You are on page 1of 13


net/publication/256695135_East_African_mid-Holocene_wetAuthor's personal copy

Earth and Planetary Science Letters 331-332 (2012) 322334

Contents lists available at SciVerse ScienceDirect

Earth and Planetary Science Letters

journal homepage:

East African mid-Holocene wetdry transition recorded in palaeo-shorelines of Lake

Turkana, northern Kenya Rift
Yannick Garcin a,, Daniel Melnick a, Manfred R. Strecker a, Daniel Olago b, Jean-Jacques Tiercelin c

Universitt Potsdam, Institut fr Erd- und Umweltwissenschaften and DFG Leibniz Center for Surface Process and Climate Studies, 14476 Potsdam, Germany
University of Nairobi, Department of Geology, PO Box 30197-00100, Nairobi, Kenya
UMR CNRS 6118 Gosciences Rennes, Universit de Rennes 1, Rennes, France

a r t i c l e

i n f o

Article history:
Accepted 2 March 2012
Available online 1 April 2012
Editor: P. DeMenocal
East African Rift System
Lake Turkana
African Humid Period
Tectonic deformation

a b s t r a c t
The wet early to mid-Holocene of tropical Africa, with its enhanced monsoon, ended with an abrupt shift toward drier conditions and was ultimately replaced by a drier climate that has persisted until the present day.
The forcing mechanisms, the timing, and the spatial extent of this major climatic transition are not well understood and remain the subject of ongoing research. We have used a detailed palaeo-shoreline record from
Lake Turkana (Kenya) to decipher and characterise this marked climatic transition in East Africa. We present
a high-precision survey of well-preserved palaeo-shorelines, new radiocarbon ages from shoreline deposits,
and oxygen-isotope measurements on freshwater mollusk shells to elucidate the Holocene moisture history
from former lake water-levels in this climatically sensitive region. In combination with previously published
data our study shows that during the early Holocene the water-level in Lake Turkana was high and the lake
overowed temporarily into the White Nile drainage system. During the mid-Holocene (~5270 300 cal. yr
BP), however, the lake water-level fell by ~ 50 m, coeval with major episodes of aridity on the African continent. A comparison between palaeo-hydrological and archaeological data from the Turkana Basin suggests
that the mid-Holocene climatic transition was associated with fundamental changes in prehistoric cultures,
highlighting the signicance of natural climate variability and associated periods of protracted drought as
major environmental stress factors affecting human occupation in the East African Rift System.
2012 Elsevier B.V. All rights reserved.

1. Introduction
Deciphering the long-term climate history and accurately identifying the mechanisms underlying variations in hydrologic budgets is
an important task in light of ongoing global change, water stress,
and the associated environmental and socioeconomic impacts in the
African tropics (Boko et al., 2007). In this context climatic variability,
climatic extremes, and the transitions between episodes with different environmental conditions, have become the focus of numerous
investigations in tropical Africa (e.g., Shanahan et al., 2009; Tierney
et al., 2008).
A period of particular interest in the climatic history of Africa is the
African Humid Period or AHP (cf. deMenocal et al., 2000; Ritchie et al.,
1985). The AHP occurred approximately between ~12,000 and ~ 5000
calendar years before present (cal. yr BP) and resulted in a northward
expansion of vegetation zones (e.g., Hoelzmann et al., 1998). Extensive parts of tropical Africa subsequently experienced pronounced
and rapid hydrologic changes associated with the termination of the
AHP during the mid-Holocene, about 5000 years ago (e.g.,
deMenocal et al., 2000). This transition toward drier conditions
Corresponding author. Tel.: + 49 331 977 5837; fax: + 49 331 977 5700.
E-mail address: (Y. Garcin).
0012-821X/$ see front matter 2012 Elsevier B.V. All rights reserved.

fundamentally impacted ecosystems across northern Africa, prompting a return to arid and semiarid vegetation in the Sahara and the
Sahel regions (Jolly et al., 1998). These environmental changes are
also believed to have led to important demographic shifts (Brooks,
2006; Kuper and Krpelin, 2006).
The origin of, and underlying mechanistic principles for, the AHP
termination are not yet fully understood and remain the subject of
ongoing investigations. This episode, which may have lasted a few
centuries, is considered to have been too rapid to be solely driven
by a linear response to gradual insolation changes (e.g., Claussen et
al., 1999; Cole et al., 2009; deMenocal et al., 2000). On the other
hand, the existence of an abrupt and spatially synchronous AHP termination has recently been called into question (e.g., Chase et al.,
2010; Krpelin et al., 2008; Marshall et al., 2011).
Here, we present a detailed study of multiple abandoned Holocene
shorelines from the Lake Turkana basin in the northern Kenya Rift of
the East African Rift System (EARS, Fig. 1). These palaeo-shorelines
provide a record of the East African moisture history that helps in
unravelling the characteristics and environmental impacts of the
AHP termination in a region that is located in the immediate vicinity
of the equator. Preliminary studies of these palaeo-shorelines provided an unprecedented insight into the environmental history of the
Holocene (e.g., Butzer et al., 1972; Owen et al., 1982). However, the

Author's personal copy

Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334


Fig. 1. (A) Map of the Turkana Basin and adjacent basins (SRTM topography). Also shown are rivers (thin blue lines), catchment boundaries (thick and thin black lines), the maximum
extent of Lake Turkana during the Holocene (MHS, thick blue line) approximated using a present-day surface elevation of 460 m, and the various overow sills in the area (arrows linked
to white circles). The overows were active during the early to mid-Holocene when the Turkana Basin received inow from the Bogoria, Baringo, and Suguta basins to the south and the
Abaya, Chamo, and Chew Bahir basins to the east; at that time Lake Turkana overowed into the Nile Basin. Bathymetric map of Lake Turkana (below 360 m) from Johnson et al. (1987).
Inset is a structural map of the EARS. (B) Location map for the radiocarbon sample sites and dGPS survey sites. Also shown are the Holocene shorelines, the approximate position of the
maximum highstand shoreline, and the main recent (late Quaternary) faults; all were mapped using eld observations, SPOT satellite imagery, and SRTM data, together with previously
published data (e.g., Baker et al., 1972; Dunkelman et al., 1989; Johnson et al., 1987; Morley et al., 1992).

exact nature and timing of some inferred environmental changes derived from lake water-level uctuations were subsequently found to
be equivocal and/or irreproducible (cf. Barton and Torgersen, 1988;
Cerling, 1986; Halfman et al., 1992; Ricketts and Johnson, 1996). In
our study we have focused primarily on the largely unexplored southernmost margin of the Turkana Basin (Fig. 1B), where steep slopes
have helped preserve a unique staircase morphology of palaeo-shorelines, between 0 and 80 m above the present-day lake water-level. In
order to reconstruct lake water-level uctuations in the Turkana

Basin and use them as indicators of past variations in the moisture regime, we mapped shoreline elevations with a high-precision differential global positioning system (dGPS) and measured oxygen-isotope
ratios of in situ fossil mollusk shells collected from shoreline deposits.
By combining 21 new 14C ages obtained from these shells with other
published ages we have constrained the timing of past lake waterlevel changes and provided insights into the environmental history
of Lake Turkana, particularly during the mid-Holocene. We also
used the abundantly available archaeological data from the Turkana

Author's personal copy


Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334

Basin to address the possible inuence of the AHP termination on the

cultural development of prehistoric populations.

2. General characteristics of the Turkana Basin

The fault-bounded Turkana Basin constitutes an area of
~ 148,000 km 2, while the lake (formerly Lake Rudolf) has a surface
area of ~ 7000 km 2. The Turkana Basin is internally drained and is
part of a low-lying area in the zone between the Kenya Rift to the
south and the Ethiopian Rift to the north (Fig. 1). The Lake Turkana
area comprises a series of halfgrabens with alternating polarity
(Morley et al., 1992). The basin accommodates uvio-lacustrine and
volcanoclastic deposits up to 7 km thick, accumulated between the
Late Oligocene-Plio-Pleistocene and the present (Dunkelman et al.,
1989; Morley et al., 1992). The faulted and tilted blocks are associated
with several volcanic centres rising above the modern lake surface,
forming three small islands (the North, Central, and South islands)
(Brown and Carmichael, 1971; Karson and Curtis, 1994).
South Island, which is ~ 12 km long and ~ 5 km wide, forms the
main focus of this study. Several distinct NS striking faults and magmatic ssures (Karson and Curtis, 1994) traverse the entire island. Active normal faulting has been documented in the recent sediments
surrounding the island (Johnson et al., 1987). Terraces, beach ridges,
and wave-cut platforms are particularly well developed in the western part of South Island (Karson and Curtis, 1994).
Lake Turkana is elongated in a NS direction, following the rift axis
(Fig. 1). It is ~ 250 km long and ~ 30 km wide. The maximum lake
depth is 120 m, with an average depth of 35 m (Johnson et al.,
1987). Lake Turkana is the most alkaline permanent lake in East Africa, with an alkalinity of 2023 meq/l (Cerling, 1979). The climate in
the southern and central part of the Turkana Basin, including the
main depression accommodating the lake, is particularly arid, with
the mean annual rainfall reaching only ~200 mm/yr and a mean potential evaporation of ~ 3000 mm/yr. The mean annual minimum
and maximum air temperatures are 23.6 and 34 C, respectively

(East African Meteorological Department, 1975; Ferguson and

Harbott, 1982). In contrast, the northern part of the Turkana Basin,
which transitions into the Ethiopian Highlands, is signicantly cooler
and wetter and receives more than 1500 mm/yr of monsoonal rainfall
(Halfman and Johnson, 1988). The Omo River (Fig. 1), which ows
into Lake Turkana from the Ethiopian Highlands, consequently supplies 8090% of the freshwater inow and thus exerts a primary control on lake water-level uctuations today (Yuretich and Cerling,
1983). The Kerio and Turkwel rivers to the south account for the
remaining inow to the lake, derived from westerly catchments.

