Quaternary Research 80 (2013) 326–340
Contents lists available at ScienceDirect
Quaternary Research
journal homepage: www.elsevier.com/locate/yqres
Short Paper
Ancient charcoal as a natural archive for paleofire regime and vegetation
change in the Mayumbe, Democratic Republic of the Congo
Wannes Hubau a, b,⁎, Jan Van den Bulcke a, Peter Kitin b, Florias Mees c, Geert Baert d, Dirk Verschuren e,
Laurent Nsenga a, b, Joris Van Acker a, Hans Beeckman b
a
Ghent University, Department of Forest and Water Management, Laboratory for Wood Technology, Coupure Links 653, B-9000 Gent, Belgium
Royal Museum for Central Africa, Laboratory for Wood Biology, Leuvensesteenweg 13, B-3080 Tervuren, Belgium
Royal Museum for Central Africa, Department of Geology and Mineralogy, Leuvensesteenweg 13, B-3080 Tervuren, Belgium
d
University College Ghent, Department of Plant Production, Schoonmeersstraat 52, B-9000 Gent, Belgium
e
Ghent University, Department of Biology, Limnology Unit, K.L. Ledeganckstraat 35, B-9000 Gent, Belgium
b
c
a r t i c l e
i n f o
Article history:
Received 11 May 2012
Available online 19 June 2013
Keywords:
Pedoanthracology
Wood anatomy
Charcoal analysis
Central Africa
Mayumbe
Vegetation history
Paleoenvironment
a b s t r a c t
Charcoal was sampled in four soil profiles at the Mayumbe forest boundary (DRC). Five fire events were recorded
and 44 charcoal types were identified. One stratified profile yielded charcoal assemblages around 530 cal yr BP
and >43.5 cal ka BP in age. The oldest assemblage precedes the period of recorded anthropogenic burning, illustrating occasional long-term absence of fire but also natural wildfire occurrences within tropical rainforest. No
other charcoal assemblages older than 2500 cal yr BP were recorded, perhaps due to bioturbation and colluvial
reworking. The recorded paleofires were possibly associated with short-lived climate anomalies. Progressively
dry climatic conditions since ca. 4000 cal yr BP onward did not promote paleofire occurrence until increasing
seasonality affected vegetation at the end of the third millennium BP, as illustrated by a fire occurring in mature
rainforest that persisted until around 2050 cal yr BP. During a drought episode coinciding with the ‘Medieval
Climate Anomaly’, mature rainforest was locally replaced by woodland savanna. Charcoal remains from pioneer
forest indicate that fire hampered forest regeneration after climatic drought episodes. The presence of pottery
shards and oil-palm endocarps associated with two relatively recent paleofires suggests that the effects of
climate variability were amplified by human activities.
© 2013 University of Washington. Published by Elsevier Inc. All rights reserved.
Introduction
Combined reconstruction of past climate variability and ecosystem
dynamics increases our understanding of ecosystem response to current
and future climate change (e.g., Willis and Birks, 2006). Substantial work
in temperate and arid regions of the northern hemisphere contrasts with
the significant knowledge gaps that continue to exist for many tropical
regions, particularly Central Africa (e.g., Leal, 2001; Emery-Barbier and
Thiébault, 2005; Tchouto et al., 2009; Théry-Parisot et al., 2010; Maley
et al., 2012; Neumann et al., 2012b). A well-known concept in tropical
paleoecology and biogeography is that of the Central African forest
refuges, which are submountainous, fluviatile or coastal regions enjoying
rather moist (micro-) climate conditions thought to protect rainforests
during dry climate anomalies (Maley, 1996, 2004; Sosef, 1996; Maley
⁎ Corresponding author at: Ghent University, Department of Forest and Water
Management, Laboratory for Wood Technology, Coupure Links 653, B-9000 Gent, Belgium.
Tel.: +32 9 264 61 23; fax: +32 9 264 90 92.
E-mail addresses: wannes.hubau@ugent.be, w.hubau@leeds.ac.uk (W. Hubau),
jan.vandenbulcke@ugent.be (J. Van den Bulcke), kitin@wisc.edu (P. Kitin),
florias.mees@africamuseum.be (F. Mees), geert.baert@hogent.be (G. Baert),
Dirk.verschuren@ugent.be (D. Verschuren), lnsenga@yahoo.fr (L. Nsenga),
joris.vanacker@ugent.be (J. Van Acker), hans.beeckman@africamuseum.be (H. Beeckman).
and Brenac, 1998; Leal, 2001, 2004). Yet paleoecological reconstructions
suggest that rainforest and other Central African ecosystems show a
remarkably high sensitivity to natural climate changes such as prolonged
drought or increased seasonality (e.g., Maley, 1996, 2001, 2004;
Ngomanda et al., 2007, 2009a,b). Also, human activity cannot be
overlooked when discussing climate and vegetation changes during the
last three millennia, when agriculture and iron smelting were introduced
in Central Africa, even though before the last millennium human impact
on Central African forests was probably marginal (e.g., Maley, 1996, 2004;
Brncic et al., 2007; Maley et al., 2012; Neumann et al., 2012b).
The most important source of evidence for the climate and vegetation
history of Central Africa is lake-sediment records (e.g., Russell and
Johnson, 2005, 2007; Stager et al., 2009) and the fossil pollen they contain
(e.g., Maley, 1996, 2004; Maley and Brenac, 1998; Brncic et al., 2007;
Hessler et al., 2010). These studies provide high-resolution temporal
records from a select few locations with undisturbed lacustrine sedimentation, but do not offer great spatial detail. In contrast, macrocharcoal
fragments can be found in terrestrial soils of any type of vegetation,
reflecting local paleofire regimes. Also, the analysis of soil charcoal from
profiles in natural environments (pedoanthracology) is not influenced
by the effects of fuelwood selection strategies, as observed in charcoal
analysis from archeological sites (archeoanthracology) (Di Pasquale et
0033-5894/$ – see front matter © 2013 University of Washington. Published by Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.yqres.2013.04.006
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
al., 2008; Théry-Parisot et al., 2010). Pedoanthracology has proven to be a
useful tool for reconstructing Holocene paleofire regimes in natural
environments, especially in the forest-boundary of submountainous
regions (Carcaillet and Thinon, 1996; Carcaillet et al., 1997; Di Pasquale
et al., 2008). Additionally, soil charcoal analysis can reveal the past
presence of woody plant taxa that are only rarely detected in pollen
assemblages (Elenga et al., 2000; Lebamba et al., 2009). As a result,
pedoanthracology is highly complementary with palynology for reconstruction of past vegetation dynamics (Emery-Barbier and Thiébault,
2005; Théry-Parisot et al., 2010; Hubau et al., 2012).
Hitherto, only a handful of studies on soil charcoal in Central Africa
included the identification of charred wood remains, due to the high
species richness of the area and lack of a scientifically sound identification procedure (e.g., Schwartz et al., 1990; Hart et al., 1996).
Recently, Hubau et al. (2012) developed a transparent identification
protocol that can allow a taxonomically more precise identification
than generally obtained by pollen analysis. The main objective of the
present study is to contribute to knowledge of Central African paleofire
and vegetation history by applying this protocol to soil charcoal assemblages from an area of the Congolese rainforest that is expected to be
sensitive to climate change. We selected a non-archeological study
site (the Luki Reserve) at the southern end of the Mayumbe hills in
the Democratic Republic of the Congo, (DRC), which are thought to
have served as a submountainous forest refuge during late Pleistocene
and Holocene episodes of climate deterioration (e.g., Maley, 1996,
2004; Sosef, 1996). Dry or more seasonal climate conditions result in
forest regression and fragmentation, particularly at forest boundaries.
However, little is known about patterns of past forest fragmentation
in Central Africa at the local scale (e.g., Tchouto et al., 2009) and about
the role of fire, which even in tropical rainforests is one of the most
important causes of forest destruction (Cochrane et al., 1999; Cochrane,
2003). Therefore, specific research questions for the present study are:
(1) what was the temporal and spatial occurrences of fire in the southern
Mayumbe? (2) How consistent are anthracological reconstructions with
known paleoclimate history? (3) Were paleofires only caused by climatic
anomalies or also by humans?
Material and methods
Study area
The Lower Guinean rainforest is separated from the West African
rainforests by the Dahomey Gap in Togo and Benin and from the eastern
part of the rainforest by the swamps of the Ubangi and Congo Rivers
(Leal, 2004). The Mayumbe is a chain of forested hills stretching along
the Atlantic Ocean from Gabon down to the Luki reserve, which is located
in the Bas-Congo province of the DRC, between 0545.00′S and 0570.00′S
and between 1305.00′E and 13 30.00′E (Fig. 1). With the establishment
of the Luki reserve in 1937, its forests, soils and possible charcoal archives
have been safeguarded from intense anthropogenic disturbance. As part
of the ‘Man and Biosphere’ (MAB) program of UNESCO, a tripartite
conservation zonation was applied, including a fully protected ‘central
zone’, which contains an important relict of the semi-evergreen subequatorial Guinean rainforest that once covered the entire Mayumbe
hills (e.g., Donis, 1948; Lebrun and Gilbert, 1954). However, its position
at the southernmost edge of the Mayumbe forest (Fig. 1A) makes this
forest relict vulnerable to climate change, involving natural forest fragmentation. As such, the Luki reserve is highly suitable for paleobotanical
research.
Charcoal sampling and profile description
Four pedoanthracological profiles were excavated in the Luki reserve (Fig. 1). One profile was located in the peripheral UH48 stand
(Couralet, 2010; Hubau et al., 2012) and three in the central zone
(CZ1, CZ2, CZ3) of mature rainforest (Fig. 1B). All profiles were only
327
a few kilometers apart and located at elevations ranging from 180 to
460 m (Fig. 1D). Soil charcoal sampling was conducted as described
by Hubau et al. (2012). For each profile a relatively flat area was chosen,
avoiding steep slopes to minimize the effects of erosion or colluvial sedimentation (see Fig. 1D, including information on relief). Furthermore,
all sampling sites were well-drained and located outside former agricultural fields (see Carcaillet and Thinon, 1996). Figure 1C presents a distribution map of soil types in the Luki reserve based on the soil map of
Bas-Congo presented by Van Ranst et al. (2010).
