Professional Documents
Culture Documents
633–658
Abstract: This paper presents an overview of paleoenvironmental changes in East Africa during the
late Quaternary based on evidence from pollen, diatoms, microscopic charcoal, and lake level records
and associated proxies. The paleoenvironmental records derived from different proxies complement
each other to provide a more accurate and complete assessment of the paleoenvironmental changes
in East Africa. The records show that the period prior to c. 42,000 14C yr BP was characterized by
warm climatic conditions similar to the present. This was followed by a change to cold dry conditions
from 42,000 to 30,000 14C yr BP, and cold and moist conditions from 30,000 to 21,000 14C yr BP.
Temperatures during the latter period leading to the Last Glacial Maximum (LGM) were probably 2
to 4.1°C lower than the present. Between c. 21,000 and 12,500 14C yr BP East Africa’s environment
was generally cool, punctuated by two significant episodes of prolonged desiccation. Warm and moist
conditions punctuated by rapid climatic changes prevailed in the region during the deglacial and middle
Holocene period. Ice core records document two significant and abrupt drought events in the region,
one at ~8300 14C yr BP and the other at 5200 14C yr BP. The onset of a longer and more extensive
desiccation period commencing ~4000 14C yr BP was registered in nearly all sites. The climate of East
Africa was generally drier than present during the Medieval Warm Period (MWP) while fairly wet
conditions prevailed during the Little Ice Age (LIA) interrupted by three episodes of aridity, more
severe than those of more recent times. Whereas this review advances our understanding of climate
and vegetational changes in East Africa beyond the Last Glacial Maximum, it also highlights limitations
of the paradigms that explain the forcing mechanisms behind the changes. However, unequivocal
interpretation of the multiproxy data from East Africa with respect to paleoenvironmental changes
becomes extremely complex and challenging especially when the anthropogenic input is considered.
records from lake sediments can provide paleohydrological evidences for lake level fluc-
detailed records of past paleoecological and tuations. Where appropriate, these records are
climatic changes (eg, Liu and Colinvaux, complemented by chemical and isotopic prox-
1988; Maley and Brenac, 1998; deMenocal ies. Although we try to cover the entire
et al., 2000; Gasse, 2000; Salzmann, 2000; breadth of the region we are limited by data
Salzmann et al., 2002; Peyron et al., gaps, especially from areas lacking suitable
2000; Wooller et al., 2000; Behling, 2002; archiving environments. Ages referred to in the
Darbyshire et al., 2003; Lamb et al., review are given in 14C years before present
2003; Behling et al., 2004). In addition to (BP), unless otherwise noted. Our review dif-
pollen, other proxies such as charcoal fers from those of Peyron et al. (2000),
(Burney, 1987a; 1987b; Long et al., 1998; Marchant and Hooghiemstra (2004), and
Whitlock and Millspaugh, 1996), phytoliths those in Battarbee et al. (2004), which, though
(Piperno, 1988; Mworia-Maitima, 1997; excellent, are either too focused on a particular
Piperno and Pearsall, 1998), diatoms (Gasse time period or too broad in regional coverage
et al., 1997; Barker et al., 2001), and grass to give the complete picture of paleoenviron-
cuticles (Mworia-Maitima, 1997; Wooller et mental changes in East Africa through the
al., 2000) have proven to be very useful in period reviewed herein.
paleoenvironmental reconstruction.
The use of multiproxy approaches
enables paleoecologists to more accurately
and completely reconstruct the paleoenvi-
ronment (Boyd and Hall, 1998). Since the
environment affects different proxies differ-
ently, the deficiencies in one proxy may be
compensated for by another. For instance,
prolonged droughts might compromise the
preservation of biological proxies such as
pollen due to the lowering of lake levels, yet
paleoenvironmental conditions that pre-
vailed during periods of low lake levels can
still be recorded through geochemical and/or
isotopic proxies such as total organic content
(TOC), hydrogen index (HI), 18O, 13C,
and 15N, among others.
We review different proxy evidences of cli-
matic changes in East Africa from beyond
42,000 yr BP to the present millennium.
Although we focus on the three East African
countries of Kenya, Uganda, and Tanzania
(Figure 1; Table 1), we include relevant infor-
mation from tropical Africa as a whole. The
region has considerable topographic relief, with
Figure 1 Map showing East Africa
stepped plateaus of 900–2000 m or more, and and the neighboring countries. The
high mountains such as Mounts Kenya, Elgon, approximate locations of some of the
Kilimanjaro, and the Ruwenzoris. We examine sites discussed in the review are also
biological proxy records such as pollen, shown in the dots numbered 1 to 21,
diatoms, and microscopic charcoal, as well as and cited in Table 1
Lawrence M. Kiage and Kam-biu Liu 635
Table 1 Sites discussed in the text and plotted in Figure 1, with references to
articles with original data and interpretations
Site Location Elevation (m) References
1 Lake Albert, Uganda 619 Beuning et al. (1997)
2 Ahakagyezi Swamp, Uganda 1830 Hamilton et al. (1986), Taylor (1990)
3 Muchoya Swamp, Uganda 2256 Hamilton (1982), Taylor (1990)
4 Lake Mahoma, (Ruwenzori 2960 Livingstone (1967)
Mountains), Uganda
5 Kamiranzovu, Rwanda 1950 Roche and Bikwemu (1989)
6 Kashiru, Burundi 2240 Bonnefille and Riollet (1988)
7 Lake Tanganyika, Tanzania 773 Vincens (1989a; 1989b; 1993)
8 Lake Rukwa, Tanzania 793 Butzer et al. (1972), Haberyan (1987),
Talbot and Livingstone (1989)
9 Lake Masoko, Tanzania 770 Vincens et al. (2003)
10a Mount Kilimanjaro, Tanzania 2650 Coetzee (1967)
10b Mount Kilimanjaro, Tanzania 5893 Thompson et al. (2002)
11 Lake Simbi, Kenya 1143 Mworia-Maitima (1997)
12 Pekington Bay, Lake Victoria, 1134 Talbot and Livingstone (1989),
Uganda Stager et al. (2003)
13 Lake Naivasha, Kenya 1890 Mworia-Maitima (1991), Verschuren
(1999), Verschuren et al. (2000)
14a Sacred Lake, Mount Kenya, 2400 Coetzee (1967), Olago (1995),
Kenya Street-Perrott and Perrott (1993),
Street-Perrott et al. (1997)
14b Simba Turn and Small Hill Turn 4595 Barker et al. (2001)
(Mount Kenya, Kenya)
14c Simba Turn and Small Hill Turn 4289 Barker et al. (2001)
(Mount Kenya, Kenya)
15 Lake Bogoria, Kenya 990 Tiercelin and Vincens (1987)
16 Loboi Swamp, Kenya 1080 Ashley et al. (2004)
17 Cherangani Hills, Kenya 2900 Bakker (1964)
18 Lake Kimilili, Mount Elgon, 4150 Hamilton and Perrott (1978; 1979)
Kenya
19 Laboot Swamp, Kenya 2880 Hamilton (1982)
20a, b Lake Turkana, Kenya 360 Bonnefille (1976),
Mohammed et al. (1995)
21 Lake Abiyata, Ethiopia 1820 Lézine (1982), Gasse et al. (1995),
Legesse et al. (2004)
highlands to less than 200 mm in the arid and mostly dry. The northeast monsoon, coming
semi-arid parts of Kenya (Figure 2) (cf. from the Arabian Desert, is a dry wind. The
Nicholson, 2000). The climate of the region is southeast monsoon, coming off the Indian
governed by air streams and convergence Ocean, constitutes the dominant air stream
zones that are linked to the seasonal shifts of over East Africa in July. Although deep and
the subtropical high pressure cells situated moist, it is not associated with much rainfall
about 20–30° north and south of the equator because the winds are divergent (McGregor
(McGregor and Nieuwolt, 1998; Nicholson, and Nieuwolt, 1998). However, the Congo
1996; 2000). The air streams include the air, which originates in the south Atlantic, is
Congo air with westerly and southwesterly humid, convergent, and thermally unstable. It
flow, the northeast monsoon (NE trade is associated with rainfall in the northwestern
winds), and the southeast monsoon (SE trade part of the region (Boucher, 1975; Hamilton,
winds) (Figure 3). The air streams are sepa- 1982; McGregor and Nieuwolt, 1998;
rated by two surface convergence zones, that Nicholson, 2000).
is, the Inter-tropical Convergence Zone The vegetational distribution pattern in
(ITCZ) and the Congo Air Boundary (CAB). East Africa (Figure 4) closely mimics the rain-
The former separates the two monsoons, fall and topographic conditions as well as the
the latter the easterlies and westerlies length of the dry season. Six major floristic
(Nicholson, 2000). regions can be delineated in East Africa (Lind
Synoptically, the climate of East Africa is and Morrison, 1974; White, 1983) (Figure 4).
dominated by the interplay between the (1) Lowland and montane forests occur in
northeast monsoon in January and the south- patches in Uganda and Kenya. These semi-
east monsoon in July. Unlike the southwest deciduous and evergreen forests are made up
monsoon of Asia, the monsoons that affect mainly of Celtis spp., Urticaceae, Myrtaceae,
East Africa are thermally stable, hence Croton, Holoptelea, Prunus, Podocarpus, Ilex,
Figure 2 Schematic of the general patterns of winds, pressure and convergence over
Africa during the months of January and July. The dotted lines indicate the inter-
tropical convergence zone (ITCZ), dashed lines, the Congo Air Boundary (CAB)
Source: Modified from Nicholson (2000).