3. Materials and methods

3.1. Kinematic dGPS survey of palaeo-shorelines
We surveyed palaeo-shorelines from the southernmost part of the
Turkana Basin (Figs. 14) following the same approach as used in our
previous investigations in the nearby Suguta Valley (Garcin et al.,
2009). We carried out high-precision topographic surveys of the
prominent palaeo-shorelines from this remote area using a dGPS
(Leica GPS 1200) and helicopter support. We focused on the palaeoshorelines on South Island as this area contains a well-preserved sequence of beach ridges (Fig. 4) previously identied on SPOT satellite
images, and surveyed their elevations along a down-slope traverse,
perpendicular to the ridges (Fig. 4B, D). We also measured elevation
proles by moving parallel to the shorelines (Fig. 4B). Kinematic
GPS data were collected continuously at a 1 Hz observation rate.
The base-rover distance ranged from 47 to 75 km. The survey data
was post-processed using Leica GeoOfce (V6.0) software, which
only provided coded solutions. The averaged height quality, which
corresponds to the standard deviation of the height element of our
21,000 post-processed eld-measurements, was 22.1 cm. Note that
the analytical errors of this measurement method are small compared
to the vertical range of surveyed shoreline elevations (~77 m for
South Island, Fig. 4E), and to the natural variability in the elevations

Abili Agituk
440 m asl


Lake Turkana: 360 m asl

South Island

Abili Agituk




Lake Turkana

Fig. 2. Aerial views of the maximum highstand shoreline (MHS: white arrows) cut into the Abili Agituk volcano (southernmost tip of Lake Turkana). Top photograph: view to the
east. Bottom photograph: view to the north, with South Island in the background. Several normal faults offset the MHS.

Author's personal copy

Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334


Fig. 3. Differential elevations for the maximum highstand shoreline (MHS) across the southern Turkana Basin. (A) SRTM shaded-relief digital topography of the surveyed area. Also
shown is the 460 m contour (dark blue line) corresponding to the present-day maximum elevation of the overow sill. Red dots are the MHS sites surveyed with dGPS (cf. Melnick
et al., this volume). (B) Same area as shown in A: map of the recent (late Quaternary) active faults. Offshore faults were extrapolated from Dunkelman et al. (1989) and Johnson
et al. (1987). (C) WE swath prole. Green line indicates mean elevation values from the swath prole shown in A; grey lines with area shaded between are minimum and
maximum elevation values. Also shown are the projected main faults (cf. map B) as well as the measured elevations of the MHS sites (red dots). Red lines show interpreted
deformed MHS. (D) Schematic structural cross-section (WE) of the Turkana Basin between the overow sill area and the eastern half of the rift axis area, showing the deformation
pattern affecting the elevation of the MHS. Approximate fault locations delineating horst and graben geometries are from Morley et al. (1992). Note the strong vertical exaggeration

of lacustrine geomorphic markers (~0.31.5 m). In order to match our

measurements with the Shuttle Radar Topography Mission 3 arc second reference frame (SRTM V4; Farr et al., 2007) we added the computed height separations (EGM96 geoidWGS84 ellipsoid) to the
dGPS elevations.
Since differential elevations of coeval palaeo-shorelines may provide clues on deformation patterns in the tectonically active settings
of the EARS (e.g., Stein et al., 1991), we also surveyed the elevation
of the highest and most prominent shoreline at several sites (including South Island) along the southern edge of the Turkana Basin
(Figs. 2 and 3). Detailed results from this complementary survey are
presented in a companion paper focusing on Holocene extensional
processes and hydro-isostasy in the Turkana and Suguta basins
(Melnick et al., this volume).
In addition to the dGPS survey of the southernmost margin of Lake
Turkana, we also surveyed palaeo-shoreline deposits from its western
margin a low-relief area using a handheld GPS (Fig. 1B). As the elevations obtained from this secondary survey were not precise we
assigned an elevation derived from the SRTM dataset to each sample
3.2. Age control
In order to estimate the possible radiocarbon-age reservoir of Lake
Turkana we obtained two accelerator mass spectrometry (AMS) 14C

dates on modern Melanoides tuberculata shells provided by Drs. H.

Scholz and M. Glaubrecht from the Museum fr Naturkunde at Humboldt University, Berlin. Modern samples (including a live snail) were
collected on August 4, 2006, along the northern shore of the Koobi
Fora spit (Fig. 1B).
Nine AMS 14C dates were obtained from fossil carbonate shells of
freshwater organisms collected on South Island: these were mostly
snails but also included an oyster and a bivalve (Table 1). The shells
sampled were from the most distinctive beach ridges (Fig. 4D). To
avoid any sampling bias such as that introduced by the presence
of reworked shells derived from higher-elevation shorelines remobilized during sheetwash erosion/deposition events, the shells
were collected from pits excavated to ~ 40 cm depth in the beach
gravels (Fig. 4F). In nearly all of the sites sampled a distinct layer
of rounded gravels (~ 1020 cm thick), identical to deposits in the
present-day lakeshore environment and intercalated with pyroclastic-fallout deposits, testied to the existence of a former lake shoreline. A snail and an encrusting oyster were also collected from small
cavities (1020 cm deep) on the surface of a bedrock abrasion platform associated with the highest palaeo-shoreline elevation
(Fig. 4B).
A further 12 conventional 14C dates were obtained from oysters
and snails, as well as one from a stromatolite, collected along the
western margin of Lake Turkana including, from south to north, the
Lothagam, Eliye Springs, Kataboi, and Nachukui areas (Fig. 1B).

Author's personal copy


Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334

Fig. 4. South Island survey area. (A) Field view to the south of the maximum highstand shoreline (MHS), which is marked by a distinctive wave-cut notch. (B) SPOT satellite image
of the area surveyed in A. Also shown are dGPS tracks, sampling sites, faults, and the position of the MHS (yellow line). (C) Oblique aerial view to the southwest of one of the main
survey areas, showing prominent fossil beach ridges. (D) SPOT satellite image of the area surveyed in C. (E) Composite topographic prole of the surveyed shorelines with projected
position of the dated samples. (F) Representative prole of sample pit in fossil beach ridges. (G) Simplied map of South Island.

These lacustrine carbonate shells were buried in coarse deposits such

as sands, gravels, and pebbles, characteristic of high-energy shoreline
to nearshore environments.
Since the sampled fossil shells were composed of aragonite, which
is a metastable mineral, they were investigated by X-ray diffraction to
ensure that recrystallisation to diagenetic calcite had not occurred:
the shells used in our analysis were all pure aragonite. AMS ages
were measured at the Leibniz Laboratory for Radiometric Dating
and Isotope Research in Kiel (Germany), while conventional ages
were determined by liquid scintillation at the LODYC in Paris (France)
(Table 1). Specimens with no dissolution or local abrasion were selected for dating. The specimens were rst cleaned using 30% H2O2
in an ultrasonic bath, rinsed with distilled water, and dried, in order
to remove adhering dust and detrital carbonate as well as any organic
surface coating. This was followed by a second cleaning step using
15% H2O2 in an ultrasonic bath. Dates were calibrated using the

OxCal 4.1 programme (Bronk Ramsey, 2001) with the IntCal09

curve (Reimer et al., 2009).
3.3. Stable isotope analysis ( 18O and 13C) of mollusk shells
A small part (150500 g) of each shell used for AMS radiocarbon
dating was further processed for stable isotope analysis. The samples
were cleaned with deionised water, dried at 40 C, and then crushed.
Oxygen and carbon isotope ratios were measured with a Finnigan
MAT 251 mass spectrometer at the Leibniz Laboratory of Kiel University. The system is coupled online with a Carbo-Kiel Device (Type I).
Samples were reacted by individual acid addition (99% H3PO4 at
73 C). Standard external error was better than 0.07 and
0.05 for 18O and 13C, respectively, according to the performance of international and laboratory internal carbonate standard
materials. Results were calibrated against the carbonate isotope

Author's personal copy

Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334


Table 1
Radiocarbon ages of lacustrine carbonates from Lake Turkana.
Sample ID


(m asl)

Corrected elevation
(m asl)b




C age
(yr BP)

Cal. age
(cal. yr BP)d

Cal. age range BP


Cal. age range BP


KIA 36856e
KIA 36857e
KIA 36858e
KIA 36859e
KIA 36860e
KIA 36861e
KIA 36862e
KIA 36864e
KIA 36865e
KIA 38214e
KIA 38215e
Pa 2226i
Pa 2227i
Pa 2224i
Pa 2223i
TOP 02/11i
LAPS 02/14i
Pa 2230i
Pa 2231i
Pa 2225i
Pa 2228i
Pa 2229i
Pa 2232i



~ 454
~ 459
~ 438
~ 427
~ 418
~ 410
~ 405
~ 414
~ 393




8180 45
9740 50
4645 35
4330 30
4680 35
4370 35
4910 35
10,025 55
4695 35
9795 100
6285 50
6630 45
9515 80
5810 70
5225 80
4810 30
4790 30
5320 35
4965 40
5260 40
5255 30

AD 1996
AD 2001





SI = South Island; KF = Koobi Fora; LO = Lothagam; ES = Eliye Springs; KA = Kataboi; NK = Nachukui.

