Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
Contents lists available at ScienceDirect
Palaeogeography, Palaeoclimatology, Palaeoecology
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p a l a e o
Late Quaternary vegetation history of southeast Africa: The molecular isotopic record
from Lake Malawi
Isla S. Castañeda a,b,c,⁎,1, Josef P. Werne a,d, Thomas C. Johnson a,c, Timothy R. Filley e
a
Large Lakes Observatory, University of Minnesota Duluth, Duluth, MN 55812, USA
Department of Geology and Geophysics, University of Minnesota, 310 Pillsbury Drive SE, Minneapolis, MN 55455, USA
Department of Geological Sciences, University of Minnesota Duluth, 1114 Kirby Drive, Duluth, MN 55812, USA
d
Department of Chemistry and Biochemistry, University of Minnesota Duluth, 1039 University Drive, Duluth, MN 55812, USA
e
Department of Earth and Atmospheric Sciences and the Purdue Climate Change Research Center, Purdue University, 550 Stadium Mall Drive, West Lafayette, IN 47907, USA
b
c
a r t i c l e
i n f o
Article history:
Received 11 November 2007
Received in revised form 7 February 2009
Accepted 13 February 2009
Keywords:
East Africa
Vegetation
n-alkane
Lignin phenol
Compound-specific carbon isotope
Paleoclimate
a b s t r a c t
Accurate reconstructions of past hydrological variability are essential for understanding the climate history of
the tropics. In tropical Africa, the relative proportion of vegetation utilizing the C3 vs. C4 photosynthetic
pathway is mainly controlled by precipitation and thus past hydrological changes can be inferred from the
vegetation record. In this study, biomarkers of terrestrial plants (lignin phenols and plant leaf wax carbon
isotopes) are examined from a well-dated sedimentary record from Lake Malawi to provide a vegetation
(aridity) record of the past 23 cal ka from southeast Africa. We suggest that the ratio of cinammyl to vanillyl
(C/V) lignin phenols in Lake Malawi sediments mainly reflects inputs of C3 trees (woody tissue) vs. C4
grasses (non-woody tissue) and find that changes in the C/V ratio generally support variability noted in the
n-alkane carbon isotope record. Together, these records provide evidence for increased C4 vegetation
(grasses) surrounding Lake Malawi during Last Glacial Maximum (LGM), the Younger Dryas, in the early
Holocene, and from ~ 2 cal ka to the present, suggesting drier conditions at these times. Elevated inputs of C3
vegetation are noted in the intervals from ~ 17–13.6 cal ka and ~ 7.7–2 cal ka, indicating wet conditions in
southeast Africa. A relationship is noted between the n-alkane average chain length (ACL) and temperature,
with longer ACLs associated with higher temperatures. Higher n-alkane carbon preference index (CPI) values
correlate with higher mass accumulation rates of biogenic silica and may result from periodic increased
northerly winds over Lake Malawi, which enhance upwelling and diatom productivity while simultaneously
increasing erosion and transport of plant leaf waxes to the lake. The molecular data produced in this study
suggest that the carbon isotopic signature of bulk sediment (δ13CTOC) in Lake Malawi is primarily a reflection
of terrestrial inputs (C3 vs. C4 vegetation) and may not mainly reflect changes in algal productivity, as
previously thought.
© 2009 Published by Elsevier B.V.
1. Introduction
Lake Malawi, the southernmost of the large lakes in the East
African Rift Valley, has been the focus of numerous recent paleoclimate investigations because the lake exhibits a strong response to
global climate events and is one of a few sites in tropical Africa that
contains a continuous sedimentary record spanning the time interval
from the Last Glacial Maximum (LGM) to the present. The modern
climate of East Africa is complex with major changes in rainfall
observed over short distances (Nicholson, 1996), and therefore, it is
perhaps not surprising that paleoclimate records obtained from
nearby sites in East Africa often contain contradictory data. For
⁎ Corresponding author. Tel.: +31 222 369568; fax: +31 222 319674.
E-mail address: [email protected] (I.S. Castañeda).
1
Present address: Netherlands Institute for Sea Research, Department of Marine
Organic Biogeochemistry, PO Box 59, 1790 AB Den Burg, Texel, The Netherlands.
0031-0182/$ – see front matter © 2009 Published by Elsevier B.V.
doi:10.1016/j.palaeo.2009.02.008
example, arid conditions characterized Lake Malawi during the
Younger Dryas cold period (Johnson et al., 2002; Filippi and Talbot,
2005; Castañeda et al., 2007; Talbot et al., 2007); however, wet
conditions existed at Lake Masoko (Garcin et al., 2006), located within
100 km of the northern end of Lake Malawi. A main focus of previous
investigations of Lake Malawi and the other East African rift lakes has
been on reconstructing changes in lake level (aridity) since tropical
climate variability is often expressed as hydrological fluctuations
(Gasse, 2000). However, relatively few studies have examined
vegetation history based on analyses of organic biomarkers, which
provide an additional method of examining past hydrological changes
because aridity is the main control on the distribution of vegetation
utilizing the C3 (Calvin–Benson) vs. the C4 (Hatch–Slack) photosynthetic pathways in tropical Africa (Schefuß et al., 2003a; Castañeda
et al., 2007).
We use molecular biomarkers and compound-specific carbon
isotopes in this study to examine the response of low-latitude tropical
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
terrestrial vegetation to changes in global climate since the LGM. At
present relatively few studies of vegetation change exist for the Lake
Malawi region and the existing pollen studies (e.g. Meadows, 1984;
DeBusk, 1998) have been hampered by poor chronology, or for certain
time intervals, poor preservation of pollen grains, making it difficult to
resolve the timing of vegetation changes noted throughout the Late
Pleistocene and Holocene. Here, we examine a sediment core from the
northern basin of Lake Malawi that contains a continuous sedimentary record of the past 23 cal ka, and for which the chronology is wellestablished (Barry et al., 2002; Johnson et al., 2002). Recent advances
in organic geochemistry have provided valuable insight into the
paleoclimatic history of East Africa by allowing for the reconstruction
of a continental temperature record from Lake Malawi (Powers et al.,
2005). This temperature record, produced from the same sediment
core that we examine in the present study, has revealed significant
temperature variability in East Africa both between the Last Glacial
Maximum (LGM) and the present, as well as major temperature
variations within the Holocene. Having a continuous and well-dated
record from a site where temperature is known provides a unique
opportunity to examine the response of tropical terrestrial vegetation
to paleoenvironmental variability. In addition, this molecular study
provides an opportunity to compare vegetation proxies based on
different compound classes and to gain further insights into the origin
and transport pathways of sedimentary organic matter in Lake
Malawi.
2. Background
2.1. Study location
Lake Malawi (9°S to 14°S) is situated between the countries of
Malawi, Mozambique, and Tanzania, and is 560 km long and up to
75 km wide (Eccles, 1974) (Fig. 1). Lake Malawi is at least 5 million
years old (Finney et al., 1996), is underlain by over 4 km of sediment
(Rosendahl, 1987), and has a maximum depth of over 700 m (Johnson
101
and Davis, 1989). The lake is permanently anoxic below ~ 200 m
(Eccles, 1974) and is characterized by relatively high sedimentation
rates of 0.5–1.5 mm year− 1 (Finney et al., 1996). The majority of water
loss in Lake Malawi is by evaporation rather than outflow, making it
extremely sensitive to minor changes in aridity (Spigel and Coulter,
1996). Lake levels in some of the East African lakes are known to have
varied by hundreds of meters in the past (Gasse, 2000). However,
even during times of severe drought when other African lakes were
completely desiccated, the deepest basins of Lake Malawi contained
water and continued to accumulate sediment (Johnson, 1996; Scholz
and Rosendahl, 1988).
In addition to having both a continuous and high-resolution
sedimentary record, Lake Malawi is also a good site for paleoenvironmental reconstructions as it is situated in a climatically sensitive
geographical location that is heavily influenced by the intertropical
convergence zone (ITCZ) (Fig. 1A). The lake is located at the southern
limit of the ITCZ (~13°S) and experiences one rainy season per year
(November to March), while the rift lakes to the north experience two
rainy seasons from the overhead passage of the ITCZ. The ITCZ is a
major feature of tropical climates (Leroux, 2001; Nicholson, 1996) and
is a main control on the distribution of tropical vegetation via its
relationship to precipitation. Presently, Lake Malawi is surrounded by
tree savanna, which mainly consists of C3 vegetation (White, 1983). To
both the north and south of Lake Malawi lies grass savanna, which is
dominated by C4 grasses (White, 1983). The proximity of Lake Malawi
to this ecosystem boundary provides a sensitive location to examine
past shifts in C3 and C4 vegetation associated with changes in climate
(Fig. 1A).
Piston core M98-1P was collected from 403 m water depth in the
northern basin of Lake Malawi (10°15.99S, 34°19.19) in 1998 by an
expedition of the International Decade for East African Lakes (IDEAL)
(Fig. 1B). The lithology and age model (based on 14C dating, varve
counting, and 210Pb dating) of this 7.8 m piston core have been
previously described in several studies (Barry et al., 2002; Johnson
et al., 2002; Filippi and Talbot, 2005) and the age model is available at
Fig. 1. A) Vegetation zones of East Africa (figure modified from Schefuß et al., 2003a) and the January and July positions of the ITCZ (based on Leroux, 2001). Presently Lake Malawi is
situated within the tree savanna vegetation zone, which is C3 dominated. Note the southernmost limit of the ITCZ at the southern end of Lake Malawi. B) Location of piston core M98-1P
in the northern basin of Lake Malawi. C) Locations of Lake Rukwa and Lake Masoko for comparison with Lake Malawi. The arrows indicate the approximate location of sediment cores
examined from Lake Rukwa (Vincens et al. 2005) and Lake Masoko (Garcin et al., 2006; Vincens et al., 2003).
102
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
the NOAA web site for the World Data Center for Paleoclimatology
(www.ncdc.noaa.gov/paleo/data.html). All ages in this study are
reported as thousands of calendar years before present (cal ka).
2.2. Terrestrial vegetation, n-alkanes and lignin phenols
The C3 pathway is the most common photosynthetic pathway and is
used by almost all trees, shrubs, herbs, and cold-season grasses and
sedges (Raven et al., 1999). The C4 pathway differs from the C3 pathway
in that it involves a carbon-concentrating mechanism, which gives C4
plants (mainly warm-season grasses and sedges) a competitive
advantage in low atmospheric CO2 conditions (Ehleringer et al., 1997).