Lawrence M. Kiage and Kam-biu Liu 637
Acacia, and Olea, among others. The (deciduous bushland and thicket), mainly con-
Afroalpine belt occurs between 1600 and sisting of Acacia and Lommiphora, are com-
3000 m and is characterized by Prunus mon in Kenya, Ethiopia, and Somalia. (3) The
africana, Olea, Hagenia abyssinica, Juniperus grasslands, which often occur in the same
procera, and Podocarpus. The Ericaceous belt ecological regions with deciduous bushland,
dominated by Philipia and Erica is found are not widespread in East Africa. (4) The
above 3000 m. (2) The drier savanna types wetter savanna types, mainly made of
638 Late Quaternary paleoenvironmental changes in East Africa
Figure 4 The spatial distribution of the major vegetation types of East Africa
Source: Modified from Lind and Morrison (1974).
Combretaceae and Acacia savannas, are woodlands (mainly miombo) occur. These
widely distributed in Uganda, Tanzania, and deciduous woodlands are basically a moist
Kenya (Figure 4). (5) In the southern parts of savanna type that is dominated by fairly tall
Tanzania, the wetter Zambezian deciduous (20 m) and densely spaced leguminous trees
Lawrence M. Kiage and Kam-biu Liu 639
belonging to such genera as Brachystegia, Swamp (Hamilton et al., 1986; Taylor, 1990),
Julbernardia, Isoberlinia, and Uapaca. (6) The and Laboot Swamp (Hamilton, 1982), among
semi-desert and desert vegetation occurs in others. However, most of the pollen records
northern Kenya and southern Ethiopia. only extend back to 30,000 14C yr BP
(Figure 5). Longer records, with a few
III Pollen records from East Africa exceptions, tend to be characterized by hia-
Investigations of East Africa’s paleoclimate tuses (Vincens, 1991; Bonnefille et al., 1995;
and paleoecology commenced in the 1960s Flenley, 1998; Elenga et al., 2000; Peyron et
largely due to pioneering studies by al., 2000; Wooller et al., 2000; Olago, 2001).
Livingstone (1962; 1967), Bakker (1964), The pollen records from this vast region
Coetzee (1964; 1967), Morrison (1968), and reveal a remarkably high diversity of both
Kendall (1969), and have continued to the paleoclimate and paleovegetation types even
present. Pollen analysis from the early studies over short distances, probably due to the
provided the first evidence to reveal that the presence of many mountainous areas (cf.
highlands of equatorial Africa, like other Peyron et al., 2000). Pollen analysis from the
parts of the world, were affected by glacia- different sites has illuminated the paleoenvi-
tion and that glacier retreat was roughly ronmental dynamics of the region beyond the
contemporaneous with those in Europe and Last Glacial Maximum (LGM).
North America (Hamilton, 1982). The
longest continuous records retrieved from 1 Period 42,000 to 30,000 14C yr BP
East Africa are from Sacred Lake, Mount The period prior to c. 42,000 14C yr BP is
Kenya (Coetzee, 1967; Olago, 2001; Street- characterized by warm climatic conditions
Perrott et al., 1997), covering the period from similar to those that presently abound in East
115,000 yr BP to the present. A record from Africa. During this period the vegetation in
two swamps (Muchoya and Ahakagyezi) in Ahakagyezi and Muchoya Swamps, Uganda,
the Rukiga Highlands, Uganda (Taylor, 1990), was dominated by moist lower montane for-
has a basal date of 42,000 yr BP while est, where Urticaceae, Myrtaceae, Croton,
those from Kashiru, Burundi (Bonnefille Minulopsis, Ilex, Macaranga, Neoboutonia,
and Riollet, 1988) and Cheshi, Zambia Nuxia, and Filcahoa occurred in the
(Stager, 1988) provide basal dates of Ahakagyezi and Muchoya catchment at the
c. 40,000 yr BP. same elevations as today (Figure 6) (Taylor,
The pollen records from the region have 1990). Contemporaneous records from
been drawn from diverse environments Sacred Lake, Mount Kenya (Coetzee, 1967;
(Figure 1; Table 1) including: small lakes on Olago, 2001) and Kamiranzovu Swamp
mountains and mountain ranges, eg, Mount (Hamilton, 1982) confirm similar trends.
Kilimanjaro (Coetzee, 1967), Mount Kenya Between 42,000 14C yr BP and 30,000 14C yr
(Coetzee, 1967; Olago, 2001; Street-Perrott BP cold dry conditions engulfed East Africa.
et al., 1997), Ruwenzori (Livingstone, 1967), Consequently, high-altitude vegetation belts
Cherangani (Bakker, 1964), and Mount Elgon descended to lower elevations. Evidence
(Hamilton and Perrott, 1978; 1979); large from charred grass cuticles from Sacred Lake,
lakes, eg, Lake Victoria (Kendall, 1969; Mount Kenya, shows a general increase in the
Stager, 1984; Stager et al., 1997), Lake Albert percentages of Bambusoid/Arundinoid-type
(Beuning et al., 1997), Lake Naivasha grasses as well as a Panicoid(Andropogonid)-
(Mworia-Maitima, 1991), Lake Turkana dominated assemblage usually associated
(Bonnefille, 1979; Mohammed et al., 1995), with mesophytic conditions. The source of
Lake Tanganyika (Vincens, 1989; 1993), and the fire that left the charred particles at the
Lake Rukwa (Butzer et al., 1972); and mires, Sacred Lake site is likely to have been natural
eg, Rukiga Swamp (Taylor, 1990), Ahakagyezi considering that humans had not established
640 Late Quaternary paleoenvironmental changes in East Africa
themselves in the lake catchment (Wooller Ericaceous belt to thrive. Similar observa-
et al., 2000). tions are made from Mount Kenya (Olago,
In the Rukiga Highlands, the vegeta- 2001; Street-Perrott and Perrott, 1993), Lake
tion that constitutes the Ericaceous belt Abiyata (Ethiopia) (Lézine, 1982), and
(eg, Anthospermum, Stoebe, Artemisia, and Kashiru Swamp (Burundi) (Bonnefille and
Cliffortia) typically at c. 3300 m today was Riollet, 1988). In Kashiru Swamp there was a
found in Muchoya Swamp at an elevation of decline in forest taxa and an increase in
2260 m above sea level (Figure 6) (Taylor, Gramineae (Bonnefille and Riollet, 1988;
1990). The present conditions in the Aucour et al., 1994), also suggesting cold, dry
Muchoya catchment suggest that it would conditions. The Mount Kenya record during
have required a much colder and drier envi- this period shows Hagenia dominated vegeta-
ronment than present, equivalent to an altitu- tion coexisting with the Ericaceous belt
dinal depression of up to 1000 m, for the components.
Lawrence M. Kiage and Kam-biu Liu 641
Figure 6 Fluctuations in % pollen and spores with depth in the Muchoya Swamp
core MC2
Source: Modified from Taylor (1990).
highlands the Ericaceous belt became 66 m respectively, reveal two distinct discon-
re-established in the Muchoya Swamp catch- tinuities that correspond to low lake levels
ment until c. 15,700 yr BP, which is coeval (Talbot and Livingstone, 1989). The physical
with a substantial reduction in temperature properties of Core Ibis 1 mark the first discon-
and precipitation in the region (Taylor, 1990). tinuity at 8.2 m (~11,710 120 14C yr BP) and
Lake level records from Lake Victoria and the second at 8.85 m (~15,390 220 14C yr
Lake Albert identify at least two separate BP) (Figure 7). The two episodes are concur-
episodes of extreme aridification during this rent with prominent layers of the littoral gas-
period (Talbot and Livingstone, 1989; Beuning tropod Bellamya unicolor, which is consistent
et al., 1997). with desiccation and oxidation of sediments
Two cores (Ibis 1 and Ibis 3) obtained from during low lake levels (cf. Stager et al., 1986;
Lake Victoria at different depths, 32 m and Talbot and Livingstone, 1989). Traces of
Figure 7 Summary of Ibis core 1 from Lake Victoria, showing changes in the
stratigraphy, especially sections indicative of significant desiccation periods
Source: Modified from Talbot and Livingstone (1989).