Elevations have been corrected for local tectonic subsidence (see main text).
M = Melanoides tuberculata; E = Etheria elliptica; B = bivalve; mM = modern Melanoides tuberculata; O = oyster; S = snail; St = stromatolite.
Calibrated radiocarbon ages given before 1950, except for the modern samples given in year Anno Domini (AD); 14C calibration method: programme OxCal 4.1 (Bronk Ramsey,
2001) with the IntCal09 curve (Reimer et al., 2009).
AMS ages.
Measured using dGPS.
Estimated from SRTM V4 data.
Expressed as % modern (%M).
Conventional ages.

standard NBS 19 and are reported on the Vienna PeeDee Belemnite

(VPDB) scale.

overowing (cf. Butzer et al., 1972; Harvey and Grove, 1982; Owen
et al., 1982), with this overow resulting in the stabilisation of the
lake's water-level over a protracted period of time.

4. Results: the palaeo-shoreline record of Lake Turkana

Series of palaeo-shorelines are ubiquitous on the anks of the Turkana Basin (Fig. 1B). They form a staircase morphology of successive
former lake water-levels, well exposed over a vertical range of
~ 100 m, from the present-day lake water-level at ~ 360 m to an elevation of ~ 455 m.
4.1. Denition of the maximum highstand shoreline
Observations from our eld survey, complemented by the analysis
of SPOT satellite images and SRTM data, indicated that the highest
shoreline is a continuous geomorphic marker spanning the entire
lake basin. Importantly, this distinct shoreline is always much wider
and better developed than the lower-elevation shorelines (Fig. 2).
We related this shoreline to the maximum highstand shoreline
(MHS) of Lake Turkana. The MHS either forms wave-cut notches
sculpted into steep-sided cinder cones (Fig. 2), or marked, gently inclined wave-cut platforms. At the northern margin of Lake Turkana
the MHS is characterised by prominent beach ridges up to 50 km
long and 700 m wide (cf. Harvey and Grove, 1982) (Fig. 1B). The
ridges there occur up to about 450 or 455 m (SRTM data), only a
few metres below the present-day elevation of the overow sill,
50 km to the southwest. This overow sill, which controlled lake
highstands in the past, has an elevation between ~ 457 and ~460 m
(SRTM data); it is an integral part of the barrier that separates the
Turkana and Nile basins (cf. Brown and Fuller, 2008). The virtually
identical elevations of the sill and the MHS along the northern margin
of the lake suggest that the MHS developed when Lake Turkana was

4.2. Inuence of tectonic deformation on shoreline elevation

Our detailed dGPS survey of the MHS in the southernmost part of
the Turkana Basin (Melnick et al., this volume), which included South
Island as well as seven additional sites distributed both on the rift
margins and along the rift axis (Fig. 3), reveals variations of up to
~9 m in the absolute elevation of the MHS. The measured elevations
ranged from ~10 to ~ 20 m below the present-day elevation of the
overow sill (Fig. 3C). The difference in elevation between these geomorphic features, which would originally have been located at the
same elevation, is a result of subsequent tectonic deformation of the
lake basin (Melnick et al., this volume). Field observations and mapping using satellite images and digital elevation models show further
evidence for recent tectonic deformation in the area, such as the presence of numerous fault scarps (Fig. 3B) affecting Holocene lacustrine
sediments, alluvial fans, shorelines, and river channels. Offshore and
onshore seismic reection proles in the area also indicate ubiquitous
recent normal faulting (Dunkelman et al., 1989; Johnson et al., 1987;
Vtel et al., 2004).
The Turkana Basin has thus remained tectonically active up to the
present, which may be the most viable explanation for the basin-wide
disparities in surveyed elevations of coeval palaeo-shorelines.
According to the regional structural model developed by Morley et
al. (1992), extension has led to subsidence along the inner rift and simultaneous exural uplift of the footwall (Fig. 3D). Such a model reconciles inconsistencies in the elevation of the overow sill with
respect to the MHS elsewhere.

Author's personal copy


Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334

4.3. Reliability of the radiocarbon ages on lacustrine carbonates

Age (cal. yr BP)







2.5 m

te: ~
ce ra


KIA 36857

Overflow sill

KIA 36856


KIA 36858

KIA 36859

KIA 36860
KIA 36864
KIA 36861


4.4. South Island: a detailed record of Holocene shorelines

The palaeo-shorelines on South Island are well developed geomorphic features (Fig. 4). They form a set of 10 major beach ridges
(up to 60 m wide and 80 cm high), each containing additional smaller-scale and lower-amplitude ridges (Fig. 4CE). The palaeo-shoreline elevations range from 360 m (the present-day lake shoreline)
to 437.5 m (Fig. 4E). The highest palaeo-shoreline, which corresponds
to the MHS, forms a distinctive wave-cut notch sculpted into volcanic
bedrock (Fig. 4A). This shoreline can be condently traced around the
entire island, although on the eastern side it is covered by extensive
lava ows.
In order to account for the effect of tectonic subsidence on South
Island (see Section 4.2), which has affected the palaeo-shoreline elevations, we translated upward the elevation of the oldest dated sample associated with the formation of the MHS (sample KIA
36857 = 11,190 cal. yr BP) to the present-day elevation of the overow sill (~ 459 m). We obtained a subsidence rate of ~2.5 mm/yr
(Fig. 5A) and, assuming a steady subsidence rate during the Holocene,
we then applied this rate to the other dated shorelines to correct their
elevations (Table 1 and Fig. 5B). The shoreline elevations at South Island, corrected for tectonic subsidence, are systematically reported
The oldest shoreline deposits on South Island (11,500 cal. yr BP)
occur at an elevation of ~ 414 m (Fig. 5B). Although the subsequent
uctuating lake water-levels (see below) and the associated wave
erosional processes should have obliterated this relatively low-elevation shoreline, its particular location at the edge of a closed depression (cf. sample KIA 36864 in Fig. 4D), isolated from the main
palaeo-lake, may have contributed to its preservation. The highest
and most distinctive shoreline (the MHS) developed between
11,190 and 9085 cal. yr BP. The absence of data between ~ 9000 and
~ 6000 cal. yr BP precludes any inference concerning the palaeo-lake




Corrected elevation (m asl)

Charcoal and other organic material is generally not preserved in

the palaeo-shorelines of Lake Turkana and radiocarbon dating of
these geomorphic features relies on lacustrine carbonates. In order
to assess the reliability of the radiocarbon ages obtained from these
deposits we dated two modern mollusk shells collected from the present-day lake shore and estimated the potential radiocarbon age
The two samples contained 14C originating from post-second
World War atomic-bomb tests and thus carbon younger than year
AD 1954 (Table 1). The 14C concentration of 111.2% modern carbon
(%M) measured in KIA 38214 (live snail in AD 2006) corresponded
to that in the atmosphere in AD 1996, while the 14C concentration
of 109.1%M measured in KIA 38215 corresponded to that in the atmosphere in AD 2001 (Levin and Kromer, 2004). The 14C concentration
in the living snail does not correspond exactly to that of the atmosphere when it was collected (~ 10-year difference). As Melanoides
has a short life span, ranging from a few months to a few years
(Leng et al., 1999), the slightly higher 14C concentrations in the living
snail may indicate that this organism was feeding on older organic
matter (i.e., with a higher 14C concentration).
Overall, our new data unambiguously document that the radiocarbon age reservoir in present-day Lake Turkana is negligible. These observations are consistent with radiocarbon dating of sediment-cores
obtained from the southern part of the lake. Indeed, Halfman et al.
(1994) found that coarse Holocene ostracod shells, which were
formed in situ, probably provide a true age for the deep sediments
of Lake Turkana. On the basis of our own observations we have assumed that all the dated fossil shells collected on palaeo-shorelines
had not been affected by the uptake of old carbon, although a radiocarbon age reservoir may have existed in the past.