Next, exploratory holes were drilled with an Edelmann auger,
down to 1 m. A pedoanthracological profile of 100 cm × 150 cm surface area was excavated on a spot where prospection yielded charcoal
and where the soil was relatively dry and penetrable. All profiles were
excavated down to a depth of 140 cm. Deeper charcoal layers were
detected and sampled by augering in the bottom of the profile pit. Charcoal fragments (largest dimension >2 mm) were carefully collected by
hand, and sorted per depth interval of 20 cm. Specific anthracomass
was calculated as described by Carcaillet and Thinon (1996). Thin sections were prepared from undisturbed soil samples, following polyester
impregnation using standard procedures (Murphy, 1986) and micromorphological features were described applying polarization microscopy,
using the concepts and terminology of Stoops (2003). These features reveal variations in texture and possible bioturbation.
Charcoal description, identification and radiocarbon dating
For each profile interval, up to 50 charcoal fragments were analyzed
using reflected light microscopy (RLM) following Hubau et al. (2012).
All charcoal fragments were grouped into charcoal types, of which
each type generally represents one species. Next, a large fragment of
each charcoal type was mounted on a stub for scanning electron microscopy (SEM). Using SEM images, charcoal types were described applying
the numbered anatomical features used for the on-line InsideWood
database (IAWA Committee, 1989; InsideWood, 2011; Wheeler, 2011;
Hubau et al., 2012). This produces two strings of numbered features.
The first string represents primary features that are easily visible, while
the second string represents secondary features that are variable or
unclear. Finally, all charcoal types were identified applying the Central
African identification protocol described by Hubau et al. (2012).
Soil features and distribution of charcoal types within the profile
revealed possible profile stratification. One charcoal fragment from
each stratigraphic interval was selected for AMS 14C measurement
at the Poznán Radiocarbon Laboratory (Poland) or Beta Analytic
(Florida, USA). In case of ambiguous stratification patterns, two or
three charcoal fragments from different intervals were selected. Calibration was performed with the OxCal v4.1.5 software (r:5) (Bronk
Ramsey, 2009) using the SHCal04 calibration curve (McCormac et al.,
2004).
Evaluation of identification reliability
The final result of the charcoal identification protocol is the association of each charcoal type with a small group of woody plant species,
ranked according to their resemblance with the charcoal type anatomy
(Hubau et al., 2012). Specifically, a 5-point ranking system was used,
whereby 5 points were attributed in the case of perfect agreement between charcoal anatomy and woody species anatomy. Finally, the charcoal type received a 9-character label composed of the three first letters
of respectively family, genus and species name of one of the best ranked
species (Hubau et al., 2012). All identifications were evaluated according
to two different reliability criteria. Criterion A concerns the phytosociological similarity of the retained species, with a good score if all highest
ranked species have similar habitat preferences and a bad score if the
highest ranked species occur in different vegetation types. Criterion B
is based on the number of highest ranked species and their anatomy
rank, with a high score for good resemblance between charcoal and
328
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Figure 1. A. Location of the Luki reserve on the currently fragmented southernmost Mayumbe forest boundary, Democratic Republic of Congo (map derived from Mayaux et al.,
1997). B. Vegetation map of the Luki reserve. All four studied soil profiles are located in relict stands of mature rainforest, although the direct surroundings are heavily fragmented
and regenerating forest patches are not far away. C. Soil types in the Luki reserve (after Van Ranst et al., 2010). Type Gb3m (profile CZ1) — soil derived from gneiss, with advanced
ABtC profile development (Acrisol; WRB, 2006), typically reddish, highly weathered, well-drained, with sandy clay texture, often limited below by gravel at 50–100 cm depth. Type
Qd3m (profiles CZ2, CZ3 and UH48) — soil derived from quartzite, with advanced ABtC development (Acrisol; WRB, 2006), typically yellowish, highly weathered, well drained, with
sandy clay-loam to sandy loam texture, often limited below by gravel at 50–100 cm depth. Type A (close to profile UH48) — deep alluvial soil, with ABwC profile development,
imperfectly to moderately drained (Gleysol; WRB, 2006), with sandy loam texture, generally not limited by gravel until at least 100 cm depth. D. Elevation map of the Luki reserve.
Profiles CZ1 and CZ2 are excavated on the relatively flat central hillcrest known as Ndiondio. Profiles CZ3 and UH48 are excavated on lower terrain, far enough from steep slopes to
avoid severe erosion or colluvial deposition.
reference wood anatomy. Using these two criteria, the charcoal types
were given a reliability rank ranging from 1 to 6.
Results
Profile description
Table 1 lists the uncalibrated and calibrated ages for all dated fragments. Figure 1C shows the distribution of soil types in the Luki reserve,
derived from the soil map of the Bas-Congo province (Van Ranst et al.,
2010), in which parent material composition is used at the highest taxonomic level. A description of the dominant soil type at each profile is
given in the caption of Figure 1. Soils of the Luki area are Humic Acrisols
in the WRB system and Kandiustalfs or Kandiustults in the USDA system
(Baert, 1995), characterized by Bt horizon development (Van Ranst et al.,
2010). Our field and soil thin section observations are largely compatible
with this characterization. Figures 2 and 3 show soil profile descriptions,
soil types, specific anthracomass and charcoal identification results.
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Table 1
Conventional radiocarbon ages and calibrated ages.
Profile
UH48
UH48
Depth
[cm]
30−40
80−90
UH48 120−130
CZI
40−60
Lab no
14C age
[14C BP]
Poz-33055
2055 ± 30
Poz-39110
Poz-39109
Poz-33051
2205 ± 30
2140 ± 35
1770 ± 30
CZI
120−140
Poz-33052
1790 ± 30
CZ2
60−80
Poz-33054
555 ± 30
CZ3
20−40
Beta-314122
580 ± 30
CZ3
60−80
Beta-314123
Calibrated age intervals
[cal BP]
2043 BP
1991 BP
1911 BP
2308 BP
2209 BP
2030 BP
2300 BP
2177 BP
2160 BP
2091 BP
2154 BP
2114 BP
2071 BP
1704 BP
1691 BP
1628 BP
1811 BP
1742 BP
1706 BP
1584 BP
554 BP
544 BP
626 BP
561 BP
554 BP
>43500
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
1872 BP
1922 BP
1900 BP
2308 BP
2036 BP
2004 BP
2250 BP
2174 BP
2098 BP
2062 BP
1946 BP
2078 BP
2000 BP
1544 BP
1655 BP
1566 BP
1753 BP
1532 BP
1594 BP
1569 BP
504 BP
515 BP
607 BP
510 BP
523 BP
Probability
[%]
95.4 %
60.1 %
8.1 %
28.9 %
62.4 %
4.1 %
22.5 %
1.0 %
33.2 %
11.6 %
95.4 %
22.3 %
45.9 %
95.4 %
24.7 %
43.5 %
6.8 %
88.6 %
60.8 %
7.4 %
95.4 %
68.2 %
8.0 %
87.4 %
68.2 %
95.4 %
68.2 %
95.4 %
68.2 %
95.4 %
68.2 %
95.4 %
68.2 %
95.4 %
68.2 %
95.4 %
68.2 %
95.4 %
68.2 %
−
Shaded portions indicate 99.4% probability intervals.
Anthracomass, profile stratification and bioturbation
The numbers presented in Figures 2 and 3 are the first attempt to
quantify anthracomass in Lower Guinean primary rainforest soils, as
previous authors (e.g., Dechamps et al., 1988; Schwartz et al., 1990;
van Gemerden et al., 2003) only present charcoal assemblage composition. Specific anthracomass of most 20-cm intervals is less than
25 ppm. Only profile UH48 contains intervals with a much higher
specific anthracomass at 30–40 and 40–50 cm depths of 121 and
184 ppm, respectively. These values are comparable to those in soil
profiles in the French Alps, where only two intervals yielded more
than 20 ppm, with a maximum of 124 ppm (Carcaillet and Thinon,
1996). Pedoanthracological profiles in the Andes yielded a maximum
of 300 ppm (Di Pasquale et al., 2008).
In profile CZ3, rock fragments are abundant below 40 cm depth
and a clear contrast exists between an upper, organic A/AB horizon
(0–40 cm) and a lower, argillic 2Bt horizon (40–80 cm). Moreover,
two distinct specific anthracomass peaks (22 and 26 ppm) are separated
by a charcoal-poor interval (40–60 cm; 3 ppm), reflecting stratification
(e.g., Carcaillet et al., 1997; Di Pasquale et al., 2008). The charcoal assemblage in the A/AB horizon is dated to between 626 and 510 cal yr BP
whereas that in the 2Bt horizon is older than 43.5 cal ka BP (Fig. 3,
Table 1). Charcoal types occurring in the A/AB horizon do not occur in
the 2Bt horizon, except for a few fragments that have probably been
transported due to bioturbation, as indicated by the occurrence of channels, passage features and zones with a pellet structure in thin sections
(Fig. 3). All other identifiable charcoal fragments in the 2Bt horizon
have a weathered and brittle appearance and all belong to only one charcoal type (CAE GUI SPP).
In contrast to profile CZ3, profiles CZ1, CZ2 and UH48 do not exhibit
a lithological discontinuity or significant clay illuviation. Still, in all three
profiles the distribution of charcoal fragments displays peaking specific
anthracomass values at the top of the profile, followed by a nearly
charcoal-free layer and a secondary anthracomass peak lower down
(Figs. 2, 3). This lower peak is weak in profiles CZ2 and UH48, but significant in profile CZ1at 120–140 cm depth. Yet, fragments from this lower
anthracomass peak belong to charcoal types that also occur in the surface
interval and their radiocarbon ages are similar (Fig. 2, Table 1). Hence,
profile CZ1 is not stratified and the local soil charcoal assemblage
329
probably originated from only one paleofire event, which occurred between 1704 and 1544 cal yr BP (Table 1). Likewise, profiles UH48 and
CZ2 are not stratified and their charcoal assemblages originate from
fire events that occurred between 2308 and 1872 cal yr BP at the
UH48 site and between 554 and 504 cal yr BP at the CZ2 site (Figs. 2, 3).