Lawrence M. Kiage and Kam-biu Liu 643
desiccation cracks and rootlets or burrows clay lined cracks are typical pedogenic fea-
extend c. 25 cm down into the mottled sedi- tures of terrestrial environments (Beuning
ments below the color change (Figure 7) sug- et al., 1997). Such features are associated
gesting exposure at these levels (cf. Stager with cotton clays, a soil type that develops on
et al., 1986). floodplains and lake margins in warm, season-
The Lake Albert record is based on a 9.2 m ally contrasted climates (cf. Blodgett, 1985).
core (core F) collected from 46 m below Further scrutiny by Beuning et al. (1997)
water level yielding a record for the last revealed that there were, in fact, two such
30,000 14C yr BP (Beuning et al., 1997). The levels in the Lake Albert stratigraphy (similar
core consists of predominantly dark grey to to core Ibis 1, Talbot and Livingstone, 1989),
black, diatomaceous mud with a thin sandy one at 7.25 to 7.9 m and the other at 6.6 to
zone at 6.6 m overlying drier, silty sediments 6.9 m (Figure 8) separated by a zone contain-
to the base of the core (9.2 m) (see details in ing diatoms, Botryococcus, ostracods, fish
Beuning et al., 1997). The key to the sedimen- fragments and rare pollen grains indicating
tation history of the core site is provided by more persistent aquatic conditions (Beuning
the sequence between 8.4 and 6.6 m where et al., 1997). Therefore, the lowstands must
low water contents and presence of rootlet have been separated by at least one period of
traces suggest terrestrial conditions. The lake transgression.
sequence is associated with calcite nodules The desiccation in East Africa must have
crumb structure, and vertical and oblique, been so severe that it led to a lowstand of at
Figure 8 Geochemical data from the core (core F) from Lake Albert. TOC total
organic content (wt%), HI hydrogen index. All determinations on bulk organic matter
Source: Beuning et al. (1997).
644 Late Quaternary paleoenvironmental changes in East Africa
least 46 m below present levels at the core Kenya is characterized by a stable abundance
site in Lake Albert. The timing of this period of Bambusoid/Arundinoid-type grasses
of maximum aridity in Lake Albert is consis- reflecting the consistent presence of mon-
tent with dated records of prolonged minimal tane forest grasses (Wooller et al., 2000).
lake levels throughout East Africa and most
of the African continent from c. 20,000 to 4 Period 12,500 to 10,000 14C yr BP
12,500 14C yr BP (eg, Livingstone, 1967; 1980; The period after 12,500 14C yr BP and into
Kendall, 1969; Talbot and Livingstone, 1989; the onset of the Holocene basically marks the
Street-Perrott and Robert, 1983; Taylor, terminal phase of the last glaciation. This
1990). Rising water in Lake Albert at 12,500 transition period was marked by an increase
14C yr BP was coincident with severe flooding in temperature and moisture, as evidenced by
along the main Nile (Said, 1993). The Lake an establishment of lower montane forest as
Albert record shows that deep-water condi- indicated by a reduction of Hagenia and
tions were restored soon after 12,500 14C yr Urticaceae and increase in Podocarpus pollen
BP although there is an apparent change in at Muchoya Swamp (c. 11,000 14C yr BP) and
sedimentation rate between 8000 and 3400 Ahakagyezi (c. 10,600 14C yr BP) (Taylor,
14C yr BP. During the latter period there was 1990). However, a major climatic fluctuation
an interruption of sedimentation at the coring involving an abrupt return to arid conditions
site that could be another, briefer period of occurred at several sites just before 10,000
low water, due to dry conditions in the catch- 14C yr BP (Figure 9), probably reflecting the
ment (Beuning et al., 1997). This dry episode Younger Dryas climatic reversal (cf. Coetzee,
may well be part of the mid-Holocene low 1967; Hamilton, 1982; Gillespie et al.,
precipitation episode that is recorded 1983; Gasse et al., 1989; Roberts et al., 1993;
throughout the tropics at c. 4000 yr 14C yr BP. Beuning et al., 1997; Johnson et al.,
The pollen record from Lake Naivasha 2000; Olago, 2001). Further evidence for the
shows a brief moist period between 17,000 Younger Dryas event affecting East Africa
and 15,000 14C yr BP characterized by some- comes from the Burundi highlands and the
what increased percentages of forest taxa, Aberdare Mountains (Barker et al., 2004). In
Podocarpus, Olea, Celtis, and Hagenia the Burundi highlands the abrupt aridity was
(Mworia-Maitima, 1991). However, the Lake marked by a shift to grassland pollen types
Naivasha record, though supported by (Bonnefille et al., 1995) while the Aberdare
diatom and ostracod data that suggest high record shows cessation of peat growth
lake levels, is inconsistent with all the other (Street-Perrott and Perrott, 1990). At the
records from the region, which show arid Lake Albert site there was an abrupt return
conditions during this period (Bakker, 1964; to drier conditions (c. 11,400 to 9900 14C yr
Coetzee, 1967; Hamilton, 1982). The anom- BP), as reflected by a 40% increase in grass
alous Naivasha record is probably a result of pollen contemporaneous with an independent
dating problems. reduction of all forest taxa percentages to
There was a slight amelioration of climate near zero (Beuning et al., 1997). Moist condi-
toward more humid conditions after 12,500 tions resumed soon after c. 9900 14C yr BP
yr BP as evident by an expansion of montane but were again interrupted by dry arid condi-
forest in the Cherangani Hills and also on tions at 8000 14C yr BP as evidenced by a
Mount Kenya (Coetzee, 1967; 1987; change in sedimentation rate.
Hamilton, 1982). Evidence from Sacred Open forest pollen types began to appear
Lake, Mount Kenya, suggests that the maxi- in Lake Tanganyika at c. 12,000 14C yr BP and
mum aridity phase occurred just before attained maximum diversity at c. 10,000 14C
13,500 14C yr BP. During this period the yr BP (Vincens, 1989a). The Zambezi wood-
charred grass cuticle record from Mount lands expanded and diversified while the
Lawrence M. Kiage and Kam-biu Liu 645
Figure 9 Enlarged subset of the pollen diagram of core F from Lake Albert.
Percentages of pollen types are based on percent total of all identified, non-aquatic
pollen grains. Selected taxa most clearly depict the 1500-year arid interval around the
lake from 11,400 to 9900 14C yr BP. This arid interval is divided into three zones with
‘P’ (11,400 to 10,900 14C yr BP) and ‘B’ (10,400 to 9900 14C yr BP) being the most arid
and a slightly moister phase during zone ‘A’ (10,900 to 10,400 14C yr BP). Calculations
of arboreal pollen percentages excluding Gramineae from the pollen sum produce
similar reductions
Source: Beuning et al. (1997).
forests and Afro-alpine vegetation belt in the the Lake Naivasha record (Mworia-Maitima,
Lake Tanganyika catchment abruptly 1991) does not show evidence of vegetation
retreated upslope to levels similar to those of change around this time (10,000 14C yr BP).
the present at the onset of the Holocene Similarly there is no evidence of temperature
(Vincens, 1989b). These changes occur dur- change in the Ruwenzori record (Livingstone,
ing a regime of rising temperatures and are 1967). Generally, there is no agreement
observed in many sites within the region, among the researchers as to whether the
albeit with a few exceptions. For instance, vegetation shifts evident during this period
646 Late Quaternary paleoenvironmental changes in East Africa
mid-Holocene due to the delayed adoption of Africa (Hamilton, 1982; Talbot, 1988; Street-
iron technology (Hamilton et al., 1986). Perrott and Harrison, 1985; Street-Perrott
A recent pollen record from Lake and Perrott, 1993; Elenga et al., 1994; Maley
Tanganyika documents increasingly arid con- and Brenac, 1998; Bonnefille and Chalié,
ditions after 5000 BP with the probable 2000; Barker et al., 2004). This abrupt aridity
replacement of forest by open grassland in has been associated with decline in the
parts of the lake’s catchment (Msaky et al., base flow of the White Nile (Talbot and
2005). The Lake Bogoria pollen sequence Brendeland, 2001) and has also been linked to
shows an abrupt disappearance of high- the collapse of the Old Kingdom in the Nile
altitude forest pollen such as Hagenia Valley (cf. Hassan, 1997; Barker et al., 2004).
abyssinica, Hypericum, Stoebe, and Ericaceae, The abrupt shift to a drier and more seasonal
and increases in more drought-adapted taxa environment at c. 4000 14C yr BP has also
such as Podocarpus, Juniperus, Acacia and been recorded at different sites in the island
Dodonaea after 4500 14C yr BP (Vincens, of Madagascar, which lies to the southeast of
1986). There was a sharp increase in the study region (Burney, 1993).
Podocarpus after 4500 14C yr BP in Mount From 2500 14C yr BP there was progres-
Kenya (Street-Perrott and Perrott, 1993) and sive degradation of arboreal cover and a con-
a similar rise in Mount Elgon at c. 3500 14C yr current increase in Gramineae pollen in the
BP (Hamilton, 1982), which signify the onset northern basin of Lake Tanganyika, which is
and establishment of dry conditions in the probably a consequence of both increasing
region. Of the four species of Podocarpus aridity and human interference. There is evi-
found in East Africa, only two are widely dis- dence of human activities documented in the
tributed – Podocarpus gracilior and P. malan- pollen sequence, such as the presence of
jianus, the former being more typical of dry the rare Elaeis guineensis (wild oil palm) in the
montane forests than the latter (Hamilton, Masoko record (Vincens, 1989a). Pollen and
1982). During this same period there was a charcoal records from Ahakagyezi Swamp
rise in dry forest components such as Olea, show evidence of deforestation associated
Macaranga, Pygeum, Celtis, Syzygium, with agricultural activities before c. 4800 14C
Neoboutonia, and Galiniera (Coetzee, 1967; yr BP (Hamilton et al., 1986). Unfortunately
Street-Perrott and Perrott, 1993; Olago, most pollen of cultivated plants has not been
2001). definitely identified (cf. Msaky et al., 2005).