Subsidence (m)


KIA 36862


KIA 36865






Age (cal. yr BP)

Fig. 5. Dated palaeo-shorelines from South Island. (A) Estimated local tectonic subsidence rate on South Island. To correct for the effect of tectonic subsidence affecting
palaeo-shoreline elevations, the oldest sample related to the MHS (i.e., KIA 36857:
white star) was translated upward to the level of the overow sill. The obtained subsidence rate (~ 2.5 mm/yr) was assumed to have been constant and was used to correct
the elevation of the other samples. (B) Radiocarbon ages of shorelines shown with
probability curves and with corrected elevations. Age error bars represent the 99% condence interval. Solid green line represents inferred main lake water-levels. Dashed
green line represents uncertain lake water-levels.

water-level during this period of time. Subsequently, within a short

time interval between 5625 and 4865 cal. yr BP, we identied a
suite of distinct shorelines covering a vertical elevation range of
~50 m (from ~438 to ~ 393 m). These shorelines record the last
major lake regression. Shoreline radiocarbon ages associated with
this regression scatter signicantly and include several age reversals
(Fig. 5B), which may suggest reworking by wave-action of older
shells from former shorelines in a high-energy lakeshore environment (cf. McGlue et al., 2010). However, the 18O values from these
dated shells argue against substantial reworking (Table 2 and
Fig. 6). The 18O values of lacustrine carbonates are generally controlled by changes in lake-water isotope composition, which may
sometimes be related to lake water-level uctuations, especially for

Author's personal copy

Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334


Table 2
Stable isotope values (18O and 13C) of freshwater mollusk shells from Lake Turkana.
Sample ID


Elevation (m asl)

Corrected elevation (m asl)b

18O ()c

13C ()c




~ 454
~ 459
~ 438
~ 427
~ 418
~ 410
~ 405
~ 414
~ 393

0.21 0.03
0.56 0.02
2.70 0.03
3.03 0.02
2.39 0.02
1.89 0.02
3.02 0.01
2.15 0.01
2.91 0.01
3.67 0.01
3.85 0.02

0.02 0.01
4.89 0.01
3.13 0.01
1.19 0.01
2.07 0.01
4.21 0.01
5.55 0.00
4.75 0.01
2.4 0.01
2.75 0.01
1.1 0.01



SI = South Island, KF = Koobi Fora.

Elevations have been corrected for local tectonic subsidence.
Values are reported on the VPDB scale as measured (aragonite), not corrected to calcite values.

closed-lake basins (e.g., Leng et al., 2006; Ricketts and Johnson, 1996).
The observed changes in 18O values from freshwater shells, which
ranged from 0.56 to +3.85 (including the modern shell data),
follow the changes in lake water-level reasonably well (Fig. 6), with
lighter 18O values associated with the higher water-levels (larger
lake volumes), and heavier 18O values associated with the lower
water-levels (smaller lake volumes). Moreover, since the largest accumulations of shells on South Island were found to be closely associated with the MHS, older shells with lighter 18O values should be present
in the lower elevation shorelines if signicant reworking had occurred.
From our observations we estimate a ~50 m fall in the lake's water-level
at ~5270 300 cal. yr BP, which probably lasted for a few centuries.
Finally, since the beach ridges on South Island are in pristine condition
the lake water-level must have remained lower than ~380 m after
~4800 cal. yr BP, since subsequent lake transgressions would have signicantly overprinted or obliterated the earlier ridges.
4.5. Synthesis of Holocene lake water-level uctuations in the Turkana
4.5.1. Age control and elevation uncertainties across the Turkana Basin
By combining our new palaeo-shoreline ages from South Island
with 12 ages from the western margin of Lake Turkana (Table 1)
and previously published ages from the greater Turkana Basin, we
are able to re-evaluate and improve the lake water-level curve at a
basin scale over the Holocene period. A total of 102 radiocarbon
ages have been obtained for this lake basin (Fig. 7) during a variety

Lake volume (km3)




early Holocene



Corrected elevation (m asl)











18O ()
Fig. 6. 18O values from freshwater mollusk shells vs. corrected elevations. Error bars
represent the 95% condence interval. Also shown is the depth-volume curve for
Lake Turkana (thick grey line).

of palaeo-environmental and archaeological investigations (e.g.,

Hildebrand et al., 2011; Owen et al., 1982; Robbins, 1972), which
are referenced in the Supplementary materials.
In order to relate those previously reported ages from the Turkana
Basin to our new ages we re-calculated several of the published ages
(cf. Brown and Fuller, 2008). This was necessary because Butzer et al.
(1969) had assumed that the CO2 in Lake Turkana was not completely
equilibrated with the atmosphere, leading to erroneous, older ages.
Butzer and Thurber (1969), Butzer et al. (1969, 1972), and Owen et
al. (1982) consequently applied an arbitrary correction of
400 years to all of their shell ages in order to correct for this effect.
However, since there is no signicant radiocarbon-age reservoir at
Lake Turkana, we added back the 400 years to each of the radiocarbon
ages published by these authors (exclusively on mollusk shells) and
calibrated them, as well as the other available radiocarbon ages,
using the same method as described in Section 3.2.
Our new age compilation (Fig. 7A) exhibits substantial variability
in the corresponding elevations for the dated samples. While lacustrine deposits may provide a precise position for the former lake
shore, archaeological data from settlements or temporary shoreline
occupations can only furnish approximations for the maximum
water-level of the palaeo-lake. Considering only those ages associated
with lacustrine deposits, the large range of estimated elevations for
one particular time could have a variety of different explanations.
For example, the fossil mollusks sampled from lake sediments may
derive from different water-depths, ranging from shallower nearshore to deeper offshore environments (Cohen, 1986), and may
therefore not be able to constrain water-depth estimates. In addition,
wave action in exposed areas may have prevented the deposition of
sediments above a certain water depth (Johnson et al., 1987) and/or
may have resulted in the redistribution of shells from shallow to deeper water (McGlue et al., 2010), thus adding noise to the original
water-level signal. Field and analytical methods used to estimate
sample elevations (e.g., handheld GPS, aneroid barometer, dumpy
level, and dGPS) may have also introduced vertical uncertainties in
the dataset since the precision and accuracy of each of these methods
vary signicantly. The ongoing tectonic deformation of the Turkana
Basin may also be responsible for a signicant part of the observed
age scatter (see Section 4.2). In view of the lack of precise dGPS measurements for the previously dated samples from the Turkana Basin,
an elevation correction for vertical tectonic displacements is not possible. Thus, in order to account for all of the issues discussed above
that may have affected the bathymetric information for the published
age data points, we assigned an arbitrary vertical uncertainty of
10 m to each of these dated samples (Fig. 7A).
4.5.2. Lake-level history
During the Last Glacial Period the lake water-level in the Turkana
Basin was probably lower than at present (Johnson, 1996). Following

Author's personal copy




50 km





Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334






Exposed land
after the midHolocene lake


el R .


Radiocarbon ages




AD 2008 lake-level










Age (cal. yr BP)




Phase I



Phase II


360 m

Phase III


390 m


Former lake
(Butzer an
d Thurber, l reconstruction
1969; Owen
et al., 1982



Elevation (m asl)

460 m

Omo R

Overflow sill

n = 102




Age (cal. yr BP)




No. of radiocarbon ages

Fig. 7. (A) Proposed lake water-level reconstruction for Lake Turkana, based on original and previously published data. Raw data and references are available in the Supplementary
materials. The lake water-level curve is a compilation of dated lacustrine deposits (radiocarbon ages mostly on mollusk shells) complemented by dated archaeological excavations
(radiocarbon ages mostly on bones, ostrich egg beads, and charcoal). Age error bars represent the 99% condence interval. All elevation uncertainties were arbitrarily xed to
10 m, except for the South Island samples which were further corrected for tectonic subsidence. Green curve represents inferred lake water-levels. Curve is dashed where uncertainties are greater. Lake highstand periods (Phases IIII) were dened by Owen et al. (1982). (B) Location map for the dated lake water-level markers used in A. The lower
and right side subplots show the distribution of the radiocarbon ages (calibrated) corresponding to longitude and latitude, respectively. Colour coding refers to the sampled material (see A). Histogram of radiocarbon age distribution is shown in the lower-right corner; n is the total number of radiocarbon ages.

this dry period, rare dated lacustrine deposits indicate a lake waterlevel reaching elevations of ~440 to ~410 m between 13,000 and
11,500 cal. yr BP. However, the lake water-level curve remains poorly
constrained for this period.
From ~ 11,500 cal. yr BP palaeo-shoreline deposits indicate that
the lake's water-level rose rapidly to its overow level, and then
remained relatively stable until ~ 8500 cal. yr BP. This lake highstand
was possibly interrupted by a brief water-level fall of ~ 15 m that
probably occurred at ~ 10,200 cal. yr BP, if evidence from prehistoric
near-shore settlements is taken into account. The protracted highstand period between ~11,500 and ~8500 cal. yr BP corresponds to
the Phase I described by Owen et al. (1982) (Fig. 7A). During this period there was uvial connectivity between Lake Turkana and the Nile
Basin, and the MHS was formed.
From 8500 cal. yr BP lacustrine deposits indicate a signicant but
short-lived lake regression (Owen et al., 1982), which may have
taken place over a few centuries. The magnitude of this regression
may have reached between ~ 50 and ~ 100 m. The subsequent transgression was equally abrupt, with the lake water-level rising again
to high levels (i.e.,~445 m) at ~ 7500 cal. yr BP.
From ~7500 to ~ 5300 cal. yr BP the presence of lacustrine deposits
indicates that the lake remained relatively stable at high levels. This
highstand period corresponds to the Phase II of Owen et al. (1982).
While Lake Turkana may have again overowed during this period,
the available data are too scarce to be certain.

