Floristic composition, diversity and structure of the forest communities in
the Kouilou Département, Republic of Congo
1,3
2,3
3
5
Margaretha van Rooyen , Noel van Rooyen , Ben Orban , Gilbert Nsongola ,
5
3,4*
Edmond Sylvestre Miabangana and Jérome Gaugris
1Department
of Plant Science, University of Pretoria, Pretoria, South Africa
Pretoria, South Africa
3FLORA FAUNA & MAN, Ecological Services Ltd., Tortola, British Virgin Islands
4Centre for Wildlife Management, University of Pretoria, Pretoria, South Africa
5Centre d’Etude sur les Ressources Végétales, Herbier National du Congo; Brazzaville, Republic of Congo
2Ekotrust,
*Corresponding Author; e-mail: jeromegaugris@florafaunaman.com
Abstract:
The objectives of the study were to classify, describe and map the forest communities in the study area and to
investigate possible gradients in plant diversity. The study area comprised approximately 166 000 ha in the
Kouilou Département, Republic of Congo, a section of land targeted for extensive future development. In total
156 forest sample sites were surveyed using the Braun-Blanquet method of phytosociology. Diversity of each
plant community was expressed in terms of species richness; Shannon-Wiener index; exponent of ShannonWiener index; evenness; Fisher's alpha; Simpson’s index; and inverse Simpson index. Eleven plant communities
were described and mapped at a much finer scale than has been done previously. The communities were
arranged along two gradients: a degraded – relatively intact gradient and a wet – dry gradient. The least
degraded communities, and consequently those with the highest conservation value, were the swamp forests
and the okoumé forest. Overall, the values reported for the diversity parameters in the current study were
within the range reported for other tropical forests in Central Africa. The study revealed a gradient between
the coast (west) and the Mayombe mountain range (east) in plant diversity linked to topography and climate
with plant diversity lowest at the coast. Furthermore, plant diversity was negatively related to the distance to
human infrastructure. The forest communities appear to be fairly resilient to past anthropogenic disturbances,
however, there is no historical analogue to the more severe types of forest destruction associated with some of
the future developments anticipated, such as mechanized logging or mining.
Key words:
Diversity, Marantaceae forest; okoumé forest, p hytosociology, plant communities,
Shannon-Wiener index
Introduction
world (Ernst et al. 2013; Malhi et al. 2013;
Maniatis et al. 2013; Mayaux et al. 2013; Rudel
2013; Singh & Sharma 2009). This situation has
been ascribed to various reasons such as political
The deforestation rate in the Congo Basin has
been notably slower than in the rest of the tropical
1
instability; inaccessibility of the region; lack of
industrialized agriculture; and oil and mineralbased economies. This pattern is however changing,
as many of the central African countries have in
the past ten years become politically relatively
stable enabling an upsurge in both public and
private sector development projects (Megevand et
al. 2013). These projects, accompanied by the
creation of access roads into the vast and resource
rich forest interior of the Congo Basin, bring their
own range of ecological and environmental
problems with local (Abernethy et al. 2013;
Laurance et al. 2006; Petrozzi et al. 2016) as well
as global consequences (Nogherotto et al. 2013).
Environmental problems are further compounded
by the fact that most of the knowledge on the
ecology and diversity of the Congo Basin is
descriptive and dated (De Wasseige et al. 2010),
furthermore large areas remain unexplored
(Corlett & Primack 2011; De Wasseige et al. 2010;
Vande Weghe 2004). The lack of knowledge
precludes proactive management, as it is
impossible to predict the environmental imp-acts
of development with any form of reliability (Willis
et al. 2013). However, as most of the Congo Basin
countries are signatories to international
conventions on biodiversity, the challenge is to
develop sustainable projects to bring economic
growth with minimal negative environmental
impacts. Consequently, research on the ecology
and biodiversity of the Congo Basin is urgently
needed (Cordeiro et al. 2007; Laurance et al. 2012;
Parmentier et al. 2013; Willis et al. 2013).
In the Republic of Congo (ROC), a recent surge
of private projects, spanning agriculture, commercial logging, afforestation, mining, urbanization
and industrialization has been observed. This
development surge is accompanied by active roadbuilding programs to link the major cities within
the country but also with neighbouring countries
(Duveiller et al. 2008). These projects play a vital
role in the growth and development of the
country’s economy and have significant political
implications that may affect the outcome of both
local and national elections. However, these
activities will generally impact negatively on the
environment and its biodiversity. Furthermore,
when such activities are associated with deforestation, this poses a serious threat to the
livelihoods of rural communities who rely directly
on the natural resources provided by the forests
(Brockington et al. 2006; Ickowitz 2006; KameriMbote and Cullet 1997; Perrings and Lovett 1999).
The challenge when exploiting an area for finan-
cial benefit is therefore to ensure that the impacts
on biodiversity and ecosystem services have been
adequately addressed and to that purpose the
habitats and their diversity need to be assessed
and understood (Laurance et al. 2012). Equipped
with knowledge on habitat types and an understanding of their environmental determinants,
dynamics and biodiversity, it becomes possible to
evaluate impacts of development and plan for
formal conservation of biodiversity and natural
resources in such landscapes (Bhagwat et al. 2008;
Cordeiro et al. 2007; Laurance et al. 2006).
The Kouilou Département spans the Atlantic
coastal plain of the ROC and provides a low
altitude savanna and forest continuum between
the rain forests of the Gabon and those of the
Democratic Republic of the Congo. A section of this
coastal plain has been earmarked for intensive
development within the next 10 years and critical
biodiversity knowledge gaps have been identified
between the Conkouati Douli National Park and
the city of Pointe Noire (WCS 2010). Forest resources on this coastal plain are poorly known and
existing inventories are dated (dating from 1985–
1991). Moreover, almost all of the area has been
logged (at least once, but in most areas more) since
the inventories were undertaken (Bayol & Atyi
2009). Overall, there is a dearth of detailed
phytosociological studies in the ROC and the
available vegetation maps at a regional scale are
not sufficient for understanding the pattern and
process governing vegetation occurrence and subsequently for developing environmental management
plans, which recent studies highlight as an
important design feature (Murthy et al. 2016).
The present
study was part of
a
multidisciplinary effort to provide an up-to-date
overview of the biodiversity and its key determinants in the northern section of the Kouilou
Département. The vegetation was chosen as the
basis for the classification of the ecosystems
because it integrates the ecological processes acting
on a site or landscape more measurably than any
other factor or set of factors (Kent 2012). The
vegetation displays patterns that reflect the
influence of environmental factors such as soil,
geology and climate and provides information on
the habitats for animals and rare plant species. It
also reflects the degree of man’s influence,
degradation and temporal aspects operating on it.
The vegetation in the study area comprises
both forests and savanna/grasslands. The current
paper will focus specifically on the forest communities. The aims of this investigation were to
2
i) classify the vegetation in order to identify,
describe and map plant communities, ii) quantify
plantdiversity of the plant communities and place
it in context to other studies conducted in tropical
Africa, iii) investigate a west to east gradient
between the coast and the Mayombe mountain
range in biodiversity linked to topography and
climate (Vande Weghe 2004), iv) investigate an
anthropogenic and utilization gradient on biodiversity, and v) investigate a northwest–southeast
gradient (parallel to coast) in biodiversity within
selected communities.
coast (west) and the Mayombe mountain range
(east) and northwest–southeast along the coast, a
second level of stratification was applied by
sampling at 5 km intervals along these gradients.
Vegetation surveys were conducted over two
seasons (May 2012 and July/August 2012) following the Zurich-Montpellier (Braun-Blanquet)
School of total floristic composition (MuellerDombois & Ellenberg 1974). In total 243 sample
plots were surveyed with 156 of these plots in
forest communities. An assessment of habitat
features (topography, aspect, slope, degree of erosion, stone/rock cover, clay content of the soil,
drainage) was made at each 25 m x 25 m sample
plot delineated. Square sample plots were used for
operational simplicity, although Newton (2007)
lists the limitations of such square plots. Nevertheless, these plots are commonly used in tropical
forest sampling (see Davidar et al. 2007). All
identifiable plant species within the delineated
area were identified and a percentage cover value
allocated to each species according to the Domin
Krajina-scale (Mueller-Dombois & Ellenberg
1974). Furthermore, representative percentage
estimates of the vegetation cover and crown height
were made for the tree, shrub and herbaceous
strata at four/two levels viz. lower, intermediate,
upper and very tall/emergent levels.
Study Area
The study area comprised approximately 165
906 ha in the Kouilou Département in the southwestern corner of the ROC. The study area lies
between two conservation areas (Conkouati-Douli
National Park and Tchimpounga Natural Reserve)
that will be affected by development in this sector.
The study’s exploration area falls within an
‘economic development zone’, which has been earmarked for commercial activities such as mining,
logging and agriculture.