C4 plants also have higher water-use efficiency than C3 plants, and thus
are common today in tropical savannas, temperate grasslands, and in
semi-arid regions (Raven et al., 1999). The main drivers of C3 vs. C4
vegetation change are thought to be temperature, aridity, and atmospheric CO2 concentrations (Cerling et al., 1993; Collatz et al., 1998;
Kuypers et al., 1999; Pagani et al., 1999; Huang et al., 2001). While the
relative importance of each of these factors remains a topic of study (e.g.,
Zhang et al., 2003; Schefuß et al., 2003a), it is generally agreed that these
forcing factors impact the distribution of vegetation because of
differences in carbon assimilation between the C3 and C4 pathways
(O'Leary, 1981). However, in tropical Africa studies have demonstrated
that aridity is the main factor driving shifts in C3 vs. C4 vegetation
(Castañeda et al., 2007; Schefuß et al., 2003a) and thus past hydrological
fluctuations can be inferred from the paleovegetation record.
In this study, two classes of higher plant biomarkers, the plant leaf
waxes (n-alkanes) and lignin phenols, are examined in order to
investigate vegetation change in East Africa during the past 23 cal ka.
The n-alkanes are straight-chain hydrocarbons that exhibit strong odd
over even carbon number predominance in living organisms and are a
major component of the epicuticular waxes of terrestrial plant leaves
(Eglinton and Hamilton, 1967). Although n-alkanes are produced by
many organisms, carbon number distributions and isotopic compositions vary depending on the source organism. Terrestrial plants are
dominated by the long-chain (C25–C33) n-alkanes while aquatic algae
are dominated by the short-chain n-alkanes (C17–C21) (Giger et al.,
1980; Cranwell et al., 1987). These compounds are generally wellpreserved in sediments (Cranwell, 1981) and differences in the
distribution patterns of n-alkanes can provide information regarding
vegetation type since the C31 n-alkane tends to be dominant in grasses
while the C27 and C29 n-alkanes are dominant in deciduous trees
(Cranwell, 1973). In addition, the carbon isotope composition of the
long-chain n-alkanes can be used to distinguish between vegetation
utilizing different photosynthetic pathways. Plants utilizing the C3
photosynthetic pathway tend to have average n-alkane carbon isotopic
compositions of around −36‰, while plants utilizing the C4 photosynthetic pathway have average n-alkane carbon isotopic compositions of around − 21.5‰ (Collister et al., 1994). A third photosynthetic
carbon-fixation pathway, the Crassulacean Acid Metabolism (CAM)
pathway, has isotopic values intermediate between those of C3 and C4
plants (Deines, 1980) but is not found in plants dominating the
environments of East Africa (Winter and Smith, 1996). The n-alkanes
are transported to lakes via water or wind but wind erosion can be the
dominant transport mechanism in arid environments (Schefuß et al.,
2003b). Although long-range transport of plant leaf waxes by eolian
processes is possible (Gagosian and Peltzer, 1986), several studies have
shown vegetation in the immediate watershed is the predominant
source of n-alkanes to lake sediments (Schwark et al., 2002; Huang
et al., 1999; Riley et al., 1991).
Lignins are large and structurally complex phenolic polymers that are
present in the cell walls of vascular plants (Hedges and Mann, 1979a,b;
Hedges and Ertel, 1982) and have been applied successfully to studies of
lacusterine sediments to provide information on vegetation dynamics
(e.g., Orem et al., 1997; Hu et al., 1999; Filley et al., 2001; Fuhrmann et al.,
2003; Ishiwatari et al., 2005, 2006; Pempkowiak et al., 2006; Tareq et al.,
2006; Hyodo et al., 2008) and to track plant source–export dynamics in
riverine systems (Dalzell et al., 2007; Bianchi et al., 2007). Differences in
abundance of vanillyl (V), syringyl (S) and cinnamyl (C) structural
groups can be used to distinguish different plant and tissue types (Goñi
and Hedges, 1992). Vanillyl-based phenols are produced by all vascular
plants, with it being the exclusive component in gymnosperm woody
tissue. Syringyl-based phenols are produced by angiosperms and are
found in both woody and non-woody tissue, although trace levels are
found in gymnosperm non-woody tissue (Goñi et al., 1993; Hedges and
Mann, 1979a,b). Cinnamyl phenols are present in the non-woody tissues
of gymnosperms and angiosperms (Hedges and Mann, 1979a,b).
Therefore, the ratio of cinnamyl/vanillyl (C/V) phenols can be used to
examine contributions of non-woody vs. woody land plant tissues while
the ratio of syringyl/vanillyl (S/V) phenols can be used to distinguish
angiosperm from gymnosperm sources (Goñi and Hedges,1992). Woody
tissues are characterized by C/V values of b0.05 while non-woody
tissues have values of 0.1 to 0.8 (Goñi, 1997). This distinction is evident
even within individual leaves as petioles and vascular tissue have little to
no cinnamyl phenols while soft body tissues of the leaf are rich in
cinnamyls (Filley et al., 2008a). Angiosperms have S/V ratios of 0.6–4
while gymnosperms have S/V values of around 0 as they lack the syringyl
phenol (Goñi, 1997). With respect to characterizing ecosystems undergoing shifts between C4 grasses and C3 woody plants, these molecular
distinctions have proven effective for determining the relative input of
these plant types to litter and soil organic matter (Filley et al., 2008b).
Lignin phenols are relatively resistant to diagenesis but are
susceptible to oxic degradation. However, it is possible to assess the
extent of oxygenic diagenetic alteration by examining the ratio of acid
phenols to aldehyde phenols (vanillic acid to vanillin (Ac/Alv) and
syringic acid to syringealdehyde (Ac/Als); Ertel and Hedges, 1985;
Hedges et al., 1982). In Lake Malawi, which has anoxic bottom waters,
it is expected that lignin phenols are mainly degraded on land or while
in transport to the lake but are well-preserved upon reaching the
sediments. In contrast to the plant leaf waxes, which are transported
by both wind and water, lignin phenols are transported to lakes
primarily by rivers and streams (e.g. Meyers, 2003; Goñi et al., 1997).
Additionally, grass and bush fires can play an important role in the
erosion and transport of terrestrial plant materials (n-alkanes and
lignin phenols) to lacustrine sediments (e.g. Finkelstein et al., 2006
and references therein).
3. Methods
3.1. TOC and C/N ratios
Total inorganic carbon (TIC) and total carbon (TC) measurements
were determined on a UIC CO2 Coulometer. In these samples, TC is
total organic carbon (TOC) because TIC was not present in any of the
samples analyzed from core M98-1P. Carbon to nitrogen ratios (C/N)
were determined on a Costech ECS 4010 elemental analyzer. Sediment
samples did not receive acid pre-treatment prior to analysis on the
elemental analyzer. Therefore, C/N ratios reported here reflect the
ratio of total organic carbon to total nitrogen (Corg/Ntot).
3.2. Bulk δ13C
Freeze dried sediment samples were treated with excess 0.1 N
hydrochloric acid for 3 h to remove inorganic nitrogen (δ15N data will
be presented elsewhere). After acidification, sediment samples were
filtered through organic-free Whatman GF/F glass fiber filters (0.7 μm
pore size) and rinsed 4x with excess distilled and deionized water
(Millipore filtration system). Sediment samples were then dried in an
oven at 35 °C and stored in a desiccator until they could be packed into
tin capsules for isotopic analysis. Bulk carbon isotope (δ13CTOC)
samples were analyzed at the College of Marine Science at the
University of South Florida on a Thermo-Finnigan Delta-Plus XL mass
spectrometer coupled to a Carlo-Erba NA2500 Elemental Analyzer.
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
3.3. Isolation of plant leaf waxes
Sixty sediment samples were selected for molecular analysis at a
sampling resolution of approximately one sample per 500 years.
During time periods in which previous studies of Lake Malawi had
documented significant changes, such as during the Younger Dryas
cold period (Johnson et al., 2002; Talbot et al., 2007), samples were
taken at higher resolution. Freeze dried sediment samples were
soxhlet extracted in groups of five with an additional blank sample
run with every batch. This extraction blank was then processed in the
same manner as the sediment samples to ensure that no contamination was introduced to the samples during any of the steps. Lipids
were extracted with 2:1 methylene chloride: methanol for 24 h to
obtain a total lipid extract (TLE). The TLE was further separated into
neutral lipid, fatty acid, and phospholipid fatty acid fractions using
Alltech Ultra-Clean SPE Aminopropylsilyl bond elute columns. Prior to
loading the sample, bond elute columns were pre-cleaned by running
10 mL of methanol and 10 mL of 1:1 methylene chloride: 2-propanol
through the column. Eight milliliter each of 1:1 methylene chloride:
2-propanol, 4% glacial acetic acid in ethyl ether, and methanol were
used to elute the neutral lipid, fatty acid, and phospholipid fatty acid
fractions, respectively. The neutral lipid fraction is the only fraction
examined in this study, and therefore the fatty acid and phospholipid
fatty acid fractions are not discussed further here. Silica gel column
chromatography was used to separate compounds in the neutral
fraction following the procedures outlined by Wakeham and Pease
(1992). The n-alkanes were present in the first apolar fraction,
which was eluted with hexane. This fraction was next passed through
an Ag+ impregnated silica pipette column to separate the saturated
and unsaturated hydrocarbons.
Mass accumulation rates of n-alkanes were calculated from the
following formula:
MARalkane = LSRTDBDTC
where MARalkane is the MAR in ng cm− 2 year− 1, LSR = linear
sedimentation rate (cm year− 1), DBD = dry bulk density (g cm− 3),
and C = the mass of compound (ng g− 1) in dry sediment.
The carbon preference index (CPI) (Bray and Evans, 1961) of nalkanes is used to examine the odd over even carbon number
predominance, which can be used to distinguish terrestrial plant from
petroleum sources. This parameter is defined as:
103
HP 5973 mass spectrometer (MS). An HP-1 capillary column
(25 m × 0.32 mm × 0.5 μm) was used with He flow rates set at 2 mL/
min. The GC/MS oven temperature program initiated at 50° C and
increased at a rate of 10° C/min to 130° C, and then at a rate of 4° C to
320° C. The final temperature of 320° C was held for 10 min. Mass
scans were made over the interval from 50 to 650. Compounds were
identified by interpretation of characteristic mass spectra fragmentation patterns and by comparison with literature.
Quantification of compounds was performed on a Hewlett-Packard
HP 6890 Gas Chromatograph with a FID detector using 5α-androstane
as an internal standard. Compound concentrations were determined
by relating chromatogram peak area to the concentration of the
internal standard. Column type and the temperature program used for
GC analysis are the same as described above for GC-MS except for that
He flow rates were set at 2.6 mL/min.
3.5. Lignin phenol isolation and quantification
Twenty-three samples were selected for lignin analysis. Lignin
phenols were analyzed in the Stable Isotope Biogeochemistry
Laboratory at Purdue University. Standard procedures for analysis of
trimethylsilyl (TMS) derivatives of lignin phenol monomers by CuO
oxidation were followed (Hedges and Mann, 1979a,b). Laboratoryspecific conditions and techniques for quantification of lignin phenol
monomers are described by Dalzell et al. (2005). Twelve of the lignin
samples were run in duplicate or triplicate (from a split of the sediment
sample) and the standard deviation of vanillyl, syringyl and cinnamyl
phenols is better than 0.01 for each of the three phenols. To be consistent
with other lignin phenol studies, we follow standard convention and
present the lignin phenol data in terms of abundance normalized to TOC
(mg lignin/100 g organic carbon (OC)). It should be noted that when
converted to mass accumulation rates (mg lignin cm− 2 year− 1), the
overall trends do not differ significantly from the TOC-normalized
abundance data.