The charcoal fragments in profiles CZ1, CZ2 and UH48 are scattered
over a depth interval of 140 cm, although at each location they were
formed during a single event. This could be explained by strong bioturbation, as evidenced by the occurrence of channels, passage features and
zones with a pellet or granular microstructure (Figs. 2, 3). Ants and
termites are abundant in semi-deciduous rainforest and can severely
disturb stratified profiles and even archeological layers (e.g., Cahen and
Moeyersons, 1977; McBrearty, 1990; Théry-Parisot et al., 2010). In
profile CZ3, a horizon with a large number of rock fragments below
40 cm depth impeded bioturbation. This may have contributed to the
preservation of two separate intervals with distinct charcoal assemblages, including one that is ancient (>43.5 cal ka BP). Besides bioturbation, micromorphological features also provide evidence for some
colluvial deposition, even though sampling sites were chosen on relatively flat areas. Examples are the presence of chert fragments (CZ1,
CZ2) and ironstone fragments (CZ2, UH48), relatively low grain angularity (UH48) and good sorting (UH48) (for terminology and interpretation, see Stoops, 2003; Stoops et al., 2010).
Since charcoal assemblages can be distributed over a thickness of more
than 140 cm over a span of 600 yr (e.g., profile CZ2, Fig. 2), redistribution
of charcoal fragments by bioturbation and colluvial deposition must be a
rather fast process. Older charcoal assemblages could be present below
140 cm depth at the CZ1, CZ2 and UH48 sites, but if these would be
spread over a large depth interval then specific anthracomass
would be low, also diminishing the chance to find charcoal. Spreading of charcoal fragments over large soil volumes could also explain
why none of the additional samples (140–180 cm) taken below the
bottom of the profiles contain charcoal fragments (Figs. 2, 3).
Charcoal assemblage composition
Analysis of a total of 935 charcoal fragments yielded 44 charcoal
types (see Supplementary Tables 1 and 2 for the complete inventory
of all retained species names and their ecology, light requirements,
morphology and geographical distribution area). The distributions
of charcoal types within each soil profile is presented in Figures 2
and 3, including an attribution for each recovered charcoal type to
one particular type of vegetation (gray shades). Specifically, some charcoal types originate from typical primary rainforest taxa, others mainly
from regenerating forest taxa and others from prominent pioneer taxa.
Some taxa have a large ecological tolerance, occurring in mature evergreen rainforest, in the forest–savanna transition zone and even in the
seasonally variable climate conditions of woodland savanna (Figs. 2, 3).
Charcoal identifications for profile UH48 were presented earlier
(Hubau et al., 2012), but one previously unidentified monocotyledon
taxon has now been identified as a Dracaena species (Fig. 3, Supplementary Table 2). Charred endocarp remains of oil palm drupes
(Elaeis guineensis) were found in all four profiles, most abundantly
in profiles UH48 and CZ3 (upper intervals). Besides these endocarps,
only charcoal type PHY ANT SPP occurs in more than one profile.
Eight charcoal types remain unidentifiable, although they are clearly
derived from mature woody species and different from all identified
charcoal types. Furthermore, 11 unidentified charcoal types are derived
from juvenile wood, fruit, bark or unidentified charred tissue (Figs. 2, 3).
These fragments can be derived from some of the same woody species
as the identified charcoal fragments.
Identification reliability
Table 2 presents the highest anatomy rank, the number of highest
ranked species and the name of one of the highest ranked species for
330
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Figure 2. Radiocarbon dates, soil profile characteristics, specific anthracomass and charcoal types recovered from sites CZ1 and CZ2. For every charcoal type, a representative species name and identification reliability evaluation is presented
in Table 2.
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Figure 3. Radiocarbon dates, soil profile characteristics, specific anthracomass and charcoal types recovered from sites CZ3 and UH48. For every charcoal type, a representative species name and identification reliability evaluation is presented in
Table 2.
331
332
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Family
Mature rainforest species
Regenerating forest species
Poineer
Woodland savanna species
Large ecological tolerance
A
B
APO ALS SPP
APO ANC PYR
APO FUN AFR
CAE ANT SPP
CAE SCO ZEN
EUP TET DID
LIN HUG PLA
PHY ANT SPP
PHY ANT SPP
AGA DRA SPP
CAE AFZ SPP
CAE GUI SPP
MYR PYC ANG
CZ1
CZ2
CZ3
CZ3
CZ3
CZ1
CZ2
CZ1
CZ2
UH48
CZ3
CZ3
CZ1
Apocynaceae
Apocynaceae
Apocynaceae
Caesalpinioideae
Caesalpinioideae
Euphorbiaceae
Linaceae
Phyllantaceae
Phyllantaceae
Agavaceae
Caesalpinioideae
Caesalpinioideae
Myristicaceae
cfr. Alstonia boonei De Wild.
cfr. Ancylobothrys pyriformis Pierre
cfr. Funtumia africana (Benth.) Stapf
cfr. Anthonotha pynaertii (De Wild.) Exell & Hillc.
cfr. Scorodophloeus zenkeri Harms
cfr. Tetrorchidium didymostemon (8aill.) Pax & K. Hoffm.
cfr. Hugonia platysepala Welw. ex Oliv.
cfr. Antidesma rufescens Tul.
cfr. Antidesma rufescens Tul.
cfr. Dracaena arborea (Willd.) Link
cfr. Afzelia bella Harms
cfr. Guibourtia demeusei (Harms) J. Leonard
cfr. Pycnanthus angolensis (Welw.) Exell
5
5
5
5
5
5
5
5
5
5
5
5
5
2
1
2
2
1
1
1
2
2
5
6
5
3
a
a
a
a
p
a
a
p
p
p
a
a
a
a
p
a
p
a
a
p
a
a
a
p
p
a
p
a
p
a
a
p
a
a
a
a
a
a
p
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
+
+
+
+
+
+
+
+
+
+
+
+
+
+++
+++
+++
+++
+++
+++
+++
+++
+++
++
++
++
++
1
1
1
1
1
1
1
1
1
1
1
1
1
ANN NEO GAB
CAE CYN MAN
COM TER SUP
MEL GUA SPP
MIM NEW SPP
RUB HEI CRI
ANN ANN LEB
CAE APH SPP
CAE GIL MAY
CAE TET BIF
MYR SYZ GUI
PHY BRI FER
RUB COF CAN
RUB COR SPP
CZ1
UH48
CZ1
UH48
CZ3
CZ1
CZ1
CZ3
UH48
UH48
UH48
CZ2
CZ2
UH48
Annonaceae
Caesalpinioideae
Combretaceae
Meliaceae
Mimosoideae
Rubiaceae
Annonaceae
Caesalpinioideae
Caesalpinioideae
Caesalpinioideae
Myrtaceae
Phyllantaceae
Rubiaceae
Rubiaceae
cfr. Neostenanthera gabonensis (Engl. & Diels) Exell
cfr. Cynometra mannii Oliv.
cfr. Terminalia superba Engl. & Diels
cfr. Guarea cedrata (A. Chev.) Pellegr.
cfr. Newtonia glandulifera (Pellegr.) Gilbert & Boutique
cfr. Heinsia crinita (Afzel.) G. Taylor
cfr. Annickia lebrunii (Robyns & Ghesq.) Sellen & Maas
cfr. Aphanocalyx microphyllus (Harms) Wieringa
cfr. Gilbertiodendron mayombense (Pellegr.) J. Leonard
cfr. Tetraberlinia bifoliolata (Harms) Hauman
cfr. Syzygium guineense (Willd.) DC.
cfr. Bridelia ferruginea Benth.
cfr. Coffea canephora Pierre ex A. Froehner
cfr. Corynanthe paniculata Welw.
4
4
4
4
4
4
4
4
4
4
4
4
4
4
1
1
1
2
2
2
5
4
4
3
4
3
6
7
p
p
a
p
a
a
p
p
p
p
a
a
a
p
a
a
a
a
p
p
a
a
a
a
a
a
a
a
a
a
p
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
a
p
a
a
a
a
a
a
a
a
a
a
a
a
p
a
p
a
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
−
−
−
−
−
−
−
−
2
2
2
2
2
2
2
2
2
2
2
2
2
2
APO TAB IBO
PHY BRI MIC
UH48
CZ1
Apocynaceae
Phyllantaceae
cfr. Tabemanthe iboga Baill.
cfr. Bridelia micrantha (Hochst.) Baill.
3
3
3
2
p
a
a
a
a
p
a
a
a
a
+
+
−−
−−
3
3
ANN XYL HYP
CAE CYN SES
FLA ONC SPP
MIM ALB SPP
CZ3
CZ3
CZ3
CZ1
Annonaceae
Caesalpinioideae
Flacourtiaceae
Mimosoideae
cfr. Xylopia hypolampra Mildbr.
cfr. Cynometra sessiliflora Harms
cfr. Oncoba mannii Oliv.
cfr. Albizia ferruginea (Guill. & Perr.) Benth.
5
5
5
5
4
4
4
5
p
p
p
p
p
p
p
p
a
a
a
a
a
a
a
a
a
a
a
a
−
−
−
−
++
++
++
++
4
4
4
4
HUA HUA GAB
IRV IRV ROB
ANN XYL AUR
CAE BER SPP
COM COM SPP
MYR COE SPP
RUB AID MIC
RUB NAU SPP
STE PTE SPP
UH48
UH48
UH48
CZ1
CZ2
UH48
CZ1
CZ2
CZ1
Huaceae
Irvingiaceae
Annonaceae
Caesalpinioideae
Combretaceae
Myristicaceae
Rubiaceae
Rubiaceae
Sterculiaceae
cfr. Hua gabonii Pierre ex De Wild.
cfr. Irvingia rabur Mildbr.
cfr. Xylopia aurantiiodora De Wild. & T. Durand
cfr. Berlinia bracteosa Benth.
cfr. Combretum mortehanii De Wild. & Exell
cfr. Coelocaryon botryoides Vermoesen
cfr. Aidia micrantha (K. Schum.) Bullock ex F. White
cfr. Nauclea diderrichii (De Wild.) Merr.
cfr. Pterygota macrocarpa K. Schum.