From ~5300 cal. yr BP the lake's water-level fell to an elevation of

~390 m. Shoreline data from South Island suggest that this regression
may have occurred abruptly, possibly within a few centuries. Additional support for a rapid regression comes from the ages derived
from archaeological sites (Fig. 7A). It is interesting to note that during
this period the ages obtained from abandoned occupation sites follow
our inferred fall in the lake's water-level, indicating that humans adjusted their settlement locations to keep pace with the receding
Following this nal regression the lake's water-level must have
remained low, particularly between ~ 4000 and 2000 cal. yr BP
when the absence of any exposed lacustrine deposits may indicate
that the water-level was even lower than at present.
Owen et al. (1982) tentatively inferred a lake highstand period
from ~ 4500 to ~3000 cal. yr BP (Phase III). However, this last highstand phase was poorly constrained as it relied on a single shell
dated at 3870 cal. yr BP (cf. Butzer et al., 1969). Analyses of sediment
cores collected across Lake Turkana, including downcore changes in
sedimentation and accumulation rate (Barton and Torgersen, 1988;
Cerling, 1986), diatom assemblages (Halfman et al., 1992), and isotopic composition of ne-grained calcite (Ricketts and Johnson,
1996), have subsequently called into question the existence of a
late Holocene lake highstand. Our new palaeo-shoreline dataset also
appears to rule out a return to a higher lake water-level during
Phase III.

Author's personal copy

Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334

From ~ 2000 to ~750 cal. yr BP, large accumulations of bones of

aquatic fauna (shes, crocodiles, and turtles) recovered from prehistoric settlements located at elevations between 370 and 380 m suggest that the water in Lake Turkana rose to slightly higher levels
(Lynch and Robbins, 1979) before falling back to its present level
(i.e., ~ 360 m).
5. Discussion
5.1. Mid-Holocene hydrology of the Turkana Basin and its relationship to
other sites in tropical Africa
The mid-Holocene ~ 50 m fall in the lake's water-level (Fig. 7), together with the 18O-enrichment (+4) of the 18O values of lacustrine carbonates from the early Holocene to the present-day (Fig. 6),
suggests that Lake Turkana experienced a major decrease in the
local hydrological balance (lower precipitation/higher evaporation)
as well as possible changes in the air-mass source areas, which affected the isotopic composition of water input into the lake from both
rivers and precipitation (cf. Levin et al., 2009).
The contribution of environmental/climatic variables (i.e., evaporation and precipitation) responsible for the mid-Holocene lake
water-level fall and resultant changes in the 18O values of lacustrine
carbonates can be derived from isotope-hydrology modelling approaches (cf. Ricketts and Anderson, 1998; Ricketts and Johnson,
1996). However, during the early Holocene the water ow both in
and out of Lake Turkana differed greatly from today. As mentioned
above, Lake Turkana overowed toward the Nile Basin during the
AHP (Butzer, 1980; Butzer et al., 1972; Harvey and Grove, 1982;
Owen and Renaut, 1986; Owen et al., 1982) and in turn received inow from the linked Bogoria, Baringo, and Suguta basins to the
south through the overtopping of the sill on the western margin of
the Suguta trough (Fig. 1) (Garcin et al., 2009; Renaut and Owen,
1980; Tiercelin and Vincens, 1987; Truckle, 1976), as well as from
the Abaya, Chamo, and Chew Bahir basins to the east (Grove et al.,
1975). These two drainage systems, which are today structurally isolated to form closed basins (Owen and Renaut, 1986), increased the
size of the Turkana Basin by ~ 40% during times of uvial connectivity
(Fig. 1). The accurate parameterisation of any isotope-based waterbalance model therefore remains elusive.
Since the catchment area of Lake Turkana covers a latitudinal
range of ~ 9, the hydrological/climatic changes recorded during the
Holocene and earlier periods must have been of regional signicance.
Our previous study of the adjacent Suguta Valley (Garcin et al., 2009)
revealed a much earlier lake highstand period (from ~17 to ~ 8.5 cal.
yr BP) than that for Lake Turkana. Although differences in local insolation could possibly be responsible for the observed temporal variations of lake highstands within the EARS, a radiocarbon age reservoir
might also have existed for the Suguta Valley and other lakes (e.g.,
Junginger, 2011). However, based on our analysis of present-day mollusk material we do not consider this to be a problem for the Lake
Turkana record.
The lake water-level curve for Lake Turkana exhibits marked similarities with other local and regional proxy-data from tropical Africa,
both in terms of the magnitude and timing of the observed changes
(Fig. 8). For example, the lake water-level curve for Lake Turkana covaries with leaf wax D data a proxy for terrestrial hydrologic conditions from Lake Tanganyika which is located ~ 1000 km to the
southwest (Fig. 8B, C). The isotopic record from this site suggests a
period of high precipitation during the early Holocene, which ended
abruptly at ~ 4700 cal. yr BP (Tierney et al., 2008). Virtually identical
changes in local hydrology have been recorded in other lakes in the
EARS, such as the ZiwayShalla (Ethiopia) and Abh (Djibouti)
lakes, where a major Holocene highstand lasting until ~ 5000 cal. yr
BP has been documented (Gasse, 1977; Gasse and Street, 1978;
Gillespie et al., 1983). A marine record of eolian dust off Cap Blanc


(Mauritania, Fig. 8D) also shares analogous patterns with the Lake
Turkana record, despite being situated ~ 6000 km to the west. Low
dust ux into the Atlantic Ocean during the early Holocene was inferred from this site, at which time the Sahara experienced a relatively wet climate (Cole et al., 2009; deMenocal et al., 2000). Dust inux
increased abruptly, however, at ~ 5500 cal. yr BP following the establishment of arid conditions and a reduction in vegetation cover in the
Sahara and Sahel, which is compatible with our observations and may
suggest a climatic teleconnection between the two regions.
According to our lake water-level reconstruction the AHP termination recorded in the Turkana Basin was probably abrupt, although age
uncertainties preclude any precise estimation of its duration. In tropical Africa the exact nature of the transition from the wet early Holocene conditions during an enhanced monsoon toward a drier climate
after the mid-Holocene remains ambiguous. Palaeo-climate records
from the whole of Africa, as well as marine records, have provided
contrasting results concerning the AHP termination, ranging from
abrupt (e.g., Cole et al., 2009; deMenocal et al., 2000; Shanahan et
al., 2006; Tierney et al., 2008) to gradual and/or stepwise (e.g.,
Chase et al., 2010; Jung et al., 2004; Krpelin et al., 2008; Marshall
et al., 2011; Vincens et al., 2010). Climate-modelling experiments
have related an abrupt termination of the AHP to a nonlinear response of the African monsoon to changing insolation, possibly
caused by a strong positive vegetationclimate feedback (e.g.,
Claussen et al., 1999; Renssen et al., 2003). In contrast, Liu et al.
(2007) simulated a minimal vegetationclimate feedback and instead
proposed that the collapse of the northern African vegetation attributed to the AHP termination was in fact the nonlinear response of
the vegetation to the crossing of a precipitation threshold. Thus, despite recent advances and the increasing size of the database for
this important event, the available data are still ambiguous. Further
studies involving new climate-proxies are therefore required in
order to resolve the mechanisms responsible for the AHP termination.