The sharp increase in Podocarpus at c. 4500 Environmental changes associated with
14C yr BP at the Mount Kenya site (Street- deforestation appeared in Muchoya Swamp
Perrott and Perrott, 1993) and at c. 3500 14C at c. 2500 yr BP and contemporaneous
yr BP at Mount Elgon (Hamilton, 1982) corre- changes are evident in the Batongo and
sponds with the third abrupt climate event Katenga mires (Morrison and Hamilton,
recorded in the Mount Kilimanjaro ice core 1974; Taylor, 1990). At Lake Tanganyika there
record (Thompson et al., 2002). In the is a consistent trend of a decrease in grass
Kilimanjaro ice core record the desiccation pollen, and increase in pteridophyte and
episode is marked by a visibly distinct (30 mm forest indicator pollen during the last few
thick) dust layer containing high concentra- centuries (Msaky et al., 2005). This trend
tions of other chemical species (F and Na), suggests an increase in soil erosion and recy-
accompanying a possible hiatus in ice accumu- cling of abundant arboreal pollen following
lation. Lake level records from Lake Rukwa, land clearance for agriculture. Increase in sed-
Tanzania (Talbot and Livingstone, 1989), and imentation rates during this period provides
Lake Cheshi, Zambia (Stager, 1988), show an further evidence for increased soil erosion in
abrupt dry episode at c. 3000 14C yr BP. The the Lake Tanganyika catchment (Cohen et al.,
abrupt dry event during c. 4500–3500 14C yr 2005; Msaky et al., 2005; Palacios-Fest et al.,
BP has now been recognized across tropical 2005).
648 Late Quaternary paleoenvironmental changes in East Africa
7 The last millennium that for Lake Victoria and Lake Albert it pro-
Verschuren et al. (2000) obtained a 6 m sedi- vides high-temporal-resolution information
ment core from the partially submerged concerning recent climate history, occasion-
Crescent Island Crater (CIC) basin in Lake ally corroborated by oral tradition.
Naivasha, Kenya, to reconstruct the record A combination of sediment-inferred
of rainfall and drought in equatorial East depth reconstruction and diatom composi-
Africa during the past 1100 years. Lake tion and distributions revealed that Lake
Naivasha is a freshwater lake with subsurface Naivasha experienced a long saline lowstand
outflow and a short (10 yr) water-residence from ~AD 1000 to 1270 interrupted by one
time (Ojiambo and Lyons, 1996). Strong freshwater interval in the early part of the
evaporation and highly variable river inflow millennium (Figure 10). This lowstand was
cause its surface elevation to fluctuate in followed from ~AD 1270 to 1550 by a mostly
response to changing climate regimes positive water balance and the establish-
(Verschuren et al., 2000). Although the lake ment of freshwater conditions, except for
level record for Lake Naivasha is shorter than one recurrence of saline conditions dated
Figure 10 A plot of lake level and salinity record for Crescent Island Crater (Lake
Naivasha) compared with the decadal record of atmospheric 14CO2 production as a
proxy for solar radiation. The grey bars are the severe drought events as recounted
from oral tradition. Bars A, B, and C represent the Wamara, Nyarubanga, and
Lapanarat-Mahlatule droughts, respectively
Source: From Verschuren et al. (2000).
Lawrence M. Kiage and Kam-biu Liu 649
synchronous with the LGM, the second aridity climate of equatorial East Africa was gener-
event recorded in the two lakes (12,500 14C yr ally drier than today during the Medieval
BP, 11,710 14C yr BP) could be associated Warm Period (MWP; ~AD 1000–1270), and
with the Younger Dryas interval. Palynological that fairly wet conditions during the Little
and other proxy evidence from Lake Albert Ice Age (LIA; ~AD 1270–1850) were inter-
(Figure 3) further provides clear evidence of sig- rupted around AD 1380–1420, 1560–1620 and
nificant and abrupt drying of climate in the 1760–1840 by episodes of persistent aridity,
Lake Albert region at ~12,500 14C yr BP. The which seem to be more severe than those of
terminal phase of the last glaciation, from more recent times (Verschuren et al., 2000;
12,500 14C yr BP into the onset of the Verschuren, 2004). Again these subtle
Holocene, was marked by an increase in tem- decadal- and century-scale events are missing
perature and moisture. However, more work in the pollen record from the lake (cf. Mworia-
needs to be done before the occurrence of the Maitima, 1991), further emphasizing the role
Younger Dryas in East Africa can be confirmed. of proxy sensitivity in revealing details of
Different proxies, eg, pollen, diatoms, paleoenvironmental conditions.
charred grass cuticles, and oxygen isotope That the oscillations in the lake levels and
records from diatomic silica, suggest that East changes in pollen and other proxy records
Africa was characterized by warm and moist reflect changes in the climate and vegetation
conditions amid rapid climatic changes during of the region is not in doubt. However, the
the early to middle Holocene period. The challenge remains in the understanding of the
ice core record from Mount Kilimanjaro forcing mechanism behind paleoenvironmen-
(Thompson et al., 2002) documents two sig- tal changes in a tropical region. Although the
nificant and abrupt drought events in the Milankovich cycles have been advanced as a
region, one at c. 8300 BP and the other at possible explanation for the changes (Short
5200 BP, which appear to correspond with an et al., 1991; Tyson, 1999; Tyson et al., 2000;
apparent change in sedimentation rate at Olago, 2001; Stager et al., 2003; Verschuren
Lake Albert sometime between 8000 and et al., 2004), they probably are not responsi-
3400 14C yr BP (Beuning et al., 1997). ble for the centennial-scale and decadal-scale
However, these drought events are not changes such as those evident from the Lake
recorded in the pollen sequence. This empha- Naivasha records. Therefore more studies
sizes the importance of proxy sensitivity and and thought ought to be given to possible
sampling resolution as it affects paleoenvi- causal mechanisms.
ronmental data interpretation. The ice core A number of hypotheses have been
record provides higher-temporal resolution offered identifying these mechanisms. One
data than the lake sediment and pollen explanation is based on the intensity of the
records can, and therefore is able to detect southeast summer monsoon and the Congo
the two abrupt drought events. The longer air stream that influence the amount and sea-
and more extensive drought event at c. 4000 sonality of rainfall in the East African region
14C yr BP (cf. Marchant and Hooghiemstra, (cf. Hamilton, 1982; Nicholson, 1996;
2004) was registered in both the pollen and Nicholson and Yin, 2001; Vincens et al.,
diatom records at different sites (Hamilton, 2003). Since the southeast summer monsoon
1982: Mworia-Maitima, 1991; Street-Perrott acts in tandem with the seasonal movement
and Perrott, 1993; Thompson et al., 2002). of the Intertropical Convergence Zone
Lake level records from Lake Naivasha (ITCZ), their character must have dramati-
emphasize the significance of high-resolution cally changed during the periods of intense
stratigraphic records in revealing decadal and aridity recorded in the East African lakes.
century-scale paleoclimatic changes. The It is possible that the summer monsoon was
Naivasha lake level record suggests that the very weak during the periods of intense
Lawrence M. Kiage and Kam-biu Liu 651
aridification (c. 20,000–18,000 14C yr BP, 12,500 At the centennial timescale, the lake level
yr BP, and after ~4000 14C yr BP), resulting in fluctuations could also be a result of the solar
intense droughts that reduced the lake levels to activity and atmospheric radiocarbon (14C)
those registered in Lakes Albert, Victoria, and production (cf. Street-Perrott et al., 1989;
Naivasha. Indeed, records from many sites in Jouzel et al., 1993; Olago, 1995). The latter
terrestrial Asia and the Arabian Sea, which phenomena are negatively linked because
correspond with those from terrestrial sites in solar activity (as measured by reduced
Africa, suggest the existence of significant sunspot numbers) increases cosmic ray bom-
century-scale monsoon weakening events bardment in the upper atmosphere, which
especially during the early to middle Holocene enhances 14C production. Indeed a compari-
(Van Campo et al., 1982; Bryson, 1989; Street- son of the Lake Naivasha record with that of
Perrott and Perrott, 1990; An et al., 1993; Van reconstructed atmospheric 14CO2 production
Campo and Gasse, 1993; Gasse and Van reveals that the inferred MWP African aridity
Campo, 1994; Overpeck et al., 1996). These and all the low lake level episodes were coeval
variations in the summer monsoon may explain with high solar activity, with the intervening
some of the abrupt climatic changes evident in highstands contemporaneous with low solar
the paleoecological records from East Africa. activity (cf. Street-Perrott et al., 1997; Van
Changes in the summer monsoon are linked to Geel et al., 2000; Verschuren et al., 2000;
changes in global circulation patterns, especially Olago, 2001; Thompson et al., 2002). For
temperature changes associated with events instance, the highest rainfall inferred from
such as the deglaciation at the end of the LGM, Lake Naivasha over the past millennium
and Younger Dryas episodes in the mid-latitude was contemporaneous with the ‘Maunder
regions of the Northern Hemisphere as well as Minimum’ of solar radiation (Verschuren
sea surface temperatures (SST). There is also et al., 2000). Therefore, variation in solar
increasing evidence linking changes in the mon- activity may have contributed to the changes
soon and the ITCZ to SST changes (Gasse and in the paleoclimate of East Africa.