4. Results
4.1. Bulk geochemical data
where Cx refers to the concentration of the n-alkane with x number of
carbon atoms.
Mass accumulation rates of total organic carbon (TOCMAR) and the
atomic ratio of organic carbon to total nitrogen (Corg/Ntot) in core M981P indicate considerable variability during both the Late Pleistocene
and Holocene (Fig. 2A, C). The lowest TOCMAR values occur in the oldest
part of the record, at ~ 22 to 20 cal ka. Low TOCMAR values are also noted
in the interval from 16 to 13.5 cal ka. Generally high TOCMAR values are
noted in the intervals from 19 to 17.5 cal ka, 13 to 11.9 cal ka, 10–7 cal ka
and from 2 to 1.5 cal ka. The Corg/Ntot record generally tracks changes in
the TOCMAR record (Fig. 2A, C), with the highest Corg/Ntot values
occurring when TOCMAR values are the highest and vice versa. A
notable feature of both the TOCMAR and Corg/Ntot records is the Younger
Dryas Cold Period (11.7–13 cal ka; Broecker et al., 1989) which is
marked by an abrupt increase in both TOCMAR and Corg/Ntot.
Bulk sediment carbon isotopic values (δ13CTOC) range from
− 22.5‰ to − 25‰, with the most enriched values noted from 22–
18 cal ka (Fig. 2D). Following the LGM, δ13CTOC values gradually
become increasingly 13C depleted until ~ 13 cal ka. Unlike the TOCMAR
and Corg/Ntot records, δ13CTOC values do not exhibit a significant
excursion during the Younger Dryas, although a trend towards
slightly 13C-depleted values is noted. Following the Younger Dryas,
δ13CTOC values generally fluctuate between − 25‰ and − 23.5‰ for
the remainder of the Holocene.
3.4. Identification and quantification of compounds
4.2. Plant leaf wax biomarkers
Molecular identification of compounds (n-alkanes) was performed
on a Hewlett-Packard 6890 gas chromatograph (GC) coupled to an
Mass accumulation rates of the long-chain (C25–C33) n-alkanes
range from ~ 0.33 to 2.47 ng cm− 2 year− 1 (Fig. 3). In most samples,
CPI= 1 = 2T½ðC25 +C27 +C29 +C31 +C33 Þ = ðC24 +C26 + C28 + C30 + C32 Þ
+ ðC25 + C27 + C29 + C31 + C33 Þ = ðC26 + C28 + C30 + C32 + C34 Þ
where Cx refers to the concentration of the n-alkane with x number of
carbon atoms. Terrestrial plants are characterized by CPIs of N3
whereas mature hydrocarbons have CPIs of ~1 (Bray and Evans, 1961).
A second parameter that can be examined from n-alkanes is the
average chain length (ACL), which has been observed to increase with
increasing aridity or temperature (e.g. Rommerskirchen et al., 2003;
Peltzer and Gagosian, 1989). For higher plants this parameter is
defined as:
ACL23 − 33 = ð23½C23 + 25½C25 +27½C27 + 29½C29 + 31½C31 + 33½C33 Þ
ð½C23 + ½C25 + ½C27 + ½C29 + ½C31 + ½C33 Þ
104
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
Fig. 2. Bulk geochemical records from core M98-1P. The interval of the Younger Dryas is highlighted. A) Mass accumulation rates of total organic carbon (TOC) plotted against age in
thousands of calibrated years before present (cal ka) are shown by the gray dots while the black line represents the smoothed TOCMAR record (5 point running average). B) Mass
accumulation rates of biogenic silica (BSi) are plotted in gray while the black line represents the smoothed BSiMAR data (BSi data are from Johnson et al., 2002 and were also measured
on core M98-1P). C) The ratio of total organic carbon to total nitrogen (Corg/Ntot) shown by the gray dots while the black line represents the smoothed Corg/Ntot data. D) Bulk carbon
isotope (δ13CTOC) data. E) Lake Malawi n-alkane δ13C record, based on the weighted mean of the C29–C33 n-alkanes in cores M98-1P and M98-2PG (data from Castañeda et al., 2007).
F) Lake Malawi temperature reconstruction based on the TEX86 paleotemperature proxy (data from Powers et al., 2005). Note that the x-axis scale is reversed to facilitate
comparisons with the bulk geochemical records.
Fig. 3. Mass accumulation rates of A) the C25 n-alkane, B) the C27 n-alkane, C) the C29 n-alkane, D) the C31 n-alkane, E) the C33 n-alkane, and F) mass accumulation rates of total
terrestrial n-alkanes (the sum of C25–C33 n-alkanes).
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
105
an average value of 3.9 (Fig. 4A). The n-alkane average chain length
(ACL) ranges ~26 to ~28.5 (Fig. 4B).
4.3. Lignin phenols
In core M98-1P, total yields of vanillyl, syringyl and cinnamyl
phenols (Λ8; includes vanillin, acetovanillone, vanillic acid, syringealdehyde, acetosyringone, syringic acid, p-hydroxycinnamic acid and
ferulic acid) range from 0.15 to 0.35 mg/100 g OC (Fig. 5E). Acid to
aldehyde ratios of vanillyl and syringyl phenols indicate that lignin
phenols present in Lake Malawi sediments are generally wellpreserved. In core M98-1P, Ac/Alv values range from 0.27–0.81
while Ac/Als values range from 0.21–0.52 (Fig. 5A). Ac/Alv and Ac/
Als ratios of greater than 0.6 are considered to indicate highly
degraded lignin (Goñi, 1997). Only the sample at 4 cal ka had an Ac/
Alv ratio greater than 0.6, and therefore, with the exception of this
sample, lignin in Lake Malawi sediments is not highly degraded with
Ac/Alv values often plotting near the range of values for fresh plant
tissue. In core M98-1P, the S/V ratio ranges from 0.97 to 1.54 with an
average value of 1.16, while the C/V ratio ranges from 0.10 to 0.64 with
an average value of 0.24 (Fig. 5B, C). Throughout the length of the
record, S/V ratios remain within the range of angiosperm tissues
while C/V ratios remain within the range of non-woody tissues. The
lignin phenol vegetation index (LPVI) provides another method for
examining vegetation shifts and is defined as:
Fig. 4. A) The Carbon Preference Index (CPI), plotted in black, compared with mass
accumulation rates of biogenic silica (BSiMAR; data from Johnson et al., 2002), plotted in
gray. B) The Average Chain Length (ACL) of n-alkanes is potted in black. The Lake
Malawi temperature record based on the TEX86 paleotemperature proxy (data from
Powers et al., 2005) is plotted in gray.
the C27 or C29 n-alkane is the most abundant of the long-chain nalkanes (Fig. 3B, C), and the C34 n-alkane is the longest homologue
present. The carbon preference index (CPI) varies from 1.7 to 7.1 with
LPVI = ½fSðS + 1Þ = ðV + 1Þ + 1g × fC ðC + 1Þ = ðV + 1Þ + 1g
With V, S and C expressed as % of the Λ 8 (Tareq et al., 2004). The
advantage of the LPVI is that it provides better separation of different
vegetation sources compared with other lignin parameters, which is
important in diverse tropical ecosystems (Tareq et al., 2004).
Gymnosperm woods, non-woody gymnosperm tissues, angiosperm
woods and non-woody angiosperm tissues have average LPVI values
Fig. 5. Lignin phenol data. The interval of the Younger Dryas (YD) is highlighted in all graphs. A) Acid to aldehyde ratios of vanillyl (Ac/Alv) and syringyl (Ac/Als) phenols. Values
above 0.6 indicate highly degraded lignin (Goñi, 1997). B) The ratio of syringyl to vanillyl phenols (S/V), which can be used to distinguish angiosperm and gymnosperm sources.
Angiosperms are characterized by S/V ratios of 0.6–4 and thus all the Lake Malawi samples plot within the range of angiosperm tissue. C) The C/V ratio, indicative of woody vs. nonwoody inputs is plotted in black. The δ13CTOC record is plotted in gray on top of the C/V record for comparison. D) The lignin phenol vegetation index (LPVI). Angiosperm woods are
characterized by LPVI values of 67 to 415 (with a mean value of 181), while non-woody angiosperm tissues are characterized by LPVI values of 378 to 2782 (with a mean value of
1090) (Tareq et al., 2004). E) Λ 8 values (the sum of vanillin, acetovanillone, vanillic acid, syringealdehyde, acetosyringone, syringic acid, p-hydroxycinnamic acid and ferulic acid) are
indicated by the black circles and are plotted next to the Lake Malawi TEX86 temperature record (Powers et al., 2005), which is indicated by the gray line.
106
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
of 1, 19, 181 and 1090, respectively. The LPVI of Lake Malawi samples
ranges from 27 to 878, and reveals similar trends to the C/V ratio. This
record indicates increased inputs of non-woody angiosperm tissue
from ~ 12–13 cal ka (Fig. 5D).
purely algal remains. Although algal inputs are undoubtedly an
important source of organic matter to Lake Malawi sediments, they do
not appear to be the main influence on δ13CTOC values.
5.2. n-alkane ACL and CPI
5. Discussion
5.1. Bulk vs. molecular geochemical records
The Lake Malawi TOC, C/N, and δ13CTOC records of core M98-1P
have been examined previously and relationships between these
parameters are complex, with multiple scenarios existing that could
explain observed trends in the data (Filippi and Talbot, 2005). It is not
always clear in parts of the Lake Malawi record whether changes
observed in bulk organic geochemical parameters are reflecting
changes in algal productivity, changes in the delivery of terrestrial
organic matter to the lake, or a combination of both. Generally,
observed trends in the bulk geochemical data are that higher TOCMAR
is associated with increased BSiMAR, higher Corg/Ntot ratios, more
enriched δ13CTOC values, and decreased temperature (Fig. 2). The
relationships between higher TOCMAR, higher Corg/Ntot ratios and
more enriched δ13CTOC values can be explained by variations in
terrestrial inputs to Lake Malawi. Increased delivery of terrestrial
organic matter to the lake could increase TOCMAR values and Corg/Ntot
ratios. In this scenario, if most of the organic matter is terrestrially
derived, shifts noted in the δ13CTOC record would reflect changes in C3
vs. C4 vegetation in the watershed. Alternatively, the trends noted in
the Lake Malawi bulk geochemical records can also be explained by
variations in aquatic productivity. The observed relationship between
high TOCMAR or high BSiMAR values and lower temperature may be a
reflection of changes in the prevalence of northerly winds or in the
thermal stratification of Lake Malawi. Northerly winds promote
upwelling in the north basin of the lake, and cooling of surface waters
could decrease thermal stratification making it easier for windinduced upwelling to occur. In this scenario, changes in algal productivity could be the dominant control on δ13CTOC values.