5.2. Possible link between hydrological changes during the Holocene and
prehistoric occupation in the Turkana Basin
Archaeological excavations within the Turkana Basin have provided a rich, but complex history of human occupation, with major cultural steps during the Holocene (Fig. 8A). Abundant bone harpoon
heads have been recovered from dated sections spanning ~10,600
to ~ 5000 cal. yr BP in the immediate vicinity of the inferred palaeolake, generally in association with wavy-line or undecorated pottery
(e.g., Beyin, 2011; Brown, 1975; Butzer et al., 1969; Phillipson,
1977; Robbins, 1972, 1984, 2006; Sutton, 1977). At the time of the
protracted lake highstand this region was populated by huntergatherers who relied mainly on shing in their subsistence lifestyle (e.g.,
Phillipson, 1977; Robbins, 1972).
The bones of domestic cattle, sheep, and goats, often in association
with Nderit Ware pottery and occasionally with stone bowls, have
been found in deposits younger than ~ 5000 cal. yr BP within the Turkana Basin (e.g., Ashley et al., 2011; Barthelme, 1977, 1984; Marshall
et al., 1984; Owen et al., 1982). The cumulative probability distribution of the radiocarbon ages related to human occupation also
shows a marked maximum between ~5000 and ~ 4000 cal. yr BP
(Fig. 8A). Although foraging persisted and sh still formed an important part of their diet, this particular period highlights the emergence
of pastoral groups in the area (Ashley et al., 2011; Barthelme, 1984;
Hildebrand et al., 2011; Marshall et al., 1984; Ndiema et al., 2011).
The rst manifestations of herding around the palaeo-lake were apparently coeval with the construction of megalithic pillar sites in the
Turkana Basin (Hildebrand and Grillo, 2012; Hildebrand et al., 2011;
Lynch and Robbins, 1978; Lynch and Robbins, 1979; Nelson, 1995;
Soper and Lynch, 1977). Possible support for this cultural change
may come from linguistic evidence suggesting that southern

Author's personal copy


Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334

Cushitic-speakers, herders of domestic livestock, entered northern

Kenya ~5000 years ago (Ehret, 1974).
It is interesting to note that the main mid-Holocene cultural transition appears to be closely related to the timing of the documented

Turkana Basin cultural record



Early Afro-Asiatic and

? Nilo-Saharan speakers?


Bone settlements

? Iron




Wavy line


Pillar sites

Occupation phases

n = 50

Lake Turkana water-level







L. Tanganyika





East Atlantic

Terrigenous (%)







Age (cal. yr BP)

10,000 12,000

D leafwax ( VSMOW)

Elevation (m asl)

6. Conclusions



C ages (CPD)


~50 m fall in lake water-level in the Turkana Basin. A possible link between the lake regression and the emergence/expansion of pastoralism in the Turkana Basin may have been the exposure of ~10,000 km 2
of fertile silty-clay lacustrine sediments (Fig. 7B). The vast majority of
these newly exposed areas, which are characterised by relatively lowrelief topography, correspond to the palaeo-deltas of the Omo, Turkwel, and Kerio rivers. In this context Robbins (2006) hypothesised
that any signicant fall in lake water-level in the area would result
in the opening up of new pastures/browsing resources in a region
suspected to have been relatively free of tsetse ies and the trypanosomiasis disease that they carry (cf. Gifford-Gonzalez, 1998). The latter is a virulent infection affecting livestock, wild game, and humans,
and is typical of regions south of the Sahara and Sahel. It is more likely, however, that the establishment of a drier and drought-prone climate in northeast Africa after ~ 5300 cal. yr BP may have forced local
herders and their domesticated livestock to converge on Lake Turkana, which was probably one of the last permanent water bodies
in the region and would have provided sufcient water and pasture
to foster the settlement of pastoralists, ultimately leading to new
forms of social and economic organisations (e.g., Marshall et al.,
2011; Wright, 2011).
The changes in landscape and environmental conditions within
the greater Turkana Basin associated with the mid-Holocene climatic
change, and the associated fall in lake water-level, may thus have encouraged the regional expansion of pastoralist cultures.

Our survey of palaeo-shorelines from the Turkana Basin reveals

new insights into past hydrological changes in the EARS. We have
documented a prolonged lake highstand with episodes of uvial connectivity to the Nile Basin during the early to mid-Holocene, supporting previous regional hydrological reconstructions. The lake waterlevel subsequently fell rapidly by ~ 50 m at ~ 5270 300 cal. yr BP,
probably in response to a major climatic transition characterised by
a rapid change in environmental conditions. This episode corresponds
to the termination of the African Humid Period, which has often been
inferred to have occurred abruptly right across tropical Africa.
Our observations indicate that ongoing tectonic segmentation of
the Turkana rift-basin has resulted in the vertical displacement of
the maximum highstand shoreline (overow stage) by up to ~ 20 m,
relative to the present-day position of the overow sill of the Turkana
palaeo-lake. When using palaeo-shorelines for high-resolution lake
water-level reconstructions and palaeoclimate assessments in active
tectonic settings it is therefore important to decipher the imprint of
tectonic deformation at various spatial and temporal scales.
Finally, since the emergence/expansion of pastoralism in the Turkana Basin was coeval with the mid-Holocene lake water-level fall,
we propose that hydrological changes such as long-term droughts
Fig. 8. Comparison of the Turkana Basin environmental and archaeological history during
the Holocene with reconstructed climate-proxy data from other African sites. (A) Turkana
Basin cultural record. Compilation derived from various archaeological sources: linguistic
evidence (Ehret, 1974); harpoon distribution and pottery tradition (Ashley et al., 2011;
Barthelme, 1977, 1984; Beyin, 2011; Brown, 1975; Butzer et al., 1969; Hildebrand et al.,
2011; Lynch and Robbins, 1979; Nelson, 1991; Owen et al., 1982; Phillipson, 1977;
Robbins, 1972, 1984, 2006); pillar sites (Hildebrand and Grillo, 2012; Hildebrand et al.,
2011; Lynch and Robbins, 1978, 1979; Nelson, 1995; Soper and Lynch, 1977); pastoralism
evidence (Ashley et al., 2011; Barthelme, 1977, 1984; Marshall et al., 1984; Owen et al.,
1982). Also shown is the cumulative probability distribution (CPD) of the calibrated radiocarbon ages related to human occupation, which includes the ages derived from archaeological excavations as well as those from lacustrine deposits associated with cultural
remains (e.g., bone harpoons). (B) Lake Turkana water-level curve. (C) Variations in Dleafwax from Lake Tanganyika recording past changes in humidity (Tierney et al., 2008).
(D) Variations in terrigenous (eolian) sediment import off Cap Blanc (Mauritania)
recording past changes in aridity over the Saharan region (deMenocal et al., 2000). Blue
and beige shadings mark humid and dry conditions, respectively, which affected several
regions of tropical Africa (see main text).

Author's personal copy

Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334

may have fundamentally inuenced prehistoric cultural changes in

the EARS.
Supplementary materials related to this article can be found online at doi:10.1016/j.epsl.2012.03.016.
Funding has been provided by the German Research Foundation
(DFG) through projects GRK1364, TR419/6-1, and STR373/16-1. Y.
Garcin was also supported by an Alexander von Humboldt Research
Fellowship. J.-J. Tiercelin was supported by grants from the INSU
CNRSECLIPSE I & II Programs. We would like to thank the University
of Nairobi, the Government of Kenya (Research Permits OP/13/001/
23C 290 and MOEST 13/001/23C 290 to J.-J.T.), the National Oil Corporation of Kenya, Wild Frontiers, and Tropic Air, for administrative and
logistical support. We also thank M.H. Trauth and A. Junginger for
eld support. We are grateful to S. Nielsen for handling our samples
during radiocarbon dating and stable isotope analysis in Kiel, to J.-F.
Salige for his part in radiocarbon age determinations, to H. Scholz
and M. Glaubrecht for generously providing samples of modern
snails, to K. Grillo for useful discussions, and to three anonymous reviewers for detailed and constructive comments. SPOT satellite images (CNES) were acquired through the ISIS-156 Project.
Ashley, G.M., Ndiema, E.K., Spencer, J.Q.G., Harris, J.W.K., Kiura, P.W., 2011. Paleoenvironmental context of archaeological sites, implications for subsistence strategies
under Holocene climate change, northern Kenya. Geoarchaeology 26, 809837.
Baker, B.H., Mohr, P.A., Williams, L.A.J., 1972. Geology of the eastern rift system of Africa. Geol. Soc. Am. Spec. Pap. 136, 167.
Barthelme, J., 1977. Holocene sites north-east of Lake Turkana: a preliminary report.
Azania 12, 3341.
Barthelme, J.W., 1984. Early evidence for animal domestication in Eastern Africa. In:
Clark, J.D., Brandt, S.A. (Eds.), From Hunters to Farmers: The Causes and Consequences of Food Production in Africa. Univ. of California Press, Berkeley, pp.
Barton, C.E., Torgersen, T., 1988. Palaeomagnetic and 210Pb estimates of sedimentation
in Lake Turkana, East Africa. Palaeogeogr. Palaeoclimatol. Palaeoecol. 68, 5359.
Beyin, A., 2011. Recent archaeological survey and excavation around the Greater Kalokol Area, west side of Lake Turkana: preliminary ndings. Nyame Akuma 75,
Boko, M., Niang, I., Nyong, A., Vogel, C., Githeko, A., Medany, M., Osman-Elasha, B., Tabo,
R., Yanda, P., 2007. Africa. In: Parry, M.L., Canziani, O.F., Palutikof, J.P., van der
Linden, P.J., Hanson, C.E. (Eds.), Climate Change 2007: Impacts, Adaptation and
Vulnerability. : Contribution of Working Group II to the Fourth Assessment Report
of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, Cambridge UK, pp. 433467.
Bronk Ramsey, C., 2001. Development of the radiocarbon calibration program OxCal.
Radiocarbon 43, 355363.
Brooks, N., 2006. Cultural responses to aridity in the Middle Holocene and increased
social complexity. Quat. Int. 151, 2949.
Brown, F.H., 1975. Barbed bone points from the Lower Omo Valley, Ethiopia. Azania 10,
Brown, F.H., Carmichael, I.S.E., 1971. Quaternary volcanoes of the Lake Rudolf region: II.
The lavas of North Island, South Island and the Barrier. Lithos 4, 305323.
Brown, F.H., Fuller, C.R., 2008. Stratigraphy and tephra of the Kibish Formation, southwestern Ethiopia. J. Hum. Evol. 55, 366403.
Butzer, K.W., 1980. The Holocene lake plain of north Rudolf, East Africa. Phys. Geogr. 1,
Butzer, K.W., Thurber, D.L., 1969. Some Late Cenozoic sedimentary formations of the
Lower Omo Basin. Nature 222, 11381143.
Butzer, K.W., Brown, F.H., Thurber, D.L., 1969. Horizontal sediments of the lower Omo
valley: the Kibish Formation. Quaternaria 11, 1529.
Butzer, K.W., Isaac, G.L., Richardson, J.L., Washbourn-Kamau, C., 1972. Radiocarbon dating of east African lake levels. Science 175, 10691076.
Cerling, T.E., 1979. Paleochemistry of Plio-Pleistocene Lake Turkana, Kenya. Palaeogeogr. Palaeoclimatol. Palaeoecol. 27, 247285.
Cerling, T.E., 1986. A mass-balance approach to basin sedimentation: constraints on the
recent history of the Turkana basin. Palaeogeogr. Palaeoclimatol. Palaeoecol. 54,
Chase, B.M., Meadows, M.E., Carr, A.S., Reimer, P.J., 2010. Evidence for progressive Holocene aridication in southern Africa recorded in Namibian hyrax middens: implications for African monsoon dynamics and the African Humid Period. Quat. Res.
74, 3645.
Claussen, M., Kubatzki, C., Brovkin, V., Ganopolski, A., Hoelzmann, P., Pachur, H.J., 1999.
Simulation of an abrupt change in Saharan vegetation in the mid-Holocene. Geophys. Res. Lett. 26, 20372040.