Van Campo, 1994; Gasse, 2002) as well as the Another hypothesis for the paleoenviron-
character of the thermohaline circulation mental changes in East Africa is the possible
(THC) (Nyberg et al., 2002). teleconnection between high and low lati-
The position of the ITCZ has been shown tudes during the early and mid-Holocene
to be closely linked to the changes in the sea (Stager et al., 2003). This could be the forcing
surface temperatures over the Atlantic mechanism behind the abrupt dry episodes at
Ocean (Gasse and Van Campo, 1994; Sirocko ~8300 and 5200 BP (Thompson et al., 2002)
et al., 1996; Marshall et al., 2001) and the El that were synchronous with major meltwater
Niño Southern Oscillation (ENSO) events and ice rafting phenomena in the North
(Nicholson, 1995; 1996; Hastenrath, 2002). Atlantic (Bond et al., 1997). However, the
Periods of lower precipitation in Africa have 8200 BP meltwater event was based on
been associated with warm SST in the north- calendar years (first detected in ice core
ern Atlantic Ocean and colder SST in the records). Firm conclusions regarding the tele-
southern Atlantic Ocean. Gasse and Van connections hypothesis are not possible due
Campo (1994) specifically link the weak mon- to uncertainties surrounding the radiocarbon
soons at c. 4000 14C yr BP to contemporane- chronologies (cf. Stager et al., 2003).
ous decreases in the North Atlantic SST. Human-induced vegetation changes in
Decrease in the latter would reduce moisture East Africa, especially during the Holocene,
in the parts of Africa where monsoonal air- cannot be ruled out. During this period
flow brings water vapour from the southern humans progressed from being relatively
subtropical Atlantic anticyclone (Marchant minor influences, restricted to a few ecosys-
and Hooghiemstra, 2004). tems, to becoming a major force across even
652 Late Quaternary paleoenvironmental changes in East Africa
ence of high counts of microscopic charcoal no anthropogenic link. Records from the
(suggesting fire occurrence) concurrent with island of Madagascar (latitude 12°S to 26°S)
forest clearance. In some cases (eg, Vincens (Burney, 1987a), similar to East Africa in
et al., 2003) forest clearance is accompanied terms of climate and vegetation communities,
by increased pollen frequencies of cultivated document fire episodes well before the pre-
crops (eg, Elaeis guineensis), Gramineae, and sumed arrival of humans c. 1500–2000 14C yr
disturbance markers (eg, Ricinus communis BP (cf. Burney, 1987b; 1993; 1997). Clearly,
that may indicate clearance for agricultural not all paleoecologically recorded fires were
and/or pastoral activities). induced by humans, since even in the absence
Rapid vegetation changes due to human of humans wildfires occur when necessary
activities have the potential of producing conditions of low fuel moisture and a natural
remarkable changes in land surface albedo ignition source (for instance, lightning and vol-
capable of inducing feedback processes that canic activities) are met. However, changes
could result in changes in climate (cf. in fire regimes and intensity may point to the
Kutzbach et al., 1996; Doherty et al., 2000). human agency, as evident in the Masoko
Climate models have shown that vegetation- records (Vincens et al., 2003). Unfortunately
albedo changes provide positive feedback most of the records from tropical Africa have
effects that cause African climate to switch temporal resolutions that are too coarse to
abruptly between wet and dry conditions fully investigate the anthropogenic effect on
(Claussen, 1997; Claussen et al., 1999). It is the vegetation.
possible that the mid-Holocene aridification Unequivocal interpretation of the various
may have been stimulated by changes in land- sedimentological, biological, and geochemical
surface conditions (cf. Lamb et al., 1989; climate proxy data extracted from the lake
DeMenocal et al., 2000). However, the fun- sediments with respect to past climatic varia-
damental problem remains to separate purely tions is an extremely complex and challenging
background-level environmental or climatic exercise. Whereas this review offers a con-
change from that caused by humans. Since tribution to our understanding of paleoenvi-
most indigenous crop plants in East Africa ronmental changes in East Africa, it also
(eg, cereals, cassava, banana, and legumes) highlights what is unknown. For instance,
have not left identifiable traces in the pollen many of the accepted paradigms are based on
record, the palynological identification of agri- coarse-temporal-resolution analyses, which
cultural influence in the region becomes diffi- may miss subtle, yet important, changes in
cult. Another problem is that natural drought vegetation and climate. Although humans and
events are often accompanied by increases in their ancestors are assumed to have inter-
grass pollen or high frequencies of micro- acted with East African environments longer
scopic charcoal. Human activities such as than in most other regions in the world, very
Lawrence M. Kiage and Kam-biu Liu 653
little is known about their impacts on the Behling, H., Pillar, V.P., Orloci, L. and Bauermann,
environment, especially during the Holocene, S.G. 2004: Late Quaternary Araucaria forest, grass-
land (Campos), fire and climate dynamics, studied by
let alone the earlier periods. There is a need high-resolution pollen, charcoal and multivariate
for more higher-temporal-resolution palyno- analysis of the Cambara do Sul core in Southern
logical studies with an emphasis on multi- Brazil. Palaeogeography, Palaeoclimatology,
proxy methodologies in order to improve our Palaeoecology 203, 277–97.
knowledge of the human/climate impact in Beuning, K.R.M., Talbot, M. and Kelts, K. 1997; A
revised 30,000-year paleoclimatic and paleohy-
the region. We still need to devise methods to drologic history of Lake Albert, East Africa.
separate purely climatic and human-driven Palaeogeography, Palaeoclimatology, Palaeoecology
environmental changes. Identification of 136, 259–79.
pollen of indigenous cultivated crops is Blodgett, R.H. 1985: Paleovertisols as indicators of cli-
certainly part of the solution. mate. American Association of Petroleum Geologists
Bulletin 69, 239–39.
Bond, G., Showers, W., Cheseby, M., Lotti, R.,
References Almasi, P., DeMenocal, P., Priore, P., Cullen,
Ashley, G.M., Mworia-Maitima, J., Muasya, A.M., H., Hajdas, I. and Bonani, G. 1997: A pervasive
Owen, R.B., Driese, S.G., Hover, V.C., Renaut, millennial-scale cycle in the North Atlantic Holocene
R.W., Goman, M.F., Mathai, S. and Blatt, S.H., and Glacial climates. Science 278, 1257–66.
2004: Sedimentation and recent history of a fresh- Bonnefille, R. 1976: Implications of pollen assemblage
water wetland in a semi-arid environment: Loboi from Koobi Fora formation, East Rudolf, Kenya.
Swamp, Kenya, East Africa. Sedimentology 51, 1–21. Nature 264, 403–407.
An, Z., Porter, S.C., Weijian, Z., Yanchou, L., — 1979: Methode palynologique et reconstitutions pale-
Donahue, D.J., Head, M.J., Xihuo, W., oclimatiques au Cenzoique dans le Rift Est Africain.
Jianzhang, R. and Hongbo, Z. 1993: Episode of Bulletin de la Société Géologique de France 21,
strengthened summer monsoon climate of Younger 331–42.
Dryas Age on the Loess Plateau of Central China. Bonnefille, R. and Chalié, F. 2000: Pollen-inferred pre-
Quaternary Research 39, 45–54. cipitation time series from equatorial mountains,
Aucour, A., Hillaire-Marcel, C. and Bonnefille, R. Africa, the last 40 kyr BP. Global and Planetary
1994: Late Quaternary biomass changes from 13C Change 26, 25–50.
measurements in a highland peat bog from equatorial Bonnefille, R. and Riollet, G. 1988: The Kashiru
Africa (Burundi). Quaternary Research 41, 225–33. sequence (Burundi) paleoclimatic implications for the
Bakker, E.M. van Z. 1964: A pollen diagram from last 40,000 yr BP in tropical Africa. Quaternary
equatorial Africa, Cherangani, Kenya. Geologie en Research 30, 19–35.
Mijnbouw 43, 123–28. Bonnefille, R., Riollet, G., Buchet, G., Icole, M.,
Bakker, E.M. van Z. and Coetzee, J.A. 1972: A reap- Lafont, R., Arnold, M. and Jolly, D. 1995:
praisal of late Quaternary climatic evidence from Glacial/interglacial record from tropical Africa, high
tropical Africa. Palaeoecology of Africa 7, 151–82. resolution pollen and carbon data at Rusaka, Burundi.