The molecular data produced in this study, including C/V ratios
and n-alkane δ13C, sheds light on the origin of organic matter in Lake
Malawi. C/V ratios, which likely reflect changes in non-woody vs.
woody vegetation (discussed further in Section 5.3), generally track
changes noted in the δ13CTOC record (Fig. 5C), suggesting a terrestrial
control on the carbon isotopic signature of Lake Malawi organic matter. Similarly, n-alkane δ13C values (hereafter referred to as δ13Calk;
Castañeda et al., 2007) also track changes noted in the δ13CTOC record
(Fig. 2), offering support to the hypothesis that the Lake Malawi
δ13CTOC record primarily reflects changes in C3 vs. C4 vegetation. An
exception to this pattern between δ13Calk and δ13CTOC values occurs
during the Younger Dryas, when the δ13Calk record indicates more 13Cenriched values while δ13CTOC values become slightly more 13Cdepleted (Fig. 2). The Younger Dryas clearly stands out in Lake Malawi
TOCMAR and BSiMAR records (Fig. 2). This event has been previously
described in Lake Malawi and is marked by a 2 °C cooling of surface
waters (Powers et al., 2005) and increased diatom productivity, which
resulted from increased northerly winds enhancing upwelling in the
northern basin of the lake (Johnson et al., 2002; Filippi and Talbot,
2005). It appears that during the Younger Dryas, algal productivity
imparted a greater signature on the δ13CTOC record than did terrestrial
inputs. However, for the majority of the record, changes in terrestrial
vegetation appear to be the main driver of variability in the δ13CTOC
record. Thus, we suggest that the δ13CTOC signal in Lake Malawi primarily
reflects changes in C3 vs. C4 vegetation in the watershed and that
terrestrial inputs to Lake Malawi are more important than previously
reported (e.g. Filippi and Talbot, 2005). These results are in contrast to
the results of Filippi and Talbot (2005) who examined the hydrogen
index (HI) of Lake Malawi sediments and concluded that the organic
matter is primarily of algal origin, although no parts of the core contain
n-alkanes are transported to marine and lacustrine sediments via
wind and runoff (Simoneit, 1977) but in arid environments wind can
be the dominant transport mechanism (Schefuß et al., 2003b). In Lake
Malawi, accumulation rates of n-alkanes may reflect transport history
rather than the amount of biomass in the watershed. Lake Malawi is
situated at the southernmost extent of the ITCZ and normally the
regional winds are from the south (Nicholson, 1996), but a strong
diurnal component of onshore/offshore wind patterns is superimposed (Hamblin et al., 2003). Previous studies of Lake Malawi
have provided strong evidence for periods in the past when the ITCZ
was located at a more southerly position than today (Johnson et al.,
2002; Filippi and Talbot, 2005; Brown and Johnson, 2005). During
these intervals, it is likely that ITCZ-driven and diurnal temperaturedriven changes in the wind regime produced stronger or more
frequent northerly winds over Lake Malawi, resulting in increased
upwelling in the northern basin of the lake, and, in turn, increased
diatom productivity (Johnson et al., 2002). It is likely that ITCZ-driven
changes in the wind regime over Lake Malawi have also affected the
transport of plant leaf waxes to the lake because winds are a major
feature of modern East African climate.
CPI values may provide some insight into the wind transport
history of n-alkanes. Plots of CPI values and mass accumulation rates
of biogenic silica (BSiMAR), a proxy for diatom productivity, display the
same general trends with higher CPI values noted during times of
increased BSiMAR values and vice versa (Fig. 4A), although major
differences in sampling resolution exist between the two records (796
BSi measurements compared to 61 n-alkane measurements). A likely
explanation for this relationship is that periodic increased northerly
winds over Lake Malawi, which lead to increased upwelling and diatom
productivity in the northern basin of the lake (Johnson et al., 2002),
enhance wind erosion and transport of plant leaf waxes to the lake. The
highest CPI values (5–7) observed in the Lake Malawi sedimentary
record are found during intervals of increased BSiMAR and clearly
indicate a terrestrial plant source. Lake Malawi samples have an average
CPI value of 3.9, also indicating a terrestrial plant source. However, in
parts of the core CPI values are as low as 1.7, which may suggest an
additional source of n-alkanes to Lake Malawi sediments (Fig. 4A). Low
CPI values (b3) can indicate hydrocarbon contamination by petroleum
(Bray and Evans,1961), and hydrocarbon seeps have been reported from
nearby Lake Tanganyika (Simoneit et al., 2000). Although hydrocarbon
seeps have not been reported from Lake Malawi, their presence cannot
be ruled out. However, vegetation shifts indicated by the δ13Calk record
are supported by the lignin phenol records (discussed in detail below),
and these trends would not be expected if Lake Malawi sediments
contained significant amounts of n-alkanes from hydrocarbon sources.
We note that wood burning is a possible source of low CPI n-alkanes
(Standley and Simoneit, 1987), and dry-season fires are a natural and
common occurrence in East Africa.
An interesting feature of the Lake Malawi n-alkane record is the
observed relationship between the ACL and temperature. In general,
shorter ACLs are present during periods of lower temperature while
longer ACLs are present during periods of higher temperature (Fig. 4B).
Relationships between ACL and temperature have been previously
suggested (e.g. Gagosian and Peltzer, 1986; Kawamura et al., 2003;
Zhang et al., 2006). Kawamura et al. (2003) found that trees growing in
tropical zones had higher ACL values compared to trees growing in
temperate or sub-temperate zones, although this observation was based
on a limited data set. Similarly, Zhang et al. (2006) note lower n-alkanol
ACL values during colder periods and higher ACL values during warm
periods in samples from the Chinese Loess Plateau that span the past
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
170 kyr. Rommerskirchen et al. (2003) suggest that leaf surface
temperature, which is related to either mean daily maximum temperature or absolute maximum temperature, may be a key factor affecting
the ACL. Increasing the ACL raises the melting point of protective leaf
surface waxes and thus plants growing at higher temperatures may
synthesize longer chain n-alkanes in order to maintain the protective
waxy coating on their leaves (Rommerskirchen et al., 2003). The
observed relationship between ACL and temperature in this Lake Malawi
core offers further support for this hypothesis.
Conversely, it has been suggested that a direct relationship exists
between longer chain n-alkanes and increased aridity (Peltzer and
Gagosian, 1989; Poynter et al., 1989; Schefuß et al., 2003b; Liu and
Huang, 2005). However, this relationship does not appear to hold true
for the Lake Malawi samples, especially in the Late Pleistocene. For
example, it is known that Lake Malawi was cooler (3.5 °C lower annual
mean lake surface temperature than today; Powers et al., 2005) and
significantly more arid during the LGM, with the lowest and most
sustained lake level lowstand occurring from 22.8–16 cal ka (Gasse et al.,
2002). This lowstand was then followed by a sharp rise in lake level from
15.7–15 cal ka, with a maximum highstand lasting until 13 cal ka (Gasse
et al., 2002). Comparing this lake level (aridity) information data to ACL
values, we note low ACL from ~22 to 18 cal ka (Fig. 4B), which centers on
the lake-level lowstand. ACL values gradually increase to higher values
until ~15.5 cal ka, and then fluctuate but remain high until ~13 cal ka
(Fig. 4B). If a relationship exists between longer chain n-alkanes and
aridity, we would expect to find higher ACL values during the LGM when
it was more arid followed by a trend towards gradually lower ACL values
from ~15–13 cal ka when more humid conditions were present, but in
fact the opposite trends are observed.
Another interesting feature of the Lake Malawi record is that
individual mass accumulation rates of C25, C27, C29 and C31 n-alkanes
display maxima at different times (Fig. 3). For example, the C25 n-alkane
reaches maximum abundance at 19.2 cal ka, the C27 n-alkane reaches
maximum abundance at 10.5 cal ka, while maximum abundances of the
C29–C31 n-alkanes are noted at 12.3, 7.8 and 5.3 cal ka. These differences
in the dominant homologue suggest varying sources of n-alkanes to
Lake Malawi over the past 23 cal ka. However, at present, no surveys of
lipid distributions in modern East African vegetation have been
conducted that might explain the observed shifts in the dominant
homologues. It should be noted that submerged and floating-leaf
aquatic macrophytes have been found to contain increased abundances
of C23 and C25 n-alkanes (Ficken et al., 2000). However, in Lake
Malawi aquatic macrophytes are not believed to be an important source
of n-alkanes as the steep sides of the northern lake basin (Fig.1B) restrict
the area of the littoral zone.
5.3. Lignin phenols
It is interesting to note that in core M98-1P, changes in the C/V
ratio, which reflect inputs of woody (i.e. trees) vs. non-woody (i.e.
grasses) tissue, may reflect variations in the amount of C3 vs. C4
vegetation present in the watershed. The overall trends observed
between the C/V ratios and the δ13Calk record (Fig. 6) suggests that
during periods of wetter climate, as reflected by depleted δ13Calk
values, woody C3 plants increased in abundance thereby decreasing C/V
values. Similarly, during periods of drier conditions, indicated by
enriched δ13Calk values, C4 grasses expanded and inputs of non-woody
tissue increased, thus increasing C/V values. Nearly all trees utilize the C3
photosynthetic pathway while in C4 grasses are a common component
of savanna vegetation zones in tropical Africa. Thus, the variations noted
in the C/V ratio are consistent with, and therefore supportive of, the
δ13Calk based C3/C4 interpretations.
Recent applications of lignin phenols to track C3/C4 shifts in
semitropical savannas in southern Texas illustrate that C/V molecular
parameters cannot be used to distinguish between input of the shoots
of C4 grasses and the leaves of C3 woody plants as they have similar
107
values (Filley et al., 2008b). However, the roots and wood of C3 tissue
is greatly lower in C/V than grasses and their associated roots.
Therefore, in order for surface litter to have provided the distinction in
plant source to core M98-1P, wood tissue would have to have been
mobilized to Lake Malawi sediments. If the plant carbon in soil organic
matter, which may be predominantly derived from root tissue (e.g.
Rasse et al., 2005) is delivered to the lake sediments, the roots may be
the plant tissues that provide the distinctions in source. It is important
to note that changing C/V ratios in sediments and decayed plant
matter can also indicate the degree of degradation of a plant source, as
was shown in a long-term decomposition study where the C/V ratio
declined with time during litter decay (Opsahl and Benner, 1995).
However, in core M98-1P the highest C/V values are not observed in
the youngest sediment samples but instead at ~12 cal ka (Fig. 5C).
Additionally, in the older section of this core (13 to 23 cal ka) the C/V
ratio displays an overall increase with increasing age, suggesting that
the C/V ratio in Lake Malawi mainly reflects vegetation shifts rather
than a decay signal.