4
4
4
4
4
4
4
4
4
2
2
5
5
11
4
7
5
10
p
p
p
a
p
p
p
p
p
p
p
a
p
a
p
a
a
p
a
a
a
a
a
a
a
a
a
a
a
a
p
p
a
p
p
a
a
a
p
a
a
a
a
a
a
−
−
−
−
−
−
−
−
−
+
+
−
−
−
−
−
−
−
5
5
5
5
5
5
5
5
5
DlC DIC MAD
ULM CEL SPP
UH48
UH48
Dichapetalaceae
Ulmaceae
cfr. Dichapetalum madagascariense Poir.
cfr. Celtis mildbraedii Engl.
3
3
2
6
p
p
a
a
a
a
a
a
p
p
−
−
−−
−−
6
6
CZ2
CZ3
ALL
9
4
5
2
7
3
29
15
57
43
69
31
71
29
70
30
67
33
0
1
4
6
3
3
2
3
4
1
3
0
0
4
0
3
5
1
1
0
9
8
8
15
4
0
7
29
43
21
23
15
23
31
8
43
0
0
57
0
30
50
10
10
0
24
18
15
35
7
almost perfect resemblance (criterion B)
very good resemblance
only moderate resemblance
almost perfect resemblance
very good resemblance
only moderate resemblance
1
4
0
0
4
1
6
1
0
4
2
4
4
1
1
3
0
3
2
0
0
2
0
5
2
0
3
0
0
13
14
2
4
9
2
7
43
7
0
29
14
31
31
8
8
23
0
43
29
0
0
29
0
50
20
0
30
0
0
33
31
4
9
20
4
Total:
14
13
7
10
44
CZ2
CZ3
&
&
&
&
&
&
CZ1
phytosociological similarity (crit. A)
phytosociological similarity
phytosociological similarity
phytosociological ambiguity
phytosociological ambiguity
phytosociological ambiguity
8
6
# species
phytosociological similarity of highest ranked species
similar habitat preferences
+
−
different habitat preferences
woody species resemblance
+++
almost perfect resemblance with only one or two species (highest anatomy rank = 5/5)
++
almost perfect resemblance with more than two species (highest anatomy rank = 5-5)
+
good resemblance with only one or two species (highest anatomy rank = 4/5)
−
good resemblance with more than two species (highest anatomy rank = 4/5)
−−
only moderate resemblance with one or more species (highest anatomy rank < 3/5)
Reliability rank:
1
2
3
4
5
6
UH48
B
ALL
A
CZ1
Reliability evaluation criteria:
UH48
Representative species
Reliability rank
Profile
# Highest ranked species
Woody species resemblance
Charcoal type
Highest anatomy rank (/5)
Phytosociological similarity
Table 2
Representative species name and evaluation of identification reliability for each identified charcoal type presented in Figs. 2 and 3. The highest anatomy rank, the number of highest
ranked species, and ecology of the highest ranked species are indicated.
% species
100 100 100 100 100
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
333
Figure 4. Left: Scanning Electron Micrographs (SEM) of charcoal type MYR PYC ANG; Right: Transmitted Light Micrographs (TLM) of a reference wood sample of Pycnanthus
angolensis (Welw.) Exell (Tw 29820) (Myristicaceae). A–B: Transversal direction; C–D: Tangential direction; E–F: Radial direction.
each of the 44 identified charcoal types, along with the vegetation
types in which the highest ranked species occur, and an identification
reliability score based on our two evaluation criteria (see Material
and methods section). In cases where only one species is ranked
highest, our identification excludes any possible confusion regarding
the ecology of the charcoal type (e.g., APO ANC PYR, EUP TET DID, LIN
HUG PLA). However, for most charcoal types, several species receive
the highest rank, with a maximum of 11 species (COM COM SPP).
Sometimes, the highest ranked species had different phytosociological
characteristics, complicating paleoecological interpretation of the charcoal type. In such cases, the charcoal type received a negative (−) score
for evaluation criterion A. Evaluation criterion B specifies how well the
charcoal type resembles reference material of the highest ranked species. Based on both criteria together, each charcoal type received an
identification reliability rank.
Overall, 64% (27 types) of our identifications are highly reliable
(ranks 1 or 2) because their highest ranked species do not have ambiguous phytosociological characteristics and resemble the charcoal
type's anatomy almost perfectly or at least very well (Table 2). Identifications are especially reliable in profiles CZ2 and CZ3 (71% and 70%
respectively), followed by profile CZ1 (62%) and finally profile UH48
(50%). As an illustration of a reliable identification, Figure 4 presents
SEM images of charcoal type MYR PYC ANG from profile CZ1
together with transmitted light images of a reference wood sample of
Pycnanthus angolensis, a prominent pioneer species. Only two charcoal
types have highest ranked species that resemble the charcoal anatomy
only moderately, although they are phytosociologically unambiguous
(rank 3, APO TAB IBO and PHY BRI MIC). Finally, 33% (15 types) of all
identifications are not reliable (ranks 4, 5 or 6) because their highest
ranked species occur in different forest types, complicating proper ecological interpretation. As an example, some of the 11 highest ranked
species of charcoal type COM COM SPP (Combretum spp.) occur in mature rainforest, whereas others are typical woodland savanna species
(Table 2 and Supplementary Table 1).
Discussion
Temporal aspects of the paleofire regime in the Luki reserve
Natural fires in moist evergreen rainforest environments are generally rather rare because high air humidity reduces combustibility
(Scott, 2000). The long period without recorded charcoal formation
at the CZ3 site, spanning more than 43 ka (Fig. 3), appears to suggests
that local paleofires did not occur on this site during the consecutive
climate variations of the late Pleistocene and Holocene (e.g., Maley
334
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Figure 5. Geographical distribution of published paleoenvironmental records throughout equatorial Africa in relation to the Luki reserve (site 1), with indication of the type of
source data and main references. The current distribution of rainforest is based on Mayaux et al. (1997) and forest area during the third millennium BP is adapted from Maley
(2004).
and Brenac, 1998), implying possible long-term absence of fire in the
Central African rainforest, even at forest boundaries (Fig. 1).
On the other hand, it is possible that the CZ3 profile does not represent
a continuous paleoenvironmental archive. Specifically, temporal gaps in
the macrocharcoal record may be due to a scarcity of woody plants combined with limited soil accumulation and profile development during dry
episodes of the last glacial cycle (e.g., Dupont et al., 2000), when fires
were probably more common in Central Africa (e.g., Bird and Cali,
1998). The absence of charcoal can also be a result of unrecognized severe
soil erosion truncating the profile. Furthermore, presence of charcoal fragments from a Guibourtia demeusei stand older than 43.5 cal ka BP implies
that naturally induced fires occurred during a period when the locality
was covered with rainforest. Based on current understanding of Congo
Basin vegetation history at this time scale, this charcoal may date from
the (later phases of) MIS 5, i.e. older than ca. 75 ka, unless episodes of
rapid climate change during Dansgaard–Oeschger cycles and Heinrich
stadials in MIS 3–4 allowed short-lived development of local rainforest,
perhaps disturbed by fire (cf. Daniau et al., 2010). The charcoal assemblages of all other recorded paleofires in our Luki profiles are a testimony
of fire within the rainforest (Figs. 2, 3, Table 2). Most probably this occurred in areas of fragmented forests where fires in dry and open savanna
patches can burn forest edges and small forest patches, thus creating
more fire-prone open spaces (Cochrane et al., 1999; Cochrane, 2003).
In paleoecological studies of the African rainforest, periods of inferred
forest fragmentation are often attributed to the occurrence of arid
climate anomalies (Vincens et al., 1998; Maley, 2001, 2002; Cochrane,
2003). On late-Quaternary time scales, Central African precipitation regimes are mainly controlled by the influence of northern high-latitude
glaciation on the Atlantic Meridional Overturning Circulation (AMOC),
reflected in differences in sea surface temperature (= ΔSST) between
the tropics and the subtropics (Schefuß et al., 2005). High ΔSST is associated with stronger Southern Hemisphere trade winds hampering the
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Figure 6. Summary of the main findings resulting from paleorecords reconstructed for the Mayumbe, the Lower Guinea, the Dahomey Gap, the Congo Basin and East Africa. Black fillings represent major disturbance periods. Dark gray indicates severe aridity, light gray indicates moderate aridity and very light gray indicates wet periods. Study site locations are presented in Fig. 5.
335
336
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
flow of moist air from the Atlantic Ocean into Central Africa, thus provoking relative aridity on the continent. Examples are MIS3 and MIS4, the
last glacial maximum (LGM or MIS2), the Younger Dryas and the 8.2 ka
event (Maley, 1996; Maley and Brenac, 1998; Alley and Ágústsdóttir,
2005; Schefuß et al., 2005). The steep rise in ΔSST starting around
3000 cal yr BP, a trend that continues to the present (Schefuß et al.,
2005), is consistent with the increased abundance of fire events
recorded within the last 3000 years in the Ituri forest (Hart et al.,
1996) and the Luki reserve (Table 1, Figs. 2, 3). Specifically, the
paleofire dated to between 2300 and 1900 cal yr BP at UH48 is
broadly coeval with a reconstructed peak in ΔSST around 2300 cal yr
BP. Also, the fire events at CZ2 and CZ3 dated to 600–500 years ago
broadly coincide with a period of increasing ΔSST (Schefuß et al., 2005).
Lack of recorded paleofires that presumably must have occurred during
other inferred arid episodes in the recent past, such as around 1200–
1000 cal yr BP (Schefuß et al., 2005), may be explained by a temporary
setback of woody vegetation interrupting soil accumulation, and thus
an absence of traces archived in the pedoanthracological record. Lack of
recorded fire events in Luki or Ituri dating from older periods of known
climatic drought (e.g., the Younger Dryas) can be explained both by this
process and by rapid burial and scattering of charcoal assemblages by bioturbation as outlined above.