Cohen, A.S., 1986. Distribution and faunal associations of benthic invertebrates at Lake
Turkana, Kenya. Hydrobiologia 141, 179197.
Cole, J.M., Goldstein, S.L., deMenocal, P.B., Hemming, S.R., Grousset, F.E., 2009. Contrasting compositions of Saharan dust in the eastern Atlantic Ocean during the last deglaciation and African Humid Period. Earth Planet. Sci. Lett. 278, 257266.
deMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Baker, L., Yarusinsky, M.,
2000. Abrupt onset and termination of the African Humid Period: rapid climate responses to gradual insolation forcing. Quat. Sci. Rev. 19, 347361.
Dunkelman, T.J., Rosendahl, B.R., Karson, J.A., 1989. Structure and stratigraphy of the
Turkana rift from seismic reection data. J. Afr. Earth Sci. 8, 489510.
East African Meteorological Department, 1975. Climatological statistics for East Africa.
Part 1: Kenya. East African Community/East African Meteorological Department,
Nairobi, Kenya. 92 pp.
Ehret, C., 1974. Ethiopians and East African: The Problem of Contacts. East African Pub.
House, Nairobi, Kenya.
Farr, T.G., Rosen, P.A., Caro, E., Crippen, R., Duren, R., Hensley, S., Kobrick, M., Paller, M.,
Rodriguez, E., Roth, L., Seal, D., Shaffer, S., Shimada, J., Umland, J., Werner, M., Oskin,
M., Burbank, D., Alsdorf, D., 2007. The shuttle radar topography mission. Rev. Geophys. 45, RG2004. doi:10.1029/2005RG000183.
Ferguson, A.J.D., Harbott, B.J., 1982. Geographical, physical, and chemical aspects of
Lake Turkana. In: Hopson, A.J. (Ed.), Lake Turkana: A Report of the Findings of
the Lake Turkana Project, 19721975. Overseas Development Administration, London, England, pp. 1110.
Garcin, Y., Junginger, A., Melnick, D., Olago, D.O., Strecker, M.R., Trauth, M.H., 2009. Late
PleistoceneHolocene rise and collapse of Lake Suguta, northern Kenya Rift. Quat.
Sci. Rev. 28, 911925.
Gasse, F., 1977. Evolution of Lake Abh (Ethiopia and TFAI), from 70,000 b.p. Nature
265, 4245.
Gasse, E., Street, F.A., 1978. Late Quaternary lake-level uctuations and environments
of the northern Rift Valley and Afar Region (Ethiopia and Djibouti). Palaeogeogr.
Palaeoclimatol. Palaeoecol. 24, 279325.
Gifford-Gonzalez, D., 1998. Early pastoralists in East Africa: ecological and social dimensions. J. Anthropol. Archaeol. 17, 166200.
Gillespie, R., Street-Perrott, F.A., Switsur, R., 1983. Post-glacial arid episodes in Ethiopia
have implications for climate prediction. Nature 306, 680683.
Grove, A.T., Street, F.A., Goudie, A.S., 1975. Former lake levels and climatic change in the
Rift Valley of Southern Ethiopia. Geogr. J. 141, 177194.
Halfman, J.D., Johnson, T.C., 1988. High-resolution record of cyclic climatic-change during the past 4 ka from Lake Turkana, Kenya. Geology 16, 496500.
Halfman, J.D., Jacobson, D.F., Cannella, C.M., Haberyan, K.A., Finney, B.P., 1992. Fossil diatoms and the mid to late Holocene paleolimnology of Lake Turkana, Kenya: a reconnaissance study. J. Paleolimnol. 7, 2335.
Halfman, J.D., Johnson, T.C., Finney, B.P., 1994. New AMS dates, stratigraphic correlations and decadal climatic cycles for the past 4 ka at Lake Turkana, Kenya. Palaeogeogr. Palaeoclimatol. Palaeoecol. 111, 8398.
Harvey, C.P.D., Grove, A.T., 1982. A prehistoric source of the Nile. Geogr. J. 148,
Hildebrand, E.A., Grillo, K.M., 2012. Early herders and monumental sites in eastern Africa: dating and interpretation. Antiquity 86, 338352.
Hildebrand, E.A., Shea, J.J., Grillo, K.M., 2011. Four middle Holocene pillar sites in West
Turkana, Kenya. J. Field Archaeol. 36, 181200.
Hoelzmann, P., Jolly, D., Harrison, S.P., Laarif, F., Bonnelle, R., Pachur, H.J., 1998. MidHolocene land-surface conditions in northern Africa and the Arabian peninsula: a
data set for the analysis of biogeophysical feedbacks in the climate system. Global
Biogeochem. Cycles 12, 3551.
Johnson, T.C., 1996. Sedimentary processes and signals of past climatic change in the
large lakes of the East African rift valley. In: Johnson, T.C., Odada, E.O. (Eds.), The
Limnology, Climatology and Paleoclimatology of the East African Lakes. Gordon
and Breach Publishers, Amsterdam, pp. 367412.
Johnson, T.C., Halfman, J.D., Rosendahl, B.R., Lister, G.S., 1987. Climatic and tectonic effects on sedimentation in a rift-valley lake: evidence from high-resolution seismic
proles, Lake Turkana, Kenya. Geol. Soc. Am. Bull. 98, 439447.
Jolly, D., Harrison, S.P., Damnati, B., Bonnelle, R., 1998. Simulated climate and biomes
of Africa during the Late Quaternary: comparison with pollen and lake status data.
Quat. Sci. Rev. 17, 629657.
Jung, S.J.A., Davies, G.R., Ganssen, G.M., Kroon, D., 2004. Stepwise Holocene aridication
in NE Africa deduced from dust-borne radiogenic isotope records. Earth Planet. Sci.
Lett. 221, 2737.
Junginger, A., 2011. East African climate variability on different time scales: the Suguta
Valley in the AfricanAsian monsoon domain. PhD Thesis, Univ. of Potsdam, Potsdam, Germany.
Karson, J.A., Curtis, P.C., 1994. Quaternary volcanic centers of the Turkana Rift, Kenya. J.
Afr. Earth Sci. 18, 1535.
Krpelin, S., Verschuren, D., Lzine, A.-M., Eggermont, H., Cocquyt, C., Francus, P., Cazet,
J.-P., Fagot, M., Rumes, B., Russell, J.M., Darius, F., Conley, D.J., Schuster, M., von
Suchodoletz, H., Engstrom, D.R., 2008. Climate-driven ecosystem succession in
the Sahara: the past 6000 years. Science 320, 765768.
Kuper, R., Krpelin, S., 2006. Climate-controlled holocene occupation in the Sahara:
motor of Africa's evolution. Science 313, 803807.
Leng, M.J., Lamb, A.L., Lamb, H.F., Telford, R.J., 1999. Palaeoclimatic implications of isotopic
data from modern and early Holocene shells of the freshwater snail Melanoides tuberculata, from lakes in the Ethiopian Rift Valley. J. Paleolimnol. 21, 97106.
Leng, M.J., Lamb, A.L., Heaton, T.H.E., Marshall, J.D., Wolfe, B.B., Jones, M.D., Holmes, J.A.,
Arrowsmith, C., 2006. Isotopes in lake sediments. In: Leng, M.J. (Ed.), Isotopes in
Palaeoenvironmental Research. : Developments in Paleoenvironmental Research.
Springer, Dordrecht, The Netherlands, pp. 147184.