Another notable feature of the Lake Malawi lignin record is that
increased Λ8 values generally correlate with increased temperatures
(Fig. 5E). This relationship may be related to changes in precipitation
(aridity) as warm/wet and cool/dry conditions are generally
associated in East Africa (compare Johnson et al., 2002; Brown and
Johnson, 2005; Powers et al., 2005). For example, during the LGM
when conditions were cooler (Powers et al., 2005) and more arid
(Gasse et al., 2002), low Λ8 values are noted (Fig. 5E). Following the
LGM a trend of generally increasing Λ8 values continues until ~ 13 cal
ka. Gradually increasing Λ8 values from the LGM until ~ 13 cal ka are
consistent with a transition to a more humid climate at this time
(Gasse et al., 2002) because enhanced lignin transport would have
occurred in response to increased runoff. Transport of lignin phenols
by wind is unlikely because these compounds are primarily
transported by water (Goñi et al., 1997; Meyers, 2003) and in Lake
Malawi delivery of terrigenous material to the lake is related to rainfall
(Johnson and McCave, 2008).
Alternatively, increased Λ8 values could result from an increase in
the amount of biomass present in the Lake Malawi watershed, with
more biomass present during wetter intervals. Another issue to
consider is that changes in the amount of stored litter, woody debris
and soil organic carbon also have been observed to occur with the
transition from grassland to woodland (Liao et al., 2006). When
grasslands are converted to woodlands, soil C and N concentrations
increase and the soil fractions responsible for C-protection shift from
free silt and clay in grasslands to microaggregate and aggregate
protected C in woodlands, which, in turn, influences organic carbon
preservation (Liao et al., 2006). Organic carbon associated with
macroaggregates and free particulate organic matter is less decomposed than carbon present in microaggregates and silt and clay
fractions (Liao et al., 2006). Such changes have likely imparted a
signature on the Lake Malawi lignin phenol record but past changes in
soil organic carbon preservation and storage are difficult to evaluate.
However, in core M98-1P, the general agreement between the δ13Calk
and lignin C/V records provides robust evidence for past vegetation
changes in the Lake Malawi watershed.
5.4. Paleovegetation history of Lake Malawi
5.4.1. Previous studies
Previous studies of vegetation history from the Malawi region are
limited with a major focus on the origin of Afromontane grasslands,
which are characterized by small patches of forest surrounded by
grasslands (Meadows and Linder, 1993). A debate focused on whether
these grasslands existed as the result of recent anthropogenic land
burning or if they occurred naturally. Meadows (1984) examined
pollen records from a series of peat cores from the Nyika Plateau in
northern Malawi and concluded that small patches of montane forest
108
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
Fig. 6. A) Summary of Lake Malawi records. The Younger Dryas (YD) is highlighted by the gray shading. The C/V ratio is indicated by the solid black circles and the weighted mean
n-alkane δ13C record (Castañeda et al., 2007) is shown with the open squares. The Lake Malawi temperature record, based on TEX86, is shown by the solid gray line (Powers et al.,
2005). Note that the x-axis scale has been reversed to facilitate comparison with the vegetation records. B) January insolation at 10°S (Berger and Loutre, 1991). C) Generalized
summary of Lake Rukwa pollen zones (data from Vincens et al., 2005). Zone IV displays a dominance of Afromontane pollen (found above 1800 m today) and increased
herbaceous taxa, indicating arid conditions. Zone III is characterized by a decrease in Afromontane taxa and an increase in woodland arboreal taxa, indicating a shift towards more
humid conditions. Zone II is characterized by diverse woodland and bushland vegetation, indicating wet conditions. Zone 1 is characterized by scarce arboreal taxa and increased
percentages of grass and sedge pollen, indicating a return to dry conditions. D) Generalized summary of conditions at Lake Masoko based on pollen data and magnetic susceptibility records
(Garcin et al., 2006). In contrast to Lake Malawi, which was dry during the LGM and YD, Lake Masoko experienced wet conditions at these times and was especially wet during the YD.
surrounded by montane grasslands were present in Malawi for at least
the past 5000 years and possibly as far back as the past 12,000 years.
These pollen records indicate significant variability in the vegetation
assemblages of the Nyika Plateau during the past 5000 years but the
limited chronology on the peat cores restricted comparisons to other
East African paleoenvironmental proxies (Meadows, 1984). DeBusk
(1998) examined the pollen records of two Lake Malawi sediment
cores from the central and southern basins of the lake. These records
extend back to 37,500 BP, possibly capturing the LGM in East Africa but
this study is also hampered by the possibility of hiatuses in the cores
and uncertainties in the chronology. DeBusk (1998) reached similar
conclusions as Meadows (1984), that small patches of forest
surrounded by montane grasslands were present throughout the
Holocene and did not result from recent anthropogenic activities. To
date, the most robust records of vegetation change from the Malawi
region come from well-dated pollen records of two lakes that lie to the
north of Lake Malawi, Lake Rukwa (Vincens et al., 2005) and Lake
Masoko (Vincens et al., 2003; Garcin et al., 2006). These two records,
located in close proximity to each other, also indicate considerable
variability in East African vegetation during the Late Pleistocene and
Holocene, and, for certain time intervals, reach differing conclusions
regarding climatic conditions in East Africa.
Many uncertainties exist regarding paleovegetation histories of
East Africa due to the climatic and topographic complexity of the
region. For example, Lake Malawi is presently situated within a C3
dominated tree savanna vegetation zone, but to both the north and
south of the lake C4 dominated grass savanna vegetation zones are
present (Fig. 1). Previous studies of tropical Africa have demonstrated
that aridity is the main factor controlling the distribution of C3 vs. C4
vegetation (Castañeda et al., 2007; Schefuß et al., 2003a), and thus it is
likely that the different vegetation zones have expanded/contracted
and migrated during the past 23 cal ka in response to changes in
precipitation. Superimposed on this regional variability is local
complexity that results from the steep topography surrounding Lake
Malawi. Mountains in the northern basin of Lake Malawi rise sharply
from the lake and reach elevations of over 2500 m. Vegetation along
this gradient is zoned and ranges from types that are limited to the
lakeshore, to closed-canopy (wetter) and open-canopy (drier)
miombo woodlands at low elevations, to montane grasslands with
patches of Afromontane forests at high elevations (above 1500 m)
(DeBusk, 1998). These vegetation zones may have migrated to higher
or lower elevations in response to climate variability.
5.4.2. Late Pleistocene–Holocene vegetation history of Lake Malawi
During the LGM, conditions in Lake Malawi were 3.5 °C cooler
(Powers et al., 2005) and significantly more arid than at present
(Gasse et al., 2002), with a lake level lowering of 100 m (Scholz et al.,
2007). Elevated C/V ratios and more enriched δ13Calk values indicate
that Lake Malawi was surrounded by a greater proportion of C4 grasses
at this time (Fig. 6). In nearby Lake Rukwa, located to the north of Lake
Malawi (Fig. 1C; ~ 280 km between coring sites), a well-dated pollen
record indicates lowering of montane taxa from 23 to 19 cal ka
(Vincens et al., 2005). Similar changes have been noted during the
LGM at other East African sites (Van Zinderen Bakker, 1969; Kendall,
1969; Livingstone, 1971; Vincens et al., 1993). In contrast, conditions at
Lake Masoko, located between Lakes Malawi and Rukwa, were wetter
during the LGM (Garcin et al., 2006).
Following the LGM, temperatures in Lake Malawi increased
steadily from ~ 24 °C at 23 cal ka to ~ 28 °C at 13.8 cal ka (Powers
et al., 2005) and precipitation appears to have increased throughout
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
this interval based on our δ13Calk and C/V data. Following low lake
levels during the LGM, Lake Malawi refilled rapidly beginning at
15.7 cal ka and reached a highstand at ~ 13 cal ka (Gasse et al., 2002).
Associated with these climate changes, C/V ratios and δ13Calk values
indicate a gradual transition from increased inputs of C4 vegetation
to increased inputs of C3 vegetation until ~ 13.6 cal ka (Fig. 6). This
shift from C4 dominated to C3 dominated vegetation following
the LGM is consistent with vegetation changes noted at other sites
in East Africa (Kendall, 1969; Vincens, 1991, 1993; DeBusk, 1998).
At Lake Rukwa, similar changes have been noted from ~ 17–11 cal ka
and this interval is characterized by an increase of arboreal taxa
accompanied by a decrease of grasses (Vincens et al., 2005). Progressive reforestation is noted beginning after 15 cal ka at Lake
Masoko (Garcin et al., 2006).
Following the post-LGM gradual warming period that lasted until
13.8 cal ka, temperatures in Lake Malawi decreased to ~25 °C at
12.5 cal ka (Powers et al., 2005). This cooling marks the Younger Dryas
in Lake Malawi, which is noted in both the bulk and molecular
geochemical records of M98-1P (Figs. 2, 5 and 6). Additionally, the
Younger Dryas is evident in almost all available geochemical records
from Lake Malawi (Johnson et al., 2002; Filippi and Talbot, 2005;
Powers et al., 2005; Talbot et al., 2007) and was clearly a period
characterized by significantly different climatic conditions. There is
strong evidence that during the Younger Dryas increased northerly
winds were present over Lake Malawi, which stimulated primary
productivity in the northern basin of the lake (Johnson et al., 2002;
Filippi and Talbot, 2005; Talbot et al., 2007). Our molecular data from
Lake Malawi also supports the Younger Dryas as a time of increased
primary productivity (Castañeda et al., in review) as well as increased
aridity. At this time, δ13Calk values indicate a major shift to more 13Cenriched values, similar to those noted during the LGM (Fig. 6). The
most 13C-enriched δ13Calk values are noted at ~ 12.5 cal ka, coincident
with the maximum cooling observed during the Younger Dryas. The C/V
ratio of lignin phenols and the LPVI both indicate major excursions to
higher values within the Younger Dryas interval (Fig. 5C, D),
suggesting increased inputs of non-woody angiosperm tissues at
this time, which is generally consistent with the δ13Calk data. It may
be possible that a drop in lake level occurred in association with the
Younger Dryas event, and thus the lignin phenol record may also
reflect reworking of terrestrial OM at this time. However, re-worked
OM might be expected to have elevated Ac/Al ratios (Dalzell et al.,
2007; Nierop and Filley, 2007) but an increase in Ac/Al values in
M98-1P is not noted during the Younger Dryas (Fig. 5). In contrast to
the major excursions noted in Lake Malawi geochemical records,
pollen records from Lake Rukwa do not indicate any major excursions
during the Younger Dryas (Vincens et al., 2005) while a decrease in
herb pollen and a simultaneous increase in Zambezian tree pollen
indicates wetter conditions at Lake Masoko during the Younger Dryas
(Garcin et al., 2006).