Topographically the study area is quite varied,
consisting of beaches and low dunes along the
Atlantic Ocean, and low undulating terrain on the
coastal plains intersected by numerous drainage
channels forming small lagoons and lakes. To the
east the study area is bounded by the Mayombe
Mountains. Most of the area belongs to the Tertiary
and Quaternary coastal sedimentary basin comprising heavily leached sandy to sandy-clay soils.
Cirque series sand sheets of the Pliocene epoch also
occur on the coastline (Dowsett and DowsettLemaire 1991; Vande Weghe 2004).
Rainfall is restricted to the wet season, which
runs from October until May with a mean of 1200
mm per annum. The mean temperature in the
region is 25°C with variations of 5°C from the
mean during the wet and dry seasons. Mean
relative humidity is 85% with variations of 2% between the seasons (Laclau et al. 2003).
Data analysis
Floristic data classification was done with the
TURBOVEG and JUICE computer programs
(Hennekens & Schaminee, 2001; Tichy, 2002).
TURBOVEG software was used to capture the
data and a TWINSPAN was run in JUICE as a
first step in the classification of the data. Furthermore, To further assist the separation of the
floristic data into groups an Incremental Sum of
Squares (ISS) cluster analysis was run in SYNTAX 2000 (Podani, 2001). For the ISS the
cover/abundance values were converted to percentages (Van der Maarel 2007) and the percentage
values standardised using a natural logarithmic
(loge) standardisation. The Bray-Curtis measure
(Podani 2001) was applied for the analysis. The
ISS of the floristic data of all 243 relevés clearly
indicated the separation between the forest (156
relevés) and savanna/grassland (87 relevés) plant
communities. A second ISS cluster analysis was
performed on only the forest relevés. The resulting
table of sample plots against species was further
refined using Braun-Blanquet tabulation procedures (Werger 1974) to produce a hierarchical
classification.
Methodology
Field survey
Homogeneous areas were delineated on satellite imagery based on vegetation cover and density,
topography, colour and texture as well from
information gathered during a reconnaissance
visit. These homogeneous units provided a first
level of stratification for selecting sample sites. In
order to establish major gradients between the
3
Fig. 1.A principal coordinate analysis of all the plots within the forest communities indicating the presence of
two groups (black line) and 11 communities. Wet–dry and degraded–natural gradients are indicated by arrows.
To visualise the relationship between the
communities the floristic data were ordinated using
principal coordinate’s analysis (PCoA) (McCune &
Grace 2002) in SYN-TAX 2000 (Podani 2001). To
indicate the vegetation succession on fallow land of
different ages, an ordination was then run within
the group of communities in which fallows
occurred.
The following diversity parameters were
calculated across all forest plots (EstimateS
version 9; Colwell, 2013) as well as for each
community (PAST, version 3.02; Hammer 2014):
i. Species richness (S), expressed as the mean
number of species per sample plot;
ii. Shannon-Wiener index (H´), H´=
-∑ ln
where n is number of individuals;
iii. Exponent H´ was calculated to convert H´ to
the effective number of species. This conversion provides a true diversity indicating the
number of equally-common species required to
give the particular value of H´ (Jost 2006);
iv. Evenness (E); calculated as the ShannonWiener index divided by the logarithm of the
number of taxa (S);
v. Fisher's alpha (a), S =a ln (1+n/a) where a is
Fisher's alpha;
vi. The complement of Simpson's index of dominance, Diversity = 1 - ∑
where pi =
proportion of individuals belonging to species I;
and
vii. Inverse Simpson calculated as 1/(1 – D).
4
Table 1. Summary of the hierarchical classification of the forest communities/subcommunities of the coastal
plains in the Kouilou Département, Republic of Congo.
Community
No.
F1.
Basic description, including sub communities if applicable
Megastachya mucronata Degraded Forest
1a.
F2.
1b.
Megastachya mucronata- Ancistrocarpus densispinosus Degraded Forest
1c.
Megastachya mucronata - Ipomoea involucrata Degraded Forest
1d.
Megastachya mucronata - Afromomum longipetiolatum – Tetrorchidium didymostemon
Degraded Forest
1e.
Megastachya mucronata - Haumania liebrechtsiana – Musanga cecropioides Degraded Forest
1f.
Megastachya mucronata - Mussaenada chippii – Argocoffeopsis eketensis Degraded Forest
1g.
Megastachya mucronata - Ricinodendron heudelotii – Trema orientalis Degraded Forest
1h.
Megastachya mucronata - Pteridium aquilinum – Manihot esculenta Degraded Forest
1i.
Megastachya mucronata - Croton haumanianus – Macaranga barteri Degraded Forest
Haumania liebrechtsiana - Cynometra lujae Marantaceae Forest
2a.
F3.
F4.
F5.
Sherbournia bignoniiflora - Argocoffeopsis eketensis Degraded Forest
Haumania liebrechtsiana - Chazaliella letouzeyi Marantaceae Forest
2b.
Haumania liebrechtsiana - Tabernaemontana crassa Marantaceae Forest
2c.
Haumania liebrechtsiana - Cynometra lujae Marantaceae Forest
2d.
Haumania liebrechtsiana - Pentaclethra eetveldeana Marantaceae Forest
Aucoumea klaineana - Klainedoxa gabonensis - Chrysobalanus icaco Okoumé Forest
3a.
Aucoumea klaineana - Sacoglottis gabonensis OkouméForest
3b.
Aucoumea klaineana - Geophila afzelii Okoumé Forest
Hymenocardia ulmoides - Milletia comosa Secondary Forest
4a.
Sorindeia juglandifolia - Chaetocarpus africanus Secondary Forest
4b.
Pentaclethra macrophylla - Hymenocardia ulmoides Secondary Forest
4c.
Carapa procera - Pentaclethra eetveldeana Secondary Forest
4d.
Millettia comosa - Thomandersia butayei Secondary Forest
Manilkara obovata - Manotes expensa - Symphonia globulifera Cirque Forest
F6.
Mangifera indica - Chrysobalanus icaco Anthropogenic Forest
F7.
Manilkara obovata - Borassus aethiopica - Chrysobalanus icaco Coastal Thicket
F8.
Manikara obovata - Syzygium guineense Ecotonal Forest
F9.
Lasimorpha senegalensis Coastal Swamp Forest
F10.
Anthostema aubryanum - Xylopia rubescens Lowland Swamp Forest
10a
F11.
Symphonia globulifera - Raphia hookeri Lowland Swamp Forest
10b
Elaeis guineensis - Hallea stipulosa Lowland Swamp Forest
10c
Uapaca guineensis - Gilbertiodendron dewevrei Lowland Swamp Forest
10d
Anthostema aubryanum - Piper guineensis Lowland Swamp Forest
Rhizophora racemosa Mangrove Forest
Map preparation
by using verified information from the sample
sites; ii) handmapping of different vegetation types
based on vegetation types highlighted, vegetation
cover and density, topography, colour and texture
of the satelite pictures used (2.5 - 5.0 m resolution
sourced from Google Earth ©), iii) for areas where
definition of different habitat types by visual obser-
Once the data analysis was complete, the
preliminary map from the reconnaissance study
was amended to its final version as appear in Fig.
2. The mapping process used a combination of the
following: i) representation of each vegetation type
5
Fig. 2. Vegetation map of the study area in the Kouilou Département, Republic of Congo.
6
Table 3. Summary of species richness and Shannon-Wiener index of diversity for the different plant
communities of Sintoukola, Republic of Congo.
For entire community
Commu- Number Total
Total
Exponent
nity no
of
area
Simpson Inverse
Species Shannonnumber
Shannonsample covered
richness Wiener
(1 - D) Simpson
of species
Wiener
plots
(ha)
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
All plots
44
19
9
21
6
17
5
5
8
17
5
156
2.75
1.19
0.56
1.31
0.38
1.06
0.31
0.31
0.50
1.06
0.31
222
202
110
138
56
101
50
14
55
152
23
466
Total for
community
Mean values per plot
23
28
28
21
20
19
20
4
14
22
7
2.6
2.6
2.5
2.4
2.2
2.1
2.1
0.8
1.8
2.5
1.0
14.84
16.63
13.35
12.20
9.75
8.98
8.00
3.08
6.47
13.19
2.91
132.8
0.86
0.84
0.82
0.83
0.77
0.75
0.78
0.39
0.71
0.83
0.44
9.95
10.16
6.16
7.65
4.58
4.73
4.59
2.96
3.90
8.72
1.92
Evenness
0.84
0.83
0.76
0.82
0.76
0.72
0.69
0.74
0.71
0.81
0.53
Fisher’s Shannon- Inverse
Alpha Wiener Simpson
31.93
44.96
23.50
19.83
18.92
16.28
10.07
0.64
8.72
20.65
2.97
112.6
4.10
3.95
3.49
4.00
3.10
3.12
2.70
1.23
3.11
3.77
1.70
4.9
60.11
52.12
32.66
54.75
22.15
22.74
14.92
3.41
22.37
43.52
5.45
61.14
vation was not possible a categorisation procedure
was performed in ArcView 9.2 over the whole area
using 15 classes: expert opinion was then used to
define whether the classes obtained in this manner
were relevant or not, iv) a land use and land use
change analysis performed for the study area was
also used to guide the differentiation between
secondary forest and mature forest.
ceous strata as well as the structure in terms of
mean crown height and mean crown cover (Table 2).