Species composition of the charcoal assemblages
Charcoal type richness as a reflection of forest species richness
The minimum species richness of the charcoal assemblages recovered
from the Luki reserve, expressed as the sum of the identified types and of
the unidentified types originating from mature wood, is on average 13
species per 1.5 m2 (12 in CZ2, 10 in CZ3, 15 in CZ1, 16 in UH48). This is
more than 10% of the 81 to 127 tree species per ha reported to occur in
Central African natural rainforest stands (van Gemerden et al., 2003;
Worbes et al., 2003). This suggests that the charcoal assemblages originated from species-rich forest types, rather than from species-poor
savanna types.
However, these forest inventories must be considered a minimum
estimate of true woody plant diversity, since they are usually limited
to trees with breast-height diameters exceeding 10 cm (Worbes et al.,
2003). Tree species dispersing seeds via wind, water or animals can produce off-site progeny. Also, the local soil seed bank can produce seedlings and young trees of species that are temporarily lacking from the
site (cf. Daïnou et al., 2011).
On the other hand, species richness recorded in pedoanthracological
profiles may also underestimate true charcoal diversity because some
charcoal types can escape burial due to post-depositional processes
such as horizontal transport by wind or water or by physical weathering
severely fragmenting the charcoal particles (e.g., Théry-Parisot et al.,
2010). Also, trees are not always completely charcoalified and they do
not always fall down after the fire (Scott, 2000). Finally, given the relatively small sample size, uncommon tree species with a small share in
the basal area are probably underrepresented in charcoal assemblages
compared to dominant species. For example, only two charcoal types
were clearly derived from liana species (APO ANC PYR, LIN HUG PLA;
see Fig. 2) although tropical forests typically have a great diversity of
lianas (Schnitzer and Bongers, 2002).
Diversity of burned forest types
Nearly all charcoal fragments in the oldest charcoal assemblage
(>43.5 cal ka BP; profile CZ3) were derived from only one species, and
probably from the same individual (Fig. 3). Guibourtia species are large
(>20 m) trees typical of old-growth evergreen and semi-deciduous
primary rainforest (Burkill, 1985; Leal, 2004; African Plants Database,
2011).
The fire event dated to between 2308 and 1872 cal yr BP in the UH48
stand undoubtedly burned a patch of mature rainforest. Thirteen out of
15 identified charcoal types yielded prominent indicators of mature
rainforest among the highest ranked species (Fig. 3). Examples are
Gilbertiodendron species (CAE GIL MAY), Guarea species (MEL GUA SPP)
and Tetraberlinia bifoliolata (CAE TET BIF) (Lebrun and Gilbert, 1954;
Leal, 2004) (Fig. 3, Table 2). Coelocaryon species (MYR COE SPP) are
also important indicators of old primary rainforest in the Luki reserve
(Donis, 1948). Endocarps of the pioneer palm E. guineensis found in
the UH48 assemblage were probably introduced by humans, given the
presence of pottery fragments (Fig. 3).
The charcoal assemblage dated to 1704–1544 cal yr BP in profile
CZ1 is dominated by charcoal types yielding prominent pioneers such
APO ALS SPP (cf. Alstonia spp.), MYR PYC ANG (cf. P. angolensis) and
COM TER SUP (cf. Terminalia superba) (Fig. 2, Table 2). These often occupy small patches within the rainforest but they sometimes also occur in
large, nearly monodominant stands within old evergreen rainforest, indicating former disturbance at the landscape scale (Donis, 1948; Lebrun
and Gilbert, 1954; Protabase, 2012). They are widely distributed and
sometimes very abundant. Also, Heinsia crinita (RUB HEI CRI), Pterygota
species (STE PTE SPP), Tetrorchidium didymostemon (EUP TET DID) and
Albizia species (MIM ALB SPP) are significant indicators for regenerating
forest (Table 2).
The CZ2 assemblage (554–504 cal yr BP) clearly indicates an open
vegetation type. Bridelia ferruginea (PHY BRI FER) is common in the
grasslands neighboring the Luki reserve, although it also occurs near
forest edges (Donis, 1948; Lebrun and Gilbert, 1954; Vincens et al.,
1998; African Plants Database, 2011). Second, in contrast to all
other charcoal assemblages it includes two liana taxa, APO ANC PYR
(cf. Ancylobotrys pyriformis) and LIN HUG PLA (cf. Hugonia platysepala).
Lianas are favored by disturbance and they are relatively more abundant in forest gaps than under closed canopy (e.g., Schnitzer and
Bongers, 2002). Finally, some of the highest ranked species for charcoal
types COM COM SPP and RUB NAU SPP occur in woodland savanna,
whereas others occur in mature rainforest (Fig. 2, Table 2). Although
identification results clearly indicate a more open vegetation type, the
relatively high species richness of the CZ2 assemblage (12 species per
1.5 m2) indicates that it was probably dry deciduous forest or dry
woodland rather than a savanna.
The upper charcoal assemblage of profile CZ3, dated to 626–510 cal yr
BP, is dominated by prominent pioneer and secondary forest taxa such
as APO FUN AFR (cf. Funtumia africana), CAE ANT SPP (cf. Anthonotha
spp.) and CAE AFZ SPP (cf. Afzelia spp.) (Fig. 3, Table 2). However, the
assemblage also includes two types belonging to prominent mature
rainforest taxa (CAE CYN SES and CAE APH SPP) (Lebrun and Gilbert,
1954; Leal, 2004; Protabase, 2012). Although fast-growing pioneer
species recruit easily in forest gaps and are initially more abundant
and more successful than more slowly regenerating mature forest species (Protabase, 2012), it is possible that the latter resprouted from a
cut- or broken-off stub after natural or human-induced disturbance
(Mwavu and Witkowski, 2008), or recruited occasionally from the seedling bank or from the local soil seed bank (Daïnou et al., 2011).
Consistency with regional paleoclimate history
Figure 5 shows the distribution of existing paleobotanical and
paleoclimatological records, with which the results of this study can be
compared. The main patterns of regional climate and vegetation
dynamics spanning the last 4000 years are summarized in Figure 6,
grouped by geographical area and arranged according to distance from
the study area. The records closest to the Luki reserve are those of
Lake Kitina in the southern Mayumbe forest (Elenga et al., 1996,
2004; Maley, 2002, 2004), Lake Sinnda in the grasslands of the
Mayumbe rainshadow area (Vincens et al., 1998), and from a marine
site near the mouth of the Congo River (Schefuß et al., 2005). However,
most information is derived from sites within the Lower Guinea farther
north (Fig. 5). We also considered high-resolution paleolimnological records from Eastern Africa that give a detailed view of late Holocene climate variability.
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
i. Third millennium BP rainforest crisis
Two successive phases of late Holocene rainforest breakdown are
documented extensively in the literature, as outlined in Figure 6.
Since the Luki reserve is located on the southernmost edge of the Central
African rainforest (Figs. 1, 5), one might expect that these events had impacted its forest considerably. The first phase occurred between 4000 and
2500 cal yr BP, when lowering SST in the Gulf of Guinea (Weldeab et al.,
2007) is associated with a more arid climate in the Lower Guinea forest
region. Although nearly all selected sites in Central and East Africa record
increasing drought during this period, it impacted only on the boundaries
of the Central African rainforest complex (Fig. 6; Vincens et al., 1998;
Elenga et al., 2004; Salzmann and Hoelzmann, 2005; Ngomanda et al.,
2009a,b; Maley et al., 2012). However, a complete absence of charcoal
from between 4000 and 2500 cal yr BP in the Luki assemblages suggests
that increasing aridity did not cause severe rainforest breakdown through
fire at the study site (Figs. 2, 3). Indeed, the closed-canopy rainforest
is remarkably resistant to drought because the canopy is able to
trap transpired moisture, thus maintaining humidity levels and hampering combustion (Cochrane, 2003). Also, in the Ituri forest of eastern
Congo (Fig. 5), only two of 28 paleofires date from this period (Hart
et al., 1996).
The second phase of forest breakdown occurred between 2500 and
2000 cal yr BP, when rising sea-surface temperatures (Weldeab et al.,
2007) resulted in a generally warmer and wetter climate, with pronounced seasonality. The wet season was characterized by the formation
of large cumuliform clouds and the occurrence of torrential rains (Maley,
2002, 2004). Increasing ΔSST between the tropical and subtropical ocean
(Schefuß et al., 2005) resulted in a strengthening of the trade winds,
causing a long and pronounced dry season (Maley and Brenac, 1998;
Elenga et al., 2004; Maley, 2004; Ngomanda et al., 2009a,b; Neumann
et al., 2012a,b). Palynological studies show that this severe climate
shift almost completely eradicated mature rainforest in certain regions
of the Lower Guinea (Maley, 2002). Yet, the UH48 charcoal assemblage
(Fig. 2) clearly indicates that despite the occurrence of fire, mature
semi-deciduous rainforest persisted until the end of this severe breakdown phase, even at the forest boundary (Figs. 1, 5).
Given strong climate seasonality, windfalls during storms may have
contributed to open up the rainforest canopy, creating additional gaps
vulnerable to drought and fire (e.g., Maley, 2002; Ngomanda et al.,
2009b). True rainforest trees generally have thinner protective bark
layers and are thus easily destroyed by fires occurring in nearby forest
gaps (Cochrane et al., 1999; Cochrane, 2003; Broadbent et al., 2008).
As such, the combined effect of gap formation during wet seasons and
severe drought with recurring fire during dry seasons may eventually
have destroyed large blocks of contiguous forest. The occurrence of a
paleofire within a mature rainforest stand at the UH 48 site indicates
that forest destruction was still ongoing by the end of the third millennium BP. At that time the site was probably located at a forest edge or in
a small mature rainforest patch within a matrix of fire-prone open
vegetation. Increasing paleofire occurrence from 2200 cal yr BP
onwards has also been reported for the Ituri forest (Fig. 6; Hart et al.,
1996).
ii. Forest recovery after the third millennium BP rainforest crisis
After 2000 cal yr BP, the Lower Guinean climate returned to wet and
relatively stable conditions (e.g., Maley, 2002, 2004; Ngomanda et al.,
2007, 2009a,b). Decreasing Gramineae and increasing abundance of
pioneer trees in pollen records indicate successful forest regeneration
(Elenga et al., 1996; Maley, 1996, 2002, 2004; Reynaud-Farrera et al.,
1996; Maley and Brenac, 1998; Elenga et al., 2004; Ngomanda et al.,
2009a; Neumann et al., 2012b; see also Fig. 6). Likewise, the CZ1 charcoal assemblage dated to this period is dominated by pioneer species
(Fig. 2, Table 2). This fire event occurred at least three centuries after
the rainforest crisis, indicating that forest recovery was most likely a
slow process. Recurring fire in scattered patches of savanna still affected
regenerating forest patches, killing the most vulnerable tree species
337
(e.g., Broadbent et al., 2008). In the southern portion of the Lower Guinea,
the combustibility of open vegetation probably remained high due to
continuing drought until at least 1300 cal yr BP, as reported for the surroundings of Lake Sinnda (Fig. 6; Vincens et al., 1998). Indeed, steep
ΔSST rises alternating with only weak ΔSST drops during the last
3000 years suggest that aridification continued (Schefuß et al., 2005).