Barker, P.A., Street-Perrott, F.A., Leng, M.J., Quaternary Science Reviews 14, 917–36.
Greenwood, P.B., Swain, D.L., Perrott, R.A., Boucher, K. 1975: Global climate. London: English
Telford, R.J. and Ficken, K.J. 2001: A 14,000-year Universities Press, 326 pp.
oxygen isotope record from diatom silica in two alpine Boyd, W.E. and Hall, V.A., editors 1998: New fron-
lakes on Mount Kenya. Science 292, 2307–10. tiers and applications in palynology: IX International
Barker, P.A., Talbot, M.R., Street-Perrott, F.A., Palynological Congress, Houston, Texas, USA,
Marret, F., Scourse, J. and Odada, E.O. 2004: 23–25 June 1996. Review of Palaeobotany and
Late Quaternary climatic variability in intertropical Palynology 103, 1–83.
Africa. In Battarbee, R.W., Gasse, F. and Stickley, Bryson, R.A. 1989: Modeling the SW India monsoon
C.E., editors, Past climate variability through Europe for the last 40000 years. Climate Dynamics 3,
and Africa, Dordrecht: Springer, 117–38. 169–77.
Battarbee, R.W., Gasse, F. and Stickley, C., editors Burney, D.A. 1987a: Late Quaternary stratigraphic
2004: Past climate variability in Europe and Africa. charcoal records from Madagascar. Quaternary
Dordrecht: Springer. Research 28, 274–80.
Behling, H. 2002: Late Quaternary vegetation and cli- — 1987b: Late Holocene vegetational change in central
mate dynamics in southeastern Amazonia inferred Madagascar. Quaternary Research 28, 130–43.
from Lagoa da Confusao in Tocantins State, northern — 1993: Late Holocene environmental changes in arid
Brazil. Amazoniana-Limnologia et Ocologia regionalis southwestern Madagascar. Quaternary Research 40,
Systemae Fluminis Amazonas 17, 27–40. 98–106.
654 Late Quaternary paleoenvironmental changes in East Africa
— 1997: Environmental change, extinction and human Gasse, F. 2000: Hydrological changes in the African
activity: evidence from caves in NW Madagascar. tropics since the Last Glacial Maximum. Quaternary
Journal of Biogeography 24, 755–67. Science Reviews 19, 189–211.
Butzer, K.W., Isaac, G.L., Richardson, J.L. and — 2002: Diatom-inferred salinity and carbonate oxygen
Washbourn-Kamau, C. 1972: Radiocarbon dating isotopes in Holocene waterbodies of western Sahara
of East African lake levels. Science 175, 1069–79. and Sahel (Africa). Quaternary Science Reviews 21,
Claussen, M. 1997: Modeling biophysical feedback in 737–67.
the African and Indian monsoon region. Climate Gasse, F. and Van Campo, E. 1994: Abrupt post-glacial
Dynamics 13, 247–57. climate events in West Asia and North Africa mon-
Claussen, M., Kubatski, C., Brovkin, V., Ganopolski, soon domains. Earth Planetary Science Letters 136,
A., Cholezmann, P. and Pachur, H.-J. 1999: 435–56.
Simulation of an abrupt change in Saharan vegetation in Gasse, F., Barker, P., Gell, P.A., Fritz, S.C. and
the mid-Holocene. Geophysical Research Letters 26, Chalie, F. 1997: Diatom inferred salinity in palaeo-
2037–40. lakes: an indirect tracer of climate change.
Coetzee, J.A. 1964: Evidence for a considerable Quaternary Science Reviews 16, 547–63.
depression of the vegetation belts during the upper Gasse, F., Juggins, S. and Khelifa, B.L. 1995: Diatom-
Pleistocene on the East African mountains. Nature based transfer functions for inferring hydrochemical
204, 564–66. characteristics of African paleolakes.
— 1967: Pollen analytical studies in East and Southern Palaeogeography, Palaeoclimatology, Palaeoecology
Africa. Palaeoecology of Africa 3, 1–46. 117, 31–54.
— 1987: Palynological intimations on the East African Gasse, F., Lédéé, V., Massault, M and Fontes, J.C.
mountains. Palaeoecology of Africa 18, 231–44. 1989: Water level fluctuations of Lake Tanganyika in
Cohen, A.C., Palacios-Fest, M.R., Msaky, E.S., phase with oceanic changes during the last glaciation
Alin, S.R., Mckee, B., O’Reilly, C.M., and deglaciation. Nature 342, 57–59.
Dettman, D.L., Nkotagu, H. and Lezzar, K.E. Gillespie, R., Street-Perrott, F.A. and Switsur, R.
2005: Paleolimnological investigations of anthro- 1983: Post-glacial arid episodes in Ethiopia have impli-
pogenic environmental change in Lake Tanganyika: cations for climate prediction. Nature 306, 680–83.
IX. Summary of paleorecords of environmental Haberyan, K.A. 1987: Fossil diatoms and the paleolim-
change and catchment deforestation at Lake nology of Lake Rukwa, Tanzania. Freshwater Biology
Tanganyika and impacts on Lake Tanganyika ecosys- 17, 429–36.
tem. Journal of Paleolimnology 34, 125–45. Hamilton, A.C. 1982: Environmental history of East
Darbyshire, I., Lamb, H. and Umer, M. 2003: Forest Africa: a study of the Quaternary. London: Academic
clearance and regrowth in northern Ethiopia during Press, 328 pp.
the last 3000 years. The Holocene 13, 537–46. Hamilton, A.C. and Perrott, R.A. 1978: Date of
DeMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., deglaciation of Mount Elgon. Nature 273, 49.
Sarnthein, M., Baker, L. and Yarusinsky, M. — 1979: Aspects of the glaciation of Mount Elgon, East
2000: Abrupt onset and termination of the African Africa. Palaeoecology of Africa 11, 153–61.
Humid Period. Quaternary Science Reviews 19, Hamilton, A.C., Taylor, D. and Vogel, J.C. 1986:
347–61. Early forest clearance and environmental degradation
Doherty, R., Kutzbach, J., Foley, J. and Pollard, D. in southwest Uganda. Nature 320, 164–66.
2000: Fully coupled climate/dynamical vegetation Hassan, F.A. 1997: Nile floods and political disorder in
model simulations over northern Africa during the early Egypt. In Dalfes, H.N., Kukula, G. and Weiss,
mid-Holocene. Climate Dynamics 16, 561–73. H., editors, Third millennium BC abrupt climate
Elenga, H., Peyron, O., Bonnefille, R., Jolly, D., change and the old world collapse, Berlin: Springer-
Cheddadi, R., Guiot, J., Andrieu, V., Bottema, S., Verlag, 1–23.
Buchet.G., de Beaulieu, J.L., Hamilton, A.C., Hastenrath, S. 2002: The intertropical convergence
Maley, J., Marchant, R., Perez-Obiol, R., Reille, zone of Eastern Pacific revisited. International Journal
M., Riollet, G., Scott, L., Straka, H., Taylor, D., Van of Climatology 22, 347–56.
Campo, E., Vincens, A., Laarif, F. and Jonson, H. Johnson, T.C., Kelts, K. and Odada, E. 2000: The
2000: Pollen-based biome reconstruction for southern Holocene history of Lake Victoria. Ambio 29, 2–11.
Europe and Africa 18,000 yr BP. Journal of Biogeography Jouzel, J., Hoffmann, G., Koster, R.D. and Mason, V.
27, 621–43. 1993: Water isotopes in precipitation: data/model
Elenga, H., Schwartz, D. and Vincens, A. 1994: comparison for present-day and past climates.
Pollen evidence for Late Quaternary vegetation and Quaternary Science Reviews 19, 363–79.
inferred climate changes in Congo. Palaeogeography, Kendall, R.L. 1969: An ecological history of Lake
Palaeoclimatology, Palaeoecology 109, 345–46. Victoria Basin. Ecological Monograph 39, 121–76.
Flenley, J.R. 1998: Tropical forests under the climates of Kutzbach, J., Bonan, G. and Harrison, S. 1996:
the last 30,000 years. Climate Research 39, 177–97. Vegetation and soil feedbacks on the response of the
Lawrence M. Kiage and Kam-biu Liu 655
African monsoon to orbital forcing in the middle Holocene: a pollen-based record from western
Holocene. Nature 373, 134–37. Uganda. The Holocene 8, 375–81.