For the differences in species composition between
the subcommunities the reader is referred to a full
differential table summarising species in each of the
communities based on cover abundance values (copy
to be requested from the corresponding author).
Results
These degraded or secondary forests were associated with exploration tracks or with abandoned
fallows. In general these forests were considered as
degraded as they had been recently (within the
past 7 years) disturbed by people, however, in some
instances these forests were already in a process of
recovery and were then considered as secondary
forests. These forests occurred in valleys at a mean
distance of 28.0 km from the coast at a mean altitude of 52 m above sea level. Because these degraded forests were either narrow bands following
tracks or small, previously cultivated patches, they
could not be mapped separately (Fig. 2). The community was found on sandy soil and erosion was low
or non-existent. The mean distance to human
infrastructure was 96 m (calculated as distance
from plot centre to nearest man made item such as
track, road or house).
The community as a whole was differentiated by
species group N (full differential table - source from
corresponding author) including Megastachya mucronata, Camoensia brevicalyx and Vernonia brazza-
F1. Megastachya mucronata Degraded Forest
Eleven forest communities (F1 - F11), of which
five were further subdivided into subcommunities,
were identified (Table 1, Fig. 1 and 2). The full
differential table of the forest communities is
available from the corresponding author (jeromegaugris@florafaunaman.com). Overall, the hierarchical classification of the floristic data was
supported by the ISS cluster analysis and the
PCoA ordination (Fig. 1). The cluster analysis of
the forest communities indicated a separation between communities F1 - F4 and communities F5 F11 (indicated in Fig. 1) with the two subcommunities of F3 divided between the two groups. Subcommunity 11a was separated from the rest of the
subcommunities of F11, possibly as a result of
differences in permanent versus seasonal inundation.
The following descriptions provide a brief
assessment of the salient habitat features, the
species composition of the tree, shrub and herba-
7
Table 2. Mean crown cover ( standard error) and crown height ( standard error) of the tree, shrub and
herbaceous stratum at four different levels in the forest communities.
Tree stratum
Cover (%)
Shrub stratum
Herbaceous stratum
Height (m)
Cover (%)
Height (m)
Cover (%)
Height (m)
26.01.1
50.00.0
150.0
-
-
F1: Megastachya mucronata Degraded Forest
Emergent/very high level
38.84.7
Upper level
49.05.0
20.51.1
-
-
90.00.0
4.00.0
Intermediate level
46.013.3
20.010.4
26.33.0
4.63.5
65.65.9
4.00.4
Lower level
24.47.5
0.70.1
32.57.8
1.60.1
58.35.8
1.30.1
F2: Haumania liebrechtsiana - Cynometra lujae Marantaceae Forest
Emergent/very high level
31.95.9
30.92.1
56.43.2
15.00.0
44.36.9
14.31.3
Upper level
56.43.2
24.711.5
39.24.0
13.31.1
31.88.0
8.30.0
Intermediate level
44.55.3
21.41.4
36.53.8
7.80.5
31.17.7
5.31.1
Lower level
49.45.8
0.50.0
46.45.9
1.30.2
46.25.0
1.40.2
-
-
F3: Aucoumea klaineana - Klainedoxa gabonensis - Chrysobalanus icaco Okoumé Forest
Emergent/very high level
45.78.4
28.91.6
25.017.6
18.31.7
Upper level
45.75.7
22.91.8
40.04.9
11.11.0
-
-
Intermediate level
46.09.3
16.81.3
35.76.5
8.10.6
20.00.0
4.52.5
Lower level
64.37.2
0.60.1
57.17.8
1.60.2
35.77.2
0.70.2
F4: Hymenocardia ulmoides - Milletia comosa Secondary Forest
Emergent/very high level
25.06.5
21.32.4
-
-
-
-
Upper level
60.63.1
17.91.2
45.07.8
9.30.4
-
-
Intermediate level
48.93.1
13.11.9
46.14.4
5.80.4
25.94.9
3.80.6
Lower level
28.84.8
0.90.1
41.16.0
1.40.2
38.56.1
1.10.2
-
-
F5: Manilkara obovata - Manotes expensa - Symphonia globulifera Cirque Forest
Emergent/very high level
60.00.0
25.00.0
-
-
Upper level
56.05.1
15.63.1
45.015.0
9.01.0
-
-
Intermediate level
57.511.1
10.33.4
46.75.7
4.50.8
50.040.0
2.00.0
Lower level
30.06.8
0.60.1
50.08.9
1.30.2
59.28.2
1.10.2
F6: Mangifera indica - Chrysobalanus icaco Anthropogenic Forest
Emergent/very high level
55.05.0
35.05.0
-
-
-
-
Upper level
59.34.1
19.51.3
30.05.8
10.72.3
-
-
Intermediate level
52.34.4
13.31.4
34.03.2
6.70.5
53.87.3
2.70.4
Lower level
30.04.1
0.80.1
37.13.7
1.80.2
61.44.4
1.00.1
-
-
F7: Manilkara obovata - Borassus aethiopica - Chrysobalanus icaco Coastal Thicket
Emergent/very high level
-
-
-
-
Upper level
53.38.8
15.74.7
-
-
-
-
Intermediate level
66.78.8
10.74.8
38.313.6
5.31.5
-
-
Lower level
17.77.9
0.80.2
40.011.5
1.30.3
50.010.0
1.00.0
F8: Manikara obovata - Syzygium guineense Ecotonal Forest
Emergent/very high level
-
-
-
-
-
-
Upper level
67.54.8
15.52.5
-
-
-
-
Intermediate level
60.05.8
10.51.7
45.08.7
5.50.9
-
-
Lower level
40.04.1
0.50.0
47.54.8
1.90.1
50.022.2
1.00.8
Contd...
8
Tree stratum
Cover (%)
Shrub stratum
Height (m)
Herbaceous stratum
Cover (%)
Height (m)
Cover (%)
Height (m)
-
-
F9: Lasimorpha senegalensis Coastal Swamp Forest
Emergent/very high level
30.010.8
26.31.3
-
-
Upper level
57.013.4
19.02.4
27.52.5
9.01.0
-
-
Intermediate level
47.54.8
12.01.8
35.81.8
5.00.3
30.020.0
3.51.5
Lower level
42.513.8
1.40.4
25.08.1
1.00.2
49.312.8
1.40.1
F10: Anthostema aubryanum - Xylopia rubescens Lowland Swamp Forest
Emergent/very high level
43.34.8
31.31.5
43.312.0
13.31.7
-
-
Upper level
55.33.5
25.31.1
28.66.3
10.91.1
-
-
Intermediate level
41.43.7
16.81.3
31.45.0
6.60.6
33.98.7
3.10.3
Lower level
41.34.8
1.30.2
38.77.0
1.70.1
59.76.8
1.30.1
20.00.0
25.00.0
-
-
-
-
Upper level
62.511.8
20.02.0
-
-
-
-
Intermediate level
47.511.1
15.02.0
48.010.7
4.00.6
50.00.0
1.00.0
Lower level
33.312.0
1.00.1
45.015.8
1.70.4
32.514.4
0.60.1
F11: Rhizophora racemosa Mangrove Forest
Emergent/very high level
villensis. Nine sub communities were distinguished. Overall, species composition was highly
variable and dependent on the time since disturbance or abandonment of the fallows.
The most common tree species included
Musanga cecropioides, Pentaclethra eetveldeana,
Millettia comosa and Xylopia aethiopica. Species
such as Trema orientalis, Cogniauxia podolaena,
Vernonia brazzavillensis and Ricinodendron heudeloti were locally prominent. The shrub layer was
represented predominantly by Alchornea cordifolia, Croton haumanianus, Mussaenda chippii,
Chromolaena odorata, Manihot esculenta and the
scandent shrub Argocoffeopsis eketensis. Megastachya mucronata and Panicum brevifolium were
the most prominent grass species and Scleria
boivinii the most abundant sedge. Although the
species composition of the herbaceous layer varied
among subcommunities, the dominant contributing species were Haumania liebrechtsiana,
Cissus oreophila, Cyathula prostrata and Aframomum longipetiolatum.
Fig. 3 represents an ordination on only F1 - F4.
Several plots were on fallow land and the age since
abandonment of the cropland has been indicated in
the figure. Recently abandoned cropland within F1
occurred predominantly towards the top of the
ordination plane, whereas the oldest fallow sites
occurred towards the bottom of the ordination
plane in communities F1 but also F2. All the fallow
sites within community F4 were closely grouped
and lay adjacent to F1.
Structurally, the tree stratum was well developed in the emergent and upper canopy levels
(Table 2). The shrub stratum had the highest cover
values for the intermediate and lower levels(Table
2). Only two levels were distinguished for the
herbaceous stratum, with the intermediate level
being best developed (Table 2).