Yet, the increasing abundance of pollen from pioneer species on several
sites indicate that forest re-establishment was eventually successful
(e.g., Fig. 2; Maley, 2002).
iii. Medieval climate anomaly and Little Ice Age
Although manifestation of the Northern Hemisphere ‘Medieval Climate Anomaly’ (MCA, 1100–700 cal yr BP) and ‘Little Ice Age’ (LIA,
650–100 cal yr BP) in various tropical regions is under debate, many
authors use this terminology as reference chronozones for tropical
climate anomalies during these broad time periods (e.g., Vincens et al.,
1998; Verschuren et al., 2000; Ngomanda et al., 2007; Russell and
Johnson, 2007; Verschuren and Charman, 2008; Stager et al., 2009).
One pattern that appears to be widespread across intertropical Africa
is a switch towards wetter climatic conditions coincident with the
MCA–LIA transition 700–650 cal yr BP (Verschuren and Charman,
2008). In eastern equatorial Africa the MCA-equivalent period was
mostly dry (Verschuren et al., 2000; Russell et al., 2003; Stager et al.,
2009), whereas close to the Atlantic coast in the Lower Guinea, the
MCA appears to be characterized by fluctuating wet–dry conditions
(Vincens et al., 1998; Schefuß et al., 2005; Ngomanda et al., 2007).
Allowing for some dating mismatch between the available records, the
start of the LIA-equivalent period is most often marked by the return of
a more consistently wet climate in tropical Africa. However, after ca.
500–450 cal yr BP these wet conditions reversed to more or less pronounced aridity throughout western and central equatorial Africa, from
near the Atlantic coast (Ngomanda et al., 2007) to the western shoulder
of the East African plateau (Russell and Johnson, 2007). Persisting until
ca. 200–150 cal yr BP, this prolonged drought coincides with coldest
LIA temperatures in Western Europe (Verschuren and Charman, 2008).
Only easternmost equatorial Africa, i.e., the portion of the continent situated beyond Atlantic Ocean influence, enjoyed relatively wet conditions
throughout the LIA-equivalent period, and even there they were occasionally interrupted by decade-long dry spells (Verschuren et al., 2000,
2009; Tierney et al., 2013).
Earlier studies detailing vegetation dynamics during the broad
MCA–LIA time interval in the Lower Guinea documented an increase
in shade-intolerant trees under fluctuating wet–dry conditions during
the MCA (Ngomanda et al., 2007), and forest disturbance with increasing
oil palm abundance between 1400 and 800 cal yr BP (Reynaud-Farrera
et al., 1996; Maley, 2002). In the Dahomey Gap, dry conditions and forest
retreat started from 1100 cal yr BP and continued until the present
(Salzmann and Hoelzmann, 2005). In the Ituri forest of eastern Congo,
the entire period from 1300 to 800 cal yr BP is marked by an unusually
high number of fire events (Hart et al., 1996). Likewise, the replacement
of mature rainforest by woodland savanna at the CZ2 site (Fig. 2) shows
that the MCA was undoubtedly a period of significant forest fragmentation in the Luki reserve. Due to recurrent fire, this open vegetation
persisted into the early LIA-equivalent period. This diverse evidence for
MCA drought on the continent is associated with a pronounced rise
in tropical Atlantic ΔSST between 1200 and 1000 cal yr BP (Fig. 6;
Schefuß et al., 2005).
Wetter conditions on land starting at the MCA–LIA transition promoted forest regeneration, as illustrated by the occurrence of pioneer
trees around 600 cal yr BP in the Luki reserve (CZ3, Fig. 3) and in the
pollen record of Lake Sinnda (Figs. 5, 6, Vincens et al., 1998). However,
the fragmented Ituri forest structure remained vulnerable to fire, as
evidenced by two fire events during this period (Hart et al., 1996). The
CZ2 and CZ3 fire events both occurred between 630 and 500 cal yr BP,
likewise during the wet early phase of the LIA (Fig. 6).
338
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Wildfires or shifting cultivation?
Charcoal fragments found in soils of natural environments are
sometimes a priori interpreted as witnesses of past slash-and-burn
activity (van Gemerden et al., 2003). However, fire is also the major
cause of natural vegetation disturbance, even in moist rainforests.
Lightning is the most important initiator of wildfires and more successfully so in canopy gaps during episodes of climatic drought (Hart et al.,
1996; Cochrane et al., 1999; Scott, 2000; Cochrane, 2003). The ancient
charcoal in the deepest layer of profile CZ3 most likely formed during
such a natural forest wildfire. While there are indications for the presence of Stone-Age human communities in the Lower Guinea before
43.5 cal ka BP (Oslisly, 2001), there is no evidence for wildfires having
been ignited by humans in the Lower Guinea before the introduction
of agriculture with the arrival of Bantu-speaking people around
2500 cal yr BP. Even then, evidence for actual farming activities remains
scarce until about 1000 cal yr BP (Fig. 6; Neumann et al., 2012b). The
early migrants were probably hunter-gatherers in the first place, and
the crops they introduced from the subtropics (e.g., Pennisetum glaucum)
needed a distinct dry season, which became more problematic after the
third millennium BP rainforest crisis as climate returned to less seasonal
conditions (Fig. 6; Neumann et al., 2012a,b). Moreover, humans confined
their activities to natural forest gaps and preferred cutting softer pioneer
trees rather than harder trees in the mature rainforest (Ngomanda et al.,
2009b; Neumann et al., 2012a).
The charcoal assemblage of the UH48 profile dated to between
2300 and 1900 cal yr BP may represent a human-set fire in the context of shifting cultivation within or at the edge of a mature rainforest
patch. Evidence in this direction includes the few small pottery shards
recovered from the 20–40 cm interval of the profile, and charred endocarps of E. guineensis mixed with the charcoal of other taxa throughout
the profile (Fig. 3). The traditional human consumption of E. guineensis
is well-documented (Maley and Chepstow-Lusty, 2001; Neumann et al.,
2012a). Furthermore, one charcoal type in the UH48 assemblage is identified as Tabernanthe iboga (APO TAB IBO), a well-known medico-magic
plant commonly used during initiation ceremonies in the Lower Guinea
(e.g., Akendengue et al., 2005; Banzouzi et al., 2008; Protabase, 2012).
The presence of both pottery shards and oil palm endocarp fragments
in the 0–20 cm interval of profile CZ3 (Fig. 3, 626–510 cal yr BP), suggests that this may have been a shifting cultivation site as well, although
at both sites these few pottery shards may well be of younger origin than
the charcoal fragments, since the distribution of charcoal and artifacts in
soil is not always equally affected by post-depositional processes (Cahen
and Moeyersons, 1977). The paleofire events recorded at CZ1 and CZ2 are
almost certainly real wildfires. No artifacts were found in either profiles
and the presence of a few charred oil palm endocarps can be explained
by the fact that both paleofires burned patches of regenerating forest,
the natural habitat of E. guineensis.
Conclusion
Four soil profiles from the Luki reserve in the southern Mayumbe
forest of DRC yielded five distinct charcoal assemblages. One of the
four profiles yielded charcoal from around 530 cal yr BP and a deeper
charcoal layer of >43.5 cal ka BP, i.e., well beyond the period of recorded
anthropogenic burning in this region. This result indicates that both
natural wildfire occurrence and long-term absence of fire are possible
in tropical rainforest. The increasing SST gradient between tropics and
subtropics during the late Holocene caused increasing aridity in the
southern Mayumbe since at least 4000 cal yr BP. However it appears to
have promoted paleofire occurrence only from the end of the third millennium BP when temporarily enhanced seasonality caused severe forest
fragmentation.
Persistence of mature rainforest (e.g., Gilbertiodendron species,
T. bifoliolata) until the end of the third millennium BP rainforest crisis
(2308–1872 cal yr BP), may illustrate the resilience of Central African
rainforest against drought, even at the forest edge. Forest regeneration
following this rainforest crisis was a slow process, as illustrated by our
documentation of a paleofire dated to 1704–1544 cal yr BP, burning a
pioneer forest stand (e.g., P. angolensis, Alstonia species). A more recent
paleofire (626–510 cal yr BP) burned an open woodland savanna patch
(e.g., B. ferruginea) and a stand of pioneer forest (e.g., F. africana, Afzelia)
during the relatively wet period immediately following the MCA
(1100–700 cal yr BP), indicating that the local MCA-equivalent period
was probably a significant drought event in Central Africa and that
forest regeneration was ongoing around 600 cal yr BP.
The documented Luki reserve fire event dated to before 43.5 cal ka BP
was almost certainly a natural wildfire ignited by lightning. In contrast,
pottery shards and oil palm endocarps associated with the paleofire
dated to 2308–1872 cal yr BP and one of the two post-MCA paleofires
suggests that these may have been set by humans practicing shifting
cultivation. As shifting cultivation by Bantu migrants was initially only a
marginal activity practised preferably in regenerating forest, the anthropogenic nature of most fires from that time is far from certain. Temporary
natural climatic drought was probably the main driving force for paleofire
occurrence, vegetation change and human migrations in the Central
African forest. During the last millennium, shifting cultivation by a growing human population increasingly amplified the destructive effects of
natural climate anomalies by slowing down forest regeneration in fireprone areas.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.yqres.2013.04.006.