Author's personal copy


Y. Garcin et al. / Earth and Planetary Science Letters 331-332 (2012) 322334

Levin, I., Kromer, B., 2004. The tropospheric 14CO2 level in mid-latitudes of the Northern Hemisphere (19592003). Radiocarbon 46, 12611272.
Levin, N.E., Zipser, E.J., Cerling, T.E., 2009. Isotopic composition of waters from Ethiopia
and Kenya: insights into moisture sources for eastern Africa. J. Geophys. Res. 114,
D23306. doi:10.1029/2009JD012166.
Liu, Z., Wang, Y., Gallimore, R., Gasse, F., Johnson, T., deMenocal, P., Adkins, J., Notaro,
M., Prenticer, I.C., Kutzbach, J., Jacob, R., Behling, P., Wang, L., Ong, E., 2007. Simulating the transient evolution and abrupt change of Northern Africa atmosphere
ocean-terrestrial ecosystem in the Holocene. Quat. Sci. Rev. 26, 18181837.
Lynch, B.M., Robbins, L.H., 1978. Namoratunga: the rst archeoastronomical evidence
in sub-Saharan Africa. Science 200, 766768.
Lynch, B.M., Robbins, L.H., 1979. Cushitic and Nilotic prehistory: new archaeological evidence from northwest Kenya. J. Afr. Hist. 20, 319328.
Marshall, F., Stewart, K., Barthelme, J., 1984. Early domestic stock at Dongodien in
Northern Kenya. Azania 19, 120127.
Marshall, M.H., Lamb, H.F., Huws, D., Davies, S.J., Bates, R., Bloemendal, J., Boyle, J., Leng,
M.J., Umer, M., Bryant, C., 2011. Late Pleistocene and Holocene drought events at
Lake Tana, the source of the Blue Nile. Global Planet. Change 78, 147161.
McGlue, M.M., Soreghan, M.J., Michel, E., Todd, J.A., Cohen, A.S., Mischler, J., O'Connell,
C.S., Castaeda, O.S., Hartwell, R.J., Lezzar, K.E., Nkotagu, H.H., 2010. Environmental
controls on shell-rich facies in tropical lacustrine rifts: a view from Lake Tanganyika's littoral. Palaios 25, 426438.
Melnick, D., Garcin, Y., Quinteros, J., Strecker, M.R., Olago, D.O., Tiercelin, J.-J., this volume. Steady rifting in northern Kenya inferred from deformed Holocene lake
shorelines of the Suguta and Turkana basins. Earth Planet. Sci. Lett.
Morley, C.K., Wescott, W.A., Stone, D.M., Harper, R.M., Wigger, S.T., Karanja, F.M., 1992.
Tectonic evolution of the northern Kenyan Rift. J. Geol. Soc. London 149, 333348.
Ndiema, K.E., Dillian, C.D., Braun, D.R., Harris, J.W.K., Kiura, P.W., 2011. Transport and
subsistence patterns at the transition to pastoralism, Koobi Fora, Kenya. Archaeometry 53, 10851098.
Nelson, C.M., 1991. Harpoon Evolution on the Spit (GaJi12) at Koobi Fora, Lake Turkana,
Kenya. Nyame Akuma 36, 1019.
Nelson, C.M., 1995. The work of the Koobi Fora eld school at the Jarigole Pillar site.
Kenya Past and Present 27, 4963.
Owen, R.B., Renaut, R.W., 1986. Sedimentology, stratigraphy and palaeoenvironments
of the Holocene Galana Boi Formation, NE Lake Turkana, Kenya. In: Frostick, L.E.
(Ed.), Sedimentation in the African Rifts. Geol. Soc. London Spec. Publ, London,
pp. 311322.
Owen, R.B., Barthelme, J.W., Renaut, R.W., Vincens, A., 1982. Palaeolimnology and archaeology of Holocene deposits north-east of Lake Turkana, Kenya. Nature 298,
Phillipson, D.W., 1977. Lowasera. Azania 12, 132.
Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey,
C., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P.,
Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac,
F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney,
C.S.M., van der Plicht, J., Weyhenmeyer, C.E., 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 050,000 years cal BP. Radiocarbon 51, 11111150.
Renaut, R.W., Owen, R.B., 1980. Late Quaternary uvio-lacustrine sedimentation and
lake levels in the Baringo basin, northern Kenya, Rift Valley. Rech. Gol. Afr. 5,

Renssen, H., Brovkin, V., Fichefet, T., Goosse, H., 2003. Holocene climate instability during the termination of the African Humid Period. Geophys. Res. Lett. 30, 1184.
Ricketts, R.D., Anderson, R.F., 1998. A direct comparison between the historical record
of lake level and the 18O signal in carbonate sediments from Lake Turkana, Kenya.
Limnol. Oceanogr. 43, 811822.
Ricketts, R.D., Johnson, T.C., 1996. Climate change in the Turkana basin as deduced from
a 4000 year long 18O record. Earth Planet. Sci. Lett. 142, 717.
Ritchie, J.C., Eyles, C.H., Haynes, C.V., 1985. Sediment and pollen evidence for an Early to
mid-Holocene humid period in the eastern Sahara. Nature 314, 352355.
Robbins, L.H., 1972. Archeology in the Turkana District, Kenya. Science 176, 359366.
Robbins, L.H., 1984. Late Prehistoric aquatic and pastoral adaptations west of Lake Turkana, Kenya. In: Desmond Clark, J., Brandt, S.A. (Eds.), From Hunters to Farmers:
The Causes and Consequences of Food Production in Africa. Univ. of California
Press, Berkeley, pp. 206211.
Robbins, L.H., 2006. Lake Turkana archaeology: the Holocene. Ethnohistory 53, 7193.
Shanahan, T.M., Overpeck, J.T., Wheeler, C.W., Beck, J.W., Pigati, J.S., Talbot, M.R., Scholz,
C.A., Peck, J., King, J.W., 2006. Paleoclimatic variations in West Africa from a record
of late Pleistocene and Holocene lake level stands of Lake Bosumtwi, Ghana.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 242, 287302.
Shanahan, T.M., Overpeck, J.T., Anchukaitis, K.J., Beck, J.W., Cole, J.E., Dettman, D.L.,
Peck, J.A., Scholz, C.A., King, J.W., 2009. Atlantic forcing of persistent drought in
West Africa. Science 324, 377380.
Soper, R., Lynch, M., 1977. The stone-circle graves at Ng'amoritung'a, Southern Turkana
District, Kenya. Azania 12, 193208.
Stein, R.S., Briole, P., Ruegg, J.-C., Tapponnier, P., Gasse, F., 1991. Contemporary, Holocene, and quaternary deformation of the Asal Rift, Djibouti: implications for the
mechanics of slow spreading ridges. J. Geophys. Res. 96, 2178921806.
Sutton, J.E.G., 1977. The African aqualithic. Antiquity 51, 2534.
Tiercelin, J.-J., Vincens, A., 1987. Le Demi-Graben de Baringo-Bogoria, Rift Gregory,
Kenya, 30,000 Ans d'Histoire Hydrologique et Sdimentaire. Bull. Centres de
Rech. Expl. Prod. Elf-Aquitaine, Pau, pp. 249540.
Tierney, J.E., Russell, J.M., Huang, Y., Sinninghe Damst, J.S., Hopmans, E.C., Cohen, A.S.,
2008. Northern hemisphere controls on tropical southeast African climate during
the past 60,000 years. Science 322, 252255.
Truckle, P.H., 1976. Geology and late Cainozoic lake sediments of the Suguta Trough,
Kenya. Nature 263, 380383.
Vtel, W., Le Gall, B., Johnson, T.C., 2004. Recent tectonics in the Turkana Rift (North
Kenya): an integrated approach from drainage network, satellite imagery and reection seismic analyses. Basin Res. 16, 165181.
Vincens, A., Buchet, G., Servant, M., 2010. ECOFIT Mbalang collaborators. Vegetation response to the African Humid Period termination in Central Cameroon (7N)
new pollen insight from Lake Mbalang: Clim. Past, 6, pp. 281294.
Wright, D.K., 2011. Frontier animal husbandry in the Northeast and East African Neolithic: a multiproxy paleoenvironmental and paleodemographic study. J. Anthropol. Res. 67, 213244.
Yuretich, R.F., Cerling, T.E., 1983. Hydrogeochemistry of Lake Turkana, Kenya: mass balance and mineral reactions in an alkaline lake. Geochim. Cosmochim. Acta 47,