F2. Haumania liebrechtsiana - Cynometra lujae
Marantaceae Forest
These Marantaceae forests covered 17% of the
study area and occurred in valleys of the Mayombe
Mountain foothills furthest inland of the current
study area. Mean distance to the coast was 29.8
km and mean altitude 51 m above sea level. The
community occupied sandy soil and erosion was
low to non-existent. Human influence such as
logging and plant and firewood collecting was
observed. The mean distance of the sample plots to
human infrastructure was 180 m.
This community (except subcommunity 2d)
was differentiated by species group S (full
differential table - source from corresponding
author) with diagnostic species such as Staudtia
kamerunensis,
Synsepalum
longecuneatum,
Diospyros hoyleana and Pausinystalia johimbe the
most noteworthy. Four subcommunities were
identified.
9
Fig. 3. Principal coordinate analysis scatter diagram of Group 1-forest communities indicating possible
successional pathways. Fallow sites (open symbols) of known age are qualified by the number of years since
abandonment and the arrows indicate possible successional sequences. F1, Degraded forest; F2, Marantaceae
forest; F3, Okoumé forest; and F4, Secondary Forest.
The most prominent tree species included the
diagnostic species mentioned above as well as
Garcinia kola, Dicranolepis laciniata, Caloncoba
welwitschii, Dialium polyanthum, Dichostemma
glaucescens, Pentaclethra eetveldeana and Trichilia heudelotii. Other conspicuous tree species
were Tabernaemontana crassa, Croton mayumbensis, Dracaena reflexa, Panda oleosa and Coula
edulis. Shrub species included Diospyros hoyleana,
Dialium polyanthum, Thomandersia butayei and
the liana Manniophyton fulvum. The herbaceous
layer was represented by Haumania liebre-
chtsiana, Agelantha villosiflorum, Mostuea brunonis and Nephthytis afzelii. Haumania liebrechtsiana was the dominant herbaceous species
often forming an almost impenetrable lower
canopy. Guaduella oblonga was the only grass
species recorded.
The multilayered tree canopy was best
developed in the upper and lower levels (Table 2).
Woody liana in the shrub layer extended to a mean
height of 15 m, but mean crown cover was highest
at the intermediate and lower levels. The
herbaceous stratum was best developed at the
10
lower level (cover 31%) primarily due to the abundance of Marantaceae species.
southfacing footslopes on sandy soil. Mean
distance from the coast was 16.6 km although the
community occurred along almost the entire coast
to inland gradient. Mean altitude was 46 m above
sea level. Human influence, such as logging and
plant collecting, was low. The mean distance of the
sample plots to human infrastructure was 488 m.
Four subcommunities were distinguished, but
there was no diagnostic species group that
differentiated the community (full differential
table - source from corresponding author).
A large number of tree species were common,
e.g. Hymenocardia ulmoides, Chaetocarpus africanus, Carapa procera, Anthocleista schweinfurthii, Manotes expansa, Greenwayodendron suaveolens, Cola heterophylla and locally Pentaclethra macrophylla and Pentaclethra eetveldeana.
Although species such as Sorindeia juglandifolia,
Greenwayodendron suaveolens and Chaetocarpus
africanus can exceed the current mean crown
height, the woody composition was dominated by
species such as Hymenocardia ulmoides and
Manotes expansa, which generally do not become
such tall trees. Prominent shrub species included
Psychotria peduncularis, Heinsia crinita, Morinda
morindoides, Caloncoba welwitschii, Brenandendron donianum and scandent shrubs such as
Tetracera alnifolia, Cnestis iomalla and Millettia
comosa were conspicuous.Thomandersia butayei
and Morinda morindoides were typically found
along the fringes of the community. No grass
species were recorded and the herbaceous layer
consisted of species such as Gnetum africanum
(liana), Aframomum citratum and Palisota spp.
The multi-layered tree stratum was best
developed in the upper canopy and the crown cover
of the shrub stratum was almost equally developed
in the intermediate and lower levels (Table 2). The
herbaceous stratum had the highest cover (31%) at
the lower level.
F3. Aucoumea klaineana - Klainedoxa gabonensis Chrysobalanus icaco Okoumé Forest
The okoumé forests covered 12% of the study
area and formed a discontinuous band seawards of
the Marantaceae forests (F2). Mean distance to the
coast was 25.6 km and mean altitude was 37 m
a.s.l. The community occupied sandy soil and
erosion was low to non-existent. Although soils were
not permanently waterlogged, seasonal inundation
could occur, especially in subcommunity 3a. This is
demonstrated by the position of these plots within
the wetland sites in the ordination (Fig. 1). Signs
of human activity were represented by logging and
plant collecting, especially of Gnetum africanum,
and firewood collecting. The mean distance to
human infrastructure was 479 m.
Community F3 was differentiated by species
group Y (full differential table - source from corresponding author) containing diagnostic species such
as Aucoumea klaineana, Klainedoxa gabonensis
and Quassia gabonensis. The two subcommunities
could be distinguished by the presence of Sacoglottis gabonensis (tree) and the creeping herb
Geophila afzelii respectively.
The most prominent tree species included the
diagnostic tree species as well as Sacoglottis
gabonensis, Chrysobalanus icaco, Symphonia globulifera, Cassipourea sp., Sorindeia juglandifolia and
Raphia hookeri. The shrubs Chrysobalanus icaco,
Cnestis iomalla, Gaertnera paniculata, Thomandersia butayei, Raphia hookeri and lianas all
featured prominently. Due to the dense cover of
the woody layer, sunlight penetration to the lower
canopy was limited and no grass species were
recorded. Vines and lianas such as Dioscorea
smilacifolia, Agelaea poggeana, Smilax anceps,
Geophila afzelii and Cnestis iomalla were
abundant.
Crown cover of the emergent (32%) and lower
tree levels (45%) was among the highest of all
communities in the study area (Table 2). The
shrub stratum was best developed in the lower
level, although high cover values were also found
at the upper and intermediate levels because of
the abundance of liana. The herbaceous stratum
was best developed in the lower level.
F5. Manilkara obovata - Manotes expansa
Psydrax moandensis Cirque Forest
-
These forests covered 4% of the study area and
were associated with the coastal zone. The
community occurred mainly on steep slopes, in
cirques, at a mean distance of 2.1 km from the coast
and a mean altitude of 56 m above sea level. Soils
had low clay content and erosion was moderate to
high. Human influence was low, but firewood and
plant collecting were noted. The mean distance to
human infrastructure was 376 m.
There was no species group that differentiated
this community, and it showed strong floristic
F4. Manotes expansa - Millettia comosa Secondary
Forest
These secondary forests covered 17% of the
study area and were found predominantly on
11
relationships with both F6 and F8 (full differential
table - source from corresponding author). The
most conspicuous woody species were Manilkara
obovata, Symphonia globulifera, Chrysobalanus
icaco, Elaeis guineensis, Syzygium guineense,
Psydrax moandensis, Anthocleista schweinfurthii,
Xylopia aethiopica, Heinsia crinita and Chaetocarpus africanus.
Although the mean height of the emergent tree
level was low, several species had the potential to
reach almost double this height. High cover values
were recorded for the upper and intermediate tree
canopy, while in the shrub stratum the intermediate and lower levels were best developed
(Table 2). Mean crown cover of the herbaceous
stratum was 51% at the lower level.
F6. Mangifera indica Anthropogenic Forest
Chrysobalanus
(40%) and lower levels (43%). The herbaceous
cover was highest at the lower level.
F7. Manilkara obovata - Borassus aethiopicum Dalbergia ecastaphyllum Coastal Thicket
These thickets covered 3% of the study area
and occurred in a narrow strip along the coast at a
mean distance of 0.32 km from the coast. The mean
altitude within the community was 15 m above sea
level. Soils had a low clay content and no erosion
was visible. Human influence was low and plant
collecting was the only activity noted. The mean
distance to human infrastructure was 1102 m.
This community was differentiated by species
group AN (full differential table - source from
corresponding author) including diagnostic species
such as Borassus aethiopicum, Heterotis rotundifolia, Dalbergia ecastaphyllum and Phoenix
reclinata.
Palms, Borassus aethiopicum and Phoenix
reclinata, were a prominent feature of this community. The most conspicuous trees were Manilkara
obovata, Chrysobalanus icaco, Anthocleista vogelii
and Syzygium guineense. In the shrub layer
Psychotria peduncularis and Tricalysia coriacea
were noticeable together with the liana Dalbergia
ecastaphyllum, which could extend into the upper
canopy to a height of 10 m. No grass species were
recorded and the most conspicuous herbaceous
species were Asystasia gangetica, Heterotis
rotundifolia, Aframomum sericeum, Sansevieria
longiflora and Abrus canescens.
The tree stratum was best developed at the
intermediate level (Table 2). Shrubs covered
around 30 % of the area in the intermediate and
lower levels, while the herbaceous stratum was
prominent only at the lower level. A substantial
amount of litter was present on the forest floor.