Acknowledgments
We are indebted to the Special Research Fund of Ghent University
for financing this study as part of the PhD project of W. Hubau. Currently,
W.H. is supported by the ERC Advanced Grant ‘Tropical Forests in the
Changing Earth System’ at Leeds University, UK. We thank the Commission for Scientific Research (Faculty of Bioscience Engineering, Ghent
University) and the King Leopold III Fund for financial support of the
fieldwork, and we thank the World Wide Fund for Nature (WWF), the
École Régionale Post-universitaire d'Aménagement et de gestion
Intégrés des Forêts et Territoires Tropicaux (ERAIFT, DRC) and the
Institut National pour l'Étude et la Recherche Agronomique (INERA,
DRC) for logistic support. Specifically, we thank Geert Lejeune and
Bruno Pérodeau for their assistance and discussions in the field and all
WWF-eco-guards who guided us through the Luki reserve. We thank
the Royal Museum for Central Africa (Tervuren, Belgium) for financing
radiocarbon dating and for organizing the SEM sessions.
References
African Plants Database, 2011. African Plants Database Version 3.3.3, Conservatoire et
Jardin botaniques de la Ville de Genève and South African National Biodiversity
Institute, Pretoria. Published on the internet [accessed 10th January 2011–26th
July 2011], from http://www.ville-ge.ch/musinfo/bd/cjb/africa/.
Akendengue, B., Lemamy, G.J., Bourobou Bourobou, H., Laurens, A., 2005. Bioactive natural
compounds from medico-magic plants of bantu area. Studies in Natural Products
Chemistry 32, 803–820.
Alley, R.B., Ágústsdóttir, A.M., 2005. The 8k event: cause and consequences of a major
Holocene abrupt climate change. Quaternary Science Reviews 24, 1123–1149.
Baert, G., 1995. Properties and Chemical Management Aspects of Soils on Different Parent
Rocks in the Lower Zaire. PhD Dissertation Ghent University (318 pp.).
Banzouzi, J.T., Makambila Koubemba, M.C., Prost, A., Mbatchi, B., Abena, A.A., 2008. Survey of
analgesic plants used by tradipractitioners in Congo Brazzaville. International Journal of
Botany 4, 176–185.
Bird, M.I., Cali, J.A., 1998. A million-year record of fire in sub-Saharan Africa. Nature
394, 767–769.
Brncic, T.M., Willis, K.J., Harris, D.J., Washington, R., 2007. Culture or climate? The relative
influences of past processes on the composition of the lowland Congo rainforest.
Philosophical Transactions of the Royal Society 362, 229–242.
Broadbent, E.N., Asner, G.P., Keller, M., Knapp, D.E., Oliveira, P.J.C., Silva, J.N., 2008. Forest
fragmentation and edge effects from deforestation and selective logging in the Brazilian
Amazon. Biological Conservation 141, 1745–1757.
Bronk Ramsey, C., 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51, 337–360.
Burkill, H.M., 1985. 1st edition. The Useful Plants of West Tropical Africa, 6. Royal Botanic
Gardens, Kew.
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Cahen, D., Moeyersons, J., 1977. Subsurface movements of stone artefacts and their implications for the prehistory of Central Africa. Nature 266, 812–815.
Carcaillet, C., Thinon, M., 1996. Pedoanthracological contribution to the study of the
evolution of the upper treeline in the Maurienne Valley (North French Alps):
methodology and preliminary data. Review of Palaeobotany and Palynology 91,
399–416.
Carcaillet, C., Barakat, H.N., Panaïotis, C., Loisel, R., 1997. Fire and late-Holocene expansion of Quercus ilex and Pinus pinaster on Corsica. Journal of Vegetation Science 8,
85–94.
Cochrane, M.A., 2003. Fire science for rainforests. Nature 421, 913–919.
Cochrane, M.A., Alencar, A., Schulze, M.D., Souza Jr., C.M., Nepstad, D.C., Lefebvre, P.,
Davidson, E.A., 1999. Positive feedbacks in the fire dynamic of closed canopy tropical
forests. Science 284, 1832–1834.
Couralet, C., 2010. Community dynamics, phenology and growth of tropical trees in the
rain forest reserve of Luki, Democratic Republic of Congo. PhD Thesis Faculty of
Bioscience Engineering.Ghent University (173 pp.).
Daïnou, K., Bauduin, A., Bourland, N., Gillet, J.-F., Fétéké, F., Doucet, J.-L., 2011. Soil seed
bank characteristics in Cameroonian rainforests and implications for post-logging
forest recovery. Ecological Engineering 37, 1499–1506.
Daniau, A.-L., Harrison, S.P., Bartlein, P.J., 2010. Fire regimes during the Last Glacial.
Quaternary Science Reviews 29, 2918–2930.
Dechamps, R., Lanfranchi, R., Le Cocq, A., Schwartz, D., 1988. Reconstruction
d'environnements quaternaires par l'étude de macrorestes vegetaux (Pays Bateke,
R.P. du Congo). Palaeogeography, Palaeoclimatology, Palaeoecology 66, 33–44.
Di Pasquale, G., Marziano, M., Impagliazzo, S., Lubritto, C., De Natale, A., Bader, M.Y.,
2008. The Holocene tree line in the northern Andes (Ecuador): first evidence
from soil charcoal. Palaeogeography, Palaeoclimatology, Palaeoecology 259,
17–34.
Donis, C., 1948. Essai d'économie forestière au Mayumbe. Publications de l'Institut National pour l'Etude Agronomique du Congo Belge (INEAC), Bruxelles, Belgique. Série
scientifique 37 (92 pp.).
Dupont, L.M., Jahns, S., Marret, F., Ning, S., 2000. Vegetation change in equatorial West
Africa: time-slices for the last 150 ka. Palaeogeography, Palaeoclimatology, Palaeoecology 155, 95–122.
Elenga, H., Schwartz, D., Vincens, A., Bertaux, J., de Namur, C., Martin, L., Wirrmann, D.,
Servant, M., 1996. Diagramme pollinique holocène du lac Kitina (Congo): mise en
évidence de changements paléobotaniques et paléoclimatiques dans le massif
forestier du Mayombe. Rendered Accounts of the Academy of Sciences, Paris, Series
II 322, 749–755.
Elenga, H., de Namur, C., Vincens, A., Roux, M., Schwartz, D., 2000. Use of plots to define
pollen–vegetation relationships in densely forested ecosystems of tropical Africa.
Review of Palaeobotany and Palynology 112, 79–96.
Elenga, H., Maley, J., Vincens, A., Farrera, I., 2004. Palaeoenvironments, palaeoclimates and
landscape development in Atlantic Equatorial Africa: a review of key sites covering
the last 25 kyrs. In: Battarbee, R.W., Gasse, F., Stickley, C.E. (Eds.), Past Climate Variability through Europe and Africa. Kluwer Academic Publishers, The Netherlands,
pp. 181–198.
Emery-Barbier, A., Thiébault, S., 2005. Preliminary conclusions on the Late Glacial
vegetation in south-west Anatolia (Turkey): the complementary nature of palynological and anthracological approaches. Journal of Archaeological Science 32, 1232–1251.
Hart, T.B., Hart, J.A., Dechamps, R., Fournier, M., Ataholo, M., 1996. Changes in forest
composition over the last 4000 years in the Ituri basin, Zaire. In: van der Maesen,
L.J.G., van der Burgt, X.M., van Medenbach de Rooy, J.M. (Eds.), The Biodiversity
of African Plants. Kluwer Academic Publishers, The Netherlands, pp. 545–563.
Hessler, I., Dupont, L., Bonnefille, R., Behling, H., González, C., Helmens, K.F., Hooghiemstra,
H., Lebamba, J., Ledru, M.-P., Lézine, A.-M., Maley, J., Marret, F., Vincens, A., 2010. Millennial-scale changes in vegetation records from tropical Africa and South America
during the Last Glacial. Quaternary Science Reviews 29, 2882–2899.
Hubau, W., Van den Bulcke, J., Mees, F., Van Acker, J., Beeckman, H., 2012. Charcoal
identification in species-rich biomes: a protocol for Central Africa optimised for
the Mayumbe forest. Review of Palaeobotany and Palynology 171, 164–178.
IAWA Committee, 1989. IAWA list of microscopic features for hardwood identification.
IAWA Bulletin 10, 219–332.
InsideWood, 2011. The InsideWood Working Group (IWG), 2004–onwards. Published
on the Internet [accessed 11th July 2011, 14:30 GMT], from http://insidewood.lib.
ncsu.edu/search.
Leal, M.E., 2001. Microrefugia, small scale ice age forest remnants. Systematics and
Geography of Plants 71, 1073–1077.
Leal, M.E., 2004. The African Rain Forest During the Last Glacial Maximum, An Archipelago
of Forests in a Sea of Grass. PhD Thesis Wageningen University, Wageningen (96 pp.).
Lebamba, J., Ngomanda, A., Vincens, A., Jolly, D., Favier, C., Elenga, H., Bentaleb, I., 2009.
Central African biomes and forest succession stages derived from modern pollen
data and functional plant types. Climate of the Past 5, 403–429.
Lebrun, J., Gilbert, G., 1954. Une classification écologique des forêts du Congo. Publications de l'Institut National pour l'Etude Agronomique du Congo Belge (INEAC),
Bruxelles, Belgique. Série scientifique 63 (89 pp.).
Maley, J., 1996. Le cadre paléoenvironnemental des refuges forestiers africains: quelques
données et hypothèses. In: van der Maesen, L.J.G., van der Burgt, X.M., van Medenbach
de Rooy, J.M. (Eds.), The Biodiversity of African plants. Kluwer Academic Publishers,
The Netherlands, pp. 519–535.
Maley, J., 2001. The impact of arid phases on the African rain forest through geological
history. In: Weber, W., White, L.J.T., Vedder, A., Naughton-Treves, L. (Eds.), African
Rain Forest Ecology and Conservation, An Interdisciplinary Perspective. Yale University
Press, New Haven, pp. 68–87.
Maley, J., 2002. A catastrophic destruction of african forests about 2500 years ago still
exerts a major influence on present vegetation formations. IDS Bulletin 33, 13–30.
