, I.G., 2012. Plio-Pleistocene synsedimentary fault compartments, foundation for the eastern Olduvai Basin paleoenvironmental mosaic, Tanzania. J. Hum. Evol. 63, 309e327. Stollhofen, H., Gerschütz, S., Stanistreet, I.G., Lorenz, V., 1998. Tectonic and volcanic controls on early Jurassic rift-valley lake deposition during emplacement of Karoo flood basalts, southern Namibia. Palaeogeogr. Palaeoclimatol. Palaeoecol. 140, 185e215. Stollhofen, H., Stanistreet, I.G., McHenry, L.J., Mollel, G.F., Blumenschine, R.J., Masao, F.T., 2008. Fingerprinting facies of the Tuff IF marker, with implications for early hominin palaeoecology, Olduvai Gorge, Tanzania. Palaeogeogr. Palaeoclimatol. Palaeoecol. 259, 382e409. Van Wagoner, J.C., Posamentier, H.W., Mitchum Jr., R.M., Vail, P.R., Sarg, J.F., Loutit, T.S., Hardenbol, J., 1988. An overview of the fundamentals of sequence stratigraphy and key definitions. In: Wilgus, C.K., Hastings, B.S., Kendall, C.G.St.C., Posamentier, H.W., Ross, C.A., Van Wagoner, J.C. (Eds.), Sea Level Changes e An Integrated Approach. SEPM Special Publication, vol. 42, pp. 39e45. Walter, R.C., Manega, P.C., Hay, R.L., 1992. Tephrochronology of Bed I, Olduvai Gorge: an application of laser-fusion 40Ar/39Ar dating to calibrating biological and climatic change. Quat. Int. 13e14, 37e46.