F8. Manilkara obovata - Syzygium guineense
Ecotonal Forest
These coastal forests were not mapped
separately since they represented a narrow
ecotone on the fringe of community F5. The
community occurred at a mean distance of 1.8 km
from the coast and a mean altitude of 52 m above
sea level. Soils had low clay content and no erosion
was recorded. Some plant collection by the local
inhabitants was noted. The mean distance to
human infrastructure was 1090 m.
This species-poor community had no diagnostic
species group and was characterised by the
dominance of Manilkara obovata and Syzygium
guineense.
The emergent layer was absent and the upper
icaco
These anthropogenic forest patches were found
on the coastal plain and covered <1% of the study
area. Mean distance from the coast was 4.9 km
and mean altitude was 51 m above sea level. Soils
had a low clay content and no erosion was visible.
Current human influence was low although signs
of plant collecting were observed. The mean
distance to human infrastructure was 917 m.
The community was differentiated by species
group AK (full differential table - source from
corresponding author) with Mangifera indica,
Millettia versicolor, Bambusa vulgaris and
Rauvolfia vomitoria representing the diagnostic
species.
The tree layer was represented by the abovementioned diagnostic tree species as well as by
Chrysobalanus icaco, Xylopia aethiopica, Anthocleista schweinfurthii, Elaeis guineensis and
Macaranga spinosa. The dominant tree was
Mangifera indica. Rauvolfia vomitoria, Tetracera
alnifolia (liana) and the alien Lantana camara
were the dominant shrub species. Manilkara
obovata occurred predominantly in a shrub form in
this plant community, although it can attain
heights exceeding 30 m. Almost no grass species
were recorded and herbaceous species included
Gnetum africanum, Vismia affinis, Abrus
canescens and the fern Nephrolepis biserrata.
Where Bambusa vulgaris had established, it
formed virtually impenetrable monocultures that
excluded all other plant species.
Mean height of the emergent tree level was the
highest of all communities, although crown cover
at this level was low (Table 2). Shrubs contributed
fairly evenly to crown cover in the intermediate
12
mean distance to human infrastructure was 956 m.
The community was differentiated by species
group AZ (full differential table - source from
corresponding author) with Anthostema aubryanum, Xylopia rubescens, Podococcus barteri,
Baphia sp., Laccosperma secundiflorum and Coelocaryon preussii some of the diagnostic species.
Four subcommunities were distinguished.
The woody layer was represented by the
diagnostic tree species as well as Chrysobalanus
icaco, Symphonia globulifera, Vitex grandifolia,
Gilbertiodendron dewevrei, Fleroya stipulosa and
Anthocleista vogelii. Palm species were prominent
e.g. Elaeis guineensis, Raphia hookeri, Podococcus
barteri and Laccosperma secundiflorum. The
dominant shrub and small tree species included
Tetracera alnifolia, Warneckea membranifolia,
Piper guineensis and Entada abyssinica. Nymphaea
lotus was found only in open water habitats.
This swamp forest had the highest crown cover
(35 %) for the emergent tree level of all forest
communities observed in the current study.
Overall, the tree layer was fairly well developed at
all levels and the high crown cover value for the
very high shrub level was noteworthy (Table 2).
The herbaceous stratum was well developed at the
lower level with a cover of 60 %.
F11. Rhizophora racemosa Mangrove Forest
These mangroves covered 1 % of the study
area and occurred in saline swampy conditions.
The sampled sites were located at a mean distance
of 3.3 km from the coast. Mean altitude within the
community was 30 m above sea level. Soils had a
low to medium clay content. No signs of animal or
human activity were observed. The mean distance
to human infrastructure was 840 m.
This mangrove community was characterised
by Rhizophora racemosa, Hibiscus tilliaceus,
Raphia subnuda, Eugenia sp. and Dracaena
mannii (species group BB, full differential table source from corresponding author).
The mangrove Rhizophora racemosa formed an
almost impenetrable monoculture. Other tree
species that were visually dominant in the emergent level were Terminalia superba and Ceiba
pentandra although neither were encountered
within the sampling sites. Typical shrub species
included Hibiscus tilliaceus and Dracaena mannii.
Although not recorded within the sampling sites,
Vossia cuspidata and Cyperus papyrus does occur
along the river embankments.
The tree stratum was best developed in the
upper canopy, while the shrub layer had high
and intermediate tree canopy levels had high cover
values although the height was not particularly
high (Table 2). The herbaceous and shrub strata
were co-dominant at the lower level.
F9. Lasimorpha
Forest
senegalensis
Coastal
Swamp
Community F9 covered 1% of the study area at
a mean distance of 3.3 km from the coast and mean
altitude of 20 m above sea level. It occurred mostly
in the southern littoral zone on flat, seasonallyflooded terrain with small raised promontories.
Soils had a low to medium clay content and no
erosion was discernible. Signs of plant collecting by
local inhabitants were visible. The mean distance to
human infrastructure was 969 m.
This community was differentiated by species
group AS (full differential table - source from
corres-ponding
author)
with
Lasimorpha
senegalensis, Voacanga thouarsii, Stipularia
africana and Panicum parvifolium representing
the diagnostic species.
The most prominent tree species were
Anthocleista vogelii, Syzygium guineense, Elaeis
guineense and Chrysobalanus icaco. Mayor contributors to the shrub layer were Voacanga thouarsii,
Stipularia africana, Landolphia incerta and
Tetracera alnifolia. The most prominent grass
species was Panicum parvifolium, which was
particularly abundant along the periphery of the
swamps. Other conspicuous herbaceous species
included Gloriosa superba, Lasimorpha senegalensis and Aframomum sericeum. The latter
species generally exhibits a preference for slightly
drier environments such as the raised promontories.
The tree stratum was best developed at the
upper canopy level, although a distinct emergent
layer was also present (Table 2). The shrub
stratum was best developed at the intermediate
and lower levels. In the herbaceous stratum cover
was highest at the lower level (49 %).
F10. Anthostema aubryanum - Xylopia rubescens
Lowland Swamp Forest
These lowland swamp forests were predominantly associated with the Kouilou River and
covered 7 % of the study area. They occurred on
flat terrain at a mean distance of 11.9 km from the
coast and a mean altitude of 23 m above sea level.
Soils were seasonally or permanently inundated
and areas of open water were often present. Very
low levels of human activity were observed. The
13
However, there was a significant positive, linear
relationship between the mean distance to the
coast and mean exponent H´ of a community
(Fig. 4a). In contrast, there was a significant
negative relationship between the distance to the
nearest human infrastructure and mean exponent
H´ of a community. This relationship was even
stronger if the swamp forest communities were
excluded because of possible problems of
accessibility (Fig. 4b). Exponent H´ did not show a
trend along the northwest-southeast coastal
gradient (not illustrated) in communities F3 and
F4, which both occurred along this gradient.
Five Red Data species were recorded in the
sample plots: Aucoumea klaineana (F1, F2 & F3)
and Fleroya stipulosa (F10) are both classified as
Vulnerable, whereas Gnetum africanum (all
communities except F5 & F11) is classified as Near
Threatened and both Irvingia gabonensis (F10)
and Milicia excelsa (F1) are classified as Near
Threatened/Lower Risk (www.iucnredlist.org).
a
b
Discussion
Phytogeography and endemism
Phytogeographically, the forest communities
fall within the Guineo-Congolian Region of
Endemism as defined by White (1979, 1983). Three
subcentres (White 1979) are recognized within this
region: Upper Guinea, Lower Guinea and Congolia.
The study area falls within the Lower Guinea
subcentre.
White (1983) described the Mayombe region as
part of the dry semi-deciduous forest (seasonal
forest), an interpretation that is not shared by
Dowsett & Dowsett-Lemaire(1991), Hecketsweiler
& Mokoko-Ikonga (1991) and Doumenge (1992),
who all regard the vegetation as wet, semi-evergreen tropical forest. In their reconstruction of the
biomes, based on a pollen analysis, Lebamba et al.
(2009) concurred with the latter authors that
floristically the Mayombe region had more affinities to the wet, semi-evergreen rainforest than
the dry, seasonal forests. Although the rainfall in
the Mayombe region is less than 1500 mm/annum
and the dry season more than three months long,
Lebamba et al. (2009) contend that the high cloud
cover and high relative atmospheric humidity
during the dry season are more important than the
annual rainfall amount and dry season length in
determining the floristic composition of the rain
forest. In their recent analysis of patterns of tree
species composition across African tropical forests
Fig. 4. Linear relationship between plant diversity
(indicated by the exponent of the Shannon-Wiener
index) and a. the distance to the coast and b. the
distance to infrastructure.
cover values at the intermediate and lower levels
(Table 2). The herbaceous stratum was present
only at a lower level.
Forest diversity and Red Data species
In total 465 taxa were recorded in the 156
forest plots sampled across eleven communities.
The mean number of species was highest in F2 and
F3 (28 species per plot), whereas the exponent H´
and inverse Simpson, which both compare effective
number of species, pointed to F2 and F1 as being
the most diverse (Table 3). Evenness, ShannonWiener (H´) and Simpson’s index were all also
highest for F1 and F2. All measures of diversity
indicated that F8 and F11 were the most speciespoor communities.