The early Holocene in Lake Malawi (~11.6 cal ka until ~7.7 cal ka) is
characterized by fluctuating but generally increased inputs of C4
vegetation, as indicated by the δ13Calk record (Fig. 6). The sampling
resolution of lignin phenols is lower than that of n-alkane δ13C
throughout this interval, but C/V ratios display increased inputs of
non-woody vegetation at ~9 cal ka. Temperature also fluctuates
throughout this interval with slightly cooler temperatures of 24–26 °C
prevailing (Powers et al., 2005). Generally enriched δ13Calk values are
consistent with previously noted Early Holocene arid conditions at
Lake Malawi (Filippi and Talbot, 2005; Gasse et al., 2002; Finney et al.,
1996). At Lake Masoko, more arid conditions are also noted beginning
at ~ 11.7 cal ka (Garcin et al., 2006). Early Holocene aridity at Lake
Malawi (and Lake Masoko) is anomalous compared to the other lakes
farther north in the East African Rift Valley, which experienced lake
level highstands or overflowing conditions at this time (Gasse, 2000).
Humid conditions are noted from 12.1 to 5.5 cal ka at Lake Rukwa
(Vincens et al., 2005) and a paleo-shoreline has been described from
109
~200 m above present day lake level, which is dated to 11–10.7 cal ka
(Delvaux et al., 1998).
From 7.7 until ~ 2 cal ka, increased inputs of C3 vegetation are noted
in Lake Malawi with the δ13Calk record indicating the highest
contributions of C3 vegetation at 4.9 cal ka and the lowest C/V value
occurring at 4 cal ka (Fig. 6). The temperature record indicates a
gradual warming from ~ 25 °C at 7.7 cal ka to ~ 29 °C at 5 cal ka (Fig. 6;
Powers et al., 2005). Expansion of forests during the middle Holocene
has been noted in many East African records (Bonnefille et al., 1995;
Ssemmanda and Vincens, 2002; Beuning et al., 1997) and has been
attributed to a strengthening of the African monsoon (DeMenocal
et al., 2000). In Lake Rukwa humid conditions initiate at 12.1 cal ka and
continue until 5.5 cal ka, with the maximum abundance and diversity
of arboreal taxa occurring during this time (Vincens et al., 2005).
The past 2 cal ka in Lake Malawi is marked by a return to greater
inputs of C4 vegetation, indicating a shift back to increasingly drier
conditions in East Africa, which appears to have initiated following peak
wet conditions at ~4–5 cal ka (Fig. 6). A gradual shift towards higher C/V
and LPVI values is noted from 4 cal ka to the present (Fig. 5C, D) while
more enriched δ13Calk values are noted from 4.9 cal ka to the present
(Fig. 6). In Lake Rukwa, increasingly arid conditions and similar
vegetation changes (decreased arboreal taxa and the expansion of
grasses) are noted from 5.5 cal ka to the present (Vincens et al., 2005). A
high-resolution pollen record from Lake Masoko, which extends back to
4.2 cal ka, indicates periods of dry conditions between 4.2 cal ka and the
present (Vincens et al., 2003). However, the Lake Masoko record, which
is of considerably higher resolution than the Lake Malawi lignin record,
also indicates periods of wetter conditions occurring within the past
4.2 cal ka (Vincens et al., 2003).
With exception of the Younger Dryas and early Holocene (11.6 to
7.7 cal ka), the Lake Rukwa pollen record and Lake Malawi molecular
records are in general agreement (Fig. 6). The often out-of-phase
behavior noted at Lake Masoko, which is situated between Lakes
Malawi and Rukwa, is intriguing. It appears that when conditions in
Lake Malawi were significantly more arid, such as during the LGM and
during the Younger Dryas, Lake Masoko experienced significantly
wetter conditions (Fig. 6). Given the extremely short distances
between these sites (b100 km between the Lake Malawi M98-1P
coring site and Lake Masoko and b200 km between Lakes Masoko and
Rukwa; Fig. 1C) it is difficult to envision such dramatically different
climate regimes. Considering the small size of Lake Masoko (area of
0.38 km2 compared with 22,490 km2 for Lake Malawi) (Barker et al.,
2003; Spigel and Coulter, 1996), and the fact that Lake Masoko is
located at an elevation of 840 m in the Rungwe volcanic highlands
surrounding northern Lake Malawi, it may be likely that Lake Masoko
is influenced more by local orographic effects whereas Lake Malawi
responds to more regional climatic forcing. Another intriguing
observation is that while rainfall at Lake Masoko is noted to be
strongly controlled by changes in low-latitude insolation (precession)
(Garcin et al., 2006), geochemical records from Lake Malawi indicate
greater complexity. Insolation is likely one of many factors influencing
the climate of southeast Africa; however, at Lake Malawi insolation
does not appear to be the dominant factor influencing climate (Fig. 6).
Instead, especially close ties are noted between conditions at Lake
Malawi and migrations of the ITCZ, with arid conditions accompanying southward ITCZ migrations (Castañeda et al., 2007). Strong ties are
also noted between conditions at Lake Malawi and the sea-surface
temperature gradient in the Atlantic Ocean off the western coast of
Africa (Castañeda et al., 2007), which controls the strength of the
southeast trade winds and the ability of Atlantic moisture to penetrate
into the interior of the African continent (Schefuß et al., 2005). The
apparent difference in the response of Lakes Malawi and Masoko to
regional climate forcing highlights the complexity of East African
(paleo)climates and demonstrates both the need to better understand
modern climatic processes in this region, and to obtain additional
well-dated and high-resolution paleoclimate records from East Africa.
110
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
6. Conclusions
1) Major vegetation changes have occurred in the Lake Malawi
watershed since the LGM and are noted in records of lignin phenols
and plant leaf wax biomarkers. The vegetation history of Lake
Malawi can be summarized into four main periods: 1) increased
abundances of C4 grasses during the LGM followed by a gradual
transition to increased abundances of C3 vegetation until 13.6 cal ka,
2) fluctuating but generally elevated abundances of C4 vegetation
from 13.6 to 7.7 cal ka, including the interval of the Younger Dryas, 3)
a middle Holocene shift towards greater C3 abundances from 7.7 to
~2 cal ka, and 4) a late Holocene shift to increased C4 abundances
from ~2 cal ka until the present. Aridity is likely the main factor
driving these vegetation changes.
2) In Lake Malawi sediments, the ratio of cinammyl to vanillyl (C/V)
lignin phenols likely reflects shifts in C4 grasses vs. C3 trees. During
periods of wetter climate, woody C3 plants increased in abundance
resulting in decreased C/V values and vice versa.
3) A relationship is noted between ACL values and temperature in
Lake Malawi samples with higher ACL values correlated with
higher temperatures. Higher temperatures may cause plants to
synthesize longer chain n-alkanes to increase the melting point of
their protective leaf surface waxes.
4) Higher CPI values correlate with higher BSiMAR values. Increased
BSiMAR values have been linked to upwelling associated with
stronger or more frequent northerly winds over Lake Malawi, and
therefore, the relationship between CPI and BSiMAR is likely reflecting
increased transport of higher plant leaf waxes to the lake during
periods of increased northerly winds.
5) The δ13CTOC record in Lake Malawi generally agrees with changes
observed in the n-alkane δ13C and lignin phenol C/V records, and
likely reflects shifts in C3 vs. C4 vegetation. Molecular data obtained
in this study indicates that terrestrial inputs have greater control
on δ13CTOC values in Lake Malawi than previously thought.
Acknowledgements
We wish to thank Yongsong Huang (Brown University) for providing
n-alkane δ13C analyses. Sarah Grosshuesch (Large Lakes Observatory),
Marcelo Alexandre (Brown University) and Dave Gamblin (Purdue
University) are thanked for providing analytical assistance. David
Hollander (University of South Florida) is thanked for providing δ13CTOC
analyses. Comments from Erik Brown (Large Lakes Observatory) on an
early version of this manuscript improved this paper. We thank the editor
and two anonymous reviewers for insightful comments that improved
and clarified this manuscript. This material is based in part upon work
supported by the National Science Foundation under Grant No. ATM9709291 to T. Johnson. Additional support was provided by a 2005
Research Corporation grant (CC6559) to J. Werne and fellowship support
from the Ford Foundation and University of Minnesota to I. Castañeda.
References
Barker, P., Williamson, D., Gasse, F., Gilbert, E., 2003. Climatic and volcanic forcing
revealed in a 50,000-year diatom record from Lake Massoko, Tanzania. Quaternary
Research 60, 368–376.
Barry, S.L., Filippi, M.L., Talbot, M.R., Johnson, T.C., 2002. Sedimentology and
geochronology of Late Pleistocene and Holocene sediments from northern Lake
Malawi. In: Odada, E.O., Olago, D.O. (Eds.), The East African Great Lakes: Limnology,
Palaeolimnology and Biodiversity: The Netherlands. Kluwer Academic Publishers,
pp. 369–391.
Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years.
Quaternary Science Reviews 10, 297–317.
Beuning, K.R.M., Talbot, M.R., Kelts, K., 1997. A revised 30,000-year paleoclimatic and
palaeohydrologic history of Lake Albert, East Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 136, 259–279.
Bianchi, T.S., Wysocki, L.A., Stewart, M., Filley, T.R., 2007. Temporal variability in
terrestrially-derived sources of particulate organic carbon in the lower Mississippi
River and its upper tributaries. Geochimica et Cosmochimica Acta 71, 4425–4437.
Bonnefille, R., Riollet, G., Buchet, G., Icole, M., Lafont, R., Arnold, M., Jolly, D., 1995.
Glacial/interglacial record from intertropical Africa, high resolution pollen and
carbon data at Rusaka, Burundi. Quaternary Science Reviews 14, 917–936.
Bray, E.E., Evans, E.D., 1961. Distribution of n-paraffins as a clue to recognition of source
beds. Geochimica et Cosmochimica Acta 22, 2–15.
Broecker, W.S., Kennett, J.P., Flower, B.P., Teller, J.T., Trumbore, S., Bonani, G., Wolfi, W.,
1989. Routing of meltwater from the Laurentide ice-sheet during the Younger Dryas
cold episode. Nature 341, 318–321.
Brown, E.T., Johnson, T.C., 2005. Coherence between tropical East African and South
American records of the Little Ice Age. Geochemistry, Geophysics, Geosystems, 6,
Q12005.
Castañeda, I.S., Werne, J.P., Johnson, T.C., 2007. Wet and arid phases in the southeast
African tropics since the Last Glacial Maximum. Geology 35, 823–826.
Castañeda, I.S., Werne, J.P., Johnson, T.C., in review. Influence of climate change on algal
community structure and primary productivity of Lake Malawi (East Africa) from
the Last Glacial Maximum to present. Limnology and Oceanography.
Cerling, T.E., Wang, Y., Quade, J., 1993. Expansion of C4 ecosystems as an indicator of
global ecological change in the late Miocene. Nature 399, 344–345.
Collatz, G.J., Berry, J.A., Clark, J.S.,1998. Effects of climate and atmospheric CO2 partial pressure
on the global grasses of C4 grasses: present, past and future. Oecologia 114, 441–454.
Collister, J.W., Rieley, G., Stern, B., Eglinton, G., Fry, B., 1994. Compound-specific δ13C
analyses from leaf lipids from differing carbon dioxide metabolisms. Organic
Geochemistry 21, 619–627.