Lamb, H., Darbyshire, L. and Verschuren, D. 2003: Marchant, R.A., Behling, H., Bush, M., Cleef, A.,
Vegetation response to rainfall variation and human Colinvaux, P.A., De Oliveira, P.E.,
impact in central Kenya during the past 1100 years. Hooghiemstra, H., Ledru, M.-P., Lozano-
The Holocene 13, 285–92. Garcia, M.S., Markgraf, V. and Salgado-
Lamb, H.F., Gasse, F., Benkaddour, A., El Lebouriau, M.L. 2002: Distribution of parent taxa
Hamnuti, N., van der Kaars, S., Perkins, W.T., commonly responsible for pollen taxa from Latin
Pearce, N.J. and Roberts, C.N. 1989: Relationship America. Review of Palaeobotany and Palynology 120,
between century scale Holocene arid intervals in 1–75.
tropical and temperate zones. Nature 373, 134–37. Marshall, J., Kushnir, Y., Battisiti, D., Chang, P.,
Legesse, D., Vallet-Coulomb, C. and Gasse, F. 2004: Czaja, A., Dickson, R., Hurrell, J., McCartney,
Analysis of the hydrological response of a tropical ter- M., Saravanan, R. and Visbeck, M. 2001: North
minal lake, Lake Abiyata (main Ethiopian Rift Valley) Atlantic variability: phenomena, impacts and mecha-
to changes in climate and human activities. nisms. International Journal of Climatology 21,
Hydrological Processes 18, 487–504. 1863–98.
Lézine, A.M. 1982: Etude palynologique des sediments McGregor, G.R. and Nieuwolt, S. 1998: Tropical cli-
Quaternaires du Lac Abiyata (Ethiopie). matology (second edition). Chichester: Wiley, 339 pp.
Palaeoecology of Africa and the Surrounding Islands 14, Mohammed, U.M., Bonnefille, R. and Johnson,
93–98. T.C. 1995: Pollen and isotopic records in Late
Lind, E.M. and Morrison, M.E.S. 1974: East African Holocene sediments from Lake Turkana, Kenya.
vegetation. London: Longman, 257 pp. Palaeogeography, Palaeoclimatology, Palaeoecology
Liu, K.-B. and Colinvaux, P.A. 1988: A 5200-year his- 119, 371–83.
tory of Amazon rainforest. Journal of Biogeography Morrison, M.E.S. 1968: Vegetation and climate change
15, 231–48. in the uplands of south-western Uganda during the
Livingstone, D.A. 1962: Age of deglaciation in the Later Holocene period, 1. Muchoya Swamp, Kigezi
Ruwenzori Range, Uganda. Nature 194, 859–60. District. Journal of Ecology 56, 363–84.
— 1967: Postglacial vegetation of the Ruwenzori Morrison, M.E.S. and Hamilton, A.C. 1974:
Mountains in equatorial Africa. Ecological Monograph Vegetation and climate in the uplands of south-west-
37, 35–52. ern Uganda during the Later Pleistocene period, 2.
— 1980: Environmental changes in the Nile headwaters. Forest clearance and other vegetational changes in
In Williams, M. and Faure, H., editors, The Sahara the Rukiga Highlands during the past 8000 years.
and the Nile, Rotterdam: Balkema, 339–60. Journal of Ecology 62, 1–31.
Long, C.J., Whitlock, C., Bartlein, P.J. and Msaky, E.S., Livingstone, D. and Davis, O.K. 2005:
Millspaugh, S.H. 1998: A 9000-year fire history Paleolimnological investigations of anthropogenic
from the Oregon Coast Range based on high-resolution environmental change in Lake Tanganyika. V.
charcoal study. Canadian Journal of Forest Research Palynological evidence for deforestation and increased
28, 774–87. erosion. Journal of Paleolimnology 34, 73–83.
Maley, J. and Brenac, C. 1998: Vegetation dynamics, Mworia-Maitima, J. 1991: Vegetation response to cli-
palaeoenvironments and climate change in the forests mate change in central Rift Valley. Quaternary
of western Cameroon during the last 28,000 years Research 35, 234–45.
BP. Review of Palaeobotany and Palynology 99, — 1997: Prehistoric fires and land-cover change in west-
157–87. ern Kenya: evidences from pollen, charcoal, grass
Maley, J., Livingstone, D.A., Giresse, P., cuticles and grass phytoliths. The Holocene 4, 409–17.
Thouveny. N., Brenac, P., Kelts, K., Kling, G., Nicholson, S.E. 1995: Environmental change within the
Stager, C., Haag, M., Fournier, M., Bandet, Y., historical period. In Goudie, A.S., Adams, W.M. and
Williamson, D. and Zogning, A. 1990: Orme, A., editors, The physical geography of Africa,
Lithostratigraphy, volcanism, paleomagnetism and Oxford: Oxford University Press, 60–75.
palynology of Quaternary lacustrine deposits from — 1996: A review of climate dynamics and climate vari-
Barombi Mbo (West Cameroon): preliminary results. ability in eastern Africa. In Johnson, T.C. and Odada,
Journal of Volcanology and Geothermal Research 42, E., editors, The limnology, climatology and paleoclima-
319–35. tology of the East African lakes, Amsterdam: Gordon
Marchant, R.A. and Hooghiemstra, H. 2004: Rapid and Breach, 25–56.
environmental change in African and South — 2000: The nature of rainfall variability over Africa on
American tropics around 4000 years before present: time scales of decades to millennia. Global Planetary
a review. Earth Science Reviews 66, 217–60. Change 26, 137–58.
Marchant, R.A. and Taylor, D. 1998: Dynamics of Nicholson, S.E. and Yin, X. 2001: Rainfall conditions
montane forest in central Africa during the late in equatorial East Africa during the nineteenth
656 Late Quaternary paleoenvironmental changes in East Africa
century as inferred from the record of Lake Victoria. Roche, E. and Bikwemu, G. 1989: Paleoenvironmental
Journal of Climate 48, 387–98. change on the Zaire-Nile ridge in Burundi; the last
Nyberg, J., Malmgren, B.A., Kuijpers, A. and 2000 years: an interpretation of palynological data
Winter, A. 2002: A centennial-scale variability of from the Kashiru Core, Ijenda, Burundi. In Mahaney,
tropical North Atlantic surface hydrology during the W.C., editor, Quaternary environmental research on
late Holocene. Palaeogeography, Palaeoclimatology, East African mountains, Balkema: Rotterdam,
Palaeoecology 183, 25–41. 231–44.
Ojiambo, B.S. and Lyons, W.B. 1996: Residence times Said, R. 1993: The River Nile. Oxford: Pergamon, 320
of major ions in Lake Naivasha, Kenya and their rela- pp.
tionship to lake hydrology. In Johnson, T.C. and Salzmann, U. 2000: Are savannas degraded forests? A
Odada, E., editors, The limnology, climatology and Holocene pollen record from the Sudanian zone of
paleoclimatology of East Africa lakes, Amsterdam: NE-Nigeria. Vegetation History and Archaeobotany 9,
Gordon and Breach, 267–78. 1–15.
Olago, D.O. 1995: Late Quaternary lake sediments Salzmann, U., Hoelzmann, P. and Morczinek, I.
Mount Kenya, Kenya. PhD thesis, University of 2002: Late Quaternary climate and vegetation of the
Oxford. Sudanian zone of northeast Nigeria. Quaternary
— 2001: Vegetation changes over palaeo-time scales in Research 58, 73–83.
Africa. Climate Research 17, 105–21. Short, D.A., Mengel, J.G., Crowley, T.J., Hyde,
Olago, D.O., Street-Perrott, F.A., Perrott, R.A., W.T. and North G.R. 1991: Filtering of Milankovich
Ivanovich, M. and Harkness, D.D. 1999: Late cycles by earth’s geography. Quaternary Research 35,
Quaternary Glacial interglacial cycle of climatic and 157–73.
environmental change on Mount Kenya. Journal of Sirocko, F., Garbe-Schönberg, C.-D., McIntyre, A.
African Earth Sciences 29, 1–26. and Molfino, B. 1996: Teleconnections between the
Overpeck, J., Anderson, D., Trumbore, S. and subtropical monsoons and high-latitude climates
Prell, W. 1996: The southwest Indian monsoon over during the last deglaciation. Nature 272, 526–29.
the last 18000 years. Climate Dynamics 12, Stager, J.C. 1984: The diatom record of Lake Victoria
213–25. (East Africa): the last 17 000 years. In Mann, D.G.,
Owen, R.B., Barthelme, J.W. and Vincens, A. 1982: editor, Proceedings of the Seventh International Diatom
Paleolimnology and archaeology of Holocene Symposium, Koenigstein: Strauss and Cramer,
deposits north-east of Lake Turkana, Kenya. Nature 455–76.
298, 523–29. — 1988: Environmental changes at Lake Cheshi,
Palacios-Fest, M.R., Cohen, A.C., Lezzar, K.E., Zambia since 40 000 yr BP. Quaternary Research
Nahimana, L. and Tanner, B.M. 2005: 29, 54–65.
Paleolimnological investigations of anthropogenic Stager, J.C., Cumming, B.F. and Meeker, L.D. 1997:
environmental change in Lake Tanganyika. III. A 17 400-year, high resolution diatom record from
Physical stratigraphy and charcoal analysis. Journal of Lake Victoria, East Africa. Quaternary Research 47,
Paleolimnology 34, 31–49. 81–89.