There was no relationship between mean
altitude of a community and the mean exponent
H´ (R2 = 0.0637, P > 0.05; not illustrated).
14
Fayolle et al. (2014b) distinguished six clusters. On
the basis of the floristic evidence the vegetation in
the study area was classified by Fayolle et al.
(2014b) as part of the Wet Central African cluster.
nated by Chromolaena odorata, followed by Hymenocardia ulmoides, Harungana madagas-cariensis,
Trema guineensis and Xylopia aethiopica after 3 4 years in the second and third stages. More
phanerophytes occur in the fourth stage, when the
vegetation corresponds to a pre-forest stage.
Several authors have commented on an apparent
change in the successional sequence in recent
years. Originally the successional pathway on
abandoned cultivated land in Mayombe was
through a Musanga cecropioides dominated stage
(De Foresta & Schwartz 1991). However, since the
permanent settling of villages and associated
increased population density, fallows of shorter
duration than previously are applied. This
increased cultivation intensity has led to fallow
fields being invaded by Chromolaena odorata or
sometimes Pteridium aquilinum (De Foresta &
Schwartz 1991, Ngobo et al. 2004) as was reported
in the current study. When considering the age of
the fallows, the successional pathway seemed to
run from F1 to F2 (Fig. 3). Alternatively, the
progression could proceed from F1 through F4 to
an okoumé forest stage (F3).
Community F4 was also a degraded or
secondary forest. It had a lower mean tree height
for the emergent, upper and intermediate levels
than F2 and was consequently considered to
represent a younger successional stage than F2
(Marantaceae forest).
Marantaceae forests, such as represented by
F2 in the current study, are common in the
lowland rainforest in the Congo Basin and have
been described by various authors (Cusset 1989,
Vande Weghe 2004). These forests are characterized by an almost impenetrable layer of Marantaceae, in particular Haumania liebre-chtsiana,
and Zingiberaceae (e.g. Aframomum species and
Costus species). These species quickly fill gaps in
disturbed forest and form extensive shrubberies in
logged areas. The aggressively growing Marantaceae can prevent further forest development,
although several young secondary species were
encountered in these forests in the study area.
Brncic et al. (2007) speculated that Marantaceae
forests indicate sites of previous human occupancy.
This view is supported by Tovar et al. (2014) who
investigated the hypothesis that the four major
forest types in central African rainforest (mixed
forest, Marantaceae forest, Gilbertiodendron
monodominant forest and swamp forest) are the
consequence of different intensities of past human
activity, in particular fire. Tovar et al. (2014)
found that the areas currently covered by
Plant communities
Eleven plant communities were described and
mapped at a much finer scale than has been done
previously. Two gradients could be established: a
disturbed - relatively undisturbed gradient (running from bottom to top in Fig. 1); and a wet - dry
gradient running from lower right to upper left in
Fig. 1). The least degraded communities, and
consequently those with the highest conservation
value were the swamp forests (F9, F10 and F11)
and the okoumé forest (F3).
Three purely descriptive botanical studies
were conducted in the Kouilou Département in the
late 1980s and 1990s (Cusset 1989, Dowsett &
Dowsett-Lemaire 1991, Hecketsweiler & MokokoIkonga 1991). Although these studies were based
on physiognomy and plant lists of earlier
vegetation descriptions, the units described could
be related fairly confidently to the communities
distinguished by the classification approach performed in the current study.
Communities F1, F2, F4 and F6 all showed
clear signs of human influence. In the previously
mentioned studies these communities were described as anthropogenically degraded rainforest.
These communities were predominantly found
within the drier (Group 1) communities to the
right in ordination space or in the lower part of the
ordination space of Group 2 communities (Fig. 1).
Community F1 was a degraded forest community
containing most fallow sites as well as sites where
recent (2010 - 2012) mine exploration activities
had occurred. Although a fair number of mature
forest species still occurred, the community was
composed primarily of pioneer species such as
Chromolaena odorata, Harungana madagascariensis, Croton haumanianus, Trema guineense,
Macaranga barteri, Xylopia aethiopica, Pteridium
aquilinum, Ricinodendron heudelotii and Musanga
cecropioides. Several of these pioneer species are
known to be common in repeatedly humandisturbed areas along roads and around settlements (Fayolle et al. 2014a) and consequently F1 is
believed to be in a young successional stage.
The successional sequence on the fallows in
the study area seems to follow a similar route to
that described by Moutsamboté et al. (2000) who
recognized four stages. The first stage is domi-
15
Marantaceae forest underwent more frequent
burning events than the other forest types over the
past 2500 years.
Mango forests (F6) clearly have an anthropogenic origin. Most villages in the region are
associated with orchards constituted of many
edible fruit-bearing species, usually including
mango trees. Old and abandoned settlements (50 100 years) can often be identified by the remaining
mango trees, although most of the other planted
and edible alien species have disappeared and
been replaced by some native species (Dowsett &
Dowsett-Lemaire 1991). After abandonment these
patches may act as nuclei for forest succession.
Although the structure of F6 resembled a mature
forest, there was a lack of many mature indigenous forest tree species, which could hinder
progression towards a more advanced successional
stage.
The coastal evergreen rainforest, or ‘Atlantic
littoral forest’ (Letouzey 1968, 1985), characterised
by Aucoumea klaineana and Sacoglottis gabonensis (F3) is one of the main wet GuineoCongolian forest types (White 1983). In the
Congolese littoral Sacoglottis gabonensis occurs at
its southern distribution limit and is no longer codominant with Aucoumea klaineana as is the case
in the Cameroono-Gabonese littoral (Kimpouni et
al. 2013). Low numbers of Sacoglottis gabonensis
were noted in the study area, with the species only
being present in one of the subcommunities of F3.
Although F3 cannot be regarded as a primary
forest it represented the most mature successional
stage in the study area. Aucoumea klaineana is a
long-lived pioneer species and therefore often a
remnant within mature or even primary forest. In
the study area the okoumé forests showed healthy
regeneration. According to Pangou et al. (2006)
successful seedling establishment of Aucoumea
klaineana occurs only every 20-30 years when an
abundance of seeds are present and favourable
climatic conditions for seedling established are
present. Furthermore, in soils with high clay
content, high grass cover and an abundance of
hardwood species and Cecropiaceae, the successful
regeneration of Aucoumea klaineana is hindered.
The cirque forests (F5) occurred within close
proximity to the coastline and were comparable to
the Symphonia gorge forest described by Dowsett
& Dowsett-Lemaire (1991) and part of the
‘melange sublittoral’ described by Cusset (1989).
Natural disturbances are common within the
cirques and could explain the young successional
stage of this community. All the dominant species
in F5 (Anthocleista schweinfurthii, Symphonia
globulifera, Elaeis guineense and Chrysobalanus
icaco) are typical of disturbed or pioneer forest
sites. The species-poor community F8 formed a
narrow band around the cirque forests and is
probably best seen as an ecotone between the
cirque forest and the adjoining savanna vegetation.
The low-canopy, coastal thicket (F7) grew up to
a few meters immediately inland of the beach
zone. The community was comparable to the
Manilkara/Chrysobalanus coastal thickets previously described by various authors (Dowsett &
Dowsett-Lemaire 1991, Mbatchou 2004).
Communities F9, F10 and F11 were wetland
communities and were grouped towards the top
left in the ordination (Group 2, Fig. 1). Community
F9 occurred in small patches along the coast and
was a seasonally-flooded forest, whereas F10 was a
permanently-flooded forest associated with Cyperus
papyrus stands. The mangroves (F11) were found
near the mouth of the Kouilou River at the
interface between land and sea. The mangrove
species possess physiological traits enabling them
to overcome high salinity and frequent tidal
inundation. Both the permanently-flooded swamp
forest (F10) and mangroves (F11) showed little
signs of human impact and could be regarded as
mature forest communities.
It must be noted that the highly preferred
timber species, Terminalia superba, was not
recorded in any sampling sites. It was however
noted while travelling through the study area in
inaccessible places on the banks of the Kouilou
River or on islands of higher ground within the
swamp forest (F10) and the mangroves (F11). In
1991, Dowset & Dowsett-Lemaire reported it as a
conspicuous species. The species decline is most
likely a result of legal and illegal logging during
the past 20 years. The species is classified as a
pioneer species and usually regenerates well after
forest exploitation. Since no saplings were encountered in the current study the lack of regeneration
could possibly be due to a shortage of seeds
(Daïnou et al. 2011).
Although several studies have shown that
African forests are fairly resilient to climate and
past anthropogenic disturbances (Brncic et al.
2007, Gourlet-Fleury et al. 2013; Kassi et al. 2008;
Oslisley et al. 2013; Rudel 2013; Willis et al. 2004)
there is no historical analogue to mechanized
logging or vegetation destruction by mining activities. Destruction of the topsoil is expected to
have severe long-lasting effects and intervention
would be necessary for forest recovery. Active
16
rehabilitation of sites where topsoil had been
removed should explore patterns of forest survival
that have allowed these forests to recover quickly
in the past and attempt to recreate suitable
conditions.
plots in Central Africa range from 3.72 to 4.00
(Gonmadje et al. 2011). Overall, the values
reported within the current study area are within
the range reported for other tropical forests in
Central Africa.