Cranwell, P.A., 1973. Chain length distribution of n-alkanes from lake sediment in
relation to post-glacial environmental change. Freshwater Biology 3, 259–265.
Cranwell, P.A., 1981. Diagenesis of free and bound lipids in terrestrial detritus deposited
in a lacustrine sediment. Organic Geochemistry 3, 79–89.
Cranwell, P.A., Eglinton, G., Robison, N., 1987. Lipids of aquatic organisms as potential
contributors to lacustrine sediments — II. Organic Geochemistry 11, 513–527.
Dalzell, B.J., Filley, T.R., Harbor, J.M., 2005. Flood pulse influences on terrestrial organic
matter export from an agricultural watershed. Journal of Geophysical Research 110,
G02011. doi:10.1029/2005JG000043.
Dalzell, B.J., Filley, T.R., Harbor, J.M., 2007. The role of hydrology in annual organic carbon
loads and terrestrial organic matter export from a midwestern agricultural
watershed. Geochimica et Cosmochimica Acta 71, 1448–1462.
DeBusk, G.H., 1998. A 37,500-year pollen record from Lake Malawi and implications for
the biogeography of afromontane forests. Journal of Biogeography 25, 479–500.
Deines, P., 1980. The isotopic composition of reduced organic carbon. In: Fritz, P., Fontes,
J.C. (Eds.), Handbook of Environmental Geochemistry: The Terrestrial Environment.
Elsevier, pp. 329–406.
Delvaux, D., Kervyn, F., Vittori, E., Kajara, R.S.A., Kilembe, E., 1998. Late Quaternary
tectonic activity and lake level change in the Rukwa Rift basin. Journal of African
Earth Sciences 26, 397–421.
deMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Barker, L., Yarusinsky, M.,
2000. Abrupt onset and termination of the African humid period: rapid climate
responses to gradual insolation forcing. Quaternary Science Reviews 19, 347–361.
Eccles, D.H., 1974. An outline of the physical limnology of Lake Malawi (Lake Nyasa).
Limnology and Oceanography 19, 730–742.
Ehleringer, J.R., Cerling, T.E., Helliker, B.R., 1997. C4 photosynthesis, atmospheric CO2,
and climate. Oecologia 112, 285–299.
Eglinton, G., Hamilton, R.J., 1967. Leaf epicuticular waxes. Science 156, 1322–1335.
Ertel, J.R., Hedges, J.I., 1985. Sources of sedimentary humic substances: vascular plant
debris. Geochimica et Cosmochimica Acta 49, 2097–2107.
Ficken, K.J., Li, B., Swain, D.L., Eglinton, G., 2000. An n-alkane proxy for the sedimentary
input of submerged/floating freshwater aquatic macrophytes. Organic Geochemistry 31, 745–749.
Filippi, M.L., Talbot, M.R., 2005. Lake Malawi: the paleolimnology of the north basin over
the last 25 ky based on the elemental and stable isotopic composition of
sedimentary organic matter. Quaternary Science Reviews 24, 1303–1328.
Filley, T.R., McCormick, M.K., Crow, S.E., Szlavecz, K., Whigham, D.F., Johnston, C.T., van
den Heuvel, R.N., 2008a. Comparison of the chemical alteration trajectory of Liriodendron tulipifera L. leaf litter among forests with different earthworm abundance.
Journal of Geophysical Research 113, G01027.
Filley, T.R., Button, T.W., Liao, J.D., Jastrow, J.D., Gamblin, D.E., 2008b. Chemical changes
to nonaggregated particulate soil organic matter following grassland-to-woodland
transition in a subtropical savanna. Journal of Geophysical Research 113, G03009.
Filley, T.R., Freeman, K.H., Bianchi, T.S., Baskaran, M., Colarusso, L.A., Hatcher, P.G., 2001.
An isotopic biogeochemical assessment of shifts in organic matter input to
Holocene sediments from Mud Lake, Florida. Organic Geochemistry 32, 1153–1167.
Finkelstein, D.B., Pratt, L.M., Brassell, S.C., 2006. Can biomass burning produce a globally
significant carbon-isotope excursion in the sedimentary record? Earth and
Planetary Science Letters 250, 501–510.
Finney, B.P., Scholz, C.A., Johnson, T.C., Trumbore, S., 1996. Late Quaternary lake-level
changes of Lake Malawi. In: Johnson, T.C., Odada, E.O. (Eds.), The Limnology,
Climatology and Paleoclimatology of the East African Lakes. Gordon and Breach
Publishers, Amsterdam, pp. 495–508.
Fuhrmann, A., Mingram, J., Lücke, A., Lu, H., Horsfield, B., Liu, J., Negendank, J.F.W.,
Schlesser, G.H., Wilkes, H., 2003. Variations in organic matter composition in
sediments from Lake Huguang Maar (Huguangyan), south China during the last
68 ka: implications for environmental and climatic change. Organic Geochemistry
34, 1497–1515.
Gagosian, R.B., Peltzer, E.T., 1986. The importance of atmospheric input of terrestrial
material to deep sea sediments. Organic Geochemistry 10, 661–669.
Garcin, Y., Vincens, A., Williamson, D., Guiot, J., Buchet, G., 2006. Wet phases in tropical
southern Africa during the last glacial period. Geophysical Research Letters 33,
L07703. doi:10.1029/2005GL025531.
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
Gasse, F., 2000. Hydrological changes in the African tropics since the Last Glacial
Maximum. Quaternary Science Reviews 19, 189–211.
Gasse, F., Barker, P., Johnson, T.C., 2002. A 24,000 yr diatom record from the northern
basin of Lake Malawi. In: Odada, E.O., Olago, D.O. (Eds.), The East African Great
Lakes: Limnology, Palaeolimnology and Biodiversity. Kluwer Academic Publishers,
The Netherlands, pp. 393–414.
Giger, W., Schaffner, C., Wakeham, S.G., 1980. Aliphitac and olefinic hydrocarbons in
recent sediments of Greifensee, Switzerland. Geochimica et Cosmochimica Acta 44,
119–129.
Goñi, M.A., Hedges, J.I., 1992. Lignin dimers: structures, distribution, and potential
geochemical applications. Geochimica et Cosmochimica Acta 56, 4025–4043.
Goñi, M.A., Nelson, B.C., Blanchette, R.A., Hedges, J.I., 1993. Fungal degradation of wood
lignins: geochemical perspectives from CuO-derived phenolic dimers and monomers. Geochimica et Cosmochimica Acta 57, 3985–4002.
Goñi, M.A., Ruttenberg, K.C., Eglinton, T.I., 1997. Sources and contribution of terrigenous
organic matter to surface sediments in the Gulf of Mexico. Nature 389, 275–278.
Goñi, M.A., 1997. Record of terrestrial organic matter composition in Amazon Fan
sediments. Proceedings of the Ocean Drilling Program, Scientific Results 155, 519–530.
Hamblin, P.F., Bootsma, H.A., Hecky, R.E., 2003. Surface meteorological observations
over Lake Malawi/Nyasa. Journal of Great Lakes Research 29, 19–33.
Hedges, J.I., Mann, D.C., 1979a. The characterization of plant tissues by their lignin
oxidation products. Geochimica et Cosmochimica Acta 43, 1803–1807.
Hedges, J.I., Mann, D.C., 1979b. The lignin geochemistry of marine sediments from the
southern Washington coast. Geochimica et Cosmochimica Acta 43, 1809–1818.
Hedges, J.I., Ertel, J.R., 1982. Characterization of Lignin by Gas Capillary Chromatography
of Cupric Oxide Oxidation-Products. Analytical Chemistry 54, 174–178.
Hedges, J.I., Ertel, J.R., Leopold, E.B.,1982. Lignin geochemistry of a late Quaternary sediment
core from Lake Washington. Geochimica et Cosmochimica Acta 46, 1869–1877.
Hu, F.S., Hedges, J.I., Gordon, E.S., Brubaker, L.B., 1999. Lignin biomarkers and pollen in
postglacial sediments of an Alaskan lake. Geochimica et Cosmochimica Acta 63,
1421–1430.
Huang, Y., Freeman, K.H., Eglinton, T.I., 1999. δ13C analyses of individual lignin phenols
in Quaternary lake sediments: a novel proxy for deciphering past terrestrial
vegetation changes. Geology 27, 471–474.
Huang, Y., Street-Perrott, F.A., Metcalfe, S.E., Brenner, M., Moreland, M., Freeman, K.H.,
2001. Climate changes as the dominant control on glacial–interglacial variations in
C3 and C4 plant abundance. Science 293, 1647–1651.
Hyodo, F., Tsugeki, N., Azuma, J., Urabe, J., Nakanishi, M., Wada, E., 2008. Changes in
stable isotopes, lignin-derived phenols, and fossil pigments in sediments of Lake
Biwa Japan: implications for anthropogenic effects over the last 100 years. Science
of The Total Environment 403, 139–147.
Ishiwatari, R., Yamamoto, S., Uemura, H., 2005. Lipid and lignin/cutin compounds in
Lake Baikal sediments over the last 37 kyr: implications for glacial–interglacial
palaeoenvironmental change. Organic Geochemistry 36, 327–347.
Ishiwatari, R., Yamamoto, S., Shinoyama, S., 2006. Lignin and fatty acid records in Lake
Baikal sediments over the last 130 kyr: a comparison with pollen records. Organic
Geochemistry 37, 1787–1802.
Johnson, T.C., Davis, T.W., 1989. High-resolution seismic profiles from Lake Malawi,
Africa. Journal of African Earth Sciences 8, 383–392.
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. 367–409.
Johnson, T.C., Brown, E.T., McManus, J., Barry, S., Barker, P., Gasse, F., 2002. A highresolution paleoclimate record spanning the past 25,000 years in southern East
Africa. Science 256, 113–132.
Johnson, T.C., McCave, I.N., 2008. Transport mechanism and paleoclimatic significance
of terrigenous silt deposited in varved sediments of an African rift lake. Limnology
and Oceanography 53, 1622–1632.
Kawamura, K., Ishimuma, Y., Yamazaki, K., 2003. Four years' observations of terrestrial
lipid class compounds in marine aerosols from the western North Pacific. Global
Biogeochemical Cycles 17, 1003. doi:10.1029/2001GB001810.
Kendall, R.L., 1969. An ecological history of the Lake Victoria basin. Ecological
Monographs 39, 121–176.
Kuypers, M.M.M., Pancost, R.D., Sinninghe Damste, J.S., 1999. A large and abrupt fall in
atmospheric CO2 concentration during Cretaceous times. Nature 399, 342–345.
Leroux, M., 2001. The Meteorology and Climate of Tropical Africa. Chichester, Springer,
Praxis Publishing, 548 p.
Liao, J.D., Button, T.W., Jastrow, J.D., 2006. Storage and dynamics of carbon and nitrogen
in soil physical fractions following woody plant invasion of grassland. Soil Biology &
Biochemistry 38, 3184–3196.