339
Maley, J., 2004. Les variations de la végétation et des paléoenvironnements du domaine forestier africain au cours du Quaternaire récent. In: Renault-Miskovsky,
J., Semah, A.M. (Eds.), Guide de la préhistoire mondiale. Artcom'/Errance, Paris,
pp. 143–178.
Maley, J., Brenac, P., 1998. Vegetation dynamics, palaeoenvironments and climatic changes
in the forests of western Cameroon during the last 28,000 years B.P. Review of
Palaeobotany and Palynology 99, 157–187.
Maley, J., Chepstow-Lusty, A., 2001. Elaeis guineensis Jacq. (oil palm) fluctuations in
central Africa during the late Holocene: climate or human driving forces for this
pioneering species? Vegetation History and Archaeobotany 10, 117–120.
Maley, J., Giresse, P., Doumenge, C., Favier, C., 2012. Comment on “Intensifying
weathering and land use in Iron Age Central Africa”. Science 337, 1040-d.
Mayaux, P., Janodet, E., Blair-Myers, C., Legeay-Janvier, P., 1997. Vegetation map of central
Africa at 1:5 000 000. Trees series D: Thematic outputs, 1 (8 pp.).
McBrearty, S., 1990. Consider the humble termite: termites as agents of post-depositional
disturbance at African archaeological sites. Journal of Archaeological Science 17,
111–143.
McCormac, F.G., Hogg, A.G., Blackwell, P.G., Buck, C.E., Higham, T.F.G., Reimer, P.J.,
2004. SHCal04 Southern Hemisphere calibration, 0–11.0 cal kyr BP. Radiocarbon
46, 1087–1092.
Murphy, C.P., 1986. Thin Section Preparation of Soils and Sediments. AB Academic Publishers,
Berkhamsted, U.K.(149 pp.).
Mwavu, E.N., Witkowski, E.T.F., 2008. Sprouting of woody species following cutting and
tree-fall in a lowland semi-deciduous tropical rainforest, North-Western Uganda.
Forest Ecology and Management 255, 982–992.
Neumann, K., Bostoen, K., Höhn, A., Kahlheber, S., Ngomanda, A., Tchiengué, B., 2012a.
First farmers in the Central African rainforest: a view from southern Cameroon.
Quaternary International 249, 53–62.
Neumann, K., Eggert, M.K.H., Oslisly, R., Clist, B., Denham, T., de Maret, P., Ozainne, S.,
Hildebrand, E., Bostoen, K., Salzmann, U., Scwartz, D., Eichhorn, B., Tchiengué, B.,
Höhn, A., 2012b. Comment on “Intensifying weathering and land use in Iron Age
Central Africa”. Science 337, 1040-c.
Ngomanda, A., Jolly, D., Bentaleb, I., Chepstow-Lusty, A., Makaya, M., Maley, J.,
Fontugne, M., Oslisly, R., Rabenkogo, N., 2007. Lowland rainforest response to hydrological changes during the last 1500 years in Gabon, Western Equatorial Africa.
Quaternary Research 67, 411–425.
Ngomanda, A., Chepstow-Lusty, A., Makaya, M., Favier, C., Schevin, P., Maley, J., Fontugne,
M., Oslisly, R., Jolly, D., 2009a. Western equatorial African forest-savanna mosaics: a
legacy of late Holocene climatic change? Climate of the Past 5, 647–659.
Ngomanda, A., Neumann, K., Schweizer, A., Maley, J., 2009b. Seasonality change and the
third millennium BP rainforest crisis in southern Cameroon (Central Africa). Quaternary
Research 71, 307–318.
Oslisly, R., 2001. The history of human settlement in the Middle Ogooué Valley (Gabon):
implications for the environment. In: Weber, W., White, L.J.T., Vedder, A., NaughtonTreves, L. (Eds.), African Rain Forest Ecology and Conservation, An Interdisciplinary
Perspective. Yale University Press, New Haven, pp. 101–118.
Protabase, 2012. Plant Resources of Tropical Africa. Published on the Internet [accessed
February–March 2012], from http://database.prota.org/search.htm.
Reynaud-Farrera, I., Maley, J., Wirrmann, D., 1996. Végétation et climat dans les
forêts du sud-ouest Cameroun depuis 4770 ans BP. Analyse pollinique des sédiments du lac Ossa. Rendered Accounts of the Academy of Sciences, Series II 232,
403–410.
Russell, J.M., Johnson, T.C., 2005. A high-resolution geochemical record from Lake Edward,
Uganda Congo and the timing and causes of tropical African drought during the late
Holocene. Quaternary Science Reviews 24, 1375–1389.
Russell, J.M., Johnson, T.C., 2007. Little Ice Age drought in equatorial Africa: intertropical
convergence zone migrations and El Niño–Southern Oscillation variability. Geology
35, 21–24.
Russell, J.M., Johnson, T.C., Kelts, K.R., Laerdal, T., Talbot, M.R., 2003. An 11 000-year
lithostratigraphic and paleohydrologic record from Equatorial Africa: Lake Edward, Uganda–Congo. Palaeogeography, Palaeoclimatology, Palaeoecology 193,
25–49.
Salzmann, U., Hoelzmann, P., 2005. The Holocene history of the Dahomey Gap: a climatic induced fragmentation of the West African Rainforest. The Holocene 15,
190–199.
Schefuß, E., Schouten, S., Schneider, R.R., 2005. Climatic controls on central African
hydrology during the past 20,000 years. Nature 437, 1003–1006.
Schnitzer, F.A., Bongers, F., 2002. The ecology of lianas and their role in forests. Trends
in Ecology & Evolution 17, 223–230.
Schwartz, D., de Foresta, H., Dechamps, R., Lanfranchi, R., 1990. Découverte d'un premier
site de l'âge du fer ancien (2110 B.P.) dans le Mayombe congolais. Implications
paléobotaniques et pédologiques. Rendered Accounts of the Academy of Sciences,
Paris, Series II 310, 1293–1298.
Scott, A.C., 2000. The Pre-Quaternary history of fire. Palaeogeography, Palaeoclimatology,
Palaeoecology 164, 281–329.
Sosef, M.S.M., 1996. Begonias and African rain forest refuges: general aspects and recent
progress. In: van der Maesen, L.J.G., van der Burgt, X.M., van Medenbach de Rooy,
J.M. (Eds.), The Biodiversity of African plants. Kluwer Academic Publishers, The
Netherlands, pp. 602–611.
Stager, C., Cocquyt, J.C., Bonnefille, R., Weyhenmeyer, C., Bowerman, N., 2009. A late
Holocene palaeoclimatic history of Lake Tanganyika, East Africa. Quaternary Research
72, 47–56.
Stoops, G., 2003. Guidelines for Analysis and Description of Soil and Regolith Thin Sections.
Soil Science Society of America, Madison, WI (184 pp.).
Stoops, G., Marcelino, V., Mees, F., 2010. Interpretation of Micromorphological Features
of Soils and Regoliths. Elsevier (752 pp.).
340
W. Hubau et al. / Quaternary Research 80 (2013) 326–340
Tchouto, M.G.P., de Wilde, J.J.F.E., de Boer, W.F., van der Maesen, L.J.G., Cleef, A.M., 2009.
Bio-indicator species and Central African rain forest refuges in the Campo-Ma'an
area, Cameroon. Systematics and Biodiversity 7, 21–31.
Théry-Parisot, I., Chabal, L., Chrzavzez, J., 2010. Anthracology and taphonomy, from
wood gathering to charcoal analysis. A review of the taphonomic processes modifying
charcoal assemblages, in archaeological contexts. Palaeogeography, Palaeoclimatology,
Palaeoecology 291, 142–153.
Tierney, J.E., Smerdon, J.E., Anchukaitis, K.J., Seager, D., 2013. Multidecadal variability in
East African hydroclimate controlled by the Indian Ocean. Nature 493, 389–392.
van Gemerden, B.S., Olff, H., Parren, M.P.E., Bongers, F., 2003. The pristine rain forest?
Remnants of historical human impacts on current tree species composition and diversity.
Journal of Biogeography 30, 1381–1390.
Van Ranst, E., Baert, G., Ngongo, M., Mafuka, P., 2010. Carte pédologique de la province
du Bas-Congo et ville de Kinshasa, échelle 1:500.000. UGent, Hogent, UNILU,
UNIKIN. 9789491114625.
Verschuren, D., Charman, D.J., 2008. Latitudinal linkages in late Holocene moisturebalance variation. In: Battarbee, R.W., Binney, H.A. (Eds.), Natural Climate Variability
and Global Warming: A Holocene Perspective. Wiley-Blackwell, United Kingdom,
pp. 189–231.
Verschuren, D., Laird, K.R., Cumming, B.F., 2000. Rainfall and drought in equatorial east
Africa during the past 1100 years. Nature 403, 410–414.
Verschuren, D., Sinninghe Damsté, J.S., Moernaut, J., Kristen, I., Blaauw, M., Fagot, M.,
Haug, G.H., 2009. CHALLACEA project members, 2009. Half-precessional dynamics
of monsoon rainfall near the East African Equator. Nature 462, 637–641.
Vincens, A., Schwartz, D., Bertaux, J., Elenga, H., de Namur, C., 1998. Late Holocene climatic
changes in Western Equatorial Africa inferred from pollen from Lake Sinnda, Southern
Congo. Quaternary Research 50, 34–45.
Weldeab, S., Lea, D.W., Schneider, R.R., Andersen, N., 2007. 155,000 years of West African
monsoon and ocean thermal evolution. Science 316, 1303–1307.
Wheeler, E.A., 2011. InsideWood — a web resource for hardwood anatomy. IAWA Bulletin
32, 199–211.
Willis, K.J., Birks, H.J.B., 2006. What is natural? The need for a long-term perspective in
biodiversity conservation. Science 314, 1261–1265.
Worbes, M., Staschel, R., Roloff, A., Junk, W.J., 2003. Tree ring analysis reveals age structure, dynamics and wood production of a natural forest stand in Cameroon. Forest
Ecology and Management 173, 105–123.
WRB, 2006. World reference base for soil resources 2006. IUSS Working Group. World
Soil Resources Reports, 103. FAO, Rome.