Plant diversity
References
The current study could quantitatively confirm
the west to east gradient between the coast and
the Mayombe mountain range with respect to
plant diversity, which is linked to topography and
climate as suggested by Hecketsweiler & MokokoIkonga (1991), Dowsett & Dowsett-Lemaire (1991)
and Vande Weghe (2004). Forest plant diversity
was lowest at the coast and increased further
inland.
The relationship between plant diversity and
distance from human infrastructure produced an
unexpected result, with the highest diversity
closest to human infrastructure where the human
pressure and disturbance is at its highest. This
relationship could be interpreted in two ways: a.
humans are selecting the most diverse plant
communities for their activities or b. human
impacts are increasing plant diversity. Previous
reports of human impacts and disturbance on
plant diversity in African rain forests have been
contradictory. Von Gemerden et al. (2003) reported
a decrease in diversity after disturbance, whereas
Kassi et al. (2008) reported that shifting cultivation could increase diversity on condition that
the fields were small and embedded in a forest
matrix, cultivation lasted only a few years,
remnant trees were preserved and fire was
excluded. The present results support this
hypothesis.
Diversity parameters are difficult to compare if
plot sizes are not equal among studies, the values
provided in Table 2 can therefore only serve to
indicate broad trends. The Shannon-Wiener index
was 4.90 if data of all 156 plots were combined (9.8
ha) and ranged from a low of 1.23 to 4.10 for
individual communities. By comparison, ShannonWiener values of 3.34 to 4.14 have been reported
for forests with gregarious species (Lope forest,
Gabon; in Senterre & Lejoly 2001) and from 5.24 to
5.46 in mixed forests (Dja forest, Cameroon; in
Senterre & Lejoly 2001) to 5.79 in rainforest
(Nsork, Equatorial Guinea, in Senterre & Lejoly
2001). In Campo-Ma’an, Cameroon, ShannonWiener index values ranged from 0.12 for
mangroves (0.2 ha) to 5.33 (1.4 ha) for entire
communities with the total for the entire CampoMa’an (14.7 ha) being 5.54. Mean values for 1 ha
Abernethy, K. A., L. Coad, G. Taylor, M. E. Lee & F.
Maisels. 2013. Extent and ecological consequences
of hunting in Central African rainforests in the
twenty-first century. Philosophical Transactions of
the Royal Society B 368: 20120303.
Bayol, N. & R. Eba’a Atyi. 2009. The forests of Congo in
2008. pp: 97-109. In: C. de Wasseige, D. Devers, P. de
Archen, R. Eba’A Atyi, R. Nasi, & P. Mayaux (eds.).
The Forests of the Congo Basin - State of the Forests
2008. Publications office of the European Union,
Luxembourg.
Bhagwat, S., K. J. Willis, H. J. B. Birks & R. Whittaker.
2008.Agroforestry: a refuge for tropical biodiversity? Trends in Ecology & Evolution 23: 261-267.
Brncic, T. M., K. J. Willis, D. J. Harris & R. Washington. 2007. Culture or climate? The relative influence
of past processes on the composition of the lowland
Congo rainforest. Transactions of the Royal Society
of London B. 362: 229-242.
Brockington, D., J. Igoe & K. Schmidt-Soltau. K. 2006.
Conservation, human rights and poverty reduction.
Conservation Biology 20: 250-252.
Colwell, R. K. 2013. EstimateS: Statistical estimation of
species richness and shared species from samples.
Version 9. User's Guide and application published
at: http://purl.oclc.org/ estimates.
Cordeiro, N. J., N. D. Burgess, D. B. K. Dovie, B. Kaplin,
A. J. Plumptre & R. Marrs. 2007. Conservation in
areas of high population density in sub-saharan
Africa. Biological Conservation 134: 155-163.
Corlett, R. & R. Primack. 2011. Tropical Rain Forests:
an Ecological and Biogeographical Comparison.
Wiley-Blackwell, Oxford.
Cusset, G. 1989. La flore et la végétation du Mayombe
congolais. Etat des connaissances. pp. 103-136. J.
Séńchal, Matuka Kabala and F. Fournier (eds.) In:
Revue des connaissances sur le Mayombe. PNUD,
MAB, UNESCO, Paris.
Daïnou, K., A. Bauduin, N. Bourland, J-F. Gillet, F.
Fétéké & J-L. Doucet. 2011. Soil seed bank characteristics in Cameroonian rainforests and implications for post-logging forest recovery. Ecological
Engineering 37: 1499-1506.
Davidar, P., D. Mohandass & S. Lalitha Vijayan. 2007.
17
forest recovery from logging: a 24 year silvicultural
experiment from Central Africa. Philosophical
Transactions of the Royal Society B 368: 20120302.
Hammer, O. 2014. PAST. PALeonotological Statistics
Version 3.02. Reference manual. Natural History
Museum, University of Olso, Olso.
Hecketsweiler, P. & J. Mokoko-Ikonga. 1991. La Reserve
de Conkouati: Congo - Le secteur sud-est. British
Petroleum Exploration Report.
Hennekens, S. M. & J. H. J. Schaminee. 2001.
TURBOVEG, A comprehensive database management system for vegetation data. Journal of Vegetation Science 12: 589-591.
Ickowitz, A. 2006. Shifting cultivation and deforestation
in tropical Africa: critical reflections. Development
and Change 37: 599-626.
Jost, L. 2006. Entropy and diversity. Oikos 113: 363-375.
Kameri-Mbote, A. P. & P. Cullet. 1997. Law, colonialism
and environmental management in Africa. Review
of European Community & International Environmental Law 6: 23-31.
Kassi, N., K. Justin & G. Decocq. 2008. Successional
patterns of plant species and community diversity
in a semi-deciduous tropical forest under shifting
cultivation. Journal of Vegetation Science 19:
809-820.
Kent, M. 2012. Vegetation Description and Analysis: a
Practical Approach. Second edition. Wiley-Blackwell,
Oxford.
Kimpouni, V., J. Loumet & J. Mizingou. 2013. Dynamic
of Sacoglottis gabonensis - Aucoumea klaineana
couple in the Congolese coast forest (Brazzaville Congo). International Research Journal of Biological
Sciences 2: 25-29.
Laclau, J-P., P. Deleporte, J. Ranger & J-P.Bouillet.
2003. Nutrient dynamics throughout the rotation of
Eucalyptus clonal stands in Congo. Annals of
Botany 91: 879-892.
Laurance, W. F., B. M. Croes, L. Tchignoumba, S. A.
Lahm, A. Alonso, M. E. Lee & C. Ondzeano. 2006.
Impacts of roads and hunting on Central African
rainforest mammals. Conservation Biology 20:
1251-1261.
Laurance, W. F. et al. 2012. Averting biodiversity
collapse in tropical forest Protected areas. Nature
489: 290-294.
Lebamba, J., A. Ngomanda, A. Vincens, D. Jolly, C.
Favier, H. Elenga H. & I. Bentaleb. 2009. A
reconstruction of Atlantic Central African biomes
and forest succession stages derived from modern
pollen data and plant functional types. Climate of
the Past Discussions 5: 153-202.
Floristic inventory of woody palms in a tropical
montane (shola) forest in the Plani Hills of the
Western Ghats, India. Tropical Ecology 48: 15-25.
De Foresta, H. & G. Schwartz. 1991. Chromolaena
odorata and disturbance of natural succession after
shifting cultivation: An example from Mayombe,
Congo, Central Africa. In: R. Muniappan & P.
Ferrar (eds.) Proceedings of the Second International Workshop on Chromolaena odorata.
Biotrop Special Publication 44: 23-41.
de Wasseige C., P. de Marcken, N. Bayol, F. Hiol Hiol, P.
Mayaux, B. Desclée, R. Nasi, A. Billand, P.
Defourny & R. Eba’a Atyi. 2012. The Forests of the
Congo Basin - State of the Forests 2010. Publications
office of the European Union, Luxembourg.
Doumenge, C. (ed.). 1992. La Réserve de Conkouati:
Congo. Le secteur sud-ouest. UICN Publication,
Gland, Suisse.
Dowsett, P. & F. Dowsett-Lemaire (eds). 1991. Flora and
Fauna of the Kouilou Basin (Congo) and their
Exploitation. Tauraco Research Report, Conoco
Congo.
Duveiller G., P. Defourny, B. Desclée, & P. Mayaux.
2008. Deforestation in Central Africa: Estimates at
regional, national and landscape levels by advanced
processing of systematically-distributed Landsat
extracts. Remote Sensing of Environment 112:
1969-1981.
Ernst, C., P. Mayaux, A. Verhegghen, C. Bodart, M.
Christophe & P. Defourny. 2013. National forest
cover change in Congo Basin: Deforestation,
reforestation, degradation and regeneration for the
years 1990, 2000 and 2005. Global Change
Biology19: 1173-1187.