Perrott, R.A. and Street-Perrott, F.A. 1982: New — 2003: A 10 000-year high-resolution diatom record
evidence for a late Pleistocene wet phase in north from Pilkington Bay, Lake Victoria, East Africa.
intertropical Africa. Palaeoecology of Africa 14, Quaternary Research 59, 172–81.
57–75. Stager, J.C., Reinthal, P.N. and Livingstone, D.A.
Peyron, O., Jolly, D., Bonnefille, R., Vincens, A. 1986: A 25,000 year history of Lake Victoria, East
and Guiot, J. 2000: Climate of East Africa 6000 14C Africa and some comments on its significance for the
yr BP. as inferred from pollen data. Quaternary evolution of cichlid fishes. Freshwater Biology 16,
Research 54, 90–101. 15–19.
Piperno, D.R. 1988: Phytolith analysis: an archaeological Street, F.A. and Grove, A.T. 1976: Environmental and
and geological perspective. San Diego: Academic climatic implications of Late Quaternary lake-level
Press. fluctuations in Africa. Nature 261, 385–90.
Piperno, D.R. and Pearsall, D.M. 1998: The silica — 1979: Global maps of lake level fluctuations since
bodies of tropical American grasses: morphology, tax- 30,000 yr B.P. Quaternary Research 12, 83–118.
onomy and implications for grass systematics and Street-Perrott, F.A. and Harrison, S.P. 1985: Lake
fossil phytolith identification. Annals of the levels and climate reconstruction. In Hecht, A.D.,
Smithsonian Institution 85, 1–40. editor, Paleoclimatic analysis and modeling, New York:
Roberts, N., Taieb, M., Barker, P., Damnati, B., Wiley, 291–340.
Icole, M. and Williamson, D. 1993: Timing of the Street-Perrott, F.A. and Perrott, R.A. 1990: Abrupt
Younger Dryas event in East Africa from lake level climatic fluctuations in the tropics: the influence of
changes. Nature 366, 146–48. Atlantic Ocean Circulation. Nature 343, 607–12.
Lawrence M. Kiage and Kam-biu Liu 657
— 1993: Holocene vegetation, lake levels and climate of in South Africa. South African Journal of Science 96,
Africa. In Wright, H.E. Jr, Kutzbach, J.E., Webb, T. III, 21–126.
Ruddiman, W.F., Street-Perrott, F.A. and Bartlein, P.J., Van Campo, E. and Gasse, F. 1993: Pollen- and
editors, Global climates since the Last Global Maximum, diatom-inferred climatic and hydrological changes in
London/Minneapolis: University of Minnesota Press, Sumxi Co Basin (western Tibet) since 13000 yr BP.
318–56. Quaternary Research 39, 300–13.
Street-Perrott, F.A. and Robert, N. 1983: Van Campo, E., Duplessy, J.C. and Rossignol-
Fluctuations in closed-basin lakes as an indicator of Strick, M. 1982: Climate conditions deduced from a
past atmospheric circulation patterns. In Street- 150 kyr oxygen isotope-pollen record from the
Perrott, F.A., Beran, M. and Ratcliffe, R.A.S., editors, Arabian Sea. Nature 296, 56–59.
Variations in global water budget, Dordrecht: Reidel, Van Geel, B., Heusser, C.J., Renssen, H.
331–45. and Schuurmans, C.J.E. 2000: Climatic change
Street-Perrott, F.A., Marchant, D.S., Roberts, N. in Chile at around 2700 BP and global evidence
and Harrison, S.P. 1989: Global lake-level variations for solar forcing: a hypothesis. The Holocene 10,
from 18,000 to 0 years ago: a palaeoclimatic analysis. 659–64.
USA DOE/ER/60304-H1 TR046. USA Department Verschuren, D. 1999: Sedimentation controls on the
of Energy Technical Report. preservation and time resolution of climate-proxy
Street-Perrott, F.A., Huang, Y., Perrott, R.A., records from shallow fluctuating lakes. Quaternary
Eglinton, G., Barker, P., Khelifa, L., Harkness, Science Reviews 18, 821–37.
D.D. and Olago, D.O. 1997: Impact of lower atmos- — 2004: The last 2000 years in tropical Africa. In
pheric carbon dioxide on tropical mountain ecosys- Battarbee, R.W., Gasse, F. and Stickley, C.E., editors,
tems. Science 278, 1422–26. Past climate variability through Europe and Africa,
Talbot, M.R. 1988: The origins of lacustrine oil source Dordrecht: Springer, 139–58.
rocks: evidence from the lakes of Africa. In Fleet, Verschuren, D., Briffa, K.R., Hoelzmann, P.,
A.J., Kelts, K. and Talbot, M.R., editors, Lacustrine Barber, K., Barker, P. and Scott, L. 2004: Climate
petroleum source rocks, Geological Society of London variability in Europe and Africa: A PAGES-PEP III
Special Publication 40, 29–43. Time Stream I synthesis. In Battarbee, R.W., Gasse,
Talbot, M.R. and Brendeland, K.I. 2001: Strontium F. and Stickley, C.E., editors, Past climate variability
isotopes as paleohydrological tracers in the White Nile through Europe and Africa, Dordrecht: Springer,
headwater lakes, East Africa. EOS Transactions, AGU 567–82.
Fall Meeting Supplement 82, Abstract PP21C–05. Verschuren, D., Laird, K.R. and Cumming, B.F.
Talbot, M.R. and Livingstone, D.A. 1989: Hydrogen 2000: Rainfall and drought in equatorial East
index and carbon isotopes of lacustrine organic Africa during the past 1,100 years. Nature 403,
matter as lake level indicators. Palaeogeography, 410–14.
Palaeoclimatology, Palaeoecology 70, 121–37. Vincens, A. 1986: Pollen diagram of a Late
Taylor, D.M. 1990: Late Quaternary pollen records Paleistocene-Holocene deposit in Lake Bogoria
from two Ugandan mires: evidence for environmental (Kenya). Review of Palaeobotany and Palynology 47,
change in the Rukiga highlands of southwest Uganda. 169–92.
Palaeogeography, Palaeoclimatology, Palaeoecology 80, — 1989a: Palaeoenvironments du basin nord-
283–300. Tanganyika (Zaïre, Burundi, Tanzanie) au cours des 13
Thompson, L.G., Mosley-Thompson, E., Davis, derniers milles ans: apport de la palynologie. Review of
M.E., Henderson, K.A., Brecher, H.H., Palaeobotany and Palaeoecology 61, 69–88.
Zagorodnov, V.S., Mashiota, T.A., Lin, P.-N., — 1989b: Les forêts zambéziennes du basin Sud-
Mikhalenko, V.N., Hardy D.R. and Beer, J. 2002: Tanganyika. Evolution entre 25000 et 6000 ans BP.
Kilimanjaro ice core records: Evidence of Holocene Comptes Rendus de l’Academie des Sciences, Paris
climate change in tropical Africa. Science 298, 308(Sér II), 809–14.
589–93. — 1991: Late Quaternary vegetation history of
Tiercelin, J.-J. and Vincens, A. 1987: The Baringo- the South Tanganyika Basin – climatic implica-
Bogoria half-graben, Gregory Rift, Kenya: 30,000 tions in South-Central Africa. Palaeogeography,
years of hydrological and sedimentary history. Bulletin Palaeoclimatology, Palaeoecology 86, 207–26.
des Centres Recherche Exploration-Production Elf — 1993: Nouvelle sequence pollinique du Lac
Aquitaine 11, 380–94. Tanganyika: 30,000 Ans d’histoire botanique et cli-
Tyson, P.D. 1999: Atmospheric circulation changes and matique du Bassin Nord. Review of Palaeobotany and
palaeoclimates of southern Africa. South African Palynology 78, 381–94.
Journal of Science 95, 194–201. Vincens, A., Williamson, D., Thevenon, F., Taieb,
Tyson, P.D., Karlen, W., Holmgren, K. and Heiss, M., Buchet, G., Decobert, M. and Thouveny, N.
G.A. 2000: The Little Ice Age and medieval warming 2003: Pollen-based vegetation changes in southern
658 Late Quaternary paleoenvironmental changes in East Africa
Tanzania during the last 4200 years: climate change Whitlock, C. and Millspaugh, S.H. 1996: Testing the
and/or human impact. Palaeogeography, Palaeo- assumptions of fire history studies: an examination of
climatology, Palaeoecology 198, 321–34. modern charcoal accumulation in Yellowstone
Webster, J.B., editor 1979: Chronology, migration and National Park, USA. The Holocene 6, 7–15.
drought in interlacustrine Africa. Dalhousie: Longman Wooller, M.J., Street-Perrott, F.A. and Agnew,
and Dalhousie University Press, 1–37. A.D.Q. 2000: Late Quaternary fires and grassland
White, F. 1983: The vegetation of Africa. A descriptive palaeoecology of Mount Kenya, East Africa: evi-
memoir to accompany the UNESCO/ dence from charred grass cuticles in lake sediments.
AETFAT/UNSO vegetation map of Africa. Paris: Palaeogeography, Palaeoclimatology, Palaeoecology
UNESCO. 164, 207–30.