Liu, W., Huang, Y., 2005. Compound-specific D/H ratios and molecular distributions of
higher plant leaf waxes as novel paleoenvironmental indicators in the Chinese
Loess Plateau. Organic Geochemistry 36, 851–860.
Livingstone, D.A., 1971. A 22,000-year pollen record from the plateau of Zambia.
Limnology and Oceanography 16, 349–356.
Meadows, M.E., 1984. Late Quaternary vegetation history of the Nyika Plateau, Malawi.
Journal of Biogeography 11, 209–222.
Meadows, M.E., Linder, H.P., 1993. A palaeoecological prespective on the origin of
Afromontane grasslands. Journal of Biogeography 20, 345–355.
Meyers, P.A., 2003. Applications of organic geochemistry to paleolimnological
reconstructions: a summary of examples from the Laurentian Great Lakes. Organic
Geochemistry 34, 261–289.
Nierop, K.G.J., Filley, T.R., 2007. Assessment of lignin and (poly-)phenol transformations
in oak (Quercus robur) dominated soils by 13C-TMAH thermochemolysis. Organic
Geochemistry 38, 551–565.
111
Nicholson, S.E.,1996. A review of climate dynamics and climate variability in Eastern Africa.
In: Johnson, T.C., Odada, E.O. (Eds.), The Limnology, Climatology and Paleoclimatology
of the East African Lakes. Gordon and Breach Publishers, Amsterdam, pp. 25–56.
O'Leary, M.H., 1981. Carbon isotope fractionation in plants. Phytochemistry 20, 553–567.
Opsahl, S., Benner, R., 1995. Early diagenesis of vascular plant tissues; lignin and cutin
decomposition and biogeochemical implications. Geochimica et Cosmochimica
Acta 59, 4889–4904.
Orem, W.H., Colman, S.M., Lerch, H.E., 1997. Lignin phenols in sediments of Lake Baikal,
Siberia: application to paleoenvironmental studies. Organic Geochemistry 27, 153–172.
Pagani, M., Arthur, M.A., Freeman, K.H., 1999. Miocene evolution of atmospheric carbon
dioxide. Paleoceanography 14, 273–292.
Peltzer, E.T., Gagosian, R.B., 1989. Organic geochemistry of aerosols over the Pacific
Ocean. In: Riley, J.P., Chester, R. (Eds.), Duce, R.A. (guest Ed.), Chemical
Oceanography 10, SEAREX: The Sea/Air Exchange Program, Academic Press, New
York, pp. 281–338.
Pempkowiak, J., Tylmann, W., Staniszewski, A., Gołębiewski, R., 2006. Lignin
depolymerization products as biomarkers of the organic matter sedimentary
record in 210Pb–137Cs-dated lake sediments. Organic Geochemistry 37, 1452–1464.
Powers, L.A., Johnson, T.C., Werne, J.P., Castañeda, I.S., Hopmans, E.C., Sinninghe Damste,
J.S., Schouten, S., 2005. Large temperature variability in the southern African tropics
since the Last Glacial Maximum. Geophysical Research Letters 32, L08706.
Poynter, J.G., Farrimond, P., Robinson, N., Eglinton, G., 1989. Aeolian-derived higher
plant lipids in the marine sedimentary record. In: Leinen, M., Sarnthein, M. (Eds.),
Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global
Atmospheric Transport. Kluwer Academic Press, Norwell, Mass, pp. 435–462.
Rasse, D.P., Rumpel, C., Dignac, M.-F., 2005. Is soil carbon mostly root carbon?
Mechanisms for a specific stabilization. Plant Soil 269, 341–356.
Raven, P.H., Event, R.F., Eichnorn, S.E., 1999. Biology of Plants. W.H. Freeman and
Company, New York. 875 p.
Riley, G., Collier, R.J., Jones, D.M., Eglington, G., Eakin, P.A., Fallick, A.E., 1991. Sources of
sedimentary lipids deduced from stable isotope analyses of individual compounds.
Nature 352, 425–427.
Rommerskirchen, F., Eglinton, G., Dupont, L., 2003. A north to south transect of Holocene
southeast Atlantic continental margin sediments: relationship between aerosol
transport and compound-specific δ13C land plant biomarker and pollen records.
Geochemistry, Geophysics, Geosystems 4, 1101. doi:10.1029/2003GC000541.
Rosendahl, B.R., 1987. Architecture of continental rifts with special reference to East
Africa. Annual Reviews of Earth and Planetary Science 15, 445–504.
Schefuß, E., Schouten, S., Jansen, J.H.F., Sinninghe Damste, J.S., 2003a. African vegetation
controlled by tropical sea surface temperatures in the mid-Pleistocene period.
Nature 422, 418–421.
Schefuß, E., Ratmeyer, V., Stuut, J.-B.W., Jansen, J.H.F., Sinninghe Damste, J.S., 2003b.
Carbon isotope analyses of n-alkanes in dust from the lower atmosphere over the
central eastern Atlantic. Geochimica et Cosmochimica Acta 67, 1757–1767.
Schefuß, E., Schouten, S., Schneider, R.R., 2005. Climatic controls on central African
hydrology during the past 20,000 years. Nature 437, 1003–1006.
Scholz, C.A., Rosendahl, B.R., 1988. Low lake stands in Lakes Malawi and Tanganyika,
East Africa, delineated with multifold seismic data. Science 240, 1645–1648.
Scholz, C.A., Johnson, T.C., Cohen, A.S., King, J.W., Peck, J.A., Overpeck, J.T., Talbot, M.R.,
Brown, E.T., Kalindekafe, L., Amoako, P.Y.O., Lyons, R.P., Shanahan, T.M., Castañeda, I.S.,
Heil, C.W., Forman, S.L., McHargue, L.R., Beuning, K.R., Gomez, J., Pierson, J., 2007. East
African megadroughts between 135 and 75 thousand years ago and bearing on earlymodern human origins. Proceedings of the National Academy of Sciences 104,
16416–16421.
Schwark, L., Zink, K., Lechterbeck, J., 2002. Reconstruction of postglacial to early
Holocene vegetation history in terrestrial Central Europe via cuticular lipid
biomarkers and pollen records from lake sediments. Geology 30, 463–466.
Simoneit, B.R.T., 1977. Diterpenoid compounds and other lipids in deep-sea sediments
and their geochemical significance. Geochimica et Cosmochimica Acta 41, 463–476.
Simoneit, B.R.T., Aboul-Kassim, T.A.T., Tiercelin, J.J., 2000. Hydrothermal petroleum from
lacustrine sedimentary organic matter in the East African Rift. Applied Geochemistry 15, 355–368.
Spigel, R.H., Coulter, G.W., 1996. Comparison of hydrology and physical limnology of the
East African Great Lakes: Tanganyika, Malawi, Victoria, Kivu and Turkana (with
reference to some North American Great Lakes). In: Johnson, T.C., Odada, E.O. (Eds.),
The Limnology, Climatology and Paleoclimatology of the East African Lakes. Gordon
and Breach Publishers, Amsterdam, pp. 103–139.
Ssemmanda, I., Vincens, A., 2002. Vegetation changes and their climatic implication for
the Lake Victoria region during the Late Holocene. In: Odada, E.O., Olago, D.O.
(Eds.), The East African Great Lakes: Limnology, Palaeolimnology and Biodiversity.
Kluwer Academic Publishers, The Netherlands, pp. 509–523.
Standley, L.J., Simoneit, B.R.T., 1987. Characterization of extractable plant wax, resin, and
thermally matured components in smoke particles from prescribed burns.
Environmental Science and Technology 21, 163–169.
Talbot, M.R., Filippi, M.L., Jensen, N.B., Tiercelin, J.-J., 2007. An abrupt change in the
African monsoon at the end of the Younger Dryas. Geochemistry, Geophysics,
Geosystems 8, Q03005. doi:10.1029/2006GC001465.
Tareq, S.M., Tanaka, N., Keiichi, O., 2004. Biomarker signature in tropical wetland: lignin
phenol vegetation index (LPVI) and its implications for reconstructing the
paleoenvironment. Science of the Total Environment 324, 91–103.
Tareq, S.M., Kitagawa, H., Tanoue, E., 2006. Isotope and lignin signature in Lake Erahi
(Southwest China): evidence for paleovegetation changes and Asian monsoon
variability during 18.5 ky BP. Geochimica et Cosmochimica Acta 70, Supplement 1, A638.
Van Zinderen Bakker, E.M., 1969. The Pleistocene vegetation and climate of the basin. In:
Clark, D.J. (Ed.), Kalambo Falls Prehistoric Site. I: Geology, Palaeoecology and
Detailed Stratigraphy of the Excavations. University Press, Cambridge, pp. 57–84.
112
I.S. Castañeda et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 275 (2009) 100–112
Vincens, A., 1991. Late Quaternary vegetation history of the South-Tanganyika Basin.
Climatic implications for South Central Africa. Palaeogeography, Palaeoclimatology,
Palaeoecology 86, 207–226.
Vincens, A., Chalie, F., Bonnefille, R., Guiot, J., Tiercelin, J.-J., 1993. Pollen-derived rainfall
and temperature estimates from Lake Tanganyika and their implication for Late
Pleistocene water levels. Quaternary Research 40, 343–350.
Vincens, A., Williamson, D., Thevenon, F., Taieb, M., Buchet, G., Decobert, M., Thouveny,
N., 2003. Pollen-based vegetation changes in southern Tanzania during the last
4200 years: climate change and/or human impact. Palaeogeography, Palaeoclimatology, Palaeoecology 198, 321–334.
Vincens, A., Buchet, G., Williamson, D., Taieb, M., 2005. A 23,000 yr pollen record from
Lake Rukwa (8°S, SW Tanzania): New data on vegetation dynamics and climate in
Central East Africa. Review of Palaeobotany and Palynology 137, 147–162.
Wakeham, S.G., Pease, T.K., 1992. Lipid Analysis in Marine Particles and Sediment
Samples. Skidaway Institute of Oceanography, Savannah, GA.
White, F., 1983. The vegetation of Africa, a descriptive memoir to accompany the
UNESCO/AETFAT/UNSO vegetation map of Africa. UNESCO, Natural Resour. Res. 20,
Paris. 356p.
Winter, K., Smith, J., 1996. Crassulacean acid metabolism: biochemistry, ecophysiology
and evolution. Ecological Studies, vol. 114. Springer, Berlin Heidelberg. 449p.
Zhang, Z., Zhao, M., Lu, H., Faiia, A.M., 2003. Lower temperature as the main cause of C4
plant declines during the glacial periods on the Chinese Loess Plateau. Earth and
Planetary Science Letters 214, 467–481.
Zhang, Z., Zhao, M., Eglinton, G., Lu, H., Huang, C.-Y., 2006. Leaf wax lipids as
paleovegetational and paleoenvironmental proxies for the Chinese Loess Plateau
over the last 170 kyr. Quaternary Science Reviews 25, 575–594.