Fayolle, A., N. Picard, J-L.Doucet, M. Swaine, N. Bayol,
F. Bénédet & S. Gourlet-Fleury.2014a. A new
insight in the structure, composition and functioning of central African moist forests. Forest Ecology
& Management 329: 195-205.
Fayolle, A., M. Swaine, J-F. Bastin, N. Bourland, J. A.
Comiskey, G. Dauby, J-L.Doucet, J-F.Gillet, S.
Gourlet-Fleury, O. J. Hardy, B. Kirunda, F. N.
Kouame & A. J. Plumptre. 2014b. Patterns of tree
species composition across tropical African forests.
Journal of Biogeography doi:10.1111/jbi.12382.
Gonmadje, C. F., C. Doumenge, D. Mckey, G. P. M.
Tchouto, T. C. H. Sunderland, M. P. B. Balinga, &
B. Sonke. 2011. Tree diversity and conservation
value of Ngovayang’s lowland forests, Cameroon.
Biodiversity and Conservation 20: 2627-2648.
Gourlet-Fleury, S., F. Mortier, A. Fayolle, F. Baya, D.
Ouédraogo, F. Bénédet & N. Picard. 2013. Tropical
18
Letouzey, R. 1968. Étude phytogéographique du Cameroun. Lechevalier, Paris.
Letouzey, R. 1985. Notice de la carte phytogéographique
du Cameroun au 1/ 500,000. Institut de la carte
international de la végétation, Toulouse, France.
Malhi, Y., S. Adu-Bredu, R. A. Asare, S. L. Lewis & P.
Mayaux. 2013. African rainforests: past, present
and future. Philosophical Transactions of the Royal
Society B 368: 20120312.
Maniatis, D., J. Gaugris, D. Mollicone, J. Scriven, A.
Corblen, C. Ndikumagenge, A. Aquino, P. Crete &
M-L. Sanz-Sanchez. 2013. Financing and current
capacity for REDD+ readiness and monitoring,
measurement, reporting and verification in the
Congo Basin. Philosophical Transactions of the
Royal Society B 368: 20120310.
Mayaux, P., J-F. Pekel, B. Desclée, F. Donnay, A. Lupi,
F. Achard, M. Clerici, C. Bodart, A. Brink, R. Nasi &
A. Belward. 2013. State and evolution of the African
rainforests between 1990 and 2010. Philosophical
Transactions of the Royal Society B 368: 20120300.
Mbatchou, G. P. T. 2004. Plant Diversity in a Central
African rain forest: Implications for Biodiversity
Conservation in Cameroon. Tropenbos-Cameroon
Series (Netherlands), No 7.
McCune, B. & J. B. Grace. 2002. Analysis of Ecological
Communities. MjM software Design, Gleneden
Beach, Oregon.
Megevand, C., A. Mosnier, J. Hourtica, K. Sanders, N.
Doetinchem & C. Streck. 2013. Deforestation Trends
in the Congo Basin: Reconciling Economic Growth
and Forest Protection. World Bank, Washington DC.
Moutsamboté, J. M., D. N’zala, & J. C. Ngondo. 2000.
Development of forest regrowth after cassava
culture in Mayombe (Congo). Cahiers Agricultures 9:
141-144.
Mueller-Dombois, D. & H. Ellenberg, 1974. Aims and
Methods of Vegetation Ecology. Wiley, New York.
Murthy I. K., S. Bhat, V. Sathyanarayan, S. Patgar, M.
Beerappa, P. R. Bhat, D. M. Bhat, N. H.
Ravindranath, M. A. Khalid, M. Prashant, S. Iyer,
D. M. Bebber & R. Saxena. 2016. Vegetation
structure and composition of tropical evergreen and
deciduous forests in Uttara Kannada District,
Western Ghats under different disturbance regimes.
Tropical Ecology 57: 77-88.
Newton, A. C. 2007. Forest Ecology and Conservation, A
handbook of Techniques. Oxford University Press,
Oxford, UK.
Ngobo, M., M. McDonald & S. Weise. 2004. Impacts of
type of fallow and invasion by Chromolaena odorata
on weed communities in crop fields in Cameroon.
Ecology and Society 9: 1. [online] URL: http://www.
ecologyandsociety.org/vol9/iss2/art1
Nogherotto, R., E. Coppola, F. Giorgi, F. & L. Mariotti.
2013. Impact of Congo Basin deforestation on the
African monsoon. Atmospheric Science Letters 14:
45-51.
Oslisley, R., L. White, I. Bentaleb, C. Favier, M.
Fontugne, J-F. Gillet & D. Sebag. 2013. Climatic and
cultural changes in the west Congo Basin forests
over the past 5000 years. Philosophical Transactions of the Royal Society B 368: 20120304.
Pangou, S. V., G. Lechon, T. Bouki & A. Mountanda.
2006. Characteristic of natural regeneration of
Aucoumea klaineana Pierre in Mayombe rain forest,
southern Congo. African Journal of Ecology 45:
156-164.
Parmentier, I., R. J. Harrigan, W. Buermann, E. T. A.
Mitchard, S. Saatchi, Y. Malhi, F. Bongers, W. D.
Hawthorne, M. E. Leal, S. L. Lewis, L. Nusbaumer,
D. Sheil, M. S. M. Sosef, K. Affum-Baffoe, A.
Bakayoko, G. B. Chuyong, C. Chatelain, J. A.
Comiskey, G. Dauby, J-L. Doucet, S. Fauset, L.
Gautier, J-L. Gillet, D. Kenfack, F. N. Kouame, E.
K. Kouassi, L. A. Kouka, M. P. E. Parren, K. S-H.
Peh, J. M. Reitsma, B. Senterre, B. Sonke, T. C. H.
Sunderland, M. D. Swaine, M. G. P. Tchouto, D.
Thomas, J. L. C. H. van Valkenburg & O.J. Hardy.
2013. Predicting alpha diversity of African rain
forests: models based on climate and satellitederived data do not perform better than a purely
spatial model. Journal of Biogeography 38:
1164-1176.
Perrings, C. & J. Lovett. 1999. Policies for biodiversity
conservation: The case of Sub-Saharan Africa.
International Affairs 75: 281-305.
Petrozzi, F., G. Amori, D. Franco, P. Gaubert, N. Pacini,
E. A. Eniang, G. C. Akani, E. Politano, L. Luiselli.
2016. Ecology of the bushmeat trade in west and
central Africa. Tropical Ecology 57: 545-557.
Podani, J. 2001. SYN-TAX Computer Programs for Data
Analysis in Ecology and Systematics. Scientia publishing, Budapest.
Rudel, T. K. 2013. The national determinants of deforestation in sub-Saharan Africa. Philosophical
Transactions of the Royal Society B 368: 20120405.
Senterre, B. & J. Lejoly. 2001. Trees diversity in the
Nsork rain forest (Rio Muni, Equatorial Guinea).
Acta Botanica Gallica: Botany Letters 148: 227-235.
Singh, S. P. & C. M. Sharma. 2009. Tropical ecology - an
overview. Tropical Ecology 50: 7-21.
Tichy, L. 2002. JUICE, software for vegetation classification. Journal of Vegetation Science.13: 451-453.
19
Tovar, C., E. Breman, T. Brncic, D. J. Harris, R. Bailey
& K. J. Willis. 2014. Influence of 1100 years of
burning on the central African rainforest. Ecography 37: 1139-1148.
Van der Maarel, E. 2007. Transformation of coverabundance values for appropriate numerical treatment - alternatives to the proposals by Podani.
Journal of Vegetation Science 18: 767-770.
Vande Weghe, J-P. 2004. Forests of Central Africa.
Nature and Man. Lanoo Publishers, Tielt, Belgique.
Von Gemerden, B. S., H. Olff, M. P. E. Parren & F.
Bongers. 2003. The pristine rain forest? Remnants
of historical human impacts on current tree species
composition and diversity. Journal of Biogeography
30: 1381-1390.
WCS. 2010. Plan d’aménagement du Parc National de
Conkouati-Douli / Project d’appui à la gestion du
Parc National de Conkouati-Douli. Draft document
obtained from the Wildlife Conservation Society.
Werger, M. J. A. 1974. On concepts and techniques
applied in the Ziirich-Montpellier method of vegetation survey. Bothalia 11: 309-323.
White, F. 1979. The Guineo-Congolian Region and its
relationships to other phytochoria. Bulletin du
Jardin Botanique National Belge 49: 11-55.
White, F. 1983. The Vegetation of Africa. UNESCO/
AETFAT/UNSO, Paris.
Willis, K. J., L. Gillson, & T. M. Brncic. 2004. How
‘virgin’ are virgin rainforest? Science 304: 402-403.
Willis, K. J., K. D. Bennett, S. L. Burrough, M. MaciasFauria & C. Tovar. 2013. Determining the response
of African biota to climate change: using the past to
model the future. Philosophical Transactions of the
Royal Society B 368: 20120491.
20